Drosophila tissue and organ: Mushroom Bodies

The Interactive Fly

Drosophila Mushroom Bodies
Sites for formation and retrieval of memories


Mushroom Body

For a review of mushroom body structure and function see Griffith, L. C. (2014). A big picture of a small brain. Elife 3: e05580. PubMed ID: 25537193

Featured Articles
Mamo decodes hierarchical temporal gradients into terminal neuronal fate

Liu, L. Y., Long, X., Yang, C. P., Miyares, R. L., Sugino, K., Singer, R. H. and Lee, T. (2019). Elife 8. PubMed ID: 31545163

Temporal patterning is a seminal method of expanding neuronal diversity. This study has unravel a mechanism decoding neural stem cell temporal gene expression and transforming it into discrete neuronal fates. This mechanism is characterized by hierarchical gene expression. First, Drosophila neuroblasts express opposing temporal gradients of RNA-binding proteins, Imp and Syp. These proteins promote or inhibit chinmo translation, yielding a descending neuronal gradient. Together, first and second-layer temporal factors define a temporal expression window of BTB-zinc finger nuclear protein, Mamo. The precise temporal induction of Mamo is achieved via both transcriptional and post-transcriptional regulation. Finally, Mamo is essential for the temporally defined, terminal identity of alpha'/beta' mushroom body neurons and identity maintenance. This study describes a straightforward paradigm of temporal fate specification where diverse neuronal fates are defined via integrating multiple layers of gene regulation. The neurodevelopmental roles of orthologous/related mammalian genes suggest a fundamental conservation of this mechanism in brain development (Liu, 2019).

The brain is a complicated organ which not only requires specific connections between neurons to form circuits, but also many neuronal types with variations in morphology, neurotransmitters and receptors. While mechanisms controlling neuronal diversity have not been globally examined, studying neural stem cells in the mouse and fruit fly have given insight into key aspects of neuronal specification. For example, in the mouse neocortex, radial glial progenitors (RGP) are multipotent-they produce a variety of neuron types organized sequentially into six layers, and then produce glia. In vivo lineage analysis demonstrated that after a stage of symmetric cell division, an individual neurogenic RGP produces an average of 8-9 progeny (range of 3-16) that can span all cortical layers. In Drosophila, clonal analysis has demonstrated a vast range of stem cell-specific lineage programs. On one extreme, lineage tracing of a single antennal lobe (AL) stem cell revealed a remarkable series of 40 morphologically distinct neuronal types generated sequentially. In light of these observations, a fundamental goal is to understand how distinct neuronal types correctly differentiate from a single progenitor. Despite a fundamental role for temporal patterning to create diverse neuronal lineages, understanding of neuronal temporal patterning is still limited. While scientists have discovered key temporal factors expressed in neural progenitors, much less is understood about how these signals are interpreted, that is what factors lie downstream of the specification signals to determine distinct neuronal temporal fates (Liu, 2019).

Despite its relatively small brain, Drosophila is leading the charge on studies of neuronal temporal fate specification. Many temporal transcription factors originally discovered in the fly have since been confirmed to have conserved roles in mouse retinal and cortical development. Moreover, temporal expression of an RNA binding protein, IGF-II mRNA-binding protein (Imp), that guides temporal patterning in the postembryonic fly brain is also implicated in mouse brain development. Drosophila brain development is an excellent model for studying neurogenesis; the neural stem cells, called neuroblasts (NB), are fixed in number, their modes of division are well characterized, and each NB produces a distinctive series of neurons which change fate based on birth order. Finally, the fruit fly is a genetically tractable system making it ideal for studying gene networks involved in cell fate decisions (Liu, 2019).

In Drosophila, cycling NBs express age-dependent genes that provide the serially derived newborn neurons with different temporal factors. In the embryonic ventral nerve cord and the optic lobe, the NBs express a rapidly changing series of four to six temporal transcription factors (tTF), some of which are directly inherited by the daughter neurons. Each tTF directly acts to specify a small number (two to four) of neuronal progeny. The neuronal progeny produced from one tTF window to the next can be quite different. The tTF series are intrinsically controlled, which ensures reliable production of all neuron types, but lacks the ability to adapt to complicated or changing conditions (Liu, 2019).

A separate mechanism is therefore required for adult brain development-both to produce very long series of related neuronal types and to coordinate with organism development. This can be accomplished utilizing protein gradients and hierarchical gene regulation, such as the mechanism used to pattern the fly's anterior/posterior (A/P) axis. In Drosophila A/P patterning, the embryo is progressively partitioned into smaller and smaller domains through layered gene regulation. This is initiated by asymmetric localization maternal mRNAs, bicoid (anterior) and nanos (posterior). The resulting opposing proteins gradients then act on maternal mRNA translation, and in the case of Bicoid, zygotic transcription. The embryo then progresses through expression of maternal morphogen gradients, then zygotic expression of gap genes to determine broad embryo regions, followed by progressive segmentation by the pair-rule and segment polarity genes, and finally specification by the homeotic selector genes (Liu, 2019).

Notably, in postembryonic brain development, two proteins in opposing temporal gradients expressed in NBs have been discovered. These proteins are Imp and Syncrip (Syp) RNA-binding proteins. Imp and Syp control neuronal temporal fate specification as well as the timing of NB termination (decommissioning). Imp promotes and Syp inhibits translation of the BTB-zinc finger nuclear protein, chinmo (chronologically inappropriate morphogenesis), so that protein levels in newborn neurons descend over time. The level of Chinmo correlates with the specification of multiple neuronal temporal fates. Discovering downstream layers in the Imp/Syp/Chinmo hierarchy is essential to fully comprehend the intricacies of temporal patterning in brain development (Liu, 2019).

Temporal regulation in the fly brain is easily studied in the relatively simple mushroom body (MB) neuronal lineages which are comprised of only three major cell types. These neuronal types are born in sequential order: beginning with γ neurons, followed by α'/β' neurons, and finally α/β neurons. Imp and Syp are expressed in relatively shallow, opposing temporal gradients in the MB NBs. Modulation of Imp or Syp expression results in shifts in the neuronal temporal fate. Imp and Syp post-transcriptionally control Chinmo so that it is expressed in a gradient in the first two temporal windows. γ neurons are produced in a high Chinmo window, α'/β' neurons are produced in a low Chinmo window, and α/β neurons are produced in a window absent of Chinmo expression. Moreover, altering Chinmo levels can shift the temporal fate of MB neurons accordingly, strongly implicating dose-dependent actions, similar to that of a morphogen. Despite its importance in temporal patterning, the mechanisms underlying the dosage-dependent effects of Chinmo on neuronal temporal identity is unknown (Liu, 2019).

This study describes Mamo (maternal gene required for meiosis), a BTB zinc finger transcription factor critical for temporal specification of α'/β' neurons. Mamo is expressed in a low Chinmo temporal window and Mamo expression can be inhibited both by high Chinmo levels and loss of chinmo. Additionally, Mamo is post-transcriptionally regulated by the Syp RNA binding protein. This layered regulation, which is utilized in both MB and AL lineages results in a discrete window of Mamo expression in young, post-mitotic neurons. In the MB lineages, this window corresponds to the middle window of neurogenesis and it was establish that Mamo codes for middle temporal fate(s); α'/β' neuronal characteristics are lost when Mamo levels are reduced and ectopic Mamo drives an increase in α'/β' neuron production. The temporal fate determination paradigm described in this study utilizes multiple levels of gene regulation. Temporal fate specification begins in the stem cell and proceeds in a hierarchical manner in successive stages where top and second-tier factors work together to specify neuronal temporal fate. These data suggest that Mamo deciphers the upstream temporal specification code and acts as a terminal selector to determine neuronal fate (Liu, 2019).

Chinmo levels in newborn neurons correlate with adult neuron identity. Based on smFISH, mamo transcription is initiated in newborn MB neurons around 72 hr ALH, which corresponds to weak Chinmo expression. Moreover, Mamo is only expressed when Chinmo levels are low, as Mamo is not expressed after either eliminating or overexpressing Chinmo. Together these data indicate that low Chinmo levels activate mamo transcription in young/maturing neurons (Liu, 2019).

Transcription initiation is not the only requirement for Mamo protein expression; Syp is also required as discussed below. This could explain why no Mamo expression turn on is seen in γ neurons, even as they age and Chinmo levels decrease, becoming quite low around wandering larval stage. γ neurons begin to express Mamo later, around pupation, despite lacking Syp. It has not yet been tested whether weak Chinmo levels are required for later Mamo expression in γ neurons. It is therefore possible that Mamo expression is controlled at this stage by an additional factor(s) (Liu, 2019).

ChIP-chip performed in embryos found five Chinmo binding sites within the mamo gene, consistent with direct activation of mamo transcription. However, the nature of Chinmo's concentration dependent actions is still unclear. Some morphogens such as Bicoid bind different targets at increasing concentrations based on the affinity of binding to different sites as well as the chromatin accessibility of the binding sites. This may also be the case with Chinmo, but would not easily explain why Mamo expression is inhibited at higher Chinmo concentrations. The gap gene Krüppel, on the other hand, has concentration dependent activities at the same binding site. Krüppel acts as an activator at lower concentrations and as a repressor at high concentrations. Krüppels C-terminus has the ability to activate genes and is also the location for dimerization. Upon dimerization, the C-terminus can no longer activate genes and Krüppel transforms from an activator to a repressor. The current data suggests that low concentrations of Chinmo activate mamo. However, in the testis, Chinmo is suspected to function as a transcriptional repressor. It is feasible that Chinmo, like Krüppel, could switch from an activator to a repressor. The protein concentration would affect whether Chinmo is a monomer (in the presence of other BTB proteins, a heterodimer) or a homodimer, and thus potentially which cofactors are recruited (Liu, 2019).

The ascending Syp RNA binding protein temporal gradient regulates Mamo expression both indirectly via its inhibition of Chinmo and also presumably directly, interacting with the mamo transcript and promoting its expression. The bi-modal, transcriptional (Chinmo) and post-transcriptional (Syp), regulation of the Mamo terminal selector is extremely advantageous. Given the finding that Mamo expression is positively autoregulated and that Mamo continues to be expressed into adult neurons, it is particularly important to control the timing of Mamo's onset. The additional layer of post-transcriptional regulation adds an extra safeguard, helping to guarantee that neuronal temporal patterning is a robust system. Indeed, as brain development needs to adapt to environmental conditions such as nutrient deprivation, it is crucial to ensure that there is no loss of neuronal diversity (Liu, 2019).

Syp is a homolog of mammalian SYNCRIP (synaptotagmin-binding cytoplasmic RNA-interacting protein) also known as hnRNP-Q. SYNCRIP is involved with multiple facets of mRNA regulation including mRNA splicing and maturation, mRNA localization and stabilization as well as inhibiting mRNA translation and miRNA-mediated repression via competition with Poly(A) binding proteins. The Drosophila ortholog seems to have corresponding functions. Drosophila Syp was isolated from the spliceosome B complex, indicating a conserved role in mRNA splicing. Syp has likewise been found to operate in mRNA localization and stabilization. Furthermore, it has clear roles in altering protein expression of its mRNA targets, both positively and negatively. The bidirectional influence on protein expression likely reflects different Syp modalities (Liu, 2019).

This study shows that Syp is required for Mamo protein expression in the MB and AL neuronal lineages. To determine the nature of this regulation, single molecule fluorescence in situ hybridization (smFISH) was performed. In the absence of Syp, mamo transcription was initiated prematurely in response to weak Chinmo levels, yet mature transcripts failed to accumulate. This leads to the belief that Syp directly binds mamo mRNA and aids in its splicing, maturation and/or stabilization. This is consistent with the finding that overexpressing a mamo cDNA (lacking 5' UTR, 3' UTR and introns) was able to promote cell fate changes despite repression of Syp (Liu, 2019).

Mamo is required to produce the α'/β' neurons in the middle temporal window of the MB lineages. Trio positive α'/β' neurons are clearly absent after RNAi depletion of Mamo during development. Cell production does not appear to be altered, as mamo-RNAi expressing MBs are a normal size. This begs the question of which, if any terminal fate the middle-born neurons adopt in the absence of Mamo. The limited markers for each MB cell type makes it difficult to determine whether the middle-born neurons undergo fate transformation or simply lack terminal fate. The presence of a Fas-II negative lobe hints that some middle-born neurons may not carry temporal fate information, but phenotypic analysis is complicated by defects in γ neuron maturation/remodeling. Removing the γ neurons with chinmo-RNAi eliminates this complication, but it is still unclear whether, without Mamo, the neurons are transformed to the α/β fate. The Fas-II positive, α/β lobe appears enlarged, but it is difficult to tell whether all s are Fas-II positive or whether Fas-II negative axons are comingled with α/β axons. Without a cell type-specific, cell body marker for α/β neurons, it is ambiguous whether the middle-born cells are transformed to α/β or whether they simply lack α'/β' temporal fate. A transformation to α/β fate would suggest that either α/β is the default fate of MB neurons (requiring no additional terminal selector) or that Mamo expression inhibits α/β specific factors (Liu, 2019).

Mamo's role in promoting α'/β' fate is further supported by Mamo overexpression phenotypes. Overexpression of Mamo in the MB is able to transform α/β and γ neurons to α'/β' neurons. In an otherwise wildtype scenario, overexpression of mamo did not transform every cell to α'/β' fate. Instead the α'/β' lobe was expanded and the other lobe seemed to be an amalgam of α and γ like lobes. This could be due to incomplete penetrance/low expression levels of the mamo transgene or it is possible that the α/β and γ cells retain their own terminal selector driven, cell-type specific gene expression, thus complicating the fate of the differentiated neuron. Mamo overexpression does not alter the specification factors Imp, Syp or Chinmo and presumably there are terminal selector genes expressed downstream of high Chinmo and possibly in Chinmo-absent cells. This seems a likely possibility when overexpressing Mamo in γ neurons. With Syp-RNAi, NBs are 'forever young' and divide into adulthood, persistently producing 'early-born' γ neurons. Interestingly when combining Syp-RNAi with the Mamo transgene, the newborn cells begin to take on a γ-like fate (expressing Abrupt) before a majority transform into an α'/β'-specific, strong Trio expression pattern and adopt α'/β'-like axon morphology. This suggests that Mamo functions downstream of the temporal fate specification genes, but is capable of overriding downstream signals in α/β and γ neurons to promote α'/β' terminal fate (Liu, 2019).

What this study describes about the BTB-ZF transcription factor, Mamo's role in α'/β' cell fate easily fits into the definition of a terminal selector gene, coined by Oliver Hobert. Terminal selector genes are a category of 'master regulatory' transcription factors that control the specific terminal identity features of individual neuronal types. Key aspects of terminal selector genes are that they are expressed post-mitotically in neurons as they mature and they are continuously expressed (often via autoregulatory mechanisms) to maintain the terminal differentiated state of the neuron. Correspondingly, mamo transcription is initiated in newborn, post-mitotic neurons and Mamo protein expression is visible beginning in young/maturing neurons. After transcription initiation, Mamo positively regulates its own expression and continues to be expressed in α'/β' neurons into adulthood. The other quintessential feature of terminal selector genes is that they regulate a battery of terminal differentiation genes, so that removing a terminal selector gene results in a loss of the specific identity features of a neuron type and misexpression can drive those features in other neurons. Indeed, removing Mamo with RNAi results in the loss of α'/β' identity, both developmentally and into adulthood. Further, overexpressing Mamo in either α/β or γ MB neurons results in shift to α'/β' fate. Individual terminal selectors do not often function alone, but in combination with other terminal selectors. Therefore, there are likely terminal selectors downstream of the MB NB-specific genes that contribute to each of the MB neuron types. In this way, the lineage-specific and temporal patterning programs can combine to define individual neuron types. This feature enables the reutilization of terminal selector genes to create disparate neuron types when used in distinct combinations (Hobert, 2016). This further suggests that temporally expressed Mamo serves as a temporally defined terminal selector gene in other lineages, such as the AL lineages that are describe in this study (Liu, 2019).

Altering Chinmo levels via upstream RNA-binding proteins or miRNAs, or by reducing Chinmo with RNAi all result in shifts in the ratio of neurons with different neuronal temporal fates. This evidence suggests a mechanism where Chinmo acts in newborn neurons to promote temporal fate specification. A recent publication suggested that Chinmo affects temporal fate via a neuronal remodeling mechanism by controlling Ecdysone signaling . Marchetti demonstrates that Chinmo is required for EcR-B1 expression; however it remains unclear whether Chinmo directly affects EcR-B1 expression or if the Chinmo-dependent EcR-B1 expression is the sole mechanism for γ neuron temporal fate specification. Moreover, neuronal temporal fate is not accurately determined by neuronal morphology alone, particularly when ecdysone signaling has known effects on MB cell morphology and fate. Ecdysone receptor signaling is highly pleiotropic, including ligand-independent functions making dominant-negative and overexpression studies difficult to interpret. Therefore, further investigation is needed to clarify the roles of Ecdysone receptor signaling in MB neuronal temporal fate and remodeling. This will be addressed in a follow-up paper. This current manuscript strongly promotes the idea that Chinmo functions in newborn neurons to promote temporal fate as weak Chinmo expression directly precedes Mamo transcription and Mamo is essential for specification and maintenance of α'/β' fate (Liu, 2019).

This study describes a multilayered hierarchical system to define distinct neuronal temporal fate that culminates in the expression of a terminal selector gene. Analogous mechanisms likely underlie temporal patterning in mammalian brains. However, whether orthologous genes play equivalent roles in mammalian temporal patterning has not been fully investigated. The Imp and Syp RNA-binding proteins are evolutionarily conserved. Both homologs are highly expressed in the developing mouse brain and play vital roles in neural development and/or neuronal morphology. The opposing functions of Imp and Syp also appear to be conserved, as the murine orthologs IMP1 and SYNCRIP bind the identical RNA to either promote or repress axon growth, respectively. Moreover, IMP1 expression in fetal mouse neural stem cells plays important roles in stem cell maintenance and proper temporal progression of neurogenesis. It would likewise be very interesting to explore SYNCRIP in the context of temporal patterning (Liu, 2019).

While Chinmo and Mamo have no clear mammalian orthologs, they are both BTB-ZF (broad-complex, tram-track and bric-à-brac - zinc finger) transcription factors. The BTB domain is a protein interaction domain that can form homo or heterodimers and also binds transcriptional regulators such as repressors, activators and chromatin remodelers. The C2H2 (Krüppel-like) zinc fingers bind DNA-providing target specificity. BTB-ZF proteins have been found to be critical regulators of developmental processes, including neural development. Indeed, the BTB-zinc finger protein, Zbtb20, appears to be essential for early-to-late neuronal identity in the mouse cortex. Zbtb20 is temporally expressed in cortical progenitors and knockout results in cortical layering defects, as the inside-out layering of the cortex follows neuronal birth order. While mutations of other brain-expressed BTB-ZF proteins also show cortical layering phenotypes, potential roles in temporal patterning have not been explored (Liu, 2019).

This study illustrates a fate specification process in which a layered series of temporal protein gradients guide the expression of terminal selector genes. The first-tier temporal gradients are expressed in neural stem cells, followed by a restricted expression window in newborn neurons to finally induce a terminal selector gene in a subset of neurons as they mature. This time-based subdivision of neuronal fate can likely be further partitioned, finally resulting in sequentially born neurons with distinct cell fates. This study demonstrates that Mamo, a BTB-ZF transcription factor, delineates α'/β' neurons, the middle temporal window of the MB lineages. Corresponding data in the AL lineages suggest that Mamo may serve as a temporally defined, terminal selector gene in a variety of lineages in the Drosophila brain. Mamo expression is regulated transcriptionally by the descending Chinmo BTB-ZF transcription factor gradient and post-transcriptionally by the Syp RNA binding protein. This multi-tiered, bimodal regulation ensures that only the progeny in a precise temporal window (those with both weak Chinmo and significant Syp levels) can effectively activate the terminal selector gene, mamo. This discovery attests to the power of gradients in creating diverse cells from a single progenitor. Utilizing layers of temporal gradients to define discrete temporal windows mirrors how in early embryos the spatial gradients of RNA-binding proteins and transcription factors specify the fly's A/P axis. This paradigm provides considerable complexity of gene network regulation, leading to abundant neural cell diversity (Liu, 2019).

Random convergence of olfactory inputs in the Drosophila mushroom body

Caron, S. J., Ruta, V., Abbott, L. F. and Axel, R. (2013). Nature 497(7447): 113-117. PubMed ID: 23615618

The mushroom body in the fruitfly Drosophila melanogaster is an associative brain centre that translates odour representations into learned behavioural responses. Kenyon cells, the intrinsic neurons of the mushroom body, integrate input from olfactory glomeruli to encode odours as sparse distributed patterns of neural activity. This study has developed anatomic tracing techniques to identify the glomerular origin of the inputs that converge onto 200 individual Kenyon cells. Each Kenyon cell integrates input from a different and apparently random combination of glomeruli. The glomerular inputs to individual Kenyon cells show no discernible organization with respect to their odour tuning, anatomic features or developmental origins. Moreover, different classes of Kenyon cells do not seem to preferentially integrate inputs from specific combinations of glomeruli. This organization of glomerular connections to the mushroom body could allow the fly to contextualize novel sensory experiences, a feature consistent with the role of this brain centre in mediating learned olfactory associations and behaviours (Caron, 2013).

Olfactory perception in the fly is initiated by the binding of an odorant to an ensemble of olfactory sensory neurons (OSNs) in the antennae, resulting in the activation of a unique and topographically fixed combination of glomeruli in the antennal lobe (AL). The discrimination of odours therefore requires the integration of information from multiple glomeruli in higher olfactory centres. AL projection neurons (PNs) extend dendrites into a single glomerulus and project axons that bifurcate to innervate two distinct brain regions, the lateral horn and the MB. The invariant circuitry of the lateral horn is thought to mediate innate behaviours, whereas the MB translates olfactory sensory information into learned behavioural responses. PN axons that innervate the MB terminate in large boutons. A given KC extends a small number of dendritic “claws”, with each claw receiving information from only one PN bouton. A single bouton connects to multiple KC claws to form a discrete anatomic structure, the microglomerulus. Each KC projects an axon to one of the three different classes of MB lobes, α/β, α’/β’, or γ, where it synapses upon a relatively small number of extrinsic output neurons (Caron, 2013).

Electrophysiological and optical imaging studies show that odorants activate sparse subpopulations of KCs distributed across the MB without spatial preference. Individual KCs could be connected to preferential combinations of glomeruli that are co-ordinately activated by behaviourally relevant odours. Alternatively, KCs may not receive structured input; rather the glomerular inputs may be random, a feature that maximizes the diversity of odour representations in the MB. This study has exploited the specialized structure of the PN-KC synapse to characterize the glomerular origin of the PNs that converge onto individual KCs (Caron, 2013).

Photoactivatable green fluorescent protein (PA-GFP) was expressed in all neurons of the fly and a single KC was photolabelled. It was observed that individual photolabelled KCs elaborate between 2 and 11 dendritic claws restricted to the main olfactory calyx. The axonal projections of a labelled KC can be traced into either the α/β, α’β’, or γ lobes of the MB. Texas red dextran was then electroporated into the centre of a single KC claw, filling the PN bouton innervating that claw. Retrograde transfer of the dye labels a single PN and its associated AL glomerulus, providing further evidence that an individual KC claw receives input from only a single glomerulus (Caron, 2013).

It was verified that this tracing method identifies functional connections between PNs and KCs. Functional imaging was performed on flies that express the calcium indicator GCaMP3 in most KCs to identify the claws activated by the stimulation of a single glomerulus. Electroporation of dye into an activated microglomerulus labels a single PN that innervates the stimulated glomerulus. Thus, the electroporation of dye into a KC claw allows faithful identification of the PN to which it is functionally connected (Caron, 2013).

The strategy of photolabelling a single KC and sequential electroporation of dye into each of its claws was used to define glomerular inputs to an individual KC. In initial experiments, PA-GFP was expressed in all neurons and 100 randomly chosen KCs were photolabelled in 100 different flies. Among the 100 photolabelled KCs, 84 α/β KCs, 14 α’/β’ KCs, but only 2 γ KCs were identified. Each MB contains about 1000 α/β KCs, 370 α’β’ KCs, and 670 γ KCs. γKCs are underrepresented in this initial data set. This is likely to result from the spatial segregation of their cell bodies, which renders γ KCs less accessible to photolabelling. Most α/βKCs, but not the α/β and α’β’ KCs, express Fruitless (Fru). An additional 100 γ KCs were targeted for photoactivation in flies expressing PA-GFP under the control of the Fru promoter (Caron, 2013).

Texas red dextran was sequentially electroporated into different claws of a photolabelled KC. It is technically difficult to fill all the claws of a KC and on average 3 glomerular inputs were identified per KC. In fewer than 5% of the samples, the number of labelled PNs differed from the number of claws filled, reflecting either unsuccessful or imprecise electroporation. Samples with more labelled PNs than expected were discarded. The low frequency of unsuccessful claw fills indicates that claws extending from a given KC were filled with equal efficiency independent of size. Thus the size of a claw was not a selection criterion in these experiments (Caron, 2013).

A total of 683 inputs that synapse on 200 KCs were identified. 654 of these inputs connect to PNs innervating 49 of the 51 AL glomeruli. PNs innervating the DA3 and VL1 are absent from the data set but boutons were identified from these PNs in the MB calyx. 29 of the claws receive input from brain regions other than the AL. Interestingly, 11 of these claws are innervated by PNs that derive from pseudoglomeruli in the proximal antennal protocerebrum, a thermosensing centre in the fly brain that receives input from distinct heat- and cold-sensing neurons in the antennae. The remaining 18 PNs innervated different uncharacterized regions of the brain (Caron, 2013).

It was observe that the distribution of the glomerular inputs to KCs is not uniform. Inputs from the DA1 and DC3 glomeruli are most frequent, with each accounting for 5.1% of the total connections. The non-uniform distribution reflects the fact that the size and number of calycal boutons formed by PNs varies across glomeruli. For instance, the PNs associated with the DA1 and DC3 glomeruli form more numerous boutons in comparison with the PNs of less frequently represented glomeruli. It was also observed that there is a small but significant difference between the inputs to the α/β and γ KCs (p < 0.001). All subsequent statistical analyses were therefore performed separately on both the α/β and γ data sets, but failed to reveal any significant difference between the two data sets. Therefore, only the results obtained from the full data set are shown (Caron, 2013).

Statistical analyses of the 665 connections allowed searching for structure among the connections between glomeruli and KCs. First, it wae determined whether the KCs receiving input from a given glomerulus have a higher probability of receiving additional input from that same glomerulus. Of the 200 KCs in the data set, only 11 receive two inputs from the same glomerulus, and none receive three or more such inputs. It was determined whether the frequency of convergent input from a single glomerulus is significantly above or below chance expectations by randomly shuffling the connections in the data set between the different KCs, while maintaining the number of connections each of them receives. This shuffling maintains the frequency of glomerular connections observed in the experimental data, but eliminates any potential, non-random patterns of inputs onto individual KCs. This shuffling is used in all subsequent statistical analyses. The frequency of multiple connections from the same glomerulus in the observed and shuffled data sets is not significantly different. Thus, no KCs were observed that receive preferential inputs from a single glomerulus. Rather, individual KCs integrate information from multiple different glomeruli (Caron, 2013).

It was next determined whether KCs are connected to any preferential pair, trio, or quartet of glomeruli. Of the 1378 (53*52/2) different pairs of glomeruli that could converge onto an individual KC, 508 distinct pairs appear in the data set. 310 of these pairs connect to only one of the 200 KCs analysed, whereas certain pairs of glomeruli connect to multiple KCs. There are combinations of glomerular trios that connect with two different KCs, and one case in which two KCs receive inputs from the same quartet of glomeruli. However, the observed frequency with which the different pairs, trios and quartets converge onto different KCs is consistent with expectations from the shuffled data set. Thus, the identity of a glomerulus connected to a KC provides no predictive information as to the identity of the remaining glomerular inputs onto that neuron (Caron, 2013).

Glomeruli can be grouped based on biological properties shared by their associated OSNs (sensilla type, odour specificity) and PNs (developmental origin and topography of their axonal projections). KCs might receive preferential input from one or another of these glomerular categories. For example, the OSNs innervating the AL are derived from three sensillar types (basiconic, coeloconic, and trichoid sensilla) that project to three classes of glomeruli tuned to different odour categories. If individual KCs were tuned to a particular class of odours, they might preferentially integrate inputs from one type of sensillum. Statistical analyses, however, reveal that KCs that receive an input from one sensillar type are no more or less likely to receive additional inputs from this or any other type of sensillum than is predicted by chance. Sensillar type, however, provides only a coarse correlate of odour tuning. Therefore, glomeruli were first grouped based on the similarity of their odour response profiles and again no structure was observed in the inputs to a KC that correlated with odour tuning (Caron, 2013).

Glomeruli were classified on the basis of the properties of their PNs. PN axons from different glomeruli project to broad but stereotyped domains in the lateral horn and calyx of the MB. Input to an individual KC could be shaped by the topography of PN projections. Analysis of the distribution of inputs to a given KC, however, fails to reveal any preferential PN connectivity that reflects the organization of their projection in either the MB calyx or lateral horn. KCs do not preferentially integrate information from glomeruli innervated by PNs sharing a developmental origin. In addition, KCs do not select their input based on topographical constraints as suggested by a previous study. Finally, three glomeruli are innervated by Fru-expressing OSNs and PNs. Preferential pairing of inputs from Fru+ PNs onto individual KCs were not observed. Moreover, although most γ KCs express Fru, there is no preferential input from Fru+ glomeruli to γ KCs (Caron, 2013).

Next, an unbiased search was performed for structure by examining correlations within the connectivity matrix between the 53 glomeruli (51 AL glomeruli and 2 pseudoglomeruli) and the 200 KCs. Correlations were extracted by performing a principal component analysis of this matrix. This analysis failed to reveal structure in the input to KCs other than that inherent in the non-uniform distribution of glomerular inputs (Caron, 2013).

These data are consistent with a model in which each KC receives input from a combination of glomeruli randomly chosen from the non-uniform distribution of glomerular projections to the MB. Classification of either glomeruli or KCs on the basis of several shared developmental, anatomic, and functional features fails to reveal structured input onto individual KCs. Members of a given PN class do not preferentially converge onto an individual KC nor do members of a KC class receive specific and distinguishing PN inputs. A given KC can integrate information from glomeruli activated by food odours, pheromones, CO2 and even temperature. Recent data suggests that the extrinsic output neurons of the MB that are responsible for the different forms of learned behaviour are anatomically segregated and synapse with KC axons within a specific MB lobe. Interestingly, similar glomerular inputs are observed for the KCs that innervate the different lobes of the MB. This random input to individual KCs provides a mechanism to contextualize a rich diversity of novel KC responses (Caron, 2013).

It is important to note that the tracing procedure that was developed only allows characterization of the inputs to a single KC per fly. It is therefore possible that the inputs to every KC are determined but this developmental program results in a distribution of glomerular inputs that appears random. However, it is difficult to conceive of a development mechanism that could dictate the identity of inputs to each of the seven claws of the 2000 KCs. Moreover, the logic of employing complex and unlikely identity codes to achieve an uncorrelated distribution of inputs is elusive. Indeed, a previous study examined the electrophysiological response of KCs to different odours in a line of flies that labels only 23 α/β neurons but failed to identify replicate KCs with shared odour response profiles. These observations support the conclusion that the complement of glomerular inputs to KCs differs in different individuals. In addition, it is not possible, from the analysis of the glomerular inputs to 200 KCs, to exclude the existence of small subsets of KCs that received determined inputs from the AL. Nonetheless, the data are most consistent with a model in which the majority of individual KCs receive input from a random collection of glomeruli, a finding with important implications for odour processing in the MB (Caron, 2013).

If the connections from AL to MB are indeed random, a given odour will activate a different ensemble of KCs in different flies. However, in an individual fly, a given odour will consistently activate the same ensemble. This representation must acquire valence through experience or unsupervised activity dependent plasticity to dictate an appropriate behavioural output. Uncorrelated glomerular input to KCs affords the fly with the ability to impart meaning to a diversity of novel and unpredictable sensory stimuli that it may encounter throughout its life. Plasticity at highly convergent synapses between KC axons and MB extrinsic neurons could mediate experience-dependent behavioural output, an elemental feature of MB function. Thus, the fly has evolved an olfactory circuit with a connectivity that optimizes its ability to contextualize and respond appropriately to a rich array of olfactory experiences (Caron, 2013).

Stochastic and arbitrarily generated input patterns to the mushroom bodies can serve as conditioned stimuli in Drosophila

Warth Perez Arias, C. C., Frosch, P., Fiala, A. and Riemensperger, T. D. (2020). Front Physiol 11: 53. PubMed ID: 32116764

Single neurons in the brains of insects often have individual genetic identities and can be unambiguously identified between animals. The overall neuronal connectivity is also genetically determined and hard-wired to a large degree. Experience-dependent structural and functional plasticity is believed to be superimposed onto this more-or-less fixed connectome. However, in Drosophila, it has been shown that the connectivity between the olfactory projection neurons (OPNs) and Kenyon cells, the intrinsic neurons of the mushroom body, is highly stochastic and idiosyncratic between individuals. Ensembles of distinctly and sparsely activated Kenyon cells represent information about the identity of the olfactory input, and behavioral relevance can be assigned to this representation in the course of associative olfactory learning. This study has tested the hypothesis that the mushroom body can learn any stochastic neuronal input pattern as behaviorally relevant, independent of its exact origin. Fruit flies can learn thermogenetically generated, stochastic activity patterns of OPNs as conditioned stimuli, irrespective of glomerular identity, the innate valence that the projection neurons carry, or inter-hemispheric symmetry (Warth Perez Arias, 2020).

Mcm3 replicative helicase mutation impairs mushroom body neuroblast proliferation and memory in Drosophila

Blumröder, R., Glunz, A., Dunkelberger, B.S., Serway, C.N., Berger, C., Mentzel, B., de Belle, J.S. and Raabe, T. (2016). Genes Brain Behav [Epub ahead of print]. PubMed ID: 27283469

This study performed the molecular and phenotypic characterization of a structural brain mutant called small mushroom bodies (smu), which was isolated in a screen for mutants with altered brain structure. Focusing on the mushroom body neuroblast lineages, it was shown that failure of neuroblasts to generate the normal number of mushroom body neurons (Kenyon cells) is the major cause of the smu phenotype. In particular, the premature loss of mushroom body neuroblasts causes a pronounced effect on the number of late-born Kenyon cells. Neuroblasts show no obvious defects in processes controlling asymmetric cell division, but generate less ganglion mother cells. Cloning of smu uncovers a single amino acid substitution in an evolutionary conserved protein interaction domain of the Minichromosome maintenance 3 (Mcm3) protein. Mcm3 is part of the multimeric Cdc45/Mcm/GINS (CMG) complex, which functions as a helicase during DNA replication. The study proposes that at least in the case of mushroom body neuroblasts, timely replication is not only required for continuous proliferation but also for their survival. The absence of Kenyon cells in smu reduces learning and early phases of conditioned olfactory memory. Corresponding to the absence of late-born Kenyon cells projecting to α'/β' and α/β lobes, smu is profoundly defective in later phases of persistent memory (Blumröder, 2016).

A requirement for mushroom body signaling during olfactory memory retrieval

McGuire, S. E., Le, P. T. and Davis, R. L. (2001). Science 293: 1330-1333. PubMed ID: 11397912

The mushroom bodies (see Anatomical organization of the olfactory nervous system in Drosophila) of the Drosophila brain are important for olfactory learning and memory. To investigate the requirement for mushroom body signaling during the different phases of memory processing, neurotransmission was transiently inactivated through this region of the brain by expressing a temperature-sensitive allele of the shibire dynamin guanosine triphosphatase, which is required for synaptic transmission. Inactivation of mushroom body signaling through alpha/beta neurons during different phases of memory processing reveal a requirement for mushroom body signaling during memory retrieval, but not during acquisition or consolidation (McGuire, 2001).

Genetic and chemical disruption of the MBs produces flies that are normal for general behaviors but are defective in olfactory learning. Many genes involved in olfactory learning and memory show enriched expression in the MBs, particularly those encoding components of the cyclic adenosine monophosphate signaling pathway. Targeting of a constitutively active G-protein alpha subunit to the MBs disrupts olfactory learning, and restoring the rutabaga-encoded adenylyl cyclase specifically to the MBs of rutabaga mutants is sufficient to restore short-term memory in these flies. The model that has emerged from these experiments posits the MBs as important centers in olfactory associative learning and the likely site of convergence of the conditional (CS) and unconditioned (US) stimuli in classical conditioning (McGuire, 2001 and references therein).

A limitation of the previous experiments is that they all involve permanent alterations to the fly's brain throughout development, leading to the possibility that some of the effects on learning might reflect developmental perturbations rather than modifications of the physiology of these neurons that subserve learning and memory processes. Additionally, the irreversible nature of these interventions has made it impossible to dissect the roles of the MBs at the different stages of memory acquisition, consolidation, and retrieval (McGuire, 2001).

To explore the roles of the MBs in the different phases of memory processing, an approach was used that allows transient inactivativation of synaptic transmission from the MBs by targeting expression of a temperature-sensitive shibirets1 transgene to the MBs by the GAL4/UAS system. The shibire gene encodes a dynamin guanosine triphosphatase (GTPase) that is essential for synaptic vesicle recycling and maintenance of the readily releasable pool of synaptic vesicles. The temperature-sensitive allele shibirets1 bears a mutation in the GTPase domain, which renders the protein inactive at restrictive temperatures (>29°C) and causes a rapid inactivation of synaptic transmission and subsequent paralysis. Restricted expression of the shibirets1 transgene in specific cells produces blindness and paralysis at restrictive temperatures. Recently, the transgene was used to demonstrate the role of the dorsal paired medial neurons in memory formation (McGuire, 2001).

A number of GAL4 lines that exhibited enriched MB expression patterns were screened for 3-min memory performance when driving the UAS-shits1 transgene at both permissive (25°C) and restrictive (32°C) temperatures in an olfactory classical conditioning paradigm. In this assay, flies are conditioned by exposure to one odor paired with electric shock (CS+) and subsequent exposure to a second odor in the absence of electric shock (CS-). Memory is then assayed at predetermined time points after training by forcing the flies to choose between the CS+ and CS-. Several MB GAL4 lines demonstrate significant memory impairment at 3 min when tested at the restrictive temperature. These lines were next analyzed for sensorimotor functions required for the conditioning assay, including locomotion, odor avoidance, and electric shock avoidance, at both the permissive and restrictive temperatures. Subsequently, focus was placed on the GAL4 lines, c739 and 247, which demonstrate intact sensorimotor functions when driving the UAS-shits1 at both permissive and restrictive temperatures. For these MB GAL4 lines, memory at 3 min at the permissive temperature is indistinguishable among flies bearing both the GAL4 element and the UAS-shits1 transgene in combination and control flies bearing the GAL4 element or the UAS-shits1 element alone. At the restrictive temperature, however, the combination of c739 or 247 with the UAS-shits1 transgene results in a significant impairment of performance. The line 201Y; UAS-shits1 shows a slight but nonsignificant decrease in memory performance under these conditions. These data indicate that the inactivation of MB neurotransmission disrupts the processes underlying the encoding, storage, or retrieval of memory tested 3 min after training (McGuire, 2001).

These data were analyzed relative to the expression patterns of the three GAL4 lines to gain insights into possible functional subdivisions of the MBs. The GAL4 line 247, in which GAL4 is under the control of a 247-base pair (bp) enhancer fragment isolated from the D-mef2 gene, drives reporter gene expression in all lobes of the MB. In the line 201Y, the gamma lobe is preferentially marked, along with a small subset of the alpha/beta neurons. In contrast, the GAL4 c739 element drives reporter gene expression preferentially in the alpha/beta lobes. The expression overlap between the two GAL4 lines that disrupt 3-min memory when combined with UAS-shits1 at the restrictive conditions is within the alpha/beta lobes, suggesting the importance of this subset of MB neurons for the expression of memory. At the restrictive temperature, the UAS-shits1 in combination with 201Y, which preferentially drives reporter gene expression principally in the gamma lobes, does not significantly impair 3-min memory. The mild memory impairment in this line could be due to insufficient levels of expression of the UAS-shits1 transgene or rather it could reflect the possibility that the neurons in which this line drives the UAS-shits1 are not necessary for memory expression at this time point (McGuire, 2001).

To determine whether the deficient performance of these flies arises from a defect in memory acquisition, consolidation, or retrieval, memory was examined at a later time point (3 hours). Prior research has shown that most of the memory measured at this time point has been consolidated into an anesthesia-resistant form. The separation of training and testing also allows MB signaling to be reversibly inactivated separately during each phase and then it can be asked whether memory performance is affected. Three-hour memory was examined at the permissive temperature throughout the experiment. Under these conditions, the performance of the c739;UAS-shits1 flies is indistinguishable from flies bearing either the c739 element or the UAS-shits1 element. The lines 247 and 201Y in combination with UAS-shits1 disrupted 3-hour memory at the permissive temperature and were not analyzed further (McGuire, 2001).

To examine the requirement for signaling through the MBs during the retrieval of olfactory memory, training was performed under permissive conditions and the flies were maintained under these conditions until just before testing, at which point they were shifted to the restrictive temperature. When the performance of these flies was examined at 3 hours under these conditions, memory was abolished in the c739;UAS-shits1 flies, whereas the memory of the control groups was intact. Whether the acquisition of olfactory memory shares a similar requirement for MB signaling was examined. Training was performed under the restrictive conditions and immediately the flies were cooled to the permissive temperature. When the performance of these flies was examined at 3 hours under these conditions, a difference was observed between the c739;UAS-shits1 flies and the control line c739 but no difference between c739;UAS-shits1 and the UAS-shits1 control, indicating a general effect of heat on lines carrying the UAS-shits1 element, but no specific disruption of memory when UAS-shits1 is combined with c739. Subsequently it was investigated whether the interval between training and testing, during which memories are consolidated and stored, would require signaling through the MBs to observe normal memory performance at 3 hours. Flies were trained and tested under permissive conditions and given a temperature shift to restrictive conditions during the interval between these events. Under these conditions, a general effect of heat on the performance of all of the lines was observed, but no significant difference between any of the groups was observed (McGuire, 2001).

By transiently blocking synaptic transmission from the MBs during memory formation, consolidation, and retrieval, the temporal requirements of MB signaling during the different phases of memory processing could be examined. The results suggest quite unexpectedly that signaling through the MB alpha/beta neurons is required during olfactory memory retrieval, but not during memory acquisition or storage. It is proposed that, in Drosophila, olfactory memory retrieval requires signaling through the alpha/beta lobes to downstream neurons for expression. This does not preclude, however, a role for other MB lobes in memory formation, consolidation, or retrieval. A recent study demonstrating the sufficiency of rutabaga expression in the MBs for rescue of the short-term memory defect in rutabaga mutants has suggested that the gamma lobes might be of particular importance in the formation of short-term memories. Recent studies have also demonstrated that fasciclinII mutants are defective in memory acquisition and this protein is predominantly expressed in the alpha/beta neurons, although it is expressed at lower levels in the gamma lobe. One hypothesis to explain the combined observations is that memory formation occurs in the gamma neurons, or in both gamma and alpha/beta neurons simultaneously, but that memory retrieval occurs principally through the output of the alpha/beta neurons. Indeed, such a scenario involving a partial redundancy of function can explain why a subset of neurons might be sufficient, but not necessary, for memory expression. However, the observation that rutabaga and fasciclinII flies are only partially impaired in short-term memory indicates the likelihood that other mechanisms and perhaps locations of signal convergence, such as the antennal lobe or the lateral protocerebrum, may additionally mediate memory acquisition or storage. Taken together, these data suggest that acquisition and consolidation occur upstream of the MB synapse upon follower neurons, either in the MB neurons themselves or in upstream circuits. Retrieval of these memories within 3 hours would then engage signaling through a subset of the MB neurons, involving the alpha/beta lobes. It remains to be determined whether long-term memories (>24 hours) are dependent on the MBs (McGuire, 2001).

Go to top

Altered representation of the spatial code for odors after olfactory classical conditioning. Memory trace formation by synaptic recruitment

Yu, D., Ponomarev, A. and Davis, R. L. (2004). Neuron 42: 437-449. PubMed ID: 15134640

In the olfactory bulb of vertebrates or the homologous antennal lobe of insects, odor quality is represented by stereotyped patterns of neuronal activity that are reproducible within and between individuals. Using optical imaging to monitor synaptic activity in the Drosophila antennal lobe, classical conditioning is shown to rapidly alter the neural code representing the learned odor by recruiting new synapses into that code. Pairing of an odor-conditioned stimulus with an electric shock-unconditioned stimulus causes new projection neuron synapses to respond to the odor along with those normally activated prior to conditioning. Different odors recruit different groups of projection neurons into the spatial code. The change in odor representation after conditioning appears to be intrinsic to projection neurons. The rapid recruitment by conditioning of new synapses into the representation of sensory information may be a general mechanism underlying many forms of short-term memory (Yu, 2004).

Drosophila can develop a robust association between an odor, the conditioned stimulus (CS), and electric shock, the unconditioned stimulus (US), if the CS and the US are paired. Flies display their memory of this association by avoiding the odor CS during a test, after previously experiencing the pairing of the CS and the US. The number, nature, and the locations of the cellular memory traces that guide this acquired avoidance behavior are unknown, but significant evidence suggests that some cellular memory traces are formed in mushroom body neurons, higher-order neurons that form part of the olfactory nervous system. Furthermore, the evidence indicates that the memory traces are formed in part by the activation of the cyclic AMP signaling system. However, the memory traces that underlie insect odor memory are probably formed in many different areas of the olfactory nervous system and in other areas of the brain as well (Yu, 2004).

Optical imaging of synaptic activity in Drosophila brains coupled with behavioral conditioning has been used to visualize and study a cellular memory trace. This trace is established as new synaptic activity after conditioning in the antennal lobe projection neurons of the olfactory system. A concept established from these results, that may generalize to other forms of memory, is that memories form by the rapid recruitment of relatively inactive synapses into the representation of the sensory information that is learned. In other words, the synaptic representation of the odor CS is changed by learning, with new synaptic activity added to the representation after learning (Yu, 2004).

The anatomical organization of the Drosophila olfactory nervous system shares many fundamental similarities to that of vertebrates, suggesting that the mechanisms for odor perception, discrimination, and learning are shared. Olfactory receptor neurons (ORNs), distributed near the surface of the antenna and maxillary palp on each side of the head, project axons to the antennal lobe, where they terminate in morphologically discrete and synapse-dense areas known as glomeruli. The projection patterns of the ORNs are stereotyped between animals; ORNs that express the same olfactory receptor gene, although distributed across the surface of the antenna and maxillary palps, project their axons to the same glomerular target in the antennal lobe. There they are thought to form excitatory synapses with at least two classes of neurons: the local interneurons (LNs), a large fraction of which are GABAergic inhibitory neurons, and the projection neurons (PNs). A unique feature of the circuitry within the insect antennal lobe is the apparent existence of reciprocal dendrodendritic connections between the PNs and the LNs. The presence of these unique junctions with both transmissive and receptive specializations indicates that each glomerulus processes and makes computations that may underlie odor perception, discrimination, and learning, rather than being a simple transit station for the throughput of olfactory information. Individual PNs generally extend dendrites into a single antennal lobe glomerulus and then convey the processed olfactory information to two higher brain centers: the mushroom bodies and the lateral protocerebrum (Yu, 2004).

The neuroanatomy thus suggests that distinct odors are represented (1) by the stimulation of distinct sets of ORNs; (2) by spatial patterns of glomerulus activation within the antennal lobe, and (3) by a distinct set of synaptic fields activated in the mushroom bodies and the lateral protocerebrum. Functional imaging experiments have suggested the existence of a spatial code for odors within the antennal lobe of insects and the olfactory bulb of vertebrates. Calcium dyes, voltage-sensitive dyes, transgenically supplied fluorescent proteins, and intrinsic optical signals have been used to visualize odor-specific patterns of glomerulus activation in Drosophila, honeybee, zebrafish, salamander, and rat (Yu, 2004 and references therein).

The search for cellular memory traces began by first asking whether synaptic transmission could be detected in antennal lobe glomeruli of intact but immobilized adult flies after stimulation with pure odors that are frequently used as conditioned stimuli in behavioral learning experiments. It is possible to detect olfactory responses with optical reporters in the antennal lobes using reduced preparations of either isolated adult heads or dissected adult brains. Intact Drosophila adults are used, immobilized in a pipette tip, with their heads and antennae exposed. A small square of cuticle is removed from the dorsal head of each animal. Flies are mounted under a laser-scanning confocal microscope to detect basal fluorescence and the change in fluorescence induced with the application of odor. Initially flies were used carrying the GH146-GAL4 transgene to drive expression of a reporter of synaptic transmission; UAS-synapto-pHluorin (UAS-spH) was used to reveal PN presynaptic specializations within antennal lobe glomeruli (Yu, 2004).

A brief application of odor through a glass micropipette directed at the antennae produced a rapid, quantifiable, and stereotypic response in glomeruli between animals. For instance, the odor 3-octanol (OCT) produced a rapid burst of fluorescence in several glomeruli that occurred with the presentation of odor. Responses were quantified as the average percent change in the intensity of the pixels that represent each glomerulus during stimulation. A spatial response observed to OCT was observed as a pseudocolor image over eight glomeruli that were unambigously identified and that formed the focus of this study. Four of the eight glomeruli were activated reproducibly by OCT, whereas four others remained unchanged. These responses were quantitatively similar at two different odor concentrations. The increased responses of the four glomeruli at the higher odor concentration indicated that responses at lower odor concentrations fell well below the dynamic response ceiling for spH. The remarkably small standard errors that were obtained for glomerular responses between flies indicate that the procedures and standards that were employed were highly consistent and accurate. The many variables that could influence reproducibility include fly dissection, fly mounting, odor application, confocal scanning, glomerulus identification, and glomerulus circumscription during data analysis (Yu, 2004).

A stimulus that is used frequently as the unconditioned stimulus for olfactory classical conditioning is mild electric shock. This shock is normally delivered to flies as they stand on an electrified grid while also being in the presence of an odor. This US is effective at conditioning flies when presented along with an odor, although the identity of the neurons within the olfactory pathway that are stimulated by both odor and shock is unknown. Neurons that can function as cellular coincidence detectors must be activated either directly or indirectly by both stimuli (Yu, 2004).

It was therefore asked whether the PNs that responded to the odor CS could also respond to the US of electric shock. Pulses of electric shock were applied (at an intensity and frequency used to behaviorally condition adult Drosophila) to the abdomens of flies that were immobilized under the microscope. Synaptic transmission was activated in all glomeruli that expressed UAS-spH. The synaptic transmission events occurred with a periodicity that matched the 5 s interstimulus interval of the electric shock. When these time-based signals were converted to the frequency domain by Fourier transformation, a major component with a frequency matching the frequency of shock delivery (0.2 Hz) was extracted. These data indicate, therefore, that PNs are activated by electric shock stimuli applied to the abdomen. The neural pathways that carry the electric shock stimulus from the abdomen to the brain and the identity of the neurons immediately presynaptic to the PNs in this pathway have not yet been identified (Yu, 2004).

Since some PNs responded to both OCT and the US of electric shock when presented separately, it was of interest to ask whether these neurons could be conditioned by simultaneously presenting both odor and shock (forward conditioning). To test this, individual flies were conditioned either with OCT paired with electric shock or with one of a series of control protocols, including the odor only, shock only, and odor with shock but separated by 30 s to 2 min (trace conditioning). The optical response of the PNs to an odor test stimulus was then monitored 3 min after these treatments. The delay of 3 min was chosen, since for normal behavioral conditioning experiments it takes 3 min after training flies to test their choice behavior in a T-maze. Focus was placed on the effect of forward conditioning compared to other conditioning protocols, since the protocols of CS only, US only, and trace conditioning failed to produce behavioral conditioning (Yu, 2004).

The responses of most PNs to the test odor of OCT after the various conditioning protocols were similar or identical to the naive response. For instance, PNs innervating glomerulus DM2 responded to OCT with a 6% increase. Conditioning with the CS only, US alone, CS and US paired, or CS + US trace did not significantly alter this response. Similarly, most PNs that failed to respond to the odor CS by itself failed to exhibit any change in response after the conditioning protocols. Surprisingly, however, there was one notable exception. PNs innervating glomerulus D responded after forward conditioning to OCT with a %deltaF/F (fluorescence within each glomerulus relative to the basal fluorescence measured prior to the test with OCT) of 7%, while the responses of these PNs after US only, CS only, or trace conditioning protocols were similar to the naive response, which was not significantly different from zero. These data indicate, therefore, that forward pairing of the odor CS and the shock US rapidly awakens the PN synapses in the D glomerulus within 3 min after conditioning. The failure to observe a conditioning effect on the OCT-responsive glomeruli -- DM6, DM2, DM3, and DL3 -- cannot be due to a ceiling effect, since the odor concentration used for conditioning was well below the ceiling of spH's dynamic range. Thus, additional PN synapses in the antennal lobe are recruited rapidly to represent the odor CS after forward conditioning (Yu, 2004). These conclusions were reproduced and extended with a second type of experimental design. Since the response of the D PNs during the CS test was not affected by prior exposure of the CS when compared between flies, a 'within-animal' design was used for the next set of experiments. Each fly was presented the odor CS for 3 s, during which PN responses were monitored. After a rest of 5 min, the fly was then conditioned, and 3 min after conditioning, the response to a 3 s odor test was again monitored. The response before conditioning was compared with the response after conditioning. As before, the response of the D PNs to OCT alone was undetectable. However, the response after forward conditioning with OCT reached a %deltaF/F of 6% when measured 3 min later (Yu, 2004).

Is the memory trace in the D PNs specific for OCT as a CS, with other odors recruiting other sets of neurons and synapses, or is the change in D PNs a general property of learning something about any odor? To address this issue, within-animal conditioning experiments were carried out using methylcyclohexanol (MCH) as the CS, a second odor that is used frequently for odor learning in Drosophila (Yu, 2004).

The responses of some PNs to MCH before any conditioning were more variable between flies than for OCT. However, PNs innervating the three glomeruli DM6, DM2, and DM3 exhibited significant responses to MCH alone applied before conditioning. Forward conditioning, however, recruited the activity of glomerulus VA1 (both VA1l and VA1m) into the representation of MCH. Like the D glomerulus for OCT responses, VA1 was insensitive to MCH prior to conditioning. Therefore, different odors recruit normally insensitive PNs into their spatial representation after conditioning (Yu, 2004).

Behavioral memories can be very short or quite enduring, depending on the nature of the task learned, the strength of the training, the saliency of the cues, and undoubtedly the nature and number of the cellular memory traces that underlie the behavioral memory trace. The stability of the cellular memory trace that was established by forward pairing with OCT and shock in the D glomerulus PNs was probed by testing at different times after conditioning. This conditioned response waned rapidly. When tested at 5 min after conditioning, the increased response at 3 min had decayed to 5%, and by 7 min after conditioning the cellular memory trace was not significantly different from zero. Attempts to extend the duration of this memory trace with multiple and spaced conditioning trials have not been successful (Yu, 2004).

The recruitment of PN synapses of the D glomerulus into the representation of OCT and those of the VA1 glomerulus into the representation of MCH after conditioning suggests that synaptic recruitment was odorant specific. Nevertheless, the conditioned animals were conditioned and challenged with only one of the two odorants. To further explore the specificity of synaptic recruitment, a discriminative, within-animal experimental design was employed in which each animal was challenged with both odors prior to and after conditioning with either OCT or MCH (Yu, 2004).

PN synapses innervating glomeruli DM6, DM2, and DM3 showed significant responses to MCH before conditioning, while the remaining glomeruli failed to show significant responses. Most importantly, there were no significant differences in the responses to MCH after conditioning compared to those before conditioning. However, the conditioning recruited the PN synapses of the D glomerulus into the naive representation of OCT, which consisted of significant responses from glomeruli DM6, DM2, DM3, and DL3. In the reciprocal experiment, conditioning with MCH did not alter the representation of OCT by glomeruli DM6, DM2, DM3, and DL3 but selectively recruited PN synapses of VA1 into the MCH representation. These results, obtained with animals that were presented with two different odors in a discriminative protocol, strongly support the contention that the recruitment is odor specific (Yu, 2004).

Since the D PNs receive synaptic inputs from ORNs and LNs, it was of interest to ask whether the memory trace induced by OCT conditioning in D PN synapses was intrinsic to these neurons or whether the trace was established in one of the presynaptic partners so that the increase in D PN synaptic activity was only a reflection of an upstream memory trace. To test whether a synaptic memory trace was established in ORNs, UAS-spH was expressed using the ORN driver OR83b-GAL4. Using imaging conditions that were designed to identify glomerulus D and other glomeruli visible with GH146-GAL4, six glomeruli, along with D, were reproducibly discernable using this driver. Stimulation of flies with OCT produced a synaptic response in three of the six identified glomeruli, but these did not include D. Thus, PNs that innervate D do not receive excitatory input from the OR83b-expressing ORNs that form synapses in D (Yu, 2004).

It was asked, nevertheless, whether the ORNs that project to D responded to the US of electric shock and whether forward conditioning could recruit the OCT-blind D ORNs into being OCT sensitive. Electrical stimulation of flies carrying both OR83b-GAL4 and UAS-spH produced no increase in fluorescence of D or other glomeruli in response to shock pulses. Furthermore, forward pairing failed to produce any detectable change in synaptic activity within the identified glomeruli (Yu, 2004).

A GAD-GAL4 driver was used to direct expression of UAS-spH in LNs to address the same issues for these neurons. LNs that innervate glomeruli DM6, DM2, DM3, and DL3 all responded to the odor CS, whereas those innervating D, DL2, DA1, and VA1 failed to respond. The sets of responding and nonresponding glomeruli matched exactly those observed using the PN GAL4 driver. However, electric shock pulses to the body failed to stimulate synaptic responses in the LNs innervating D, and the synaptic responses of these neurons also could not be conditioned. The failure of the D PN synaptic trace to be transmitted to the LNs, which may be both presynaptic and postsynaptic to PNs, may indicate that the recruited D PNs may synapse on other PNs or interneurons rather than on the GAD-expressing LN or that the threshold for LN activation is simply too high for the memory trace to be transferred from the D glomerulus PNs (Yu, 2004).

Therefore, forward conditioning directly recruits D PNs into the representation of the CS of OCT. This recruitment is not the manifestation of a conditioned memory trace in the presynaptic ORNs or the LNs, since neither the ORNs nor the LNs that are presynaptic to the D PNs responded to the shock US, and neither neuron type exhibited a conditioned response (Yu, 2004).

The forward conditioning protocol used for most of the imaging experiments employed a single odor as the CS, paired with the US of electric shock pulses. Behavioral conditioning experiments, however, have often employed discriminative conditioning protocols with two different odors. The two-odor, discriminative, behavioral conditioning paradigm was modified into a single-odor classical conditioning paradigm to test the behavioral effects of the various conditioning protocols used for imaging. Flies were presented with CS only, US only, CS + US paired, or CS + US with a trace interval of 30 s, 1 min, or 2 min. They were then tested for their avoidance of the odor CS in a T-maze against a second odor to which they were naive and under conditions in which animals naive to any conditioning protocol distribute equally between the two odors (Yu, 2004).

The GH146-GAL4/UAS-spH flies were behaviorally conditioned using the new single-odor conditioning protocol and the effects of this behavioral conditioning was compared at 3 min posttraining to the conditioned synaptic responses of D PNs. The CS only, US only, or trace conditioning protocols produced small or no behavioral changes, similar to the lack of effect at the synaptic level. In contrast, forward conditioning produced a high behavioral performance score, similar to the robust synaptic change observed in D PNs. Therefore, the synaptic changes that were observed in D PNs produced by the conditioning protocols correlate well with the behavioral changes produced by the same protocols at 3 min after conditioning. Although the relative effectiveness of the various conditioning protocols correlated well between the imaged memory trace and behavioral performance, the duration of the behavioral memory after single-odor CS/US coincidence was much more enduring (>2 hr) than the enhanced synaptic activity of the D glomerulus PNs. Therefore, the D glomerulus memory trace would be capable of driving behavior for only the first few minutes after conditioning. Other memory traces of longer duration must be formed for more enduring behavioral performance (Yu, 2004).

The results offer two main conceptual advances. First, it is shown that forward conditioning of living Drosophila alters the representation of the odor CS in the PN synapses in the antennal lobe. Prior studies with the honeybee have suggested that memory traces are laid down in the antennal lobes, but these studies have employed pharmacological manipulations, calcium imaging, or physical insults to the entire antennal lobe without discriminating the roles of individual glomeruli, specific neuron types, or their synapses. In this study the GAL4 system of Drosophila to drive reporter expression in subsets of neurons, which provided resolution between types of neurons, and the reporter synapto-pHluorin, which provided a specific readout of synaptic activity in response to odorants. This approach was extended by imaging living flies before and after conditioning. This extension led to the specific finding that a short-lived cellular memory trace forms in Drosophila PNs after conditioning (Yu, 2004).

The existence of the short-term cellular memory trace in PNs and the correlated behavioral responses lends strong support to the idea that transient olfactory memories are formed in the insect antennal lobe. Much evidence has now accumulated to support the hypothesis that mushroom body neurons are centrally involved in odor learning, using the cAMP signaling cascade, in part, for the integration of sensory information. However, memories are distributed, and neurons other than mushroom body neurons are clearly involved in olfactory learning. The data provide evidence that the distributed memory system in Drosophila includes the antennal lobes. An attractive hypothesis is that the antennal lobes and the mushroom bodies are both sites for memory formation but that the earliest memories are formed in the antennal lobes by altering the representation of the sensory stimulus and that this altered representation is then transferred to and perhaps strengthened by the mushroom bodies (Yu, 2004).

The evidence offers the surprising conclusion that the PNs likely function as integrators of the CS and US. The ORNs, LNs, and PNs that innervate glomeruli recruited by conditioning did not respond to the odor CS. Of the three, only the PNs responded to the US of electrical shock. Thus, the available evidence suggests that PNs are the first point in the CS pathway that intersects functionally with the US pathway, although the possibility cannot be eliminated that ORNs and LNs receive US information via neuromodulatory rather than excitatory inputs, nor can the possibility be eliminated that some unknown neuron presynaptic to the recruited PNs integrates the CS and US. There is no neuroanatomical information about the US pathway from peripheral receptors or the identity of the presynaptic neurons providing US input to the PNs. However, it seems likely that the stimulus of electric shock must itself be processed by higher-order neurons in order to acquire its negative value attribute, which can then be stamped onto the PNs as associated with the CS. The CS pathway to the recruited PNs also remains unknown, since the odor CS (OCT or MCH) does not appear to be conveyed to glomerulus D or VA1 via the OR83b-expressing ORNs. It is possible that some ORNs that fail to express OR83b may project to these glomeruli and convey the CS stimulus. An alternative and more attractive possibility is that some local interneurons may convey the CS information from other glomeruli by synapsing on PNs innervating the recruited glomeruli. Such excitatory, interglomerular local interneurons have been discovered in the vertebrate olfactory bulb (Yu, 2004).

The second major conceptual advance is that the evidence suggests that memory traces are formed by the recruitment of synapses that are relatively silent to the odor CS, within the sensitivity of optical imaging, into the ensemble of synapses whose activity represents the odor CS in naive animals and that the selection of recruited synapses is odor specific. The possibility cannot be excluded, however, that some synaptic activity exists within the recruited PNs that is below the sensitivity of that detectable by optical imaging. Nevertheless, the results and the emerging evidence that cellular synaptic plasticity may occur from the activation of normally silent synapses suggest that some forms of behavioral memory may occur through a large synaptic gain mechanism, perhaps approaching an 'off-on' switch mechanism, rather than through smaller graded changes in synapses that represent the stimuli in naive animals. Thus, memory formation involves the recruitment of synapses to represent the sensory cues that are learned (Yu, 2004).

In addition to these advances, these findings also pose new and intriguing puzzles. Is the short-term memory trace established in the PNs independent of other memory traces, so as to directly guide behavior for a short period after learning, or is it transferred to the mushroom bodies or the lateral protocerebrum, perhaps to be consolidated there into a more enduring trace, with behavior being guided from these higher-ordered brain centers? A related question is whether the PN synaptic memory trace is specific to the connections made in the antennal lobe or whether this occurs on a cell-wide basis, with conditioning also stamping its effects on PN synapses made in the mushroom bodies and the lateral protocerebrum. Do any of the known memory mutants disrupt the formation or stability of the PN memory trace? How is it that the recruitment of new synapses in the antennal lobe produces a new representation of the learned odor? Is it just simply that more activated synapses represent the learned odor, or does the synaptic activation of PNs alter the coding of the odor CS, perhaps by influencing the coherency or timing of PN and LN oscillations that may contribute to odor encoding (Yu, 2004)?

Go to top

Drosophila DPM neurons form a delayed and branch-specific memory trace after olfactory classical conditioning

Yu, D., Keene, A. C., Srivatsan, A., Waddell, S. and Davis, R. L. (2005). Cell 123(5): 945-57. PubMed ID: 16325586

Formation of normal olfactory memory requires the expression of the wild-type amnesiac gene in the dorsal paired medial (DPM) neurons. Imaging the activity in the processes of DPM neurons revealed that the neurons respond when the fly is stimulated with electric shock or with any odor that was tested. Pairing odor and electric-shock stimulation increases odor-evoked calcium signals and synaptic release from DPM neurons. These memory traces form in only one of the two branches of the DPM neuron process. Moreover, trace formation requires the expression of the wild-type amnesiac gene in the DPM neurons. The cellular memory traces first appear at 30 min after conditioning and persist for at least 1 hr, a time window during which DPM neuron synaptic transmission is required for normal memory. DPM neurons are therefore 'odor generalists' and form a delayed, branch-specific, and amnesiac-dependent memory trace that may guide behavior after acquisition (Yu, 2005).

Memory traces together represent the memory engram that directs behavior of the organism after learning or conditioning events. Classical conditioning is one form of learning whereby a conditioned stimulus (CS) becomes predictive of an unconditioned stimulus (US) when the two stimuli are paired in an appropriate way. The prototypic example of classical conditioning stems from studies on dogs conducted by Ivan Pavlov in which tone cues (CS) paired with a food reward (US) became predictive of the food reward, shown by the dog's salivation upon hearing the tone cue after conditioning. In Drosophila, olfactory classical conditioning is a robust and well-studied type of learning in which olfactory cues (CS) are usually paired with electric shock (US), such that conditioning leads to learned avoidance behavior of the CS. Learning to associate two forms of sensory information likely involves specific neurons that respond to both sensory cues and can integrate the information to produce learning. Thus, memory traces for olfactory classical conditioning in Drosophila are expected to form in neurons positioned at the intersections of the olfactory nervous system, the pathway that conveys and processes the CS (CS pathway), and the pathways that convey and process the US (US pathway) (Yu, 2005 and references therein).

The insect olfactory nervous system begins with olfactory receptor neurons (ORNs) distributed between the antennae and maxillary palps. The ORNs project axons to the antennal lobe, where they terminate in morphologically discrete and synapse-dense areas known as glomeruli. There, the ORNs are thought to form excitatory synapses with at least two classes of neurons, one of these being the projection neurons. The projection neurons then convey the olfactory information along their axons in the antennal cerebral tract to at least two higher brain centers, the mushroom bodies and the lateral horn. A large body of evidence has accumulated indicating the importance of the mushroom bodies for olfactory learning. Thus, odors are represented first in the olfactory nervous system by the activation of overlapping sets of ORNs; second by the activation of overlapping sets of projection neurons, and third by the activation of mushroom body and lateral horn neurons (Yu, 2005 and references therein).

An olfactory memory trace forms in the projection neurons after olfactory classical conditioning. Synaptic release of neurotransmitter from presynaptic specializations of projection neurons in the antennal lobe was monitored optically using the transgenically supplied indicator of synaptic transmission, synapto-pHluorin (spH). The memory trace was detected in this case by a rapid but short-lived recruitment of new synaptic activity into the representation of the learned odor. More specifically, distinct odors stimulate distinct sets of projection neurons in naive animals. Within 3 min after conditioning, additional sets of projection neurons become activated by the learned odor. This recruitment is odor specific; different odors recruit different sets of projection neurons into the representation of the learned odor. The recruitment of new sets of projection neuron synapses into the representation of the learned odor, however, is short lived, lasting only 5 min before the synaptic release from the recruited sets of projection neurons decays to the undetectable levels observed prior to conditioning. The short-lived memory trace of projection neurons, although potentially important for guiding behavior for a few minutes after conditioning, cannot account for the time course for behavioral memory, which can last for days. Thus, memory traces in other areas of the nervous system must provide for the persistence of behavioral memory (Yu, 2005).

The dorsal paired medial (DPM) neurons are large neurons that express neuropeptides encoded by the amnesiac (amn) gene and are critical for normal memory. The encoded neuropeptides are related to pituitary adenylyl cyclase-activating peptide (PACAP). The DPM neurons have been widely hypothesized to be part of the US pathway through the release of the expressed modulatory neuropeptides, in part because their processes invade the mushroom body neuropil and are thought to intersect the olfactory nervous system. There is also evidence that these neurons release acetylcholine as a co-neurotransmitter along with neuropeptides (Yu, 2005).

The hypothesis that DPM neurons are solely part of a US pathway predicts that the US but not the CS should activate them and that their response properties should not change after olfactory classical conditioning. A change in their response pattern after conditioning would indicate the presence of a memory trace. A surprising observation is reported that not only do DPM neurons respond to the US of electric shock -- predicted by the hypothesis that they are part of the US pathway -- but they are odor generalists, responding to all odors that were tested. Moreover, they form odor-specific memory traces as registered by increased odor-evoked calcium influx and synaptic transmission. In contrast to the memory trace that forms immediately after conditioning in projection neurons, the memory trace that forms in the DPM neurons is delayed, appearing at 30 min after olfactory classical conditioning. Temporally distinct memory traces that form within projection neurons and DPM neurons after classical conditioning may be partly responsible for guiding behavior during different time windows after learning (Yu, 2005).

There are two DPM neurons, each with a large cell body residing in the dorsal aspect of each brain hemisphere. They have no obvious dendritic field and extend a single neurite in an anterior direction toward the neuropil regions (lobes) that contain the axons of the mushroom body neurons. The neurite from each DPM neuron splits, and one branch broadly innervates the vertical mushroom body lobes while the other innervates the horizontal mushroom body lobes. Careful examination of 12 different confocal stacks highlighting the DPM neurons with the DPM neuron driver, c316-GAL4, revealed that all of the fluorescence in the vertical and horizontal lobes of the mushroom bodies in c316-GAL4/UAS-mCD8GFP flies can be traced to the DPM neuron cell bodies rather than other c316-GAL4-expressing neurons in the brain (Yu, 2005).

It was first asked whether DPM neurons respond to electric-shock pulses delivered to the abdomen of living flies. The processes of DPM neurons in the mushroom body lobes of flies carrying c316-GAL4 and the synaptic transmission reporter UAS-synapto-pHluorin (UAS-spH) or the calcium reporter UAS-G-CaMP were visualized before and during the application of electric-shock pulses delivered to the abdomen. Electric-shock pulses were used of the same intensity, duration, and frequency as those used for behavioral conditioning. The calcium influx in the DPM neuron processes innervating the vertical mushroom body lobes that occurs with electric shock was examined. There was a dramatic response in the processes at the distal tip of the vertical lobes as well as at the vertical-lobe stalk. The change in fluorescence (ΔF/Fo) was examined that occurs with 12 shock pulses delivered at a rate of 1 shock pulse every 5 s. There was an increase in ΔF/Fo coincident with each shock pulse in the vertical lobes and the horizontal lobes with both G-CaMP and spH. These data indicate therefore that electric-shock pulses to the abdomen produce both calcium influx into the DPM neuron processes and synaptic release from their terminals, observations consistent with the possibility that DPM neurons provide US input to the mushroom body neurons (Yu, 2005).

The hypothesis that DPM neurons provide US input into the mushroom body neurons for olfactory memory formation predicts that these neurons should not be activated by the CS of an olfactory stimulus. To test this prediction, DPM neuron calcium influx and synaptic transmission was examined in flies presented with odor stimuli to their antennae and maxillary palps. Stimulation with pure odors like 3-octanol (OCT), 4-methylcyclohexanol (MCH), and benzaldehyde (BEN) elicited robust calcium influx into the DPM neuron processes innervating the vertical mushroom body lobes. The magnitude of the response was dependent on the odor concentration; no responses were observed using air blown over mineral oil, which was used as an odorant diluent. Moreover, these pure odors elicited synaptic activity of the DPM neuron as well as increased calcium influx. Odor responses were also observed in the DPM neuron processes innervating the horizontal mushroom body lobes. Finally, the generality of the DPM odor-evoked response was tested using a battery of 17 different odorants, ranging from pure odors to complex odors such as apple, banana, and grape. In all cases, the DPM neurons responded with increased calcium influx into their processes. Therefore, the DPM neurons are odor generalists in the sense that they apparently respond to all odors administered to the fly. However, the circuitry that provides odorant information to the DPM neurons is unknown (Yu, 2005).

Since the DPM neurons responded to both an odor conditioned stimulus and an electric-shock unconditioned stimulus, the possibility was considered that the neurons might form a memory trace and exhibit a changed response to the CS after olfactory classical conditioning. Either calcium influx into the DPM neuron processes or synaptic release after olfactory classical conditioning were measured. For all experiments, each animal was used for only one measurement in order to avoid potential complications produced by odor habituation, adaptation, or generalization that could occur with multiple exposures (Yu, 2005).

A within-animal experimental design was employed in which the response of the DPM neuron processes within each animal was first evaluated with a single, 3 s presentation of odor. This was followed by forward conditioning, in which a 60 s odor stimulus was presented simultaneously with 12 electric-shock pulses or by backward conditioning in which the 60 s odor stimulus was presented after the onset of the electric-shock stimuli. Forward conditioning leads to robust behavioral conditioning, whereas backward conditioning does not. The response of the DPM neuron processes was then tested at various times after conditioning, again within each animal, and the postconditioning response was compared to the preconditioning response (Yu, 2005).

The postconditioning responses of the DPM neurons to any of three conditioned odors, OCT, MCH, or BEN, did not differ from the preconditioning responses when assayed at 3 min after forward conditioning. This is in marked contrast to the odor-specific memory trace responses that occur in the projection neurons of the antennal lobe within 3 min after conditioning. However, amn mutant flies are only slightly impaired in 3 min memory and have a more pronounced impairment at later times beginning 10-60 min after training. Furthermore, synaptic transmission from DPM neurons is not required for 3 min memory and is dispensable during acquisition and retrieval for robust 3 hr behavioral memory. DPM synaptic transmission is instead required during the interval between training and testing. These observations led to a consideration of the possibility that the DPM neurons might form a memory trace with a delayed onset (Yu, 2005).

The postconditioning responses at 30 min after forward conditioning to three different odors compared to the preconditioning responses revealed that a delayed memory trace registered by increased calcium influx was detectable at this time. The increase in calcium influx was not observed at 30 min after backward conditioning, indicating that the memory trace is dependent on the order in which the CS and US are presented, like behavioral conditioning. Furthermore, the increased calcium influx after forward conditioning was also detectable at 60 min after conditioning but not at 15 or 120 min after forward conditioning. However, the variability of the postconditioning responses at 120 min was larger than at other time points, probably because the physiological state of the flies becomes compromised and more variable from the prolonged immobilization. Thus, the DPM neuron memory trace, detectable first at 30 min postconditioning, extends to at least 1 hr and perhaps 2 hr after conditioning. Moreover, the delayed memory trace registered by increased calcium influx into the DPM neuron processes at 30 min after conditioning was also registered as increased synaptic transmission using spH as a reporter. Therefore, odor evokes both increased calcium influx and increased synaptic transmission from DPM neurons 30 min after forward conditioning (Yu, 2005).

Two general models were considered for the role of the amn gene and the DPM neurons in the process of olfactory memory formation in Drosophila. The first model, the possibility that the amn gene and the DPM neurons provide solely US information for the process of acquisition, is unlikely for several different reasons. (1) amn mutants have normal levels of memory acquisition, shown by memory growth curves with multiple training trials relative to control flies. Impairment in the processing of the US information would likely cause the mutant flies to exhibit performance scores that reach asymptote at levels lower than controls. (2) The parallel nature of the memory growth curves also suggests that the processing of CS information is unimpaired since CS impairment should slow the memory growth rate relative to control flies. (3) These data suggest that the association process itself, or acquisition, is unimpaired since a defect in the association of the CS with the US would also alter the memory growth curve. (4) The discovery of a delayed olfactory memory trace within the DPM neurons themselves, unless fortuitous, is inconsistent with a role specific to US processing. Rather, the data are strongly consistent with a second model alternative envisioning amn and DPM neuron involvement in the formation of intermediate-term memory. The amn mutants exhibit no obvious deficit in acquisition but are impaired in memory. Synaptic transmission is required from the DPM neurons during the interval between training and testing but not at the time of training or testing. The latter observation indicates either that DPM neurons are chronically active or that acquisition itself leads to sustained DPM neuron activity since blocking synaptic activity after acquisition produces a memory impairment at 3 hr. The delayed memory trace formed in DPM neurons that is coincident in time with their requirement for normal memory formation argues for their involvement in an intermediate stage of memory (Yu, 2005).

The delayed olfactory memory trace that forms in the DPM neurons is different from previously observed projection neuron memory trace in several interesting ways. (1) The memory trace formed by antennal-lobe projection neurons occurs by the recruitment of new synaptic activity into the representation of the learned odor. In other words, there is a qualitative change in the brain's representation of the learned odor as represented by projection neuron activity. The memory trace formed by DPM neurons, in contrast, is a quantitative one, being manifest as an increase in calcium influx and synaptic release with CS stimulation after acquisition. Despite this, the trace formed in the DPM neurons is odor specific. (2) The memory trace formed by projection neurons is detectable very early (as little as 3 min) after training, whereas the memory trace formed by DPM neurons is delayed, forming between 15 and 30 min after training. (3) The memory trace formed by projection neurons is very short lived, existing for about 5 min after training. The memory trace established in DPM neurons persists for at least 2 hr after training. The existence of multiple memory traces in distinct areas of the olfactory nervous system with different times of formation and duration leads to the interesting hypothesis that memory of a singular event over time is due to multiple and distinct memory traces that guide behavior during different windows of time after learning, a conclusion also reached from studies with the honeybee (Menzel, 2001; Yu, 2005).

These observations show that the delayed olfactory memory trace is established in the DPM neuron branch that innervates the vertical mushroom body lobes and not in the branch that innervates the horizontal mushroom body lobes. Thus, there exists an intriguing branch specificity to the formation of the delayed olfactory memory trace. The significance of this observation is not yet clear. However, other studies have pointed to the possibility that mushroom body neurons have branch-specific information processing. Some flies mutant for the α lobes absent (ala) gene lack the vertical branch or the horizontal branch of the mushroom body neurons. Intriguingly, mutant animals missing only the vertical branch of the mushroom body neurons have been reported to exhibit normal short-term memory but no long-term memory. Thus, long-term memory may form only in the vertical branch of the mushroom body neurons or be retrieved specifically from this branch. The formation of a delayed olfactory memory trace in the DPM neuron branch that innervates the vertical mushroom body lobes is consistent with the possibility that branch-specific long-term memory processes occurring in the vertical branch of the mushroom body lobes are dependent on the delayed memory trace that forms in this DPM neuron branch (Yu, 2005).

The delayed memory trace that forms in the DPM neurons is dependent on the normal function of the amn gene product since the trace fails to form in amn mutants but can be rescued by expression of the wild-type amn gene in the DPM neurons. This observation raises at least three possibilities for the role of the amn-encoded neuropeptides in the formation of the delayed olfactory memory trace. (1) It is possible that the released neuropeptides exert their effects in an autocrine fashion, interacting with neuropeptide receptors on the DPM neurons themselves in order to initiate the formation of the memory trace. (2) It is also possible that the released neuropeptides interact with receptors on postsynaptic neurons, such as mushroom body neurons, and that this stimulates a retrograde signal that leads to the formation of the DPM neuron memory trace. (3) It is possible that the amn-encoded neuropeptides are not employed for physiological changes in the adult brain but are required in a developmental capacity for DPM neurons to be competent to form the memory trace (Yu, 2005).

There exist at least two broad explanations for the role of the DPM neurons and the amn-encoded neuropeptides in olfactory learning. One possibility is that the DPM neurons integrate CS and US information independently of integration events that occur elsewhere in the nervous system. In this scenario, the CS information may be transmitted to the DPM neurons via unknown interneurons from the antennal lobe or lateral horn, or, alternatively, the DPM neurons might receive CS information from the mushroom body axons. In other words, DPM neurons may be postsynaptic to the mushroom body neurons. This could explain why the DPM neurons are odor generalists since their broad innervation of the mushroom body lobes would allow them to sample the odorant-stimulated activity of many or all mushroom body neurons. This possibility predicts that the DPM neurons should exhibit postsynaptic specializations on some of their processes -- perhaps those that innervate the horizontal lobes, as one possibility. The strengthening of specific mushroom body-DPM neuron synapses after olfactory learning could explain how the DPM neurons form odor-specific memory traces despite being odor generalists. Other DPM neuron processes may be presynaptic to the mushroom bodies such that the CS/US integration events that occur within the DPM neurons might be passed on to the mushroom bodies to reinforce their output. The presynaptic interactions may be through synapses onto the mushroom body fibers in the vertical lobes, reinforcing mushroom body output over the intermediate term and perhaps establishing the permissive signaling events for long-term memories to form in the vertical lobes. The DPM neurons may also receive US information indirectly from the mushroom body neurons or from other neurons. The contributions of the two putative DPM neuron neurotransmitters -- acetylcholine and neuropeptides -- to these processes remain to be clarified. Both acetylcholine and amn neuropeptides are required for behavioral memory (from experiments with Shibire and amn mutants, respectively). The amn neuropeptides are also required autonomously for the formation of the DPM neuron memory trace (Yu, 2005).

The second broad explanation envisions the DPM neurons as maintaining already integrated information through a networked association with the mushroom bodies. The complete integration of CS and US information may occur in the projection neurons and mushroom body neurons. DPM neurons, in a postsynaptic role to the mushroom bodies, would receive integrated information leading to increased excitability. The transfer of the CS/US-integrated information from the mushroom bodies to the DPM neurons may occur immediately after learning, initiating a process intrinsic to the DPM neurons that produces a delayed increase in odor-evoked transmission 30 min later, or the transfer of the integrated information itself from the mushroom bodies to the DPM neurons may occur through a delayed process after learning. In either case, the increased excitability of the DPM neurons would feed back onto and strengthen the output of the mushroom body neurons, leading to robust intermediate-term memory (Yu, 2005).

Go to top

Sequential use of mushroom body neuron subsets during Drosophila odor memory processing

Krashes, M. J., et al. (2007). Neuron 53: 103-115. PubMed ID: 17196534

Drosophila mushroom bodies (MB) are bilaterally symmetric multilobed brain structures required for olfactory memory. Previous studies suggested that neurotransmission from MB neurons is only required for memory retrieval. An unexpected observation that Dorsal Paired Medial (DPM) neurons, which project only to MB neurons, are required during memory storage but not during acquisition or retrieval, led to a revisiting of the role of MB neurons in memory processing. Neurotransmission from the α′β′ subset of MB neurons is required to acquire and stabilize aversive and appetitive odor memory, but is dispensable during memory retrieval. In contrast, neurotransmission from MB αβ neurons is only required for memory retrieval. These data suggest a dynamic requirement for the different subsets of MB neurons in memory and are consistent with the notion that recurrent activity in an MB α′β′ neuron-DPM neuron loop is required to stabilize memories formed in the MB αβ neurons (Krashes, 2007).

It is often said that form follows function. According to this postulate, the striking multilobed arrangement of the insect MBs would imply functional differences between the different types of MB neurons: αβ, α′β′, and γ, but very limited data describing the individual function of these anatomical subdivisions exists. Although several complex behaviors in insects appear to require the MBs and a differential role for distinct MB neuron groups has been suggested, most conceptual models of memory treat the MBs as a single unit (Krashes, 2007).

One of the most detailed examinations of MB function has been in the context of Drosophila aversive olfactory memory, where flies are trained to associate specific odors with the negative reinforcement of electric shock. Genetic studies over the last thirty years have suggested that the MBs play an essential role in fly olfactory memory, but the role of the MBs in memory acquisition, storage, and retrieval has only been examined recently. Taking advantage of a dominant, temperature-sensitive dynamin transgene, uas-shits1, a number of laboratories concluded that MB output was required only for recall, but not for acquisition or storage. These and other findings have led to a simple model wherein Drosophila olfactory memory is formed and “stored” at MB output synapses (Krashes, 2007).

Functional studies of DPM neurons, MB extrinsic neurons that ramify throughout the MB lobes, demonstrated they were specifically required during consolidation, but not acquisition or storage. Furthermore, genetically modified DPM neurons that primarily innervate the MB α′β′ lobes retain function, implying that MB α′β′ neurons might also have a similar function in memory consolidation (Krashes, 2007).

Examination of the GAL4 enhancer-trap lines used to express the uas-shits1 transgene in the earlier MB studies revealed that c309, c747, and MB247 only express in a few MB α′β′ neurons compared to αβ and γ neurons, while c739 expresses exclusively in αβ neurons. Thus, it seems likely that prior studies utilizing these drivers did not observe requirements for MB activity during either olfactory memory acquisition or storage because of insufficient expression in α′β′ neurons (Krashes, 2007).

Subsequently two GAL4 enhancer-traps that strongly express in MB α′β′ neurons were identified to test this hypothesis. The expression of c305a appears to be entirely restricted to α′β′ neurons within the MBs whereas c320 expresses in α′β′, αβ, and a few γ neurons. Both of these lines also express in additional non-MB neurons, so MB{GAL80} tool was employed to more rigorously test the requirement for MB activity in these {GAL4} lines. With these reagents the role of MB α′β′ neurons in memory was investigated and it was found that MB α′β′ neuron output during and after training is critical for the formation and consolidation of both appetitive and aversive odor memory from a labile to a more stable state. For comparison the requirements were also examined for MB αβ neurons using c739, confirming previous results. Thus, output from the MB α′β′ neuron subset is required for memory acquisition and stabilization, whereas output from αβ neurons is apparently dispensable during training and consolidation but is required for memory retrieval (Krashes, 2007).

Based on c305a and c739 data, it is recognized that c320 flies, which express in both α′β′ and αβ neurons, might be expected to exhibit memory loss if MB neuron output was blocked during both the consolidation and recall time windows. However, it is possible that a retrieval effect was not observed with c320 because it expresses GAL4 in fewer αβ neurons, or is in a different subset of αβ neurons relative to the c739 driver (Krashes, 2007).

Despite this caveat, these data suggest that different lobes of the MB have different roles in memory and provide a significant shift in the current understanding of the role of the MB in memory. Older models implied that MB αβ, α′β′, and γ neurons were largely interchangeable, and that each of the MB neurons that responded to a particular odor received coincident CS and US input and modified their presynaptic terminals to encode the memory. The data presented in this study suggest that MB αβ and α′β′ neurons are functionally distinct (Krashes, 2007).

In this study, the role of the unbranched γ lobe neurons was not investigated. Previous work with c309, c747, and MB247 suggests that neurotransmission from γ neurons is likely dispensable for acquisition and consolidation. In addition, a prior study indicated that γ neurons are minimally involved in middle-term memory (MTM) and anesthesia-resistant memory (ARM). However, it is possible that experiments to date have not employed odors that require γ neuron activation. The response of γ neurons may be tailored to ethologically relevant odors such as pheromones. It is notable that fruitless, a transcription factor required for male courtship behavior, is expressed in MB γ neurons, and blocking expression of the male-specific fruM transcript in the MB γ neurons impairs courtship conditioning. If the relevant odors can be identified, it will be interesting to determine if MB α′β′ neurons and Dorsal Paired Medial (DPM) neurons are required to stabilize these odor memories in the γ neurons. Recent work is supportive of the idea that odor identity may be a factor in determining the requirement for the subsets of MB neurons in olfactory learning (Krashes, 2007).

Stable aversive and appetitive odor memory requires prolonged DPM neuron output during the first hour after training, and DPM neuron output is dispensable during training and retrieval. DPM neurons ramify throughout the MB lobes, but DPM neurons that have been engineered to project mostly to the MB α′β′ lobes retain wild-type capacity to consolidate both aversive and appetitive odor memory. This study has demonstrated that, similar to wild-type DPM neurons, blocking output from these modified DPM neurons for 1 hr after training abolishes memory. Thus, finding a specific role for both DPM neuron output to MB α′β′ lobes and MB α′β′ neuron output during the first hour after training is consistent with the notion that a direct DPM-MB α′β′ neuron synaptic connection is important for memory stability. It should be reiterated that the focus of this paper has been on protein synthesis-independent memory, and whether or not a similar processing circuit is utilized for protein synthesis-dependent LTM remains an open question (Krashes, 2007).

Beyond simply attributing an additional function to the MBs, when taken in conjunction with work on the role of DPM neurons in memory, the data presented here suggest a new model for how olfactory memories are processed within the MBs. It is proposed that olfactory information received from the second-order projection neurons (PNs) is first processed in parallel by the MB αβ and α′β′ neurons during acquisition. Activity in α′β′ neurons establishes a recurrent α′β′ neuron-DPM neuron loop that is necessary for consolidation of memory in αβ neurons, and subsequently, memories are stored in αβ neurons, whose activity is required during recall. It is plausible that MB α′β′ neurons are directly connected to MB αβ neurons and/or that DPM neurons provide the conduit between MB neurons. However, the finding that DPM neurons that project primarily to MB α′β′ neurons are functional implies that only a few connections from DPM neurons to MB αβ neurons are necessary (Krashes, 2007).

The requirement for α′β′ neuron output during training also potentially provides a source for the activity that drives DPM neurons. DPM neuron activity is not required during training, and the current data are consistent with the idea that olfactory conditioning triggers activity in MB α′β′ neurons, which in turn elicits DPM neuron-dependent activity. It is proposed that after training, recurrent MB α′β′ neuron-DPM neuron activity is self-sustaining for 60-90 min. This recurrent network mechanism is similar to models for working memory in mammals. It is also conceivable that MB α′β′ neurons receive prolonged input after training from the antenna lobes via the PNs. Olfactory conditioning has been reported to alter the odor response of Drosophila PNs in the AL, but the observed effects were short-lived. Nevertheless, AL plasticity for a few minutes after training could contribute to the required MB α′β′ neuron activity. If continued activity from the AL is required for consolidation, blocking PN transmission with shits1 for 1 hr after training should abolish memory. The bee AL and MB are clearly involved in olfactory memory and may function somewhat independently in learning and memory consolidation, respectively. However, biochemical manipulation of the bee AL can also induce LTM, and therefore it is possible that either plasticity in the AL alone can support LTM, or that the AL and MB interact during acquisition and consolidation. A differential role for the AL and MBs has also been suggested from neuronal ablation studies of courtship conditioning in Drosophila. Short-term courtship memory can be supported by the AL, but memory lasting longer than 30 min requires the MBs (Krashes, 2007).

This work also has significant implications for the organization of aversive and appetitive odor memories in the fly brain. Stability of both appetitive and aversive memory is dependent on DPM neurons and MB α′β′ neurons. It therefore appears that processing of aversive and appetitive odor memories may bottleneck in the MBs. It has been demonstrated that aversive memory formation requires dopaminergic neurons whereas appetitive memory relies on octopamine to provide a possible mechanism to distinguish valence (see Tyramine β hydroxylase). However, it was also found that MB output is required to retrieve aversive and appetitive odor memory, suggesting that both forms of memory involve MB neurons and that both US pathways may converge on MB neurons. It will be important to understand how the common circuitry is organized to independently process the different types of memory and to establish if, and how, such memories coexist (Krashes, 2007).

These data imply that stable memory may reside in MB αβ neurons because blocking output from MB αβ neurons impairs retrieval of MTM and ARM (both components of 3 hr memory). It has been proposed that AMN peptide(s) released from DPM neurons cause prolonged cAMP synthesis in MB neurons that is required to stabilize memory. The finding that genetically engineered DPM neurons mostly projecting to the MB α′β′ lobes are functional, taken with the idea that stable memory resides in MB αβ neurons, is somewhat inconsistent with the notion that crucial AMN-dependent memory processes occur in MB αβ neurons. However, it is plausible that AMN, or another DPM product that is released in a shibire-dependent manner, could diffuse locally from the aberrant DPM neurons to MB αβ neurons (Krashes, 2007).

This work demonstrates that MB αβ neurons and α′β′ neurons have different roles in memory. Beyond gross structural and gene expression differences, it will be essential to establish the precise connectivity, relative excitability, and odor responses of the different MB neurons. Future study may also reveal further functional subdivision within the MB lobes, and it should be possible to refine the current MB α′β′ neuron GAL4 lines with appropriate GAL80 transgenes and FLP-out technology (Krashes, 2007).

In the mammalian brain memories that initially depend on the function of the hippocampus lose this dependence when they are consolidated. This transient involvement of the hippocampus has led to the idea that consolidation of memory results in the transfer of memory from the hippocampal circuits to the cortex. An alternate view is that aspects of the memory are always in the cortex but are dependent on the hippocampus because recurrent activity from cortex to hippocampus to cortex is required for consolidation. Hence, disrupting hippocampal activity during consolidation leads to memory loss (Krashes, 2007).

The current data suggest the simpler fruit fly brain similarly employs parallel and sequential use of different regions to process memory. MB α′β′ neuron activity is required to form memory, MB α′β′ neurons and DPM neurons are transiently required to consolidate memory, and output from αβ neurons is exclusively required to retrieve memory. It is therefore propose that aversive and appetitive odor memories are formed in MB αβ neurons and are stabilized there by recurrent activity involving MB α′β′, DPM neurons, and the MB αβ neurons themselves (Krashes, 2007).

It is becoming increasingly apparent that neural circuit analysis will play an important role in understanding how the brain encodes memory. The ease and sophistication with which one can manipulate circuit function in Drosophila, combined with the relative simplicity of insect brain anatomy, should ensure that the fruit fly will make significant contributions to this emerging discipline (Krashes, 2007).

Go to top

In vivo large-scale analysis of Drosophila neuronal calcium traces by automated tracking of single somata

Delestro, F., Scheunemann, L., Pedrazzani, M., Tchenio, P., Preat, T. and Genovesio, A. (2020). In vivo large-scale analysis of Drosophila neuronal calcium traces by automated tracking of single somata. Sci Rep 10(1): 7153. PubMed ID: 32346011

How does the concerted activity of neuronal populations shape behavior? In Drosophila melanogaster, the mushroom body (MB) represents an excellent model to analyze sensory coding and memory plasticity. This work presents an experimental setup coupled with a dedicated computational method that provides in vivo measurements of the activity of hundreds of densely packed somata uniformly spread in the MB. This study exploited spinning-disk confocal 3D imaging over time of the whole MB cell body layer in vivo while it is exposed to olfactory stimulation. Importantly, to derive individual signal from densely packed somata, a fully automated image analysis procedure was developed that takes advantage of the specificities of our data. After anisotropy correction, this approach operates a dedicated spot detection and registration over the entire time sequence to transform trajectories to identifiable clusters. This enabled discarding spurious detections and reconstruct missing ones in a robust way. It was demonstrated that this approach outperformed existing methods in this specific context and made possible high-throughput analysis of approximately 500 single somata uniformly spread over the MB in various conditions. Applying this approach, it was found that learned experiences change the population code of odor representations in the MB. After long-term memory (LTM) formation,an increase in responsive somata count and a stable single neuron signal were quantified. It is predicted that this method, which should further enable studying the population pattern of neuronal activity, has the potential to uncover fine details of sensory processing and memory plasticity (Delestro, 2020).

Go to top

Mapping olfactory representation in the Drosophila mushroom body

Jefferis, G. S., et al. (2007). Cell 128(6): 1187-203. PubMed ID: 17382886

The first olfactory relay in the brain contains a spatial map. Olfactory receptor neurons (ORNs) expressing a specific odorant receptor (and therefore having precisely defined olfactory tuning properties) send axon projections to discrete and reproducibly positioned glomeruli in the vertebrate olfactory bulb or insect antennal lobe. In Drosophila, most ORN classes express one specific odorant receptor and send axons to one of ~50 glomerular targets (Jefferis, 2007 and references therein).

Persistent spatial organization deep within the brain is a motif in many sensory systems. For example, adjacent regions of the somatosensory cortex respond to stimuli from neighboring body parts. Does the spatial organization evident in the first olfactory relay also persist at deeper levels? In flies, the branching patterns have been described of the axons of second order projection neurons (PNs, equivalent to vertebrate mitral cells) in higher olfactory centers: the mushroom body (MB) and lateral horn (LH) of the protocerebrum. In the LH axon branching patterns of PNs of the same glomerular class were highly stereotyped across animals, while such stereotypy was less evident in the MB. Several putative output neurons of the LH have been described. Understanding how these neurons integrate olfactory information is a key problem in the neural basis of olfactory perception. In mice, the existence of some spatial organization in higher olfactory centers has been reported by following the targets of 2 of the 1000 ORN classes to the olfactory cortex. The integrative properties of olfactory cortical neurons have also been studied. However, the anatomical basis of this integration remains challenging because of the numerical complexity of the rodent olfactory system (Jefferis, 2007 and references therein).

Neuroanatomy is the foundation of both developmental and functional studies of the brain. In order to understand the development of neuronal wiring, it is necessary to describe the degree of wiring precision across individuals. Similarly, high-resolution neuroanatomy makes predictions about information transfer and transformation, constraining models of neural processing. Two anatomical approaches have been particularly influential in constructing wiring diagrams. The first is exemplified by the classic work of Cajal using the Golgi method. A small fraction of the neurons within a piece of tissue are stained to reveal their dendritic and axonal projection patterns; the information from many specimens is compared and integrated to give a global picture of the circuit. While this approach was enormously successful in defining the basic logic of connectivity, it lacks comprehensiveness and precision: comprehensiveness because only a small fraction of the neuronal elements are used to construct the global picture; precision because integrating information across sample brains has allowed only qualitative comparisons. The second method is a complete reconstruction of all the connections in a small number of specimens through serial electron microscopy. While new EM technologies are under development, traditional serial section transmission EM approaches are so labor intensive that this has only been achieved once - the reconstruction of the nervous system of C. elegans hermaphrodites (Jefferis, 2007 and references therein).

This study describes an approach that has merits of both methods. By combining genetic single-cell labeling with state-of-the-art image registration techniques, comprehensive maps have been produced of the LH and MB, the two higher olfactory centers of Drosophila. Projections of individual neuronal classes with their neighbors can be visualized and directly compared. These three dimensional maps directly demonstrate the spatial stereotypy of input to the LH and MB. Probabilistic synaptic density maps have been devised and used to identify and quantify the organizational principles of these two centers. It was found, for example, that fruit odors and pheromones are represented in distinct compartments of the LH. Finally, postsynaptic neurons of the LH have been characterized at the single-cell level and the density maps have been used to predict connectivity with input PNs. All the raw and derived data and the necessary software tools are available on the project website, providing a resource that will be integrated with future anatomical, physiological and behavioral data to understand the neural basis of olfactory perception in Drosophila (Jefferis, 2007).

Previous studies have revealed aspects of the spatial organization of higher olfactory centers - the MB calyx and the LH. Of particular relevance to the principles of olfactory information processing, single PNs of different classes have highly stereotyped LH projections. Using five Gal4 enhancer trap lines each labeling 1-3 PN classes, it has been found that PNs from 9 glomeruli project to 3 corresponding zones in the MB calyx and LH; MB output neurons integrate information from each of these zones whereas 6 groups of putative LH output neurons maintain the segregation of these 3 zones (Jefferis, 2007).

This study contains several advances over previous approaches: (1) the projection patterns of 11 new PN classes are described at single-cell resolution, qualitatively extending previous results; (2) all single neuron tracings were digitized, and transformed onto a common reference brain; (3) at the single neuron level, the distribution was determined of PN presynaptic terminals in the MB and LH. Fourth and most importantly, combining the above information allowed generation of quantitative synaptic density maps for 35 PN classes, representing 32 of ~50 unique olfactory channels defined by the projection of ORN classes to antennal lobe glomeruli. This allowed decomposition of MB and LH input into individual channels and then the reassemblage for most of the olfactory system, providing a global view of these higher order centers. Lastly, projection patterns are described of three groups of LHNs at single-cell resolution, and predictions are made about their physiological properties based on their potential connectivity with specific PN classes (Jefferis, 2007).

The concentric zonal organization of PN input into the MB calyx was quantitatively confirmed. However, LH organization is more complex and cannot simply be described as zonal, with the exception of the segregation of pheromone projections from the rest of the channels. This is evident from the single neuron projections of many classes that send stereotyped and divergent branches to multiple areas of the LH, as well as the synaptic density maps. Together with the extensive branching of individual LHNs, characterizing the LH as providing relatively little integration across glomeruli is now considered inaccurate (Jefferis, 2007).

Comparing PN branching patterns in the LH and MB suggests that the LH is likely to support more stereotyped integration. This proposal is consistent with the view that the LH mediates innate olfactory behaviors while the MB participates in odor-mediated learning. However, a clear stereotypy of PN terminals in the MB calyx has now been demonstrated. This is likely to explain observations that certain odors can evoke spatially stereotyped activity in MB neurons. Thus the MB calyx and LH receive different levels of stereotyped input that can be integrated by third order coincidence detectors that combine information from different input channels (Jefferis, 2007).

The most striking biological insight obtained from this study is the segregation in the LH between putative pheromone representing PNs and almost all other PNs in the apparently homogeneous LH neuropil. Interestingly, the highest degree of LH volumetric sexual dimorphism that was quantified coincides with the presynaptic terminals of the GABAergic vVA1lm and vDA1 PNs. It is important to note that in addition to the PNs that express the GAL4 driver GH146 that were characterized in this study, there may be other PNs that relay pheromone information from VA1lm and DA1 glomeruli to higher brain centers and contribute to the sexual dimorphism that was found in the LH (Jefferis, 2007).

The convergence of excitatory and inhibitory projections from these putative pheromone representing glomeruli at overlapping or adjacent locations may allow postsynaptic neurons to respond to the presence of a signal that activates these two glomeruli in a particular ratio or to allow signals from these two glomeruli to have opposing effects on LH neurons that initiate particular behaviors. Behaviorally, male flies appear to integrate information both from attractive and inhibitory pheromones produced by other males. Furthermore, new data show that Fru+ Or67d ORNs innervating the DA1 glomerulus detect a male sex pheromone that has a negative effect on other males and a positive effect on females. It is speculated that balanced excitation and inhibition in these pathways may regulate LHNs that contribute to the appropriate behavioral alternative. Sex-specific integration in the lateral horn may underlie sex-specific behaviors (Jefferis, 2007).

The spatial segregation of pheromone representation contrasts with the representation of glomeruli that receive input from ORNs of the basiconic sensilla, which are generally activated by fruit odorants. Many of these PN classes have extensive overlap in their LH synaptic density maps. This property, coupled with the fact that many fruit odorants activate multiple classes of basiconic ORNs, makes the representations of different fruit odorants and natural fruit odors quite overlapping. These data thus support the following principles: olfactory information concerning food has extensive structural intermixing at the LH compared to the glomerular organization of the antennal lobe, but rather discrete channels are retained for pheromones all the way from the sensory periphery to the LH. It is proposed that the LH is globally organized according to biological values rather than chemical nature of the odorant information (Jefferis, 2007).

This finding is reminiscent of the male silkworm moth, Bombyx mori, where PNs from the macroglomeruli representing sex pheromones send axon projections to a discrete area in the lateral protocerebrum defined by a high level of anti-cGMP staining. Spatial segregation of the pheromone representation in higher olfactory centers may therefore be a conserved feature in insects. This segregation is exaggerated into two entirely separate pathways in mammals, where the nasal epithelium and main olfactory bulb process general odorants and some pheromones, while the vomeronasal organ and accessory olfactory bulb are more specific to pheromone sensation. Furthermore, mitral cells originating from the main and accessory olfactory bulbs project to distinct areas of the cortex (Jefferis, 2007).

Having generated a comprehensive and quantitative map of PN input to the LH, a future challenge is to identify and characterize third order LHNs: where are their dendritic fields in the LH, with which PNs do they form synapses, where do they send their axonal outputs, and what are their physiological properties and functions in olfactory behavior? This effort has been started by identification of Gal4 lines labeling neurons with projections in the vicinity of the LH. Three groups of LHN were characterized at single-cell resolution and their potential connectivity with different PN classes was predicted. However this is clearly only a beginning. The widespread distribution of LHN cell bodies and their potential output to different parts of the brain along with the difficulty of identifying large groups of LHNs labeled by new Gal4 enhancer traps suggest that LHNs are heterogeneous genetically, anatomically and, in all likelihood, functionally. One tractable avenue will be to find LHNs that send dendrites to DA1/VA1lm PN target areas and may therefore respond to pheromones and instruct mating behavior. Two LHN groups that were characterized project to this LH region, and single-cell and potential synapse analyses indicate that some of these LHNs may form strong connections with pheromone responsive PN channels. Further characterization of these and other LHNs will bring an understanding the neural circuit basis of olfactory perception and behavior (Jefferis, 2007).

Go to top

A map of olfactory representation in the Drosophila mushroom body

Lin, H. H., Lai, J. S., Chin, A. L., Chen, Y. C. and Chiang, A. S. (2007). Cell 128(6): 1205-17. PubMed ID: 17382887

Neural coding for olfactory sensory stimuli has been mapped near completion in the Drosophila first-order center, but little is known in the higher brain centers. This study reports that the antenna lobe (AL) spatial map is transformed further in the calyx of the mushroom body (MB), an essential olfactory associated learning center, by stereotypic connections with projection neurons (PNs). Kenyon cell (KC) dendrites are segregated into 17 complementary domains according to their neuroblast clonal origins and birth orders. Aligning the PN axonal map with the KC dendritic map and ultrastructural observation suggest a positional ordering such that inputs from the different AL glomeruli have distinct representations in the MB calyx, and these representations might synapse on functionally distinct KCs. These data suggest that olfactory coding at the AL is decoded in the MB and then transferred via distinct lobes to separate higher brain centers (Lin, 2007).

The greatest challenge facing the field of sensory biology at present is to address how sensory coding is represented from the first-order center to the higher brain centers, where neuronal activity must be computed to elicit appropriate behavior responses. This study reports a spatial map of olfactory representations in the MBs of adult Drosophila brains. It was shown that KC dendrites are segregated into 17 complementary domains defined by both clonal origins and birth orders. When viewed from the posterior, PN axonal termini of DL3/D/DA1/VA1d, of DM2, and of DM1/VA4 form three concentric zones corresponding to KC dendrites from early α/β, late α/β, and α'/β' neurons in the posterior calyx, respectively. The spatial organization of PN-to-KC connectivity suggests that olfactory coding in the AL is maintained in the MB calyx where signal processing is more versatile (Lin, 2007).

One question that now can be addressed is, are odorants carrying similar biological information processed by the same class of KCs? A chemotopic map of OR responses to 110 odorants indicates that excitatory and inhibitory responses are chemical-class dependent. By integrating the PN-to-KC map with existing electrophysiological data, it was observe that KC responses to chemicals are likely also class specific. For example, many aromatic odorants are excitatory to Or10a and many terpenes are inhibitory to Or49b, which connects via DL1 and VA5, respectively, to γ neurons. Since MBs are essential for odor discrimination, it is predicted that a fly will find it more difficult to discriminate two odorants that are relayed by the PNs to the same class(es) of KCs. It has been observed that odors of a particular chemical class are often clustered, as shown by the 'odor space' constructed from the response of 24 ORs to 110 odorants. It was shown that three odorants in different chemical classes mapped to three distinct points in this space: pentyl acetate (an ester) and 2-hepatanone (a ketone) elicited similar patterns of activation maps together, distant and different from that elicited by methyl salicylate (an aromatic compound). Consistently in the anatomical study, it was found that pentyl acetate- and 2-hepatanone-induced signals are likely processed by the same class of KCs (i.e., late α/β neurons), while those induced by methyl salicylate reach other classes of KCs, excluding late α/β neurons. It is noted that this is a working model of odor discrimination, since it does not yet include all PN-to-KC connections in the adult olfactory system and only six ORs have yet been linked with the PN-to-KC map. Additional functional subsets of KCs are expected in each of the five classes. While a more complete PN-to-KC map is needed, the model of odor space versus KC class provides an important advance in the understanding of neural computation underlying behavioral responses to odors (Lin, 2007).

The findings are in congruence with functional imaging studies, which indicate that odor-evoked activity occurs in specific regions in the calyx. As odor concentration increases, more glomeruli are activated in the AL and more KCs are activated in the calyx. The anatomical data suggest that perception of odor identity may require integration among five classes of KCs, while the number of responsive KCs may reflect the perception of odor intensity (Lin, 2007).

Confocal imaging of specific GAL4-driven reporter expression patterns reveals axonal segregation for each of the five KC classes. This topology implies that stereotyped olfactory representation in the AL glomeruli received from OSNs (first order) are further relayed by PNs (second order) to a fixed combination of five KC classes (third order) in the calyx; such an implication is supported by findings on connectivity. It is surmised that information processing is further achieved by segregating axon bundles to different MB lobes, where output neurons (fourth order) diverge the processed information to separate higher brain centers (fifth order). Incomplete as this model may be, the possibility is acknowledged of (1) crosstalk among KCs and (2) modulatory innervation of MB calyx and lobes. Because this study was focused only on a subset of PNs, the possibility cannot be ruled out that further functionally important differences in PNs may yet be discerned. Nonetheless, this anatomical model serves to advance the notion that an olfactory map in the MBs helps to guide olfactory-driven behaviors (Lin, 2007).

All animals are born with a set of innate behavioral responses, 'hardwired' in the nervous system. In Drosophila, innate behaviors such as sleep and courtship require proper functioning of the MBs. It is notable that Fruitless, a transcription factor required for male courtship behavior, is expressed in OSNs and PNs innervating the same set of AL glomeruli (VL2a, DA1, and VA1), suggesting interconnections between these two sets of olfactory neurons. Intriguingly, fruitless expresses also in the MBs of the ? and α/β lobes, and courtship conditioning is impaired when expression of the male-specific fru transcript is disrupted in MB γ neurons. In likelihood connectivity assignment, VL2a- and DA1-PNs connect with γ and early α/β lobes. It is possible that the fru-expressing PNs and KCs also are interconnected in the MB calyx, as are the fru-expressing OSNs and PNs in the AL. These data suggest that stereotyped connectivity in the PN-to-KC map is likely involved in fru-expressing circuits, which are essential for proper behavioral responses to volatile sex pheromones (Lin, 2007).

A central question in olfaction is how the brain discriminates different odors to elicit an appropriate behavioral response. Stereotypic connectivity maps of odorant-to-OR, OSN-to-PN, and PN-to-KC at three consecutive levels allow further construction of a neural computation of odor discrimination in the adult Drosophila brain. Stereotypic PN-to-KC connectivity and functional imaging suggest differential representation of the odors in the AL is maintained in the MB calyx and possibly further processed in the different MB neurons/lobes. If so, how does the same class of KCs discriminate odorants carrying different biological information, such as a sex pheromone and an aggregation pheromone? A single class of KCs might be sufficient to discriminate two different odors in some cases, since Drosophila larvae can discriminate different odors with only γ neurons. Thus, additional spatial and/or temporal complexity for neural computation must exist among KCs of the same birth-order class. Consistent with this notion, the data show that PNs connecting with the same class of KCs may have different projecting patterns among K1-K5 dendritic divisions, suggesting differential functions for each of the four KC clones. Even with the same developmental history of clonal origin and birth order, KCs are likely divided into different identities further based on differential gene expression. For example, Gal4 line G0050 labels the entire α'/β' lobe but c305a labels only the frontal-half α'/β' lobe. Even with such developmental specification, odor discrimination also may require additional integration among different classes of KCs (Lin, 2007).

Although stereotypic connectivity maps from ORNs to PNs to KCs give the impression of a straight and simple path, olfactory coding clearly will be modulated by both stimulatory and inhibitory signals as it makes its way through the brain. A single ORN can exhibit both excitatory and inhibitory responses to different odorants. In the ALs, odor responses of the PNs are reshaped by inhibition from local neurons. In the MBs, KCs may receive both stimulatory and inhibitory stimuli from PNs, since most of them are cholinergic but some of them are GABAergic. Immunohistochemical labeling and GFP expression patterns in Cha-GAL4 and GAD-GAL4 lines indicate that KCs are also composed of both cholinergic and GABAergic neurons. The distribution of odor responses across different classes of KCs and the imposition of odor-sensitive excitatory and inhibitory responses both appear to enhance distinct neural representations of different odors. Such complexity of odor representations greatly reduces the possibility of overlap between spatiotemporal patterns elicited by two different odorants, making them easier to discriminate or to memorize and recall (Lin, 2007).

In conclusion, the data offer specific and testable hypotheses that olfactory coding at the ALs is likely further represented and decoded in the MBs and then transferred via distinct lobes to separate higher brain centers. It would be important now to complete the PN-to-KC map, to identify further subclasses within each of the five KC classes, and to answer how different classes of KCs communicate with each other during olfactory neural computation (Lin, 2007).

Go to top

Shared mushroom body circuits underlie visual and olfactory memories in Drosophila

Vogt, K., Schnaitmann, C., Dylla, K. V., Knapek, S., Aso, Y., Rubin, G. M. and Tanimoto, H. (2014). Elife 3: e02395. PubMed ID: 25139953

In nature, animals form memories associating reward or punishment with stimuli from different sensory modalities, such as smells and colors. It is unclear, however, how distinct sensory memories are processed in the brain. This study established appetitive and aversive visual learning assays for Drosophila that are comparable to the widely used olfactory learning assays. These assays share critical features, such as reinforcing stimuli (sugar reward and electric shock punishment), and allow direct comparison of the cellular requirements for visual and olfactory memories. It was found that the same subsets of dopamine neurons drive formation of both sensory memories. Furthermore, distinct yet partially overlapping subsets of mushroom body intrinsic neurons are required for visual and olfactory memories. Thus, these results suggest that distinct sensory memories are processed in a common brain center. Such centralization of related brain functions is an economical design that avoids the repetition of similar circuit motifs (Vogt, 2014).

Devising a transparent electric shock grid module made it possible to apply the same visual stimulation in aversive and appetitive conditioning assays. Also an integrated platform was developed for fully automated high-throughput data acquisition using customized software to control the presentation of electric shock and visual stimuli while making video recordings of behavior. In these assays, memory performance is based on altered visual preference in walking flies, a task likely to be less demanding than the constant flight required for flight simulator learning. These advantages facilitate behavioral examination of many genotypes (Vogt, 2014).

Circuits underlying olfactory and visual memory can be optimally compared when the sugar reward and electric shock punishment are matched between the two modalities. Visual and olfactory memories share the same subsets of dopamine neurons that convey reinforcing signals. This shared requirement of the transmitter system between visual and olfactory learning has been described in crickets. However, the pharmacological manipulation used in these studies does not allow further circuit dissection (Vogt, 2014).

For electric shock reinforcement, identified neurons in the PPL1 cluster, such as MB-MP1, MB-MV1 and MB-V1, drive aversive memories in both visual and olfactory learning, while the MB-M3 neurons in the PAM cluster seem to be involved specifically in aversive olfactory memory. Thus, overlapping sets of dopamine neurons appear to represent electric shock punishment in both visual and olfactory learning with olfactory aversive memory probably recruiting a larger set. Previous studies have shown that the MB-M3 neurons induce aversive olfactory memory that increases stability of other memory components. Olfactory memories last longer than visual memories potentially due to the recruitment of additional dopamine neurons (Vogt, 2014).

In appetitive conditioning, PAM cluster neurons play crucial roles in both olfactory and visual memories. Which cell types in these clusters are involved and whether there is a cellular distinction between olfactory and visual memory requires further analysis at the single cell level. Most importantly, all these neurons convey dopamine signals to restricted subdomains of the MB. The blockade of octopamine neurons did not impair appetitive visual memories with sucrose. The involvement of octopamine neurons may be more substantial when non-nutritious sweet taste rewards are used, as has been shown in olfactory learning (Vogt, 2014).

In addition to these shared reinforcement circuits in the MB, the necessity of MB output for visual memory acquisition and retrieval is also consistent with olfactory conditioning. Taken together, these results suggest that the MBs harbor associative plasticity for visual memories and support the conclusion that similar coincidence detection mechanisms are used to form memories within the MBs. Centralization of similar brain functions spares the cost of maintaining similar circuit motifs in different brain areas and may be an evolutionary conserved design of information processing. Such converging inputs of different stimuli into a multisensory area have even been described in humans (Vogt, 2014).

'Flight simulator' visual learning was shown to require the central complex but not the MBs. Although this appears to contradict the current study, it is noted that there are important differences between the behavioral paradigms employed. In the flight simulator, a single tethered flying Drosophila is trained to associate a specific visual cue with a laser beam punishment, to later on avoid flying towards this cue in the test. Although this study controlled for visual context consistency and the 'operant component' of the flight simulator training, any other difference could account for the differential requirement of brain structures. Given that flies during flight show octopamine-mediated modulation of neurons in the optic lobe, similar state-dependent mechanisms might underlie different requirement of higher brain centers. Thus, it is critical to design comparable memory paradigms (Vogt, 2014).

This study together with previous results in associative taste learning highlights the fact that the role of the MB in associative learning is not restricted to one sensory modality or reinforcer. This study found that olfactory and visual memories recruit overlapping, yet partly distinct, sets of Kenyon cells (see Circuit model of olfactory and visual short-term memories). In contrast to the well-described olfactory projection neurons, visual inputs to the MB remain unidentified. No anatomical evidence has been reported in Drosophila for direct connections between optic lobes and MBs although such connections are found in other insects. Also afferents originating in the protocerebrum were found to provide multi-modal input to the MB lobes of cockroaches. Thus, Drosophila MBs may receive indirect visual input from optic lobes, and the identification of such a visual pathway would significantly contribute to understanding of the MB circuits (Vogt, 2014).

Given the general requirement of the γ lobe neurons, visual and olfactory cues may be both represented in the γ neurons. Consistently, the dopamine neurons that convey appetitive and aversive memories heavily project to the γ lobe. In olfactory conditioning, the γ lobe was shown to contribute mainly to short-term memory. This converging evidence from olfactory and visual memories suggests a general role for the γ lobe in short-lasting memories across different sensory modalities. Previous studies found that the MB is also involved in sensorimotor gating of visual stimuli or visual selective attention. Therefore, the MB circuits for visual associative memories might be required for sensorimotor gating and attention (Vogt, 2014).

Interestingly, the contribution of the α'/β' lobes is selective for olfactory memories. This Kenyon cell class is more specialized to odor representation, as the cells have the broadest odor tuning and the lowest response threshold among the three Kenyon cell types (Vogt, 2014).

The role of α/β neurons in visual memories is also limited. The α/β neurons might play more modulatory roles in specific visual memory tasks, such as context generalization, facilitation of operant learning and occasion setting. This modulatory role of the α/β neurons is corroborated in olfactory learning, where they are preferentially required for long-lasting memories (Vogt, 2014).

Differentiated but overlapping sensory representations by KCs may be conserved among insect species. In honeybees, different sensory modalities are represented in spatially segregated areas of the calyx, whereas the basal ring region receives visual and olfactory inputs. The MB might thus have evolved to represent the sensory space of those modalities that are subject to associative modulation (Vogt, 2014).

Go to top

Visual input into the Drosophila melanogaster mushroom body

Li, J., Mahoney, B. D., Jacob, M. S. and Caron, S. J. C. (2020). Cell Rep 32(11): 108138. PubMed ID: 32937130

Visual input into the Drosophila melanogaster mushroom body

The patterns of neuronal connectivity underlying multisensory integration, a fundamental property of many brains, remain poorly characterized. The Drosophila melanogaster mushroom body-an associative center-is an ideal system to investigate how different sensory channels converge in higher order brain centers. The neurons connecting the mushroom body to the olfactory system have been described in great detail, but input from other sensory systems remains poorly defined. This study used a range of anatomical and genetic techniques to identify two types of input neurons that connect visual processing centers-the lobula and the posterior lateral protocerebrum-to the dorsal accessory calyx of the mushroom body. Together with previous work that described a pathway conveying visual information from the medulla to the ventral accessory calyx of the mushroom body, this study defines a second, parallel pathway that is anatomically poised to convey information from the visual system to the dorsal accessory calyx (Li, 2020).

Sensory systems use different strategies to detect specific physical features of the outside world. For instance, the olfactory system contains many different types of sensory neuron that are each specialized in detecting a specific class of volatile chemicals. Through only two neuronal layers, olfactory information-the identity of an odor and its concentration-is relayed to higher brain centers. In contrast, the visual system contains far fewer types of sensory neuron, but through numerous neuronal layers, it relays a range of highly processed information-for instance, color, brightness, motion, and shape-to higher brain centers. Thus, higher brain centers have to integrate different types of processed information, bind that information into a coherent representation of the outside world, and use such representations to guide behavior. How higher brain centers achieve this feat remains largely unknown. This gap in knowledge mainly stems from the fact that higher brain centers are formed by a large number of neurons and that the projection neurons conveying information from different sensory systems to these centers often remain poorly characterized. This makes it difficult to understand whether there are specific patterns of neuronal connectivity that enable multisensory integration and what the nature of these patterns are. Deciphering the fundamental neuronal mechanisms that underlie multisensory integration requires a model system such as the Drosophila melanogaster mushroom body, which consists of a relatively small number of neurons whose connections can be charted reliably (Li, 2020).

The Drosophila mushroom body is formed by ∼2,000 neurons-called the Kenyon cells-and has long been studied for its essential role in the formation of olfactory associative memories. The identity of the projection neurons that connect the olfactory system to the mushroom body-and the way Kenyon cells integrate input from these neurons-has been characterized in great detail, highlighting fundamental connectivity patterns that enable this higher brain center to represent olfactory information efficiently. Evidence in Drosophila melanogaster shows that the mushroom body is more than an olfactory center, as it is also required for the formation of visual and gustatory associative memories. However, the identity of the neurons that connect the mushroom body to other sensory systems remains poorly characterized. Thus, a first step toward understanding how the mushroom body integrates multisensory information is to identify such non-olfactory mushroom body input neurons and the genetic tools necessary to manipulate these neurons (Li, 2020).

The mushroom body receives its input through its calyx and sends its output through its lobes. The calyx-a morphologically distinct neuropil containing the synapses formed between projection neurons and Kenyon cells-can be divided into four, non-overlapping regions: one main calyx as well as three accessory calyces named the dorsal, lateral, and ventral accessory calyces. The five output lobes-the α, α', β, β', and γ lobes-contain the synapses formed between Kenyon cells, mushroom body output neurons, and dopaminergic neurons. With respect to these input and output regions, Kenyon cells can be divided into seven distinct types. Of these seven types, five types-the α/βc, α/βs, α'/β'ap, α'/β'm, and γmain Kenyon cells-extend their dendrites only into the main calyx and their axons along one or two lobes. Most of the neurons that project to the main calyx emerge from the antennal lobe, the primary olfactory center in the Drosophila brain. Thus, α/βc, α/βs, α'/β'ap, α'/β'm, and γmain Kenyon cells receive input primarily from the olfactory system (Li, 2020).

In contrast, the two other classes of Kenyon cells do not extend their dendrites into the main calyx. Instead, the α/βp Kenyon cells extend their dendrites into the dorsal accessory calyx-avoiding completely the main, lateral, and ventral accessory calyces-and their axons along the α and β lobes. Likewise, the γd Kenyon cells extend their dendrites exclusively into the ventral accessory calyx and their axons along the γ lobe. Thus, both the α/βp and γd Kenyon cells are anatomically poised to receive non-olfactory input. There is evidence suggesting that the ventral accessory calyx receives input from the medulla, a region of the optic lobe that specializes in processing brightness and color. Furthermore, a recent study suggests that the dorsal accessory calyx is a multisensory center that integrates input from multiple sensory pathways, including the olfactory, gustatory, and visual systems (Li, 2020).

This study reports a strategy that uses a combination of genetic tools-including transgenic lines that drive expression in few neurons and a photo-labeling technique used to identify individual neurons and their pre-synaptic partners-to characterize the input neurons of the α/βp Kenyon cells. Two types of mushroom body input neuron were identified in that, together, form about half of the total input the α/βp Kenyon cells receive in the dorsal accessory calyx. The first neuronal type-henceforth referred to as LOPNs-consists of a neuron that projects from the lobula, a region of the optic lobe specialized in detecting visual features, such as shape and motion. The second type of neuron-henceforth referred to as PLPPNs-consists of projection neurons that emerge from the posterior lateral protocerebrum, a brain region that receives input from the optic lobe. Interestingly, LOPN and PLPPNs do not project to the ventral accessory calyx and do not connect to the γd Kenyon cells. Based on these findings, it is concluded that there are two parallel pathways that convey visual information to the mushroom body: a pathway projecting from the medulla to the γd Kenyon cells and another pathway projecting from the lobula and posterior lateral protocerebrum to the α/βp Kenyon cells (Li, 2020).

This study has identified and characterized neurons projecting to the dorsal accessory calyx of the mushroom body and show that these neurons are pre-synaptic to the α/βp Kenyon cells. Using a combination of genetic and anatomical techniques, it was possible to distinguish two different types of projection neuron: LOPN projecting from the lobula-an area of the optic lobe processing visual features, such as shape and motion-and the PLPPNs projecting from the posterior lateral protocerebrum. Although the posterior lateral protocerebrum remains poorly characterized in D. melanogaster, evidence from other insects shows that this brain region receives input from the optic lobe. Interestingly, it was found that the dendrites formed by the PLPPNs in the posterior lateral protocerebrum are in close proximity to neurons that project from the ventral medulla. Based on these results-and considering insights from the connectome-it is estimated that LOPNs and PLPPNs account for half of total input that α/βp Kenyon cells receive in the dorsal accessory calyx. LOPNs and PLPPNs do not extend axonal terminals into the ventral accessory calyx, the other calyx known to receive visual input, but rather extend axonal terminals into the dorsal accessory calyx and into the superior lateral protocerebrum. Likewise, the α/βp Kenyon cells do not connect to the visual projection neurons that are associated with the ventral accessory calyx. These findings suggest that the visual system is connected to the mushroom body via two parallel pathways: the α/βp Kenyon cells receive input from the lobula and the posterior lateral protocerebrum, whereas the γd Kenyon cells receive input directly from the medulla. Further functional studies are necessary to determine what kind of visual information is processed by the α/βp Kenyon cells (Li, 2020).

In Drosophila melanogaster, the mushroom body has long been studied as an olfactory processing center. However, evidence from many insects, including the honeybee Apis mellifera, shows that the mushroom body integrates sensory information across different modalities. In honeybees, the input region of the mushroom body, also called the calyx, is divided into different layers, and each layer receives input from either the olfactory or visual system. Because the dendrites of Kenyon cells are also restricted to specific layers, it has been suggested that, in the honeybee, multisensory integration does not occur at the level of individual Kenyon cells but rather at the population level. Although the honeybee mushroom body differs greatly from the Drosophila mushroom body-it contains about a hundred times as many Kenyon cells and its input region is divided in multiple complex layers-it appears that both mushroom bodies share a common fundamental connectivity principle: the segregation of input based on sensory modality. This connectivity mechanism is immediately apparent in the structural organization of the Drosophila melanogaster mushroom body: the Kenyon cells receiving input from the olfactory system all extend their dendrites into the main calyx, whereas the Kenyon cells receiving input from the visual system extend their dendrites either in the dorsal accessory calyx or the ventral accessory calyx. Many studies have demonstrated that the Kenyon cells that process olfactory information-those associated with the main calyx-integrate input broadly across the different types of olfactory projection neuron. Interestingly, it appears that the Kenyon cells that process visual information are wired differently (Li, 2020).

A thorough understanding is available of how olfactory Kenyon cells integrate input from the antennal lobe: most Kenyon cells receive, on average, input from seven projection neurons, and the projection neurons connecting to the same Kenyon cell share no apparent common features. Theoretical studies have shown that this random-like connectivity pattern enables the mushroom body to form sparse and decorrelated odor representations and thus maximizes learning. Randomization of sensory input is a connectivity pattern that is well suited for representing olfactory information-as an odor is encoded based on the ensemble of olfactory receptors it activates-and might not be suitable for representing visual information. Indeed, the results of this study suggest that specific visual features-the signals processed by the medulla and the ones processed by the lobula and the posterior lateral protocerebrum-need to be represented by two separate subpopulations of Kenyon cells. This observation mirrors anatomical studies of the honeybee brain: the neurons projecting from the lobula terminate in a different layer than the neurons projecting from the medulla. This arrangement might be essential to preserve distinct visual features when forming associative memories. Functional and behavioral studies are required to determine whether indeed the mushroom body represents multisensory stimuli in this manner (Li, 2020).

Go to top

The neuronal architecture of the mushroom body provides a logic for associative learning

Aso, Y., Hattori, D., Yu, Y., Johnston, R. M., Iyer, N. A., Ngo, T. T., Dionne, H., Abbott, L., Axel, R., Tanimoto, H. and Rubin, G. M. (2014b). Elife 3. PubMed ID: 25535793.

Animals discriminate stimuli, learn their predictive value and use this knowledge to modify their behavior. In Drosophila, the mushroom body (MB) plays a key role in these processes. Sensory stimuli are sparsely represented by approximately 2000 Kenyon cells, which converge onto 34 output neurons (MBONs) of 21 types (see Circuit diagrams of the mushroom body). The role of MBONs were studied in several associative learning tasks and in sleep regulation, revealing the extent to which information flow is segregated into distinct channels and suggesting possible roles for the multi-layered MBON network. Optogenetic activation of MBONs can, depending on cell type, induce repulsion or attraction in flies. The behavioral effects of MBON perturbation are combinatorial, suggesting that the MBON ensemble collectively provides an adaptive mechanism to assign valence-positive or negative survival value-to a sensory stimulus then store that information, and recall it when that same stimulus is encountered again. It is proposed that local, stimulus-specific dopaminergic modulation selectively alters the balance within the MBON network for those stimuli. These results suggest that valence encoded by the MBON ensemble biases memory-based action selection (Aso, 2014b: PubMed).

To survive in a dynamic environment, an animal must discover and remember the outcomes associated with the stimuli it encounters. It then needs to choose adaptive behaviors, such as approaching cues that predict food and avoiding cues that predict danger. The neural computations involved in using such memory-based valuation of sensory cues to guide action selection require at least three processes: (1) sensory processing to represent the identity of environmental stimuli and distinguish among them; (2) an adaptive mechanism to assign valence-positive or negative survival value-to a sensory stimulus, store that information, and recall it when that same stimulus is encountered again; and (3) decision mechanisms that receive and integrate information about the valence of learned stimuli and then bias behavioral output. To understand such decision-making processes, one approach is to locate the sites of synaptic plasticity underlying memory formation, identify the postsynaptic neurons that transmit stored information to the downstream circuit and discover how their altered activities bias behavior (Aso, 2014b).

The mushroom body (MB) is the main center of associative memory in insect brains. While the MB processes several modalities of sensory information and regulates locomotion and sleep, MB function has been most extensively studied in the context of olfactory memory-specifically, associating olfactory stimuli with environmental conditions in order to guide behavior. In Drosophila, olfactory information is delivered to the MB by projection neurons from each of ~50 antennal lobe glomeruli. Connections between the projection neurons and the ~2000 Kenyon cells (KCs), the neurons whose parallel axonal fibers form the MB lobes, are not stereotyped; that is, individual flies show distinct wiring patterns between projection neurons and KCs. Sparse activity of the KCs represents the identity of odors. The output of the MB is conveyed to the rest of the brain by a remarkably small number of neurons-34 cells of 21 cell types per brain hemisphere (Aso, 2014b).

The information flow from the KCs to the MB output neurons (MBONs) has been proposed to transform the representation of odor identity to more abstract information, such as the valence of an odor based on prior experience (see discussion in Aso, 2014a). In contrast to KCs, MBONs have broadly tuned odor responses; any given odor results in a response in most MBONs, although the magnitude of the response varies among MBON cell types. Unlike the stereotyped response to odors of the olfactory projection neurons that deliver odor information to the MB, the odor tuning of the MBONs is modified by plasticity and varies significantly between individual flies, suggesting that MBONs change their response to odors based on experience (Aso, 2014b).

For olfactory associative memory in Drosophila, multiple lines of evidence are consistent with a model in which dopamine-dependent plasticity in the presynaptic terminals of KCs alters the strength of synapses onto MBON dendrites. This is thought to provide a mechanism by which the response of MBONs to a specific odor could represent that odor's predictive value. D1-like dopamine receptors and components of the cAMP signaling pathway, such as the Ca2+/Calmodulin-responsive adenylate cyclase encoded by the rutabaga gene, are required specifically in the KCs for memory formatio and rutabaga was shown to be required for the establishment of the differences in MBON odor tuning between individuals. Reward and punishment recruit distinct sets of dopaminergic neurons (DANs) that project to specific regions in the MB lobes. Moreover, exogenous activation of these DANs can substitute for reinforcing stimuli to induce either appetitive or aversive memory, depending on DAN cell type. In sum, while the identity of the learned odor is likely encoded by the small subset of KCs activated by that odor, whether dopamine-mediated modulation assigns positive or negative valence to that odor would be determined by where in the MB lobes KC-MBON synapses are modulated and thus which MBON cell types alter their response to the learned odor (Aso, 2014ab).

Combining the above observations with the comprehensive anatomical characterization of MB inputs and outputs lays the groundwork for testing models of how the MB functions as a whole. It is suggested that each of the 15 MB compartments-regions along the MB lobes defined by the arborization patterns of MBONs and DANs (see Circuit diagrams of the mushroom body)-functions as an elemental valuation system that receives reward or punishment signals and translates the pattern of KC activity to a MBON output that serves to bias behavior by altering either attraction or aversion. This view implies that multiple independent valuation modules for positive or negative experiences coexist in the MB lobes, raising the question of how the outputs across all the modules are integrated to result in a coherent, adaptive biasing of behavior (Aso, 2014b).

Although several MBON cell types have been shown to play a role in associative odor memory, the functions of most MBONs have not been studied. Based on anatomical analyses (Aso, 2014a), it is believed that just 34 MBONs of 21 types provide the sole output pathways from the MB lobes. To gain mechanistic insight into how the ensemble of MBONs biases behavior, it would be first important to know the nature of the information conveyed by individual MBONs and the extent to which their functions are specialized or segregated into different information channels. Then it would be necessary to discover how the activities of individual MBONs contribute to influence the behavior exerted by the complete population of MBONs. Thus, in order to understand how memory is translated into changes in behavior, experimental access would be needed to a comprehensive set of MBONs and investigate how the outputs from different MBONs bias behavior, singly and in combination (Aso, 2014b).

In the accompanying paper (Aso, 2014a), the detailed anatomy is described of the DANs and MBONs (see Circuit diagrams of the mushroom body and the generation of intersectional split-GAL4 driver lines to facilitate their study. All but one of the 21 MBON cell types consists of only one or two cells per hemisphere. Dendrites of MBONs that use the same neurotransmitter-GABA, glutamate or acetylcholine-are spatially clustered in the MB lobes. Intriguingly, this spatial clustering resembles the innervation patterns of modulatory input by two clusters of dopaminergic neurons, PPL1 and PAM. MBONs have their axonal terminals in a small number of brain regions, but their projection patterns also suggest pathways for relaying signals between compartments of the MB lobes; three MBONs send direct projections to the MB lobes and several other MBONs appear to target the dendrites of specific DANs (Aso, 2014b).

Split-GAL4 drivers give the capability to express genetically encoded effectors in identified MBONs to modify their function. This study examine the roles of specific MBONs in various learning and memory tasks as well as in the regulation of locomotion and sleep. Whether direct activation of specific MBONs are sufficient to elicit approach or avoidance was also studied. The results indicate that the ensemble of MBONs does not directly specify particular motor patterns. Instead, MBONs collectively bias behavior by conveying the valence of learned sensory stimuli, irrespective of the modality of the stimulus or the specific reward or punishment used during conditioning (Aso, 2014b).

In the insect brain, a sparse and non-stereotyped ensemble of Kenyon cells represents environmental cues such as odors. The behavioral response to these cues can be neutral, repulsive or attractive, influenced by the prior experiences of that individual and how dopaminergic and other modulatory inputs have changed the weight of its KC-MBON synapses. MBONs are thought to encode the predictive value associated with a stimulus. A fly would then use that information to bias its selection of behavioral responses in an ever-changing environment. The accompanying paper (Aso, 2014), describes in detail the projection patterns of the MBON and DAN cell types that comprise the MB lobes in Drosophila. This report begins the process of determining the nature of the information conveyed by MBONs. Specific MBON cell types were correlated with roles in associative memory and sleep regulation (see A map of MBON functions). These anatomical and behavioral results lay the groundwork for understanding the circuit principles for a system of memory-based valuation and action selection (Aso, 2014b).

Optogenetic activation of MBONs in untrained flies can induce approach or avoidance. The ability of the MBONs to induce changes in behavior in the absence of odors suggests that MBONs can bias behavior directly. This observation is consistent with a recent study showing that flies are able to associate artificial activation of a random set of KCs-instead of an odor stimulus-with electric shock, and avoid reactivation of the same set of KCs in the absence of odors (Vasmer, 2014), a result that recapitulates a finding in the potentially analogous piriform cortex of rodents. This study found that the sign of the response to MBON activation was highly correlated with neurotransmitter type; all the MBONs whose activation resulted in avoidance were glutamatergic, whereas all the attractive MBONs were cholinergic or GABAergic (Aso, 2014b).

By tracking flies as they encounter a border between darkness and CsChrimson-activating light, activation of an MBON was shown to bias walking direction. Although activation of glutamatergic MBONs repelled flies, the avoidance behaviors were not stereotyped; flies showed a variety of motor patterns when avoiding the red light. This observation implies that MBONs are unlikely to function as command neurons to drive a specific motor pattern, as has been observed, for example, in recently identified descending neurons that induce only stereotyped backward walking (Aso, 2014b).

Rather, fly locomotion can be considered as a goal-directed system that uses changes in MBON activity as an internal guide for taxis. For example, walking in a direction that increases the relative activity of aversive-encoding MBONs, which would occur as a fly approaches an odorant it had previously learned to associate with punishment (or when a fly expressing CsChrimson in an avoidance-inducing MBON approaches the CsChrimson activating light), signals the locomotive system to turn around and walk the other direction. Detailed studies of locomotor circuitry will be required to determine the mechanisms of executing such taxic behaviors and should help elucidate how MBON inputs guide this system (Aso, 2014b).

In this view, the MBON population functions as neither a purely motor nor a purely sensory signal. From the motor perspective, as described above, MBONs bias locomotive outcomes rather than dictate a stereotyped low-level motor program. From the sensory perspective, this study has shown that the same MBONs can be required for experience-dependent behavioral plasticity irrespective of whether a conditioned stimulus is a color or an odor, and irrespective of the specific identity of the odor. Taken together with the fact that MBONs lie immediately downstream of the sites of memory formation, these observations support the proposal that MBONs convey that a stimulus has a particular value-but not the identity of the stimulus itself. This contrasts with sensory neurons whose activity can also induce approach or avoidance, but which do convey the stimulus per se. In mammals, neural representations of abstract variables such as 'value', 'risk' and 'confidence' are thought to participate in cognition leading to action selection. From the point of view of this framework, the MBON population representing the value of a learned stimulus and informing locomotion might be operationally viewed as a cognitive primitive (Aso, 2014b).

Co-activating multiple MBON cell types revealed that the effects of activating different MBONs appear to be additive; that is, activating MBONs with the same sign of action increases the strength of the behavioral response, whereas activating MBONs of opposite sign reduces the behavioral response. Thus, groups of MBONs, rather than individual MBONs, likely act collectively to bias behavioral responses. Consistent with the idea of a distributed MBON population code, all 19 MBON cell types imaged so far show a calcium response to any given odor (Aso, 2014b).

If it is the ensemble activity of a large number of MBONs that determines memory-guided behavior, how can local modulation of only one or a few MB compartments by dopamine lead to a strong behavioral response? Activation of a single DAN such as PPL1-γ1pedc that innervates a highly localized region of the MB can induce robust aversive memory, yet the odor associated with the punishment will activate MBONs from all compartments, including MBONs that can drive approach as well as those that drive avoidance. It is proposed that, in response to a novel odor stimulus, the activities of MBONs representing opposing valences may initially be 'balanced', so that they do not impose a significant bias. Behavior would then be governed simply by any innate preference a fly might have to that odor, using neuronal circuits not involving the MB. Now suppose an outcome associated with that stimulus is learned. Such learning involves compartment-specific, dopamine-dependent plasticity of the KC-MBON synapses activated by that stimulus. If that occurs, the subsequent ensemble response of the MBONs to that stimulus would no longer be in balance and an attraction to, or avoidance of, that stimulus would result. Consistent with this idea, eliminating MB function by disrupting KCs, which are nearly 10% of neurons in the central brain, had surprisingly minor effects on odor preference (Aso, 2014b).

Recent studies of dopamine signaling have implicated distinct sites of memory formation within the MB lobes. Consistent with this large body of work, this study found that one type of PPL1 cluster DAN, PPL1-γ1pedc, played a central role in formation of aversive memory in both olfactory and visual learning paradigms. This DAN also mediates aversive reinforcement of bitter taste. For appetitive memory, PAM cluster DANs that innervate other regions of the MB lobes, in particular the compartments of glutamatergic MBONs, are sufficient to induce appetitive memory. These results strongly suggest that the synaptic plasticity underlying appetitive and aversive memory generally occurs in different compartments of the MB lobes (Aso, 2014b).

The sign of preference observed in response to CsChrimson activation of particular MBONs was, in general, opposite to that of the memory induced by dopaminergic input to the corresponding MB compartments. For example, activation of MBON-γ1pedc>α/β and MBON-γ2α'1 attracted flies, whereas DAN input to these regions induced aversive memory. Conversely, activation of glutamatergic MBONs repelled flies, while DAN input to the corresponding regions is known to induce appetitive memory. These results are most easily explained if dopamine modulation led to synaptic depression of the outputs of the KCs representing the CS + stimulus. Consistent with this mechanism, the PE1 MBONs in honeybees as well as the V2 cluster MBONs in Drosophila reduce their response to a learned odor and depression of KC-MBON synapses has been shown for octopamine modulation in the locust MB. Moreover, long-term synaptic depression is known to occur in the granular cell synapses to Purkinje cells in the vertebrate cerebellum, a local neuronal circuit with many analogies to the MB. Other mechanisms are also possible and multiple mechanisms are likely to be used. For example, dopamine may modulate terminals of KCs to potentiate release of an inhibitory cotransmitter such as short neuropeptide F, which has been demonstrated to be functional in KCs and hyperpolarizes cells expressing the sNPF receptor. RNA profiling of MBONs should provide insights into the molecular composition of synapses between KCs and MBONs. It is also noteworthy that the effect of dopamine can be dependent on the activity status of Kenyon cells; activation of PPL1-γ1pedc together with odor presentation induces memory, while its activation without an odor has been reported to erase memory. In the vertebrate basal ganglia, dopamine dependent synaptic plasticity important for aversive and appetitive learning is known to result in both synaptic potentiation and synaptic depression (Aso, 2014b).

This study sought the effects of selectively and specifically manipulating the activities of a comprehensive set of MBONs on several behaviors. As a consequence, some insights were gained into the extent to which the relative importance of particular MBONs differed between behaviors. Most obvious was the segregation between appetitive and aversive behaviors. For example, this study found that blocking MBON-γ1pedc>α/β impaired both short-term aversive odor and visual memory, suggesting a general role in aversive memory independent of modality. Conversely, a subset of glutamatergic MBONs was required in all three appetitive memory assays. It still remains to be demonstrated that the outputs of these MBONs are required transiently during memory retrieval. Nevertheless, CsChrimson activation experiments demonstrate that activation of these MBONs can directly and transiently induce attraction and avoidance behaviors (Aso, 2014b).

In the cases described above, the DANs and MBONs mediating a particular behavior innervate the same regions of MB lobes. Cases were found where the DANs and MBONs required for a behavior do not innervate the same compartments of the MB lobes. For example, even though several cholinergic MBONs are required for appetitive memory, the compartments with cholinergic MBONs do not receive inputs from reward-mediating PAM cluster DANs, but instead from PPL1 cluster DANs that have been shown to be dispensable for odor-sugar memory. What accounts for this mismatch? Perhaps these cholinergic MBONs' primarily function is in memory consolidation rather than retrieval. But the fact that CsChrimson activation of the cholinergic MBON-γ2α'1 and V2 cluster MBONs resulted in attraction, strongly suggests that at least some of the cholinergic MBONs have a role in directly mediating the conditioned response. Indeed, previous studies found a requirement for cholinergic MBONs (the V2 cluster and MBON-α3) during memory retrieva. One attractive model is that requirement of cholinergic MBONs originates from the transfer of information between disparate regions of the MB lobes through the inter-compartmental MBONs connections within the lobes or by way of connections outside the MB, like those described in the next two sections (Aso, 2014b).

The multilayered arrangement of MBONs provides a circuit mechanism that enables local modulation in one compartment to affect the response of MBONs in other compartments. Once local modulation breaks the balance between MBONs, these inter-compartmental connections could amplify the differential level of activity of MBONs for opposing effects. For example, the avoidance-mediating MBON-γ4>γ1γ2 targets the compartments of attraction-mediating MBON-γ2α'1 and MBON-γ1pedc>α/β (Aso, 2014b).

This network topology might also provide a fly with the ability to modify its sensory associations in response to a changing environment. Consider the α lobe. Previous studies and the current results indicate that circuits in the α lobe play key roles in long-term aversive and appetitive memory. The α lobe is targeted by MBONs from other compartments and comprises the last layer in the layered output model of the MB. The GABAergic MBON-γ1pedc>α/β and the glutamatergic MBON-β1>α both project to α2 and α3, where their axonal termini lie in close apposition. DAN input to the compartments housing the dendrites of these feedforward MBONs induces aversive and appetitive memory, respectively. This circuit structure is well-suited to deal with conflicts between long-lasting memory traces and the need to adapt to survive in a dynamic environment where the meaning of a given sensory input may change. To test the proposed role of the layered arrangement of MBONs in resolving conflicts between old memories and new sensory inputs, behavioral paradigms will be needed that, unlike the simple associative learning tasks used in the current study, assess the neuronal requirements for memory extinction and reversal learning (Aso, 2014b).

The neuronal circuits that are downstream of the MBONs and that might read the ensemble of MBON activity remain to be discovered. However, the anatomy of the MBONs suggests that, at least in some cases, summation and canceling effects may result from convergence of MBON terminals on common targets. For example, the terminals of the sleep-promoting cholinergic MBON-γ2α'1 overlap with terminals of wake-promoting glutamatergic MBONs (γ5β'2a, β'2mp and β'2mp bilateral) in a confined area in CRE and SMP. In addition, some MBONs appear to terminate on the dendrites of DANs innervating other compartments, forming feedback loops. Using these mechanisms, local modulation in a specific compartment could broadly impact the ensemble of MBON activity and how it is interpreted (Aso, 2014b).

Testing these and other models for the roles of the MBON network, both within the MB lobes and in the surrounding neuropils, will be facilitated by an EM-level connectome to confirm the synaptic connections we have inferred based on light microscopy. Physiological assays will be needed to confirm the sign of synaptic connections and to measure plasticity. For example, the sign of action of glutamate in the targets of glutamatergic MBONs are not known, as this depends on the receptor expressed by the target cells. In this regard, it is noted that previous studies demonstrated a role for NMDA receptors in olfactory memory (Aso, 2014b).

Neurons that are thought to mediate innate response to odors (a subset of projection neurons from the antennal lobes and output neurons from the lateral horn) also project to these same convergence zones (see Convergence zone of MBON terminals as network nodes to integrate innate and learned valences). It is proposed that these convergence zones serve as network nodes where behavioral output is selected in the light of both the innate and learned valences of stimuli. What are the neurons downstream to these convergence zones? One obvious possibility is neurons that project to the fan-shaped body of the central complex whose dendrites are known to widely arborize in these same areas. It would make sense for the MB to provide input to the central complex, a brain region involved in coordinating motor patterns (Aso, 2014b).

Using inactivation to uncover the roles of specific cell types is inherently limited by redundancy and resiliency within the underlying neural circuits. For example, consider the MBONs from the α/β lobes. The output of the α/β Kenyon cells is known to be required for retrieval of aversive memory. The current anatomical and behavioral results show that MBON-γ1pedc>α/β, a cell type that was found to be critical for aversive memory, has terminals largely confined inside the α/β lobes, well-positioned to regulate a total 6 types of MBONs from the α/β lobes. Yet no requirement was detected for any of these MBONs in short term aversive memory when tested individually. The ability to detect phenotypes also depends on the strength of the effector; for example, four glutamatergic MBON drivers showed aversive memory impairment in initial screening assays with a strong inhibitor of synaptic function, but these effects could not be confirmed using a weaker effector (Aso, 2014b).

The failure to see effects when inactivating individual cell types is most easily explained by combinatorial roles and redundancy between MBONs. It is noted that this high level of resiliency is very reminiscent of observations made with genetic networks, where less than half of gene knockouts of evolutionarily conserved Drosophila genes result in a detectable phenotype. Whether or not a requirement is detected for a particular MBON in a particular learning paradigm is likely to depend on which DANs are recruited by the unconditioned stimulus used in that paradigm as well as the degree of redundancy in the MBON representation of valence. It will be informative to test systematically whether blocking combinations of MBONs, which did not show significant behavioral effect when blocked separately, results in significant memory impairment. It will also be important in future experiments to employ imaging and electrophysiological methods, in which the activities of individual neurons, and the consequences of plasticity, can be observed without being obscured by redundancy (Aso, 2014b).

The MBs are implicated in functions beyond processing of associative memory. MBONs that influence approach to, or avoidance of, a learned stimulus may also have roles in innate preference behaviors for temperature and hunger-dependent CO2 avoidance. Moreover, the behavioral repertoire that MBONs govern are expected to go beyond simple approach and avoidance; the MB is known to play a role in experience-dependent regulation of proboscis extension as well as regulation of sleep and post-mating behaviors such as oviposition. Intriguingly, this study found that MBONs whose activation was repulsive promoted wakefulness, whereas MBONs whose activation was attractive promoted sleep; it would make sense for flies to be awake and attentive in an adverse environment. Other internal states, in addition to sleep, are likely to affect the decision to carry out a particular memory-guided behavior; for example, the state of satiety has been shown to regulate memory expression. The diverse influences of MBONs on behavior can be most easily explained if it is assumed that the activity of the ensemble of MBON conveys an abstract representation of both valence and internal state. In this view, the ensemble of MBONs may represent internal states along axes such as pleasant-unpleasant or aroused-not aroused. It is upon these axes that primitive forms of emotion are thought to have evolved (Aso, 2014b).

Go to top

A connectome of a learning and memory center in the adult Drosophila brain

Takemura, S. Y., Aso, Y., Hige, T., Wong, A., Lu, Z., Xu, C. S., Rivlin, P. K., Hess, H., Zhao, T., Parag, T., Berg, S., Huang, G., Katz, W., Olbris, D. J., Plaza, S., Umayam, L., Aniceto, R., Chang, L. A., Lauchie, S., Ogundeyi, O., Ordish, C., Shinomiya, A., Sigmund, C., Takemura, S., Tran, J., Turner, G. C., Rubin, G. M. and Scheffer, L. K. (2017). Elife 6. PubMed ID: 28718765

Understanding memory formation, storage and retrieval requires knowledge of the underlying neuronal circuits. In Drosophila, the mushroom body (MB) is the major site of associative learning. This study reconstructed the morphologies and synaptic connections of all 983 neurons within the three functional units, or compartments, that compose the adult MB's alpha lobe, using a dataset of isotropic 8 nm voxels collected by focused ion-beam milling scanning electron microscopy. It was found that Kenyon cells (KCs), whose sparse activity encodes sensory information, each make multiple en passant synapses to MB output neurons (MBONs) in each compartment. Some MBONs have inputs from all KCs, while others differentially sample sensory modalities. Only 6% of KC>MBON synapses receive a direct synapse from a dopaminergic neuron (DAN). Two unanticipated classes of synapses, KC>DAN and DAN>MBON, were identified. DAN activation produces a slow depolarization of the MBON in these DAN>MBON synapses and can weaken memory recall (Takemura, 2017).

Associative memory helps animals adapt their behaviors to a dynamically changing world. The molecular mechanisms of memory formation are thought to involve persistent changes in the efficiency of synaptic transmission between neurons. In associative learning, persistent changes in synaptic efficacy correlated with memory formation have been found at points of convergence between two neuronal representations: one providing information from sensory inputs about the outside world and a second indicating whether the current environment is punitive or rewarding. Such sites of convergence have been identified for multiple forms of associative learning. However, a comprehensive synaptic level description of connectivity at such a site of convergence is not available for an animal as complex as the fruit fly, Drosophila (Takemura, 2017).

The mushroom body (MB) is the center of associative learning in insects. Sensory information enters the MB via the calyx, where the dendritic claws of Kenyon cells (KCs) receive synaptic inputs from projection neurons of olfactory and other modalities including visual, gustatory and thermal. The parallel axonal fibers of the KCs form the MB-lobes, the output region of the MB. A pattern of sparse activity in the KC population represents the identity of the stimulus. This sparseness is maintained through two mechanisms. First, individual KCs generally only spike when they receive simultaneous inputs from multiple projection neurons. Second, overall KC excitability is regulated by feedback inhibition from a GABAergic neuron, MB-APL, that arborizes throughout the MB. Thus, only a small subset of KCs respond to a given sensory stimulus. Upon this representation of the sensory world, dopaminergic or octopaminergic neurons convey information of punishment or reward and induce memories that associate the sensory stimulus with its valence (Takemura, 2017).

The functional architecture of the MB circuit is best understood in adult Drosophila (see Diagram of the α lobe of the mushroom body). In each MB, the parallel axonal fibers of ~2000 KCs can be divided into 16 compartmental units by the dendrites of 21 types of MB output neurons (MBONs) and the axon terminals of 20 types of dopaminergic neurons (DANs). A large body of behavioral and physiological studies suggests that these anatomical compartments are also parallel units of associative learning. In each compartment, the dendrites of a few MBONs overlap with axon bundles of hundreds of KCs. Punishment and reward activate distinct sets of DANs. DAN input to a compartment has been shown to induce enduring changes in efficacy of KC>MBONs synapses in those specific KCs that were active in that compartment at the time of dopamine release. The valence of the memory appears to be determined by which compartment receives dopamine during training, while the sensory specificity of the memory is determined by which KCs were active during training (Takemura, 2017).

Compartments can have distinct rates of memory acquisition and decay, and the 16 compartments together appear to form a set of parallel memory units whose activities are coordinated through both direct and indirect inter-compartmental connections. The DANs which project to the α1 compartment, the ventral-most compartment of the vertical lobe, play a key role in the formation of appetitive long-term memory of nutritional foods. DANs that project to the other α lobe compartments, α2 and α3, play roles in aversive long-term memory. All three of these compartments receive feedforward inputs from GABAergic and glutamatergic MBONs whose dendrites lie in other MB compartments known to be involved in aversive or appetitive memory. In addition, two types of MB-intrinsic neurons send arbors throughout the MB-lobes: a large GABAergic neuron, MB-APL, which provides negative feedback important for sparse coding, and the MB-DPM neuron, which is involved in memory consolidation and sleep regulation (Takemura, 2017).

Previous EM studies in the MB lobes of cockroaches, locusts, crickets, ants, honey bees and Drosophila identified KCs by their abundance, fasciculating axons and small size. Additionally, large GABA immunoreactive neurons that contact KC axons were identified in the locust pedunculus. While these data provided early insights to guide modeling of the MB circuit, the volumes analyzed were limited and most neuronal processes could not be definitively assigned to specific cell types. This paper reports a dense reconstruction of the three compartments that make up the α lobe of an adult Drosophila male. Because a dense reconstruction was performed, with the goal of determining the morphology and connectivity of all cells in the volume, it can be confidently stated that all cell types with processes in the α lobe have been identified (Takemura, 2017).

Comprehensive knowledge of the connectivity in the α lobe has allowed addressing of several outstanding issues. The first concerns the nature of KC>MBON connectivity. Although each KC passes through all three compartments, it is not known if individual KCs have en passant synapses in each compartment. Thus, it remains an open question whether the sensory representation provided to each compartment and each MBON within a compartment is the same or whether different MBONs within a compartment might sample from non-overlapping sets of KCs, and thus use independent sensory representations for learning. It was also not known which, if any, other cell types are direct postsynaptic targets of KCs (Takemura, 2017).

The second concerns dopamine modulation. What are the locations of dopaminergic synapses and what does this distribution imply about the targets of dopaminergic modulation as well as volume versus local transmission? Cell-type-specific rescue of dopamine receptor mutants suggests that dopamine acts presynaptically in the KCs of KC>MBON synapses. However, postsynaptic mechanisms have also been proposed and a recent study detected expression of dopamine receptors in MBONs, raising the possibility that MBONs might also be direct targets of DAN modulation. Behavioral, imaging and electrophysiological data indicate that dopamine modulation respects the borders between compartments, but it is not known whether these borders have a distinct structure, such as a glial sheet (Takemura, 2017).

The third concerns the two MBON types that send feedforward projections into the α lobe. These MBONs have important roles in associative learning as revealed by behavioral assays and have been postulated to integrate memories of opposing valence and different time scales. However, it is not known which cell types these feedforward MBON projections targets within the MB (Takemura, 2017).

The fourth concerns the two neurons, MB-APL and MB-DPM, which arborize throughout the MB and are thought to regulate MB function globally. What is their local synaptic connectivity within the α lobe and what can this inform about how they perform their roles (Takemura, 2017)?

Finally, the three compartments of the α lobe differ in important aspects, including valence of the memory formed, the time course of memory formation and retrieval, and the numerical complexity of their DAN inputs and MBON outputs. Are there obvious differences in the microcircuits of different compartments (Takemura, 2017)?

In this paper reports the answers to these questions. In addition, the utility of detailed anatomy at the electron microscopic level to provide novel insights is demonstrated: It is shown that nearly all cell types in the α lobe contain more than one morphological class of synaptic vesicle, raising the possibility that these cells utilize multiple neurotransmitters. In addition, two prevalent sets of synaptic motifs - from DANs to MBONs and from KCs to DAN s- are described that were unanticipated despite the extensive anatomical, physiological, behavioral and theoretical studies that have been performed on the insect MB. These novel DAN to MBON connections are characterized using behavioral and physiological assays and find that DAN activation produces a slow depolarization of postsynaptic MBONs and can weaken memory recall (Takemura, 2017).

The connections between the neurons observed in this study are are summarized in a Summary diagram of the connectome reconstruction of the α lobe. In each of the lobe's three compartments, parallel axonal fibers of ~1000 KCs project through the dendrites of a few MBONs and the terminal arbors of a few DANs. The results provide support for several aspects of the generally accepted model for MB circuit function. First, it was found that each KC forms en passant synapses with multiple MBONs down the length of its axon, making it possible for parallel processing across the different compartments of the MB lobes. Secondly, with the assumption that released dopamine diffuses locally, KC>MBON synapses would receive dopaminergic input close to the sites of vesicle release, consistent with the prevailing hypothesis that plasticity occurs at the presynaptic terminals of KCs. However, several circuit motifs were found that were not anticipated by previous work. For example, synaptic connections were found from KCs to DANs, indicating that DANs get axo-axonal inputs within the MB lobes themselves. A recent report provides evidence that these KC>DAN synapses are functional (Cervantes-Sandoval, 2017). An even more unexpected motif was the direct synaptic contacts from DAN to MBON found in every compartment. Functional connectivity experiments confirmed that these connections are monosynaptic, and showed that they give rise to a slow depolarization in the MBON. Moreover, stimulating DANs in freely behaving flies yields effects consistent with a net excitatory DAN>MBON connection. Finally, the synaptic connections are described of two feedforward MBONs, which have been proposed to mediate the interaction of the various parallel memories within the MB lobes, as well as two intrinsic MB neurons, APL and DPM (Takemura, 2017).

This work not only provides definitive evidence for, and quantitative detail about, many previously observed circuit motifs, but also reveals several motifs not anticipated by prior anatomical, behavioral or theoretical studies. These additional circuit motifs provide new insights and raise new questions about the computations carried out by the MB. It is noted that these same novel connections were also found in a parallel study of the larval MB. Not only were the same circuit motifs found in the larval MB and adult α lobe, but also the relative prevalence of these connections was strikingly similar: DAN>MBON synapses were 4.5% the number of KC>MBON synapses in the adult α lobe and 3.4% in the larval MB. KC>DAN synapses were 1.5 times as prevalent as DAN>KC synapses in the adult α lobe, as compared with 1.1 in the larval MB. KCs make 48% of their synapses onto other KCs in the adult α lobe and 45% in the larval upper vertical lobe compartments. It is tempting to speculate that the conservation of the relative abundances of these connections across developmental stages reflects important functional constraints on the circuit (Takemura, 2017).

A large body of work supports the idea that individual KC>MBON synapses are the elemental substrates of associative memory storage in the MB. The dominant hypothesis in the field is that coincidence detection occurs within the presynaptic terminals of the KCs. The Conditioned Stimulus (CS, for example an odor) evokes a spiking response in a sparse subset of KCs, which in turn leads to Ca2+ influx. The Unconditioned Stimulus (US, for example electric shock) activates dopaminergic inputs to the MB lobes, where they likely activate G-protein-coupled dopamine receptors on the KC cell membrane. The coincidence of these two events is thought to be detected by the Ca2+ sensitive, calmodulin-dependent adenylate cyclase rutabaga, which initiates a cAMP signaling cascade that leads to the biochemical changes underlying synaptic plasticity (Takemura, 2017).

The tiling of MBON and DAN projections down the length of the KC axons suggests that each of these compartments serves as an independent module, with the association of reinforcement with sensory input taking place in parallel across several different modules. One important assumption in this model is that each KC sends parallel input to each compartment by making synapses all the way down the length of its axon. Light microscopic imaging established that the axons of individual α/β KCs do indeed run through all three compartments of the α lobe. However, they also revealed that the axonal branching patterns differ between KC classes. For example, the axons of α/βp KCs branch in α2, whereas those of α/βc and α/βs KCs do not, raising the question of how extensive KC outputs are across the different compartments. The dense EM reconstruction established that in fact all α/β KCs form en passant synapses on MBONs in each of the three α lobe compartments (Takemura, 2017).

In many cases, these synapses were found at enlarged boutons that contained the presynaptic machinery. However, output sites were also found on the smooth axons of the α/βc KCs, which lack obvious bouton-like swellings. Only occasional, short (generally <5 μm) segments of KC axons where the axon became thinner than 300 nm in diameter lacked presynaptic sites. Of course, it is not known whether all these synapses are functional. EM analysis showed that within each compartment, every KC passing through a layer of the compartment that was extensively innervated by an MBON made at least one synapse with that MBON. Previous electrophysiological measurements of connectivity in the α2 compartment indicated that only about 30% of KCs connect to MBON-α2sc, suggesting the possibility that the majority of KC>MBON synapses are functionally silent, as they are in cerebellar cortex, where 98% of the parallel fiber-to-Purkinje cell synapses are believed to be silent. However, a more trivial explanation cannot be ruled out: These measurements were made in the presence of cholinergic antagonists that could have partially blocked synaptic events and lead to an underestimate of total connectivity levels (Takemura, 2017).

The EM data revealed that the the number of synapses made by individual KCs was well-described by a Poisson distribution, where each synapse connects with a uniform, independent, and random probability to one of the KCs. Although the predicted distributions strongly depend on the number of connections between two cell types, almost all KC connections to other cells obeyed Poisson statistics. This was true of every KC in the α1 and α3 compartments, where each MBON has compartment-filling dendrites. The α2 compartment is somewhat unusual in that its MBONs innervate only subzones of the compartment. While light microscopy showed that MBON-α2sc primarily innervates the surface and core of the compartment, MBON-α2sp was found to project more to the surface and posterior. The connectome results bore out these observations from the light and electron microscopy, although EM reconstructions also showed that these borders were not sharp, and these MBONs receive less extensive and weaker connections outside these subzones. Nevertheless, within the primary area of innervation, it was again the case that every KC made synapses with all MBONs along its axon. Thus each of the 949 α/β KCs can deliver information to the MBONs in each of the three α lobe compartments (Takemura, 2017).

A strictly feed-forward view of the circuit may miss important processing, however, as earlier studies suggested, and the current results re-emphasize. Firstly, gap junctions between KCs have been reported. This opens up the possibility for lateral propagation of signals across KCs, either biochemical or electrical. For example, in mammalian systems, axo-axonal gap junction coupling can synchronize firing between neurons. Secondly, chemical synapses between KCs have been reported in the MB pedunculus in the locust. The reconstructions show that such KC>KC connections are also present in the lobes, where they are surprisingly prevalent. In fact, the most frequent outputs of the α/βs KCs are other α/βs KCs, assuming the morphologically defined KC>KC connections are functional synapses (Takemura, 2017).

A high percentage (55%) of these putative KC>KC synapses occur in rosette-like structures where multiple KCs also converge on a single dendritic process of an MBON . These are relatively unusual structures, not observed in EM reconstructions of the Drosophila visual system and, indeed, there is no direct evidence that they are functional synapses. At present it is only possible to speculate on their role. As points of heavy convergence, they might allow the effects of synapses from different KCs onto the same dendrite to act synergistically. Activity of a single KC may spread to its neighbors within the rosette, potentially generating a large compound synaptic release event onto the MBON in the middle. Such a signal amplification mechanism may be important to ensure that individual KCs can have a significant impact on MBON membrane potential by recruiting their rosette partners. How the specificity of learning could be maintained in this scenario is, however, unclear. Several basic questions will need to be answered before it is possible to begin to understand the functional significance of these rosettes. For example, can a single KC in the rosette indeed activate its neighbors? And how similar are the response properties of the different KCs that contribute to one rosette (Takemura, 2017)?

In conclusion, the connectivity of the KCs that carry olfactory and other sensory representations supports a model where parallel distributed memory processing occurs in each compartment. However, several circuit motifs that seem designed to spread and possibly amplify signals at the sites of KC output indicate that this circuit is likely more complicated than a simple feed-forward view of the system suggests (Takemura, 2017).

Dopamine-induced plasticity of the KC>MBON synapse is thought to be central to associative learning in this system. The reconstructions showed that dopaminergic neurons make well-defined synaptic contacts within the α lobe, with closely apposed post-synaptic membranes. This contrasts somewhat with dopaminergic innervation in the mammalian system, where there is typically not such close contact with a single clear post-synaptic partner, and volume transmission is the predominant model for dopamine release. It is not known whether the direct and indirect dopaminergic release sites have different functional consequences. Nevertheless, it seems likely that some type of volume transmission happens in the mushroom body. First, ~10 times more KC>MBON synapses than presynaptic sites of dopamine release were found in the α lobe, but previous work showed that learning-induced plasticity depresses MBON responses so strongly that most inputs are likely affected. Second, dopamine would need to diffuse only ~2 μm to reach every KC>MBON synapse within a compartment, but would also be sufficiently short range to prevent significant spill-over of dopamine to neighboring compartments, ensuring that the modularity of plasticity is maintained (Takemura, 2017).

Functional connectivity measurements showed that stimulating the DANs elicits large amplitude calcium signals from MBONs, similar to previous results. Intracellular recordings revealed that this was a surprisingly strong connection, sufficient to elicit spikes in the MBON. The response persisted when both spiking and nicotinic transmission was blocked, to limit the possibility that the DANs act through the KCs, which are cholinergic. Conversely, the response was strongly reduced by adding a dopamine receptor antagonist. Taken together, these results indicate that the response is likely a direct action of dopamine released by the DANs on the MBON, although a more complex mechanism or a role for the transmitter contained in the dense core vesicles observed in the DANs cannot be formally rule out. The depolarization exhibited markedly slow dynamics, peaking >2 s after stimulation offset, and then decaying over tens of seconds. Dopaminergic responses of similar amplitude and time course have been reported in both mammalian systems and in Aplysia, where it is mediated by cAMP-driven changes in a non-selective cation conductance (Takemura, 2017).

It is possible to induce memory formation in this circuit by pairing odor delivery with artificial activation of DANs. Targeting this optogenetic training procedure to DANs that innervate different compartments within the α lobe gives rise to memories with different valence, induction threshold and persistence. In the α1 compartment, a single pairing for 1 min induces an appetitive memory that lasts for 1 day. In contrast, optogenetic training focused on the α3 compartment requires multiple 1 min pairings, repeated at spaced intervals, and induces an aversive memory that lasts for 4 days. Although it seems likely that the different valences reflect the different projection sites of the MBONs for each of these compartments, where the differences in induction threshold and memory persistence might arise is less clear. There is no simple explanation for these differences from the EM-level circuit structure, as the basic wiring motifs were very similar in each compartment. Moreover, any explanation that invokes biochemical differences in KC>MBON synapses would require crisp spatial localization of the signaling pathway machinery that triggers plasticity, as exactly the same KCs participate in memory formation in different compartments. However, the observation that there are DAN>MBON synapses raises the possibility that biochemical differences in the MBONs might contribute to these differences in plasticity induction and maintenance. Indeed, RNAseq data from a set of four different MBONs showed expression of dopamine receptors. An alternative possibility, suggested by the findings in this study, is that the cotransmitter found in the dense core vesicles in the DANs is responsible for these differences. The size of these vesicles differs between DANs innervating the different compartments. Thus, these cells might release distinct co-transmitters, as has been observed in mammalian brain, which could trigger different signaling cascades in either the KCs or the MBONs to differentially modulate the induction and expression of plasticity across compartments (Takemura, 2017).

Models of MB function have generally considered the role for DANs to be confined to relaying signals about punishment or reward to the MB. However, in the mammalian brain, DANs can dynamically change their responses to both US and CS. In this study, it was found that the axonal terminals of the DANs receive many inputs from KCs within the lobes. In other words, both MBONs, DANs and even KCs receive extensive synaptic input from KCs in each compartment. If the current model that plasticity is pre-synaptic proves to be correct, this suggests that the responses of the DANs themselves would be subject to plasticity. If the synaptic depression observed at KC>MBON synapses also acts at KC>DAN connections, odor-evoked DAN responses would be diminished as a result of learning. This would serve as a negative feedback loop, reducing the strength of plasticity on successive training cycles with the same odor. Indeed, a gradually plateauing of the learning curve is a common feature of memory formation in different systems, including olfactory conditioning in Drosophila (Takemura, 2017).

One of the more surprising findings of this study was the observation that there are many direct DAN>MBON synaptic connections. Moreover, the functional connectivity measures indicate that these were relatively strong excitatory inputs. The excitatory sign of the DAN>MBON connection is also consistent with the behavioral effects of DAN activation that was observed. What role these DAN>MBON connections play in overall circuit function is an important question for future work. There are two general possibilities that are felt to be interesting to consider. Dopaminergic modulation has been proposed to play a general role in routing of information through the MB to different downstream neurons. Although changes in KC>MBON strength contribute to this process, the current results suggest that such state changes could also potentially be conveyed to the MBONs directly from the DANs. State-dependent changes in DAN activity have indeed been observed with calcium imaging. The slow synaptic dynamics observed in the DAN>MBON connection in MBON-α1 suggest the possibility that small changes in DAN firing might be capable of producing sustained changes in MBON membrane potential reflecting the current internal state of the animal (Takemura, 2017).

A second possibility, suggested from the framework of reinforcement learning established in vertebrates, is related to motivation and the comparison of expected versus actual reward. In Drosophila, prior work on odor-sugar conditioning in larvae provided evidence that flies form a comparison between the current state of reward and the reward expected from the conditioned cue. This work showed that animals behaviorally express memories only when the expected reward intensity is higher than the currently available reward. This is similar to the results presented in this study; just as the presence of reward diminished memory expression in the larvae, stimulating the DANs suppressed performance of animals trained by the optogenetic conditioning. The need to compare current and expected reward could potentially explain why there is an opponent relationship between the depression of KC>MBON synapses that drives associative learning and the excitatory effects of the DAN>MBON connection. If depression dominates, the association drives behavior, but this can be overridden by sufficient levels of DAN activity. In this respect, it is noteworthy that DANs appear to be able to act directly on the MBON, without participation of the KCs. Overall, this comparison could ensure that learned behavior is motivated not strictly by the expectation of reward, but rather the expected increase in reward, assessed at the moment of testing (Takemura, 2017).

The organization of the MB into a set of compartments arranged in series along the KC axons is well suited for simultaneously storing multiple independent memories of a given sensory stimulus. However, there must be some means by which these modules interact with one another to ensure coordinated, coherent expression of memory. Feedforward connections that link different compartments, first discovered by light microscopic anatomy, have recently been shown to be important for mediating such interactions. In particular, MBON-γ1pedc>α/β is an inhibitory neuron that connects aversive and appetitive learning compartments; it ensures that the circuit can readily toggle between different behavioral outputs (Takemura, 2017).

The EM reconstructions included both MBON-γ1pedc>α/β and MBON-β1>α, two feedforward neurons which project from their respective compartments to widely innervate other parts of the MB. Memories stored in the α lobe compartments are long-term and relatively inflexible, whereas the short-term memories formed in β1 and γ1pedc are readily updated by recent experiences. The feedforward connections are thought to enable the short-term memories in β1 and γ1pedc to temporarily mask expression of the stable memories stored in the α lobe. Indeed training an animal with either a multi-component aversive/appetitive food stimulus, or by simultaneous optogenetic activation of a composite set of DANs covering both appetitive and aversive compartments results in a compound memory that is initially aversive and later transitions to appetitive. The connectome results show that the primary synaptic targets of these feedforward neurons are the MBONs in the downstream compartment. By contrast, relatively few connections onto KCs were observed. Overall, this suggests that the feedforward connections can strongly influence the output from a compartment, but likely have little impact on the sensory information delivered to each compartment from the KCs. This is consistent with observations that MBON-γ1pedc>α/β strongly modulates activity of glutamatergic neurons at the tip of the horizontal lobe, but not their dendritic responses. Targeting these feedforward connections to the MBON may ensure that conflicting memories can form simultaneously in response to a complex sensory input, but with the behavioral manifestation of those memories capable of undergoing a crisp switch (Takemura, 2017).

This study has provided synapse level anatomical information on neuronal circuits involved in learning and memory in Drosophila. The comprehensive nature of this dataset should enable modeling studies not previously possible and suggests many experiments to explore the physiological and behavioral significance of the circuit motifs that were observed. That many of these motifs were not anticipated by over 30 years of extensive anatomical, experimental and theoretical studies on the role of the insect MB argues strongly for the value of electron microscopic connectomic studies (Takemura, 2017).

A dense (complete) reconstruction of neurons and synapses is resource intensive, so it is reasonable to ask if tracing a subset of cells or synapses could have yielded similar results with less effort. This is hard to answer in general, since there are many sparse tracing strategies, and each can be pursued to differing degrees of completeness. It is likely that most sparse tracing strategies would have discovered the new pathways reported in this, as the connections are numerous and connect well known cell types. Conversely, the conclusions that all cell types in this circuit had been identified would have been more difficult to make with confidence and a rare cell type, such as the SIFamide neuron, might have been missed. Perhaps, most importantly, statistical arguments, particularly those that require an accurate assessment of which cells are not connected, such as the absence of network structures such as rings or chains, would have been hard to make from sparse tracing. More generally, the model independent nature of dense tracing helps to discover any 'unknown unknowns', provides the strongest constraints on how neural circuits are constructed, and allows retrospective analysis of network properties not targeted during reconstruction (Takemura, 2017).

Go to top

Localized inhibition in the Drosophila mushroom body
Amin, H., Apostolopoulou, A. A., Suarez-Grimalt, R., Vrontou, E. and Lin, A. C. (2020). Localized inhibition in the Drosophila mushroom body. Elife 9. PubMed ID: 32955437

Many neurons show compartmentalized activity, in which activity does not spread readily across the cell, allowing input and output to occur locally. However, the functional implications of compartmentalized activity for the wider neural circuit are often unclear. This problem was addressed in the Drosophila mushroom body, whose principal neurons, Kenyon cells, receive feedback inhibition from a non-spiking interneuron called APL. This study used local stimulation and volumetric calcium imaging to show that APL inhibits Kenyon cells' dendrites and axons, and that both activity in APL and APL's inhibitory effect on Kenyon cells are spatially localized (the latter somewhat less so), allowing APL to differentially inhibit different mushroom body compartments. Applying these results to the Drosophila hemibrain connectome predicts that individual Kenyon cells inhibit themselves via APL more strongly than they inhibit other individual Kenyon cells. These findings reveal how cellular physiology and detailed network anatomy can combine to influence circuit function (Amin, 2020).

This study has shown that activity in APL is spatially restricted intracellularly for both sensory-evoked and local artificial stimulation. This local activity in APL translates into a spatially non-uniform inhibitory effect on Kenyon cells that is strongest locally and becomes weaker farther from the site of APL stimulation. Finally, combining physiological and anatomical data -- APL's estimated space constant and the spatial arrangement of KC-APL and APL-KC synapses -- predicts that each Kenyon cell disproportionately inhibits itself more than other individual Kenyon cells (Amin, 2020).

This study showed that APL can locally inhibit Kenyon cells in multiple locations in the mushroom body. Remarkably, inhibition of Ca2+ influx in Kenyon cell axons could be stronger closer to the axon initial segment (in the posterior peduncle than farther from it, for example, stronger at the lower vertical lobe than the tip of the vertical lobe, both when stimulating APL with ATP or when directly applying GABA. At first glance, this observation is puzzling given the direction of action potential travel. GABA should act on Kenyon cells by suppressing depolarization through shunting inhibition in both dendrites and axons, as the GABAA receptor Rdl is expressed in both. Therefore, the observation of stronger proximal than distal inhibition appears to suggest that as action potentials travel from the axon initial segment toward the distal tip of the axon, they can enter a zone of shunting inhibition and be suppressed, yet recover on the other side. Such a scenario could occur if depolarization is suppressed enough to reduce voltage-gated Ca2+ influx in a local zone, yet remains sufficient on the other side of the inhibitory zone to trigger enough voltage-gated Na+ channels to regenerate the action potential. However, this interpretation seems unlikely. More likely, local inhibition might particularly suppress Ca2+ influx rather than depolarization, perhaps by acting through GABAB receptors; Kenyon cells express GABABR1 and GABABR2 and APL inhibits KCs partly via GABAB receptors, although their subcellular localization is unknown. Given that synaptic vesicle release requires Ca2+ influx, such inhibition would still locally suppress Kenyon cell synaptic output (Amin, 2020).

Local inhibition of Kenyon cell output predicts that activity of MBONs near the site of APL activation would be more strongly inhibited than MBONs far away. This prediction may be tested in future experiments, for example locally stimulating APL in the tip of the vertical lobe and comparing the inhibitory effect on MBONs in nearby compartments like α3 and α'3 vs. on MBONs in distant compartments like γ5 and β2 (Amin, 2020).

While APL inhibits Kenyon cells locally, the inhibitory effect spreads somewhat more widely than APL's own activity, though weakly. Local GABA application produces similar results to locally activating APL with ATP, suggesting that GCaMP6f signals in APL accurately predict GABA release. How can APL inhibit Kenyon cells where it itself is not active? Wider inhibition when inhibiting Kenyon cell dendrites can be easily explained as blocking action potentials, but the wider inhibition when inhibiting the axons is more puzzling. It is speculated that this result might arise from wider network activity. For example, Kenyon cells form recurrent connections with DPM and dopaminergic neurons and form synapses and gap junctions on each other. Through such connections, an odor-activated Kenyon cell might excite a neighboring Kenyon cell's axon; the neighbor might passively spread activity both forward and backward or, not having fired and thus not being in a refractory period, it might even be excited enough to propagate an action potential back to the calyx. Alternatively, Kenyon cells indirectly excite antennal lobe neurons in a positive feedback loop. In these scenarios, locally puffing GABA or activating APL artificially in the vertical lobe tip would block the wider network activity, thus reducing Ca2+ influx in the calyx. Although the calyx was not activated when stimulating Kenyon cells in the vertical lobe tip, this might be because simultaneous activation of all Kenyon cells (but not odor-evoked activation of ~10% of Kenyon cells) drives strong-enough local APL activation to block wider network activity. Future experiments may test these possibilities (Amin, 2020).

By combining physiological measurements of localized activity with the detailed anatomy of the hemibrain connectome, a model was built that predicts that the average Kenyon cell inhibits itself more than it inhibits other individual Kenyon cells. This prediction is supported by previous experimental results that some Kenyon cells can inhibit some Kenyon cells more than others, and that an individual Kenyon cell can inhibit itself. The model goes beyond these results in predicting that the average Kenyon cell actually preferentially inhibits itself, and by explaining how differential inhibition can arise from local activity in APL and the spatial arrangement of KC-APL and APL-KC synapses (Amin, 2020).

Note that these results do not contradict previous findings that Kenyon cell lateral inhibition is all-to-all. The finding of all-to-all inhibition was based on findings that whereas blocking synaptic output from all Kenyon cells vastly increases Kenyon cell odor responses (by blocking negative feedback via APL), blocking output from one subset of Kenyon cells has no effect on (for α'β' and γ Kenyon cells), or only weakly increases (for αβ Kenyon cells), odor responses of that subset. This finding is consistent with findings that predict only preferential, not exclusive, self-inhibition. For example, while blocking output from only γ Kenyon cells would remove γ-to-γ inhibition, γ Kenyon cells (only 1/3 of all Kenyon cells) should still get enough lateral inhibition from the other 2/3 of Kenyon cells to suppress their activity to normal levels (Amin, 2020).

The findings show that localized activity within APL has two broader implications for mushroom body function. First, local activity in APL leads to stronger inhibition of Kenyon cells nearby than far away, to the point that different compartments of the mushroom body lobes defined by dopaminergic and output neuron innervation can receive different inhibition. This local inhibition would allow the single APL neuron to function effectively as multiple inhibitory interneurons, much like mammalian and fly amacrine cells. Each 'sub-neuron' of APL could locally modulate the function of one mushroom body 'compartment', that is, a unit of dopaminergic neurons/Kenyon cells / mushroom body output neurons. Different compartments are innervated by different dopaminergic neurons (e.g., reward vs. punishment) and different mushroom body output neurons (e.g., avoid vs. approach), and they govern synaptic plasticity and hence learning by different rules (e.g., different speeds of learning/forgetting). Thus, APL could locally modulate different compartment-specific aspects of olfactory learning, especially given that different regions of APL respond differently to dopamine and electric shock punishment. If such local inhibition is important for learning, the fact that spatial attenuation of APL's inhibitory effect is gradual and incomplete could explain why mushroom body compartments are arranged in their particular order, with reward and punishment dopaminergic neurons segregated into the horizontal and vertical lobes, respectively. Under this scenario, APL would serve two distinct, spatially segregated functions: enforcing Kenyon cell sparse coding in the calyx, and modulating learning in the compartments of the lobes (Amin, 2020).

Second, the finding of disproportionate self-inhibition compared to other-inhibition provides a new perspective on APL's function. Inhibition from APL sparsens and decorrelates Kenyon cell odor responses to enhance learned odor discrimination but in general, decorrelation is better served by all-to-all lateral inhibition than by self-inhibition. Self-inhibition does help decorrelate population activity by pushing some neurons' activity below spiking threshold; this could occur if APL can be activated by subthreshold activity in Kenyon cells, for example, from KC-APL synapses in Kenyon cell dendrites. However, this effect of self-inhibition is better thought of, not as decorrelation per se, but rather as gain control, effectively equivalent to adjusting the threshold or gain of excitation according to the strength of stimulus. Of course, lateral inhibition is still a strong force in the mushroom body: given that there are ~2000 Kenyon cells, the sum total lateral inhibition that an individual Kenyon cell receives would still be stronger than its own self-inhibition, even with the predicted imbalance. Why might the predicted 'bonus' self-inhibition be useful? Beyond its role in sparse coding, APL inhibition is also thought to function as a gating mechanism to suppress olfactory learning; for such a function it would make sense for Kenyon cells to disproportionately inhibit themselves. Future work will address how lateral inhibition interacts with other functions for APL in this local feedback circuit (Amin, 2020).

Go to top

The complete connectome of a learning and memory centre in an insect brain

Eichler, K., Li, F., Litwin-Kumar, A., Park, Y., Andrade, I., Schneider-Mizell, C. M., Saumweber, T., Huser, A., Eschbach, C., Gerber, B., Fetter, R. D., Truman, J. W., Priebe, C. E., Abbott, L. F., Thum, A. S., Zlatic, M. and Cardona, A. (2017). Nature 548(7666): 175-182. PubMed ID: 28796202

Associating stimuli with positive or negative reinforcement is essential for survival, but a complete wiring diagram of a higher-order circuit supporting associative memory has not been previously available. This study reconstructed one such circuit at synaptic resolution, the Drosophila larval mushroom body. Most Kenyon cells were found to integrate random combinations of inputs, but a subset receives stereotyped inputs from single projection neurons. This organization maximizes performance of a model output neuron on a stimulus discrimination task. A novel canonical circuit in each mushroom body compartment with previously unidentified connections is reported: reciprocal Kenyon cell to modulatory neuron connections, modulatory neuron to output neuron connections, and a surprisingly high number of recurrent connections between Kenyon cells. Stereotyped connections found between output neurons could enhance the selection of learned behaviours. The complete circuit map of the mushroom body should guide future functional studies of this learning and memory centre (Eichler, 2017).

Massively parallel, higher-order neuronal circuits such as the cerebellum and insect mushroom body (MB) serve to form and retain associations between stimuli and reinforcement in vertebrates and evolutionarily complex invertebrates. Although these systems provide a biological substrate for adaptive behaviour, no complete synapse-resolution wiring diagram of their connectivity has been available to guide analysis and aid understanding. The MB is a higher-order parallel fibre system in many invertebrate brains, including hemimetabolous as well as holometabolous insects and their larval stages. MB function is essential for associative learning in adult insects and in Drosophila larvae from the earliest larval stages onwards. Indeed, the basic organization of the adult and the larval MB and their afferent circuits is very similar; however, larvae have about an order of magnitude fewer neurons. Thus, to systematically investigate the organizational logic of the MB, this study used serial section electron microscopy to map with synaptic resolution the complete MB connectome in a first-instar Drosophila larva. L1 are foraging animals capable of all behaviours previously described in later larval stages including adaptive behaviours dependent on associative learning. Their smaller neurons enable fast electron microscopy imaging of the entire nervous system and reconstruction of complete circuits (Eichler, 2017).

Models of sensory processing in many neural circuits feature neurons that fire in response to combinations of sensory inputs, generating a high-dimensional representation of the sensory environment. The intrinsic MB neurons, the Kenyon cells (KCs), integrate in their dendrites inputs from combinations of projection neurons (PNs) that encode various stimuli, predominantly olfactory in both adult and larva1, but also thermal, gustatory, and visual in adult and larva. Previous analyses in adults and larvae suggest that the connectivity between olfactory PNs and KCs is random, but they do not eliminate the possibility of some degree of bilateral symmetry, which requires access to the full PN-to-KC wiring diagram in both hemispheres (Eichler, 2017).

The MB contains circuitry capable of associating reward or punishment with the representation of the sensory environment formed by KCs. KCs have long parallel axons that first run together, forming the peduncle, and then bifurcate, forming the so-called lobes, in both larvae and adults. KCs receive localized inputs along their axonal fibres from dopaminergic as well as octopaminergic modulatory neurons (DANs and OANs, respectively) that define separate compartments. DANs and OANs have been shown to convey reinforcement signals in adult insects and larval Drosophila. The dendrites of the mushroom body output neurons (MBONs) respect the DAN compartments in adults and larvae. It has been shown in adult Drosophila that co-activation of KCs and DANs can associatively modulate the KC-MBON synapse. Thus, the compartments represent anatomical and functional MB units where sensory input (KCs) is integrated with internal reinforcement signals (DANs/OANs) to modulate instructive output for behavioural control (MBONs). However, the synaptic connectivity of KCs, DAN/OANs, and MBONs at this crucial point of integration was previously unknown (Eichler, 2017).

Furthermore, studies in adult Drosophila have shown that, despite the compartmental organization of the MB, many MBONs interact with MBONs from other compartments, suggesting that the MBON network functions combinatorially during memory formation and retrieval. However, a comprehensive account of all MB neuron connections is lacking. Thus, to provide a basis for understanding how the MB, a prototypical parallel fibre learning and memory circuit, functions as an integrated whole, this study provides a full, synapse-resolution connectome of all MB neurons of an L1 Drosophila larvaxxWe provide the first complete wiring diagram of a parallel fibre circuit for adaptive behavioural control. Such circuits exist in various forms, for example the cerebellum in vertebrates and the MB in insects. They contribute to multiple aspects of behavioural control including stimulus classification, the formation and retrieval of Pavlovian associations, and memory-based action selection. A comprehensive wiring diagram of such a multi-purpose structure is an essential starting point for functionally testing the observed structural connections and for elucidating the circuit implementation of these fundamental brain functions (Eichler, 2017).

Even though individual neurons may change through metamorphosis, many of the basic aspects of the MB architecture are shared between larval and adult Drosophila stages and with other insects. It is therefore expected that the circuit motifs identified in this study are not unique to the L1 developmental stage, but instead represent a general feature of Drosophila and insect MBs (Eichler, 2017).

Electron microscopy reconstruction revealed a canonical circuit in each MB compartment featuring two unexpected motifs in addition to the previously known MBIN-to-KC and KC-to-MBON connections. First, it was surprising to observe that the number of KC-to-MBIN and KC-to-MBON synapses are comparable. As KCs were shown to be cholinergic in adults, KC-to-MBIN connections could be potentially depolarizing. Untrained, novel odours can activate DANs in adult Drosophila and OANs in bees. Similar brief short-latency activations of dopamine neurons by novel stimuli are observed in monkeys, too, and are interpreted as salience signals. Learning could potentially modulate the strength of the KC-to-MBIN connection, either weakening it or strengthening it. The latter scenario could explain the increase in DAN activation by reinforcement-predicting odours observed in adult Drosophila, bees, and monkeys. In addition, dopamine receptors have been shown to be required in Drosophila KCs for memory formation. Another unexpected finding was that MBINs synapse directly onto MBONs, rather than only onto KCs. Such a motif could provide a substrate for neuromodulator-gated Hebbian spike-timing dependent plasticity, which has been observed in the locust MB (Eichler, 2017).

In addition to random and bilaterally asymmetric olfactory and structured non-olfactory PN-to-KC connectivity, the analysis identified single-claw KCs whose number and lack of redundancy are inconsistent with random wiring. Random wiring has previously been shown to increase the dimension of sensory representations when the number of neurons participating in the representation is large compared with the number of afferent fibres, as in the cerebellum or adult MB. However, the current model shows that when the number of neurons is limited, random wiring alone is inferior to a combination of random and structured connectivity that ensures each input is sampled without redundancy. The presence of single-claw KCs may reflect an implementation of such a strategy. In general, the results are consistent with a developmental program that produces complete and high-dimensional KC sensory representations to support stimulus discrimination at both larval and adult stages (Eichler, 2017).

This study reveals the complete MBON-MBON network at synaptic resolution. Previous studies in the larva have shown that odour paired with activation of medial and vertical lobe DANs leads to learned approach8 and avoidance, respectively. This connectivity analysis reveals that glutamatergic MBONs from the medial lobe laterally connect to MBONs of the vertical lobe. The glutamatergic MBON-MBON connections could be inhibitory, although further studies are needed to confirm this. Furthermore, inhibitory GABAergic MBONs from the vertical lobe laterally connect to MBONs of the medial lobe. An example is the feedforward inhibition of medial lobe MBON-i1 output neuron by the vertical lobe GABAergic MBON-g1, -g2 output neurons. A similar motif has been observed in the Drosophila adult, where aversive learning induces depression of conditioned odour responses in the approach-promoting MBON-MVP2 (MBON-11), which in turn disinhibits conditioned odour responses in the avoidance-promoting MBON-M4/M6 (MBON-03) because of the MBON-MVP2 to MBON-M4/M6 feedforward inhibitory connection (Eichler, 2017).

Combining the present connectivity analysis of the MBON-MBON network in the larva and previous studies in the adult Drosophila, the rule seems to be that MBONs encoding opposite learned valance laterally inhibit each other. Such inhibitory interactions have been proposed as a prototypical circuit motif for memory-based action selection (Eichler, 2017).

Go to top

Carbon monoxide, a retrograde messenger generated in postsynaptic mushroom body neurons, evokes noncanonical dopamine release

Ueno, K., Suzuki, E., Naganos, S., Ofusa, K., Horiuchi, J. and Saitoe, M. (2017). J Neurosci 40(18): 3533-3548. PubMed ID: 32253360

Dopaminergic neurons innervate extensive areas of the brain and release dopamine (DA) onto a wide range of target neurons. However, DA release is also precisely regulated. In Drosophila melanogaster brain explant preparations, DA is released specifically onto α3/α'3 compartments of mushroom body (MB) neurons that have been coincidentally activated by cholinergic and glutamatergic inputs. The mechanism for this precise release has been unclear. This study found that coincidentally activated MB neurons generate carbon monoxide (CO), which functions as a retrograde signal evoking local DA release from presynaptic terminals. CO production depends on activity of heme oxygenase in postsynaptic MB neurons, and CO-evoked DA release requires Ca(2+) efflux through ryanodine receptors in DA terminals. CO is only produced in MB areas receiving coincident activation, and removal of CO using scavengers blocks DA release. It is proposed that DA neurons use two distinct modes of transmission to produce global and local DA signaling (Ueno, 2020).

Dopamine (DA) is required for various brain functions, including the regulation of global brain states, such as arousal and moods. To perform these functions, individual DA neurons innervate extensive areas of the brain, and release DA onto a wide range of target neurons through a processes known as volume transmission. However, this extensive innervation is not suitable for precise, localized release of DA, and it has been unclear how widely innervating dopaminergic neurons can also direct DA release onto specific target neurons (Ueno, 2020).

In Drosophila, olfactory associative memories are formed and stored in the mushroom bodies (MBs) where Kenyon cells, MB intrinsic neurons which are activated by different odors, form synaptic connections with various MB output neurons, which regulate approach and avoidance behaviors. Dopaminergic neurons (DA neurons) modulate plasticity of synapses between Kenyon cells and MB output neurons. However, while there are ~2000-2500 Kenyon cells that form thousands of synapses with MB output neurons, plasticity at these synapses is regulated by relatively few DA neurons. This indicates that canonical action potential-dependent release cannot fully explain DA release and plasticity. It was recently determined that in Drosophila, synaptic vesicular (SV) exocytosis from DA terminals is restricted to MB neurons that have been activated by coincident inputs from odor-activated cholinergic pathways, and glutamatergic pathways, which convey somatosensory (pain) information. Odor information is transmitted to the MBs by projection neurons from the antennal lobe (AL), while somatosensory information is transmitted to the brain via ascending fibers of the ventral nerve cord (AFV). AL stimulation evokes Ca2+ responses in the MB by activating nAChRs, and AFV stimulation evokes Ca2+ responses in the MBs by activating N-Methyl-D-aspartate receptors (NMDARs) in the MBs. Significantly, when the AL and AFV are stimulated simultaneously (AL 1 AFV) or the AL and NMDARs are stimulated simultaneously (AL 1 NMDA), plasticity occurs such that subsequent AL stimulations causes increased Ca2+ responses in the a3/a93 compartments. This plasticity is known as long-term enhancement (LTE) of MB responses and requires activation of D1 type DA receptors (D1Rs) in the MBs. Furthermore, while activation of D1Rs alone is sufficient to produce LTE, neither AL nor AFV stimulation alone is able to cause SV exocytosis from presynaptic DA terminals projecting onto the a3/a93 compartments of the vertical MB lobes. Instead, exocytosis from DA terminals occurs only when postsynaptic Kenyon cells are activated by coincident AL 1 AFV or AL 1 NMDA stimulation. Strikingly, while MBs are bilateral structures and DA neurons project terminals onto both sides of MBs, SV exocytosis occurs specifically in MB areas that have been coincidently activated. Based on these results, it is proposed that coincident inputs specify the location where DA is released, whereas DA induces plastic changes needed to encode associations. However, it has been unclear how activated Kenyon cells send a retrograde 'demand' signal to presynaptic DA terminals to induce SV release (Ueno, 2020).

This study used a Drosophila dissected brain system to examine synaptic plasticity and DA release, and found that coincidentally activated postsynaptic Kenyon cells generate the retrograde messenger, carbon monoxide (CO). CO is generated by heme oxygenase (HO) in postsynaptic MB neurons, and induces DA release from presynaptic terminals by evoking Ca2+ release from internal stores via ryanodine receptors (RyRs). Thus, while individual DA neurons extensively innervate the MBs, on-demand SV exocytosis allows DA neurons to induce plasticity in specific target neurons (Ueno, 2020).

CO functions as a retrograde on-demand messenger for SV exocytosis in presynaptic DA terminals A central tenet of neurobiology is that action potentials, propagating from the cell bodies, induce Ca2+ influx in presynaptic terminals to evoke SV exocytosis. However, recent mammalian studies have shown that only a certain fraction of a large number of presynaptic DA release sites is involved in canonical SV exocytosis. In this study, a novel mechanism of SV exocytosis was identified in which activity in postsynaptic neurons evokes presynaptic release to induce plastic changes. This mechanism allows the timing and location of DA release to be strictly defined by activity of postsynaptic neurons (Ueno, 2020).

On-demand SV exocytosis uses CO as a retrograde signal from postsynaptic MB neurons to presynaptic DA terminals. CO fulfills the criteria that have been proposed for a retrograde messenger: (1) it was demonstrated that HO, which catalyzes CO production, is highly expressed in postsynaptic MB neurons, indicating that MB neurons have the capacity to synthesize the messenger; (2) it was shown that pharmacological and genetic suppression of HO activity in the MBs inhibits CO production, presynaptic DA release, and LTE, and (3) using a CO fluorescent probe, COP-1, it was demonstrated that CO is generated in the MBs following coincident stimulation of the MBs, and CO generation is restricted to lobes of MB neurons that receive coincident stimulation. It was further shown that direct application of CO, or a CO donor, induces DA release from presynaptic terminals, whereas addition of a CO scavenger, HemoCD, suppresses release. Fourth, it was demonstrated that CO activates RyR in presynaptic terminals to induce SV exocytosis. Strikingly, CO-dependent SV exocytosis does not depend on influx of extracellular Ca2+ but instead requires efflux of Ca2+ from internal stores via RyR. Finally, it was shown that pharmacological inhibition and genetic suppression of RyR in DA neurons impair DA release after coincident stimulation and CO application (Ueno, 2020).

Other retrograde signals, such as NO and endocannabinoids, enhance or suppress canonical SV exocytosis, but this study finds that CO-dependent DA release occurs even in conditions that block neuronal activity and Ca2+ influx in presynaptic DA terminals. This suggests that CO does not function to modulate canonical SV exocytosis, but may instead evoke exocytosis through a novel mechanism. Several previous studies have indicated that CO and RyR-dependent DA release also occurs in mammals. A microdialysis study has shown that CO increases the extracellular DA concentration in the rat striatum and hippocampus, either through increased DA release or inhibition of DA reuptake. Also, pharmacological stimulation of RyRs has been reported to induce DA release in the mice striatum. This release is attenuated in RyR3-deficient mice, while KCl-induced DA release, which requires influx of extracellular Ca2+, is unaffected, suggesting that RyR-dependent release is distinct from canonical DA release. However, it has been unknown whether and how CO is generated endogenously. Physiologic conditions that activate RyR to evoke DA release have also been unclear (Ueno, 2020).

While most neurotransmitters are stored in synaptic vesicles and released on neuronal depolarization, the release of gaseous retrograde messengers, such as NO and CO, is likely coupled to activation of their biosynthetic enzymes, NOS and HO. In mammals, an HO isoform, HO-2, is selectively enriched in neurons, and HO-2-derived CO is reported to function in plasticity. HO-2 is activated by Ca2+/calmodulin (CaM) binding, and by casein kinase II (CKII) phosphorylation. Previously, it was shown that coincident AL 1 NMDA stimulation induces a robust Ca2+ increase in the MBs that is greater than the increase from either stimulation alone. It is proposed that this increase may activate Drosophila HO in the MB to generate CO during associative stimulation. While Drosophila has a single isoform of RyR, mammals have three isoforms, RyR1-RyR3. Skeletal muscle and cardiac muscle primarily express RyR1 and RyR2, and the brain, including the striatum, hippocampus, and cortex, expresses all three isoforms. RyRs are known to be activated by Ca2+ to mediate Ca2+ induced Ca2+ release. However, CO-evoked DA release occurs even in the presence of Ca2+-free extracellular solutions containing TTX and EGTA, suggesting that CO activates RyR through a different mechanism. In addition to Ca2+, RyR can be activated by calmodulin, ATP, PKA, PKG, cADP-ribose, and NO. NO can directly stimulate RyR1 nonenzymatically by S-nitrosylating a histidine residue to induce Ca2+ efflux. CO has been reported to activate Ca2+-activated potassium channels (KCa) through a nonenzymatic reaction in rat artery smooth muscle, raising the possibility that it may activate RyR through a similar mechanism. Alternatively, both NO and CO can bind to the heme moiety of soluble guanlylate cyclase leading to its activation. Activated soluble guanlylate cyclase produces cGMP, and cGMP-dependent protein kinase (PKG) rapidly phosphorylates and activates RyRs. Interestingly, NO increases DA in the mammalian striatum in a neural activity-independent manner. Since activation of RyRs also increases extracellular DA in the striatum, hippocampus, and cortex, NO may play a pivotal role in RyRs activation and DA release in mammals. However, NOS expression has not been detected in the MBs, suggesting that, in Drosophila, CO rather than NO may function in this process (Ueno, 2020).

Previous studies have shown that electrical activity from the AL and AFV is transmitted to the MBs by cholinergic and glutamatergic neurons acting on nAChRs and NMDARs, respectively. Although the cholinergic inputs from the AL are known to be delivered by projection neurons, the glutamate inputs are still unclear. Previous work identified glutamatergic neurons that innervate a3/a93 compartments of the MBs and show SV release on electrical stimulation of the AFV. It is proposed that these neurons may transmit information regarding AFV stimuli to the MBs. Alternatively, while NMDARs are localized throughout MB lobes, vesicular glutamate transporter-positive terminals are found only sparsely on the MBs. This suggests that neurons expressing a currently uncharacterized vesicular glutamate transporter may convey information from the AFV to MBs (Ueno, 2020).

DA plays a critical role in associative learning and synaptic plasticity. In flies, neutral odors induce MB responses by activating sparse subsets of MB neurons. After being paired with electrical shocks during aversive olfactory conditioning, odors induce larger MB responses in certain areas of the MBs. This plastic change was modeled in ex vivo brains as LTE, and it was shown that DA application alone is sufficient to induce this larger response. However, in the Drosophila brain, only a small number of DA neurons (~12 for aversive and ~100 for appetitive) regulate plasticity in ~2000 MB Kenyon cells. Thus, to form odor-specific associations, there should be a mechanism regulating release at individual synapses. CO-dependent on-demand DA release provides this type of control. If on-demand release is involved in plasticity and associative learning, knockdown of genes associated with release should affect learning. Indeed, this study shows that knocking down either dHO in the MBs or RyR in DA neurons impairs olfactory conditioning. While these knockdowns did not completely abolish olfactory conditioning, this may be due to inefficiency of the knockdown lines. Alternatively, on-demand release may not be the only mechanism responsible for memory formation, but may instead be required for a specific phase of olfactory memory (Ueno, 2020).

In ex vivo studies, this study found that DA release requires coincident activation of postsynaptic MB neurons by cholinergic and glutamatergic stimuli. However, other in vivo imaging studies have shown that DA neurons can be activated and release DA on odor stimulation or shock application alone. Notably, projection of DA terminals is compartmentalized on the MB lobes and shows distinct responses and DA release during sensory processing. In these studies, dopaminergic neurons innervating the the a3/a93 compartments at the tips of the MB vertical lobes were examined, whereas other studies focused on compartments located on the MB horizontal lobes. This suggests that plasticity in different MB compartments may be regulated by different mechanisms. Unfortunately, due the location of the microelectrode for AL stimulation which caused interference in fluorescent imaging of the horizontal lobes, it was not possible to obtain reliable imaging data from these lobes in this study. Another difference between ex vivo and in vivo studies is that in vivo imaging studies use living, tethered, dissected flies that are likely in different states of arousal/distress, are exposed to many different stimuli, and can form unintended associations. In contrast, brains in ex vivo preparations are in a more controlled environment and are likely exposed to fewer unintended sensory stimuli. This may also explain apparent discrepancies between ex vivo and previous in vivo results (Ueno, 2020).

In mammals, the role of CO in synaptic plasticity is unclear. Application of CO paired with low-frequency stimulation induces LTP, while inhibiting HO blocks LTP in the CA1 region of the hippocampus. However, HO-2-deficient mice have been reported to have normal hippocampal CA1 LTP. In contrast to CO, a role for NO in synaptic plasticity and learning has been previously reported. Thus, at this point, it is an open question whether CO or NO evokes DA release in mammals. Downstream from CO or NO, RyRs have been shown to be required for hippocampal and cerebellar synaptic plasticity (Ueno, 2020).

The current results suggest that DA neurons release DA via two distinct mechanisms: canonical exocytosis and on-demand release. Canonical exocytosis is evoked by electrical activity of presynaptic DA neurons, requires Ca2+ influx, and may be involved in volume transmission. This mode of release can activate widespread targets over time, and is suited for regulating global brain functions. In contrast, on-demand release is evoked by activity of postsynaptic neurons, requires Ca2+ efflux via RyR, and can regulate function of specific targets at precise times. DA neurons may differentially use these two modes of SV exocytosis in a context-dependent manner. Understanding how DA neurons differentially use these modes of transmission will provide new insights into how a relatively small number of DA neurons can control numerous different brain functions (Ueno, 2020).

Go to top

Long-term memory requires sequential protein synthesis in three subsets of mushroom body output neurons in Drosophila

Wu, J. K., Tai, C. Y., Feng, K. L., Chen, S. L., Chen, C. C. and Chiang, A. S. (2017). Sci Rep 7(1): 7112. PubMed ID: 28769066

Creating long-term memory (LTM) requires new protein synthesis to stabilize learning-induced synaptic changes in the brain. In the fruit fly, Drosophila melanogaster, aversive olfactory learning forms several phases of labile memory to associate an odor with coincident punishment in the mushroom body (MB). It remains unclear how the brain consolidates early labile memory into LTM. This study surveyed 183 Gal4 lines containing almost all 21 distinct types of MB output neurons (MBONs) and showed that sequential synthesis of learning-induced proteins occurs at three types of MBONs. Downregulation of oo18 RNA-binding proteins (ORBs) in any of these MBONs impaired LTM. And, neurotransmission outputs from these MBONs are all required during LTM retrieval. Together, these results suggest an LTM consolidation model in which transient neural activities of early labile memory in the MB are consolidated into stable LTM at a few postsynaptic MBONs through sequential ORB-regulated local protein synthesis (Wu, 2017).

Go to top

Selective dendritic localization of mRNA in Drosophila mushroom body output neurons

Mitchell, J., Smith, C. S., Titlow, J., Otto, N., van Velde, P., Booth, M., Davis, I. and Waddell, S. (2021). Elife 10. PubMed ID: 33724180

Memory-relevant neuronal plasticity is believed to require local translation of new proteins at synapses. Understanding this process requires the visualization of the relevant mRNAs within these neuronal compartments. This study used single-molecule fluorescence in situ hybridization to localize mRNAs at subcellular resolution in the adult Drosophila brain. mRNAs for subunits of nicotinic acetylcholine receptors and kinases could be detected within the dendrites of co-labeled mushroom body output neurons (MBONs) and their relative abundance showed cell specificity. Moreover, aversive olfactory learning produced a transient increase in the level of CaMKII mRNA within the dendritic compartments of the γ5β'2a MBONs. Localization of specific mRNAs in MBONs before and after learning represents a critical step towards deciphering the role of dendritic translation in the neuronal plasticity underlying behavioral change in Drosophila (Mitchell, 2021).

Mammalian CaMKII mRNA is transported to neuronal dendrites, where it is locally translated in response to neuronal activity. Drosophila CAMKII is critical for behavioral plasticity and is also thought to be locally translated. However, fly CAMKII mRNAs have not been directly visualized within individual neurons. Therefore this study first hybridized CaMKII smFISH probes to whole-mount brains and imaged the mushroom body (MB) calyx, a recognizable neuropil containing the densely packed dendrites of ~2000 KCs and their presynaptic inputs from ~350 cholinergic olfactory projection neurons using a standard spinning disk confocal microscope. To detect and quantify mRNA within the 3D volume of the brain, a FIJI-compatible custom-built image analysis tool was developed that segments smFISH image data and identifies spots within the 3D volume using a probability-based hypothesis test. This enabled detection of mRNAs with a false discovery rate of 0.05. CaMKII smFISH probes labeled 56 ± 5 discrete puncta within each calyx. In comparison, smFISH probes directed to the α1 nicotinic acetylcholine receptor (nAChR) subunit labeled 33 ± 2 puncta in the calyx. Puncta were diffraction limited and the signal intensity distribution was unimodal, indicating that they represent single mRNA molecules (Mitchell, 2021).

Drosophila learning is considered to be implemented as plasticity of cholinergic KC-MBON synapses. To visualize and quantify mRNA specifically within the dendritic field of the γ5β'2a and γ1pedc>α/β MBONs, a membrane-tethered UAS-myr::SNAP reporter transgene was expressed using MBON-specific GAL4 drivers. This permitted simultaneous fluorescent labeling of mRNA with smFISH probes and the MBON using the SNAP Tag. To correct for chromatic misalignment that results from imaging heterogenous tissue at depth, brains were also co-stained with the dsDNA-binding dye Vybrant DyeCycle Violet (VDV). VDV dye has a broad emission spectrum so labeled nuclei can be imaged in both the SNAP MBON and smFISH mRNA channels. This triple-labeling approach allowed quantification and correction of any spatial mismatch between MBON and smFISH channels in x, y, and z planes, which ensures that smFISH puncta are accurately assigned within the 3D volume of the MBON dendritic field (Mitchell, 2021).

Using this smFISH approach, an average of 32 ± 2 CaMKII mRNAs was detected within the dendrites of γ5β'2a MBONs. However, in contrast to the calyx, no nAChRα1 was detected in γ5β'2a MBON dendrites. This differential localization of the CaMKII and nAChRα1 mRNAs within neurons of the mushroom body is indicative of cell specificity. To probe mRNA localization in MBONs more broadly, a single YFP smFISH probe set and a collection of fly strains harboring YFP insertions in endogenous genes were used. YFP insertions in the CaMKII, PKA-R2, and Ten-m genes were selected as test cases and the localization of their YFP-tagged mRNAs was compared between γ5β'2a MBON and γ1pedc>α/β MBON dendrites (Mitchell, 2021).

The CaMKII::YFP allele is heterozygous in flies also expressing myr::SNAP in MBONs. Therefore, YFP smFISH probes detected half the number of CaMKII mRNAs in γ5β'2a MBON dendrites compared to CaMKII-specific probes. Importantly, YFP probes hybridized to YFP-negative control brains produced background signal that was statistically distinguishable in brightness from genuine smFISH puncta. Comparing data from YFP-negative and YFP-positive samples allowed definition of the false discovery rate to be 14% when using YFP-directed probes. These results indicate that the YFP probes are specific and that the YFP insertion does not impede localization of CaMKII mRNA. A similar abundance of CaMKII::YFP was detected in the dendritic field of γ5β'2a and the γ1 dendritic region of γ1pedc>α/β MBONs. In contrast, more PKA-R2 mRNAs were detected in the dendrites of γ5β'2a MBONs compared to γ1pedc>α/β MBONs. Importantly, the relative abundance of dendritically localized CaMKII and PKA-R2 mRNAs did not simply reflect the levels of these transcripts detected in the MBON somata. In addition, Ten-m mRNAs was not detected in either γ5β'2a or γ1pedc>α/β MBON dendrites, although they were visible in neighboring neuropil and at low levels in the MBON somata. These results suggest that CaMKII and PKA-R2 mRNAs are selectively localized to MBON dendrites (Mitchell, 2021).

Although nAChRα1 mRNA was not detected within γ5β'2a MBON dendrites, prior work has shown that nAChR subunits, including nAChRα1, are required in γ5β'2a MBON postsynapses to register odor-evoked responses and direct odor-driven behaviors. Since the YFP insertion collection does not include nAChR subunits, nAChRα5 and nAChRα6-specific smFISH probes were designed. These probes detected nAchRα5 and nAchRα6 mRNAs within γ5β'2a and γ1pedc>α/β MBON dendrites, with nAchRα6 being most abundant. Importantly, nAchRα1, nAchRα5, and nAchRα6 were detected at roughly equivalent levels in the γ5β'2a and γ1pedc>α/β MBON somata. Therefore, the selective localization of nAchRα5 and nAchRα6 mRNA to MBON dendrites indicates that these receptor subunits may be locally translated to modify the subunit composition of postsynaptic nAChR receptors (Mitchell, 2021).

Localized mRNAs were on average 2.8x more abundant in γ5β'2a relative to the γ1 region of γ1pedc>α/β MBON dendrites. Therefore whether this apparent differential localization correlated with dendritic volume and/or the number of postsynapses between these MBONs was tested. Using the recently published electron microscope volume of the Drosophila 'hemibrain', the dendritic volume of the γ5β'2a MBON was calculated to be 1515.36 nm3 and the γ1 region of the γ1pedc>α/β MBON was calculated to be 614.20 nm3. In addition, the γ5β'2a regions of the γ5β'2a MBON dendrite contain 30,625 postsynapses, whereas there are only 17,020 postsynapses in the γ1 region of the γ1pedc>α/β MBON. Larger dendritic field volume and synapse number is therefore correlated with an increased number of localized nAchRα5, nAchRα6, and PKA-R2 mRNAs. The correlation, however, does not hold for CaMKII mRNA abundance. Selective localization of mRNAs to MBON dendrites therefore appears to be more nuanced than simply reflecting the size of the dendritic arbor, the number of synapses, or the level of transcripts detected throughout the cell (Mitchell, 2021).

Whether CaMKII::YFP mRNA abundance in γ5β'2a and γ1pedc>α/β MBONs was altered following aversive learning was tested. mRNA in the somata and nuclei of these MBONs was quantified. Transcriptional activity is indicated by a bright nuclear transcription focus. Flies were initially subjected to four conditions: (1) an 'untrained' group that was loaded and removed from the T-maze but not exposed to odors or shock; (2) an 'odor only' group, exposed to the two odors as in training but without shock; (3) a 'shock only' group that was handled as in training and received the shock delivery but no odor exposure; and (4) a 'trained' group that was aversively conditioned by pairing one of the two odors with shock. Fly brains were extracted 10 min, 1 hr, or 2 hr after training and processed for smFISH (Mitchell, 2021).

CaMKII mRNA increased significantly in γ5β'2a MBON dendrites 10 min after training compared to all control groups. Including an additional 'unpaired' experiment, where odor and shock presentation was staggered, confirmed that the increase at 10 min after training requires coincident pairing of odor and shock. Moreover, levels returned to baseline by 1 hr and remained at that level 2 hr after training. CaMKII mRNAs in γ5β'2a MBON somata showed a different temporal dynamic, with transcripts peaking 1 hr after training, albeit only relative to untrained and odor only controls. The proportion of γ5β'2a nuclei containing a CaMKII transcription focus did not differ between treatments, suggesting that the transcript increase in the somata is not correlated with the number of actively transcribing γ5β'2a nuclei, at least at the timepoints measured. In addition, the mean brightness of γ5β'2a transcription foci did not change across treatments, although the variation was substantial. An increase of dendritically localized CaMKII mRNAs could result from enhanced trafficking or through the release of transcripts from protein bound states, which would increase smFISH probe accessibility and hence spot brightness. Since the brightness of CaMKII mRNA spots detected in the dendrites of γ5β'2a MBONs did not change with treatment, it is concluded that the increased abundance likely results from altered traffic (Mitchell, 2021).

Assessing CaMKII mRNA abundance in γ1pedc>α/β MBONs after learning did not reveal a change in mRNA abundance in the dendrites or somata between trained flies and all control groups at all timepoints measured. These results indicate specificity to the response observed in the γ5β'2a MBONs (Mitchell, 2021).

Since CaMKII protein is also labeled with YFP in CaMKII::YFP flies, protein expression was assessed by measuring YFP fluorescence intensity specifically within the MBON dendrites. This analysis did not reveal a significant difference in fluorescence intensity across treatments. However, since smFISH provides single-molecule estimates of mRNA abundance, a similar level of single-molecule sensitivity may be required to detect subcellular resolution changes in protein copy number. Moreover, new synthesis and replacement of specific isoforms of CaMKII could radically change local kinase activity, even without an observable change in overall abundance (Mitchell, 2021).

Early studies in Drosophila demonstrated that broad disruption of CAMKII function impaired courtship learning. In contrast, later studies that manipulated activity more specifically in olfactory projection neurons or particular classes of KCs reported a preferential loss of middle-term or long-term olfactory memory. This study focused on two subtypes of MBONs, that are known to exhibit changes in odor-evoked activity after a single trial of aversive olfactory conditioning. Whereas γ1pedc>α/β MBON responses to the previously shock-paired odor are depressed immediately after aversive learning, prior studies observed a learning-related increase of the conditioned odor response of γ5β'2a MBONs, likely resulting from a release of feedforward inhibition from γ1pedc>α/β MBONs. It is therefore speculated that the specific change in CaMKII mRNA abundance in the γ5β'2a MBONs after aversive learning might be a consequence of network-level potentiation of their activity, such as would result from a release from inhibition. Since CAMKII local translation-dependent plasticity is expected to underlie more extended forms of memory, it will be interesting to investigate whether the training-evoked change in CaMKII mRNA abundance in the γ5β'2a MBON dendrites contributes to later aversive memory formation and maintenance. This may be possible with MBON-specific targeting of CAMKII mRNAs that contain the long 3'UTR, which is essential for dendritic localization and activity-dependent local translation (Mitchell, 2021).

Go to top

Local translation provides the asymmetric distribution of CaMKII required for associative memory formation

Chen, N., Zhang, Y., Adel, M., Kuklin, E. A., Reed, M. L., Mardovin, J. D., Bakthavachalu, B., VijayRaghavan, K., Ramaswami, M. and Griffith, L. C. (2022). Curr Biol 32(12): 2730-2738. PubMed ID: 35545085 BioArchives

How compartment-specific local proteomes are generated and maintained is inadequately understood, particularly in neurons, which display extreme asymmetries. This study shows that local enrichment of Ca(2+)/calmodulin-dependent protein kinase II CaMKII) in axons of Drosophila mushroom body neurons is necessary for cellular plasticity and associative memory formation. Enrichment is achieved via enhanced axoplasmic translation of CaMKII mRNA, through a mechanism requiring the RNA-binding protein Mub and a 23-base Mub-recognition element in the CaMKII 3' UTR. Perturbation of either dramatically reduces axonal, but not somatic, CaMKII protein without altering the distribution or amount of mRNA in vivo, and both are necessary and sufficient to enhance axonal translation of reporter mRNA. Together, these data identify elevated levels of translation of an evenly distributed mRNA as a novel strategy for generating subcellular biochemical asymmetries. They further demonstrate the importance of distributional asymmetry in the computational and biological functions of neurons (Chen, 2022).

Local protein synthesis at synapses has been studied extensively in the context of specialized processes like activity-dependent plasticity and axon guidance. Recent theory and experimental work, however, suggests that local translation occurs much more generally and may be used to establish differential proteomes in functionally-specialized subcellular regions. This study resolves two long-standing questions about CaMKII: how and why it achieves extraordinary levels in axons. It was demonstrated that resting adult levels of CaMKII protein are translationally accrued, and that the high levels in this compartment form a computational scaffold critical for formation of associative memory and the cellular memory trace. While previous studies using mutants and RNAi have shown a role for CaMKII in plasticity, the current manipulations of the 3'UTR, which do not affect somatic kinase levels, establish the necessity of synaptic enrichment. This enrichment requires cis-elements present only in the long form of the 3'UTR and Mub, the Drosophila poly-C-binding-protein homolog demonstrating a new, activity-independent function for the CaMKII 3'UTR (Chen, 2022).

Activity-dependent translation and differential polyadenylation are ancient conserved features of CaMKII mRNAs. For mammalian CAMK2A, early work in which the 3'UTR was deleted demonstrated its requirement for mRNA stability and dendritic localization, and also for protein accumulation and activity-dependent synthesis (Chen, 2022).

A handful of studies attempted to identify cis-elements regulating dendritic CAMK2A mRNA localization and transport, but there is as yet no information on 3'UTR cis-elements controlling translation, though in silico prediction suggests that the CAMK2A 3'UTR may have polyC-binding protein motifs (Chen, 2022).

At the Drosophila larval neuromuscular junction, it has been shown that the CaMKII 3'UTR controls activity-dependent synthesis of CaMKII. The fact that the rodent CAMK2A 3'UTR can support activity-dependent protein synthesis in the fly suggests that there will be shared mechanisms for this aspect of CaMKII regulation. But while there are many similarities between mammals and flies, there are also differences. In Drosophila, the 3'UTR appears to have little effect on mRNA localization, and only a small effect on stability that is ascribable to a proximal cis-element. How CaMKII mRNA reaches synapses in Drosophila is yet to be determined, but the differences in localization mechanism may reflect the ca. 100-fold difference in distances that mRNAs need to travel to reach synapses (Chen, 2022).

The ability of Mub, which is present at low levels in MB axons and at high levels in MB and other cell bodies, to specifically regulate axonal accumulation of CaMKII protein without affecting somatic protein levels suggests several models. One possibility is that MB axons have either compartment-specific translational machinery or a distinct set of auxiliary proteins that allow Mub to regulate axonal ribosomes. The presence of Mub protein in MB axons, but not in other neuropils, may indicate the existence of unique translational complexes in that compartment. Another possibility is that Mub is a general translation enhancer, but MB soma contain repressor proteins that locally inhibit its actions. This would be consistent with the finding that there are cis elements that appear to act as general repressors in the CaMKII 3'UTR. While these ideas remain speculative, the robust interaction of Mub with CaMKII provides an opportunity to deepen understanding of how local protein synthesis can shape neuronal function and build the synaptic proteome (Chen, 2022).

Go to top

Transsynaptic mapping of Drosophila mushroom body output neurons

Scaplen, K. M., Talay, M., Fisher, J. D., Cohn, R., Sorkac, A., Aso, Y., Barnea, G. and Kaun, K. R. (2021). Elife 10. PubMed ID: 33570489

The mushroom body (MB) is a well-characterized associative memory structure within the Drosophila brain. Analyzing MB connectivity using multiple approaches is critical for understanding the functional implications of this structure. Using the genetic anterograde transsynaptic tracing tool, trans-Tango, this study identified divergent projections across the brain and convergent downstream targets of the MB output neurons (MBONs). This analysis revealed at least three separate targets that receive convergent input from MBONs: other MBONs, the fan-shaped body (FSB), and the lateral accessory lobe (LAL). A multilayer circuit is described, both anatomically and functionally, in which inhibitory and excitatory MBONs converge on the same genetic subset of FSB and LAL neurons. This circuit architecture enables the brain to update and integrate information with previous experience before executing appropriate behavioral responses. This use of trans-Tango provides a genetically accessible anatomical framework for investigating the functional relevance of components within these complex and interconnected circuits (Scaplen, 2021).

Go to top

An optogenetic analogue of second-order reinforcement in Drosophila

Konig, C., Khalili, A., Niewalda, T., Gao, S. and Gerber, B. (2019). An optogenetic analogue of second-order reinforcement in Drosophila. Biol Lett 15(7): 20190084. PubMed ID: 31266421

In insects, odours are coded by the combinatorial activation of ascending pathways, including their third-order representation in mushroom body Kenyon cells. Kenyon cells also receive intersecting input from ascending and mostly dopaminergic reinforcement pathways. Indeed, in Drosophila, presenting an odour together with activation of the dopaminergic mushroom body input neuron PPL1-01 leads to a weakening of the synapse between Kenyon cells and the approach-promoting mushroom body output neuron MBON-11. As a result of such weakened approach tendencies, flies avoid the shock-predicting odour in a subsequent choice test. Thus, increased activity in PPL1-01 stands for punishment, whereas reduced activity in MBON-11 stands for predicted punishment. Given that punishment-predictors can themselves serve as punishments of second order, whether presenting an odour together with the optogenetic silencing of MBON-11 would lead to learned odour avoidance was tested, and this was found to be the case. In turn, the optogenetic activation of MBON-11 together with odour presentation led to learned odour approach. Thus, manipulating activity in MBON-11 can be an analogue of predicted, second-order reinforcement (Konig, 2019).

Go to top

Integration of parallel opposing memories underlies memory extinction

Felsenberg, J., Jacob, P. F., Walker, T., Barnstedt, O., Edmondson-Stait, A. J., Pleijzier, M. W., Otto, N., Schlegel, P., Sharifi, N., Perisse, E., Smith, C. S., Lauritzen, J. S., Costa, M., Jefferis, G., Bock, D. D. and Waddell, S. (2018). Cell 175(3):709-722. PubMed ID: 30245010

Accurately predicting an outcome requires that animals learn supporting and conflicting evidence from sequential experience. In mammals and invertebrates, learned fear responses can be suppressed by experiencing predictive cues without punishment, a process called memory extinction. This study shows that extinction of aversive memories in Drosophila requires specific dopaminergic neurons, which indicate that omission of punishment is remembered as a positive experience. Functional imaging revealed co-existence of intracellular calcium traces in different places in the mushroom body output neuron network for both the original aversive memory and a new appetitive extinction memory. Light and ultrastructural anatomy are consistent with parallel competing memories being combined within mushroom body output neurons that direct avoidance. Indeed, extinction-evoked plasticity in a pair of these neurons neutralizes the potentiated odor response imposed in the network by aversive learning. Therefore, flies track the accuracy of learned expectations by accumulating and integrating memories of conflicting events (Felsenberg, 2018).

Extinction was first described by Pavlov in his experiments with dogs. Although extinction is broadly believed to result from new inhibitory learning, rather than erasure of the original memory, the underlying neural mechanisms have remained elusive. This study describes how competing memories arise and are integrated to extinguish aversive memory in Drosophila (Felsenberg, 2018).

Extinction of aversive memory required PAM dopaminergic neurons during the period of odor re-exposure. Some of these DANs provide teaching signals when flies are trained with odor and sugar or water reward. Importantly, sugar reward learning mediated by these DANs induces relative depression of CS+ odor-evoked responses in M4β'/M6 MBONs, which was also observed following extinction of aversive memory. Since reduced odor-driven activity in M6 MBONs is enough to convert odor avoidance behavior into attraction, plasticity of aversive memory extinction can be considered to be appetitive. These results together suggest that absence of predicted punishment is coded in the fly brain in a similar way to positive experience. But how can lack of punishment lead to a potential reward signal (Felsenberg, 2018)?

Previous data and those presented in this study suggest that aversive learning reconfigures the MBON network into a state primed to preferentially drive a reward teaching signal, when the flies re-experience trained odor without punishment. Prior work showed that aversive learning depresses conditioned odor drive to the KC-MVP2 MBON pathway, that favors approach behavior. Furthermore, like the role for disinhibition in mice, aversive learning reduces MVP2-mediated feedforward inhibition in the network and thereby also indirectly potentiates M4β'/M6 MBON odor responses that drive avoidance behavior. Since some avoidance directing MBONs can provide recurrent input to PAM DANs, odor re-exposure after aversive learning should preferentially drive a positive teaching signal via these MBONs. When directly triggered, glutamatergic M4β' and M6 neurons selectively activated DANs releasing dopamine in the γ5 compartment. Finding that extinction induced a corresponding depression of conditioned odor drive to M6 neurons is therefore also consistent with the previously trained odor activating γ5 DANs, to direct odor-specific plasticity at KC-M6 synapses (Felsenberg, 2018).

It is not known whether extinction-relevant γ5 DANs are the same as those providing water or sugar reward-teaching signals. Despite expectations, it was not possible to observe increased odor-evoked activity in γ5 DANs after aversive learning, using GCaMP6m. R58E02-GAL4-labeled γ5 DANs exhibited robust oscillatory activity, which impeded reliable recording of odor-evoked events. Some γ5 DANs may oscillate and others be cue evoked, but the genetic tools to direct transgene expression to meaningful subsets are lacking. Nevertheless, there are between 8 and 21 γ5 DANs and γ5 presynaptic innervation within that MB compartment may be further segregated. If individual γ5 DANs have input and output specificity, different KC-M6 synapses along the same odor-activated KC would be modified by sugar reward learning and aversive memory extinction, thereby expanding the coding range within the KC-MBON network. Nevertheless, this level of potential synaptic specificity of reward learning and extinction would still generate a similar odor-specific depression when recording broad odor-evoked signals from M6 dendrites. Although anatomical specificity is appealing and not at odds with the current and prior data, it will be essential to determine how individual γ5 DANs operate and analyze KC-M6 dendritic plasticity at higher resolution (Felsenberg, 2018).

Without knowing the specific location of extinction-driven synaptic plasticity, the model predicts that if punishment does not follow conditioned odor presentation, extinction plasticity triggered at the KC-M6 MBON junction readjusts the balance in the MBON network. Whereas if shock were to follow, extinction plasticity would be offset by additional modification made to the site of the original aversive memory. An opposite scenario is assumed to underlie the extinction of appetitive memory, which is initially coded as depression of conditioned odor drive to avoidance directing MBONs. Re-exposing flies to the conditioned odor, without sugar, neutralizes odor driven approach. However, this process instead required aversively reinforcing DANs, some of which are functionally connected to approach directing MBONs. Omission of predicted reward therefore appears to be coded as aversive experience. Taken with data in this study, it is proposed that DAN-driven formation of a competing memory of opposite valence is a general, and likely conserved, feature of memory extinction (Felsenberg, 2018).

Prediction error, an unexpected change in reward or punishment contingency, has a strong theoretical and experimental foundation in mammalian dopaminergic neurons. However, it is not clear how errors are registered and how dopaminergic activity alters the underlying network. By coding valence of learning as a particular skew in the MBON network, the fly can use opposing arms of the DAN system to keep track of when expected contingencies between odors and positive or negative events are not met. Such a model predicts that odors that are learned to be avoided will preferentially trigger appetitively reinforcing DANs if punishment does not follow, whereas odors learned to be approached will more strongly activate aversive DANs and be registered as bad, if the expected reward is omitted (Felsenberg, 2018).

Physiological traces of the original aversive memory, and new extinction memory were observed in different nodes of the MBON network at the same time after training. An aversive memory trace measurable in the dendrites of MVP2 neurons survived extinction, while a new extinction trace arose in the odor responsiveness of M6 neurons. Although functional imaging suggests that the change in relative odor drive from KCs to MVP2 MBONs that accompanies aversive learning remains after extinction, it is not certain sure that it results from the same unaltered synaptic or neural mechanism (Felsenberg, 2018).

Flies simultaneously form parallel memories of opposite valence, if trained with odor and sugar laced with bitter taste. These separate aversive and appetitive memories compete to guide either learned odor avoidance or approach behavior. Since aversive memory followed by extinction is equivalent to sequential formation of parallel memories, it follows that a new extinction memory written in the KC-M6 MBON connection by γ5 DANs, can partially neutralize behavioral expression of the original aversive memory, formed at the KC-MVP2 junction. Since multiple MBON pathways (e.g., MVP2 and V2α) are modified by aversive learning, but only the KC-M6 junction is modified by extinction (not KC-M4β'), an imbalanced number of plastic connections might account for the partial nature of aversive memory extinction (Felsenberg, 2018).

The apparent stability of learning induced changes in odor-evoked activity in MVP2 neurons after extinction, taken with retraining experiments indicate that flies can accumulate information across training, extinction, and retraining trials. It is proposed that retention of learned information following extinction is a fundamental feature of a memory network. Combining supporting and conflicting information from consecutive experience is certainly a prerequisite for more complex probabilistic learning (Felsenberg, 2018).

MVP2 neurons innervate multiple compartments of the MB and appear to make different connections with vertical and horizontal lobe MBONs. Ultrastructure shows that an MVP2 neuron forms distinct synaptic connections with M4β' and M6 MBONs. Whereas MVP2 makes large bouton-type synapses onto M4β' distal dendrites, MVP2 forms en passant synapses along M6 primary neurites. These connections are reminiscent of those made by unique types of mouse GABA-ergic neurons (Felsenberg. 2018).

Recent EM reconstruction of the larval MB wiring diagram described connections between MBONs, and convergence neurons pooling collections of MBON inputs. This study found that aversive and extinction memories are already integrated within the MBON network and specifically in M6 neurons, that promote avoidance. The learning induced potentiated odor-response in M6, resulting from reduced MVP2 mediated inhibition, appeared nullified by addition of odor-specific depression of the KC-M6 connection. This suggests that extinction memory can suppress expression of the original aversive memory and consequently learned odor avoidance behavior (Felsenberg. 2018).

It is not known how Drosophila appetitive memories are countered by their corresponding extinction memory to suppress conditioned approach. At present the MBON network architecture looks more complex than a straightforward 'winner-takes-all' scenario involving direct reciprocal inhibitory connections between approach and avoidance directing pathways (Felsenberg, 2018).

This work exclusively studied extinction soon after training. Prior studies in flies and other animals suggest processes might differ at later times. Given expression of longer-term memories is apparently more reliant on αβ than γ KCs, it is possible odor re-exposure at later times will drive a different imbalanced MBON network configuration than that earlier on. In this case, other appetitively reinforcing DANs, and plasticity at different KC-MBON junctions, might be required to acquire a competing extinction memory at that time (Felsenberg, 2018).

Sometimes extinguished memories spontaneously recover with time, consistent with a new memory temporarily suppressing previous learned behavior (Rescorla, 2004, Bouton, 2006). In Drosophila, spontaneous recovery of extinguished aversive memory is time dependent. Memories extinguished 2 days after training remain low for 4 days, whereas those extinguished at 5 days recover 4 days later (Hirano, 2016). Recovery of extinguished memories could be accompanied by loss of odor-specific plasticity in KC-M6 dendrites. Furthermore, the ability of extinguished memories to recover might result from the relative strength of KC-MBON connections in which the original aversive memory resides, and the extinction memory is formed, at the time the fly re-encounters the CS+ without punishment (Felsenberg, 2018).

Some reward-activated mammalian DANs also respond to absence of an expected aversive stimulus. Therefore, fear extinction also could be triggered by appetitively reinforcing DANs. Acquisition and extinction of fear memory involves plasticity in basolateral amygdala (BLA), which contains distinct neural paths for fear and reward memories. Perhaps an analogous arrangement of parallel competing memories, driven by teaching signals from BLA-projecting DANs, extinguishes mammalian fear (Felsenberg, 2018).

An early mechanistic study of Drosophila extinction concluded that aversive learning and its extinction both occur within the same subset of KCs. In addition, it was proposed that extinction involved intracellular antagonism with cAMP signaling that is required for memory formation. These data suggest initial aversive learning and subsequent extinction are coded as consecutive learning events within the same odor-activated KCs. However, two parallel memories are formed within anatomically separate output compartments of the same KCs where they synapse onto different MBONs. Learned behavior is therefore extinguished as a result of intercellular antagonism within the output layer of the MB network. This process is likely reliant on the extended architecture of KCs that separates KCs' primary sensory input layer in the MB calyx from a compartmentalized error adjustment layer in the lobes. Activity in populations of KCs therefore represents specific odors, whereas associated values, such as unexpected shock and absence of predicted shock, can be independently and locally assigned to odors by altering the weights of synapses in different output compartments from the same KCs (Felsenberg. 2018).

Go to top

Activity of defined mushroom body output neurons underlies learned olfactory behavior in Drosophila

Owald, D., Felsenberg, J., Talbot, C. B., Das, G., Perisse, E., Huetteroth, W. and Waddell, S. (2015). Neuron 86: 417-427. PubMed ID: 25864636

During olfactory learning in fruit flies, dopaminergic neurons assign value to odor representations in the mushroom body Kenyon cells. This study identified a class of downstream glutamatergic mushroom body output neurons (MBONs) called M4/6, or MBON-β2β'2a, MBON-β'2mp, and MBON-γ5β'2a, whose dendritic fields overlap with dopaminergic neuron projections in the tips of the β, β', and γ lobes. This anatomy and their odor tuning suggests that M4/6 neurons pool odor-driven Kenyon cell synaptic outputs. Like that of mushroom body neurons, M4/6 output is required for expression of appetitive and aversive memory performance. Moreover, appetitive and aversive olfactory conditioning bidirectionally alters the relative odor-drive of M4β' neurons (MBON-β'2mp). Direct block of M4/6 neurons in naive flies mimics appetitive conditioning, being sufficient to convert odor-driven avoidance into apprroach, while optogenetically activating these neurons induces avoidance behavior. It is therefore proposed that drive to the M4/6 neurons reflects odor-directed behavioral choice. See Three Pairs of Glutamatergic Output Neurons Innervate the Tips of the Horizontal Mushroom Body Lobes (Owald, 2015).

Many prior studies have concluded that mushroom body neurons are dispensable for naive odor-driven behavior and subsets are either required or are dispensable for particular memory functions. However, these experiments simultaneously blocked all the outputs from a given population of KCs using cell-wide expression of shits1. The current results suggest that these models should be reconsidered. Blocking the specific M4β/MBON-β2β'2a, M4β'/MBON-β'2mp, and M6/MBON-γ5β'2a output from the mushroom body, as opposed to blocking all outputs, has a radical effect on naive odor-driven behavior. It is proposed that ordinarily, in naive flies, the multiple mushroom body output channels are ultimately pooled and contribute a net zero to odor-driven behavior. Therefore, if one uses a mushroom body neuron-driven UAS-shits1 that simultaneously blocks all outputs, there is no apparent effect on naive behavior. If, however, one blocks only one channel, or alters its efficacy by learning, the odor-driven behavior can be changed. A similar logic could also account for why clear memory retrieval defects are seen when blocking M4β'/MBON-β'2mp and M6/MBON-γ5β'2a neurons that presumably pool outputs from the tip of the γ and β' lobe, yet blocking all α'β' neuron outputs did not demonstrably disrupt later memory retrieval. Others have shown a role for α'β' neuron output to retrieve earlier forms of memory (Owald, 2015).

Both the physiological and behavioral results are consistent with a depression of the M4β'/MBON-β'2mp and M6/MBON-γ5β'2a output being sufficient to code learned approach. Learning-related plasticity has been reported at the β-lobe outputs in both bees and locusts, although the importance of these synaptic connections in the behavior of these insects is not known. At this stage it is not certain that the observed decrease in the relative odor drive reflects plasticity of the synapses between odor-specific KCs and the M4/6 neurons. However, it seems plausible, because this synaptic junction is addressed by the relevant rewarding dopaminergic neurons. Given that blocking M4β'/MBON-β'2mp and M6/MBON-γ5β'2a neurons converts avoidance to approach, other mushroom body output channels, perhaps some of which lie on the vertical α-lobe projection, must drive the approach behavior. It is therefore conceivable that a similar plasticity of odor drive to these putative approach outputs could be critical for aversive conditioning. Such an idea is consistent with several prior reports of aversive memory traces that are specific to the vertical α-branch of the mushroom body. In addition, aversive learning has been reported to depress odor drive in the vertical lobe of downstream MB-V2α/MBON-α2sc and MB-V2α'/MBON-α'3 neurons and to potentiate odor drive of MB-V3/MBON-α3 output neurons. However, it is notable that blocking either the MB-V2α/MBON-α2sc and MB-V2α'/MBON-α'3 neurons or MB-V3/MBON-α3 neurons did not affect naive odor avoidance behavior in the current experiments or those of others. Therefore, although MB-V2α/MBON-α2sc, MB-V2α'/MBON-α'3, and MB-V3/MBON-α3 neurons are required for memory expression, it is not currently known which reinforcing neurons address MB-V2α/MBON-α2sc, MB-V2α'/MBON-α'3, and MB-V3/MBON-α3 connections and how these outputs specifically contribute to odor-guided behavior (Owald, 2015).

The physiological analyses suggest bidirectional plasticity of odor-evoked responses, with aversive learning increasing the relative conditioned odor drive to the M4β'/MBON-β'2mp neurons. This could account for why output from M4/6 neurons is also required for expression of aversive memory. Moreover, whereas blocking the M4β/MBON-β2β'2a, M4β'/MBON-β'2mp, and M6/MBON-γ5β'2a neurons converts odor avoidance into approach, activation of M4β/MBON-β2β'2a, M4β'/MBON-β'2mp, and M6/MBON-γ5β'2a neurons drives avoidance. It therefore seems likely that plasticity of the relative odor drive to M4β'/MBON-β'2mp neurons is also part of the aversive memory engram. Again, it is not known that the increased odor drive after training reflects synaptic potentiation between odor-specific KCs and the M4β'/MBON-β'2mp neurons. Increased odor drive to M4β'/MBON-β'2mp neurons could, for example, also result from plasticity elsewhere in the KCs that enhances signal propagation along the horizontal KC arbor. Nevertheless, the MB-M3 dopaminergic neurons that are required to reinforce aversive memory also innervate the tips of the β and β' lobe. In addition, a recent study reported that aversive learning specifically decreased unconditioned odor-evoked neurotransmission from the γ neurons, a result that presumably would mirror a relative increase in the response to the reinforced odor. Lastly, aversive conditioning using relative shock intensity utilizes the rewarding dopaminergic neurons that occupy the same zones on the mushroom body as the M4β'/MBON-β'2mp and M6/MBON-γ5β'2a neuron dendrites. With the caveat that GRASP is only an indicator of proximity, the anatomical studies suggest that dendrites of rewarding dopaminergic neurons may connect to the M4β'/MBON-β'2mp and M6/MBON-γ5β'2a neuron presynaptic terminals, forming a potential feedback or forward loop that could serve such a relative-judgment function (Owald, 2015).

It is perhaps noteworthy that KC outputs in the vertical lobe are onto excitatory cholinergic MB-V2α/MBON-α2sc and MB-V2α'/MBON-α'3 neurons, whereas the horizontal outputs are onto glutamatergic, potentially inhibitory, M4β/MBON-β2β'2a, M4β'/MBON-β'2mp, and M6/MBON-γ5β'2a neurons. This suggests that distinct signaling modes may be driven from the bifurcated collaterals of KCs. It will be crucial to understand how these outputs from the different branches, and those from discrete lobes, are ultimately pooled to guide appropriate behavior (Owald, 2015).

Go to top

Mushroom body output differentiates memory processes and distinct memory-guided behaviors

Musso, P. Y., Lampin-Saint-Amaux, A., Tchenio, P. and Preat, T. (2017). Nat Commun 8(1): 1803. PubMed ID: 29180783

The mushroom body (MB) of Drosophila melanogaster has multiple functions in controlling memory and behavior. This study systematically probed the behavioral contribution of each type of MB output neuron (MBON) by blocking during acquisition, retention, or retrieval of reward or punishment memories. The contribution was evaluated using two conditioned responses: memory-guided odor choice and odor source attraction. Quantitative analysis revealed that these conditioned odor responses are controlled by different sets of MBONs. The valence of memory, rather than the transition of memory steps, has a larger impact on the patterns of required MBONs. Moreover, it was found that the glutamatergic MBONs forming recurrent circuits commonly contribute to appetitive memory acquisition, suggesting a pivotal role of this circuit motif for reward processing. These results provide principles how the MB output circuit processes associative memories of different valence and controls distinct memory-guided behaviors (Ichinose, 2021).

Distinct MBONs are required for odor choice guided by appetitive and aversive memories Neural structures in the central nervous system can have functional roles across a wide range of cognitive tasks. The mushroom body (MB) in Drosophila melanogaster has served as a unique circuit model to study such pleiotropy. So far, many aspects of behavior and physiology, such as sensory processing, learning, and sleep/arousal, have been shown to be under the control of the MB. Particularly, MBs play important roles in different forms of associative learning: appetitive and aversive learning of odor and color. Transient blockade of the MB intrinsic neurons, Kenyon cells (KCs), revealed that neurotransmission from the MB controls not only retrieval but also acquisition and retention of appetitive and aversive olfactory memories. Investigation of MB output should thus be a key to understand how MBs process diverse memory-related functions (Ichinose, 2021).

The anatomy of the MB output is well characterized. The entire MB lobes can be subdivided into 15 compartments, based on the dendritic arbors of 21 types of MB output neurons (MBONs). These MBONs are post-synaptic to KCs and dopamine neurons and project their axons to defined neuropils of the brain. Selective requirement and plasticity in specific KC-MBON synapses during memory processing further corroborate distinct functions of the MB. However, because these studies have been focused on specific MBON subtypes and based on experiments under different conditions, direct comparisons are difficult. Therefore, this study sought to systematically examine the pattern of the MBON requirements under the same experimental condition (Ichinose, 2021).

To this end, each MBON type was blocked using a set of 13 split-GAL4 driver lines that collectively cover all the identified MBON subtypes. The temperature-sensitive, dominant-negative dynamin Shibirets1 (Shits1) was expressed by each split-GAL4 driver. Using different temperature-shifting protocols, the blockade of target MBONs was restricted to each step of memory processing-acquisition, retention, or retrieval-or the blockade (permissive control) was omitted. Performance indices (PIs) were calculated based on the conditioned odor choice at 1 h after appetitive or aversive conditioning using sucrose reward or electric shock punishment. Flies were starved for appetitive conditioning and test (Ichinose, 2021).

To quantitatively compare patterns of MBON requirement among different memory steps, an index called the contribution score was introduced. It was calculated by comparing the PI of an experimental group to that of the control genotype without the driver. The experimental PI was normalized to eliminate potential genetic and temperature effects. The score represents the memory loss caused by the MBON blockade. Zero indicates no memory impairment, and one denotes the complete memory loss. An improved performance causes a negative score (Ichinose, 2021).

The MBON contribution scores in the three steps of reward and punishment memories revealed striking distinctions of the requirement patterns: each memory step is mediated by a unique but partially overlapping combination of MBONs. Therefore the similarity of MBON contribution patterns was quantified among the 6 different memory steps. Interestingly, a hierarchical cluster analysis grouped the three steps in reward and punishment memories together. This suggests that the valence of memory, compared to the transition of memory steps, has more influence on the pattern of MBON contribution. Moreover, this study found that MBON contribution patterns are more similar for retrieval and acquisition than retention both in reward and punishment memories. This symmetric result suggests similar MB output usage during encoding and decoding of memory, possibly reflecting the presence of olfactory stimulation (Ichinose, 2021).

Next the behavioral results of each MBON type was projected on a standard brain by representing the contribution scores of all three steps as its intensity. The MBON types were color coded according to the clusters of dopamine neurons (DANs) innervating to them. DANs in the MB originate from the three distinct cell clusters, PAM, PPL1, and PPL2ab. The PAM and PPL1 cluster DANs generally convey the positive and negative valence, respectively, were found in different forms of associative learning. MBONs under the direct control of PAM cluster DANs exhibit remarkably higher contributions to reward than punishment memory. In contrast, those MBONs corresponding to the terminals of the PPL1 or PPL2ab clusters were selectively required for punishment memory. These results generally fit to the local modulation in the MB by individual DANs. Therefore, the valence representation of DANs is globally maintained in the MB output layer for memory-guided odor choice (Ichinose, 2021).

The MBONs can be categorized into three classes based on the neurotransmitters they release: acetylcholine, γ-aminobutyric acid (GABA), and glutamate. To examine the preferential role of neurotransmitters, MBON contributions were grouped according to the neurotransmitter systems presented in different colors. Reward memory preferentially depends on glutamatergic MBONs, whereas the contribution of GABAergic MBON-γ1pedc is most conspicuous in punishment memory. These results are largely consistent with previous investigations about compartment-specific roles of MBONs and DANs (Ichinose, 2021).

Many MBONs have been shown to send their terminals to the dendrites of DANs. These MBON-DAN connections were classified into the recurrent (i.e., MBON dendrites and DAN terminals share the same MB compartment) and the others and found that the recurrent connections are predominantly glutamatergic. Interestingly, all those glutamatergic MBONs with major contributions to reward memory (MBON-α1, -β1, -γ5β'2a, and -β'2mp) form recurrent circuits. These recurrent connections were verified with multiple approaches. Imaging physically expanded immunolabelled brains using a water-dipping objective with a high numerical aperture resolved that Bruchpilot-marked active zones of MBON boutons abut on DAN dendrites, suggesting direct synapses. Furthermore, an anterograde transsynaptic circuit tracing technique, trans-Tango, corroborated the connectivity. A recent connectome study reached the same conclusion. These classes of glutamatergic MBONs are all required during the acquisition of reward memory, and the corresponding DANs are implicated in mediating reward. It is suggested that the glutamate-dopamine recurrent connection in the MB is a circuit motif that controls reward processing (Ichinose, 2021).

It was hypothesized that the MBONs with the glutamatergic feedback motif plays a role in sustaining intense reward signals, as previously proposed for the feedback from MBON-α1 to PAM-α1 neurons. Therefore, the roles of MBON-α1, -β1, -γ5β'2a, and -β'2mp were tested for appetitive memory acquisition by varying conditioning lengths. Interestingly, this study found that MBON-α1 and -β1 were selectively required for longer conditioning. Therefore, it is proposed that reverberating input from these MBONs could sustain DAN activity for prolonged reward presentations (Ichinose, 2021).

Despite a wide variety of behaviors modulated by learning, much attention has been paid to conditioned odor choice in Drosophila olfactory learning. This study focused on memory-guided positioning bias from the upwind odor source, because positioning to the odor source should be ecologically important for efficient maneuver with potential food and danger. Indeed, wild-type flies showed remarkable changes in the distributions with respect to the paired and unpaired odor sources (CS+ and CS-, respectively): in the test of reward memory, flies approach more toward the CS+ source than CS- and vice versa for punishment memory. By introducing different numbers of flies to the T-maze, it was confirmed that a memory-guided distribution bias is not explained by the fly density (Ichinose, 2021).

To evaluate such conditioned odor source attraction, an attraction index (AI) was devised based on receiver operating characteristic (ROC) curve analysis for relative fly distributions in the two arms of the T-maze. AI becomes positive or negative if flies are relatively attracted to or repelled from the CS+ odor source, respectively. As for PIs, AIs in the tests of appetitive and aversive memories were significantly different from zero in the opposite directions. As well as conditioned odor choice, odor source attraction increased with the training durations. Blockade of all three KC subtypes in the MB010B/UAS-shits1 flies abolished conditioned odor source attraction for both reward and punishment memories. These results together indicate that conditioned odor choice and odor source attraction correlate well with each other and are both dependent on the MB (Ichinose, 2021).

Next it was asked whether MBONs are also required for conditioned odor source attraction. To this end, the AI of the Shits1-expressing flies was calculated. Consistent with the observation of the wild-type flies, most of the genotypes showed bi-directional conditioned odor source attraction at the permissive temperature. The transient blockade of certain MBON types attenuated the distribution bias. Because both conditioned odor choice and odor source attraction require the MB output, it was asked whether these two behaviors are controlled by the common circuits. To this end, the contribution scores of AI were calculated and compared to those of PI. Interestingly, no significant correlation was found between these two behavioral variables. For example, the contribution of MBON-α1 during reward memory acquisition was selective to PI, although MBON-β2β'2a was selectively required for AI. These results indicate that conditioned odor choice and attraction are mediated by, at least in part, independent sets of MBONs (Ichinose, 2021).

To understand how the MB differentiates its output in memory steps for odor source attraction, the patterns of MBON contribution were looked into in each phase of reward and punishment memories. In contrast to the case of PI, memory retrieval required a more distinct set of MBONs than acquisition and retention. Indeed, what stood out in the matrix was that requirement of most MBONs is highest during the retrieval of reward memory. The impaired AI in the reward memory retrieval was further examined by comparing the median density profiles of the experimental and control genotypes. This analysis revealed that the MBON blockade altered both, but often differentially, the affinity to the paired odor and the retreat from the unpaired odor source. Altogether, these results highlight the particular importance of the MB output to localize food-predicting odor source and the distinct neuronal regulation for memory-guided odor choice and source attraction (Ichinose, 2021).

Quantitative analysis of MBON contribution resolved distinct patterns during different memory functions. This approach revealed that valence, conveyed by DANs, was a key driving factor to dissociate the circuit usage, corroborating the widely accepted model of compartment-based functional dissociation in the MB. One should, however, be cautious when interpreting each data point, especially negative data, because of the fluctuating nature of the behavioral assays and relatively 'weak' UAS-shits1 strain used in this study. Comprehensive examination of the MB output circuit under the same experimental condition allowed direct and quantitative comparisons of requirements and successfully uncovered the pattern how the MB multiplexes distinct memory functions (Ichinose, 2021).

Interestingly, MBONs that are most required for reward memory acquisition are all glutamatergic and form recurrent circuits by projecting their terminals to dendrites of the corresponding DANs. These results are consistent with a previous proposal of recurrent dopamine reward signals and further expand it as a common circuit motif, which may help to sustain the reward signals. Simplest interpretation of these feedback loops would be mutual potentiation of MBON and DAN activity. Because glutamate and dopamine can also suppress neuronal activity, mutual inhibition between MBONs and DANs may enhance the DAN activity. The glutamatergic recurrent circuits in the MB thus facilitate appetitive memory formation in response to substantial reward input (Ichinose, 2021).

The MB output biases was found not only for odor choice but also their positions toward upwind odor sources. Comparisons of MBON contributions for choice and attraction revealed that the MBON circuit differentially controls these two memory-guided odor responses. Remarkably, it was found that the vast majority of MBONs were required during retrieval of reward memory. Consistently, activation of PAM cluster neurons paired with an odor presentation was reported to trigger plastic changes throughout the MB lobes. Engagement of nearly the whole MB circuit in appetitive memory retrieval might reflect the intricacy of the task: flies need to integrate reward memory, wind direction, and feeding state to execute approach to the food-related odor source. Indeed, the Drosophila MB was shown to play roles in innate odor attraction, airflow, and foraging. Therefore, the rewarded odor may acquire the access to exploit these functions of the MBON circuit. An open question is how different types of information are processed in the MB network and how it is modulated by the past experiences (Ichinose, 2021).

Go to top

Ingestion of artificial sweeteners leads to caloric frustration memory in Drosophila

Musso, P. Y., Lampin-Saint-Amaux, A., Tchenio, P. and Preat, T. (2017). Nat Commun 8(1): 1803. PubMed ID: 29180783

Non-caloric artificial sweeteners (NAS) are widely used in modern human food, raising the question about their health impact. This study asked whether NAS consumption is a neutral experience at neural and behavioral level, or if NAS can be interpreted and remembered as negative experience. Behavioral and imaging approaches were used to demonstrate that Drosophila melanogaster learn the non-caloric property of NAS through post-ingestion process. These results show that sweet taste is predictive of an energy value, and its absence leads to the formation of what we call Caloric Frustration Memory (CFM) that devalues the NAS or its caloric enantiomer. CFM formation involves activity of the associative memory brain structure, the mushroom bodies (MBs). In vivo calcium imaging of MB-input dopaminergic neurons that respond to sugar showed a reduced response to NAS after CFM formation. Altogether, these findings demonstrate that NAS are a negative experience for the brain (Musso, 2017).

Go to top

Memory-relevant mushroom body output synapses are cholinergic

Barnstedt, O., Owald, D., Felsenberg, J., Brain, R., Moszynski, J. P., Talbot, C. B., Perrat, P. N. and Waddell, S. (2016). Neuron 89: 1237-1247. PubMed ID: 26948892

Memories are stored in the fan-out fan-in neural architectures of the mammalian cerebellum and hippocampus and the insect mushroom bodies. However, whereas key plasticity occurs at glutamatergic synapses in mammals, the neurochemistry of the memory-storing mushroom body Kenyon cell output synapses is unknown. This study demonstrates a role for acetylcholine (ACh) in Drosophila. Kenyon cells express the ACh-processing proteins ChAT and VAChT, and reducing their expression impairs learned olfactory-driven behavior. Local ACh application, or direct Kenyon cell activation, evokes activity in mushroom body output neurons (MBONs). MBON activation depends on VAChT expression in Kenyon cells and is blocked by ACh receptor antagonism. Furthermore, reducing nicotinic ACh receptor subunit expression in MBONs compromises odor-evoked activation and redirects odor-driven behavior. Lastly, peptidergic corelease enhances ACh-evoked responses in MBONs, suggesting an interaction between the fast- and slow-acting transmitters. Therefore, olfactory memories in Drosophila are likely stored as plasticity of cholinergic synapses (Barnstedt, 2016).

Despite decades of work on learning and memory and other functions of the MB, the identity of the fast-acting neurotransmitter that is released from the KCs has remained elusive. Much of the insect brain was considered to be cholinergic, but the MB was thought to be unique. Histological studies concluded that the MB did not express ChAT but that subsets of KCs contained glutamate, aspartate, or taurine. However, conclusive evidence that these molecules are released as neurotransmitters has not materialized (Barnstedt, 2016).

This study presents multiple lines of evidence that ACh is a KC transmitter. (1) KCs express the ChAT and VAChT proteins that synthesize and package ACh into synaptic vesicles, and the expression of these genes is required for MB-dependent learned behavior. (2) Stimulation of KCs triggers responses in MBONs that are similar to those evoked by direct ACh application. (3) Reducing ACh processing in KCs impairs KC-evoked responses in MBONs. (4) ACh- and KC-evoked responses in MBONs are both sensitive to antagonism of nicotinic ACh receptors. (5) Odor-evoked responses in MBONs are attenuated by reducing the expression of several nicotinic ACh receptor subunits. Taken together, these data provide compelling support that ACh is a major neurotransmitter released from Drosophila KCs (Barnstedt, 2016).

The anatomy of ACh-responsive MBONs suggests that many αβ, α'β', and γ lobe KCs are likely to be cholinergic. Calcium imaging may miss subtle or inhibitory effects, so it remains possible that subclasses of KC might also release or corelease other small molecule transmitters. It is, for example, notable that the MB neurons express an atypical putative vesicular transporter. Furthermore, taurine histology specifically labels the αβ core neurons. Anatomy suggests that αβ core and αβ surface outputs are pooled by MBONs with dendrites in the α lobe tip and throughout the β lobe, but that the dendrites of MBONs in the α lobe stalk preferentially innervate αβ surface neurons. It will be important to understand how ACh signals from different KCs are integrated by MBONs. The αβ and γ, but not α'β', KCs can corelease ACh with the sNPF neuropeptide. The current data raise the possibility that coreleased sNPF may facilitate ACh-evoked responses. sNPF drives autocrine presynaptic facilitation of certain olfactory sensory neurons in the adult fly. Conversely, sNPF decreased the resting membrane potential of larval motor neurons that ectopically express sNPFR. MBONs with dendrites in certain lobes therefore receive different combinations of transmitters and may vary in responding to sNPF (Barnstedt, 2016).

Finding that ACh is the KC transmitter has important implications for learning-relevant plasticity at KC-MBON synapses. Current models suggest that valence-specific and anatomically restricted reinforcing dopaminergic neurons drive presynaptically expressed plasticity between KCs and particular MBONs. Reward learning skews KC-MBON outputs toward driving approach by depressing the odor drive to MBONs that direct avoidance, whereas aversive learning enhances drive to avoidance by reducing drive to approach MBONs and increasing drive to avoidance pathways. The results here indicate that learning is represented as dopaminergic tuning of excitatory cholinergic KC-MBON synapses (Barnstedt, 2016).

Learning requires dopamine receptor function in the KCs, which implies a presynaptic mechanism of plasticity at the KC-MBON junction. Presynaptic plasticity of odor-activated KCs provides a simple means to retain odor specificity of memory in the highly convergent anatomy of the MB-where 2,000 KCs converge onto single or very few MBONs per zone on the MB lobes. The anatomically analogous mammalian cerebellar circuits, to which the insect MBs have been compared, exhibit presynaptic glutamatergic plasticity that is cAMP dependent. Finding that the KC transmitter is ACh suggests that cAMP-dependent mechanisms can modulate synaptic connections, regardless of transmitter identity. The MB KCs appear to be strikingly similar to the large parallel ensemble of cholinergic amacrine cells in the vertical lobe of the cuttlefish. These Cephalopod amacrine cells also share the same fan-out input and fan-in efferent anatomy of the Drosophila KCs, and plasticity occurs at the cholinergic connection between amacrine cells and downstream large efferent neurons (Barnstedt, 2016).

Work in the locust suggested that spike-timing-dependent plasticity (STDP) marks the relevant conditioned odor-activated KC-MBON synapses so that they are susceptible to reinforcing modulation. STDP relies on coincidence of pre- and postsynaptic activity and influx of postsynaptic Ca2+ through NMDA-type glutamate receptors. Recent work in Drosophila pairing odor presentation with dopaminergic neuron activation reported odor-specific synaptic depression at a KC-MBON junction that did not require postsynaptic MBON depolarization. It will be important to determine whether this holds for all DAN-MBON compartments or whether some learning-induced plasticity involves synaptic Ca2+ influx through an ACh-triggered nAChR, rather than the more traditional glutamate-gated NMDA receptors (Barnstedt, 2016).

This study identified roles for the Dα1, Dα4, Dα5, and Dα6 nAChR subunits in M4/6 MBONs. Reducing the expression of these subunits lowered odor-evoked signals in MBONs and converted naive odor avoidance into approach behavior. Dα5 and Dα6 subunits can form functional heteromeric channels in vitro. Different MBONs may express unique combinations of AChRs and therefore have characteristic physiological responses to KC-released ACh, as well as perhaps different learning rules and magnitudes of plasticity. Pre- or postsynaptically localized muscarinic AChRs could provide additional memory-relevant modulation (Barnstedt, 2016).

Beyond important roles in memory formation, consolidation, and expression, the MB- and DAN-directed modulation of specific MBON pathways has also been implicated in controlling hunger, thirst, temperature, and sleep/wake state-dependent locomotor behaviors. It will therefore be important to understand how plasticity of cholinergic KC transmission serves these discrete functions (Barnstedt, 2016).

Go to top

Suppression of a single pair of mushroom body output neurons in Drosophila triggers aversive associations

>Ueoka, Y., Hiroi, M., Abe, T. and Tabata, T. (2017). FEBS Open Bio 7(4): 562-576. PubMed ID: 28396840

Memory includes the processes of acquisition, consolidation and retrieval. In the study of aversive olfactory memory in Drosophila melanogaster, flies are first exposed to an odor (conditioned stimulus, CS+) that is associated with an electric shock (unconditioned stimulus, US), then to another odor (CS-) without the US, before allowing the flies to choose to avoid one of the two odors. The center for memory formation is the mushroom body which consists of Kenyon cells (KCs), dopaminergic neurons (DANs) and mushroom body output neurons (MBONs). However, the roles of individual neurons are not fully understood. This study focused on the role of a single pair of GABAergic neurons (MBON-gamma1pedc) and found that it could inhibit the effects of DANs, resulting in the suppression of aversive memory acquisition during the CS- odor presentation, but not during the CS+ odor presentation. It is proposed that MBON-gamma1pedc suppresses the DAN-dependent effect that can convey the aversive US during the CS- odor presentation, and thereby prevents an insignificant stimulus from becoming an aversive US (Ueoka, 2017).

Go to top

Propagation of homeostatic sleep signals by segregated synaptic microcircuits of the Drosophila mushroom body

Sitaraman, D., Aso, Y., Jin, X., Chen, N., Felix, M., Rubin, G. M. and Nitabach, M. N. (2015). Curr Biol 25: 2915-2927. PubMed ID: 26455303.

The Drosophila mushroom body (MB) is an associative learning network that is important for the control of sleep. Particular intrinsic MB Kenyon cell (KC) classes have been identified that regulate sleep through synaptic activation of particular MB output neurons (MBONs) whose axons convey sleep control signals out of the MB to downstream target regions. Specifically, it was found that sleep-promoting KCs increase sleep by preferentially activating cholinergic sleep-promoting MBONs, while wake-promoting KCs decrease sleep by preferentially activating glutamatergic wake-promoting MBONs. By using a combination of genetic and physiological approaches to identify wake-promoting dopaminergic neurons (DANs) that innervate the MB, it was shown that they activate wake-promoting MBONs. These studies reveal a dopaminergic sleep control mechanism that likely operates by modulation of KC-MBON microcircuits (Sitaraman, 2015).

This study used a combination of sophisticated cell-specific genetic manipulations with behavioral sleep analysis and optical electrophysiology to identify compartment-specific wake-promoting MB DANs that activate wake-promoting microcircuits. Previous studies have implicated DANs innervating the central complex (CX) - a brain region involved in locomotor control - in regulating sleep, and other non-dopamingeric CX neurons have been implicated in homeostatic control of sleep. In addition, it has recently been shown that manipulations of dopamine signaling in the MB alter sleep, although the precise DANs involved remains unclear. This study has now identified specific wake-promoting MB DANs and shown that they innervate lobe compartments also innervated by wake-promoting KCs and MBONs. Importantly, this study has also shown that dopamine secretion by DANs innervating a particular MB lobe compartment acts through D1 subtype receptors to activate the wake-promoting microcircuit specific to that compartment to a much greater extent than it activates the sleep-promoting microcircuit residing in different compartments. This provides direct physiological evidence for compartment-specific dopamine signaling in the regulation of sleep by the MB, and is consistent with a previous study in the context of learning and memory. Future studies are required to determine additional cellular and molecular details of how dopamine signals modulate sleep-regulating microcircuits (Sitaraman, 2015).

On the basis of recently published studies of MB control of sleep and the results presented in this study, a unified mechanistic model is proposed for homeostatic control of sleep by excitatory microcircuits in the Drosophila MB. Wake-promoting MBON-γ5β'2a/β'2mp/β'2mp_bilateral and sleep-promoting γ2α'1 each receive anatomical inputs from both wake-promoting γm and α'/β' KCs KCs and sleep-promoting γd KCs. However, segregation of sleep control information into separate microcircuits is enforced by greater synaptic weights between γ and γm and α'/β' KCs and MBON-γ5β'2a/β'2mp/β'2mp_bilateral, and between γm and α/β' KCs and MBON-γ5β'2a/β'2mp/β'2mp_bilateral, and between γd KCs and MBON-γ2α'1 (Sitaraman, 2015). Thus it is hypothesize that compartment-specific dopamine signals from MB DANs could potentially determine these differences in synaptic weight. Future studies will test this hypothesis (Sitaraman, 2015).

Interestingly, other fly behaviors have recently been found to be regulated by sleep-controlling compartment-specific MB microcircuits. For example, the integration of food odor to suppress innate avoidance of CO2 is mediated by MBON-γ5β'2a/β'2mp/β'2mp_bilateral and PAM DANs that innervate the β'2 compartment. Optogenetic activation experiments reveal that wake-promoting γ5β'2a/β'2mp/β'2mp_bilateral mediates innate avoidance, while MBON-γ2α'1 mediates attraction. However, thermogenetic inactivation studies reveal that both MBON-γ5β'2a/β'2mp/β'2mp_bilateral and MBON-γ2α'1 are important for various forms of associative memory formation. These diverse waking behaviors that involve the activity of sleep-regulating neurons raises the interesting question whether such roles are independent, or causally linked, which future studies can address (Sitaraman, 2015).

Importantly, this study has provided for the first time a cellular and molecular mechanism for for dopaminergic control of sleep through modulation of an associative network. While dopaminergic projections to cerebral cortex are known to be important for regulating sleep and arousal in mammals, underlying cellular and molecular mechanisms remain poorly understood, although D2 subtype dopamine receptors have been implicated in the control of REM sleep. Because of the possible evolutionary relationship between the MB and vertebrate forebrain associative networks (such as mammalian cerebral cortex), these studies thus provide a framework for the design of analogous experiments in genetically tractable vertebrate model systems such as zebrafish and mice (Sitaraman, 2015)

Go to top

Control of sleep by dopaminergic inputs to the Drosophila mushroom body

Sitaraman, D., Aso, Y., Jin, X., Chen, N., Felix, M., Rubin, G. M. and Nitabach, M. N. (2015). Curr Biol 25: 2915-2927. PubMed ID: 26455303.

The Drosophila mushroom body (MB) is an associative learning network that is important for the control of sleep. Particular intrinsic MB Kenyon cell (KC) classes have been identified that regulate sleep through synaptic activation of particular MB output neurons (MBONs) whose axons convey sleep control signals out of the MB to downstream target regions. Specifically, it was found that sleep-promoting KCs increase sleep by preferentially activating cholinergic sleep-promoting MBONs, while wake-promoting KCs decrease sleep by preferentially activating glutamatergic wake-promoting MBONs. By using a combination of genetic and physiological approaches to identify wake-promoting dopaminergic neurons (DANs) that innervate the MB, it was shown that they activate wake-promoting MBONs. These studies reveal a dopaminergic sleep control mechanism that likely operates by modulation of KC-MBON microcircuits.

Go to top

Pairing-Dependent Plasticity in a Dissected Fly Brain Is Input-Specific and Requires Synaptic CaMKII Enrichment and Nighttime Sleep

Adel, M., Chen, N., Zhang, Y., Reed, M. L., Quasney, C. and Griffith, L. C. (2022). J Neurosci 42(21): 4297-4310. PubMed ID: 35474278

In Drosophila, in vivo functional imaging studies revealed that associative memory formation is coupled to a cascade of neural plasticity events in distinct compartments of the mushroom body (MB). In-depth investigation of the circuit dynamics, however, will require an ex vivo model that faithfully mirrors these events to allow direct manipulations of circuit elements that are inaccessible in the intact fly. The current ex vivo models have been able to reproduce the fundamental plasticity of aversive short-term memory, a potentiation of the MB intrinsic neuron (Kenyon cells [KCs]) responses after artificial learning ex vivo However, this potentiation showed different localization and encoding properties from those reported in vivo and failed to generate the previously reported suppression plasticity in the MB output neurons (MBONs). This study developed an ex vivo model using the female Drosophila brain that recapitulates behaviorally evoked plasticity in the KCs and MBONs. This plasticity accurately localizes to the MB α'3 compartment and is encoded by a coincidence between KC activation and dopaminergic input. The formed plasticity is input-specific, requiring pairing of the conditioned stimulus and unconditioned stimulus pathways; hence, it was named pairing-dependent plasticity. Pairing-dependent plasticity formation requires an intact CaMKII gene and is blocked by previous-night sleep deprivation but is rescued by rebound sleep. In conclusion, this study showed that the ex vivo preparation recapitulates behavioral and imaging results from intact animals and can provide new insights into mechanisms of memory formation at the level of molecules, circuits, and brain state (Adel, 2022).

A neutral experience (conditioned stimulus [CS]) can be remembered as positive or negative if closely followed by rewarding or punishing reinforcement (unconditioned stimulus [US]). The ability to form this type of 'associative' memory is phylogenetically conserved; Drosophila form robust associative memories, most of which are encoded and stored in the mushroom body (MB). The MB is a higher brain structure made of 15 distinct compartments. Each compartment is built on a scaffold of axons of one of the three main types of Kenyon cells (KCs; αβ, α'β', and γ). The KCs connect to MB output neurons (MBONs), which project out of the MB to bias behavior. The KC->MBON synapses are modulated by dopaminergic neurons (Adel, 2022).

During aversive olfactory associative learning, an odor (the CS) activates a sparse group of KCs, such that this odor identity is represented across all MB compartments. Simultaneously, dopaminergic neurons from the protocerebral posterior lateral (PPL1) cluster are activated by the US, encoding negative prediction errors in MB compartments. When KC activation and the dopaminergic signal coincide within a compartment, the KC->MBON synapses in that compartment are depressed, biasing the circuit output to aversion (Adel, 2022).

Many studies have investigated the properties of this circuitry using in vivo calcium imaging in intact animals. In contrast, explanted brains have been used mostly for establishing connectivity between neurons or interrogating a specific biochemical pathway; only a few studies have attempted to understand memory circuit logic ex vivo. In the best-developed paradigm, a previous study observed a potentiation of KC responses in the tips of the MB vertical lobes which they termed 'long-term enhancement' (LTE). This laid the groundwork for developing ex vivo models of this circuit, but there were major differences between LTE and associative memory observed in intact animals. The most significant were that the plasticity was not specific to the α'β' KCs and that dopamine release by the US was not observed; it was only seen after CS+US coincidence (Ueno et al., 2013, 2017) (Adel, 2022).

This study establish an ex vivo paradigm that resolves these discrepancies and exhibits the cardinal features of associative learning. Pairing odor and punishment pathway activation in dissected brains results in a localized potentiation of the α'β' KCs and suppression of their postsynaptic MBONs in the α'3 compartment. Because both KC potentiation and MBON suppression are strictly dependent on temporal coincidence of the CS and US, this paradigm was termed 'pairing-dependent plasticity' (PDP). Like the CS specificity of associative memories, PDP is specific to the subset of odor-representing projection neurons activated during the artificial training. Evidence is also provided that dopamine is released by activation of the US pathway and does not require CS+US coincidence (Adel, 2022).

This ex vivo paradigm can be used for obtaining new mechanistic insight into memory formation at the molecular and circuit levels. Data is presented indicating that the 3'UTR of the CaMKII gene is critical for short-term memory (STM) formation and that the primacy of α' compartment plasticity in learning is because of differences in input/response relationships between α and α'. Finally, it was demonstrated that the ability of the ex vivo brain to be plastic can be influenced by prior in vivo experience, as this study reports that brains of sleep-deprived flies fail to form PDP, but as little as 2 h of recovery sleep rescues this learning impairment (Adel, 2022).

Drosophila neural circuits are traditionally studied by relating in vivo genetic and chemical manipulations with their consequent behavioral outcomes, from which circuit information can then be inferred. More recently, the advent of in vivo calcium imaging allowed for tracing neural activity in actively behaving flies. Over more than a decade of such in vivo studies, the general circuit mechanisms of associative memory have been discovered, but there are limitations imposed by imaging the brain of an active intact fly. These include the relatively low signal-to-noise ratio, the inaccessibility of multiple brain regions because of restrictions on imaging angles, the difficulty of doing acute pharmacological studies, and the possible confounds of studying the brain of a movement-restricted fly experiencing ongoing stress. Taking inspiration from the way the LTP hippocampal slice model revolutionized understanding of mammalian memory, this study provides an ex vivo model of Drosophila memory which can overcome these limitations and offer a powerful preparation for studying Drosophila memory circuits. Importantly, this model provides a framework for investigating the dynamics of neural circuits in the fly brain (Adel, 2022).

Most of the previous studies investigating the associative learning circuit ex vivo have focused on mapping connectivity or characterizing a specific biochemical pathway. Only a few ex vivo studies have focused on understanding MB circuit logic. In the LTE model, pairing a stimulation of the CS and US pathways induced a potentiation of KC responses in the tips of the MB vertical lobes, but LTE did not fully recapitulate other characteristics of associative memory observed in intact flies. This study developed a modified ex vivo model that resolves these discrepancies, showing that the paired activation of odor and punishment pathways induces appropriate plasticity at multiple nodes in the circuit: potentiation of KCs and suppression of MBONs. Several mechanisms for encoding those opposite forms of the plasticity have been proposed, including spike timing-dependent plasticity and activation of distinct dopaminergic receptors. Spike timing-dependent plasticity mechanisms appear less likely as MBON suppression was shown to not require MBON spiking. Perhaps the strongest model so far comes a previous demonstration that differences in the order of KCs activation and dopaminergic input activate distinct dopaminergic receptors, DopR1 or DopR2, which encode MBON suppression or potentiation, respectively. It is important to note that previous work studied the plasticity in MB medial lobes, while the current study focused on MB vertical lobes, so this paradigm may be useful in gaining greater mechanistic insight into this sign transformation in the vertical lobes (Adel, 2022).

PDP is localized to the MB α'3 compartment and not in α3, in alignment with most imaging studies in intact flies. Importantly, in this ex vivo preparation, punishment information is relayed to the MB through dopaminergic release from the PPL1 subset. Bath application of dopamine in the current preparation does not interfere with the specificity of associative learning since PDP is exclusively formed in the cells that were active during the dopamine application. These data settle several inconsistencies between previous ex vivo studies and the majority of in vivo reports. It is suggest that the genesis of the discrepancies was not because of any inherent difference between intact and ex vivo brains but was rather a consequence of technical considerations, including stimulation strength, dopamine concentration, and the sensor tools used (Adel, 2022).

An ex vivo preparation that recapitulates the cardinal features of the circuits underlying associative memory formation should be useful for mechanistic studies at the molecular, cellular, and systems levels. This model was used to ask a new question about the innerworkings of the circuit at each of these levels. At the molecular level, the importance of normal levels of CaMKII was demonstrated by manipulating the 3'UTR of CaMKII mRNA. Deletion of this region of the CaMKII gene drastically reduces the amount of CaMKII protein in synaptic regions and blunts the ability to form STM and to generate a potentiation PDP in KC axons. The data argue that the role of this molecule is downstream of the CS+US coincidence detector, as a much weaker PDP was observed in CaMKIIUdel flies. Whether the behavioral defect is due solely to the KC PDP defect is not completely clear since CaMKII likely has active roles at other circuit nodes (Adel, 2022).

At the cellular level, it was asked why STM and PDP form in the α'3 but not the nearby α3 compartment when both compartments respond to odors and AL stimulation, and both receive dopaminergic input from the same PPL1 cluster. Previous work found that real odors cause activity in only 5%-12% of KCs and elicit a much higher spike rate in the α'β' KCs than in the αβ KCs. This study found that low-intensity AL stimulation (100 μAmps) elicits a stronger response in the α'3 than in the α3 compartment, while high-intensity AL-stimulation (200 μAmps) causes strong responses in the α3 compartment and recruits it to the learning circuit. Coupling this with the observation of lower dopamine release in α3 suggests a model in which odor presentation during associative learning causes subthreshold responses in αβ cells such that the CS+US coincidence detector is not triggered, while the stronger responses in the α'β' cells bypass this threshold, allowing plasticity in the α'β' cells only. This notion is in agreement with the previous finding that α'β' cells have a lower firing threshold than αβ cells. Further, It is possible that long-term memory and the enhancement memory trace in the αβ KCs after repetitive space training require a gradual potentiation of the αβ KC responses with every training session such that the responses bypass the coincidence detection threshold after several training sessions. Whether repetition of AL+DA pairings recruits PDP in the α3 compartment remains unclear. It is also yet to be determined whether shortcutting the circuit and recruiting αβ cells in the first training session reduces the need for multiple spaced training sessions in long-term memory formation (Adel, 2022).

In conclusion, this study looked at the ability of the effects of prior experience, or brain state, on the memory circuit to be retained in the ex vivo preparation. Excitingly, it was found that sleep-deprived flies could not form PDP, but that as little as 2 h of rest before dissection allowed the brain to recover PDP formation. The complete abolition of PDP in sleep-deprived flies at first and the gradual recovery in plasticity afterward suggest that sleep converges on the memory circuit upstream of the CS+US coincidence detector. Whether this involves regulation of dopamine receptors in the MB during sleep remains to be determined. The ability to retain in some functional way the internal state of the brain will allow this preparation to be used to understand how memory formation is altered by global system alterations (Adel, 2022).

Go to top

A neural circuit linking learning and sleep in Drosophila long-term memory

Lei, Z., Henderson, K. and Keleman, K. (2022). Nat Commun 13(1): 609. PubMed ID: 35105888

Animals retain some but not all experiences in long-term memory (LTM). Sleep supports LTM retention across animal species. It is well established that learning experiences enhance post-learning sleep. However, the underlying mechanisms of how learning mediates sleep for memory retention are not clear. Drosophila males display increased amounts of sleep after courtship learning. Courtship learning depends on Mushroom Body (MB) neurons, and post-learning sleep is mediated by the sleep-promoting ventral Fan-Shaped Body neurons (vFBs). This study shows that post-learning sleep is regulated by two opposing output neurons (MBONs) from the MB, which encode a measure of learning. Excitatory MBONs-γ2α'1 becomes increasingly active upon increasing time of learning, whereas inhibitory MBONs-β'2mp is activated only by a short learning experience. These MB outputs are integrated by SFS neurons, which excite vFBs to promote sleep after prolonged but not short training. This circuit may ensure that only longer or more intense learning experiences induce sleep and are thereby consolidated into LTM (Lei, 2022).

This study has identified a neural circuit that regulates learning-induced sleep for LTM consolidation. This circuit links neurons essential for learning and memory in Drosophila, the MB neurons, with those critical for post-learning sleep, the vFBs8. It is proposed that only a longer learning experience is sufficient to induce sleep, and thereby be consolidated into LTM. Given that the increasing duration of a learning experience correlates with the total amount of time males spend on futile courtship towards mated females during training, selective activation of vFBs likely depends on the amount or intensity of a learning experience, rather than just its duration. Post-learning sleep induction requires integration of two MB outputs, previously implicated in courtship memory in SFSs. Post-learning activity of the excitatory MBONs-γ2α'1 increases linearly with the duration of the prolonged learning experience. In contrast, activity of the inhibitory MBONs-β'2mp peaks after a short experience sufficient to induce STM. As a result, only when the males court mated females sufficiently long or intensely, the activity of MBONs-γ2α'1 reaches the threshold required to activate SFSs. This in turn leads to activation of vFBs to promote post-learning sleep and the reactivation of those dopaminergic neurons (DANs) that were involved in memory encoding. Consequently, biochemical processes essential for LTM consolidation become engaged (Lei, 2022).

How might MBONs-γ2α'1 and MBONs-β'2mp measure the learning experience to control post-learning sleep? In homeostatic sleep regulation, the potentiation of R2 neurons reflects a measure of sleep loss that is sensed by dFBs, likely in response to the accumulation of byproducts of oxidative stress during sleep loss. In the case of learning-induced sleep, it is envisioned that learning results in lasting changes in the molecular pathways essential for memory formation in the MB. For example, the cAMP pathway along with the dopamine receptor are activated during sleep in a discrete 3-h time window after learning in rodents and Drosophila males lacking a dopamine receptor, and hence unable to learn, do not display increased post-learning sleep. Thus, the accumulation of changes in the cAMP signaling pathway upon increasing learning experience with mated females might lead to the increasing potentiation of MBONs-γ2α'1 and MBONs-β'2mp after learning. Interestingly, MBONs-γ2α'1 and MBONs-β'2mp display distinct temporal activity patterns upon learning which likely reflects their distinct neuronal properties (Lei, 2022).

This study reveals a circuit mechanism that ensures that only persistent, and thus likely significant, learning experiences generate post-learning sleep to consolidate LTM. Recent findings suggest that dFBs, involved in sleep homeostasis, might mediate a paradoxical type of sleep, in humans also called Rapid Eye Movement (REM) sleep. This in conjunction with the current data, provide an opportunity to investigate whether the post-learning sleep, mediated by vFBs, might represent another type of sleep implicated in mammals in memory consolidation (Lei, 2022).

Go to top

rdgB knockdown in neurons reduced nocturnal sleep in Drosophila melanogaster

Kobayashi, R., Yamashita, Y., Suzuki, H., Hatori, S., Tomita, J. and Kume, K. (2023). Biochem Biophys Res Commun 643: 24-29. PubMed ID: 36586155

Recent studies revealed behaviorally defined sleep is conserved across broad species from insect to human. For evolutional analysis, it is critical to determine how homologous genes regulate the homologous function among species. Drosophila melanogaster shares numerous sleep related genes with mammals including Sik3, salt-inducible kinase 3, whose mutation caused long sleep both in mouse and fruit fly. The Drosophila rdgB (retinal degeneration B) encodes a membrane-associated phosphatidylinositol transfer protein and its mutation caused light-induced degeneration of photoreceptor cells. rdgB mutation also impaired phototransduction and olfactory behavior, indicating rdgB is involved in the normal neural transmission. Mammalian rdgB homologue, Pitpnm2 (phosphatidylinositol transfer protein membrane-associated 2) was discovered as one of SNIPPs (sleep-need index phosphoproteins), suggesting its role in sleep. This study shows that rdgB is involved in sleep regulation in Drosophila. Pan-neuronal and mushroom body (MB) specific rdgB knockdown decreased nocturnal sleep. MB neurons play a dominant role, since the rescue of rdgB expression only in MB neurons in pan-neuronal knockdown reversed the sleep reducing effect of rdgB knockdown. These results revealed the sleep-related function of rdgB in Drosophila which may be conserved across species (Kobayashi, 2023).

Go to top

Heterosynaptic plasticity underlies aversive olfactory learning in Drosophila

Hige, T., Aso, Y., Modi, M. N., Rubin, G. M. and Turner, G. C. (2015). Neuron 88(5): 985-998. PubMed ID: 26637800

Although associative learning has been localized to specific brain areas in many animals, identifying the underlying synaptic processes in vivo has been difficult. This study provides the first demonstration of long-term synaptic plasticity at the output site of the Drosophila mushroom body. Pairing an odor with activation of specific dopamine neurons induces both learning and odor-specific synaptic depression. The plasticity induction strictly depends on the temporal order of the two stimuli, replicating the logical requirement for associative learning. Furthermore, dopamine action was shown to be confined to and distinct across different anatomical compartments of the mushroom body lobes. Finally, overlap between sparse representations of different odors defines both stimulus specificity of the plasticity and generalizability of associative memories across odors. Thus, the plasticity found in this study not only manifests important features of associative learning but also provides general insights into how a sparse sensory code is read out (Hige, 2015).

By developing new tools for precisely manipulating neural circuitry in Drosophila, this study provides the first characterization of synaptic plasticity linked to associative learning at the output of the MB. Focusing on the γ1pedc compartment, which is critically involved in memory acquisition, Long-term depression was observed of the synaptic inputs to these MBONs. The fact that a roughly 90% reduction is seen in synaptic currents supports the interpretation that the effect is at KC-MBON synapses, which are expected to be the most prominent input to the MBONs. However, there is still the possibility that other circuit elements presynaptic to the MBONs, including potentially the DPM neuron that widely innervates the MB lobes, also contribute to the changes seen in this study. The plasticity was robust, long-lasting, and depended on the temporal order of CS and US with sub-second precision. Thus, the minimum logical requirements for associative learning are implemented by dopamine-induced heterosynaptic plasticity (Hige, 2015).

MBONs have been proposed to convey the valence of olfactory stimuli, since direct activation of different MBONs can elicit approach or avoidance behavior, depending on the cell type. In particular, MBON-γ1pedc is thought to signal positive valence, since optogenetically activating this neuron evokes attraction to the light. If the plasticity is related to the behavior, then when an animal learns to avoid an odor, the response of MBON-γ1pedc to that particular odor should go down. Indeed, the plasticity we found in this neuron during aversive learning was LTD, and it was odor specific. Thus, these observations provide a simple, unifying explanation for how behavior is modified during learning (Hige, 2015).

In general, the direction of the behavioral response triggered by MBON activation (i.e., approach versus avoidance) is opposite in sign to the valence signaled by the corresponding DANs (i.e., reward versus punishment). This opponent relationship suggests that dopamine-induced plasticity is depression, so that DANs act by turning down activity of MBONs that signal the opposite valence. This study indeed found that heterosynaptic LTD takes place not only in the γ1pedc compartment, but also in α2. However, the results also indicate that the rules for inducing plasticity are not the same across all compartments, as it required much longer pairing of odor and DAN activation to induce plasticity in the α2 compartment than in γ1pedc. What is the functional significance of this differential sensitivity to dopamine? The extremely high sensitivity in the γ1pedc compartment matches well with the observations that activation of PPL1-γ1pedc can substitute aversive US at much higher efficiency compared to other PPL1-DANs. Although the difference in these behavioral scores could potentially be explained by other factors, such as differential strength of the link from the MBON to behavioral output, it is consistent with the fact that the γ1pedc compartment or, more generally, γ KCs play a critical role in acquisition of aversive memory. On the other hand, the behavioral evidence suggests the α2 compartment may be more heavily involved in retrieval of memories (Hige, 2015).

The results show that the anatomically defined MBON-DAN modules in the MB lobes reflect the modularity of circuit function, compartmentalizing the synaptic changes that accompany learning into these discrete zones. This matches well with the picture developed from recent behavioral studies indicating the different compartments are related to distinct motivational drives. These studies showed that the dopaminergic neurons required for learning with different types of reinforcement project to different compartments. For example, the DANs necessary and sufficient for appetitive conditioning project to different compartments than those for learning driven by thirst. Even reinforcement by sweet taste versus caloric intake is localized to distinct compartments. These studies lead to the hypothesis that there is a mechanism whereby information represented by the MB is independently read out by multiple types of MBONs. The results in this study provide the first evidence that this is achieved by compartmentalizing the synaptic changes driven by different reinforcement pathways into these discrete anatomical zones. Dopamine released in one compartment induces robust plasticity in that compartment, but not in its neighbor. This modularity is all the more noteworthy since the dopamine receptors involved in memory acquisition reside in the KC axons, and each KC axon makes en passant synapses with dendrites of multiple MBONs in different compartments. Nevertheless, this study found that the action of dopamine is spatially well confined -- even though MBON-γ1pedc and MBON-γ2α'1 likely share most of the same KC inputs, inducing plasticity in the γ1pedc compartment does not alter the responses in γ2α'1. This indicates that neither dopamine nor its downstream intracellular signaling molecules can spread into the synaptic boutons of the same KC axon in the very next compartment. The circuit organization of the MB represents an ideal format for the parallel read out of information, since large numbers of KCs make converging connections with multiple MBONs at different points down the length of their axons. Thus, each MBON likely has access to much of the olfactory information present in the KC population. The compartmentalization of plasticity would allow the circuit to form a series of different, highly odor-specific associations in each of the MBON-DAN modules, making it possible for the flies to make the complex context-dependent choices they need to cope with a changing environment (Hige, 2015).

The compartmental specificity seen in this study is broadly consistent with previous observations that thermogenetically activating DANs leads to an elevation in cAMP levels in KC axons that is localized to the specific compartments innervated by those DANs. Moreover, abundant genetic evidence suggests that cAMP signaling is central to induction of plasticity in the MB. However, one study no changes were observed in odor-evoked calcium responses in the KC axons in the γ1 region after pairing odor with DAN activation using TH-GAL4, even though this study observed very robust LTD in MBON-γ1pedc using the same GAL4 line. In fact, there was not a close correlation between the spatial patterns of cAMP elevation and of altered calcium responses; some compartments that did not show a cAMP elevation exhibited a change in odor-evoked calcium responses, while other compartments that did show cAMP increases did not show a change in calcium responses. This lack of correlation might arise simply because the molecular machinery used for plasticity lies downstream of calcium influx, or it may be that the change in calcium concentration is so small and local that it is difficult to detect with a cytosolic calcium reporter. It is also possible that plasticity is predominantly expressed postsynaptically. Several pioneering studies have demonstrated learning-related changes in odor responses of KC axons using calcium imaging. Given the possible discrepancy between synaptic plasticity and changes in axonal calcium signals, this issue needs further investigation (Hige, 2015).

Much like other higher-order sensory areas, stimulus representations in KCs involve sparse activation of small numbers of cells, each of which has highly specific response properties. Having established methods to induce plasticity in a neuron that receives heavily converging input from a well-characterized sparse coding area, this study was presented a unique opportunity to directly test a long-held hypothesis about sparse coding. That is, by reducing the overlap between ensembles of cells responding to different stimuli, sparse representations minimize the problem of synaptic interference. By testing multiple odor pairs that evoke similar or dissimilar responses in the KC population, this study indeed found a clear relationship between the degree of overlap in KC representations and the odor specificity of the plasticity. These results suggest a simple model for learning, where those KC-MBON synapses that are active upon the arrival of dopamine (or shortly prior to its arrival) undergo plasticity. The observation that plasticity does not rely on MBON spiking also supports the idea that it is the coincident activity of KCs and DANs that is the sole determinant for plasticity. When these synapses overlap with those activated by similar odors, the stimulus specificity is concordantly reduced (Hige, 2015).

These parallel behavioral experiments showed that synaptic interference carries an important biological meaning. For pairs of odors where synaptic interference was observed in physiological experiments, an association formed with one of those stimuli generalizes to the other odor. In other words, generalization arises because learning one association modifies representations of stimuli with overlapping response patterns. Thus, the results reveal important aspects of learning in a system with distributed population-level representations of sensory inputs, likely to be widely applicable to other memory-related brain areas (Hige, 2015).

Go to top

Associative learning drives longitudinally graded presynaptic plasticity of neurotransmitter release along axonal compartments

Stahl, A., Noyes, N. C., Boto, T., Botero, V., Broyles, C. N., Jing, M., Zeng, J., King, L. B., Li, Y., Davis, R. L. and Tomchik, S. M. (2022). Elife 11. PubMed ID: 35285796

Anatomical and physiological compartmentalization of neurons is a mechanism to increase the computational capacity of a circuit, and a major question is what role axonal compartmentalization plays. Axonal compartmentalization may enable localized, presynaptic plasticity to alter neuronal output in a flexible, experience-dependent manner. This study shows that olfactory learning generates compartmentalized, bidirectional plasticity of acetylcholine release that varies across the longitudinal compartments of Drosophila mushroom body (MB) axons. The directionality of the learning-induced plasticity depends on the valence of the learning event (aversive vs. appetitive), varies linearly across proximal to distal compartments following appetitive conditioning, and correlates with learning-induced changes in downstream mushroom body output neurons (MBONs) that modulate behavioral action selection. Potentiation of acetylcholine release was dependent on the Ca(V)2.1 calcium channel subunit Cacophony. In addition, contrast between the positive conditioned stimulus and other odors required the inositol triphosphate receptor, which maintained responsivity to odors upon repeated presentations, preventing adaptation. Downstream from the MB, a set of MBONs that receive their input from the γ3 MB compartment were required for normal appetitive learning, suggesting that they represent a key node through which reward learning influences decision-making. These data demonstrate that learning drives valence-correlated, compartmentalized, bidirectional potentiation, and depression of synaptic neurotransmitter release, which rely on distinct mechanisms and are distributed across axonal compartments in a learning circuit (Stahl, 2022).

Compartmentalized plasticity in neurotransmitter release expands the potential computational capacity of learning circuits. It allows a set of odor-coding MB neurons to bifurcate their output to different downstream approach- and avoidance-driving downstream output neurons, independently modulating the synaptic connections to alter action selection based on the conditioned value of olfactory stimuli. The KCs modify the encoded value of olfactory stimuli through bidirectional plasticity in odor responses, which vary in a compartment-specific manner along the length of the axons. These changes were observed following pairing an olfactory CS with gustatory/somatosensory US (sucrose feeding or electric shock) in vivo. The CS+ and CS- drive unique patterns of plasticity in each compartment, demonstrating that olfactory stimuli are reweighted differently across compartments following learning, depending on the temporal associations of the stimuli. Different molecular mechanisms govern the potentiation of trained odor responses (CaV2/Cac) and maintenance of responsivity over time (IP3R). Finally, one set of γ output neurons, γ3/γ3β'1, is important for appetitive short-term memory (Stahl, 2022).

Learning-induced plasticity of ACh release in the MB was bidirectional within the compartment, depending on the valence of the US, and was coherent with the valence of the MBON downstream of the compartment. Notably, the γ2 and γ3 MB compartments, which relay information to approach-promoting MBONs, exhibited plasticity that was coherent with promoting behavioral approach following appetitive conditioning and avoidance after aversive conditioning. There was an increase in the relative CS+:CS- ACh responses after appetitive conditioning, and conversely reduced CS+:CS- ACh responses following aversive conditioning. Fhis study focused on the time point 5 min following conditioning, which is consistent with behavioral short-term memory. Aversive conditioning was previously reported to decrease neurotransmitter release from KCs. Indirect evidence, via Ca2+ imaging in presynaptic KCs, suggested that increases in presynaptic neurotransmission could also be associated with learning. Pairing odor with stimulation of appetitive PAM dopaminergic neurons potentiates odor-evoked cytosolic Ca2+ transients across the KC compartments. Appetitive conditioning increases odor-evoked Ca2+ transients across KC compartments. Stimulation of dopaminergic circuits associated with reward learning potentiate MB γ4 connections with the respective γ4 MBON. Statistically significant effect was not observed in γ4 with appetitive or aversive classical conditioning, though the CS+ and CS- trended in the same direction as the adjacent γ5 compartment following conditioning. Overall, the present data demonstrate that there are bidirectional changes in neurotransmitter (ACh) release from MB compartments following appetitive vs aversive learning and provide a window into the spatial patterns of plasticity across compartments following associative learning (Stahl, 2022).

Behavioral alterations following conditioning involve changes in responses among the MBONs. As the KCs provide presynaptic olfactory input to the MBONs, it was a logical a priori assumption that presynaptic plasticity in the KCs could be altered in a compartmental manner and contribute to the changes in MBON responses after conditioning. Yet data from previous Ca2+ imaging experiments have not completely supported this model. Compartmentalized effects have been observed in KCs with non-associative learning protocols and within the γ4 compartment following associative learning. In contrast, classical conditioning produces no compartmentalized differences in odor-evoked Ca2+ responses. Appetitive conditioning with odor + sucrose pairing increases odor-evoked cytosolic Ca2+ transients in KCs across the γ lobe compartments. Aversive conditioning produces no net change across the compartments, but alters synapse-specific Ca2+ responses at the individual bouton level. If the compartmental effects of conditioning (observed with Ca2+ imaging) in KCs drove a proportional change in neurotransmitter release, both the approach- and avoidance-promoting MBONs would be simultaneously potentiated. Extracellular influx of Ca2+ through voltage-gated calcium channels is a primary driver of neurotransmitter release; however, there are multiple sources of Ca2+ in the cytosol that could contribute to the GCaMP signals. A major conclusion of the present study is that learning drives compartmentalized plasticity in neurotransmitter release that is coherent with the behavioral valence of the corresponding MBON (Stahl, 2022).

At least two major molecular mechanisms govern the spatial patterns of plasticity across the MB compartments: a Cac-dependent CS+ potentiation and an IP3R-dependent maintenance of sensory responses over trials/time. This suggests that different sources of Ca2+ play different roles in regulating KC synaptic responses. Cac is the pore-forming subunit of the voltage-sensitive, presynaptic CaV2 Ca2+ channel in Drosophila. CaV2 channels regulate several forms of synaptic plasticity, including paired-pulse facilitation, homeostatic plasticity, and long-term potentiation. The current data suggest that these channels regulate the spatial patterns of learning-induced plasticity in the MB unidirectionally (from baseline), with Cac underlying potentiation but not depression. CaV2 channel activity is modulated by presynaptic calcium and G protein-coupled receptor activity, and channel localization in the active zone dynamically regulates synaptic strength. Thus, Cac insertion into, or increased clustering within, the active zones may underlie learning-induced potentiation (e.g. in the γ1-γ2 compartments following appetitive conditioning). Conditional knockdown of Cac, which reduced Cac levels by ~29%, impaired this potentiation, likely by decreasing the number of available channels for modulation. Baseline stimulus-evoked neurotransmitter release was maintained during Cac knockdown, mediated either by the significant residual Cac expression or compensation by other intracellular Ca2+ channels/sources. In contrast to the Cac effect on potentiation, IP3R was necessary to maintain normal odor responsivity when odors were presented repeatedly across multiple trials (whether those were pre/post trials in the conditioning protocol or 10x odor presentations in the adaptation protocol). This is broadly consistent with the temporal role of IP3R in maintenance of presynaptic homeostatic potentiation at the neuromuscular junction. In addition, dopaminergic circuits associated with reward learning drive release of Ca2+ from the endoplasmic reticulum when activated with KCs in a backward pairing paradigm ex vivo, potentiating MB γ4 connections with the respective γ4 MBON. This is consistent with a role for ER calcium in positively regulating synaptic strength (Stahl, 2022).

Potentiation and depression of ACh release was observed across multiple MB compartments following conditioning, providing a presynaptic mechanism that potentially contributes to shaping conditioned MBON responses. Importantly, by comparing the CS+ and CS- responses to those of untrained odors, plasticity was ascribed to potentiation or depression (accounting for any non-associative olfactory adaptation) within each compartment. This is relevant for modeling efforts, where it has been unclear whether to include potentiation (along with depression) in the learning rule(s) at KC-MBON synapses. In addition, the experiments revealed an additional layer of spatial regulation in the γ1-γ3 compartments: a gradient of CS+ potentiation to CS- depression following appetitive conditioning. Specifically, the CS+/CS- relationship changed in a linear gradient down the γ1-γ3 compartments following appetitive conditioning. Appetitive conditioning increased CS+ responses in the γ1 compartment, while decreasing the CS- responses in the γ3 compartment. The γ2 compartment yielded a mix of these responses. These patterns of plasticity have the net effect of increasing the relative response to the CS+ odor (↑CS+:CS-). Since the MBONs postsynaptic to these compartments drive behavioral approach, this would bias the animal to approach the CS+ if it encountered both odors simultaneously. Such a situation occurs at the choice point of a T-maze during retrieval in a classical conditioning assay. The CS+ and CS- produce different patterns of plasticity at different loci (e.g. γ1 vs γ3), which presumably coordinate to regulate behavior via temporal integration of the odor and US cues. The CS+ is temporally contiguous with the US, while the CS- is nonoverlapping. Therefore, the timing of CS/US pairing drives plasticity differently in each compartment. These patterns of plasticity presumably coordinate to regulate memory formation and action selection during retrieval. For instance, while the γ1-γ3 compartments exhibited ↑CS+:CS- following appetitive conditioning, the γ5 compartment exhibited plasticity in the opposite direction: decreasing the relative response to the CS+ odor (↓CS+:CS-). As the γ5 compartment is presynaptic to an avoidance-promoting MBON, this plasticity pattern would coherently contribute to biasing the animal toward CS+ approach (reducing CS+ avoidance). Thus, it would work in concert with the plasticity in γ1-γ3 to bias the animal toward behavioral approach. Overall, plasticity is regulated in each MB compartment individually by the timing of events and the valence of the US, with the changes coordinated across multiple compartments to coherently drive behavior (Stahl, 2022).

Behaviorally, MBONs innervating the γ lobe variably drive behavioral approach or avoidance when stimulated. Despite the approach-promoting valence of the γ2α'one and γ3/γ3β'1 MBONs, among them, only the γ3/γ3β'1 MBONs produced a loss-of-function phenotype in behavioral appetitive conditioning. This suggests that redundancy and/or different weighting across approach promoting MBONs renders the system resilient to silencing some of them. A previous study found effects of blocking the γ2α'1 MBONs, though not γ3/γ3β'1 MBONs, when blocking individual steps of memory processing (acquisition, retention, and/or retrieval) with a 1 hr appetitive memory protocol. This suggests that the different MBONs have differing roles across time, with some redundancy in appetitive processing. Blocking synaptic output of γ3/γ3β'1 MBONs reduced appetitive conditioning performance in these experiments immediately following conditioning, suggesting that these neurons play a specific role in appetitive short-term memory (Stahl, 2022).

The present and previous studies suggest that alterations of MBON activity following learning are the product of both presynaptic and postsynaptic plasticity at the KC-MBON synapses, as well as feedforward inhibition. Blocking synaptic output from KCs impairs the acquisition of appetitive memories (30-60 min after conditioning), suggesting a role for postsynaptic plasticity. However, this does not rule out presynaptic plasticity, as blocking KC output (with R13F02) leaves signaling from reinforcing dopaminergic neurons partially intact, which likely shapes the presynaptic KC responses via heterosynaptic plasticity. At the circuit level, polysynaptic inhibition can convert depression from select MB compartments into potentiation in MBONs following learning; in one established example, reduction of odor-evoked responses in the GABAergic γ1pedc MBON following aversive conditioning disinhibits the downstream γ5β'2 a MBON (Stahl, 2022).

KC-MBON synapses represent one node of learning-related plasticity, which is distributed across multiple sites during learning. Short-term memory-related plasticity has been observed in multiple olfactory neurons, such as the antennal lobes. In addition, connectomics studies have revealed complex connectivity within and beyond the MB, which is a multi-layered network including circuit motifs that influence the propagation of information and generation of plasticity during learning. Such connections include recurrent feedback. Some of these recurrent connections are from cholinergic MBONs that synapse within the MB, which could have contributed to the ACh signals observed in this study. For instance, the γ2α'1 MBON is a cholinergic MBON that sends ~6% of its output back to the γ lobe. Some of the recurrent connections are formed by dopaminergic neurons, such as the PAM γ4<γ1/y2. In addition, reciprocal connections between KCs and dopaminergic neurons in the vertical lobes are necessary for memory retrieval. This adds another layer of recurrent circuitry that may participate in reinforcement during associative learning. Across these circuits, some neurons corelease several neurotransmitters and act on an array of postsynaptic receptors, which contribute to plasticity distributed across multiple sites (Stahl, 2022).

Overall, plasticity between KCs and MBONs may guide behavior through biasing network activation to alter action selection in a probabilistic manner. Appetitive conditioning drives compartmentalized, presynaptic plasticity in KCs that correlates with postsynaptic changes in MBONs that guide learned behaviors. Prior studies documented only depression at these synapses at short time points following conditioning. This study observed both potentiation and depression in ACh release in the MB, suggesting that bidirectional presynaptic plasticity modulates learned behaviors. These bidirectional changes likely integrate with plasticity at downstream circuit nodes that also undergo learning-induced plasticity to produce network-level alterations in odor responses across the olfactory pathway following salient events. Thus, plasticity in ACh release from KCs functions to modulate responsivity to olfactory stimuli features across graded plasticity maps down the MB axons (Stahl, 2022).

Go to top

Transient active zone remodeling in the Drosophila mushroom body supports memory

Turrel, O., Ramesh, N., Escher, M. J. F., Pooryasin, A. and Sigrist, S. J. (2022). Curr Biol 32(22): 4900-4913. PubMed ID: 36327980

Elucidating how the distinct components of synaptic plasticity dynamically orchestrate the distinct stages of memory acquisition and maintenance within neuronal networks remains a major challenge. Specifically, plasticity processes tuning the functional and also structural state of presynaptic active zone (AZ) release sites are widely observed in vertebrates and invertebrates, but their behavioral relevance remains mostly unclear. This study provides evidence that a transient upregulation of presynaptic AZ release site proteins supports aversive olfactory mid-term memory in the Drosophila mushroom body (MB). Upon paired aversive olfactory conditioning, AZ protein levels (ELKS-family BRP/(m)unc13-family release factor Unc13A) increased for a few hours with MB-lobe-specific dynamics. Kenyon cell (KC, intrinsic MB neurons)-specific knockdown (KD) of BRP did not affect aversive olfactory short-term memory (STM) but strongly suppressed aversive mid-term memory (MTM). Different proteins crucial for the transport of AZ biosynthetic precursors (transport adaptor Aplip1/Jip-1; kinesin motor IMAC/Unc104; small GTPase Arl8) were also specifically required for the formation of aversive olfactory MTM. Consistent with the merely transitory increase of AZ proteins, BRP KD did not interfere with the formation of aversive olfactory long-term memory (LTM; i.e., 1 day). These data suggest that the remodeling of presynaptic AZ refines the MB circuitry after paired aversive conditioning, over a time window of a few hours, to display aversive olfactory memories (Turrel, 2022).

Synapses are key sites of information processing and storage in the brain. Notably, synaptic transmission is not hardwired but adapts through synaptic plasticity to provide appropriate input-output relationships as well as to process and store information on a circuit level. Still, there are fundamental gaps in understanding of exactly how the dynamic changes of synapse performance intersect with circuit operation and consequently define behavioral states. This is partly due to the inherent complexity of synaptic plasticity mechanisms, which operate across a large range of timescales (sub-second to days) and use a rich spectrum of both pre- and post-synaptic molecular and cellular mechanisms. Lately, refinement processes following the immediate engram formation have been described, which might promote specific neuronal activity patterns to select neurons for longer-term information display and storage (Turrel, 2022).

Synaptic transmission across chemical synapses is evoked by action potentials that activate presynaptic Ca2+ influx through voltage-gated Ca2+ channels to trigger the fusion of synaptic vesicles (SVs) containing neurotransmitter at sites called active zones (AZs). AZs assemble from conserved scaffold proteins, including ELKS (Drosophila ortholog: BRP), RIM, and the RIM-binding protein (RBP) family. Recent work in Drosophila showed that discrete SV release sites form at AZ. In the AZ, the ELKS-family BRP master scaffold protein localizes the critical Munc13 family release factor Unc13A in defined nanoscopic clusters around Ca2+ channels (BRP/Unc13A nanomodules). This AZ architecture of the nanoscale organization between BRP/Unc13 release machinery and the AZ-centric Ca2+ channels is present across all Drosophila synapses, including Kenyon cell (KC) derived AZs, and munc13-clusters also define release sites at central mammalian synapses. Importantly, AZ structure and function is dynamic and can remodel within 10 min, as shown at Drosophila neuromuscular junction (NMJ) synapses (Turrel, 2022).

The Drosophila mushroom body (MB) forms and subsequently stores olfactory memories. Importantly, a depression of SV release from the AZ of intrinsic KCs within specific compartments of the MB lobes was found to promote the formation of olfactory memories within a few minutes of paired conditioning. Indeed, Ca2+ in vivo imaging experiments indicate that dopamine bidirectionally tunes the strength of KC synapses to output neurons, with forward conditioning driving depression of those synapses and backward conditioning generally driving potentiation. How this tuning is executed at AZ level is not yet known (Turrel, 2022).

This study present evidence for AZ remodeling (BRP, Syd1, and Unc13A) to take place within MB lobes after paired conditioning for a few hours and provide genetic evidence that this AZ remodeling within the MB-intrinsic KCs is crucial for mid-term aversive olfactory memories. To identify candidate mechanisms of presynaptic remodeling to then be tested in MB-dependent olfactory memory, the role of AZ remodeling was studied during extended larval NMJ plasticity and relevant transport factors were identified. These data suggest that broad but transient changes of presynaptic AZs depending on the transport of new biosynthetic material support refinement processes within KC and MB circuitry and are specifically needed for stable formation of mid-term olfactory memories (Turrel, 2022).

Historically, postsynaptic plasticity mechanisms have been analyzed extensively, and molecular and cellular processes targeting postsynaptic neurotransmitter receptors have been convincingly connected to learning and memory. At the same time, the necessity of using postsynaptic neurons as reporters of presynaptic activity (and, thus, setup paired recordings) has imposed an additional obstacle specific to the functional study of presynaptic forms of mid- and long-term plasticity. Furthermore, the cellular and molecular processes remodeling presynaptic AZs are not characterized as extensively as those at the postsynapse. Consequently, although widely expressed by excitatory and inhibitory synapses of mammalian brains, the behavioral relevance of longer-term presynaptic plasticity remains largely obscure (Turrel, 2022).

This study combined the possibility of genetically analyzing memory formation and stabilization within discrete neuron populations of the Drosophila MB with the identification of molecular machinery remodeling presynaptic AZs in vivo. Evidence is provided for an extended but temporally restricted (a few hours post training) upregulation of presynaptic AZ proteins across the MB lobes, a process seemingly needed in MB intrinsic neurons to display olfactory MTM (Turrel, 2022).

Notably, the acute formation of aversive STM was previously shown to trigger synaptic depression at the KC::MBON synapse in the respective MB compartments. It is emphasized that the exact relation of the AZ remodeling described in this study to this STM-controlling short-term depression is presently unknown. Particularly, it is not possible to tell whether the conditioning-associated presynaptic remodeling described in this study is indeed potentiating KCs and MB AZs or whether overlapping sets of synapses are involved in STM and MTM formation and display. What can be concluded, however, is that molecular machinery that executes structural remodeling at NMJ AZs is critically needed for MTM within the MB intrinsic neurons. Establishing the degree to which synaptic weight changes are associated with the mechanism of MB presynaptic remodeling will have to await the development of protocols to directly follow synapses in vivo for hours after conditioning. Different from presynaptic remodeling being part of the memory trace or engram itself, the idea is favored that synaptic upregulation might instead execute a refinement function extending over larger parts of the MB AZ populations. Refinement is an emerging concept stating that stable propagation and maintenance of memory traces might depend on homeostatic regulations of neuronal circuitry. Sleep-dependent synaptic plasticity is suggested to similarly play an important role in neuronal circuit refinement after learning (Turrel, 2022).

Notably, it has been recently shown that a similar upregulation of AZ proteins (BRP/Unc13A) is indeed a functional part of Drosophila sleep homeostasis, where it suffices to trigger rebound sleep patterns. It thus appears conceivable that the AZ changes associated with conditioning reported in this study might promote specific MB activity patterns instrumental for MTM. An alternative, not mutually exclusive, option is that the initial synaptic depression associated with aversive conditioning must, on a longer term, be compensated by the MB AZ changes (and potential potentiation) described in this study (Turrel, 2022).

Notably, compartment-specific synaptic changes occur in the MB in response to sheer odor presentation or DAN activity although AZ remodeling in this study behaved strictly conditioning dependent, meaning it was not observed after unpaired conditioning, and appeared broadly distributed. It cannot be excluded, however, that smaller size, compartment-specific AZ changes, have been missed, given the limited resolution of the staining assays (Turrel, 2022).

Cell biological processes remodeling presynaptic AZs at larval NMJ synapses can also be of relevance for memory formation in the adult fly KCs. Concretely, this study found that the MB KC-specific KD of transport factors, which at the NMJ level provoked plasticity profiles similar to BRP, also specifically affected MTM but spared STM. Given that several molecular factors, including transport proteins not directly physically associated with the AZ, fulfilled this relation, it indeed appears likely that retrieving axon-transported biosynthetic AZ precursor material is what is critical here (Turrel, 2022).

Speaking of the specificity of rthe MTM phenotypes in relation to AZ remodeling, this study found STM formation undisturbed, but at the same time, MTM to be severely affected after BRP and transport factor KD. This is strong evidence against the possibility of baseline synaptic defects being responsible for the observed MTM deficits. It is also emphasized that this study achieved behavioral phenotypes by comparatively mild and strictly post-developmental KD and that odor Ca2+ responses in MBON neurons postsynaptic to KC appeared normal in BRP KD flies (Turrel, 2022).

When analyzing in a MB-lobe-specific manner, α/&betal and α'/β' neurons showed stronger and more sustained upregulation of BRP/Unc13A than the γ lobes. This might indicate that the extent and role of refinement across the MB lobes is adapted to their specific roles in memory acquisition and retrieval. This is also in accordance with previous observations showing heterogeneity in the exact AZ protein composition across synapses of the Drosophila brain (Turrel, 2022).

Interestingly, Syd-1 levels are significantly increased 1 h after conditioning in the α/β and α'/β' lobes, whereas it has been shown that Syd-1 levels are not increased 10 min after PhTx treatment at the NMJ. This finding indicates that some of the AZ proteins may be affected differently in those two plasticity processes (Turrel, 2022).

Given the generally observed sparse representation of odors within the MB KCs, one might expect initial synaptic changes to be specific to only a few odor-response KCs. Still, this analysis apparently reveals more extended changes of synaptic AZs across the lobes. Potentially, upon successful conditioning, the initial, more restricted, synaptic changes might be followed by an extended communication between the neurons involved in the memory circuit, potentially including KC::KC communication. Indeed, there is ample evidence for a transfer of requirement between different subsets of KCs in the temporal evolution of olfactory memory. This communication seemingly involves gap junctions between KCs but might in parallel also use chemical synapses and their AZs. Concerning the broad distribution of the AZ changes across compartments, it is interesting to mention that KC-global, conditioning-dependent metabolic changes have been observed, being critical for LTM but also MTM (Turrel, 2022).

It is tempting to speculate that the initial, compartment-specific changes, confined to a few odor-responding KCs, might overcome a threshold to also trigger more global synaptic changes. Also interesting in this context, dorsal paired medial (DPM) neurons' odor response increase following spaced conditioning, also indicating that opposite synaptic strength changes might counterbalance the initial synaptic changes occurring in the memory-relevant compartment or depending on post-synaptic partner neurons provoke either potentiation or depression (Turrel, 2022).

As mentioned above, this study found that KD of BRP in the adult MB lobes did not affect LTM, whereas MTM was decreased both at 1 and 3 h. Such a phenotype, a deficit of MTM but subsequent memory phases being intact, was only rarely observed before (Nep2-RNAi in adult DPM neurons, synapsin mutants with memory deficits up to 1 h but normal memory later on). On one hand, this reinforces the idea that MTM and LTM might form using separate circuits, and on the other hand, that cell types other than KCs might contribute to aversive olfactory LTM formation. Different sets of proteins in the same lobes might operate in parallel circuits similar to what has been observed in the honeybee. However, it might also well be that the presynaptic AZ remodeling observed in this study is indeed specific for the display of MTM and that the synaptic memory traces orchestrating the later recall of LTM are mediated by independent parallel molecular/synaptic mechanisms or distinct circuit (Turrel, 2022).

Go to top

Mechanisms underlying homeostatic plasticity in the Drosophila mushroom body in vivo

Apostolopoulou, A. A. and Lin, A. C. (2020). Proc Natl Acad Sci U S A 117(28): 16606-16615. PubMed ID: 32601210

Neural network function requires an appropriate balance of excitation and inhibition to be maintained by homeostatic plasticity. However, little is known about homeostatic mechanisms in the intact central brain in vivo. Homeostatic plasticity was studied in the Drosophila mushroom body, where Kenyon cells receive feedforward excitation from olfactory projection neurons and feedback inhibition from the anterior paired lateral neuron (APL). Prolonged (4-d) artificial activation of the inhibitory APL causes increased Kenyon cell odor responses after the artificial inhibition is removed, suggesting that the mushroom body compensates for excess inhibition. In contrast, there is little compensation for lack of inhibition (blockade of APL). The compensation occurs through a combination of increased excitation of Kenyon cells and decreased activation of APL, with differing relative contributions for different Kenyon cell subtypes. This findings establish the fly mushroom body as a model for homeostatic plasticity in vivo (Apostolopoulou, 2020).

Go to top

Mitochondrial haplotypes affect metabolic phenotypes in the Drosophila Genetic Reference Panel

Bevers, R. P. J., Litovchenko, M., Kapopoulou, A., Braman, V. S., Robinson, M. R., Auwerx, J., Hollis, B. and Deplancke, B. (2019). Nat Metab 1(12): 1226-1242. PubMed ID: 32694676

The nature and extent of mitochondrial DNA variation in a population and how it affects traits is poorly understood. This study resequenced the mitochondrial genomes of 169 Drosophila Genetic Reference Panel lines, identifying 231 variants that stratify along 12 mitochondrial haplotypes. 1,845 cases of mitonuclear allelic imbalances were identified, thus implying that mitochondrial haplotypes are reflected in the nuclear genome. However, no major fitness effects are associated with mitonuclear imbalance, suggesting that such imbalances reflect population structure at the mitochondrial level rather than genomic incompatibilities. Although mitochondrial haplotypes have no direct impact on mitochondrial respiration, some haplotypes are associated with stress- and metabolism-related phenotypes, including food intake in males. Finally, through reciprocal swapping of mitochondrial genomes, it was demonstrated that a mitochondrial haplotype associated with high food intake can rescue a low food intake phenotype. Together, these findings provide new insight into population structure at the mitochondrial level and point to the importance of incorporating mitochondrial haplotypes in genotype-phenotype relationship studies (Bevers, 2020).

Go to top

Dopaminergic neurons write and update memories with cell-type-specific rules

Aso, Y. and Rubin, G. M. (2016). Elife 5. PubMed ID: 27441388

Associative learning is thought to involve parallel and distributed mechanisms of memory formation and storage. In Drosophila, the mushroom body (MB) is the major site of associative odor memory formation. The anatomy of the adult MB and 20 types of dopaminergic neurons (DANs) that each innervate distinct MB compartments have been previously defined and described. This study compared the properties of memories formed by optogenetic activation of individual DAN cell types. Extensive differences were found in training requirements for memory formation, decay dynamics, storage capacity and flexibility to learn new associations. Even a single DAN cell type can either write or reduce an aversive memory, or write an appetitive memory, depending on when it is activated relative to odor delivery. These results show that different learning rules are executed in seemingly parallel memory systems, providing multiple distinct circuit-based strategies to predict future events from past experiences. The mechanisms that generate these distinct learning rules are unknown. They could arise from differences in the dopamine release properties of different DAN cell types or from local differences in the biochemical response to dopamine signaling in each MB compartment. For example, KCs express four distinct dopamine receptors, which might be deployed differently in each compartment. Or they could originate from circuit properties (Aso, 2016).

Punitive or rewarding stimuli such as electric shock, heat, cold, bitter taste and sugar generally activate a complex pattern of DANs. Believing that a reduction in complexity will be essential to understand the roles played by DAN inputs to different MB compartments, intersectional split-GAL4 drivers were used to express CsChrimson, a red-shifted channelrhodopsin, in specific cell types. While activation of CsChrimson in these driver lines provides a stimulus that is unlikely to occur naturally, it allowed separate examination of the memory components induced by individual DAN cell types (Aso, 2016).

An olfactory arena was developed that allows fine temporal control of both odor delivery and DAN activation using optogenetics (see Learning rules in one MB compartment). In this arena, freely moving flies can be repeatedly trained and tested without the manual handling or temperature changes required in previous assays, thereby minimizing variability that might obscure subtle behavioral effects. These methods allowed systematic examination of the properties of memories induced in different MB compartments, including: (1) the temporal pairing requirements of odor presentation and DAN activation; (2) the amount of training required for memory formation; (3) retention time; (4) weakening of the conditioned response induced by either DAN activation in the absence of odor presentation or by odor presentation in the absence of DAN activation; (5) the ability to learn new associations; and (6) the capacity to store multiple memories (Aso, 2016).

Dopamine signaling to KCs has been implicated in both learning and forgetting. However, it has not been determined if a single DAN cell type can drive both processes within the same compartment. Pairing one of two odors with activation of PPL1-γ1pedc, results in robust aversive memory to the paired odor. That memory is fully retained after 10 min but has largely decayed by 24 hr. Presentation of the odors alone a few minutes after training resulted in a modest reduction in the conditioned response. A second activation of the same DAN a few minutes after training in the absence of odor can almost completely abolish the conditioned response. Recent imaging data of the γ4 MBON suggest that this reduction most likely results from restoration of the response of the MBON to the odor; that is, erasure of the memory. If, after the first training, contingencies are reversed such that the other odor is presented paired with DAN activation, the first memory is reduced and a memory of the new association is formed. Taken together, these data indicate that the same DANs can write a new memory or reduce an existing conditioned response, enabling the flexibility to rapidly change the associations formed between a conditioned stimulus (CS), the odor, and an unconditioned stimulus (US) represented by dopamine release (Aso, 2016).

In classical conditioning, both the rate of learning and the valence of the resultant memory depend on the relative timing between the CS and the US. The ability to precisely control DAN stimulation and odor presentation enabled examination of the CS-US timing relationship. PPL1-γ1pedc stimulation fell within a 30-s time window following the onset of a 10-s odor presentation, an aversive memory was formed. Interestingly, it was observed that DAN stimulation that precedes odor presentation by 20 to 60 s induced an appetitive memory. These observations are consistent with the notion that it is the predictive aspect of CS-US timing that matters. When the timing is such that the CS predicts a subsequent aversive US, animals learn to avoid the CS. However, animals can also learn that the CS predicts the end of an aversive US, and are consequently attracted to the CS. It has been suggested that this timing dependency could result from the dynamics of the biochemical signaling cascade acting downstream of dopamine receptors (Aso, 2016).

Pairing activation of PPL1-γ1pedc with odor has been shown to depresses the subsequent spiking rate of the MBON from the γ1pedc compartments in response to the trained odor (Hige, 2015). In behavioral assays, optogenetic activation of this MBON was shown to attract flies. Taken together, these observations suggest that DAN activation paired with an odor produces an aversive behavioral response to that odor by decreasing the MBON's attractive output. Thus the data can be most easily explained if this single DAN can bi-directionally alter the strength of KC-MBON synapses depending on the presence and relative timing of odor-driven KC activity; a full testing of this model awaits additional physiological measurements (Aso, 2016).

In order to compare the parameters of learning in different MB compartments, a set of additional split-GAL4 drivers were selected that express at similar and high levels in different DAN cell types. In addition, the three DAN cell types, which express CsChrimson using the MB320C and MB099C drivers, have been shown to have similar spiking responses to CsChrimson activation (Hige, 2015). In three of six cases, drivers were chosen expressing in a combination of two cell types because it was found that activation of only a single DAN cell type did not produce a sufficiently robust memory (see Rules for writing memory). It was confirmed that the lines used in the optogenetic experiments showed comparable memory formation when trained with electric shock or sugar reward. Together, the selected drivers innervate 11 of the 15 MB compartments. Below we describe the results obtained in a number of different learning assays by activating these split-GAL4 drivers (Aso, 2016).

Long-term aversive memory in flies requires repetitive electric shock conditioning with resting intervals, so-called spaced training. Two sets of DANs, PPL1-α3 alone (MB630B) or the combination of PPL1-γ2α'1 and PPL1-α'2α2 (MB099C) can induce 1-day and 4-day aversive memory after spaced training, suggesting that the effects of spaced training can be implemented in individual compartments. Making memory formation dependent on repetitive training might be beneficial by allowing an animal to ignore spurious one-time events. Recent work has shown that the γ2α'1 compartments play key roles in both sleep regulation and long-term memory. The observation that co-activation of PPL1-γ2α'1 and other DANs synergistically prolongs memory retention raises the possibility that PPL1-γ2α'1 might act broadly to facilitate memory consolidation by promoting sleep after learning (Aso, 2016).

A particular DAN's ability to induce the formation of immediate, 1-day and 4-day memories was not correlated. For example, immediate memory after a single pairing with activation of PPL1-α3 (MB630B) was barely detectable, although multiple activations resulted in 4-day memory. In contrast, PPL1-γ1pedc (MB320C) activation resulted in robust immediate memory acquisition after a single round of training, but its activation failed to induce 4-day memory even after extensive spaced training. These results imply the stability of memory is an intrinsic property of the MB compartment, rather than a consequence of the training protocol. In this view, repetitive training with naturalistic stimuli that activate many DAN cell types would recruit additional compartments with slower acquisition rates and the behaviorally assayed retention of memory would reflect the combined memories formed in different compartments. It remains an open question whether short-term memories are converted into long-term memories as biochemical changes in the same synapses or whether these memories are formed separately and in parallel. For olfactory learning in Drosophila, the data are consistent with a model in which memory formation and consolidation can occur independently and in parallel in individual MB compartments; this view does not exclude the possibility that network activity facilitates memory consolidation (Aso, 2016).

The memories induced in different compartments have different stabilities, displaying different dynamics of spontaneous memory decay over a 1-day period. Memories in each compartment also differed in the extent to which they were reduced by a second presentation of the trained odor without reinforcement (Aso, 2016).

Unlike with immediate memory that was induced by a single training, memory induced by spaced training with MB099C or MB630B and tested after 1 d does not show a significant reduction. Likewise, DAN activation without odor presentation significantly reduced immediate memory for four of the five sets of DANs tested. These two effects might be mechanistically linked as odor presentation alone can result in activation of a subset of dopaminergic neurons (Aso, 2016).

In both the case of presentation of odor without dopamine and of dopamine without odor, the association the fly had previously learned is not confirmed. It would make sense for a memory to be diminished when the contingency upon which it is based is found to be unreliable. Consistent with this idea, repetitive spaced training with these same DANs can induce 1-day memory that is resistant to DAN activation. The differences observed between compartments suggest that they weigh the importance of the reliability of the correlation between CS and US differently (Aso, 2016).

The α1 compartment differed from the other compartments tested in that it was resistant to memory reduction by DAN activation. This compartment plays a key role in long-term appetitive memory of nutritious foods and has an unusual circuit structure: its MBON (MBON-α1) appears to form synapses on the dendrites of the DAN that innervates the α1 compartment (PAM-α1) forming a recurrent circuit necessary for long-term memory formation. The α1 compartment also showed the least ability to replace an older association with a new one. This observation suggests that the initial memory may not be affected by the second training, resulting in co-existing appetitive memories for both odors. Indeed, flies were able to retain associations between each of two odors and PAM-α1 (MB043C) activation, while only the most recently learned association was remembered with PPL1-γ1pedc (MB320C) activation. The higher memory capacity of the α1 compartment is not due to generalization, since training with one odor pair did not affect the innate odor preference observed with a different, untrained odor pair. Thus two distinct strategies for updating memories appear to be used in different MB compartments: (1) writing a new memory, while diminishing the old memory; or (2) writing a new memory, while retaining the old memory (Aso, 2016).

The results suggest that memory formation in each compartment is largely parallel and independent, with compartmental specific rules for updating memories. Such a model of independent memory storage should allow appetitive and aversive memories to be simultaneously formed for the same odor in different compartments. This idea was tested by simultaneously activating DANs to α1 and γ1pedc while exposing flies to an odor. When flies were tested immediately after training, the odor was strongly aversive, but the same odor became appetitive after 1 day. These results are most easily explained by simultaneous formation of an aversive memory in γ1pedc and an appetitive memory in α1, with rapid decay of the memory in γ1pedc and slow decay in α1 resulting in a shift in valence of the conditioned response over time. However, the fact that strongly aversive immediate memory was observed, rather than an intermediate response, suggests that the MB network non-linearly integrates these conflicting signals. The known feedforward connection between γ1pedc and α1 provides a possible circuit mechanism. Recent studies provide further examples most easily explained by parallel induction of conflicting memories of different decay rates. It was also found that wild type flies are capable of efficiently switching odor preference when they had conflicting sequential experiences of sugar reward followed by shock punishment with the same odor (Aso, 2016).

These results demonstrate that different MB compartments use distinct rules for writing and updating memories of odors. By analyzing individual memory components-or engrams-induced by local dopamine release, this study found that the interpretation of a common odor representation carried by sparse KC activity to multiple compartments could be modified differently in each of those compartments. The mechanisms that generate these distinct learning rules are not known. They could arise from differences in the dopamine release properties of different DAN cell types or from local differences in the biochemical response to dopamine signaling in each MB compartment. For example, KCs express four distinct dopamine receptors, which might be deployed differently in each compartment. Or they could originate from circuit properties: it is known from anatomical, behavioral and functional imaging studies that MB compartments can communicate through connections between their extrinsic neurons, the DANs and MBONs, as well as by a layered network within the MB. In the mammalian brain, associative memories are also stored as distributed and parallel changes with partially overlapping functions; for example, different populations of dopaminergic neurons develop representations of a visual objects' value with distinct learning rules. It is expected that many of the underlying strategies and mechanisms may be shared between flies and other species. This work provides a foundation for experiments aimed at understanding the molecular and circuit mechanisms by which distributed memory components are written with distinct rules and later integrated to guide memory-based behaviors (Aso, 2016).

Go to top

Re-evaluation of learned information in Drosophila

Felsenberg, J., Barnstedt, O., Cognigni, P., Lin, S. and Waddell, S. (2017). Nature 544(7649): 240-244. PubMed ID: 28379939

Animals constantly assess the reliability of learned information to optimize their behaviour. On retrieval, consolidated long-term memory can be neutralized by extinction if the learned prediction was inaccurate. Alternatively, retrieved memory can be maintained, following a period of reconsolidation during which it is labile. Although extinction and reconsolidation provide opportunities to alleviate problematic human memories, a detailed mechanistic understanding of memory updating is lacking. This study identified neural operations underpinning the re-evaluation of memory in Drosophila. Reactivation of reward-reinforced olfactory memory can lead to either extinction or reconsolidation, depending on prediction accuracy. Each process recruits activity in specific parts of the mushroom body output network and distinct subsets of reinforcing dopaminergic neurons. Memory extinction requires output neurons with dendrites in the α and α' lobes of the mushroom body, which drive negatively reinforcing dopaminergic neurons that innervate neighbouring zones. The aversive valence of these new extinction memories neutralizes previously learned odour preference. Memory reconsolidation requires the γ2α'1 mushroom body output neurons. This pathway recruits negatively reinforcing dopaminergic neurons innervating the same compartment and re-engages positively reinforcing dopaminergic neurons to reconsolidate the original reward memory. These data establish that recurrent and hierarchical connectivity between mushroom body output neurons and dopaminergic neurons enables memory re-evaluation driven by reward-prediction error (Felsenberg, 2017).

Go to top

Coincident postsynaptic activity gates presynaptic dopamine release to induce plasticity in Drosophila mushroom bodies

Ueno, K., Suzuki, E., Naganos, S., Ofusa, K., Horiuchi, J. and Saitoe, M. (2017). Elife 6. PubMed ID: 28117664

Simultaneous stimulation of the antennal lobes (ALs) and the ascending fibers of the ventral nerve cord (AFV), two sensory inputs to the mushroom bodies (MBs), induces long-term enhancement (LTE) of subsequent AL-evoked MB responses. LTE induction requires activation of at least three signaling pathways to the MBs, mediated by nicotinic acetylcholine receptors (nAChRs), NMDA receptors (NRs), and D1 dopamine receptors (D1Rs). This study demonstrates that inputs from the AL are transmitted to the MBs through nAChRs, and inputs from the AFV are transmitted by NRs. Dopamine signaling occurs downstream of both nAChR and NR activation, and requires simultaneous stimulation of both pathways. Dopamine release requires the activity of the rutabaga adenylyl cyclase in postsynaptic MB neurons, and release is restricted to MB neurons that receive coincident stimulation. These results indicate that postsynaptic activity can gate presynaptic dopamine release to regulate plasticity (Ueno, 2017).

Go to top

Carbon Monoxide, a Retrograde Messenger Generated in Postsynaptic Mushroom Body Neurons, Evokes Noncanonical Dopamine Release

Ueno, K., Morstein, J., Ofusa, K., Naganos, S., Suzuki-Sawano, E., Minegishi, S., Rezgui, S. P., Kitagishi, H., Michel, B. W., Chang, C. J., Horiuchi, J. and Saitoe, M. (2020) J Neurosci 40(18): 3533-3548. PubMed ID: 32253360

Dopaminergic neurons innervate extensive areas of the brain and release dopamine (DA) onto a wide range of target neurons. However, DA release is also precisely regulated. In Drosophila melanogaster brain explant preparations, DA is released specifically onto alpha3/alpha'3 compartments of mushroom body (MB) neurons that have been coincidentally activated by cholinergic and glutamatergic inputs. The mechanism for this precise release has been unclear. This study found that coincidentally activated MB neurons generate carbon monoxide (CO), which functions as a retrograde signal evoking local DA release from presynaptic terminals. CO production depends on activity of heme oxygenase in postsynaptic MB neurons, and CO-evoked DA release requires Ca(2+) efflux through ryanodine receptors in DA terminals. CO is only produced in MB areas receiving coincident activation, and removal of CO using scavengers blocks DA release. It is proposed that DA neurons use two distinct modes of transmission to produce global and local DA signaling (Ueno, 2020).

Go to top

Presynapses in Kenyon cell dendrites in the mushroom body calyx of Drosophila

Christiansen, F., et al. (2011). J. Neurosci. 31(26): 9696-707. PubMed ID: 21715635

Plastic changes at the presynaptic sites of the mushroom body (MB) principal neurons called Kenyon cells (KCs) are considered to represent a neuronal substrate underlying olfactory learning and memory. It is generally believed that presynaptic and postsynaptic sites of KCs are spatially segregated. In the MB calyx, KCs receive olfactory input from projection neurons (PNs) on their dendrites. Their presynaptic sites, however, are thought to be restricted to the axonal projections within the MB lobes. This study shows that KCs also form presynapses along their calycal dendrites, by using novel transgenic tools for visualizing presynaptic active zones and postsynaptic densities. At these presynapses, vesicle release following stimulation could be observed. They reside at a distance from the PN input into the KC dendrites, suggesting that regions of presynaptic and postsynaptic differentiation are segregated along individual KC dendrites. KC presynapses are present in γ-type KCs that support short- and long-term memory in adult flies and larvae. They can also be observed in α/β-type KCs, which are involved in memory retrieval, but not in α'/β'-type KCs, which are implicated in memory acquisition and consolidation. It is hypothesized that, as in mammals, recurrent activity loops might operate for memory retrieval in the fly olfactory system. The newly identified KC-derived presynapses in the calyx are, inter alia, candidate sites for the formation of memory traces during olfactory learning (Christiansen, 2011).

This study used several approaches to provide evidence that the KC dendrites within the calyx of larval and adult Drosophila are not exclusively postsynaptic. They also form presynaptic active zones (AZs), which was named KCACs. These findings are supported by data from two previous studies, which reported the presence of a presynaptic vesicle protein, Synaptobrevin, in KCs within the calyx. This study used a functional imaging approach to show that KCACs are able to release SVs. Furthermore, which different KC subtypes form KCACs was examined and a detailed description of the KCAC location within the calyx is provided. The presence of these previously undescribed KC-intrinsic presynaptic elements adds a new layer of complexity to the MB microcircuitry (Christiansen, 2011).

Within KC dendrites, AZs and PSDs are clearly organized into discrete subdomains. Here, the question emerges whether a given KC dendrite is either exclusively presynaptic or postsynaptic, or whether both presynaptic and postsynaptic domains can be present within the same KC dendrite in a consecutive fashion. MARCM identified single KCs, which showed Bruchpilot (BRP) puncta spatially segregated from claw-like regions that are thought to harbor the postsynaptic specializations of cholinergic PN::KC synapses. In parallel experiments, these claws were shown clustered the acetylcholine receptor Dα7. Thus, it appears likely that presynaptic and postsynaptic domains can be present within the same KC dendrite (Christiansen, 2011).

Based on BRP-RNAi analysis, it is estimated that ~20%-30% of all presynapses in the calyx are KCACs, in both adults and larvae. These synapses are apparently part of the general calyx microcircuitry. They might synapse onto PNs, KCs themselves, the anterior paired lateral (APL) neuron, modulatory neurons, or so-far-undescribed cells. From this analysis, it appears unlikely that PN boutons are direct postsynaptic partners of the KCACs, as the KCACs appear to be clearly physically segregated from the PN boutons. KCACs might, however, project onto PN axons (Christiansen, 2011).

It appears well possible that KCACs project onto the GABAergic APL neuron, which arborizes in the whole calyx. Within the insect antennal lobe, reciprocal dendrodendritic connections between the PNs and the partially GABAergic local interneurons (LNs) have been described. The PN neurites and the LNs are both transmissive and receptive in the antennal lobe, suggesting a computation between them. KCACs might be involved into similar computations in the calyx. This would be in accordance with EM studies in crickets that suggest presynapses in KCs that connect to GABAergic fibers in the MB calyx (Christiansen, 2011).

KCACs might also mediate KC::KC communication. In fact, dendritic segments of KCs that harbor presynapses appear to run in a parallel fashion. This arrangement could promote the communication between dendritic segments of KCs via dendrodendritic synapses. Such KC::KC synapses could therefore modulate signals originating from the distal segments of the arborizations, which carry odor-evoked signals. By these means, an effective computation between KCs could be accomplished before they transmit their input signals downstream (Christiansen, 2011).

Unfortunately, at the moment no general PSD markers are available in Drosophila. Moreover, the neurotransmitter used by KCs remains unknown. With a general postsynaptic marker or knowledge about the KC transmitter, it would have been possible to generated tools to identify the postsynaptic partners of KCACs. Yet currently, despite efforts, it is only possible to speculate about the postsynaptic partners of KCs in the calyx (Christiansen, 2011).

Memory traces are typically thought to be manifested as plastic changes in neuronal anatomy and physiology that occur in specific brain regions. Several lines of evidence indicate that MBs are causally involved in associative learning of olfactory stimuli. Flies with chemically ablated KCs or mutants lacking the MBs are deficient in olfactory learning. Learning was investigated in flies mutant for the adenylyl cyclase rutabaga (rut), which is suggested to act as a coincidence detector between conditioned stimulus (odor) and unconditioned stimulus (e.g., electric shock). Reexpression of a rut cDNA in a rut- background within a subpopulation of KCs sufficed to restore odor learning. For appetitive learning, reexpression of rut in either PNs or KCs is sufficient to restore the mutant defect, whereas aversive learning is rescued only by rut reexpression in KCs. Reversible disruption of transmitter release in Drosophila KCs, using a temperature-sensitive dynamin transgene, UAS-shibirets1, was shown to block memory retrieval in α/β neurons and acquisition and stabilization of memory in α'/β' neurons. Together, these data imply that MBs play a major role in learning and memory. To form, stabilize, and retrieve memory, KCs use their presynapses. The KC presynapses are so far believed to reside in the lobes (Christiansen, 2011 and references therein).

The biogenic amines octopamine and dopamine are thought to mediate the unconditioned stimulus signal for learning olfactory associations, with octopamine representing appetitive stimuli and dopamine representing aversive stimuli. It has been shown that, in honeybees, sugar can be replaced by octopamine application to the calyx to trigger the conditioned proboscis-extension reflex. In the fruit fly, the amines octopamine and dopamine are released onto MB lobes and calyx. This holds also true for the larva. Therefore, the KCACs might be involved in appetitive learning as well as in aversive learning in fly and larva (Christiansen, 2011).

Notably, this study found that the KC subpopulations α and α/β, but not α'/β', form KCACs. This dichotomy correlates with functional differences in learning and memory that have been assigned to these KC classes in previous studies. For example, α'/β' KCs were reported to be required during and after training to acquire and stabilize olfactory memory, whereas output from α/β neurons was postulated to be required to retrieve memory. It has been proposed that, during acquisition, olfactory information received from PNs is first processed in parallel by the α/β and α'/β' KCs. Notably, activity in α'/β' KCs (which do not form KCACs) is supposed to trigger a recurrent loop between α'/β' KCs and dorsal paired medial neurons, which project to the MB lobes. This loop, in turn, might be necessary for memory consolidation in α/β neurons. Subsequently, memories could be stored in α/β neurons, whose activity is required during recall. As α'/β' neurons are devoid of KCACs, KCACs cannot be involved in the circuit described above. Instead, it is likely that additional, similar recurrent loops exist, which are mediated via KCACs. However, it remains unresolved how exactly KC::KC communication is organized anatomically and functionally. This study now proposes a newly discovered synapse population as candidate sites for KC::KC communication (Christiansen, 2011).

In the mammalian olfactory system, major feedback pathways exist, which project onto neurons one level lower in hierarchy. It has been shown that likewise in Drosophila activation of KCs induced a depolarization in cell bodies of PNs and LNs within the antennal lobes. It was thus suggested that MB lobes provide feedback to the ALs. Moreover, an additional memory trace appears to exist in the antennal lobe, in the PNs. It may therefore well be the case that KCs project onto PNs or onto feedback neurons via their KCACs (Christiansen, 2011).

An urgent question of the field concerns the identification of the postsynaptic partner cells of KC presynapses, which harbor memory traces during olfactory conditioning. It is generally assumed that MB-extrinsic downstream neurons involved in behavioral execution of learned behavior serve as postsynaptic partners here. The current findings raise the possibility that microcircuits inside the MB could be places for further modulation and computation of olfactory processing and/or memory formation and modulation. As a consequence, not only the communication to downstream neurons but also the representation of sensory information within the MB circuitry might be changed by experience. Future analysis using optophysiological tools at the KCACs, together with further anatomical work, should provide answers to these questions (Christiansen, 2011).

Go to top

Integration of the olfactory code across dendritic claws of single mushroom body neurons

Gruntman, E. and Turner, G. C. (2013). Nat Neurosci. 16: 1821-1829. PubMed ID: 24141312

In the olfactory system, sensory inputs are arranged in different glomerular channels, which respond in combinatorial ensembles to the various chemical features of an odor. This study investigated where and how this combinatorial code is read out deeper in the brain. The unique morphology was exploited of neurons in the Drosophila mushroom body, which receive input on large dendritic claws. Imaging odor responses of these dendritic claws revealed that input channels with distinct odor tuning converge on individual mushroom body neurons. How these inputs interact to drive the cell to spike threshold was determined using intracellular recordings to examine mushroom body responses to optogenetically controlled input. The results provide an elegant explanation for the characteristic selectivity of mushroom body neurons: these cells receive different types of input and require those inputs to be coactive to spike. These results establish the mushroom body as an important site of integration in the fly olfactory system (Gruntman, 2013).

There are two basic requirements that must be fulfilled to show that neurons read the combinatorial code of the early olfactory layers: they receive convergent input from functionally distinct glomerular channels and require activation of multiple inputs to respond. Using in vivo imaging of the individual synaptic inputs, this study found that Kenyon cells receive inputs with functionally distinct odor response properties. By optogenetically stimulating a defined subset of projection neurons and intracellularly recording from postsynaptically connected Kenyon cells, it was found that several of those inputs must be activated to evoke a spiking response in a downstream Kenyon cell. These results indicate that Kenyon cells respond to specific combinations of coactive glomerular channels. This is likely the basis for their highly stimulus-specific response properties. Notably, this study found that synaptic summation is essentially linear in Kenyon cells, and multiple lines of evidence suggest that Kenyon cell dendrites have purely passive cable properties. Using patterned photostimulation, others have shown that neurons in olfactory cortex respond selectively to particular patterns of glomerular activation. Thus, the integration of different channels is likely to be a fundamental aspect of the transformation at the third layer of the olfactory system. (Gruntman, 2013).

Anatomical studies originally suggested that the mushroom body might be an important site for the convergence of different glomerular channels. Projection neurons send wide-ranging projections in the calyx of the mushroom body, and the synaptic terminals of different projection neuron types intermingle in this area. Moreover, single-cell labeling suggests that the dendrites of individual Kenyon cells extend rather widely in the mushroom body calyx, suggesting they could collect input from different projection neuron types. Retrogradely tracing the inputs to individual Kenyon cells showed that projection neurons of different glomerular origin converge onto individual Kenyon cells. This study found no statistical structure in the probability of different projection neurons converging onto the same Kenyon cell, suggesting that connectivity at this layer is random. This result contrasts somewhat with earlier anatomical studies that found that different projection neuron types show coarsely regionalized projections, although there was extensive overlap between different projection zones. It has also been shown that different subtypes of Kenyon cells tend to innervate particular regions of the mushroom body calyx. Consistent with this loose regionalization, one study reported a correlation in the projection patterns of particular types of projection neurons and Kenyon cells. Altogether, the evidence from these global mapping studies supports a model in which projection neuron-Kenyon cell connectivity is probabilistic and regionally biased rather than completely random (Gruntman, 2013).

This study took a functional approach to the question of convergence by directly imaging odor responses of individual synaptic sites. The results indicate that functionally distinct inputs converge onto the dendritic trees of individual Kenyon cells. When the similarity of odor response profiles was examined for the claws of a given Kenyon cell, it was found that they were more similar to one another than they were to claws from different cells. This suggests that projection neurons with similar tuning properties have a tendency to converge onto the same Kenyon cells. In addition, when ChR2 stimulation was used to examine the relationship between anatomical and functional connectivity, the levels of connectivity that were observed were significantly different from that predicted by random convergence. Overall, the results were more consistent with the model of regionalized, probabilistic connectivity derived from global mapping of projection neuron and Kenyon cell projections than that of entirely random convergence supported by retrograde labeling. The functional approach may have proved more sensitive to detecting correlations in connectivity than a purely anatomical one; correlated response properties are important determinants of wiring in many neural circuits. Nevertheless, the dominant theme of the current results was that functionally distinct inputs converge onto the dendrites of individual Kenyon cells. Convergence is likely a core feature of this layer of the olfactory circuit, as trans-synaptic tracing has shown that neurons in the piriform cortex receive mitral cell input originating from several different glomeruli (Gruntman, 2013).

A recent study investigated the input-output transformation of Kenyon cells by imaging activity in projection neuron boutons and Kenyon cell axons (Li, 2013). Examining odor-evoked activity in these different cells, the authors found that the summed activity of the projection neuron inputs correlates well with the likelihood of an axonal response in a postsynaptic Kenyon cell. In contrast, this study used an electrophysiological approach to examine synaptic responses to optogenetically controlled projection neuron inputs, enabling investigation of the integration of inputs from the dendritic claws. By recording from randomly selected Kenyon cells and reconstructing their morphology post hoc, both functional and anatomical measures of connectivity were establised for each cell recorded. Comparing responses of Kenyon cells with different levels of connectivity allowed determinination of how different synaptic inputs interact and allowed establishment of the relationship between synaptic input and spiking in these cells (Gruntman, 2013).

Dendritic inputs from different sites could interact synergistically; for example, by summating to activate voltage-gated channels that boost the synaptic response. Such synergistic interactions have been observed in mammalian barrel cortex, where coordinated input to different dendritic sites from different sensory modalities elicits a distinct bursting response mode in these cells. Synergistic interactions could make it easier for Kenyon cells to respond selectively to particular input patterns; they could enable a clear distinction between activation of a small set of inputs that is not sufficient to bring the cell to spike and activation of a larger, supra-threshold number of inputs. Electrical stimulation experiments in locust Kenyon cells have shown that synaptic responses are amplified and temporally sharpened by a voltage-dependent mechanism. In contrast, the current results with Drosophila Kenyon cells revealed that claws interact linearly when small numbers are coactive, and sublinearly when larger numbers are stimulated, likely as a result of shunting effects on the dendritic tree. Note that the stimulation conditions were designed to deliver high-frequency projection neuron input in a narrow time window, conditions that are well-suited to reveal synergistic interactions between coincident input. In separate experiments, the interaction between two precisely timed inputs was examined: synaptic stimulation at the dendrites and current injection at the soma. The additivity of these inputs was examined at a variety of different relative timings, and again no evidence was found for synergistic effects that would indicate heightened sensitivity to coincident input. Rather, the results are consistent with a model in which Kenyon cell dendrites serve as passive cables that convey and integrate synaptic input. Notably, this may enable Kenyon cells to take advantage of the format of projection neuron population activity. The antennal lobe transforms olfactory signals so that population representations are more linearly separable in the projection neurons than in the ORNs. By acting as linear integrators, the Kenyon cell dendrites would be well-suited to separate distinct projection neuron activity patterns (Gruntman, 2013).

The results are reminiscent of findings from third-order neurons in vertebrates. A recent study of cells in the dorsal telencephalon of zebrafish showed that these cells exhibit no special sensitivity to synchronous mitral cell inputs. Rather, input synchronization was important to control the precise timing of dorsal telencephalon spikes. Similarly, in mammalian olfactory cortex, synaptic inputs arrive in alternating waves of excitation and inhibition, which enforce temporally precise spiking, phase-locked to this oscillation cycle. In the case of the dorsal telencephalon neurons, the main factor driving cells to spike threshold is a large, slow depolarization, whereas the oscillatory synaptic drive governs the timing of spikes. This is very similar to intracellular recordings of Kenyon cell odor responses: large depolarizations are observed, with high-frequency (although non-periodic) fluctuations riding on top. As such, it seems likely that the main factor driving Kenyon cells to spike threshold in Drosophila, similar to dorsal telencephalon neurons, is the slow wave of depolarization that arises from the summation of projection neuron input from several dendritic claws (Gruntman, 2013).

The linear and sublinear additivity of claws is likely an important factor contributing to the sparsity of Kenyon cell spikes. When examining synaptic responses, only a few cases were found in which projection neuron stimulation evoked a response that was consistently above spike threshold in the postsynaptic Kenyon cell. These Kenyon cells were connected via several claws: three claws for one cell, five for the other. To characterize the relationship between connectivity and spiking, an extensive series of recordings was carried out in which Kenyon cells that spike in response to photostimulation were searched for specifically. A range of connectivity was found in the set of responding cells. Comparison was made across these different cells to examine the effects of increasing connectivity on spiking characteristics of Kenyon cells. Surprisingly, no correlation was found between the number of contacted claws and the magnitude of the spiking response. Kenyon cells tended to either respond or not, similar to the odor-evoked responses of these cells, which also do not span a wide range of spike rates. However, when the likelihood of Kenyon cell responses was examined as a function of the number connected claws, a strong relationship was observed. This analysis revealed that there was a marked increase in the proportion of responding cells receiving input on four or more claws. Kenyon cells typically have five to seven claws, so this suggests that the majority of a Kenyon cell's claws need to be coactive to evoke an odor-like response (Gruntman, 2013).

This requirement was not absolute, however, as cells with reliable spiking responses with lower levels of connectivity were found, and even a very small number of Kenyon cells were found that spiked when connected via only a single claw (2 of 191 total recorded Kenyon cells). Although it is possible that some Kenyon cells require activation of all their claws, it seems likely that most Kenyon cells require only a subset of their inputs to be active to spike. As there are many different projection neuron input patterns that could activate a subset of the Kenyon cells claws, these results indicate that Kenyon cells encode input patterns in a degenerate manner and that several different input patterns could effectively drive the cell. This is consistent with the odor response properties of Kenyon cells; although these cells are much more odor-selective than projection neurons, they do occasionally respond to multiple odors. Thus, although Kenyon cells' requirement for multiple coactive inputs certainly contributes to their stimulus selectivity, that selectivity is not absolute. Moreover, inhibitory circuit elements could potentially be important for controlling this selectivity, a possibility that was not addressed in this study. Together, these considerations indicate that Kenyon cells are likely to be degenerate decoders that respond to several related projection neuron response patterns (Gruntman, 2013).

It was found that Kenyon cells derive their highly specific response properties by integrating over diverse input channels, effectively reading the combinatorial odor code. Why would such selective responses be constructed from synaptic inputs with widely divergent tuning? One possibility is that it gives the mushroom body the flexibility to support olfactory learning. Given that each Kenyon cell is connected to inputs with very different odor tuning, adjusting the synaptic strength of one of the inputs could markedly alter the cell's response properties. In addition, this convergent connectivity is likely important for diversifying the odor response properties of the Kenyon cells relative to the projection neurons. This diversification could enable the network to transition from broadly tuned projection neurons to narrowly tuned Kenyon cells while maintaining the capacity to represent many different odors (Gruntman, 2013).

Go to top

Semaphorin 1a-mediated dendritic wiring of the Drosophila mushroom body extrinsic neurons

Lin, C. H., Senapati, B., Chen, W. J., Bansal, S. and Lin, S. (2022). Proc Natl Acad Sci U S A 119(12): e2111283119. PubMed ID: 35286204

=

The adult Drosophila mushroom body (MB) is one of the most extensively studied neural circuits. However, how its circuit organization is established during development is unclear. This study provides an initial characterization of the assembly process of the extrinsic neurons (dopaminergic neurons and MB output neurons) that target the vertical MB lobes. The cellular mechanisms guiding the neurite targeting of these extrinsic neurons were probed, and it was demonstrate that Semaphorin 1a is required in several MB output neurons for their dendritic innervations to three specific MB lobe zones. This study reveals several intriguing molecular and cellular principles governing assembly of the MB circuit (Lin, 2022).

The MB is one of the most intensively studied structures in the fly brain. Its complex and organized circuit architecture has provided important clues to its operational logic. However, in contrast to the extensive investigations of its functions, how the MB circuit architecture is established during development has been little explored. This study provides an initial characterization of MB circuit assembly and identifies Sema1a as an important guidance molecule that directs dendritic innervations of multiple MBONs in three MB lobe zones. Below, several implications of this study relating to the wiring principles of the MB circuit are discussed, and a hypothetical model for how DAN axons and MBON dendrites are modularly assembled into the MB lobes is presented (Lin, 2022).

The most intriguing feature of the organization of the MB circuit is the zonal innervation of the MB lobes by DAN axons and MBON dendrites. The borders of the zones are distinct, with minimal overlap between DAN axons or MBON dendrites in the neighboring zones. Given such a highly organized neural network, elaborate interactions among the extrinsic neurons might be expected. For example, dendritic tiling, as observed between dendritic arborization (da) neurons in fly larvae, might be required for the formation of zonal borders between MBON dendrites, and match-ups between DAN axons and MBON dendrites in the same zone might be important for these neurites to establish proper zonal innervation patterns. However, the results suggest that the targeting and elaboration networks of DAN axons and MBON dendrites are largely independent, at least for those projecting to the MB vertical lobes. In the α'2 zone, where innervation by DAN axons precedes that by MBON dendrites, ablation of DANs does not affect zonal elaboration of the MBON dendrites. Moreover, upon ablation of one type of DAN or MBON in a given zone, morphologies of the neighboring neurites appear to be normal. Therefore, the extent and location of zonal network elaboration by DAN axons and MBON dendrites in the vertical lobes do not depend on interactions between these extrinsic neurons (Lin, 2022).

In each MB lobe zone, the DAN axons and MBON dendrites form synapses with each other and the KC axons. Given that DANs and MBONs do not depend on each other to form zonal networks, could KCs be responsible? The results support the importance of the KCs in the zonal organization of the DAN axons and MBON dendrites. Aberrant branching of KCs in alpha lobes absent (ala) mutant brains resulted in some MBs lacking the vertical lobes. When this occurred, most DAN axons and MBON dendrites that normally innervate these lobes do not form zonal arborization. Without the MB vertical lobes, MBON-α'2 dendrites are rerouted to other zones in the horizontal lobes and form potential synaptic connections with the local DAN axons. Importantly, this reorganization of the MBON dendrites requires the presence of the KCs (Lin, 2022).

It is still possible that the KC axons and the lobes they form simply provide an anchoring point on which DAN axons and MBON dendrites grow, and that the extent of their arborization is determined cell-autonomously as an intrinsic property. However, the axonal innervation pattern of PPL1 DANs argues against this possibility. PPL1-α'3 and PPL1-α'2α2 axons enter the MB vertical lobes at almost the same location but specifically occupy distinct zones on opposite sides of the entry point, suggesting the existence of local positional cues in the lobes to guide the innervation of DAN axons. Furthermore, overexpression of sema1a in DANs directs their dendrites to specific MB lobe zones, and importantly, the arborization of these rerouted dendrites is confined to their respective zones. Since these DAN dendrites normally do not innervate the MB lobe, there likely exist local positional cues that interact with Sema1a-expressing dendrites to guide their zonal arborization (Lin, 2022).

What could be the sources of these positional cues? KCs are good candidates because they synapse with DAN axons and MBON dendrites and are essential for zonal arborization of these neurites. However, since KCs provide the main framework of the MB lobes, their manipulation may affect the organization of other cell types in the MB lobes that could also be potential sources of the positional cues. Electron microscopy-based reconstructions of the MB circuit have provided a comprehensive catalog for the neurons that innervate the MB lobe. In addition to KCs, DANs, and MBONs, the MB lobes are innervated by one dorsal paired medial (DPM) neuron, one anterior paired lateral (APL) neuron, two SIFamide-expressing neurons, and two octopaminergic neurons. These neurons do not exhibit zonal innervation patterns in the MB lobes; the SIFamide- and octopamine-expressing neurons only sparsely innervate the MB lobes, and the neurites of APL and DPM ramify the entire MB lobes. The MB lobes are also populated by glia, which is another potential source of the positional cues. Although their sources remain undetermined, the results suggest that the MB lobes are likely prepatterned with positional cues to guide the zonal elaboration of the MBON and DAN neurites (Lin, 2022).

Sema1a is an evolutionarily conserved guidance molecule that functions as a ligand or receptor depending on the cellular context. The data suggest that Sema1a functions as a receptor in MBONs to regulate dendritic targeting in a zone-specific manner. Loss of sema1a activity preferentially affects MBON dendrites in the α'3, α'1, and β'2 zones. Even for MBONs that innervate multiple zones, such as MBON-β2β'2a and MBON-γ5β'2a, reducing sema1a activity in these neurons selectively impacts their dendrites in the β'2a zone. Since Sema1a is not differentially localized in the dendrites of these MBONs, the guidance cues that Sema1a responds to might primarily be present in the β'2a zone, with additional guidance signals working collaboratively to sort the dendrites from these MBONs into Sema1a-sensitive and -insensitive zones. Not all MBONs innervating these Sema1a-sensitive zones require Sema1a. For MBON-γ2α'1 and MBON-α'1 that both innervate the α'1 zone, loss of sema1a only affects dendritic innervation by MBON-α'1 but not MBON-γ2α'1. Therefore, how Sema1a functions is also cell-type-specific. Taken together, these results imply that multiple positional cues may be present in each MB lobe zone, with each MBON being equipped with multiple sensors that work in concert to respond to those cues. Moreover, given that Sema1a is broadly expressed in many neurons in the developing brain, including the KCs, Sema1a likely acts with other proteins or signaling molecules to determine guidance specificity (Lin, 2022).

Sema1a has been shown to mediate both neurite attraction and avoidance. Currently, it is unclear which of the two mechanisms underlies its guidance of MBON dendrites. For MBONs whose zonal dendritic innervation requires sema1a, their dendrites can still project to areas nearby their target zones when sema1a activity was removed. Hence, the cues that guide these MBON dendrites are likely to be short-ranged. However, overexpression of sema1a in PPL1-α'2α2 DANs can redirect their dendrites to innervate zones far away from their original location, suggesting that the guidance cues may also exert long-distance functionality. The data indicate that Sema1a functions as a receptor in MBONs. Identification of Sema1a ligands and determining their distributions in the MB lobes are critical steps toward understanding how Sema1a instructs the zonal innervation of MBON dendrites. Plexin A (PlexA) or secreted Semaphorin 2a and 2b (Sema2a and Sema2b) are known ligands for the Sema1a receptor. This study has tested if these canonical ligands of Sema1a are required for the dendritic innervations of β'2- and α'3-projecting MBONs. However, dendritic innervations by these MBONs were minimally affected in homozygous sema2a/2b double mutant flies or when PlexA was knocked down either pan-neuronally or in glia. Therefore, the canonical Sema1a ligands do not seem to play a role in MBON dendritic targeting. However, it remains to be determined if PlexA and Sema2a/2b function redundantly in this system or if an unidentified noncanonical ligand is involved (Lin, 2022).

Although the molecular nature of the positional cues in the MB lobes that organize the zonal patterns of DAN and MBON neurites awaits discovery, the data suggest that these cues likely work in a combinatorial manner. Supporting evidence for this notion comes from the observation that MBON-α'1 and MBON-γ2α'1 use sema1a-dependent and -independent mechanisms to innervate the α'1 zone, indicating that this zone may present at least two different guidance cues. Furthermore, sema1a is expressed in multiple MBONs that innervate distinct MB lobe zones. This pattern could potentially be explained if the positional cues attracting Sema1a-positive neurites appear sequentially in these zones (i.e., so that the zone an MBON innervates is determined by the developmental timing of the MBON). However, the finding that the ectopic innervations of MBON-α'2 in the β'2 and α'1-like zones of the ala brain occur simultaneously argues against that possibility. Therefore, the Sema1a-sensitive zones likely harbor additional zone-specific guidance cues that work in combination with Sema1a to diversify guidance specificity (Lin, 2022).

The observation that the mistargeted MBON-α'2 dendrites in ala mutant brains innervate other zones in the α'β' lobe, but not those in the αβ and γ lobes, has also prompted a hypothesis that there might be general attraction cues emanating from the α'β' lobes for all α'β' lobe-projecting MBONs, separating them from MBONs targeting αβ and γ lobes. Therefore, a hypothetical model is proposed whereby multiple hierarchically-organized positional cues are presented in the MB lobe zones, with these cues acting in concert to pattern zonal innervation by DAN axons and MBON dendrites in the MB lobes (Lin, 2022).

Go to top

Presynaptic inhibition of γ lobe neurons is required for olfactory learning in Drosophila

Zhang, J., Tanenhaus, A. K., Davis, J. C., Hanlon, B. M. and Yin, J. C. (2014). Neurobiol Learn Mem 118C: 80-88. PubMed ID: 25460038

The loss of heterotrimeric Go signaling through the expression of pertussis toxin (PTX) within either the α/β or γ lobe mushroom body neurons of Drosophila results in the impaired aversive olfactory associative memory formation. This study focused on the cellular effects of Go signaling in the γ lobe mushroom body neurons during memory formation. Expression of PTX in the γ lobes specifically inhibits Go activation, leading to poor olfactory learning and an increase in odor-elicited synaptic vesicle release. In the γ lobe neurons, training decreases synaptic vesicle release elicited by the unpaired conditioned stimulus minus, while leaving presynaptic activation by the paired conditioned stimulus plus unchanged. PTX expression in γ lobe neurons inhibits the generation of this differential synaptic activation by conditioned stimuli after negative reinforcement. Hyperpolarization of the γ lobe neurons or the inhibition of presynaptic activity through the expression of dominant negative dynamin transgenes ameliorated the memory impairment caused by PTX, indicating that the disinhibition of these neurons by PTX was responsible for the poor memory formation. The role for γ lobe inhibition, carried out by Go activation, indicates that an inhibitory circuit involving these neurons plays a positive role in memory acquisition. This newly uncovered requirement for inhibition of odor-elicited activity within the γ lobes is consistent with these neurons serving as comparators during learning, perhaps as part of an odor salience modification mechanism (Zhang, 2013).

Activation of Go is inhibited by the expression of PTX. The inhibition of Go activation by the expression of PTX within the γ lobe neurons of the mushroom bodies leads to a significant decrease in short term aversive memories. This study has now shown that this memory defect is caused by the disinhibition of the γ lobe neurons. Inhibition of Go increases odor-induced presynaptic activity. The inhibition of γ lobe neurons by either hyperpolarization or synaptic vesicle depletion reverses the PTX learning phenotype. It is proposed that the Go-mediated presynaptic inhibition of γ lobe neurons is required to generate differential conditioned stimulus salience during discriminative leaning (Zhang, 2013).

Go to top

Activity-dependent plasticity in an olfactory circuit

Sachse, S., Rueckert, E., Keller, A., Okada, R., Tanaka, N. K., Ito, K. and Vosshall, L. B. (2007). Neuron 56: 838-850. PubMed ID: 18054860

Olfactory sensory neurons (OSNs) form synapses with local interneurons and second-order projection neurons to form stereotyped olfactory glomeruli. This primary olfactory circuit is hard-wired through the action of genetic cues. It was asked whether individual glomeruli have the capacity for stimulus-evoked plasticity by focusing on the carbon dioxide (CO2) circuit in Drosophila. Specialized OSNs detect this gas and relay the information to a dedicated circuit in the brain. Prolonged exposure to CO2 induced a reversible volume increase in the CO2-specific glomerulus. OSNs showed neither altered morphology nor function after chronic exposure, but one class of inhibitory local interneurons showed significantly increased responses to CO2. Two-photon imaging of the axon terminals of a single PN innervating the CO2 glomerulus showed significantly decreased functional output following CO2 exposure. Behavioral responses to CO2 were also reduced after such exposure. It is suggested that activity-dependent functional plasticity may be a general feature of the Drosophila olfactory system (Sachse, 2007).

Neuroanatomical, functional, and behavioral analysis suggests that the Drosophila olfactory system has the capacity for reversible activity-dependent plasticity. Evidence of this plasticity is readily seen by measuring the volume of the V glomerulus. Because the volume increase can be induced by odor activation of ORs ectopically expressed in the CO2-activated OSNs, it is concluded that persistent stimulus-evoked activity in these neurons underlies these anatomical changes. It has been shown that stimulus-evoked plasticity is a general feature of the Drosophila olfactory system and not a peculiarity of the CO2 circuit. For instance, the volume of DM2 is increased by chronic exposure to ethyl butyrate, a ligand for the Or22a-expressing neurons that target DM2 (Sachse, 2007).

Drosophila, CO2 is detected by a population of approximately 25-30 OSNs in the antenna that express the chemosensory receptor Gr21a, which along with Gr63a comprises the Drosophila CO2 receptor. These OSNs project axons that terminate in the V glomerulus in the ventral antennal lobe. The Drosophila CO2 circuit is ideal for studying odor-evoked plasticity because Gr21a-expressing OSNs are the only neurons in the fly that respond to CO2, and they do not respond to any other stimuli. In this work, stimulus-evoked changes in the anatomy and function were examined of the Drosophila CO2 circuit. The results provide functional evidence that a primary olfactory center is capable of activity-dependent plasticity (Sachse, 2007).

The data are consistent with a model in which one class of inhibitory LNs and the output of the V glomerulus are the major targets of plasticity induced by sensory exposure. Under conditions of ambient CO2, the Gr21a circuit forms normally and small amounts of CO2 produce robust behavioral responses. When flies are exposed to elevated CO2 early in life, it is postulated that chronic activation of Gr21a neurons promotes functional changes in the LN2 subtype of inhibitory local interneurons without affecting either the functional properties of the OSNs or the CO2-evoked response of the LN1 neurons. It is suggested that the volume increases seen with CO2 exposure may result from neuroanatomical changes in the LNs, although their extensive glomerular arborization made this hypothesis difficult to test experimentally. Since a majority of the LN2 population in Drosophila has been shown to be GAD1 positive and thus to release GABA, as known for antennal lobe LNs in other insects, greater CO2-evoked activity of LN2s may lead to an increased inhibition of the PN postsynaptic to Gr21a OSNs. The finding of reduced activity in the output region of the PN innervating the V glomerulus supports this hypothesis. Thus, CO2-evoked activity would be attenuated in the antennal lobe circuit in these animals, producing a corresponding decrease in the intensity of the behavioral response (Sachse, 2007).

It has recently been shown that LNs are not only inhibitory, as has been assumed so far. A newly described population of excitatory cholinergic LNs forms a dense network of lateral excitatory connections between different glomeruli that may boost antennal lobe output (Olsen, 2007; Shang, 2007). Future studies are necessary to investigate if excitatory LNs are also subject to activity-dependent plasticity (Sachse, 2007).

Stimulus-dependent plasticity can be induced and reversed in a critical period early in the life of a fly. Similar critical periods have been documented in selective deafferentation periods in mammalian somatosensory and visual cortex. In all these model systems, the critical period likely allows the animal to compare the genetically determined network template with external conditions and make activity-dependent adjustments that reflect the external environment. For instance, visual cortex 'expects' binocular input when it is wired in utero. If monocular input is experimentally imposed, the system is rewired to reflect this. The same rewiring occurs in the barrel cortex, in which the receptive fields of missing whiskers are invaded by neighboring whiskers, allowing the animal to maintain a continuous representation of external somatosensory space. Drosophila pupae have no sensory input during development and develop an olfactory system that relies neither on evoked activity nor the expression of ORs. The time following adult eclosion may represent a period in which the functional set point of the Drosophila olfactory system is evaluated and adapted to the local environment (Sachse, 2007).

What elements of the antennal lobe circuit are responsible for the stimulus-dependent volume increases seen here? No evidence was found that OSNs modulate their number, morphology, branching pattern, or functional properties in response to CO2 exposure. The same neuroanatomical properties of single LNs or PNs could not be assayed due to the dense processes of these neurons in a given glomerulus. Since the observed net increase in volume cannot be ascribed to anatomical changes in OSNs, morphological plasticity is most likely occurring either at the level of LN or PN. A model is favored in which changes in the LNs underlie the observed volume increases because clear functional differences were found in LN2 responsivity in CO2-exposed animals and because PN dendrites and axons have been shown to be extremely stable in size and morphology when deprived of OSN input. Similar stability in mitral/tufted cells has been shown in rodent olfactory bulb. The possibility that other cells, such as glia, contribute to these activity-dependent volume changes cannot be excluded (Sachse, 2007).

This work suggests that antennal lobe LNs marked with two different Gal4 enhancer traps, Gal4-LN1 and Gal4-LN2, are functionally distinct. The arborization of LN1 and LN2 processes in the V glomerulus suggests that they interact differentially with the antennal lobe circuitry. LN1 processes appear to innervate the core of a given glomerulus, while LN2 processes innervate the glomerulus more uniformly. Both LN1 and LN2 neurons show weakly concentration-dependent tuning to odor stimuli. Thus, compared to the OSNs or PNs, which transmit a precise spike-timing code that reflects absolute CO2 concentration, these LNs appear to respond in a binary fashion, showing similar levels of activity regardless of stimulus concentration (Sachse, 2007).

There is a clear difference in how the responses of these two LN populations are modulated by CO2 exposure. While the activity of LN1 neurons was not significantly affected by CO2 exposure, LN2 neurons exhibited robust and significant increases in CO2-evoked activity after CO2 exposure. It will be of interest to examine the functional properties of these neurons in greater detail using electrophysiological approaches. It is plausible that circuit plasticity as evidenced in the LN2 neurons can be detected with electrophysiology at even lower CO2 concentrations for shorter exposure periods (Sachse, 2007).

How might chronic activation of CO2-sensitive OSNs specifically affect the physiology of LN2 neurons? It is speculated that due to the broader innervation of LN2 processes, these neurons would receive greater presynaptic innervation from Gr21a-expressing OSNs. Thus, with chronic CO2 exposure, the LN2 neurons would be chronically activated. This might cause long-term plasticity leading to greater GABA release from LN2 neurons. In cerebellar stellate cells, such an increase in inhibitory transmitter release has been documented and coined 'inhibitory-long term potentiation' (I-LTP). I-LTP is induced in stellate cells by glutamate released from parallel fibers acting on presynaptic NMDA receptors in these inhibitory interneurons and producing a long-lasting increase in the release of GABA from these cells. Like stellate neurons, at least one population of Drosophila LNs is pharmacologically GABAergic (Sachse, 2007).

How might alterations in LN2 pharmacology affect downstream circuit elements and ultimately CO2-evoked behavior? Drawing on the same cerebellar analogy discussed above, it is plausible that PNs exhibit a type of 'rebound potentiation' that has been observed in Purkinje cells responding to inhibitory input. GABA released from LNs would regulate the excitability of PNs, such that greater GABA release from LN2 would tend to decrease the excitability of CO2-specific PNs. The finding that the output from the V glomerulus to the lateral horn is reduced following CO2 exposure supports the idea that downstream activity in higher processing centers is modulated by the antennal lobe network. However, it still needs to be shown that LN2 neurons form direct inhibitory synapses onto PNs in the V glomerulus. Reduced PN activity in the lateral horn in turn may produce a reduced behavioral sensitivity to this stimulus. Future experiments that examine this stimulus-dependent plasticity at the cellular level using pharmacology and electrophysiology will be necessary to test this model (Sachse, 2007).

Go to top

AKAPs act in a two-step mechanism of memory acquisition

Scheunemann, L., Skroblin, P., Hundsrucker, C., Klussmann, E., Efetova, M. and Schwarzel, M. (2013). J Neurosci 33: 17422-17428. PubMed ID: 24174675

Defining the molecular and neuronal basis of associative memories is based upon behavioral preparations that yield high performance due to selection of salient stimuli, strong reinforcement, and repeated conditioning trials. One of those preparations is the Drosophila aversive olfactory conditioning procedure where animals initiate multiple memory components after experience of a single cycle training procedure. This study explored the analysis of acquisition dynamics as a means to define memory components and revealed strong correlations between particular chronologies of shock impact and number experienced during the associative training situation and subsequent performance of conditioned avoidance. Analyzing acquisition dynamics in Drosophila memory mutants revealed that rutabaga (rut)-dependent cAMP signals couple in a divergent fashion for support of different memory components. In case of anesthesia-sensitive memory (ASM) this study identified a characteristic two-step mechanism that links rut-AC1 to A-kinase anchoring proteins (AKAP)-sequestered protein kinase A at the level of Kenyon cells, a recognized center of olfactory learning within the fly brain. It is proposed that integration of rut-derived cAMP signals at level of AKAPs might serve as counting register that accounts for the two-step mechanism of ASM acquisition (Scheunemann, 2013).

Conditioned odor avoidance is subject to a general dichotomy since multiple memory components are engaged in control of behavior. This is usually analyzed at two time points, i.e., 3 min and 3 h after training. At 3 min, basal and dynamic STM are separable by genetic means as revealed by opposing phenotypes of rut1 and dnc1 mutants. Moreover, those components are also separable due to characteristic differences in their acquisition dynamics as revealed by different effects of shock number. A similar dichotomy applied to 3 h memory when ASM and ARM were separable by means of amnestic treatment. It was striking that basal STM and consolidated ARM were instantaneously acquired, resulting in a front line of protection by eliciting conditioned avoidance after a singular experience of a CS/US pairing. Interestingly, consolidated ARM (as defined by means of resisting amnestic treatment) was installed quickly after training. However, it remains to be addressed at the genetic and molecular level, whether 3 h ARM linearly results from the functionally similar basal SMT component (Scheunemann, 2013).

In contrast, the two components of dynamic STM and labile ASM acquire in a dynamic fashion but are clearly dissociated from each other by characteristic chronologies of CS/US pairings required for their acquisition. However, either component contributes to behavioral performance in addition to the appropriate instantaneous component, and hence, increases avoidance probability during the test situation. Considering a potential benefit from avoiding aversive situations this overall dichotomy of behavioral control seems plausible and is also reflected at the genetic level since rut-dependent cAMP signals are limited to support of dynamic but not instantaneous memory components. Rut-dependent STM and ASM, however, are dissociated by means of shock impact and discontinuous formation of ASM is limited to situations where animals repeatedly experience high-impact CS/US pairings within a predefined time window. Experience that does not meet this criterion, however, is not discounted but adds to continuously acquired dynamic STM. By functional means these two components are thus clearly separated but commonly dependent on rut-derived cAMP signals within the KC layer, forming ties between genetically and functionally defined memory components (Scheunemann, 2013).

Genetic dissection of memory has a long-standing history in Drosophila and provided a powerful means to define molecular, cellular, and neuronal networks involved in regulation of conditioned odor avoidance. Among others, the cAMP-signaling cascade has been identified as one central tenet of aversive odor memory foremost by means of two single-gene mutants affecting either a Ca2+-sensitive type 1 adenylyl cyclase (AC1) and/or a cAMP-specific type 4 phosphodiesterase (PDE4) affected in the Drosophila learning mutants rutabaga (rut-AC1) and dunce (dnc-PDE4). Although originally isolated due to poor performance in the aversive odor learning paradigm, a general dichotomy has been recently revealed that separates memory components by their dependency on either rut-AC1 or dnc-PDE4 function, and the view was established that two different types of cAMP signals are engaged during the single-cycle training procedure (Scheunemann, 2012). A similar dichotomy is observed at level of acquisition dynamics and suggests that rut-dependent cAMP signals are limited to formation of dynamically acquired memory components, i.e., dynamic STM and ASM. Interestingly, rut-dependent cAMP is also required for long-term memory (LTM), which acquires after spaced and repeated training sessions. Downstream the signaling cascade, however, appropriate cAMP signals are differently channeled to either support LTM in a CREB-dependent manner, ASM via tomosyn-dependent plasticity, or basal STM via synapsin-dependent regulation of synaptic efficacy. It appears that the chronology of CS/US pairings is an important determinant of which downstream effect is triggered and hence molecular mechanisms must be installed that are sensitive to the temporal order of training (Scheunemann, 2013).

At the level of molecular scaffolds, literature suggests that AKAPs serve the integration of cAMP with other signaling processes and are crucially involved in the control of a plethora of cellular functions in any organ. For example, AKAP79 coordinates cAMP and Ca2+ signaling in neurons to control ion channel activities. The recognized design principle of AKAPs to serve as molecular switch is well in line with the recognized two-step register mechanism involved in ASM formation. An increasing body of evidence shows that AKAPs are involved in memory processing across phyla and accordantly those studies revealed a contribution for support of matured, but not immediate memories. Communality among all those AKAP-dependent memories is the need for repeated and temporally organized training sessions, i.e., only spaced training sessions are effective to induce protein synthesis-dependent LTM in flies and mammals. Similarly, ASM requires the precise timing of two training sessions and mechanistically acts via an 'activated' state generated by the initial CS/US pairing that persists within the brain for ~5 min. Such temporal integration might well take place at level of AKAPs within the KC layer to operate rut-AC1-dependent cAMP signals finally onto phosphorylation of tomosyn. Identification of the particular AKAPs involved in two-step ASM formation will require further analysis of appropriate mutants. To date, only four Drosophila AKAPs are characterized, i.e., rugose, a 550 kDa protein that impacts on STM performance probably via molecular domains other than its AKAP function; yu/spoonbill that supports LTM; and Nervy and AKAP200 have not been tested for their impact on aversive odor memory (Scheunemann, 2013).

Together, the benchmarking of Drosophila aversive odor memory performance by means of acquisition dynamics that were demonstrated in this study will provide a valuable tool since dynamic aspects of acquisition are obviously informative and add to the steady-state condition determined by the single-cycle training procedure (Scheunemann, 2013).

Go to top

Hebbian STDP in mushroom bodies facilitates the synchronous flow of olfactory information in locusts

Cassenaer, S. and Laurent, G. (2007). Nature 448: 709-713. PubMed ID: 17581587

Odour representations in insects undergo progressive transformations and decorrelation from the receptor array to the presumed site of odour learning, the mushroom body. There, odours are represented by sparse assemblies of Kenyon cells in a large population. Using intracellular recordings in vivo, this study examined transmission and plasticity at the synapse made by Kenyon cells onto downstream targets in locusts. It was found that these individual synapses are excitatory and undergo hebbian spike-timing dependent plasticity (STDP) on a +-25 ms timescale. STDP is a phenomenon in which the gain of a connection can be changed according to the temporal relationship between pre- and post-synaptic spikes. When placed in the context of odour-evoked Kenyon cell activity (a 20-Hz oscillatory population discharge), this form of STDP enhances the synchronization of the Kenyon cells' targets and thus helps preserve the propagation of the odour-specific codes through the olfactory system (Cassenaer, 2007).

Olfactory processing in insects begins in an array of receptor neurons that express collectively many tens of olfactory receptor genes (~60 in Drosophila; ~150 in honeybees). The representations of general odours are then decorrelated by local circuits of projection neurons and local neurons in the antennal lobe. In locusts and other insects, the antennal lobe output is distributed in space and time and can be described as stimulus-specific time-series of projection-neuron activity vectors, updated at each cycle of a 20-Hz collective oscillation. Distributed projection-neuron activity is then projected to Kenyon cells, the intrinsic neurons of the mushroom body. In contrast to projection neurons, Kenyon cells respond very specifically and fire extremely rarely. The mechanisms underlying this sparsening are starting to be understood. Such sparse representations are advantageous for memory and recall, consistent with established roles of the mushroom bodies in learning. In Drosophila, experiments combining molecular inactivation with behaviour indicate that synaptic output from Kenyon cells in the lobes is required for memory retrieval. Little is known, however, about the electrophysiological properties of these synapses (Cassenaer, 2007).

The connections made by Kenyon cells onto a small population of extrinsic neurons have been studied in the β-lobe of the locust mushroom body, using an intact, in vivo preparation. β-lobe neurons (β-LNs) respond to odours; their responses are odour-specific and their tuning is sensitive to input synchrony. This study recorded intracellularly from pairs of Kenyon cells and β-LNs: randomly selected Kenyon cells were impaled in their soma; β-LNs were impaled in a dendrite in the β-lobe. Focused was placed on one β-LN anatomical subtype, which comprises many individual neurons. Neurons of this subtype, called β-LNs here, could be recognized also by their physiological characteristics. Each β-LN has extensive dendrites that intersect many of 50,000 Kenyon cell axons. Monosynaptic connections were found in ~2% of tested Kenyon cell (KC)-β-LN pairs. All were excitatory. The delay between Kenyon cell spike and β-LN-excitatory post-synaptic potential (EPSP) onset was 6.5 +- 0.70 ms, including 5.4 +- 0.25 ms for spike propagation from Kenyon cell soma to the β-lobe. The remaining (synaptic) delay (~1 ms) is similar to that at another chemical synapse in the locust brain. Unitary EPSPs were large (1.58 mV +- 1.11), in contrast to those generated in Kenyon cells by individual projection neurons (86 microV +- 44). The fact that Kenyon cell outputs are powerful is consistent with Kenyon cell spikes being rare and therefore highly informative. EPSP amplitude varied greatly across connected pairs (0.55-4 mV). This could reflect a distribution of electrotonic distances between synapses and recording sites. Simultaneous impalements of different dendrites in the same β-LN, however, show that the amplitudes of most events were the same across recording sites. Consistent with this, unitary EPSP kinetics (10-90% rise time, 8.3 ms +- 2.3; time to 1-(1/e) of peak, 13.2 ms +- 4.4) were independent of the β-LN recorded and, thus, of the impalement site. Simultaneous dendritic recordings of different β-LNs, however, revealed that their synaptic backgrounds overlapped only partly. Common EPSPs rarely had the same amplitude. Hence, β-LNs may each receive inputs from hundreds to thousands (~2% of 50,000 Kenyon cells) of Kenyon cells, in overlapping subsets; KC-β-LN connections are strong on average, with target-specific strength (Cassenaer, 2007).

Odour-evoked activity in projection neurons and Kenyon cells consists principally of sequential volleys of synchronized spikes-generally, one spike per responding neuron per oscillation cycle. β-LN responses to odours also consisted typically of sequences of single phase-locked spikes, timed around the trough of several local field potential (LFP) oscillation cycles. The cycles when a spike was produced (usually with probability <1) depended on β-LN and stimulus identity. It is concluded that, to each oscillation cycle corresponds a particular activity vector in the projection neuron, Kenyon cell and β-LN populations. By recording from pairs of β-LNs simultaneously during odour trials, it was also observed that, when the two β-LNs fired one action potential during the same oscillation cycle, those action potentials were tightly synchronized (+-2 ms) (Cassenaer, 2007).

A fortuitous observation provided hints of plasticity at the KC-β-LN synapse. At trial 4 of a Kenyon cell stimulus sequence intended to explore β-LN integration, the β-LN fired a spontaneous action potential roughly at the time of the first (of 2) Kenyon-cell-evoked EPSP. At trial 5, 10 seconds after this single fortuitous pairing, the first EPSP of the pair was greatly enhanced. This suggested the possibility of spike-timing-dependent plasticity (STDP), a phenomenon thus far unknown in invertebrates but well characterized in vertebrates, in which the gain of a connection can be changed according to the temporal relationship between pre- and post-synaptic spikes. The consequence of pre-post temporal relationships was explored on the KC-β-LN synapse. A β-LN was impaled and stimulated alternately by two independent Kenyon cell pathways-one for pairing, one for unpaired control. Each stimulus was repeated every 10 s, with a 5-s delay between pairing and control stimuli. Pairing consisted of a single Kenyon cell (pre) stimulus and a 5-ms supra-threshold β-LN (post) current pulse, timed such that the delay (dt = tpost-tpre) between pre- and post-synaptic spikes varied between -60 and +50 ms. Test trials, used to measure connection strength before and after pairing, were identical to the pairing trials in all respects except in the temporal relationship between pre- and post-synaptic spike times (2.5 s apart). Two examples (for dt = 10 ms and -4 ms, 25 pairings each) are given. For dt = 10 ms, the paired input underwent potentiation; for dt = -4 ms, it underwent depression. For both conditions, the control pathway (same β-LN, different Kenyon cell input) remained unchanged. The changes were thus input-specific; they were often detectable after a single pairing, and could be maintained for up to 25 min. 26 values of dt between -60 and +50 ms were tested. The resulting changes define a classical hebbian profile: the synapse is potentiated when pre- precedes post-, and depressed when post- precedes pre-, with symmetrical profiles. The changes could be fitted well with two exponential decays flanking a narrow linear range around t = +4 ms. Several connections were tested successively with two (or more) values of dt (some positive, others negative): the same connections could undergo both depression and potentiation, depending on the value of dt. The STDP profile thus seems to be a property of each connection and not only a collective one (Cassenaer, 2007).

It was observed that the values of dt over which synaptic weights change correspond to the period of single odour-evoked oscillation cycles; hence, only within-cycle 'coincidences' may modify the connections between a Kenyon cell and its targets. The features of the STDP curve, when considered together with the timing of Kenyon cells and β-LNs during odour-evoked activity, have interesting consequences. Consider the phases of Kenyon cell and β-LN spikes. Owing to propagation delays, Kenyon cell spikes reach their targets just before the trough of the LFP, a little before β-LN firing. Consider a cycle in which a β-LN spikes early: some KC-β-LN connections will undergo depression; at the next trial, β-LN spike time at this cycle should be delayed. If, in contrast, a β-LN spikes late, STDP should potentiate Kenyon cell drive to it, and thus advance spike time for that cycle. In short, the cycle-by-cycle action of STDP suggests adaptive control of β-LN spike phase. The need for such regulation is not unique to this system: models of cortical networks indicate that, as activity propagates through successive 'layers', accumulating noise can rapidly smear the temporal structure that may exist. Modelling studies predict that STDP, given appropriate parameters, could preserve the temporal discretization of activity through such layers (Cassenaer, 2007).

A reduced model was generated of the KC-β-LN circuit, and the STDP rule derived from these experiments was introduced. To control the relative phases of Kenyon cells and β-LNs, Kenyon cell spike phases were drawn from experiments and input weights from uniform distributions with different means: with low weights, β-LN spikes tended to occur late (dt > 0); with larger weights, they occurred early (dt < 0). After several trials (each with a random draw of inputs from the same distribution), STDP was allowed to modify synaptic weights for the following trials: when β-LN spikes occurred late (dt > 0), Kenyon cell outputs became potentiated and β-LN spikes were advanced; for dt < 0, time shifts were inverted. The histograms shown represent spike-time distributions for 1,000 trials before and after STDP, for each of three conditions. These simulations were repeated 200 times (50 trials each), with 11 different Kenyon cell input distributions. Once STDP was turned on (trial 1), the evolution was systematic and rapid, leading to the adaptive up- or downregulation of input weights, firing phase and response intensity. Given that the model is entirely constrained by experiments, it is noteworthy that the mean phase of the first β-LN spike at steady state, matches precisely that measured experimentally (Cassenaer, 2007).

To test directly the effect of STDP on β-LN output, β-LN spike timing was manipulated during responses to odours in vivo: if the model is correct, such manipulations should change the output of the odour-activated Kenyon cells onto that β-LN and, thus, generate predictable shifts in its spike phase. During odour stimuli, short current pulses locked to selected cycles of the LFP were injected in a β-LN: a negative pulse was injected during the cycles and phase when the β-LN would naturally fire (to prevent stimulus-evoked spikes), and a positive pulse was injected at a desired phase, for those same cycles (that is, at an abnormal time relative to the Kenyon cell inputs that would normally drive the recorded β-LN). An example is shown for four consecutive cycles. After several such pairing trials, current injection was terminated and β-LN-firing phase over the next trials was compared to that before pairing. The effects of one such manipulation (dt > 0) were plotted: as predicted, an artificial phase-delay caused a corrective phase-advance. Twenty distinct experiments were carried out in six β-LNs; the expected phase shifts were observed in 16 of those 20. This is consistent with an adaptive role for STDP in the fine-tuning of β-LN spike-phase, and may explain the tight synchronization of β-LNs. Hence, STDP helps preserve the discrete and periodic structure of olfactory representations as they flow through the mushroom bodies (Cassenaer, 2007).

This study showed that the connections made by Kenyon cells to β-LNs are excitatory, strong on average, variable across pairs, and plastic. Plasticity follows time-sensitive hebbian associativity rules and is constrained to within-cycle interactions between pre-and post-synaptic neurons. STDP is therefore not specific to vertebrates or cortical architectures. The molecular underpinnings of STDP in this system, or whether STDP might confer the associative features usually ascribed to mushroom bodies, are not known. The fly and honeybee genomes both reveal coding sequences for N-methyl-D-aspartate (NMDA) receptor subunits and some Drosophila behavioural results are compatible with STDP learning rules. One hypothesis, readily testable, is that STDP provides associativity by tagging transiently the subset of synapses activated simultaneously by the odour, before the conditional arrival of a slower, non-specific reward signal (Cassenaer, 2007).

The results reinforce the proposed importance of spike timing for this, and possibly other, olfactory system(s): Kenyon cells act as coincidence detectors for synchronized projection neuron input, β-LNs act as coincidence detectors for Kenyon cell input; because STDP helps enhance β-LN synchronization, it is inferred that spike timing must be relevant also for the processing of β-LN output. These results indicate that the oscillation cycle-a temporal unit of processing first defined by negative feedback in the antennal lobe-is actively preserved in at least three successive layers of processing (projection neurons, Kenyon cells and β-LNs). It will be interesting to assess whether all Kenyon cell outputs obey the same STDP rules, and if these rules are themselves subject to learning-related modifications. Indeed, Kenyon cells seem to communicate with one another through axo-axonal chemical synapses. Given the dynamics of projection neuron/Kenyon cell activity vectors in response to odours, the possibility that Kenyon cell-Kenyon cell synapses also undergo STDP suggests a mechanism for sequence learning, similar to principles proposed for spatial map formation in rodents; in this study, however, the learned sequences have no relation to movement in physical space. The existence of such similarities (synaptic learning rules, and synchronized and sequential neural activity patterns) may bring us closer to understanding the relationships between circuit dynamics, architecture and learning in the brain (Cassenaer, 2007).

Go to top

A mechanistic model for reward prediction and extinction learning in the fruit fly

Springer, M. and Nawrot, M. P. (2021). eNeuro. PubMed ID: 33785523

Extinction learning, the ability to update previously learned information by integrating novel contradictory information, is of high clinical relevance for therapeutic approaches to the modulation of maladaptive memories. Insect models have been instrumental in uncovering fundamental processes of memory formation and memory update. Recent experimental results in Drosophila melanogaster suggest that, after the behavioral extinction of a memory, two parallel but opposing memory traces coexist, residing at different sites within the mushroom body. This study proposes a minimalistic circuit model of the Drosophila mushroom body that supports classical appetitive and aversive conditioning and memory extinction. The model is tailored to the existing anatomical data and involves two circuit motives of central functional importance. It employs plastic synaptic connections between Kenyon cells and mushroom body output neurons (MBONs) in separate and mutually inhibiting appetitive and aversive learning pathways. Recurrent modulation of plasticity through projections from MBONs to reinforcement-mediating dopaminergic neurons implements a simple reward prediction mechanism. A distinct set of four MBONs encodes odor valence and predicts behavioral model output. Subjecting this model to learning and extinction protocols reproduced experimental results from recent behavioral and imaging studies. Simulating the experimental blocking of synaptic output of individual neurons or neuron groups in the model circuit confirmed experimental results and allowed formulation of testable predictions. In the temporal domain, this model achieves rapid learning with a step-like increase in the encoded odor value after a single pairing of the conditioned stimulus with a reward or punishment, facilitating single-trial learning (Springer, 2021).

Go to top

A fly-inspired mushroom bodies model for sensory-motor control through sequence and subsequence learning

Arena, P., Cali, M., Patane, L., Portera, A. and Strauss, R. (2016). Int J Neural Syst: 1650035. PubMed ID: 27354193

Classification and sequence learning are relevant capabilities used by living beings to extract complex information from the environment for behavioral control. The insect world is full of examples where the presentation time of specific stimuli shapes the behavioral response. On the basis of previously developed neural models, inspired by Drosophila melanogaster, a new architecture for classification and sequence learning is here presented under the perspective of the Neural Reuse theory. Classification of relevant input stimuli is performed through resonant neurons, activated by the complex dynamics generated in a lattice of recurrent spiking neurons modeling the insect Mushroom Bodies neuropile. The network devoted to context formation is able to reconstruct the learned sequence and also to trace the subsequences present in the provided input. A sensitivity analysis to parameter variation and noise is reported. Experiments on a roving robot are reported to show the capabilities of the architecture used as a neural controller (Arena, 2016).

Go to top

The GABAergic anterior paired lateral neuron suppresses and is suppressed by olfactory learning

Liu, X. and Davis, R. L. (2009). Nat. Neurosci. 12(1): 53-59. PubMed ID: 19043409

GABAergic neurotransmitter systems are important for many cognitive processes, including learning and memory. A single neuron was identified in each hemisphere of the Drosophila brain, the anterior paired lateral (APL) neuron, as a GABAergic neuron that broadly innervated the mushroom bodies. Reducing GABA synthesis in the APL neuron enhanced olfactory learning, suggesting that the APL neuron suppressed learning by releasing the inhibitory neurotransmitter GABA. Functional optical-imaging experiments revealed that the APL neuron responded to both odor and electric-shock stimuli that was presented to the fly with increases of intracellular calcium and released neurotransmitter. Notably, a memory trace formed in the APL neuron by pairing odor with electric shock. This trace was detected as a reduced calcium response in the APL neuron after conditioning specifically to the trained odor. These results demonstrate a mutual suppression between the GABAergic APL neuron and olfactory learning, and emphasize the functional neuroplasticity of the GABAergic system as a result of learning (Liu, 2009).

Using single neuron labeling techniques and immunohistochemistry, the APL within the GH146-Gal4 expression domain neuron was identified as the first GABAergic neuron that innervated the mushroom bodies of Drosophila. The innervation was surprisingly broad, with this single neuron accounting for GABAergic processes that extend across the complete three-dimensional volume of the calyx, peduncle, and lobes. Knocking down GABA synthesis in the APL neuron enhanced olfactory learning, indicating that the role of APL was to suppress olfactory learning by releasing the inhibitory neurotransmitter GABA. Functional optical imaging revealed that the APL neuron responded to both CS and US stimuli used for training. It was further demonstrated that a memory trace registered as a reduced response specifically to the trained odor formed in the APL neuron after conditioning, suggesting that olfactory learning somehow suppressed the activity of this inhibitory neuron (Liu, 2009).

These observations meshed well with observations made from altering the expression level of the RDL receptor in the mushroom bodies. It was discovered that overexpression of RDL in the mushroom bodies inhibits learning, whereas reducing RDL expression in the mushroom bodies enhances learning, similar to the effect of reducing GABA synthesis in the APL neuron. Furthermore, the calcium responses to odor observed in the mushroom body neurons of flies that overexpress RDL are reduced, whereas the responses observed in flies with reduced expression of RDL are increased. Thus, increased learning is observed by either reducing RDL expression in the mushroom body neurons, or by decreasing GABA synthesis in the APL neuron that innervates the mushroom body neuropil. The logical conclusion is that the APL neuron provides the GABAergic input to the RDL receptor expressed on the mushroom body neurons, and that this neurotransmitter:receptor dynamic establishes the probability for learning to occur (Liu, 2009).

GABAergic feedback neurons projecting to the mushroom bodies have been reported in the honeybees. The morphology of these feedback neurons and their innervation patterns in the mushroom bodies are similar to the Drosophila APL neuron described in this study. Pairing an odor with a sucrose reward induces a decreased spike activity in the GABAergic feedback neurons towards the trained odor shortly after training, similar to the decreased response observed by optical imaging in the APL neuron after training. These observations suggest that the APL neuron in Drosophila might be the equivalent of the honeybee GABAergic feedback neurons. The processes of the GABAergic feedback neurons in the mushroom body lobes of the honeybee are considered to be postsynaptic and their processes in the mushroom body calyces are considered to be presynaptic. However, the processes of the APL neuron in the mushroom body lobes of Drosophila clearly contained presynaptic specializations, since synaptic vesicle release was observed from these processes by functional imaging. Thus, the functional relationship between the Drosophila APL neuron and the Apis GABAergic feedback neurons remains uncertain (Liu, 2009).

Functional optical imaging experiments have revealed multiple memory traces formed after olfactory conditioning in different areas of the Drosophila brain. The APL neuron memory trace was unique compared to previously described traces, since it was registered as a decrease rather than an increase of neuronal activity. This is not surprising given that the APL neuron releases the inhibitory neurotransmitter GABA. However, an important issue is raised by the combined observations. Is the increased activity in the mushroom bodies after training inducing the decreased activity in the APL neuron, or is the later serving as a permissive event for the former to take place? Temporally, the APL memory trace observed in this study forms within a similar time window as the early memory trace recently reported to form in the α'/β' mushroom body neurons, so these two scenarios remain equally possible. Another more complicated scenario is that these memory traces could form synergistically and in parallel rather than sequentially, since many insect neurons have mixed axons and dendrites and communicate bi-directionally with connected neurons (Liu, 2009).

The APL neuron exhibited a depression in activity after training that was specific to the trained odor compared to a control odor. The mechanism underlying this specificity is unclear. One of the simpler possibilities is that the APL neuron is both pre- and post-synaptic to mushroom body neurons, similar to models proposed for the dorsal paired medial (DPM) neuron. Training may produce a synaptic depression at the synapses between mushroom body neurons conveying the information about the trained odor and the postsynaptic APL neuron, but not at synapses between mushroom body neurons conveying information about other odors and the postsynaptic APL neuron. Such depression would reduce the activity of the APL neuron specifically to the trained odor. This depression of APL activity to the trained odor would also be registered as increased activity in the mushroom body neurons representing the trained odor, since the mushroom body neurons would then receive reduced inhibitory signals from the APL neuron acting presyaptically. A second possibility is that the increased activity of the mushroom body neurons conveying information about the trained odor might induce retrograde signaling causing a depression in specific APL presynaptic, inhibitory fibers. Recent studies of endocannabinoid-mediated hippocampal metaplasticity have revealed that focal stimulation of CA1 pyramidal neurons triggers a long-term depression at inhibitory synapses (I-LTD) restricted to a very small dendritic area (~10 microm), mediated by the postsynaptic release of endocannabinoid that binds to the presynaptic CB-1 receptor on the inhibitory neuron presynaptic terminals31. It remains unknown whether a similar retrograde signaling system exists in flies to mediate a similar effect, although a Ca2+ and synaptotagmin 4 dependent retrograde signaling mechanism has been discovered at the Drosophila neuromuscular junction that functions in a synapse-specific fashion. If selective suppression of inhibitory inputs exists in the central nervous system of Drosophila, then it may serve as a novel mechanism to code and store information in the brain (Liu, 2009).

Go to top

Parallel processing of appetitive short- and long-term memories in Drosophila

Trannoy, S., Redt-Clouet, C., Dura, J. M. and Preat, T. (2011). Curr Biol 21: 1647-1653. PubMed ID: 21962716

It is broadly accepted that long-term memory (LTM) is formed sequentially after learning and short-term memory (STM) formation, but the nature of the relationship between early and late memory traces remains heavily debated. To shed light on this issue, this study used an olfactory appetitive conditioning in Drosophila, wherein starved flies learned to associate an odor with the presence of sugar. Advantage was taken of the fact that both STM and LTM are generated after a unique conditioning cycle to demonstrate that appetitive LTM is able to form independently of STM. More specifically, it was shown that (1) STM retrieval involves output from γ neurons of the mushroom body (MB), i.e., the olfactory memory center, whereas LTM retrieval involves output from αβ MB neurons; (2) STM information is not transferred from γ neurons to αβ neurons for LTM formation; and (3) the adenylyl cyclase RUT, which is thought to operate as a coincidence detector between the olfactory stimulus and the sugar stimulus, is required independently in γ neurons to form appetitive STM and in αβ neurons to form LTM. Taken together, these results demonstrate that appetitive short- and long-term memories are formed and processed in parallel (Trannoy, 2011).

Short-term memory (STM) forms right after learning and is based on transient molecular and cellular events lasting from a few minutes to a few hours, whereas long-term memory (LTM) forms later on and involves gene expression and de novo protein synthesis following conditioning. The nature of the links between STM and LTM has long been debated, but there is consensus that STM and LTM are sequential processes and that LTM formation is built on the short-term trace. However, other studies have led to the conclusion that the mechanisms underpinning STM and LTM in vertebrates are at least partially independent (Trannoy, 2011).

Studies in insects have highlighted that mushroom bodies (MBs) play a major role in learning and memory. In particular, in Drosophila, MBs play a key role in olfactory learning and memory. The MBs in each brain hemisphere of Drosophila consist of approximately 2,000 neurons called Kenyon cells that can be classified into three major types: αβ, whose axons branch to form a vertical (α) and a medial (β) lobe, α'β', which also form a vertical (α') and a medial (β') lobe, and γ, which form a single medial lobe in the adult (Trannoy, 2011).

Several molecular-level studies have demonstrated that the cyclic AMP (cAMP) pathway plays a pivotal role in associative learning. In particular, calcium/calmodulin-dependent adenylyl cyclase (AC) encoded by the rutabaga (rut) gene is necessary to aversive olfactory conditioning where an odorant is associated to electric shock. Rut AC was proposed to function as a coincidence detector, integrating both the olfactory information carried by projection neurons to the MB and the electric shock carried by dopaminergic neurons. Interestingly, Rut cAMP signaling is required in γ neurons to form aversive STM (Zars, 2000; Blum, 2009) and in αβ neurons to form LTM (Blum, 2009), suggesting an independence of these two memory phases. However, several results suggest that aversive STM and LTM are not processed by fully independent neuronal pathways. Thus, a more efficient rescue of rut STM or LTM defect is observed when Rut is expressed in both γ and αβ neurons, suggesting that Rut is also involved in αβ neurons for aversive STM and in γ neurons for LTM. In addition, blocking αβ neuron synaptic transmission during memory retrieval impairs both aversive STM and LTM. Moreover, the induction of aversive STM and LTM requires different conditioning protocols, because STM is induced by one cycle of conditioning, whereas LTM formation requires spaced conditioning consisting of repeated training sessions separated by 15 min rest periods. These different training protocols may induce different physiological states within the relevant neurons, making it more difficult to interpret whether LTM is or is not built upon STM (Trannoy, 2011).

In Drosophila, appetitive STM and LTM are both generated after a single session of odorant-plus-sugar association, offering a powerful situation to study the link between the short- and the long-term trace. Rut AC has been hypothesized to be the coincidence detector in olfactory appetitive memory, because rut mutants exhibit poor immediate memory. It was shown that Rut AC in MB αβ and γ neurons or in projection neurons is sufficient for appetitive learning and STM, but it remains unknown which brain structure involves Rut activity for appetitive LTM (Trannoy, 2011).

Consolidation of appetitive STM and LTM requires output from α'β' neurons for 1 hr after training but not from αβ neurons. The role of γ neurons in appetitive STM or LTM consolidation has not yet been addressed. STM retrieval involves output from αβ and/or γ neurons, but the role of α'β' neurons in STM retrieval remains unknown. LTM retrieval involves output from αβ neurons but not from α'β' neurons, and the role of γ neurons in LTM retrieval has not yet been addressed. Thus a full picture of the role of MB neurons in appetitive memory processing has yet to emerge (Trannoy, 2011).

To clarify the role of the different MB neurons in appetitive STM and LTM, the c739-GAL4 driver and the UAS-shi c739-GAL4ts (shi) transgene were used to block synaptic transmission in αβ neurons during memory retrieval. The dominant temperature-sensitive SHITS protein blocks dynamin-dependent membrane recycling and synaptic vesicle release at restrictive temperature (33°C). This effect is reversible when the temperature is shifted back to 25°C. It was previously reported that c739/+; shi/+ flies have normal olfactory acuity. This study first checked that c739/+; shi/+ flies present a normal sugar response at restrictive temperature. It was then found that blocking synaptic transmission from MB αβ neurons did not affect appetitive STM retrieval. This was a surprising observation, because it was previously shown that output from MB αβ neurons is required for appetitive LTM retrieval, which was confirmed in this study. To rule out the possibility that the LTM retrieval defect might be due to c739-driven expression of shits outside of the MB, the MB247-GAL80+ construct (MB-GAL80) was used to inhibit GAL4 activity in αβ neurons. As expected, c739/MB-GAL80; shi/+ flies showed normal LTM performance when tested at restrictive temperature. Furthermore, appetitive LTM performance of c739/+; shi/+ flies at permissive temperature was normal. Hence, these data indicate that MB αβ neuron output is required for appetitive LTM retrieval but not for STM retrieval (Trannoy, 2011).

It was previously shown that output from MB αβ and γ neurons is required for STM retrieval. Because the data established that output from αβ neurons is not required for this process, the role of γ neurons in STM retrieval was examined. shits was expressed in γ neurons using the NP21-GAL4 driver. First, it was checked whether NP21/shi flies present normal sugar response and olfactory acuity at restrictive temperature. Blocking γ neuron output during the memory test significantly impaired appetitive STM. Inhibition of NP21 activity in the MB by MB-GAL80 rescued the STM defect of NP21/shi flies. Furthermore, the STM performance of NP21/shi flies was indistinguishable from controls at permissive temperature. Interestingly, blocking γ neuron output during the test did not affect LTM retrieval. The specific role of γ neurons in appetitive STM retrieval was confirmed with another γ neuron driver, 1471-GAL4. Taken together, the results indicate that MB γ neuron output is indispensable for appetitive STM retrieval but dispensable for LTM retrieval (Trannoy, 2011).

Appetitive STM and LTM retrieval each mobilize specific subsets of MB neurons, namely γ neurons for STM and αβ neurons for LTM. This might be due to the fact that STM and LTM are actually mutually independent, being formed and stored in spatially distinct compartments. Alternatively, γ and αβ neurons might be sequentially recruited: in this scenario, STM would form in γ neurons and information would be further transferred from γ to αβ neurons during the consolidation phase to build LTM. Under this latter assumption, blocking output from γ neurons during the LTM consolidation phase should lead to an LTM defect. To discriminate between the two hypotheses, γ neuron neurotransmission was blocked during training and consolidation and then LTM performance was measured. First, it was observed that blocking γ neuron output during training and consolidation did not affect STM. Then, to test the putative role of γ neurons in LTM formation, NP21/shi flies were trained at restrictive temperature and maintained at this temperature for 14 hr during the memory consolidation phase (the flies were kept at 33oC for 14 hr and not for the full 24 hr consolidation period because they started to die after 14 hr; given that appetitive LTM is being detectable 6 hr after training, it is likely that LTM consolidation takes place during the 14 hr time-period of γ neuron neurotransmission blockade). NP21/shi flies showed a normal 24 hr memory in this condition, suggesting that γ neuron output is dispensable during appetitive LTM acquisition and consolidation (Trannoy, 2011).

To further prove that LTM could be formed independently of STM, neurotransmission was constitutively blocked from γ neurons using UAS-TNT (TNT) construct encoding the tetanus toxin. TNT/+; NP21/+ flies are viable and present normal sugar response and olfactory acuity. Interestingly, a continuous blockade of γ neurons abolished STM but left LTM unaffected. Thus, TNT/+; NP21/+ flies trained with a single protocol showed no appetitive STM but a normal LTM at 24 hr. These results indicate that appetitive LTM formation is independent of STM and does not require synaptic communication between γ and αβ neurons (Trannoy, 2011).

Rut AC has been hypothesized to be the coincidence detector in olfactory appetitive memory. rut appetitive STM defect can be rescued by expressing Rut in αβ and γ neurons, but it had not yet been addressed whether Rut is specifically involved in γ or αβ neurons. Because circuit blocking experiments suggested that STM and LTM operate independently and involve different subsets of MB neurons, whether rut STM and LTM defects could be rescued independently by expressing Rut in γ and αβ neurons, respectively, was investigated. NP21 and c739 transactivators were used to express UAS-rut were used inrut2080 mutant flies. Expressing Rut in γ neurons fully rescued the rut STM defect, whereas expressing Rut in αβ neurons failed to rescue the rut STM defect. Conversely, expressing Rut in γ neurons failed to rescue the rut LTM defect, whereas expressing Rut in αβ neurons fully rescued the rut LTM defect. These results indicate that Rut AC is specifically required in γ neurons to form STM and in αβ neurons to form LTM, which further argues that appetitive STM and LTM are formed independently of each other (Trannoy, 2011).

The data suggest that appetitive STM and LTM are processed independently in γ and αβ neurons, respectively. Accordingly, immediate appetitive memory processing should involve γ neurons. To test this hypothesis, neurotransmission was constitutively blocked from γ neurons. As expected, TNT/+; NP21/+ flies displayed a 3 min memory defect. The involvement of γ neurons was further confirmed as 1471/+; shi/+ flies displayed a 3 min memory defect at restrictive temperature but not at permissive temperature. Strikingly, blocking neurotransmission from αβ neurons did not affect immediate memory. These results are in agreement with previous observations, suggesting that γ neurons support appetitive STM and αβ neurons support appetitive LTM. It has been shown that appetitive immediate memory is abolished by expressing SHITS in αβ and γ neurons under the MB247 driver. The partial inhibition observed with NP21 and 1471 GAL4 drivers might be due to the fact that MB247 shows a very strong expression in γ neurons, unlike 1471 or NP21. To further prove that the immediate appetitive memory forms in γ neurons, whether rut defect could be rescued was investigated by expressing Rut in γ neurons. Indeed, Rut expression under the NP21 driver restored rut immediate memory defect. On the contrary, expressing Rut in αβ neurons failed to rescue the rut immediate memory defect (Trannoy, 2011).

Appetitive conditioning offers a powerful situation for studying the link between STM and LTM, because both are formed after a single training cycle. It remained unknown whether the same MB neurons process both appetitive STM and LTM formation or whether these two memory phases are underpinned by specialized pathways. This study leads to a new understanding of the role of αβ and γ neurons in appetitive STM and LTM. Using distinct GAL4 drivers to specifically express SHITS or the tetanus toxin in either αβ or γ neurons, this study has shown that appetitive STM and LTM involve γ and αβ neurons, respectively. This study found the following: (1) immediate memory and STM processing involves Rut AC specifically in γ neurons, whereas LTM formation involves Rut in αβ neurons; (2) MB γ neuron output is required to retrieve immediate memory and STM but not LTM, and conversely, αβ neuron output is required to retrieve LTM but neither immediate memory nor STM; (3) γ neuron output is dispensable during memory consolidation, and therefore short-term information is not transferred from γ to αβ neurons to form LTM. Blocking γ neurons using tetanus toxin resulted in a striking phenotype, because flies completely deprived of appetitive STM exhibited normal LTM at 24 hr. In conclusion, this study provides strong evidence that in Drosophila, appetitive STM and LTM are two parallel and independent processes, involving different subsets of neurons within the MB (Trannoy, 2011).

The dynamics of the appetitive memory phase involve other neural circuits than just αβ and γ neurons. Blocking output from α'β' neurons for 1 hr after training affects both STM and LTM. Similarly, blocking output from dorsal paired medial (DPM) neurons, which project onto the MB lobes, for 1 hr after appetitive conditioning affects both STM and LTM. And it was recently shown that blocking GABAergic anterior paired lateral (APL) neurons, which project onto the MB lobes and dendrites, for 2 hr after appetitive conditioning affects STM but not LTM. It has been proposed that α'β'-DPM neurons form a recurrent loop that stabilizes STM and LTM and that APL activity regulates this loop for STM-related processes (Pitman, 2011). Because α'β' neurons are not required for either LTM or STM retrieval, the current results are in agreement with this scheme, where independent STM and LTM traces in γ and αβ neurons are maintained by output from α'β' neurons and MB-extrinsic neurons (Trannoy, 2011).

This model of STM and LTM independence is supported by several studies in other species. In Aplysia, synaptic connection between tail sensory neurons and motor neurons exhibits short- and long-term synaptic facilitation following learning. It has been shown that the induction of short-term synaptic plasticity is not necessary for the induction of long-term plasticity. Studies in vertebrates indicate that STM and LTM involve different biochemical pathways or distinct connected brain areas. This study goes one step further, because it identified neuronal structures that independently process STM and LTM, providing a unique opportunity to analyze biochemical and cellular processes specifically associated with STM and LTM (Trannoy, 2011).

Go to top

A simple computational model of the bee mushroom body can explain seemingly complex forms of olfactory learning and memory

Peng, F. and Chittka, L. (2016). Curr Biol [Epub ahead of print]. PubMed ID: 28017607

Honeybees are models for studying how animals with relatively small brains accomplish complex cognition, displaying seemingly advanced (or "non-elemental") learning phenomena involving multiple conditioned stimuli. These include "peak shift"-where animals not only respond to entrained stimuli, but respond even more strongly to similar ones that are farther away from non-rewarding stimuli. Bees also display negative and positive patterning discrimination, responding in opposite ways to mixtures of two odors than to individual odors. Since Pavlov, it has often been assumed that such phenomena are more complex than simple associate learning. This paper presents a model of connections between olfactory sensory input and bees' mushroom bodies, incorporating empirically determined properties of mushroom body circuitry (random connectivity, sparse coding, and synaptic plasticity). The model's parameters were not optimized to replicate specific behavioral phenomena, because the study was interested in the emergent cognitive capacities that would pop out of a network constructed solely based on empirical neuroscientific information and plausible assumptions for unknown parameters. The circuitry mediating "simple" associative learning was shown to also replicate the various non-elemental forms of learning mentioned above and can effectively multi-task by replicating a range of different learning feats. It was found that projection neuron-kenyon cell (PN-KC) synaptic plasticity is crucial in controlling the generalization-discrimination trade-off - it facilitates peak shift and hinders patterning discrimination - and that PN-to-KC connection number can affect this trade-off. These findings question the notion that forms of learning that have been regarded as "higher order" are computationally more complex than "simple" associative learning (Chittka, 2016).

Go to top

Two independent mushroom body output circuits retrieve the six discrete components of Drosophila aversive memory

Bouzaiane, E., Trannoy, S., Scheunemann, L., Placais, P. Y. and Preat, T. (2015). Cell Rep 11(8): 1280-1292. PubMed ID: 25981036

Understanding how the various memory components are encoded and how they interact to guide behavior requires knowledge of the underlying neural circuits. Currently, aversive olfactory memory in Drosophila is behaviorally subdivided into four discrete phases. Among these, short- and long-term memories rely, respectively, on the γ and α/β Kenyon cells (KCs), two distinct subsets of the ~2,000 neurons in the mushroom body (MB). Whereas V2 efferent neurons retrieve memory from α/β KCs, the neurons that retrieve short-term memory are unknown. This study identified a specific pair of MB efferent neurons, named M6, that retrieve memory from γ KCs. Moreover, network analysis revealed that six discrete memory phases actually exist, three of which have been conflated in the past. At each time point, two distinct memory components separately recruit either V2 or M6 output pathways. Memory retrieval thus features a dramatic convergence from KCs to MB efferent neurons (Bouzaiane, 2015).

In Drosophila, memory phases have historically been characterized behaviorally through the identification of specific mutants or through experimental features (e.g., resistance to cold shock or sensitivity to protein synthesis inhibitors). Based on these approaches, four aversive memory phases were previously documented: STM, MTM, ARM, and LTM. This study aimed to bridge the network and behavioral levels through a comprehensive dissection of the circuits involved in aversive memory retrieval at different time points after training, for both labile and anesthesia-resistant forms of memory. The role of MB neurons in all of these memories has long been established, but the identification of MB efferent neurons mediating memory retrieval is much more fragmented. In particular, STM is thought to be encoded in γ KCs, although the only output neurons that have been described so far project from the MB vertical lobes. Importantly, this study has established the characterization of M6 neurons and, to a lesser extent, the M4 neurons as an additional type of MB output neuron for memory retrieval, particularly in STM retrieval. The other main conclusion from this work is that ARM, previously considered as a singular memory phase, can be split into three distinct temporal phases: ST-ARM, MT-ARM, and LT-ARM, which rely on distinct sets of KCs and MB output neurons. Interestingly, a recent independent study also identified M4 and M6 neurons as necessary for the retrieval of immediate memory, both aversive and appetitive. This study investigated in detail the recruitment of the different retrieval circuits by the distinct spatiotemporal components of aversive memory. Altogether, this study strikingly confirms that the behavioral distinction of memory phases is clearly reflected at the level of neural networks, since a specific circuit can be assigned to each form of memory at a given time point (Bouzaiane, 2015).

A single aversive training cycle generates pairs of memories that are independently encoded and retrieved in time and space. Immediately after training, STM is retrieved from γ KCs via the M4/M6 neurons. Simultaneously, ST-ARM is retrieved from α/β KCs by the V2α neurons. It is currently technically impossible to image odor responses in these cell types within the timescale dictated by the short persistence of STM and ST-ARM. Indeed, this would require development of a setup to train flies directly under the microscope, but such experiments could be revealing. Blocking M4 neurons (projecting from the β' lobes) and M6 neurons (projecting from the γ and β' lobes) impaired STM retrieval; however, blocking either M4 or M6 neurons alone surprisingly failed to block STM retrieval. This indicates that these two neurons can serve redundant functions in STM retrieval. Consistent with this hypothesis, another study also reported that simultaneous blocking of M4 and M6 neurons much strongly impaired immediate aversive or appetitive memory than blocking of M6 neurons alone (Owald, 2015). It is possible that the α'/β' KC-M4 circuit is recruited as an alternative in case the default γ KC-M6 circuit is disrupted or damaged. Nonetheless, this redundancy does not exist at the KC level, since blocking γ KCs alone was sufficient to alter STM. Thus, the alternative circuit should also involve communication between γ and α'/β' KCs, through a mechanism that remains to be identified. Such a functional redundancy could guarantee the robustness of STM retrieval despite a minimum number of M6 neurons (Bouzaiane, 2015).

During the 3-hr range after training, MTM is retrieved from α/β KCs by V2α neurons, and MT-ARM is retrieved from γ KCs via the M6 neurons (see Spatiotemporal Distribution of Six Aversive Memory Phases in Drosophila). The respective assignments of labile and anesthesia-resistant memory components are thus inverted, as compared to immediate memory. Whether distinct forms of memory involve the same sets of neurons, and therefore act on the same synapses, has long been a subject of interest. Although understanding of the physiological processes underlying labile and anesthesia-resistant forms of memory is still limited, the identification of the separate output circuits for distinct forms of simultaneous memories reported in this study provides an insight into their distinguishing features. Calcium imaging of odor responses 3 hr after training has been performed in V2 neurons (for the retrieval of MTM), as well as in M6 neurons (for the retrieval of MT-ARM). Interestingly, training has dramatically opposite effects on the olfactory responses of these two cell types. V2 neurons respond to odors with a strong phasic increase in activity, and training decreases the response to the CS+ odorant as compared to CS-. On the contrary, M6 neurons display a moderate but prolonged increase in the relative response to CS+. This major mechanistic difference could explain why these distinct forms of memory cannot involve the same synapses, and hence the same circuits of KC-output neurons. Further studies are required to confirm whether this spatial segregation results from mutually antagonist or incompatible mechanisms (Bouzaiane, 2015).

In the 24-hr range, the LT-ARM formed after massed training is retrieved from α'/β' KCs by M6 neurons. In a previous study, a memory retrieval defect was recorded 24 hr after massed training by blocking V2 neurons with MZ160 or NP2492 driver (Séjourné, 2011). In contradiction with this previous report, this study did not measure a defect with the MZ160 driver 24 hr after massed training. The fact that no defect was observed with the 71D08 driver strongly suggests that an unfortunate error occurred in the crosses used in the former massed training experiment with the MZ160 driver (for example, that the NP2492 driver was used instead of the MZ160 one). The present study additionally showed that the LT-ARM defect observed with the NP2492 driver is due to non-cholinergic signaling and hence attributable to neurons other than V2. Collectively, these results indicate that LT-ARM is retrieved by M6 neurons (Bouzaiane, 2015).

LTM, which forms after spaced training, is encoded in α/β KCs according to several convergent reports and is retrieved via the V2 neurons. Previously, it was reported that LTM formation is gated during the inter-trial intervals of a spaced training by the activity of at most three pairs of PPL1 dopaminergic neurons projecting on the MB lobes. As the activity of the same neurons has an adverse effect on ARM, a model of LTM formation is proposed in which ARM and LTM are the products of antagonist consolidation pathways. The ARM pathway is fully inhibited during spaced training to allow for LTM formation. Now that this study has established that ARM is divided into three distinct phases, it can be asked which ARM phase inhibits LTM formation. Two separate lines of arguments advocate ST-ARM for this role. First, ST-ARM occurs on a timescale that is highly compatible with the gating that occurs over the 1.5-hr duration of the spaced training. Second, the location of LTM in the α/β KCs is firmly documented, and ST-ARM also relies on the α/β KCs. It is thus possible that cellular mechanisms underlying ST-ARM antagonize LTM formation through intra-α/β KCs processes. Overall, this study revealing composite memory circuits sheds light on how to address the questions of memory phase interaction and systems consolidation (Bouzaiane, 2015).

Go to top

Two components of aversive memory in Drosophila, anesthesia-sensitive and anesthesia-resistant memory, require distinct domains within the Rgk1 small GTPase

Murakami, S., Minami-Ohtsubo, M., Nakato, R., Shirahige, K. and Tabata, T. (2017). J Neurosci 37(22):5496-5510. PubMed ID: 28416593

For aversive olfactory memory in Drosophila, multiple components have been identified that exhibit different stabilities. Intermediate-term memory generated after single cycle conditioning is divided into anesthesia-sensitive memory (ASM) and anesthesia-resistant memory (ARM), with the latter being more stable. This study determined that the ASM and ARM pathways converged on the Rgk1 small GTPase and that the N-terminal domain-deleted Rgk1 was sufficient for ASM formation, whereas the full-length form was required for ARM formation. Rgk1 is specifically accumulated at the synaptic site of the Kenyon cells (KCs), the intrinsic neurons of the mushroom bodies (MBs), which play a pivotal role in olfactory memory formation. A higher than normal Rgk1 level enhanced memory retention, which is consistent with the result that Rgk1 suppressed Rac-dependent memory decay; these findings suggest that rgk1 bolsters ASM via the suppression of forgetting. It is proposed that Rgk1 plays a pivotal role in the regulation of memory stabilization by serving as a molecular node that resides at KC synapses, where the ASM and ARM pathway may interact (Murakami, 2017).

Drosophila olfactory learning and memory, in which an odor is associated with stimuli that induce innate responses such as aversion, has served as a useful model with which to elucidate the molecular basis of memory. Olfactory memory is divided into several temporal components and the intermediate-term memory (ITM) generated after single cycle conditioning is further classified into two distinct phases, anesthesia-sensitive memory (ASM) and anesthesia-resistant memory (ARM). Evidence has suggested that ASM and ARM are distinctly regulated at the neuronal level and at the molecular level (Murakami, 2017).

Mushroom bodies (MBs) represent the principal mediator of olfactory memory. Kenyon cells (KCs) are the intrinsic neurons of MBs, which are bilaterally located clusters of neurons that project anteriorly to form characteristic lobe structures and are a platform of MB-extrinsic neurons that project onto or out of the MBs. To elucidate the molecular mechanisms that underlie olfactory memory, screenings for MB-expressing genes have been a useful strategy. A technique used to examine gene expression in a small amount of tissue samples has enabled the investigation of the expression profile in MBs with a substantial dynamic range of expression levels and high sensitivity, thereby representing a promising approach with which to identify novel genes responsible for memory. This study deep sequenced RNA isolated from adult MBs and identified rgk1 as a KC-specific gene (Murakami, 2017).

The RGK protein family, for which Drosophila Rgk1 exhibits significant protein homology, belongs to the Ras-related small GTPase subfamily, which is composed of Kir/Gem, Rad, Rem, and Rem2. Their roles include the regulation of Ca2+ channel activity and the reorganization of cytoskeleton. Notably, mammalian REM2 is expressed in the brain and has been shown to be important for synaptogenesis, as well as activity-dependent dendritic complexity. These findings raise the possibility that RGK proteins may have a role in the synaptic plasticity that underlies memory formation. Drosophila has several genes that encode proteins homologous to the RGK family, including rgk1. Therefore, based on the ample resources available in Drosophila for the investigation of neuronal morphology and functions, Drosophila Rgk proteins will provide a good opportunity to elucidate the function of RGK family proteins (Murakami, 2017).

This study describes the analysis of Drosophila rgk1, which exhibited specific expression in KCs. Rgk1 accumulated at synaptic sites and was required for olfactory aversive memory, making the current study the first to demonstrate the role of an RGK family protein in behavioral plasticity. These data suggest that Rgk1 supports ASM via the suppression of Rac-dependent memory decay, whereas the N-terminal domain has a specific role in ARM formation. Together, these findings indicated that Rgk1 functions as a critical synaptic component that modulates the stability of olfactory memory (Murakami, 2017).

It is proposed that the ITM is genetically divided into three components: the rut-, dnc-, and rgk1-dependent pathways. The rut and dnc pathway act specifically for ASM and ARM, respectively, whereas rgk1 acts for both ASM and ARM, albeit partially. Consistent with this notion, it is noteworthy that the ASM and ARM pathways converge on Rgk1, yet the functional domains may be dissected; the full-length form of Rgk1 is required for ARM, whereas the molecule that lacks the N-terminal domain is capable of generating ASM, which suggests that the protein(s) required for ARM formation may interact with the N-terminal domain of Rgk1 (Murakami, 2017).

The data suggested that Rgk1 acts for both ASM and ARM, whereas the rgk1 deletion mutant, which was shown to be a protein null, exhibited only a partial reduction in ITM; these findings imply that Rgk1 regulates an aspect of each memory component. This idea may be explained by the expression pattern of Rgk1. Rgk1 exhibited exclusive expression and cell-type specificity in the KCs, whereas the memory components have been shown to be regulated by the neuronal network spread outside of the MBs and are encoded by multiple neuronal populations. For example, two parallel pathways exist for ARM and ASM is modulated, not only by MB-extrinsic neurons, but also by the ellipsoid body that localizes outside of the MBs. dnc-dependent ARM requires antennal lobe local neurons and octopamine-dependent ARM requires α'/β' KCs, in neither of which was Rgk1 detected. Therefore, Rgk1 may support a specific part of memory components that exists in a subset of KCs (Murakami, 2017).

The specific expression of Rgk1 in KCs suggests its dedicated role in MB function. Rgk1 exhibited cell-type specificity in KCs from anatomical and functional points of view. Rgk1 is strongly expressed in α/β and γ KCs and weakly expressed in α'/β' KCs and the expression of the rgk1-sh transgene in α/β and γ KCs was sufficient to disrupt memory. Several genes required for memory formation have been shown to be expressed preferentially in the KCs and the notable genes include dunce, rutabaga, and DC0. Although a recent study in KC dendrites showed that the modulation of neurotransmission into the KCs affects memory strength, KC synapses are thought to be the site in which memory is formed and stored. The current analyses with immunostaining and GFP fusion transgenes indicated that Rgk1 is localized to synaptic sites of the KC axons, which raises the possibility that Rgk1 may regulate the synaptic plasticity that underlies olfactory memory. Among the RGK family proteins, Rem2 is highly expressed in the CNS and regulates synapse development through interactions with 14-3-3 proteins, which have been shown to be localized to synapses and are required for hippocampal long-term potentiation and associative learning and memory. In Drosophila, 14-3-3Ζ is enriched in the MBs and is required for olfactory memory. In addition, the C-terminal region of Drosophila Rgk1 contains serine and threonine residues that exhibit homology to binding sites for 14-3-3 proteins in mammalian RGK proteins. Therefore, Rgk1 and 14-3-3Ζ may act together in the synaptic plasticity that underlies olfactory memory (Murakami, 2017).

The roles of RGK family proteins in neuronal functions have been investigated extensively. The current data, when combined with the accumulated data on the function of RGK family proteins, provide novel insights into the mechanism that governs two distinct intermediate-term memories, ASM and ARM. Regarding the regulation of ASM, the data showed that Rgk1 suppressed the forgetting that was facilitated by Rac. Rac is a major regulator of cytoskeletal remodeling. Similarly, mammalian RGK proteins participate in the regulation of cell shape through the regulation of actin and microtubule remodeling. Rgk1 may affect Rac activity indirectly by sharing an event in which Rac also participates because there have been no reports showing that RGK proteins regulate Rac activity directly; further, it was determined that rgk1 transgene expression did not affect the projection defect of KC axons caused by RacV12 induction during development. Therefore, it is suggested that Rgk1 signaling and Rac signaling may merge at the level of downstream effectors in the regulation of forgetting. A member of the mammalian RGK1 proteins, Gem, has been shown to regulate Rho GTPase signaling through interactions with Ezrin, Gimp, and Rho kinase. Rho kinase is a central effector for Rho GTPases and has been shown to phosphorylate LIM-kinase. In Drosophila, the Rho-kinase ortholog DRok has been shown to interact with LIM-kinase. Furthermore, Rac regulates actin reorganization through LIM kinase and cofilin and the PAK/LIM-kinase/cofilin pathway has been postulated to be critical in the regulation of memory decay by Rac. It was shown recently that Scribble scaffolds a signalosome consisting of Rac, Pak3, and Cofilin, which also regulates memory decay. Therefore, Rgk1 may counteract the consequence of Rac activity (i.e., memory decay) through the suppression of the Rho-kinase/LIM-kinase pathway. DRok is a potential candidate for further investigation of the molecular mechanism in which Rgk1 acts to regulate memory decay (Murakami, 2017).

The data indicated that Rgk1 is required for ARM in addition to ASM. It has been shown that Synapsin and Brp specifically regulate ASM and ARM, respectively. The functions of Synapsin and Brp may be differentiated in a synapse by regulating distinct modes of neurotransmission. The exact mechanism has not been identified for this hypothesis; however, the regulation of voltage-gated calcium channels may be one of the key factors that modulate the neurotransmission. Voltage-gated calcium channels are activated by membrane depolarization and the subsequent Ca2+ increase triggers synaptic vesicle release. The regulation of voltage-gated calcium channels has been shown to be important in memory; a β-subunit of voltage-dependent Ca2+ channels, Cavβ3, negatively regulates memory in rodents. Importantly, Brp regulates the clustering of Ca2+ channels at the active zone. Moreover, it has been demonstrated extensively that mammalian RGK family proteins regulate voltage-gated calcium channels. Kir/Gem and Rem2 interact with the Ca2+ channel β-subunit and regulate Ca2+ channel activity. In addition, the ability to regulate Ca2+ channels has been shown to be conserved in Drosophila Rgk1. Therefore, both Brp and Rgk1 may regulate ARM through the regulation of calcium channels, the former through the regulation of their assembly and the latter through the direct regulation of their activity. The finding that Rgk1 localized to the synaptic site and colocalized with Brp lends plausibility to the scenario that Rgk1 regulates voltage-gated calcium channels at the active zone (Murakami, 2017).

Several memory genes identified in Drosophila, including rutabaga, PKA-R, and CREB, have homologous genes that have been shown to regulate behavioral plasticity in other species. The identification of Drosophila rgk1 as a novel memory gene raises the possibility for another conserved mechanism that governs memory. There is limited research regarding the role of RGK proteins at the behavioral level in other species; however, the extensively documented functions of RGK proteins with respect to the regulation of neuronal functions, combined with the data presented in this study regarding Drosophila Rgk1, raise the possibility of an evolutionally conserved function for RGK family proteins in memory (Murakami, 2017).

Go to top

Neural control of startle-induced locomotion by the mushroom bodies and associated neurons in Drosophila

Sun, J., Xu, A. Q., Giraud, J., Poppinga, H., Riemensperger, T., Fiala, A. and Birman, S. (2018). Front Syst Neurosci 12: 6. PubMed ID: 29643770

Neural control of startle-induced locomotion by the mushroom bodies and associated neurons in Drosophila

Startle-induced locomotion is commonly used in Drosophila research to monitor locomotor reactivity and its progressive decline with age or under various neuropathological conditions. A widely used paradigm is startle-induced negative geotaxis (SING), in which flies entrapped in a narrow column react to a gentle mechanical shock by climbing rapidly upwards. This study combined in vivo manipulation of neuronal activity and splitGFP reconstitution across cells to search for brain neurons and putative circuits that regulate this behavior. The activity of specific clusters of dopaminergic neurons (DANs) afferent to the mushroom bodies (MBs) modulates SING, and DAN-mediated SING regulation requires expression of the DA receptor Dop1R1/Dumb, but not Dop1R2/Damb, in intrinsic MB Kenyon cells (KCs). Previous observations were confirmed that activating the MB α'β', but not αβ, KCs decreased the SING response, and further MB neurons implicated in SING control were identified, including KCs of the γ lobe and two subtypes of MB output neurons (MBONs). Co-activating the αβ KCs antagonizes α'β' and γ KC-mediated SING modulation, suggesting the existence of subtle regulation mechanisms between the different MB lobes in locomotion control. Overall, this study contributes to an emerging picture of the brain circuits modulating locomotor reactivity in Drosophila that appear both to overlap and differ from those underlying associative learning and memory, sleep/wake state and stress-induced hyperactivity (Sun, 2018).

This study has identified MB afferent, intrinsic and efferent neurons that underlie modulation of startle-induced locomotion in the Drosophila brain. Using in vivo activation or silencing of synaptic transmission in neuronal subsets, specific compartments of the MBs were shown to be central to this modulation. Implicated neurons include α'β' and γ KCs, subsets of PAM and PPL1 DANs, and the MBONs V2 and M4/M6. Some of the potential synaptic connections between these elements were characterized using splitGFP reconstitution across cells. Although the picture is not complete, these results led to a proposal of a scheme of the neuronal circuits underlying the control of locomotor reactivity in an insect brain (Sun, 2018).

It has been previously reported that the degeneration of DANs afferent to the MBs in the PAM and PPL1 clusters is associated with accelerated decline of SING performance in aging flies. This study has specifically addressed the role of these and other DANs in SING modulation. The initial observation was that thermoactivation of TH-Gal4-targeted DANs consistently led to decreased locomotor reactivity, while silencing synaptic output from these neurons had no effect. This result was verified by rapid optogenetic photostimulation, indicating that indeed DAN activation affects locomotor reactivity during the execution of the behavior. In contrast, blocking selectively synaptic output of the PAM DANs neurons resulted in a slight increase in SING performance, suggesting that a subset of spontaneously active neurons in the PAM inhibits SING. It should be noted, however, that this effect appeared small probably in part because SING performance was already very high for the control flies in the assay condition. This issue may have prevented detection of other modulatory neurons in the course of this study. Interestingly, the data suggest that those PAM neurons that inhibit SING are targeted by NP6510-Gal4, a driver that expresses in 15 PAM DANs that project to the MB β1 and β'2 compartments. The degeneration of these neurons also appears to be largely responsible for α-synuclein-induced decline in SING performance in a Parkinson disease model. Moreover, one observation is provided in this study, using DAN co-activation with TH-Gal4 and R58E02-Gal4, suggesting that other subsets of the PAM cluster may modulate locomotor reactivity with opposite effects, i.e., increase SING when they are stimulated (Sun, 2018).

This study further indicated that thermoactivation of two DANs of the PPL1 cluster, either MB-MP1 that projects to the γ1 peduncle in the MB horizontal lobes or MB-V1 that projects to the α2 and α'2 compartments of the MB vertical lobes, was sufficient to significantly decrease SING performance. This suggests that the MB-afferent DANs of the PPL1 cluster are also implicated in SING modulation. Other DAN subsets could play a role and are still to be identified. However, inactivation of a DA receptor, Dop1R1/Dumb, in MB KCs precluded DAN-mediated SING modulation, strongly suggesting that DANs afferent to the MBs play a prominent role in the neuronal network controlling fly's locomotor reactivity. In contrast, inactivating Dop1R2/Dumb in KCs did not show any effect on DA-induced SING control (Sun, 2018).

Therefore, these results suggest that DA input to the MBs can inhibit or increase the reflexive locomotor response to a mechanical startle, allowing the animal to react to an instant, sudden stimulus. In accordance with this interpretation, previous reports have shown that the MB is not only a site for associative olfactory learning, but that it can also regulate innate behaviors. By combining synaptic imaging and electrophysiology, a previous study demonstrated that dopaminergic inputs to the MB intrinsic KCs play a central role in this function by exquisitely modulating the synapses that control MB output activity, thereby enabling the activation of different behavioral circuits according to contextual cues (Sun, 2018).

A decrease in SING performance has been previously reported when KCs in the α'β' lobes, but not in the αβ and γ lobes, were thermogenetically stimulated or their synaptic output silenced. Using a set of specific drivers, the contribution of the various MB lobes in the modulation of this innate reflex was precisely studied. It was confirmed that the α'β' KCs down-regulate SING when they are activated but not when their output is inhibited. Other unidentified neurons, targeted by the rather non-selective c305a-Gal4 and G0050-Gal4 drivers, trigger a decrease in SING performance when they are inhibited by Shits1, and are therefore potential SING-activating neurons. It was further found that the MB γ lobes contain KCs that strongly inhibit SING when activated, both by thermogenetic and optogenetic stimulation, as shown with the γ-lobe specific driver R16A06-Gal4. However, thermoactivation of γ neurons with other drivers, like mb247-Gal4, which express both in the αβ and γ lobe, did not decrease SING. This could result from an inhibitory effect of αβ neuron activation on SING modulation by γ neurons. To test this hypothesis, a double-driver was generated by recombining mb247-Gal4 with R16A06-Gal4. Because both drivers express in the γ lobes, one would expect a stronger effect on SING modulation after thermoactivation with the double-driver than with R16A06-Gal4 alone. The opposite was observed, i.e., that SING was decreased to a lesser extent with the double-driver than with R16A06-Gal4 alone. Activation of mb247-Gal4 αβ neurons therefore likely counterbalanced the effect of γ neuron activation with R16A06-Gal4 on SING modulation. A similar and even more obvious result was obtained when mb247-Gal4 was recombined with the α'β' driver R35B12-Gal4: co-activation of the neurons targeted by these two drivers prevented the strong SING modulation normally induced by R35B12-Gal4 alone. These results suggest the existence of an inter-compartmental communication process for locomotor reactivity control in the Drosophila MB. Comparably, it was recently suggested, in the case of memory retrieval, that MB output channels are ultimately pooled such that blockade (or activation) of all the outputs from a given population of KCs may have no apparent effect on odor-driven behavior, while such behavior can be changed by blocking a single output. Such a transfer of information could occur, as was previously reported, through connections involving the MBONs within the lobes or outside the MB (Sun, 2018).

Finally, the activation of two sets of MB efferent neurons, cholinergic MBON-V2 and glutamatergic MBON-M4/M6, consistently decreased SING performance of the flies. In contrast, silencing these neurons had no effect on locomotor behavior. The dendrites of these MBONs arborize in the medial part of the vertical lobes (α2, α'3) and the tips of the horizontal lobes (β'2 and γ5), respectively, as further evidence that the prime and γ lobes, and DANs efferent to these compartments, are involved in SING modulation. Results are also shown from GRASP observations suggesting that the PAM DANs either lay very close or in some other manner make potential synaptic connections with the MBON-M4/M6 neurons in their MB compartments, as well as the M4/M6 with the PAM in the SMP. The results also provide evidence that the PPL1 DANs and MBON-V2 contact each other in the vertical lobes and that axo-axonic synaptic contacts may occur between the MBON-V2 and M4/M6 neurons in their common projection region in the SMP (Sun, 2018).

These MBONs are known to be involved in opposite ways in olfactory memory: DAN-induced synaptic repression of cholinergic or glutamatergic MBONs would result in the expression of aversive or attractive memory, respectively. This study finds, in contrast, that the activation of these two sets of MBONs had similar depressing effects on SING behavior. Interestingly, it has been recently reported that the glutamatergic MBONs and PAM neurons that project to the MB β'2 compartment are also required for modulation of another innate reflex, CO2 avoidance (Lewis, 2015). CO2 exposure, like mechanical startle, represents a potential danger for the flies, thus triggering an avoidance behavior that can be suppressed by silencing these MBONs in specific environmental conditions. However, it is the activation of glutamatergic MBONs that inhibits SING. This apparent discrepancy might be explained if the downstream circuits were different for these two escape behaviors (CO2 avoidance and fast climbing). Overall, the current results further support the hypothesis of a primary role of the MB as a higher brain center for adapting innate sensory-driven reflex to a specific behavioral context (Sun, 2018 and references therein).

Even though the model remains to be confirmed and elaborated, a proposed scheme summarizing the current working hypothesis is presented of the neural components underlying SING control. Sensory information from mechanical stimulation triggers an innate climbing reflex (negative geotaxis) that can be regulated by signals transmitted from MB-afferent DANs (in the PAM and PPL1 cluster) to select KCs and two sets of MBONs (V2 and M4/M6) in specific MB compartments. Processing of this information could occur through synergistic or antagonistic interactions between the MB compartments and, finally, the MBON neurons would convey the resulting modulatory signal to downstream motor circuits controlling the climbing reflex. It was observed that the axonal projections of these MBONs make synaptic contacts with each other and converge together to the SMP where the dendrites of DANs lie, suggesting that these projections might form feedback loops to control DA signaling in the circuits (Sun, 2018).

DA signals for SING modulation originate from PAM neuron subsets and neurons inside the PPL1 cluster (MB-MP1 and MB-V1) that project to the MB lobes (see Schematic representation of MB-associated neural components modulating startle-induced locomotion). Axon of MB-V1 is shown as a dashed line because a driver specific for this neuron could not be tested in this study. The α'β' and γ KCs appear to be the main information integration center in this network, while their effect on SING modulation is opposed by the activity of αβ lobe KCs. Processed SING modulation signals are then transferred by two subtypes of MB efferent neurons, MBON-V2 and M4/M6, to the downstream SING reflex motor circuits. These two MBON subtypes have their axons converging together in the SMP where they may form axo-axonic synaptic connections, in which MBON-V2 would be presynaptic to MBON-M4/M6. The SMP also contains dendrites of the PAM and PPL1 DANs, thereby potentially forming instructive feedback loops on DA-mediated SING modulation. Most neurons identified here downregulated SING performance when they were activated, except for a subset of the PAM clusters that appeared constitutively inhibitory (represented as darker neurons in the figure) and the αβ lobe KCs that seem to antagonize SING modulation by other MB neurons. The different MB lobes are shown in various shades of green as indicated. The PAM DANs, PPL1 DANs and MBONs are drawn in magenta, light blue and dark gray, respectively. PAM: protocerebral anterior medial; PPL1: protocerebral posterior lateral; MBON: mushroom body output neuron; SMP superior medial protocerebrum; ped: peduncle; pre: presynaptic; pos: postsynaptic (Sun, 2018).

SING performance can be affected by a collection of factors including the arousal threshold of the fly, the ability to sense gravity and also climbing ability. 'Arousal' is defined as a state characterized by increased motor activity, sensitivity to sensory stimuli, and certain patterns of brain activity. A distinction can be made between endogenous arousal (i.e., wakefulness as opposed to sleep) and exogenous arousal (i.e., behavioral responsiveness). In Drosophila, DA level and signaling control all known forms of arousal. Because the MB plays an important role in sleep regulation, sleep- or wake-promoting networks might indeed in part interact or overlap with those controlling locomotor reactivity. However, this study observed that thermoactivation with various drivers had in a number of cases opposite effects on sleep/wake state and SING. First, neuronal thermoactivation with TH-Gal4 suppresses sleep but decreases the SING response. Second, extensive thermogenetic activation screen revealed that α′β′ and γm KCs are wake-promoting and γd KCs are sleep-promoting. In the current experiments, neuronal activation of α′β′ or γ KCs both led to strongly decreased locomotor reactivity. Third, stimulating MBON-M4 and M6, which are wake-promoting, decreased SING performance (Sun, 2018).

Another brain structure, the EB, plays important roles in the control of locomotor patterns and is also sleep-promoting. Furthermore, the EB is involved in the dopaminergic control of stress- or ethanol-induced hyperactivity, which can be considered as forms of exogenously-generated arousal. Several drivers labeling diverse EB neuronal layers were used, and no noticeable effects of thermoactivation of these neurons on the SING response was found. It is concluded that the circuits responsible for SING modulation, although they apparently share some similarities, are globally different from those controlling sleep/wake state and environmentally-induced hyperactivity (Sun, 2018).

Overall, this work identified elements of the neuronal networks controlling startle-induced locomotion in Drosophila and confirmed the central role of the MBs in this important function. Future studies are required to complete this scheme and explore the intriguing interactions between the different MB compartments in SING neuromodulation (Sun, 2018).

Go to top

One-trial learning in larval Drosophila
Weiglein, A., Gerstner, F., Mancini, N., Schleyer, M. and Gerber, B. (2019). Learn Mem 26(4): 109-120. PubMed ID: 30898973

Animals of many species are capable of "small data" learning, that is, of learning without repetition. Ghis study introduces larval Drosophila melanogaster as a relatively simple study case for such one-trial learning. Using odor-food associative conditioning, it was first shows that a sugar that is both sweet and nutritious (fructose) and sugars that are only sweet (arabinose) or only nutritious (sorbitol) all support appetitive one-trial learning. The same is the case for the optogenetic activation of a subset of dopaminergic neurons innervating the mushroom body, the memory center of the insects. In contrast, no one-trial learning is observed for an amino acid reward (aspartic acid). As regards the aversive domain, one-trial learning is demonstrated for high-concentration sodium chloride, but is not observed for a bitter tastant (quinine). Second, follow-up, parametric analyses are provided of odor-fructose learning. Specifically,its dependency on the number and duration of training trials was ascertained, as well as the requirements for the behavioral expression of one-trial odor-fructose memory, its temporal stability, and the feasibility of one-trial differential conditioning. The results set the stage for a neurogenetic analysis of one-trial learning and define the requirements for modeling mnemonic processes in the larva (Weiglein, 2019).

Go to top

Independent contributions of discrete dopaminergic circuits to cellular plasticity, memory strength, and valence in Drosophila
Boto, T., Stahl, A., Zhang, X., Louis, T. and Tomchik, S. M. (2019). Cell Rep 27(7): 2014-2021. PubMed ID: 31091441

Dopaminergic neurons play a key role in encoding associative memories, but little is known about how these circuits modulate memory strength. This study reports that different sets of dopaminergic neurons projecting to the Drosophila mushroom body (MB) differentially regulate valence and memory strength. PPL2 neurons increase odor-evoked calcium responses to a paired odor in the MB and enhance behavioral memory strength when activated during olfactory classical conditioning. When paired with odor alone, they increase MB responses to the paired odor but do not drive behavioral approach or avoidance, suggesting that they increase the salience of the odor without encoding strong valence. This contrasts with the role of dopaminergic PPL1 neurons, which drive behavioral reinforcement but do not alter odor-evoked calcium responses in the MB when stimulated. These data suggest that different sets of dopaminergic neurons modulate olfactory valence and memory strength via independent actions on a memory-encoding brain region (Boto, 2019).

Dopaminergic neurons are involved in associative learning across taxa. In Drosophila, activation of certain dopaminergic neurons during associative learning tasks drives conditioned approach or avoidance, suggesting that they function as part of the reinforcement pathway and may encode stimulus valence (positive or negative). There are eight clusters of dopaminergic neurons in the fly brain. Neurons in three clusters project to the mushroom body (MB), a region that receives olfactory information and is required for olfactory learning. PAM dopaminergic neurons project to the horizontal lobes (β, β', and γ); PPL1 neurons project to the vertical lobes (α and α'), heel, and peduncle; and PPL2ab neurons project to the calyx. Different subsets of PPL1 and PAM neurons modulate reinforcement during learning (Boto, 2019).

Different sets of dopaminergic neurons play discrete roles in reinforcement during learning. Activating PPL1 dopaminergic neurons in lieu of reinforcement induces behavioral aversion to a paired odor. Conversely, activation of PAM neurons is sufficient to generate appetitive memories. These dopaminergic neurons respond strongly to the unconditioned stimulus during conditioning, releasing dopamine into the MB that integrates with odor-evoked spiking activity to drive learning-induced, cyclic AMP (cAMP)-dependent plasticity in the MB. Little is known about the third MB-innervating cluster, PPL2ab. These neurons innervate the ipsilateral MB calyx, as well as the lateral horn, lobula, optical track and esophagus, and medial and posterior protocerebrum. They have been implicated in the regulation of courtship behaviors but have no known role in learning and memory. How the multiple dopaminergic circuits that converge on the MB regulate learning is a major question (Boto, 2019).

This study has investigated the distinct roles of MB-innervating dopaminergic circuits in neuronal plasticity and behavioral memory. PPL2 neurons were found to play a role in learning, modulating neuronal gain and memory strength without imparting a strong valence. Thus, different subsets of dopaminergic neurons converging on memory-encoding neurons (the MB neurons) play roles in modulation of memory strength and valence (Boto, 2019).

This study provides insight into how PPL2 dopaminergic neurons regulate neuronal plasticity in the MB and behavioral learning. PPL2 neurons project to the MB calyx, where intrinsic MB neurons receive input from olfactory projection neurons. This places them in a position to exert strong influence over MB olfactory responses. The present data suggest that they both act as a gain control that modulates the MB olfactory responses and increase the strength of aversive short-term memory. Activation of PPL2 neurons in a differential conditioning protocol increased the relative responsivity to the paired odor in γ neurons. Behaviorally, this did not drive memory on its own but increased the strength of memory if paired with odor-shock conditioning. Therefore, PPL2 neurons appear to modulate the strength of aversive memory, rather than dictating its content. One mechanism underlying this effect could be that PPL2 neurons enhance MB responses to the odor during training, facilitating the generation of synaptic plasticity that has been observed at the MB output synapses. The memory enhancement effect that this study observed in flies may reflect a more general role that dopaminergic circuits play in other species. For instance, in mice, dopaminergic projections to the medial prefrontal cortex are not sufficient to induce memory, but they improve learning via effects on stimulus discrimination (Popescu et al., 2016) (Boto, 2019).

Previous studies have suggested that PPL2 neurons could regulate motivation and arousal. Increased responses in the MB do not likely represent the valence of a memory directly, but they may reflect a salience or motivational component of memory. This is supported by PPL1 stimulation failing to induce changes in the odor representation in the MB but inducing conditioned aversion that drives heterosynaptic depression at certain MB-mushroom body output neuron (MBON) synapses. In contrast, PPL2 neurons drive strong Ca2+ response plasticity in the MB but do not encode strong valence on their own. The effect was limited to aversive memory, possibly because the starvation necessary for appetitive protocols had already maximized the animals' arousal state and/or salience of the sensory cues (though other possibilities are discussed later) (Boto, 2019).

One function of PPL2 dopaminergic neurons may be to regulate the net responsivity of MB γ neurons to odorants and thereby alter the potential for stimuli to drive memory strength. Alternatively, the plasticity could regulate the balance of excitation across downstream MBONs that innervate spatially discrete zones of the MB and drive approach or avoidance behavior. For instance, increasing responses of MB γ neurons alone could increase the net excitatory drive to aversive MBONs relative to appetitive MBONs. In a previous study, appetitive conditioning was found to robustly increase Ca2+ responses to CS+ across the MB lobes (including both γ and α/β). This could be interpreted to indicate either that the motivational component of appetitive conditioning differentially engages MB circuitry relative to aversive conditioning or that the appetitive valence is encoded as a bona fide cellular-level memory trace, comprising an increase in Ca2+ responses across all MB lobes. If the latter is true, perhaps a selective increase in Ca2+ responses in γ reflects a more aversive signature. Previous studies have demonstrated a critical role of γ neurons in short-term memory. Rescue of Rutabaga in the γ lobe of rut mutants is sufficient to restore performance in short-term memory, and rescue of the D1-like DopR receptor in the γ lobe is sufficient to rescue both short- and long-term memory. In addition, aversive learning induces plasticity in synaptic vesicle release from the MB γ lobes (Boto, 2019).

Several caveats in experimental interpretations should be noted. First, it is not known whether the MB plasticity forms in parallel to memory enhancement or directly drives it. Contributions of polysynaptic circuit elements to the physiological effects (MB plasticity) and/or behavioral effects (enhanced memory) are possible. Future mapping studies may identify additional circuit elements contributing to the memory networks underlying these phenomena. Nonetheless, anatomical innervation of the MB calyx by PPL2 neurons positions them to provide strong modulatory input to the MB dendrites and associated neuronal circuitry. Thus, while valence is layered at the MB output synapses, the data suggest that PPL2 neurons may be a control mechanism that influences how responsive the MB is to odors, potentially altering the propensity for synaptic plasticity downstream (Boto, 2019).

Go to top

Optimal degrees of synaptic connectivity

Litwin-Kumar, A., Harris, K. D., Axel, R., Sompolinsky, H. and Abbott, L. F. (2017). Neuron 93(5): 1153-1164. PubMed ID: 28215558

Synaptic connectivity varies widely across neuronal types. Cerebellar granule cells receive five orders of magnitude fewer inputs than the Purkinje cells they innervate, and cerebellum-like circuits, including the insect mushroom body, also exhibit large divergences in connectivity. In contrast, the number of inputs per neuron in cerebral cortex is more uniform and large. This study investigated how the dimension of a representation formed by a population of neurons depends on how many inputs each neuron receives and what this implies for learning associations. This theory predicts that the dimensions of the cerebellar granule-cell and Drosophila Kenyon-cell representations are maximized at degrees of synaptic connectivity that match those observed anatomically, showing that sparse connectivity is sometimes superior to dense connectivity. When input synapses are subject to supervised plasticity, however, dense wiring becomes advantageous, suggesting that the type of plasticity exhibited by a set of synapses is a major determinant of connection density (Litwin-Kumar, 2017).

Go to top

Effect of circuit structure on odor representation in the insect olfactory system

Rajagopalan, A. and Assisi, C. (2020). eNeuro. PubMed ID: 32345734

In Neuroscience, the structure of a circuit has often been used to intuit function - an inversion of Louis Kahn's famous dictum, 'Form follows function'. However, different brain networks may utilize different network architectures to solve the same problem. The olfactory circuits of two insects, the Locust, Schistocerca americana, and the fruit fly, Drosophila melanogaster, serve the same function - to identify and discriminate odors. The neural circuitry that achieves this shows marked structural differences. Projection neurons (PN) in the antennal lobe (AL) innervate Kenyon cells (KC) of the mushroom body (MB). In locust, each KC receives inputs from approximately 50% PNs, a scheme that maximizes the difference between inputs to any two of approximately 50,000 KCs. In contrast, in drosophila, this number is only 5% and appears sub-optimal. Using a computational model of the olfactory system, this study shows the activity of KCs is sufficiently high-dimensional that it can separate similar odors regardless of the divergence of PN-KC connections. However, when temporal patterning encodes odor attributes, dense connectivity outperforms sparse connections. Increased separability comes at the cost of reliability. The disadvantage of sparse connectivity can be mitigated by incorporating other aspects of circuit architecture seen in Drosophila. These simulations predict that Drosophila and locust circuits lie at different ends of a continuum where the Drosophila gives up on the ability to resolve similar odors to generalize across varying environments, while the locust separates odor representations but risks misclassifying noisy variants of the same odor (Rajagopalan,, 2020).

Go to top

Memory elicited by courtship conditioning requires mushroom body neuronal subsets similar to those utilized in appetitive memory

Montague, S. A. and Baker, B. S. (2016). PLoS One 11: e0164516. PubMed ID: 27764141

In Drosophila courtship conditioning, male flies learn not to court females during training with an unreceptive female. He retains a memory of this training and for several hours decreases courtship when subsequently paired with any female. Courtship conditioning is a unique learning paradigm; it uses a positive-valence stimulus, a female fly, to teach a male to decrease an innate behavior, courtship of the female. As such, courtship conditioning is not clearly categorized as either appetitive or aversive conditioning. The mushroom body (MB) region in the fruit fly brain is important for several types of memory; however, the precise subsets of intrinsic and extrinsic MB neurons necessary for courtship conditioning are unknown. This study disrupted synaptic signaling by driving a shibirets effector in precise subsets of MB neurons, defined by a collection of split-GAL4 drivers. Out of 75 lines tested, 32 showed defects in courtship conditioning memory. Surprisingly, there were no hits in the γ lobe Kenyon cells, a region previously implicated in courtship conditioning memory. Several γ lobe extrinsic neurons were necessary for courtship conditioning memory. Overall, the memory hits in the dopaminergic neurons (DANs) and the mushroom body output neurons were more consistent with results from appetitive memory assays than aversive memory assays. For example, protocerebral anterior medial DANs were necessary for courtship memory, similar to appetitive memory, while protocerebral posterior lateral 1 (PPL1) DANs, important for aversive memory, were not needed. Overall, these results indicate that the MB circuits necessary for courtship conditioning memory coincide with circuits necessary for appetitive memory (Montague, 2016).

Go to top

Persistent activity in a recurrent circuit underlies courtship memory in Drosophila

Zhao, X., Lenek, D., Dag, U., Dickson, B. and Keleman, K. (2018). Elife 7. PubMed ID: 29322941

Persistent activity in a recurrent circuit underlies courtship memory in Drosophila

Recurrent connections are thought to be a common feature of the neural circuits that encode memories, but how memories are laid down in such circuits is not fully understood. This study presents evidence that courtship memory in Drosophila relies on the recurrent circuit between mushroom body γ (MBγ), M6 output, and aSP13 dopaminergic neurons. Persistent neuronal activity of aSP13 neurons was demonstrated; it transiently potentiates synaptic transmission from MBγ>M6 neurons. M6 neurons in turn provide input to aSP13 neurons, prolonging potentiation of MBγ>M6 synapses over time periods that match short-term memory. These data support a model in which persistent aSP13 activity within a recurrent circuit lays the foundation for a short-term memory (Zhao, 2018).

As animals pursue their goals, their behavioral decisions are shaped by memories that encompass a wide range of time scales: from fleeting working memories relevant to the task at hand, to short-term and long-term memories of contingencies learned hours, days, or even years in the past. Working memory is thought to reflect persistent activity generated within neural networks, including recurrent circuits. In contrast, short-term memory (STM) and long-term memory (LTM) involves changes in synaptic efficacy due to functional and structural modification of synaptic connections. However, the neural circuit mechanisms involved in the formation, persistence and transitions between these distinct forms of memory are not fully known (Zhao, 2018).

A robust form of memory in Drosophila is courtship memory, which can last from minutes to days, depending on the duration and intensity of training. Naïve Drosophila males eagerly court both virgin females, which are generally receptive, and mated females, which are not. However, upon rejection by mated females, they become subsequently less likely to court other mated females. This selective suppression of courtship towards mated females, called courtship conditioning, can be attributed to the enhanced sensitivity of experienced males to an inhibitory male pheromone deposited on the female during mating, cis-vaccenyl acetate (cVA) (Zhao, 2018).

Olfactory memory in insects relies on the function of a central brain structure called the mushroom body (MB). The principal MB cells, the cholinergic Kenyon cells (KCs), receive input from sensory pathways in the dendritic calyx region and from dopaminergic neurons (DANs) in the axonal lobes of the MB. These MB lobes are compartmentalized, with each compartment innervated by specific classes of DANs and MB output neurons (MBONs). MBONs receive input from both KCs and DANs (Zhao, 2018).

Previously work established that short-term courtship conditioning is mediated by the aSP13 class of DANs (also known as the PAM-γ5 neurons, which innervate the MBγ5 compartment. The activity of aSP13 neurons is essential for courtship conditioning in experienced males and sufficient to induce conditioning in naïve males. This study demonstrates that courtship memory also requires the corresponding MBγ KCs and the MBγ5 MBONs, the glutamatergic M6 neurons (also known as MBON-γ5&β;'2a neurons). Furthermore, this study presents evidence that MBγ, M6, and aSP13 neurons form a recurrent circuit and that persistent activity of the aSP13 neurons mediates plasticity at the MBγ to M6 synapses that can last from minutes to hours. Consistent with this model, M6 activity is required not only for memory readout but also, like aSP13, for memory formation. These data support a model in which persistent aSP13 activity within the MBγ>M6>aSP13 recurrent circuit lays the foundation for short-term courtship memory (Zhao, 2018).

This study has identified and characterized a tripartite MBγ>M6>aSP13 recurrent circuit that is essential for courtship memory in Drosophila. Behavioral and physiological data suggest the following model for the function of this feedback loop in short-term courtship memory. When a naïve male courts a mated female, the aSP13 and MBγ neurons may both be activated, perhaps in response to behavioral rejection and olfactory stimuli presented by the female, respectively. Dopamine released by aSP13 neurons potentiates transmission from MBγ to M6 neurons, which in turn provide a recurrent excitatory glutamatergic input back onto aSP13 neurons. Upon activation by M6, aSP13 activity persists for several minutes, providing a short time window during which continued MBγ activity can further drive M6 and aSP13. Thus sustained, aSP13 activity can lead to a longer-lasting accumulation of dopamine in the γ5 compartment, facilitating MBγ>M6 neurotransmission for up to 2-3 hr (Zhao, 2018).

The timescales for these physiological processes in ex vivo brain preparations broadly match the dynamics of courtship training and short-term memory formation. In the standard training paradigm, the male typically courts the female over several minutes, during which he performs a series of courtship bouts, each lasting for several seconds. As a result, a behavioral memory forms that lasts for several hours. Memory formation during training requires both M6 and aSP13, consistent with the notion that it reflects activation of the recurrent circuit. Memory readout requires M6 but not aSP13, and so evidently does not involve the recurrent circuit. It is infered that M6 suppresses courtship through other, aSP13-independent, pathways, and that its ability to do so is independent of experience. The consequence of training is to provide MBγ neurons with access to this M6-dependent courtship suppression pathway (Zhao, 2018).

Two important open questions are, first, what mechanism underlies the persistent calcium response in aSP13, and second, how does potentiation of MBγ>M6 synapses result in enhanced sensitivity to cVA (cis-vaccenyl acetate, cVA, a major component of the male cuticular hydrocarbon profile). The persistent calcium response the hallmark of courtship memory. The persistent response in aSP13 is evidently not an intrinsic property of aSP13, as it is not induced when aSP13 neurons themselves are activated. This observation would also likely exclude reciprocal excitation between aSP13 and other DANs. Persistent aSP13 activity is induced in response to transient M6 activation, and is not associated with any persistent activity of M6 neurons themselves. Thus, it is also unlikely to involve feedback from aSP13 and M6, although aSP13 >M6 synapses likely do exist. One possibility is that aSP13 persistence reflects unusually prolonged activation of the glutamatergic M6 >aSP13 synapses, or perhaps lies within interposed but still unidentified circuit elements (Zhao, 2018).

Given that M6 neurons activate a courtship suppression pathway, the potentiation of MBγ>M6 neurotransmission may explain why MBγ activation suppresses courtship in trained but not naïve flies. But MBγ neurons likely do not specifically respond to cVA, so this change alone cannot account for the enhanced sensitivity of trained flies to cVA. A small and variable subset of MB γneurons do receive input from the olfactory pathway that processes cVA, but cVA is not required during training and it is difficult to envision any other mechanism by which aSP13-dependent plasticity could be specifically restricted to the cVA-responsive MBγ neurons. It is formally possible that, despite the broad potentiation of MBγ output synapses upon training, it is only the contribution of the cVA-responsive MBγ neurons that drives courtship suppression when the male subsequently encounters as mated female. Alternatively, it has been suggested that M6 neurons encode a generic aversive signal, and so specificity to cVA might instead arise in downstream circuits that selectively integrate M6 output with the innate cVA-processing pathway from the lateral horn. In this regard, it is interesting to note that other MBONs have been implicated in courtship learning or general aversion, but M6 is the only MBON common to both (Zhao, 2018).

Late activation of the same aSP13 neurons in the time window of 8-10 hr after training is both necessary and sufficient to consolidate STM to LTM. Thus, in the time window when STM would otherwise decay, reactivation of the same MBγ>M6>aSP13 recurrent circuit may instead consolidate it into LTM. The mechanism by which aSP13 neurons are reactivated is unknown, but is evidently dependent upon their activation within the MBγ>M6>aSP13 recurrent circuit during training. It will be interesting to find out how this late aSP13 reactivation mechanism might relate to the mechanism that underlies persistent aSP13 activity during training (Zhao, 2018).

In summary, the data suggest that a brief persistent activity of aSP13 neurons represents a neural correlate of courtship working memory, while the prolonged potentiation of MBγ>M6 synapses corresponds to STM. It is proposed that persistent activity of the dopaminergic neurons in the MBγ>M6>aSP13 feedback loop lays the foundation for formation of short-term courtship memory in Drosophila, and that later reactivation of the same recurrent circuit consolidates STM into LTM. Thus, in contrast to the prevailing view of memory progression in the Drosophila MB that distinct memory phases are located in different compartments or lobes, the current data suggest that in the context of courtship conditioning, working memory, STM, and LTM all reside in the same γ5 compartment. These conclusions do not preclude however, the involvement of other MB neurons in courtship memory as it is conceivable that modulation, potentially of the opposite sign, of the appetitive memory pathways could be critical for courtship learning. This study therefore envisions that distinct courtship memory types are not located in distinct circuits, but rather mediated by distinct processes within a common circuit. Encoding distinct memory phases within a common circuit may be an efficient mechanism for encoding memories for which the behavioral consequence is largely independent of timing and context (Zhao, 2018).

Go to top

A pair of inhibitory neurons are required to sustain labile memory in the Drosophila mushroom body

Pitman, J. L., Huetteroth, W., Burke, C. J., Krashes, M. J., Lai, S. L., Lee, T., Waddell, S. (2011). Curr Biol 21: 855-861. PubMed ID: 21530258

Labile memory is thought to be held in the brain as persistent neural network activity. However, it is not known how biologically relevant memory circuits are organized and operate. Labile and persistent appetitive memory in Drosophila requires output after training from the α'β' subset of mushroom body (MB) neurons and from a pair of modulatory dorsal paired medial (DPM) neurons. DPM neurons innervate the entire MB lobe region and appear to be pre- and postsynaptic to the MB, consistent with a recurrent network model. This study identified a role after training for synaptic output from the GABAergic anterior paired lateral (APL) neurons. Blocking synaptic output from APL neurons after training disrupts labile memory but does not affect long-term memory. APL neurons contact DPM neurons most densely in the α'β' lobes, although their processes are intertwined and contact throughout all of the lobes. Furthermore, APL contacts MB neurons in the α' lobe but makes little direct contact with those in the distal α lobe. It is proposed that APL neurons provide widespread inhibition to stabilize and maintain synaptic specificity of a labile memory trace in a recurrent DPM and MB α'β' neuron circuit (Pitman, 2011).

Fruit flies form robust aversive or appetitive olfactory memory following a training session pairing odorant exposure with electric shock punishment or sucrose reward, respectively. Olfactory memories are believed to be stored in the output synapses of third-order olfactory system neurons in the mushroom body (MB), a symmetrical structure comprised of roughly 2500 neurons on each side of the brain that can be structurally and functionally dissected into αβ, α′β′, and γ neuron systems (Pitman, 2011).

Similar to aversive memory, appetitive memory measured 3 hr after training is referred to as middle-term memory and is comprised of a labile anesthesia-sensitive memory (ASM) and an anesthesia-resistant memory (ARM) component. Both of these phases and later long-term memory (LTM) require the action of the dorsal paired medial (DPM) neurons. DPM neurons exclusively innervate the lobes and base of the peduncle regions of the MB, where functional imaging suggests they are pre- and postsynaptic to MB neurons. DPM neuron projections to the α′β′ MB neuron subdivision appear to be of particular importance, and blocking output from α′β′ neurons themselves during a similar time period after training phenocopies a DPM neuron block. These data led to a proposal that reverberant activity in a recurrent MB α′β′-to-DPM neuron circuit is required to hold labile memory and for consolidation to LTM within αβ neurons (Pitman, 2011).

As part of a screen for additional neurons contributing to appetitive memory processing after training, A role for the second-order olfactory projection neurons (PNs) was tested. A uas-shibirets1 transgene was tested with the two most frequently utilized PN GAL4 drivers, GH146 and NP225. The uas-shits1 transgene allows one to temporarily block synaptic transmission from specific neurons by shifting the flies from the permissive temperature of <25°C to the restrictive temperature of >29°C. Appetitive olfactory memory was tested in GH146;uas-shits1 and NP225;uas-shits1 flies in parallel with control flies harboring the GAL4 drivers or the uas-shits1 transgene alone. c316/uas-shits1 flies, in which the DPM neurons were blocked for comparison, was also tested. No defects were apparent when the flies were trained and tested at the permissive temperature. To test for a role after training, all flies were trained at 23°C and immediately after training shifted to 31°C for 2 hr to disrupt neurotransmission from PN or DPM neurons. All flies were then returned to 23°C and tested for 3 hr memory. Memory was significantly impaired by GH146;uas-shits1 and c316/uas-shits1 manipulation, but not by NP225;uas-shits1. The performance of GH146;uas-shits1 and c316/uas-shits1 flies was significantly different from their respective control flies. In contrast, the performance of NP225;uas-shits1 flies was not significantly different from control flies. GH146 and NP225 label a large number of largely overlapping PNs. However, because NP225;uas-shits1 flies did not exhibit a memory defect, it is concluded that other neurons labeled by GH146 that are downstream of PNs could be responsible for the observed memory defect. GH146 most obviously differs from NP225 by also expressing in two anterior paired lateral (APL) neurons that innervate the MB. Each APL ramifies throughout the entire ipsilateral MB. This anatomy is similar to the DPM neurons, which project ipsilaterally throughout the MB lobes and base of the peduncle. Therefore whether the APL neurons were required for memory processing after training was further investigated (Pitman, 2011).

The NP5288 and NP2631 GAL4 lines have also been reported to label the APL neurons. NP5288 is expressed in a subset of PNs similar to that of NP225, as well as a few other distributed neurons in the brain. NP2631 does not label PNs but labels many other neurons in the brain including those in the median bundle, protocerebral bridge, and subesophageal ganglion. The consequence on memory of blocking synaptic output after training was tested from the neurons labeled in these additional APL-expressing lines. As before, no apparent defects were observed when the flies were trained and tested at the permissive temperature. However, flies trained at 23°C, shifted to 31°C for 2 hr after training, and tested for 3 hr memory at 23°C revealed defective memory. Memory performance of NP5288;uas-shits1 and NP2631;uas-shits1 flies was statistically different from the performance of their genetic control groups. These data are consistent with a role for APL neurons in memory processing after training (Pitman, 2011).

Others have reported that combining a ChaGAL80 transgene with GH146 inhibits expression in the APL neurons but leaves expression in PNs relatively intact. This approach was used to further test the requirement of uas-shits1 expression in APL neurons for observed memory defects. The ChaGAL80 transgene was combined with the GH146, NP5288, and NP2631 GAL4 drivers and uas-mCD8::GFP to visualize the extent of GAL4 inhibition by ChaGAL80 in these flies. As described for GH146;ChaGAL80 flies, confocal imaging of the GFP-labeled brains revealed that the ChaGAL80 transgene efficiently suppressed APL expression. The APL neurons were evident in all flies lacking ChaGAL80 but were not labeled in any of the three genotypes containing ChaGAL80. ChaGAL80 affected the expression in other neurons labeled by each GAL4 line to varying degrees. This analysis revealed a strong inhibition in GFP expression in the PNs labeled by GH146 and NP5288, although more PNs retained expression in NP5288 than in GH146, consistent with these two GAL4 drivers labeling partially nonoverlapping PN populations. ChaGAL80 inhibited APL expression in NP2631 and also removed expression from several other neurons. Expression was lost in some neurons innervating the subesophageal ganglion, whereas robust expression remained in the median bundle and protocerebral bridge of the central complex. Unfortunately, several intersectional approaches to create more specific control of APL neurons were unsuccessful (Pitman, 2011).

Next ChaGAL80 was combined with each APL-expressing GAL4 driver and the uas-shits1 transgene to test whether APL expression was necessary for the observed memory phenotypes when GH146, NP5288, and NP2631 neurons were blocked after training. Memory performance of GH146, NP5288, and NP2631 flies expressing uas-shits1 was assayed with or without the ChaGAL80 transgene along with GAL4;ChaGAL80 and uas-shits1 control flies for comparison. Again flies were trained at 23°C, shifted to 31°C for 2 hr after training, and tested 3 hr appetitive memory at 23°C. This manipulation significantly impaired memory performance in all flies without the ChaGAL80 transgene but not in flies with the ChaGAL80 transgene. Memory performance of GH146;uas-shits1, NP5288;uas-shits1, and NP2631;uas-shits1 flies was significantly different from uas-shits1 and GAL4;ChaGAL80 flies. In contrast, memory performance of all flies also harboring the ChaGAL80 transgene was not significantly different from the performance of the genetic control flies. These data suggest that expression in APL neurons is critical to disrupt 3 hr memory when blocking neurotransmission after training (Pitman, 2011).

The memory experiments described did not disrupt synaptic transmission during training or testing. Nevertheless, to control for possible confounding effects, the olfactory acuity and motivation to seek sucrose was further investigated in naive flies following a 2 hr disruption of synaptic transmission and 1 hr recovery as employed in the memory experiments. No olfactory acuity defects were observed in GH146;uas-shits1 or NP5288;uas-shits1 flies. However, NP2631;uas-shits1 flies exhibited a pronounced defect, which questions the validity of the memory experiments with this line. Reliance was therefore placed on the GH146;uas-shits1 and NP5288;uas-shits1 flies and the comparison to NP255;uas-shits1 flies to draw the conclusions. GH146;uas-shits1 and NP5288;uas-shits1 flies also exhibited sucrose acuity that was statistically indistinguishable from uas-shits1 controls. NP5288;uas-shits1 flies performed better than NP5288, which had an apparent defect. These data suggest that 3 hr appetitive memory requires synaptic output from the APL neurons after training, similar to the requirement for output from DPM and MB α′β′ neurons [5635563">5 (Pitman, 2011).

DPM neuron output is also required after training for appetitive LTM. Therefore GH146 and NP5288 were used to test whether APL block disrupted LTM. APL output was blocked for 2 hr after training and 24 hr memory was tested. Surprisingly, performance of GH146;uas-shits1 and NP5288;uas-shits1 flies was not significantly different from uas-shits1 or GAL4 flies, suggesting that APL output is specifically required for an earlier memory phase. Appetitive memory at 3 hr has been shown to be sensitive to cold-shock anesthesia delivered 2 hr after training. Therefore whether APL block affected this labile component was tested by performing experiments with cold shock. Wild-type flies were trained, half of them were subjected 2 hr afterward to a 2 min cold shock, allowed them to recover at room temperature, and tested for 3 hr memory. Performance of these flies was significantly different from those not receiving a cold shock. Interestingly, the performance of GH146;uas-shits1 flies in which APL neurons were blocked for 2 hr after training was statistically indistinguishable from cold-shocked wild-type flies. To further test whether APL-blocked flies were missing the cold-shock-sensitive memory component, the shits1 block and cold-shock treatments were combined. GH146;uas-shits1 flies were trained, APL was blocked after training by shifting flies to 31°C for 105 min, they were returned to 25°C for 15 min, given a 2 min cold shock, and tested for 3 hr memory. The performance of these flies was statistically different from GH146;uas-shits1 flies that received all treatment except the cold shock, suggesting that some ASM was present in GH146;uas-shits1-blocked flies. Importantly, memory performance was not totally abolished. Because significant memory remained following the uas-shits1 block and the cold shock, it was concluded that APL neuron block largely affects the labile anesthesia-sensitive appetitive memory. However, it is worth noting that APL block and cold shock cannot be considered to be operationally equivalent because the 2 min cold shock at 2 hr reduced memory observed at 24 hr, whereas blocking APL for 2 hr did not impact 24 hr memory. Therefore, blocking GH146 and NP5288 neurons appears to be more specific to labile appetitive memory than cold-shock treatment at this time (Pitman, 2011).

To determine possible sites of cell-cell contact, GFP reconstitution across synaptic partners (GRASP) was used. GRASP is detectable when neurons expressing complementary parts of an extracellular split-GFP are close enough that functional GFP is reconstituted. Flies were constructed that express lexAop-mCD4::spGFP11 in MB with 247-LexA and uas-mCD4::spGFP1-10 in APL or DPM with NP5288 or c316-GAL4. This analysis revealed distinct innervation of the MB by DPM and APL. DPM-MB GRASP was very dense and punctate throughout the MB lobes and peduncle and generally resembled the mCD8::GFP pattern covering all the major MB lobe regions. APL-MB GRASP was most notable for structure that is absent. The regular net-like appearance of APL seen with mCD8::GFP was not apparent, and label mostly decorated fibers running in parallel with MB neurons in the lobes. APL-MB GRASP in the vertical lobes was particularly revealing. Whereas the APL mCD8::GFP network extended throughout the vertical lobes and for APL mCD8::mCherry, APL-MB GRASP labeled processes extending in the α′ lobe but very little in the α lobe. Because GRASP is most reliably an indicator of proximity rather than connectivity, these data indicate that much of the APL network is distant to the MB neurons in the α lobe. GRASP was also used to visualize contact between APL and DPM neurons using NP5288-GAL4 for APL and L0111-LexA for DPM. APL-DPM GRASP revealed punctate labeling throughout the MB lobes that was most dense in the α′β′ lobes and base of the peduncle region. It is concluded that APL contacts DPM and MB neurons preferentially in the α′ lobe. In the horizontal lobes, APL contacts DPM throughout and makes dense contact with proximal portions of the MB β, β′, and γ neurons. The density of contact decreases toward the distal end of each horizontal lobe. It seems plausible that APL contacts other unidentified neurons, especially in the areas where they are apparently avoiding MB neurons (Pitman, 2011).

Presynaptic active zones were labelled in APL and DPM neurons by expressing a uas-Bruchpilot::GFP with a mCD8::mCherry transgene that should label the entire cell surface. Brp::GFP driven in APL with GH146 revealed presynaptic zones throughout the MB lobes with elevated levels in the α′β′ lobes. In contrast, Brp::GFP driven in DPM neurons with c316 revealed presynaptic zones throughout the lobes but very pronounced labeling in the αβ lobes (Pitman, 2011).

The anatomical data are consistent with a model of a recurrent MB α′β′-DPM-APL circuit and flow of activity from the α′β′ lobes through the DPM neurons to the αβ lobes. Importantly, GRASP suggests that APL and DPM contact is most dense within the α′β′ lobes, and Brp::GFP indicates strongest APL neurotransmitter release in α′β′. APL-MB GRASP indicates that APL preferentially contacts α′β′ MB neurons (most apparent in the vertical lobes). Interestingly, others have found that APL and DPM neurons are electrically coupled via heterotypic gap junctions. It will therefore be important to determine whether APL-DPM contact in α′β′ is exclusively electrical or a mixture of electrical and chemical (Pitman, 2011).

In conclusion, this study has identified a role after training for synaptic output from the GABAergic APL neurons. APL neurons appear to be specifically required for labile memory, and not for consolidation of long-term memory. APL and DPM neurons are functionally connected, yet outside of labile memory described in this study, disrupting either neuron can have different consequences. First, reducing GABA synthesis in APL neurons enhances learning, whereas DPM neurons are not required during acquisition. Functional imaging data suggests that learning specifically increases DPM neuron activity but reduces APL activity driven by the conditioned odor. It is suspected that these differences relate to APL also having processes in the MB calyx, where GRASP suggests that APL directly contacts MB neurons. Second, APL neurons are only required for earlier labile memory, whereas DPM neurons are required for labile and consolidated memory. It is suspected that this reflects the mode and function of their respective transmitters. It is proposed that APL provides broad nonselective cross-inhibition to maintain synaptic specificity in the recurrent DPM-MB-APL circuit that was originally set by the conditioned odor at acquisition. DPM in contrast might return activity to MB α′β′ neurons and supply consolidating signals to MB αβ neurons. It is expected that additional neurons contribute to the network and await identification. It will also be important to gain exclusive control of APL neurons (Pitman, 2011).

Active memory storage is thought of mostly on a seconds-to-minute timescale in mammals. ASM in Drosophila suggests a prolonged-duration active memory system. It will be important to determine the physiological property that is 'held' in the putative recurrent network. A step change in membrane potential accompanies periods of persistent activity in the oculomotor neural integrator of the goldfish. Such a change in the MB neurons coding olfactory memory would render them more easily excited by the conditioned odorant. Physiology will be needed to definitively add ASM in Drosophila to goldfish gaze stabilization and head direction and prefrontal cortical circuits in mammals as models to understand how memory is stored as persistent activity in recurrent neural networks. Nevertheless, the architecture and prolonged requirement for neurotransmission within the MB-DPM-APL neural circuit are suggestive. In addition, a recent gene profiling study of developing vertebrate cortex and annelid MB indicates a common evolutionary origin (Pitman, 2011).

Go to top

Two pairs of mushroom body efferent neurons are required for appetitive long-term memory retrieval in Drosophila

Placais, P. Y., Trannoy, S., Friedrich, A. B., Tanimoto, H. and Preat, T. (2013). Cell Rep. 5(3): 769-80 PubMed ID: 24209748

One of the challenges facing memory research is to combine network- and cellular-level descriptions of memory encoding. In this context, Drosophila offers the opportunity to decipher, down to single-cell resolution, memory-relevant circuits in connection with the mushroom bodies (MBs), prominent structures for olfactory learning and memory. Although the MB-afferent circuits involved in appetitive learning were recently described, the circuits underlying appetitive memory retrieval remain unknown. This study has identified two pairs of cholinergic neurons efferent from the MB alpha vertical lobes, named MB-V3, that are necessary for the retrieval of appetitive long-term memory (LTM). Furthermore, LTM retrieval was correlated to an enhanced response to the rewarded odor in these neurons. Strikingly, though, silencing the MB-V3 neurons did not affect short-term memory (STM) retrieval. This finding supports a scheme of parallel appetitive STM and LTM processing (Placais, 2013).

This study identified two pairs of cholinergic neurons, MBV3 neurons, that are efferent from the tip of the MB α lobes and are required for the retrieval of appetitive LTM, but not STM. It was previously established that appetitive STM and LTM are formed and retrieved through different sets of KCs (Trannoy, 2011). STM formation involves the rutabaga adenylylcyclase, probably as a coincidence detector, in γ KCs, and STM retrieval requires output from the same γ KCs. Conversely, LTM formation requires the same cyclase in α/β KCs, and LTM retrieval requires the output from α/β KCs (Trannoy, 2011). The fact that MB-V3 cholinergic neurons are specifically recruited for the retrieval of appetitive LTM, but not STM, and that they are efferent from the tip of a lobes is fully consistent with this scheme of parallel processing of STM and LTM. On this basis, it is anticipated that other as yet unidentified γ lobe-efferent neurons could mediate the transmission of the STM trace to relevant downstream areas (Placais, 2013).

This work, has also addressed the question of the specific requirement of MB-V3 neurons for the retrieval of appetitive LTM, as opposed to aversive forms of memory. MB-V3 neurons were found to be dispensable for the retrieval of aversive STM and ARM, in accordance with a recently published study by Pai (2013). However, the current results diverge from theirs on the retrieval of aversive LTM because they found that blocking MB-V3 neurons, with the same G0239 GAL4 driver this study used and Shits-thermosensitive neural blocker, yielded a mild but significant defect after spaced training. In addition, calcium-imaging experiments were performed with the MB-V3- specific driver G0239 that revealed no alteration of MB-V3 neuron olfactory responses after spaced training, consistent with behavioral results. In contrast, Pai (2013) reported an increase of the response to CS+ compared to CS after spaced training, but not massed training. It should be noted that they used for their imaging experiments a GAL4 driver that is not specific for MB-V3 neurons and especially labels other MB-extrinsic neurons that have projections on the vertical lobes close to MB-V3 dendrites (MB-V4 neurons according another the nomenclature). The conclusions are drawn from a lack of effect, which as a negative result, should be interpreted with caution. However, the imaging data are consistent with behavior experiments using three different effectors (Shits, ChATRNAi, and dTrpA1) that were all strong enough to completely abolish or very strongly affect (ChATRNAi, dTrpA1) appetitive LTM retrieval. Thus, it seems unlikely that the effects of all these manipulations could have been below detection limits. Two hypotheses are proprosed that would explain the discrepancy between the two reports. First, this study showed that MB-V3 neurons are required for the retrieval of aversive fLTM. This form of aversive LTM shares molecular featurescurrent and retrieval circuit with appetitive LTM. The formation of fLTM requires that flies are mildly fasted before training and put back on food immediately after training. Longer fasting periods before training and/or prolonged starvation after training prevents fLTM formation. Aversive LTM in the study of Pai (2013) might have contained fLTM because their spaced-training protocol takes twice as long as the current protocol (ten versus five cycles), during which flies may be mildly fasted. Being put back on food after training, flies could form some fLTM, which would result in the mild memory impairment they observed. In addition, it could be that other neurons (for example, the MB-V4 neurons, labeled in two other GAL4 drivers they used for behavior experiments and for imaging) are actually involved in aversive LTM retrieval, by themselves or maybe in combination with MB-V3 neurons. Testing this latter hypothesis would require another GAL4 driver that would target MB-V4 neurons independently of MB-V3 neurons. Unfortunately, no such tool has been reported so far (Placais, 2013).

Retrieval of appetitive LTM was functionally correlated to an increased response of MB-V3 neurons to the odor that was associated with sugar ingestion during training, an increase that did not occur in the hour range after training when only STM is formed, and that was abolished under two conditions that disrupt appetitive LTM formation or retrieval. In order to know if this increased response in MB-V3 neurons was the direct consequence of a similar phenomenon occurring upstream in the circuit, similar calcium-imaging experiments were performed in the a branch of α/β KCs, which are directly presynaptic to MB-V3 neurons. No trace of LTM formation was detected in these neurons, either by comparing responses to CS+ and CS within a fly or by comparing flies that are trained to make LTM and flies that undergo an unpaired protocol. These data contradict the conclusion from a recent study claiming that appetitive training induces an increase in the CS+ response in the MB α lobes 24 hr after training (Cervantes-Sandoval, 2013). However, what is shown in this latter study is that the average of the CS+/CS ratio is higher than one, but this effect may in fact be a bias toward high ratio values due to an inappropriate mathematical method of data analysis: the correct way for analyzing the ratio of experimental measurements is to average the logarithm of the ratio, as performed by several groups for similar imaging experiments. Furthermore, Cervantes-Sandoval (2013) did not show in their article the most relevant control data: comparing trained flies with flies that underwent an unpaired protocol would have been more accurate than statistical comparison of KC response between naive and trained flies. Of course, one cannot exclude that a calcium trace might eventually be described in KCs that this study would have missed. However, in the current situation, it seems likely that the LTM retrieval trace observed in MB-V3 neurons is not a simple readout of a similar calcium trace already present in upstream KC axons. In this scheme, appetitive LTM formation likely results in plasticity located at the level of the synapses between α/β KCs and MB-V3 neurons. When the conditioned odor is perceived, potentiation of these synapses would result in an increased response in MB-V3 neurons, which in turn could alter olfactory information to subsequent brain structures. The current data show that blocking the output of MB-V3 neurons is sufficient to fully abolish appetitive LTM retrieval. This of course does not exclude that other MB-extrinsic neurons may also be necessary, but it remains striking that appetitive LTM retrieval depends on signaling from as few as two neurons per brain hemisphere, which represents a huge convergence from the 1,000 α/β KCs. This drastic convergence is consistent with proposed models of memory encoding in insects, where the specificity of memory toward a given odor is conferred by the sparse representation of olfactory stimuli in the KCs, whereas an altered -- in this case increased -- response in the restricted number of output neurons is sufficient to encode an alteration of the valence of an odorant. The specificity of memory expression toward the conditioned odor is preserved provided that synaptic plasticity occurs only in the conditioned odor-responsive KCs; in the present case, at the synapse with MB-V3 neurons. At present, there is no straightforward way to target the KCs specifically responding to a given odor. The identification of MB-V3 should prove a key step in unraveling the precise mechanisms of synaptic plasticity underlying appetitive LTM (Placais, 2013).

Go to top

Plasticity of local GABAergic interneurons drives olfactory habituation

Das, S., Sadanandappa, M. K., Dervan, A., Larkin, A., Lee, J. A., Sudhakaran, I. P., Priya, R., Heidari, R., Holohan, E. E., Pimentel, A., Gandhi, A., Ito, K., Sanyal, S., Wang, J. W., Rodrigues, V. and Ramaswami, M. (2011). Proc Natl Acad Sci U S A 108: E646-654. PubMed ID: 21795607

Despite its ubiquity and significance, behavioral habituation is poorly understood in terms of the underlying neural circuit mechanisms. This study presents evidence that habituation arises from potentiation of inhibitory transmission within a circuit motif commonly repeated in the nervous system. In Drosophila, prior odorant exposure results in a selective reduction of response to this odorant. Both short-term (STH) and long-term (LTH) forms of olfactory habituation require function of the rutabaga-encoded adenylate cyclase in multiglomerular local interneurons (LNs) that mediate GABAergic inhibition in the antennal lobe; LTH additionally requires function of the cAMP response element-binding protein (CREB2) transcription factor in LNs. The odorant selectivity of STH and LTH is mirrored by requirement for NMDA receptors and GABAA receptors in odorant-selective, glomerulus-specific projection neurons (PNs). The need for the vesicular glutamate transporter in LNs indicates that a subset of these GABAergic neurons also releases glutamate. LTH is associated with a reduction of odorant-evoked calcium fluxes in PNs as well as growth of the respective odorant-responsive glomeruli. These cellular changes use similar mechanisms to those required for behavioral habituation. Taken together with the observation that enhancement of GABAergic transmission is sufficient to attenuate olfactory behavior, these data indicate that habituation arises from glomerulus-selective potentiation of inhibitory synapses in the antennal lobe. It is suggested that similar circuit mechanisms may operate in other species and sensory systems (Das, 2011).

A key observation is that rut function is uniquely required in adult-stage GABAergic local interneurons for STH and LTH. This observation contrasts with the rut requirement in mushroom-body neurons for olfactory aversive memory. The demonstration of fundamentally different neural mechanisms used in olfactory habituation and olfactory-associative memory elegantly refutes a proposal of the Rescorla-Wagner model that habituation (and extinction) may be no more than associations made with an unconditioned stimulus of zero intensity (Das, 2011).

The requirement for rut in inhibitory LNs also indicates that intrinsic properties of multiglomerular LNs change during habituation. However, logic, as well as anatomical and functional imaging data, indicate that glomerulus-selective plasticity must be necessary if LN changes produce odorant-selective habituation. A potentially simple mechanism for glomerulus-specific potentiation of LN terminals is suggested by the specific requirement for postsynaptic NMDAR in odorant-responsive glomeruli (Das, 2011).

The observation that LTH and STH show similar dependence on rut, NMDAR, VGLUT, GABAA receptors, and transmitter release from LN1 cells indicates a substantially shared circuit mechanism for the two timescales of habituation. The data point to a model in which transient facilitation of GABAergic synapses underlies STH; long-lasting potentiation of these synapses through CREB and synaptic growth-dependent processes underlies LTH. This finding differs in three ways from synaptic facilitation that underlies Aplysia sensitization. First, it refers to inhibitory synapses, with potentiation that may involve a specific heterosynaptic mechanism similar to that used for inhibitory Long Term Potentiation (iLTP) in the rodent ventral tegmentum. Second, by presenting evidence for necessary glutamate corelease from GABAergic neurons, it proposes the involvement of a relatively recently discovered synaptic mechanism for plasticity. Third, it posits an in vivo mechanism to enable glomerulus- specific plasticity of LN terminals (Das, 2011).

It is pleasing that, in all instances tested, physiological and structural plasticity induced by 4-d odorant exposure requires the same mechanisms required for behavioral LTH. When taken together, these different lines of experimental evidence come close to establishing a causal connection between behavioral habituation and accompanying synaptic plasticity in the antennal lobe (Das, 2011).

It is important to acknowledge that, although the current experiments show that plasticity of LN-PN synapses contributes substantially to the process of behavioral habituation, it remains possible that plasticity of other synapses, such as of recently identified excitatory inputs made onto inhibitory LNs, also accompany and contribute to olfactory habituation (Das, 2011).

The conserved organization of olfactory systems suggests that mechanisms of olfactory STH and LTH could be conserved across species. Although this prediction remains poorly tested, early observations indicate that a form of pheromonal habituation in rodents, termed the Bruce effect, may arise from enhanced inhibitory feedback onto mitral cells in the vomeronasal organ (Das, 2011).

Less obviously, two features of the circuit mechanism that we describe suggest that it is scalable and generalizable. First, selective strengthening of inhibitory transmission onto active glomeruli can be used to selectively dampen either uniglomerular (CO2) or multiglomerular (EB) responses; thus, the mechanism is scalable. Second, the antennal lobe/olfactory bulb uses a circuit motif commonly repeated throughout the brain, in which an excitatory principal cell activates not only a downstream neuron but also local inhibitory interneurons, which among other things, limit principal cell excitation (Das, 2011).

It is possible that, in nonolfactory regions of the brains, a sustained pattern of principal neuron activity induced by a prolonged, unreinforced stimulus could similarly result in the specific potentiation of local inhibition onto these principal neurons. Subsequently, the pattern of principal cell activity induced by a second exposure to a now familiar stimulus would be selectively gated such that it would create only weak activation of downstream neurons. In this manner, the circuit model that is proposed for olfactory habituation could be theoretically generalized. More studies are expected to test the biological validity of this observation (Das, 2011).

Go to top

The GABAergic anterior paired lateral neurons facilitate olfactory reversal learning in Drosophila

Wu, Y., Ren, Q., Li, H. and Guo, A. (2012). Learn Mem. 19: 478-486. PubMed ID: 22988290

Reversal learning has been widely used to probe the implementation of cognitive flexibility in the brain. Previous studies in monkeys identified an essential role of the orbitofrontal cortex (OFC) in reversal learning. However, the underlying circuits and molecular mechanisms are poorly understood. This study used the T-maze to investigate the neural mechanism of olfactory reversal learning in Drosophila. In reversal learning flies selectively avoid the odor more recently paired with shock in the second cycle of learning, regardless of what they learned during cycle 1. By adding a reversal training cycle to the classical learning protocol, it was shown that wild-type flies are able to reverse their choice according to the alteration of conditioned stimulus (CS)-unconditioned stimulus (US) contingency. The reversal protocol induced a specific suppression of the initial memory, an effect distinct from memory decay or extinction. GABA down-regulation in the anterior paired lateral (APL) neurons, which innervate the mushroom bodies (MBs), eliminates this suppression effect and impairs normal reversal. These findings reveal that inhibitory regulation from the GABAergic APL neurons facilitates olfactory reversal learning by suppressing initial memory in Drosophila (Wu, 2012).

The one-trial instant reversal paradigm was chosen as the primary reversal protocol for several reasons in addition to its being concise and simple to execute. First, if cycle 2 immediately follows cycle 1, the memory decay factor of cycle 1 is minimized, and the contribution of the reversal factor is highlighted. In fact, an elegant early analysis of the interaction between cycle 1 memory and cycle 2 training already disclosed that, although the reversal learning PI increases along with the delay of the cycle 2 training, the nonadditive effect of the two cycles is most salient when the delay is shorter than 30 min. Therefore, it would be more suitable to adopt the short delay protocol when examining the reversal effect (Wu, 2012).

Second, the necessity was demonstrated of the protocol (A+B– B+A–) for the investigation of reversal learning. Since CS+ odor presentation alone is sufficient to produce an optimal learning result in classical learning, CS– odor is dispensable in classical learning. Therefore, it brought the concern about whether CS– odor is also dispensable in reversal learning. A timeline-aligned direct comparison between the basic reversal learning protocol (A+B– B+A–) and the CS– odor absent protocol (A+B+) indicated that the reversal protocol exerted a stronger inverting power than CS– absent protocol. Therefore, CS– odor is indispensable in reversal learning protocol. Additional experiments showed that the flies developed distinguishable memory strengths for the two trained odors before testing, as significantly different PIs were observed when the two odor memories were separately evaluated. Moreover, a detailed investigation of a 'two-event choice' protocol, which actually is also A+B+, revealed that the flies are unable to make a consistent choice when the time delay between the two conditioning sessions is shorter than 2 min. This confirmed the effectiveness of the protocol and further differentiated the reversal learning from a decision-making process, which was probed through the 'two-event choice' protocol (Wu, 2012).

Third, the serial reversal paradigm that has been used in other model systems demonstrated that increased reversal experience expedited the reversal acquisition. Sequential reversal learning was also assayed, but so far improved reversal PIs with increasing reversal cycle numbers in Drosophila have not been observed. The difference could be due to a lower efficiency of Pavlovian training on the T-maze platform compared with operant training (Wu, 2012).

The importance of MBs for olfactory learning and memory in Drosophila has been extensively addressed. Genetic suppression of Rac expression in MBs disturbed reversal learning. In honeybees, MBs are required for reversal learning. Our experimental results further demonstrated that Drosophila olfactory reversal learning also requires the MBs. It is not surprising that the MBs have such a vital role since the reversal learning protocol is based on classical associative learning. For the same reason, it is difficult to distill reversal-specific mechanisms in MBs. Therefore, the role of GABA regulation in reversal learning is emphasized (Wu, 2012).

Previous studies have shown that down-regulation of GABA through knock- down of the GABAA receptor, resistance to dieldrin (RDL), or through reducing GABA synthesis in the APL neurons improves the associative learning ability of flies. The role of the GABAergic APL neurons in Drosophila learning and memory has recently garnered attention. The APL neurons are required to sustain labile memory, and gap junctions between the APL and the DPM neurons play a critical role in olfactory memory. Moreover, in locusts, the giant GABAergic neurons (GGN), which are anatomical equivalents to the Drosophila APL neurons, were found to form a normalizing negative-feedback loop within the MBs. Taken together with the finding that reducing GABA synthesis in APL neurons disrupted the reversal learning ability in flies, it is proposed that the APL neurons mediate the olfactory reversal learning, which might be acquired inside MBs. Specifically, it is speculated that when the flies face reversal training that contradicts previous learning, the APL neurons inhibit the initial memory and facilitate the reversal acquisition by releasing an appropriate amount of GABA. Since the APL neurons innervate the MBs broadly, it is difficult to imagine how the release of GABA would occur specifically on the MB neurons that represent the initially learned odor. Perhaps there exists some retrograde information from the MB neurons, representing the initially learned odor, that is sent to the presynaptic dendrites of the APL neurons, thus regulating the GABA release accordingly (Wu, 2012).

Attempts to manipulate RDL expression and up-regulate GABA synthesis in APL neurons failed to yield meaningful results; no significant change of reversal learning was observed in those experiments. The reason might be that the efficiencies of those manipulations were not sufficient to generate detectable differences in reversal learning. Nevertheless, enhanced GABAergic innervation has been shown to improve reversal learning in mice, which supports the speculation regarding GABA modulation in reversal learning (Wu, 2012).

Memory extinction occurs when the CS+ that was previously associated with US is presented without US pairing. In the reversal protocol, the memory extinction component could not fully account for the reversal learning PI. In honeybees, the blockade of MBs led to reversal defects without affecting extinction, which suggests an obvious divergence between extinction and reversal. Despite these apparent differences, there appears to be some common underlying mechanisms. Overlapping neural systems mediating extinction and reversal were reported in humans. In Drosophila, odorant memory extinction is supposed to be an intracellular process and antagonizes the previous memories at the molecular level, affecting cAMP signaling. The experimental results demonstrated that reversal is also cAMP-dependent, which indicates that there is a similar foundation for extinction and reversal at the molecular level. Based on those evidences, the relationship between memory extinction and reversal learning seems to be complicated and might involve different but not completely separate neuronal mechanisms (Wu, 2012).

Go to top

The GABA system regulates the sparse coding of odors in the mushroom bodies of Drosophila

Lei, Z., Chen, K., Li, H., Liu, H. and Guo, A. (2013). Biochem Biophys Res Commun 436: 35-40. PubMed ID: 23707718

In the mushroom bodies (MBs) of Drosophila, an analogue of the mammalian olfactory cortex, olfactory stimuli are sparsely encoded by Kenyon cells (KCs) that exhibit a high level of odor selectivity. Sparse coding of olfactory stimuli has significant advantages for maximizing the discrimination power and storage capacity of MBs. The inhibitory gamma-aminobutyric acid (GABA) system is important for regulating information processing in MBs, but its specific role in the sparse coding of odors is unclear. This study investigated the role of the GABA system in the sparse coding of odors using an in vivo calcium imaging strategy, which allowed measurment of the activity of the KC population at single cell resolution while the components of the GABA system were genetically manipulated. It was found that the down-regulation of GABAA but not GABAB receptors in KCs reduced the sparseness of odor representations in the MB, as shown by an increase in the population response probability and decrease in the odor selectivity of single KCs. Furthermore, the down-regulation of GABA synthesis in a pair of large GABAergic neurons innervating the entire MB reduced the sparseness of odor representations in KCs. In conclusion, the sparse coding of odors in MBs is regulated by a pair of GABAergic neurons through the GABAA receptors on KCs, thus demonstrating a specific role of the inhibitory GABA system on information processing in the MB (Lei, 2013).

Go to top

Origins of cell-type-specific olfactory processing in the Drosophila mushroom body circuit

Inada, K., Tsuchimoto, Y. and Kazama, H. (2017). Neuron 95(2): 357-367. PubMed ID: 28728024

How cell-type-specific physiological properties shape neuronal functions in a circuit remains poorly understood. This issue has been addressed in the Drosophila mushroom body (MB), a higher olfactory circuit, where neurons belonging to distinct glomeruli in the antennal lobe feed excitation to three types of intrinsic neurons, α/β, α'/&beta', and γ Kenyon cells (KCs). Two-photon optogenetics and intracellular recording revealed that whereas glomerular inputs add similarly in all KCs, spikes were generated most readily in α'/β' KCs. This cell type was also the most competent in recruiting GABAergic inhibition fed back by anterior paired lateral neuron, which responded to odors either locally within a lobe or globally across all lobes depending on the strength of stimuli. Notably, as predicted from these physiological properties, α'/β' KCs had the highest odor detection speed, sensitivity, and discriminability. This enhanced discrimination required proper GABAergic inhibition. These results link cell-type-specific mechanisms and functions in the MB circuit (Inada, 2017).

Go to top

Dopamine receptor DAMB signals via Gq to mediate forgetting in Drosophila

Himmelreich, S., Masuho, I., Berry, J. A., MacMullen, C., Skamangas, N. K., Martemyanov, K. A. and Davis, R. L. (2017). Cell Rep 21(8): 2074-2081. PubMed ID: 29166600

Dopamine receptor DAMB signals via Gq to mediate forgetting in Drosophila

Prior studies have shown that aversive olfactory memory is acquired by dopamine acting on a specific receptor, dDA1, expressed by mushroom body neurons. Active forgetting is mediated by dopamine acting on another receptor, Damb, expressed by the same neurons. Surprisingly, prior studies have shown that both receptors stimulate cyclic AMP (cAMP) accumulation, presenting an enigma of how mushroom body neurons distinguish between acquisition and forgetting signals. This study surveyed the spectrum of G protein coupling of dDA1 and Damb, and it was confirmed that both receptors can couple to Gs to stimulate cAMP synthesis. However, the Damb receptor uniquely activates Gq to mobilize Ca(2+) signaling with greater efficiency and dopamine sensitivity. The knockdown of Galphaq with RNAi in the mushroom bodies inhibits forgetting but has no effect on acquisition. These findings identify a Damb/Gq-signaling pathway that stimulates forgetting and resolves the opposing effects of dopamine on acquisition and forgetting (Himmelreich, 2017).

This study provides biochemical and behavioral evidence that the Drosophila DA receptor Damb couples preferentially to Gαq to mediate signaling by Damb for active forgetting. This conclusion offers an interesting contrast to the role of the dDA1 receptor in MBns for acquisition, and it resolves the issue of how MBns distinguish DA-mediated instructions to acquire memory versus those to forget. Prior studies had classified both dDA1 and Damb as cAMP-stimulating receptors, similar to mammalian D1/D5 DA receptors that work through Gαs/olf. The results extend prior studies of dDA1 by examining coupling of this receptor with multiple heterotrimeric G proteins to show that the receptor strongly and preferentially couples to Gs proteins. This affirms the receptor's role in the acquisition of memory, consistent with the tight link between acquisition and cAMP signaling. This study found that the Damb receptor can weakly couple to Gs proteins but preferentially engages Gq to trigger the Ca2+-signaling pathway, a feature not displayed by dDA1. Comparing the two Gαq paralogs of Drosophila (G and D) with a human ortholog shows that Drosophila GαqG and human Gαq share a conserved C terminus, essential for selective coupling to GPCRs, but quite distinct in sequence compared to the GαqD C terminus. Since GαqD is a photoreceptor-selective G protein that couples with rhodopsin, it is proposed that GαqG is the isoform that relays Damb's signals to spur forgetting (Himmelreich, 2017).

It is envisioned that memory acquisition triggered by strong DA release from electric shock pulses used for aversive conditioning drives both cAMP and Ca2+ signaling through dDA1 and Damb receptors in the MBns. Forgetting occurs from weaker DA release after the acquisition through restricted Damb/Gαq/Ca2+ signaling in the MBns. The coupling of Damb to Gs at high DA concentrations also explains why Damb mutants have a slight acquisition defect after training with a large number of shocks. Although the model allows the assignment of acquisition and forgetting to two distinct intracellular signaling pathways, it does not preclude the possibility that other differences in signaling distinguish acquisition from forgetting. These include the possible use of different presynaptic signals, such as a co-neurotransmitter released only during acquisition or forgetting (Himmelreich, 2017).

Go to top

Rac1 impairs forgetting-induced cellular plasticity in mushroom body output neurons

Cervantes-Sandoval, I., Davis, R. L. and Berry, J. A. (2020). Front Cell Neurosci 14: 258. PubMed ID: 33061890

Active memory forgetting is a well-regulated biological process thought to be adaptive and to allow proper cognitive functions. Recent efforts have elucidated several molecular players involved in the regulation of olfactory forgetting in Drosophila, including the small G protein Rac1, the dopamine receptor DAMB, and the scaffold protein Scribble. Similarly, recent work has reported that dopaminergic neurons mediate both learningand forgetting-induced plasticity in the mushroom body output neuron MBON-β2α'1. Two open questions remain: how does forgetting affect plasticity in other, highly plastic, mushroom body compartments and how do genes that regulate forgetting affect this cellular plasticity? This study shows that forgetting reverses short-term synaptic depression induced by aversive conditioning in the highly plastic mushroom body output neuron MBON-β1pedc>α/β. In addition, the results indicate that genetic tampering with normal forgetting by inhibition of small G protein Rac1 impairs restoration of depressed odor responses to learned odor by intrinsic forgetting through time passing and forgetting induced acutely by shock stimulation after conditioning or reversal learning. These data further indicate that some forms of forgetting truly erase physiological changes generated by memory encoding (Cervantes-Sandoval, 2020).

New insights have demonstrated that associative olfactory learning changes the output weight of KC synapses onto the corresponding MBON, suggesting a model in which dopamine-induced plasticity tilts the overall MBON network to direct appropriate behavior. In fact, recent physiological studies have shown that learning alters odor drive to specific MBONs. As a whole, these changes can be described as memory traces. Interestingly, reward learning appears to reduce the drive to output pathways that direct avoidance, whereas aversive learning increases drive to avoidance pathways while reducing the drive to approach pathways. this study choose to explore how forgetting and its genetic disruption affected memory traces formed in MBON-γ1pedc>α/β. This trace was selected because it can be easily induced after a very short stimulation of odor (1 s) along with optogenetic stimulation of dopaminergic neuron innervating the same MB compartment. In addition, it has been shown, using optogenetics, and behavior, that MB-γ1 compartment is the fastest to encode new memories, the most unstable or susceptible to memory decay and shock interference. Furthermore, it was shown that memories in this compartment are highly vulnerable to retroactive interference induced by a formation of additional olfactory memory. These features increased chances to observe forgetting related changes after memory encoding (Cervantes-Sandoval, 2020).

Using electrophysiology whole-cell recordings of MBON-γ1pedc>α/β, it has been shown that pairing and odor with specific artificial activation of dopaminergic PPL1-γ1pedc induced odor-specific synaptic depression. In addition, it has been shown that training the flies by pairing 1 min of CS+ with 12 shocks followed by 1 min presentation of CS-, induced a decreased response to the CS+ relative to CS- compared to no change in mock trained animals. This study first tried to confirm that this depression is observed when individual flies are imaged before and after learning and is observable using calcium reporter GCaMP6f. For this, the fly was trained under a confocal microscope, and calcium responses to odors was recorded in MBON-γ1pedc>α/β before and after training using split-gal4 driver MB112c. Pairing 20 sec of methylcyclohexanol (MCH) presentation with electric shock delivered to the fly legs by a floating electric grid platform induced a robust depression of calcium response to the learned odor. This depression was specific to the paired odor and was not observed in octanol (OCT), which was used as a non-paired odor. Additionally, this decreased response was not observed when flies were trained by a mock training (no shock) or backwards training (shock presented before odor onset). Finally, training flies with the reciprocal odor (OCT) showed similar results. These results confirm previous results and demonstrated that aversive olfactory conditioning induces under the microscope induced a robust memory trace represented as a depression of MBON-γ1pedc>α/β calcium responses to trained odors (Cervantes-Sandoval, 2020).

Next, it was asked how is this memory trace affected when forgetting occurs either intrinsically (as time passes) or is induced by interfering-electric shock or reversal learning. For this experiment GCaMP6f was expressed in MBON- γ1pedc>α/β using R12G04-lexA driver. Flies were trained as above and post-training responses were recorded 5, 15, or 30 min after conditioning. Similar to prior result, full depression to learned odor was observed 5 min after conditioning. This depression showed increasing recovery and was no longer significant from preconditioning responses 15 or 30 min after training. No significant changes were detected in the non-paired odor (OCT). This data demonstrate that at least for the memory trace observed in MBON-γ1pedc>α/β under these training conditions, intrinsic forgetting restitutes MCH calcium responses to normal levels after 30 min. It is important to indicate that a previous study showed that the decreased response to CS+ observed after 1 min CS+ odor pairing, lasted for at least 3 to 4 h after training. These differences, of course, could be attributed to the fact that the current study used a reduced training protocol (20s pairing) intending to improve chances of detecting rapid changes in the observed plasticity (Cervantes-Sandoval, 2020).

Previous work has shown that mechanical stimulation mediated by dopaminergic neurons can promote forgetting if presented after learning. Similarly, another study showed a decrease in conditioned response as a result of DAN activation after artificially induced aversive learning. These results suggested a model where dopamine bidirectionally regulates connectivity between KC > MBON; this regulation would be contingent on dopamine release in the context of odor presentation or not. This study asked whether the presentation of electric shock pulses presented after learning restored memory trace observed in MBON-γ1pedc>α/β to preconditioning levels. Results indicate that 12, 90 V shocks presented after conditioning was enough to restore responses to the paired MCH odor back to preconditioning levels. Responses to the non-paired odor were not affected by any of the protocols followed. Additionally, presenting the four shocks alone after conditioning was not enough to restore responses to CS+, indicating that the effect of shocks alone and reversal learning are somehow different (Cervantes-Sandoval, 2020).

Inducing acute memory forgetting can also be achieved by retroactive interference. In flies, it has been demonstrated that training with reversal conditioning, where flies are trained by presenting a first odor paired with electric shock followed by a second non-paired odor as a CS- and then immediately trained with the reverse contingency, showed decrease memory performance to the first CS+. This study now shows that retroactive interference induced forgetting by reversal learning also restored MCH responses to preconditioning levels. Responses to OCT after reversal conditioning were scarcely significantly decreased. Additionally, analysis of the CS+/CS- ratio showed that reversal conditioning not only restored responses to initial associated odor but also interfered with the synaptic depression of the newly learned contingency, namely no difference between responses ratios pre and post-conditioning. These results contrast with findings in plasticity induced in MBON-γ2 α'1, where reversal learning restores responses to initial CS+ and simultaneously depresses responses to the new CS+. The current results suggest the presence of not only retroactive interference to initial memory but also forgetting of secondary memory induced by proactive interference, which has been previously reported behaviorally. This difference can be attributed to the fact that different MB domains have different properties (Cervantes-Sandoval, 2020).

A previous study identified one of the central molecular regulators of active forgetting, the small G protein Rac1; overexpression of dominant negative (DN) form of Rac1 (RacN17) was found to impair normal memory forgetting. The current study tested how the memory trace in MBON-γ1pedc>α/β is affected by genetic disruption of this active forgetting regulation. For this flies that express GCaMP6f on MBON-γ1pedc>α/β using lexA driver, R12G04-lexA, were trained while expressing DN form of Rac1 in KC using gal4 driver R13F02-gal4. Expression of RacN17 in KC was further confined to adulthood using target system. Flies expressing RacN17 expression in adulthood showed a normal complete depression to the learned MCH odor. Nevertheless, these flies showed impaired recovery of memory-induced plasticity in MBON-γ1pedc>α/β after 15 and 30 min after conditioning with MCH. Unexpectedly, a mild non-specific depression to the non-paired odor was observed. This non-specific depression might be a result of Rac1 inhibition broadening odors representation and therefore increase in generalization; other explanations might also be possible. Despite this, two-way Anova analysis with Sidak's multiple comparisons test showed that the depression observed to the paired odor is significantly higher that the non-specific depression to the CS- (Cervantes-Sandoval, 2020).

Control flies carrying all genetic insertion but the uas-RacN17 and subjected to the same temperature conditions, showed normal depression to the learned odor and full recovery of odor response 30 min after conditioning. Control flies did not show a depressed odor response to the non-paired odor. Flies kept at 18°C to keep target system at non-permissive temperature showed normal learning-induced odor depression as well as normal recovery of odor calcium responses. These results suggest that, at least partially, RacN17 inhibits forgetting by impairing the bidirectional regulation of KC > MBON plasticity that is to say the restoration of the depressed odor responses to CS+ (MCH) in MBON-γ1pedc>α/β (Cervantes-Sandoval, 2020).

The above results indicate that the recovery of depressed olfactory responses in MBON-γ1pedc>α/β to a learned odor mediated by intrinsic forgetting, or normal memory decay trough time passing, is impaired when the DN form of Rac1 is expressed in KC. Next, attempts were made to investigate if RacN17 also affected memory trace loss when this is induced by interfering-electric shocks presented after learning. For that flies were trained as before and then 12, 90 V electric shocks were delivered to fly legs to induce acute forgetting. Flies expressing DN form of RacN17 in KC showed no recovery in learned-odor calcium responses (MCH) as compared to control flies. These results indicated that genetically interfering with memory forgetting by the expression of DN RacN17 in KC impairs not only intrinsic forgetting but also acute dopamine-mediated forgetting induced by strong electric shock stimulation after learning (Cervantes-Sandoval, 2020).

Finally, the effects were investigated of DN RacN17 expression on retroactive interference forgetting provoked by reversal conditioning. For this, the flies were trained presenting a first odor paired with electric shock (MCH, CS+) followed by air and a second odor not paired with electric shock (OCT, CS-). After learning flies were subjected to reversal training in which previous CS- odor was now paired with electric shock and the former conditioned odor was now presented as CS-. This protocol acutely induced a complete recovery of cellular memory trace in MBON- γ1pedc>α/β in control animals. Surprisingly, once again analysis of CS+/CS- ratio showed that reversal conditioning not only restored responses to initial associated odor but also interfered with the synaptic depression of the newly learned contingency, namely no difference between responses ratios pre and post conditioning. Again, these results reinforce the suggestion of proactive interference. Expression of RacN17 in KC during adulthood not only impaired memory trace restoration of initial contingency but also induced a strong depression to the secondary paired odor. These results indicated that genetically interfering with memory forgetting by expression of DN RacN17 in KC impairs restoration of olfactory responses in MBON-γ1pedc>α/β induced by intrinsic memory loss (time passing), acute forgetting induced by a non-associative stimuli (electric shock), and acute forgetting by new associations or memory updating (reversal learning) (Cervantes-Sandoval, 2020).

This study indicates that forgetting reverses synaptic depression induced by aversive conditioning in MBON-γ1pedc>α/β. This is true for intrinsic memory forgetting through time passing, and acute forgetting by both interfering-electric shock and retroactive interference provoked by reversal learning. The results also show physiological evidence of proactive interference in MBON-γ1pedc>α/β, previously observed behaviorally in Drosophila, where prior learning interferes with the formation of new learning. Results also indicate that genetic tampering with normal forgetting by inhibition of small G protein Rac1 impairs restoration of depressed odor responses to learned odor by the three mechanisms described above. It has been recently reported that Rac1 partially regulates forgetting through time passing as well as forgetting induced by reversal learning but it does not affect forgetting induced by non-associative experiences like heat stress, electric shocks or odor presented alone. The current results indicate that at least at physiological level Rac1 inhibition does affect odor responses restoration induced by electric shock in MBON-γ1pedc>α/β. It is possible that this apparent discrepancy is due to the fact that this study only explored the memory trace of a single MBON whereas, as mentioned before, behavior arises, most likely, as a combinational effect of the whole KC > MBON network. Therefore, a single compartment analysis does not necessarily reflect final behavior. It is also important to indicate that in this study, a reduced training protocol (20 s odor with four shocks) was used when compared to the classical training paradigm used for behavioral studies (1 min odor with 12 shocks). This mild training session was used to increase chances of observing the reversal of synaptic plasticity. The dynamic of these physiological changes when flies are trained with classical 1-min protocol remains to be studied. It was recently reported that training flies with a single training cycle (1 min odor presentation along with 12 shocks) induces an independent contextual memory that resides in the lateral horn. It is very likely that the forgetting described in this study and others have different dynamics and/or rules to this context-dependent memory (Cervantes-Sandoval, 2020).

In memory research, one school of thought holds that nothing is ever lost from storage and that forgetting represents only a temporal failure or inhibition of access to memory. The other school holds that memory is not completely preserved and that forgetting is a true erasure of information from storage. The current findings indicate that normal forgetting reverses plasticity generated by aversive learning in MBON-γ1pedc>α/β suggesting that forgetting, in the case of short-term non-protein-synthesis dependent memories, truly erases at least some of physiological changes caused by memory encoding. This finding does not exclude the possibility that other compartments have different properties nor that the same phenomenon is true for long-term memories. It is possible that memories that had undergone protein-synthesis dependent memory consolidation are more resistant to permanently reverse the physiological changes that form part of the long-term memory trace. In that case, when talking about forgetting, it is not possible to talk about erasure but rather a transient blockage of memory retrieval (Cervantes-Sandoval, 2020).

Go to top

Synapsin function in GABA-ergic interneurons is required for short-term olfactory habituation

Sadanandappa, M. K., Redondo, B. B., Michels, B., Rodrigues, V., Gerber, B., Vijayraghavan, K., Buchner, E. and Ramaswami, M. (2013). J Neurosci 33: 16576-16585. PubMed ID: 24133261

In Drosophila, short-term (STH) and long-term habituation (LTH) of olfactory avoidance behavior are believed to arise from the selective potentiation of GABAergic synapses between multiglomerular local circuit interneurons (LNs) and projection neurons in the antennal lobe. However, the underlying mechanisms remain poorly understood. This study shows that synapsin (syn) function is necessary for STH and that syn97-null mutant defects in STH can be rescued by syn+ cDNA expression solely in the LN1 subset of GABAergic local interneurons. As synapsin is a synaptic vesicle-clustering phosphoprotein, these observations identify a presynaptic mechanism for STH as well as the inhibitory interneurons in which this mechanism is deployed. Serine residues 6 and/or 533, potential kinase target sites of synapsin, are necessary for synapsin function suggesting that synapsin phosphorylation is essential for STH. Consistently, biochemical analyses using a phospho-synapsin-specific antiserum show that synapsin is a target of Ca2+ calmodulin-dependent kinase II (CaMKII) phosphorylation in vivo. Additional behavioral and genetic observations demonstrate that CaMKII function is necessary in LNs for STH. Together, these data support a model in which CaMKII-mediated synapsin phosphorylation in LNs induces synaptic vesicle mobilization and thereby presynaptic facilitation of GABA release that underlies olfactory STH. Finally, the striking observation that LTH occurs normally in syn97 mutants indicates that signaling pathways for STH and LTH diverge upstream of synapsin function in GABAergic interneurons (Sadanandappa, 2013).

Mechanisms for synaptic plasticity have been intensively analyzed in reduced preparations, e.g., in hippocampal or cortical-slice preparations and in neuromuscular synapses, e.g., of Aplysia, crayfish, frog, lamprey, and Drosophila. In contrast, synaptic mechanisms underlying specific forms of behavioral learning are less well understood in vivo, in terms of the signaling pathways engaged by experience as well as the cell types in which these mechanisms operate. The analysis of such in vivo mechanisms of plasticity requires an accessible neural circuit, whose properties are measurably altered by experience (Sadanandappa, 2013).

The olfactory response in Drosophila is initiated by the activation of odorant receptors expressed in the olfactory sensory neurons (OSNs), which project into the antennal lobe (AL) and form synapses with glomerulus-specific projection neurons (PNs) that wire to both, the mushroom bodies (MB) and the lateral protocerebrum. OSNs and PNs also activate multiglomerular excitatory and inhibitory local interneurons (LNs), which mediate lateral and intraglomerular inhibition in the AL. Strong evidence has been proved for a model in which Drosophila olfactory habituation, i.e., reduced olfactory avoidance caused by previous exposure to an odorant arises through potentiation of inhibitory transmission between GABAergic LNs and PNs of the AL. This study analyzes the molecular basis of this potentiation and propose a mechanism that could lead to increased presynaptic GABA release after odorant exposure (Sadanandappa, 2013).

Synapsins are a conserved family of synaptic vesicle-associated proteins. They are predominantly associated with the reserve pool of synaptic vesicles and their phosphorylation by kinases such as calcium-dependent protein kinases (CaMKs), protein kinase A (PKA), and MAPK/Erk, result in the mobilization of these vesicles and thereby induces presynaptic facilitation. While synapsin may play key roles in behavioral plasticity in mammals, its functions in learning and memory remain mysterious in part because mammals have three synapsins (I, II, and III) encoded by three different genes (Sadanandappa, 2013).

Recent studies of the single synapsin gene in Drosophila show that it is required for adult anesthesia-sensitive memory of odor-shock association (Knapek, 2010). Larval associative memory requires synapsin with intact phosphorylation target sites (Ser6 and Ser533) likely operating in a subset of MB neuron (Sadanandappa, 2013).

This study asked whether, where, and how synapsin functions in olfactory habituation. The observations indicate that short-term habituation (STH) requires synapsin function as well as its CaMKII-dependent phosphorylation in the LN1 subset of inhibitory interneurons in the AL. These point to presynaptic facilitation of GABA release as being a crucial mechanism for STH. The observation that long-term habituation (LTH) forms normally in syn97 mutants indicates that this form of long-term memory can be encoded without transition through a short-term memory stage (Sadanandappa, 2013).

Previous studies have established the important roles for vertebrate and invertebrate synapsins in short-term synaptic plasticity. The conclusion that increased GABA release underlies olfactory STH is built on previous studies that have established the role of synapsin in synaptic vesicle mobilization and presynaptic facilitation. Several in vivo and in vitro studies have provided evidence that synapsin phosphorylation and dephosphorylation regulate the effective size of different vesicle pools and thereby control neurotransmitter release. In the mollusks Aplysia californica and Helix pomatia/aspersa, the phosphorylation and redistribution of synapsin induced by the PKA or MAPK/Erk-pathways mediates short-term facilitation of transmitter release. In Drosophila larval neuromuscular junctions, post-tetanic potentiation of transmitter release requires synapsin function and is accompanied by mobilization of a reserve pool of synaptic vesicles. And in the mouse CNS , enhanced ERK signaling in the inhibitory neurons results in increased levels of synapsin-1 phosphorylation and enhanced GABA release from hippocampal interneurons (Sadanandappa, 2013).

In context of these prior studies of synapsin, the current observation that synapsin and its phosphorylation are necessary in GABAergic local interneurons for STH indicates that the facilitation of GABA release from LNs attenuates excitatory signals in the AL and thereby results in the reduced behavioral response, characteristic of habituation. The molecular reversibility of phosphorylation could potentially account for the spontaneous recovery of the olfactory response after STH. These in vivo observations on synapsin function constitute a substantial advance as they show that synapsin-mediated plasticity, usually observed in response to experimentally enforced electrophysiological stimulations, can underlie behavioral learning and memory induced by sensory experience. In addition, by placing these changes in an identified subpopulation of local inhibitory interneurons in the AL, they implicate presynaptic plasticity of LNs as a mechanism necessary for habituation and thereby substantially clarify a neural circuit mechanism for this form of nonassociative memory (Sadanandappa, 2013).

A significant question is how synapsin phosphorylation is regulated in vivo for synaptic plasticity and how this contributes to altered circuit function that underlies behavioral learning. This has been difficult to address for two reasons: first, due to the complexity and potential promiscuity of kinase signaling pathways, and second, this requires not only the identification of neurons that show synapsin-dependent plasticity, but also concurrent understanding of the circuit functions of these neurons and their postsynaptic target(s) in vivo. The latter has been a particular challenge in neural networks for behavioral memory. For instance, although synapsin- and S6/S533-dependent odor-reward memory trace localized to the MBs, the downstream targets of the MB to mediate learned behavior are only beginning to be unraveled. This makes it difficult to comprehensively interpret the complex role of synapsin in associative memory illustrated by the finding that 3 min memory and anesthesia-sensitive 2 h memory require synapsin, whereas 5 h memory and anesthesia-resistant 2 h memory do not (Knapek, 2010). In the Drosophila neural circuit that underlies olfactory habituation, this study presents evidence that is most parsimoniously explained by a model in which CaMKII regulates synapsin phosphorylation in GABAergic LNs of the AL, neurons that are known to inhibit projection neurons that transmit olfactory input to higher brain centers (Sadanandappa, 2013).

in vivo CaMKII studies not only show the requirement of its function in olfactory habituation, but also demonstrate that the predominant form of synapsin, which arises from pre-mRNA editing, is a potent substrate for CaMKII. Thus, expression of an inhibitory CaMKII peptide in neurons not only reduces (yet does not abolish) the S6-phosphorylated form of synapsin detected using S6 phospho-specific antibody, it also blocks olfactory habituation, thereby providing experimental evidence for presynaptic function of CaMKII. These findings provide in vivo support for a model that proposed the primacy of CaMKII regulation of synapsin function; however, the structural basis for this regulation may be slightly different as invertebrate synapsins lack the D domain found on mammalian homologs that appear to be the major targets of CaMKII-dependent synapsin phosphorylation (Sadanandappa, 2013).

These conclusions on the role of CaMKII must be qualified in two ways. First, though tight correlations were shown between CaMKII phosphorylation of synapsin and the protein's function in STH, the data fall short of establishing causality. Second, it was not possible to directly demonstrate that odorant exposure that induces STH results in CaMKII-dependent synapsin phosphorylation in LNs. Attempts to address the latter issue failed because of technical difficulties in using phospho-specific antibodies for in vivo immunohistochemistry. Finally, although it is proposed that CaMKII phosphorylation is necessary for synapsin function during STH, it remains possible that other additional kinases, e.g., PKA and CaMKI, also contribute to synapsin regulation in vivo required for STH (Sadanandappa, 2013).

Several studies have discriminated between mechanisms of short-term and long-term memory (STM and LTM) formation. Most have focused on the distinctive requirement for protein synthesis in LTM and have not established whether STM is a necessary step in the formation of LTM or whether STM and LTM arise via distinctive, if partially overlapping molecular pathways. However, a few reports describe experimental perturbations that greatly reduce STM, without altering LTM (Sadanandappa, 2013).

In cultured sensorimotor synapses, which provide a surrogate model for behavioral sensitization in Aplysia, it has been shown that synaptic application of a selective 5-HT receptor antagonist blocked short-term facilitation while leaving long-term facilitation unaffected. In contrast, somatic application of the same antagonist selectively blocked long-term facilitation. This showed that long-term synaptic plasticity, though triggered by similar inputs (5-HT in this case), can progress through a pathway that does not require short-term plasticity (Sadanandappa, 2013).

In mammalian systems, it has been shown that a variety of pharmacological infusions into the hippocampal CA1 region or in entorhinal/ parietal cortex inhibited STM without affecting LTM of a shock avoidance memory task. In Drosophila, different isoforms of A-kinase anchor protein (AKAP) interacts with cAMP-PKA and play a distinct role in the formation of STM and LTM (Lu, 2007; Zhao, 2013). While these studies indicate that LTM is not built on STM, they do not rule out the possibility that they share an early common synaptic mechanism, but differ in subsequent mechanisms of consolidation, which occur in anatomically distinct brain regions (Sadanandappa, 2013).

The current observations on the absolute requirement for synapsin in STH but not LTH extend a model in which LTM can be encoded without transition through STM, particularly because STH and LTH involve different timescales of plasticity in the very same olfactory neurons. In the simple learning circuit for STH and LTH, the current observations are interpreted in a biochemical model. While both STH and LTH occur through potentiation of iLN-PN synapses, it is proposed that the signaling mechanisms for short- and long-term synaptic plasticity diverge before the stage of synapsin phosphorylation in GABAergic local interneurons. Thus, odorant exposure results in the activation of key signaling molecules such as PKA and CaMKII in LNs required for both forms of olfactory habituation. However, the kinases affect STH through synapsin phosphorylation, which results in a rapid but transient increase in evoked GABA release. Meanwhile, the same signaling molecules also participate in the activation of translational and/or transcriptional control machinery, which act, on a slower timescale, to cause the formation of additional GABAergic synapses that persist stably for much longer periods of time (Sadanandappa, 2013).

The unusual behavioral state, in which an animal is unable to form an STM but capable of longer term memory may have some clinical significance. Recently, a comprehensive mouse model for Down syndrome, which is trisomic for ~92% of the human chromosome 21, was shown to have defective STM of novel object recognition, while still showing normal LTM in this task, which is conceptually similar to behavioral habituation. Thus, it is suggested that further studies may eventually identify several other molecules which, like synapsin, are selectively necessary for short-term but not for long-term memory (Sadanandappa, 2013).

Go to top

Aging impairs protein-synthesis-dependent long-term memory in Drosophila

Tonoki, A. and Davis, R.L. (2015). J Neurosci 35: 1173-1180. PubMed ID: 25609631

Although aging is known to impair intermediate-term memory in Drosophila, its effect on protein-synthesis-dependent long-term memory (LTM) is unknown. This study shows that LTM is impaired with age, not due to functional defects in synaptic output of mushroom body (MB) neurons, but due to connectivity defects of dorsal paired medial (DPM) neurons with their postsynaptic MB neurons. GFP reconstitution across synaptic partners (GRASP) experiments revealed structural connectivity defects in aged animals of DPM neurons with MB axons in the α lobe neuropil. As a consequence, a protein-synthesis-dependent LTM trace in the α/β MB neurons fails to form. Aging thus impairs protein-synthesis-dependent LTM along with the α/β MB neuron LTM trace by lessening the connectivity of DPM and α/β MB neurons (Tonoki, 2015).

The data presented in this study offer several important findings about the neural circuitry and the forms of memory disrupted by aging. First, it shows that aging impairs only one of the two mechanistically distinct forms of LTM generated by spaced, aversive classical conditioning in Drosophila. LTM that is independent of protein synthesis remains unaffected by age, whereas that form of LTM requiring protein synthesis becomes impaired. Therefore, there is mechanistic specificity in the effects of aging on LTM. Although aging, in principal, could disrupt processes like protein synthesis at the molecular level leading to a LTM deficit, these results indicate that the problem is traceable to the circuitry involved in generating protein-synthesis-dependent LTM (Tonoki, 2015).

The normal synaptic transmission from DPM neurons onto follower neurons during spaced training that is required for generating LTM is lost with age. This is attributable to the reduction of synaptic contacts between DPM neuron processes and MB axons specifically in the tip of the α lobe neuropil as revealed by GRASP signals. The loss of synaptic contacts between DPM and MB neurons in this region also may explain why synaptic blockade of DPM neurons during acquisition disrupts protein-synthesis-dependent LTM in young but not old flies. Therefore, a second major finding is that neural contacts and subsequent synaptic activity between DPM and α/β MB neurons are required for generating protein-synthesis-dependent LTM, and aging impairs this process. Consistent with this model, it is found that aging blocks the formation of a calcium-based, protein-synthesis-dependent memory trace in the α/β MB neurons (Tonoki, 2015).

It was found previously that ITM is impaired in flies of 30 d of age along with the capacity to form an ITM trace in the DPM neurons. Nevertheless, aging does not compromise the capacity to form an STM trace in the α'/β' MB neurons. Therefore, aging disrupts specific temporal forms of memory, including ITM and protein synthesis LTM, but not STM and protein-synthesis-independent LTM. It is possible that the loss of connectivity of DPM neurons with the α tip neuropil is responsible for the loss of both ITM and protein-synthesis-dependent LTM, along with their respective memory traces. Previous and this study's data indicate that STM appears to bypass the DPM neurons, whereas the reciprocal activity between DPM and MB neurons is required for ITM and LTM. Aging puts a kink in this neural system by impairing connectivity (Tonoki, 2015).

The study offers a model to explain the neural circuitry involved in protein-synthesis-dependent LTM formation and how aging impairs this form of memory. Although DPM neurons make contacts widely throughout the MB lobe neuropil with processes of many cell types, the critical interaction for LTM formation occurs in the vertical lobes of the MB through contacts onto the axons of α/β MB neurons. DPM neuron synaptic activity during spaced training, which occurs due to their stimulation by MB neurons, promotes synaptic changes in the postsynaptic α/β MB neurons and leads to the formation of memory trace in the α/β MB neurons. Aging impairs protein-synthesis-dependent LTM along with a LTM trace that normally forms in the α/β MB neurons by lessening the connectivity of DPM and α/β MB neurons. Identifying the mechanisms by which the DPM neurons lose their connectivity with only the tips of α/β MB neurons might reveal how aging impairs protein-synthesis-dependent LTM (Tonoki, 2015).

Go to top

Acetylcholine deficit causes dysfunctional inhibitory control in an aging-dependent manner

Sabandal, P. R., Saldes, E. B. and Han, K. A. (2022). Acetylcholine deficit causes dysfunctional inhibitory control in an aging-dependent manner. Sci Rep 12(1): 20903. PubMed ID: 36463374

Inhibitory control is a key executive function that limits unnecessary thoughts and actions, enabling an organism to appropriately execute goal-driven behaviors. The efficiency of this inhibitory capacity declines with normal aging or in neurodegenerative dementias similar to memory or other cognitive functions. Acetylcholine signaling is crucial for executive function and also diminishes with aging. Acetylcholine's contribution to the aging- or dementia-related decline in inhibitory control, however, remains elusive. This was addressed in Drosophila using a Go/No-Go task that measures inhibition capacity. Inhibition capacity was reported to decline with aging in wild-type flies, which is mitigated by lessening acetylcholine breakdown and augmented by reducing acetylcholine biosynthesis. The mushroom body (MB) γ neurons were identified as a chief neural site for acetylcholine's contribution to the aging-associated inhibitory control deficit. In addition, it was found that the MB output neurons MBON-γ2α'1 having dendrites at the MB γ2 and α'1 lobes and axons projecting to the superior medial protocerebrum and the crepine is critical for sustained movement suppression per se. This study reveals, for the first time, the central role of acetylcholine in the aging-associated loss of inhibitory control and provides a framework for further mechanistic studies (Sabandal, 2022).

Go to top

Functional architecture of reward learning in mushroom body extrinsic neurons of larval Drosophila

Saumweber, T., Rohwedder, A., Schleyer, M., Eichler, K., Chen, Y. C., Aso, Y., Cardona, A., Eschbach, C., Kobler, O., Voigt, A., Durairaja, A., Mancini, N., Zlatic, M., Truman, J. W., Thum, A. S. and Gerber, B. (2018). Nat Commun 9(1): 1104. PubMed ID: 29549237

The brain adaptively integrates present sensory input, past experience, and options for future action. The insect mushroom body exemplifies how a central brain structure brings about such integration. This study used a combination of systematic single-cell labeling, connectomics, transgenic silencing, and activation experiments to study the mushroom body at single-cell resolution, focusing on the behavioral architecture of its input and output neurons (MBINs and MBONs), and of the mushroom body intrinsic APL neuron. The results reveal the identity and morphology of almost all of these 44 neurons in stage 3 Drosophila larvae. Upon an initial screen, functional analyses focusing on the mushroom body medial lobe uncover sparse and specific functions of its dopaminergic MBINs, its MBONs, and of the GABAergic APL neuron across three behavioral tasks, namely odor preference, taste preference, and associative learning between odor and taste. These results thus provide a cellular-resolution study case of how brains organize behavior (Saumweber, 2018).

Go to top

Elongator complex is required for long-term olfactory memory formation in Drosophila

Yu, D., Tan, Y., Chakraborty, M., Tomchik, S. and Davis, R. L. (2018). Learn Mem 25(4): 183-196. PubMed ID: 29545390

The evolutionarily conserved Elongator Complex associates with RNA polymerase II for transcriptional elongation. Elp3 is the catalytic subunit, contains histone acetyltransferase activity, and is associated with neurodegeneration in humans. Elp1 is a scaffolding subunit and when mutated causes familial dysautonomia. This study shows that elp3 and elp1 are required for aversive long-term olfactory memory in Drosophila. RNAi knockdown of elp3 in adult mushroom bodies impairs long-term memory (LTM) without affecting earlier forms of memory. RNAi knockdown with coexpression of elp3 cDNA reverses the impairment. Similarly, RNAi knockdown of elp1 impairs LTM and coexpression of elp1 cDNA reverses this phenotype. The LTM deficit in elp3 and elp1 knockdown flies is accompanied by the abolishment of a LTM trace, which is registered as increased calcium influx in response to the CS+ odor in the alpha-branch of mushroom body neurons. Coexpression of elp1 or elp3 cDNA rescues the memory trace in parallel with LTM. These data show that the Elongator complex is required in adult mushroom body neurons for long-term behavioral memory and the associated long-term memory trace (Yu, 2018).

Go to top

Mushroom body glycolysis is required for olfactory memory in Drosophila

Wu, C. L., Chang, C. C., Wu, J. K., Chiang, M. H., Yang, C. H. and Chiang, H. C. (2018). Neurobiol Learn Mem 150: 13-19. PubMed ID: 29477608

Glucose catabolism, also known as glycolysis, is important for energy generation and involves a sequence of enzymatic reactions that convert a glucose molecule into two pyruvate molecules. The glycolysis process generates adenosine triphosphate as a byproduct. This study investigated whether glycolysis plays a role in maintaining neuronal functions in the Drosophila mushroom bodies (MBs), which are generally accepted to be an olfactory learning and memory center. The data showed that individual knockdown of glycolytic enzymes in the MBs, including hexokinase (HexA), phosphofructokinase (Pfk), or pyruvate kinase (PyK), disrupts olfactory memory. Whole-mount brain immunostaining indicated that pyruvate kinase is strongly expressed in the MB alphabeta, α'β', and γ neuron subsets. It is concluded that HexA, Pfk, and PyK are required in each MB neuron subset for olfactory memory formation. The data therefore indicates that glucose catabolism in the MBs is important for olfactory memory formation in Drosophila (Wu, 2018).

Go to top

Two parallel pathways assign opposing odor valences during Drosophila memory formation

Yamazaki, D., Hiroi, M., Abe, T., Shimizu, K., Minami-Ohtsubo, M., Maeyama, Y., Horiuchi, J. and Tabata, T. (2018). Cell Rep 22(9): 2346-2358. PubMed ID: 29490271

During olfactory associative learning in Drosophila, odors activate specific subsets of intrinsic mushroom body (MB) neurons. Coincident exposure to either rewards or punishments is thought to activate extrinsic dopaminergic neurons, which modulate synaptic connections between odor-encoding MB neurons and MB output neurons to alter behaviors. However, this study identifies two classes of intrinsic MB γ neurons based on cAMP response element (CRE)-dependent expression, γCRE-p and γCRE-n, which encode aversive and appetitive valences. γCRE-p and γCRE-n neurons act antagonistically to maintain neutral valences for neutral odors. Activation or inhibition of either cell type upsets this balance, toggling odor preferences to either positive or negative values. The mushroom body output neurons, MBON-;gamma'5β'2a/&beta'2mp and MBON-γ2α'1, mediate the actions of γCRE-p and γCRE-n neurons. The data indicate that MB neurons encode valence information, as well as odor information, and this information is integrated through a process involving MBONs to regulate learning and memory (Yamazaki, 2018).

Go to top

The interruptive effect of electric shock on odor response requires mushroom bodies in Drosophila melanogaster

Song, W., Zhao, L., Tao, Y., Guo, X., Jia, J., He, L., Huang, Y., Zhu, Y., Chen, P. and Qin, H. (2018). Genes Brain Behav: e12488. Pubmed ID: 29808570

Nociceptive stimulus involuntarily interrupts concurrent activities. This interruptive effect is related to the protective function of nociception that is believed to be under stringent evolutionary pressure. Background noxious electric shock (ES) dramatically interrupts Drosophila odor response behaviors in a T-maze, termed blocking odor-response by electric-shock (BOBE). ES could interrupt both odor avoidance and odor approach. To identify involved brain areas, focus placed on the odor avoidance to 3-OCT. By spatially abolishing neurotransmission with temperature sensitive Shibire(TS1), This study found that mushroom bodies (MBs) are necessary for BOBE. Among the three major MB Kenyon cell (KCs) subtypes, α/β neurons and γ neurons but not α'/β' neurons are required for normal BOBE. Specifically, abolishing the neurotransmission of either α/β surface (α/&betas;), α/β core (α/βc) or γ dorsal (γd) neurons alone is sufficient to abrogate BOBE. This pattern of MB subset requirement is distinct from that of aversive olfactory learning, indicating a specialized BOBE pathway. Consistent with this idea, BOBE wasn't diminished in several associative memory mutants and noxious ES interrupted both innate and learned odor avoidance. Overall, These results suggest that MB α/β and γ neurons are parts of a previously unappreciated central neural circuit that processes the interruptive effect of nociception (Song, 2018).

Go to top

Calcium in Kenyon cell somata as a substrate for an olfactory sensory memory in Drosophila

Ludke, A., Raiser, G., Nehrkorn, J., Herz, A. V. M., Galizia, C. G. and Szyszka, P. (2018). Front Cell Neurosci 12: 128. PubMed ID: 29867361

Animals can form associations between temporally separated stimuli. To do so, the nervous system has to retain a neural representation of the first stimulus until the second stimulus appears. The neural substrate of such sensory stimulus memories is unknown. This study searched for a sensory odor memory in the insect olfactory system and characterized odorant-evoked Ca(2+) activity at three consecutive layers of the olfactory system in Drosophila: in olfactory receptor neurons (ORNs) and projection neurons (PNs) in the antennal lobe, and in Kenyon cells (KCs) in the mushroom body. The post-stimulus responses in ORN axons, PN dendrites, PN somata, and KC dendrites are odor-specific, but they are not predictive of the chemical identity of past olfactory stimuli. However, the post-stimulus responses in KC somata carry information about the identity of previous olfactory stimuli. These findings show that the Ca(2+) dynamics in KC somata could encode a sensory memory of odorant identity and thus might serve as a basis for associations between temporally separated stimuli (Ludke, 2018).

Go to top

Communication from learned to innate olfactory processing centers is required for memory retrieval in Drosophila

Dolan, M. J., Belliart-Guerin, G., Bates, A. S., Frechter, S., Lampin-Saint-Amaux, A., Aso, Y., Roberts, R. J. V., Schlegel, P., Wong, A., Hammad, A., Bock, D., Rubin, G. M., Preat, T., Placais, P. Y. and Jefferis, G. (2018). Neuron. PubMed ID: 30244885

Communication from learned to innate olfactory processing centers is required for memory retrieval in Drosophila

The behavioral response to a sensory stimulus may depend on both learned and innate neuronal representations. How these circuits interact to produce appropriate behavior is unknown. In Drosophila, the lateral horn (LH) and mushroom body (MB) are thought to mediate innate and learned olfactory behavior, respectively, although LH function has not been tested directly. This study identified two LH cell types (PD2a1 and PD2b1) that receive input from an MB output neuron required for recall of aversive olfactory memories. These neurons are required for aversive memory retrieval and modulated by training. Connectomics data demonstrate that PD2a1 and PD2b1 neurons also receive direct input from food odor-encoding neurons. Consistent with this, PD2a1 and PD2b1 are also necessary for unlearned attraction to some odors, indicating that these neurons have a dual behavioral role. This provides a circuit mechanism by which learned and innate olfactory information can interact in identified neurons to produce appropriate behavior (Dolan, 2018).

Go to top

Drosophila SLC22A transporter is a memory suppressor gene that influences cholinergic neurotransmission to the mushroom bodies

Gai, Y., Liu, Z., Cervantes-Sandoval, I. and Davis, R.L. (2016). Neuron 90: 581-595. PubMed ID: 27146270

The mechanisms that constrain memory formation are of special interest because they provide insights into the brain's memory management systems and potential avenues for correcting cognitive disorders. RNAi knockdown in the Drosophila mushroom body neurons (MBn) of a newly discovered memory suppressor gene, Solute Carrier DmSLC22A, a member of the organic cation transporter family, enhances olfactory memory expression, while overexpression inhibits it. The protein localizes to the dendrites of the MBn, surrounding the presynaptic terminals of cholinergic afferent fibers from projection neurons (Pn). Cell-based expression assays show that this plasma membrane protein transports cholinergic compounds with the highest affinity among several in vitro substrates. Feeding flies choline or inhibiting acetylcholinesterase in Pn enhances memory, an effect blocked by overexpression of the transporter in the MBn. The data argue that DmSLC22A is a memory suppressor protein that limits memory formation by helping to terminate cholinergic neurotransmission at the Pn:MBn synapse (Gai, 2016).

Genetic studies have now identified hundreds of genes required for normal memory formation. Some of these genes regulate the development of the cells and circuits required for learning; some mediate the physiological changes that occur with acquisition and storage. Of particular interest are gene functions that suppress normal memory formation and, by analogy with tumor suppressor genes, are referred to as memory suppressor genes. These genes and their products can, in principle, suppress memory formation by antagonizing the process of acquisition, limiting memory consolidation, promoting active forgetting, or inhibiting retrieval. Recently, a large RNAi screen of ∼3,500 Drosophila genes has bee carried out, and several dozen new memory suppressor genes were identified (Walkinshaw, 2015), identified as such, because RNAi knockdown produces an enhancement in memory performance after olfactory conditioning (Gai, 2016).

Aversive olfactory classical conditioning is a well-studied type of learning in Drosophila and consists of learning a contingency between an odor conditioned stimulus (CS) and most often an unconditioned stimulus (US) of electric shock. Many cell types in the olfactory nervous system are engaged in this type of learning, including antennal lobe projection neurons (Pn), several different types of mushroom body neurons (MBn), dopamine neurons (DAn), and others, but a focused model of olfactory memory formation holds that MBn are integrators of CS and US information with the CS being conveyed to the MBn dendrites by the axons of cholinergic, excitatory Pn of the antennal lobe, and the US conveyed to the MBn by DAn Gai, 2016).

A memory suppressor gene identified and describe in this report encodes a member of the SLC22A transporter family. The Solute Carrier (SLC) family of transporters in humans consists of 395 different, membrane-spanning transporters that have been organized into 52 different families. Some of these are localized pre-synaptically and involved in neurotransmitter recycling, others localize to glia for clearance of neurotransmitter from the synapse. In addition, glutamate transporters can be localized post-synaptically to regulate neurotransmission strength via clearance mechanisms. Some of these SLC transporters have prominent roles in neurological and psychiatric disorders and in drug design, including SLC1A family members that are responsible for glutamate uptake and clearance of this neurotransmitter from the synaptic cleft and SLC6A2-4 proteins that transport monoamines into cells. Inhibitors of these proteins, which include the serotonin-specific reuptake inhibitors (SSRIs) and serotonin-noradrenaline reuptake inhibitors (SNRIs), increase monoamine dwell time at the synapse and are used to treat depression and several other neuropsychiatric disorders (Gai, 2016).

The SLC22A family of transporters is distinguished into two major classes that carry either organic cations (SLC22A1-5, 15, 16, and 21) or anions (SLC22A6-13 and 20) across the plasma membrane, with generally low substrate binding affinity and high capacity. They transport numerous molecules with diverse structures, including drugs, acetylcholine, dopamine, histamine, serotonin, and glycine among others. Ergothioneine has been identified as a high-affinity substrate for SLC22A4 and spermidine for SLC22A16. Mice mutant in the two organic cation transporters, SLC22A2 and SLC22A3, exhibit behavioral phenotypes suggestive of functions in anxiety, stress, and depression. These observations point out the importance of the SLC22A family for brain function and cognition. Recently, a Drosophila SLC22A family member, CarT (CG9317) was identified and found to transport carcinine into photoreceptor neurons for the recovery of essential visual neurotransmitter histamine (Gai, 2016).

This study shows that the Drosophila gene, CG7442, functions as a memory suppressor gene and is a member of the SLC22A family. This transporter is expressed most abundantly in the dendrites of the MBn, at the synapses with the cholinergic antennal lobe Pn. Cell-based expression assays show that Drosophila SLC22A transports choline and acetylcholine with the highest affinity among several substrates. Pharmacological and genetic data support the model that Drosophila SLC22A functions at the Pn:MBn synapse to terminate cholinergic neurotransmission, differing from well-characterized presynaptic choline transporters for neurotransmitter recycling, and mechanistically explaining its role in behavioral memory suppression (Gai, 2016).

These data connect the SLC22A family of transporters and memory suppression. DmSLC22A, located on the dendrites of the adult α/β and α'/β' MBn, removes ACh from the Pn:MBn synapses in the calyx. The normal expression level of this plasma membrane transporter limits the transference of olfactory information to the MBn by removing neurotransmitter from the synapse. Overexpression of DmSLC22A hardens this limit, weakening the CS representation and weakening memory formation. Reducing DmSLC22A expression has the opposite effect of softening the limit, producing a stronger CS representation and stronger memory formation. Thus, the data indicate that acetylcholinesterase and postsynaptic SLC22A transporter function jointly to regulate neurotransmitter persistence at the synapse. This conclusion is notable, given the longstanding emphasis on ACh degradation as the primary route for termination of the cholinergic synaptic signal. Although the evidence is strong for the proposed mechanism shown in Figure 8A, the transporter exhibits broad substrate specificity and expression outside of the Pn:MBn synapse. Alternative or additional mechanisms of action in memory suppression thus remain a possibility (Gai, 2016).

The current data are consistent with the model that ACh persistence at the Pn:MBn synapse is a surrogate for the strength of the CS and therefore a primary effector of olfactory memory strength. Other data similarly point to the strength of stimulation of MBn as an important variable for regulating memory strength. The MBn also receive inhibitory input through GABAA receptors expressed on the MBn. Overexpressing the MBn-expressed GABAA receptor Rdl impairs learning, while RNAi knockdown of this receptor in the MBn enhances memory formation. This regulation of memory strength is independent of the US pathway involved in classical conditioning, functioning similarly for both aversive and appetitive USs. However, it is noted that the in vivo functions for the SLC22A class of transporters must be broader than the focused model presented above. For instance, the data indicate that the Drosophila SLC22A protein transports both acetylcholine and dopamine in ex vivo preparations. Moreover, the protein's memory suppressor function maps to both MBn and the DAn. How DmSLC22A might function in DAn to suppress memory formation has not been explored, but one reasonable hypothesis is that DmSLC22A transports acetylcholine at the synapse between upstream and putative cholinergic neurons that provide input to the DAn that convey the US in classical condition. Testing this hypothesis requires identifying the presynaptic neurons to the DAn that carry the US information (Gai, 2016).

One unexplained observation is that although DmSLC22A knockdown enhances the duration of memory produced from stronger memory traces instilled at acquisition, it slows the rate of acquisition as measured by acquisition curves. However, this observation has been made with another memory suppressor gene as well. A knockdown of the pre- and post-synaptic scaffolding protein, Scribble, has the same effect of producing more enduring memories but slowing acquisition. In addition, similar observations have been made in mouse: injection of muscarinic acetylcholine receptor antagonists impairs memory acquisition but enhances retention (Easton, 2012) (Gai, 2016).

These studies bring a focus on the SLC22A family of plasma membrane transporters as potential targets for neurotherapeutics. Of the 24 members of this family, only a few have been studied in some detail in the nervous system. RNA expression experiments have shown that SLC22A1-5 are all expressed in the brain, with SLC22A3 and A4 being the most abundant, and immunohistochemistry experiments have revealed that SLC22A4-5 are localized at dendrites within the hippocampus. Mammalian members of this family of transporters and, by extension, probably DmSLC22A, are subject to regulation by multiple signaling molecules including protein kinase A, calcium/calmodulin-dependent protein kinase II, and the mitogen-activated protein kinases. Knockout mice for SLC22A2 and A3 show reduced basal level of several neurotransmitters in a region-dependent manner and decreased anxiety-related behaviors, although the effects of SLC22A3 on anxiety-related behaviors is debated. In addition, the knockouts or antisense insults reveal behavioral changes in depression-related tasks, with SLC22A2 knockouts exhibiting increased behavioral despair, and SLC22A3 antisense-treated animals exhibiting decreased behavioral despair. Little is known about the biological or behavioral functions of the other members of the SLC22A family. The current results show that the SLC22A family of transporters is also involved in memory suppression (Gai, 2016).

DmSLC22A is a unique and new type of memory suppressor gene. There are, to date, about two dozen memory suppressor genes identified in the mouse and about three dozen such genes in Drosophila. The mechanisms by which all of these genes suppress memory formation are not yet known, but a few themes have emerged. For instance, several of the genes suppress memory formation by limiting excitatory neurotransmitter release and function, or the expression and function of post-synaptic receptors. DmSLC22A appears to fall into this category. Another example is Cdk5, which negatively influences the expression of NR2B and limits memory formation. Knockouts of some GABA receptors reduce inhibitory tone of learning circuitry so as to facilitate memory formation. Several of the known memory suppressor genes are known to function in active forgetting processes. These include damb, a dopamine receptor involved in forgetting mechanisms; scribble, a pre- and post-synaptic scaffolding gene; and rac, a small G protein involved in the biochemistry of active forgetting. Memory suppressor genes can also encode signaling molecules that negatively regulate transcription factors required for long-term memory and the transcription factors themselves, such as repressing isoforms of Aplysia Creb; ATF4, a transcription factor homolgous to ApCreb-2; and protein phosphatase I. Elucidating all of the genetic constraints on memory formation and their mechanisms will have profound consequences for understanding of how the brain forms and stores memories and for the development of cognitive therapeutics (Gai, 2016).

Go to top

Long-term memory engram cells are established by c-Fos/CREB transcriptional cycling

Miyashita, T., Kikuchi, E., Horiuchi, J. and Saitoe, M. (2018). Cell Rep 25(10): 2716-2728. PubMed ID: 30517860

Training-dependent increases in c-fos have been used to identify engram cells encoding long-term memories (LTMs). However, the interaction between transcription factors required for LTM, including CREB and c-Fos, and activating kinases such as phosphorylated ERK (pERK) in the establishment of memory engrams has been unclear. Formation of LTM of an aversive olfactory association in flies requires repeated training trials with rest intervals between trainings. This study finds that prolonged rest interval-dependent increases in pERK induce transcriptional cycling between c-Fos and CREB in a subset of KCs in the mushroom bodies, where olfactory associations are made and stored. Preexisting CREB is required for initial c-fos induction, while c-Fos is required later to increase CREB expression. Blocking or activating c-fos-positive engram neurons inhibits memory recall or induces memory-associated behaviors. These results suggest that c-Fos/CREB cycling defines LTM engram cells required for LTM (Miyashita, 2018).

This study has found that activation of CREB is only part of a c-Fos/CREB cycling program that occurs in specific cells to generate memory engrams. Previous studies have shown that LTM is encoded in a subset of neurons that are coincidently activated during training. The data suggest that these coincidently activated neurons differ from other neurons because they activate c-Fos/CREB cycling, which then likely induces expression of downstream factors required for memory maintenance. Thus, memory engram cells can be identified by the colocalization of c-Fos, CREB, and pERK activities. Inhibiting synaptic outputs from these neurons suppresses memory-associated behaviors, while artificial activation of these neurons induces memory-based behaviors in the absence of the conditioned stimulus (Miyashita, 2018).

The importance of rest intervals during training for formation of LTM is well known. 10x spaced training produces LTM in flies, while 48x massed trainings, which replace rest intervals with further training, does not. It has been shown that pERK is induced in brief waves after each spaced training trial, and it has been proposed that the number of waves of pERK activity gates LTM formation. While the current results are generally consistent with previous studies, this study found that LTM is formed after 48x massed training in CaNB2/+ and PP1/+ flies, which show sustained pERK activity instead of wave-like activity. Thus, it is suggest that either sustained pERK activity or several bursts of pERK activity are required, first to activate endogenous CREB, then to activate induced c-Fos, and later to activate induced CREB (Miyashita, 2018).

In this study, 10x massed training of CaNB2/+ flies produces an intermediate form of protein synthesis-dependent LTM that declines to baseline within 7 days. This result is consistent with results from a previous study, which identified two components of LTM: an early form that decays within 7 days and a late form that lasts more than 7 days. 10x massed training takes the same amount of time as 3x spaced training, which is insufficient to produce 7-day LTM and instead produces only the early form of LTM from preexisting dCREB2. It is proposeed that long-lasting LTM requires increased dCREB2 expression generated from c-Fos/CREB cycling. This increased dCREB2 expression allows engram cells to sustain expression of LTM genes for more than 7 days (Miyashita, 2018).

Although it is proposed that c-Fos/CREB cycling forms a positive feedback loop, this cycling does not result in uncontrolled increases in c-Fos and dCREB2. Instead, spaced training induces an early dCREB2-dependent increase in c-fos and other LTM-related genes, and subsequent c-Fos/CREB cycling maintains this increase and sustains LTM. It is believed that c-Fos/CREB cycling does not cause uncontrolled activation, because dCREB2 activity depends on an increase in the ratio of activator to repressor isoforms. The data indicate that splicing to dCREB2 repressor isoforms is delayed relative to expression of activator isoforms, leading to a transient increase in the activator-to-repressor ratio during the latter half of spaced training. However, the ratio returns to basal by the 10th training cycle, suggesting that the splicing machinery catches up to the increase in transcription. The transience of this increase prevents uncontrolled activation during c-Fos/CREB cycling and may explain the ceiling effect observed in which training in excess of 10 trials does not further increase LTM scores or duration (Miyashita, 2018).

Why does ERK activity increase during rest intervals, but not during training? ERK is phosphorylated by MEK, which is activated by Raf. Amino acid homology with mammalian B-Raf suggests that Drosophila Raf (DRaf) is activated by cAMP-dependent protein kinase (PKA) and deactivated by CaN. The current results indicate that ERK activation requires D1-type dopamine receptors and rut-AC, while a previous study demonstrates that ERK activation also requires Ca2+ influx through glutamate NMDA receptors. Thus, training-dependent increases in glutamate and dopamine signaling may activate rut-AC, which produces cAMP and activates PKA. PKA activates the MAPK pathway, resulting in ERK phosphorylation. At the same time, training-dependent increases in Ca2+/CaM activate CaN and PP1 to deactivate MEK signating in increased ERK activation during the rest interval after training (Miyashita, 2018).

This study examined the role of ERK phosphorylation and activation in LTM and did not observe significant effects of ERK inhibition in short forms of memory. However, a previous study reported that ERK suppresses forgetting of 1-hr memory, suggesting that ERK may have separate functions in regulating STM and LTM. c-Fos/CREB cycling distinguishes engram cells from non-engram cells, and it is suggested that this cycling functions to establish and maintain engrams. However, studies in mammals indicate that transcription and translation after fear conditioning is required for establishing effective memory retrieval pathways instead of memory storage. Thus, c-Fos/CREB cycling may be required for establishment and maintenance of engrams or for retrieval of information from engrams (Miyashita, 2018).

The engram cells identified in this study consist of α/β KCs, a result consistent with previous studies demonstrating the importance of these cells in LTM. Although some α/β neurons are seen expressing high amounts of dCREB2 in naive and massed trained animals (6.5% ± 0.5% of pERK-positive cells in massed trained animals), few c-fos-positive cells are seen and no overlap between c-fos expression and dCREB2 in these animals. After spaced training, the percentage of cells that express both c-fos and dCREB2 jumps to 18.9% ± 1.2%, and these cells fulfill the criteria for engram cells, because they are reactivated upon recall and influence memory-associated behaviors. The phosphatase pathway may predominate during training, inhibiting ERK phosphorylation. However, phosphatase activity may deactivate faster at the end of training compared to the Rut/PKA activity, While some mammalian studies suggest that neurons that express high amounts of CREB are preferentially recruited to memory engrams, this study found that the percentage of neurons that express high dCREB2 and low c-fos remains relatively unchanged between massed trained and spaced trained flies. Furthermore, this study finds that the increase in neurons expressing high amounts of dCREB2 after spaced training corresponds to the increase in c-Fos/CREB cycling engram cells. Thus, in flies, LTM-encoding engram cells might not be recruited from cells that previously expressed high amounts of dCREB2 but instead may correspond to cells in which c-Fos/CREB cycling is activated by coincident odor and shock sensory inputs (Miyashita, 2018).

Go to top

CREBB repression of protein synthesis in mushroom body gates long-term memory formation in Drosophila

Lin, H. W., Chen, C. C., Jhang, R. Y., Chen, L., de Belle, J. S., Tully, T. and Chiang, A. S. (2022). Proc Natl Acad Sci U S A 119(50): e2211308119. PubMed ID: 36469774

Learned experiences are not necessarily consolidated into long-term memory (LTM) unless they are periodic and meaningful. LTM depends on de novo protein synthesis mediated by cyclic AMP response element-binding protein (CREB) activity. In Drosophila, two creb genes (crebA, crebB) and multiple CREB isoforms have reported influences on aversive olfactory LTM in response to multiple cycles of spaced conditioning. How CREB isoforms regulate LTM effector genes in various neural elements of the memory circuit is unclear, especially in the mushroom body (MB), a prominent associative center in the fly brain that has been shown to participate in LTM formation. This study reports that 1) spaced training induces crebB expression in MB α-lobe neurons and 2) elevating specific CREBB isoform levels in the early α/β subpopulation of MB neurons enhances LTM formation. By contrast, learning from weak training 3) induces 5-HT1A serotonin receptor synthesis, 4) activates 5-HT1A in early α/β neurons, and 5) inhibits LTM formation. 6) LTM is enhanced when this inhibitory effect is relieved by down-regulating 5-HT1A or overexpressing CREBB. These findings show that spaced training-induced CREBB antagonizes learning-induced 5-HT1A in early α/β MB neurons to modulate LTM consolidation (Lin, 2022).

Recurrent spaced learning has been shown to relieve inhibition and gate LTM formation in animal models. However, gene regulatory mechanisms that act to filter relevant signals of repeated events and override inhibitory constraints in identified circuit elements remain unknown. The current data suggest that MB neurons in Drosophila provide a compelling cellular gating mechanism for LTM formation. Weak learning is sufficient to increase 5-HT1A synthesis in early α/β neurons, and these neurons produce a downstream inhibitory effect on LTM formation. After spaced training, CREBB expression represses further 5-HT1A synthesis, thereby relieving the inhibitory effect on LTM formation. These conclusions are supported by several lines of evidence: i) CREBB transcription increased after 5xS or 10xS but not after 1x (Fig. 1); and ii) RNAi-mediated knockdown of CREBB in α/β impaired LTM (Fig. 1), while overexpression of a crebB-a or crebB-c transgene enhanced LTM. iii) Conversely, RNAi-mediated knockdown of 5-HT1A in early α/β neurons enhanced LTM, while overexpression of a 5-HT1A transgene impaired LTM; and iv) 1x was sufficient to activate 5-HT1A, and this activation was inhibited by expression of CREBB proteins. v) Furthermore, overexpression of 5-HT1A-mediated LTM impairment was fully rescued by CREBB overexpression. Together, these findings suggest that synthesis of 5-HT1A and CREBB proteins in response to training operate like an opposing molecular switch to inhibit or disinhibit downstream LTM formation, respectively (Lin, 2022).

Previous reports suggested that expression of a chimeric CREBB-a transcriptional activator and a CREBB-b transcriptional repressor throughout whole fly enhanced and impaired LTM formation, respectively. Subsequently, CREBB-a-dependent enhancement of LTM was not observed using a hs-Gal4 driver that has low expression in MB. Chronic expression of a CREBB-b in all α/β neurons was shown to impair 1-d memory after spaced training. It has been documented, however, that these chronic disruptions of CREBB-b produced developmental abnormalities in MB structure. In contrast, acute induced expression of CREBB-b only in adult α/β neurons did not impair 1-d memory after spaced training (and did not produce structural defects). Using a different inducible system (MB247-Switch) to acutely expresses CREBB-b in γ and α/β neurons showed a mild impairment of 1-d memory after spaced training. More interestingly, various molecular genetic tools were used to show that interactions among CREBB, CREB-binding protein, and CREB-regulated transcription coactivator in MB were clearly involved in LTM formation or maintenance, respectively. Using the same inducible gene switch tool, a positive regulatory loop has been shown between Fos and CREBB in MB during LTM formation - but that study did not show behavioral data pertaining to manipulation of CREBB per se - nor did that study restrict experiments to early α/β neurons (Lin, 2022).

In another study CRE-luciferase transgene was expressed in different subpopulations of MB neurons and then monitored luciferase activity in live flies at various times after spaced training. Immediately after spaced training, some patterns of luciferase expression decreased (OK107 expressing in all MB neurons; c739 expressing in all α/β neurons; 1471 expressing in γ neurons), or increased (c747 and c772 expressing variably in all MB neurons), or showed no detectable change (c320 expressing variably in γ, α'/β' and α/β subpopulation, 17d expressing primarily in late α/β and in early α/β neurons). Indeed, the Zhang paper pointed out that, because the CRE-reporter was expressed in more than one subpopulation of MB neurons, only net effects of CREB function could be quantified. Furthermore, this study did not elucidate which CREBB isoforms might increase or decrease after spaced training. Obviously, this information would be critical if different isoforms have opposing activator and repressor functions in specific MB neuron subpopulations. The current study provides a dramatic example of this point. By restricting manipulation to early α/β neurons in adult stage animals, this study showed that enhanced LTM formation after acute CREBB-c overexpression is comparable to the net effect of chimeric CREBB-a overexpression in whole flies, and that spaced training serves to increase the expression of CREBB in these early α/β neurons (Lin, 2022).

It has been reported that the CREBB-a isoform functions as a PKA-responsive transcriptional activator and the CREBB-b isoform functions as a repressor of CREBB-a-induced gene activation. Using new KAEDA synthesis as a reporter for temporal gene activation, it has been previously shown that CREBB-b in DAL neurons represses CREBA-mediated gene activation to inhibit LTM formation. In early α/β MB neurons, KAEDA experiments indicate that CREBB-a and CREBB-c, but not CREBB-b, both repress 5-HT1A-mediated inhibition to gate LTM formation. These findings demonstrate a neuron- and training-specific CREBA activation and CREBB repression of effecter genes involved in modulating LTM formation. Although crebB promoter-driven Gal4 expression, crebBRNAi downregulation, and cell-type specific transcriptomes show CREBB expression in early α/β neurons, it remains unclear whether specific naturally occurring CREBB isoforms in these neurons serve to modulate LTM formation (Lin, 2022).

How is the learning-induced LTM gating mechanism differentially regulated by different [1x, 10xM (ten massed cycles of training without rest intervals) or 10xS (spaced trials)] training protocols? Expression of both 5-HT1A and crebB in early α/β MB neurons was elevated 24 h after 10xS, whereas only 5-HT1A was induced after 1x, and neither gene was induced after 10xM. Why is elevated 5-HT1A seen after 10xS, when constitutive expression of CREBB proteins suppresses 5-HT1A expression? A possible explanation is that 5-HT1A may be normally activated as an early response to 1x, whereas crebB induction by 10xS is not evident for about 3 h. Gradual cessation of 5-HT1A transcription by the delayed 10xS-induced CREBB expression may account for lower KAEDE levels observed in one odor/shock pairing experiment. Interestingly, the data showed that even with elevated 5-HT1A, CREBB proteins can still enhance 1-d memory, suggesting that CREBB-mediated inhibition is rather complex (Lin, 2022).

Massed training appears not to activate or suppress learning-induced transcriptional activity in early α/β neurons, and 5-HT1A nor crebB is activated after 10xM. Nevertheless, massed training may antagonize LTM formation. For instance, in MB neurons, spaced training induces repetitive waves of Ras/mitogen-activated protein kinase (MAPK) activity, activates MAPK translocation to the nucleus mediated by importin-7 (29), increases CREBB expression and, in dorsal-anterior-lateral (DAL) neurons, training induces activity-dependent crebA, CamKII, and per gene expression - all of which are not activated after massed training. These notions above suggest that massed training produces a more upstream general suppression of these 1x- and 10xS-induced genes required for inhibitory/gating mechanisms allocated in MB and DAL neurons, respectively (Lin, 2022).

An LTM enhancing role associated with CREBB expression and protein synthesis inhibition is a novel aspect of this gating mechanism. A previous study showed that inhibition of protein synthesis in MB after strong spaced training did not reduce LTM. Since it would not be possible to detect enhanced performance in these experiments, the possibility cannot be excluded that this inhibition might eliminate downregulation of LTM effector genes, with a net effect of promoting the formation of LTM rather than impairing it. This study estalished that synthesis of new 5-HT1A proteins in early α/β neurons after weak learning provides negative regulation and produces a downstream inhibitory effect on LTM formation. Surprisingly, CREBB protein synthesis in early α/β neurons after strong spaced training provides positive regulation by antagonizing this negative effect of 5-HT1A on LTM . Thus, CREBB-mediated repression is equivalent to the net effect of blocking protein synthesis in MB. Both relieve downstream inhibition and enhance rather than impair LTM formation. It is proposed that CREBB-mediated inhibition operates both directly by repressing gene transcription and indirectly through activating their downstream translational suppression (Lin, 2022).

Together, these experiments uncover a biochemical LTM gating mechanism that requires delicate regulation of protein synthesis and repression after training within identified neurons. More broadly, these observations also highlight the need to confirm the regulatory functions of specific CREB isoforms in identified neuronal subtypes before making conclusions about their roles in LTM formation (Lin, 2022).

The discovery that molecules in early α/β neurons inhibit LTM formation is relevant to future studies. Another persistent anesthesia-resistant form of memory (ARM) is also mediated by α/β neurons and has been shown to inhibit LTM formation. 5-HT1A appears to be a key protein involved in both ARM and LTM. Furthermore, the interaction of serotonin released from dorsal paired medial neurons and 5-HT1A in α/β neurons is necessary for sleep. CREBB expression in MB is also under circadian regulation, which together suggests mechanistic links between ARM, LTM, sleep, and circadian timing in early α/β neurons (Lin, 2022).

Go to top

Stromalin constrains memory acquisition by developmentally limiting synaptic vesicle pool size

Phan, A., Thomas, C. I., Chakraborty, M., Berry, J. A., Kamasawa, N. and Davis, R. L. (2018). Neuron. PubMed ID: 30503644

Stromalin, a cohesin complex protein, was recently identified as a novel memory suppressor gene, but its mechanism remained unknown. This study shows that Stromalin functions as a negative regulator of synaptic vesicle (SV) pool size in Drosophila neurons. Stromalin knockdown in dopamine neurons during a critical developmental period enhances learning and increases SV pool size without altering the number of dopamine neurons, their axons, or synapses. The developmental effect of Stromalin knockdown persists into adulthood, leading to strengthened synaptic connections and enhanced olfactory memory acquisition in adult flies. Correcting the SV content in dopamine neuron axon terminals by impairing anterograde SV trafficking motor protein Unc104/KIF1A rescues the enhanced-learning phenotype in Stromalin knockdown flies. These results identify a new mechanism for memory suppression and reveal that the size of the SV pool is controlled genetically and independent from other aspects of neuron structure and function through Stromalin (Phan, 2018).

Learning and memory are tightly regulated processes that require the activity of hundreds of genes to orchestrate the proper development of neural circuits and the underlying physiological changes necessary for cellular and synaptic plasticity. While many genes are known that define mechanisms required for the formation and consolidation of memory, far less is known about the genetic factors that constrain memory formation and their molecular and cellular mechanisms. Important conceptual insights about memory formation might come from elucidating the cellular mechanisms underlying this class of genetic element. Memory suppressor genes, named so by analogy to tumor suppressor genes, could, in principal, function by limiting memory acquisition, consolidation, or retrieval or by participating in active forgetting processes (Phan, 2018).

Several dozen and novel memory suppressor genes were recently identified in a large RNAi screen for effects on 3 hr aversive olfactory memory expression in Drosophila) They were classed as such because knockdown of these genes led to increased memory expression. A cohesin complex member, stromalin, was one such gene identified in the screen. The highly conserved cohesin complex is comprised of Stromalin (STAG1/2 in mammals) and three other subunits named structural maintenance of chromosomes 1 (SMC1), SMC3, and Rad21 (Phan, 2018).

Although the complex was first identified for its role in the proper segregation of chromosomes during cell, evidence has emerged showing that the complex has other important biological functions. Cohesin complex mRNAs and proteins are present at moderate to high levels in both the Drosophila and mouse nervous systems, revealing potential roles beyond chromosome segregation. Elegant studies from two research groups have clearly shown that members of the complex have a post-mitotic role in the proper pruning of axons in Drosophila mushroom body neurons. Other studies have provided evidence for roles in gene expression, DNA repair, and cancer susceptibility. It is notable that absent from this list are clear and specific roles for the complex in learning and memory processes, other than the general cognitive disturbances observed in humans with cohesinopathies (Phan, 2018).

This study reports that RNAi knockdown of Stromalin in mushroom body and dopamine neurons leads to enhanced aversive olfactory memory in adult flies. stromalin functions during development as a negative regulator of both synaptic and dense core vesicle (DCV) number in the nervous system, limiting the strength of synaptic connections to suppress memory acquisition. Reducing Stromalin levels specifically increases the number of vesicles in neurons without detectably altering other features of the targeted neurons in adult flies, including synapse number, synapse volume, or neurite branching. These observations offer evidence that the size of the synaptic vesicle pool is regulated independently of other structural features of the neuron (Phan, 2018).

Memory suppressor genes offer a unique window for understanding the molecular and cellular mechanisms that constrain memory formation. In contrast to the many genes and gene products known to be required for acquisition and memory consolidation, there are but a handful of memory suppressor genes studied to the point of providing new conceptual insights into the processes of memory formation. Some function at the transcriptional level to control the formation of protein synthesis-dependent long-term memory (LTM). For instance, isoforms of the Creb transcription factor (Creb repressors) exist that inhibit the normal function of Creb activators to limit LTM. These are thought to function after initial memory acquisition through biochemical cascades that mobilize new protein synthesis required for LTM. Other memory suppressor genes actively repress communication between neurons. For instance, Drosophila SLC22A encodes a plasma membrane transporter that removes neurotransmitter from the synaptic cleft to terminate synaptic communication. Cyclin-dependent kinase 5 (Cdk5) promotes proteolysis of the NMDA receptor subunit NR2B, attenuating NMDA receptor signaling in mammalian neurons. It is notable that these and other previously described memory suppressor genes limit the memory capacity of adult organisms, while developmental negative regulation of adult memory is rare. Stromalin is unique, acting during a critical developmental window to constrain the strength of synaptic communication between neurons by limiting the size of the synaptic vesicle pool. This phenotype then persists into adult life (Phan, 2018).

The data argue that Stromalin regulates the SVs in DAn independent of other structural features of the neuron, such as cell number, the apparent ramification of DAn neuropil in the MB, and synapse number or size. These observations lead to the novel and important conclusion that the SV pool is under its own genetic regulation through Stromalin function. Prior to these results, SV pool size was thought to be a function of synapse or active zone size or some other aspect of neuronal morphology (Phan, 2018).

Surprisingly, Stromalin alters the number of both SVs and DCVs, suggesting that it has a shared role in the biosynthetic or degradative pathways for both types of vesicles that are distinct from the piccolo-bassoon transport vesicles that contain active zone proteins. Components of SVs and DCVs are generated in the endoplasmic reticulum (ER) and processed through the Golgi apparatus but are sorted separately into SV transport precursor vesicles containing SV proteins and into DCVs for anterograde transport toward the axon terminals. This study study provides the first evidence for a developmental genetic program that specifically controls the strength of synaptic connections by constraining the SV pool in neurons. It is hypothesized that Stromalin regulates SV and DCV number through its role in regulating gene expression (Phan, 2018).

Interestingly, the critical window for Stromalin's effects on the SV pool size occurs during the third-instar larval period. This developmental time point is well after the integration of DAn into the larval olfactory memory circuit and after the initial onset of DAn synaptogenesis onto MBn, since these synapses are already present at the first-instar larval stage. The γ MBn that are present in the larval brain prior to the mid-third-instar developmental stage undergo extensive axonal and dendritic restructuring during the pupal stage, such that the structural organization and connectivity of the larval γ MBn is distinct from that in the adult. It is during the mid-third-instar larval stage that the α'β' MBn develop, and these appear to persist into the adult fly relatively unchanged in structure. This developmental transition maps directly onto the critical window for Stromalin's effects on limiting synaptic vesicle pool size in DAn. Stromalin does not affect SV number at the earliest stages of neural circuit development and synaptogenesis but rather only upon emergence of the first set of MBn that persist and integrate into the adult neural circuitry. Stromalin may thus be specifically involved in a developmental program that adjusts the strength of DAn synaptic connectivity for adult-relevant neural circuitry and functions (Phan, 2018).

Insults that produce a loss of function of cohesin complex genes have been previously shown to cause developmental axonal and dendritic pruning defects in the γ subset of MBn. Membrane-GFP data, as well as the EM analysis on DAn neuropil volumes, failed to find similar differences in the DAn axonal ramifications of adult fly brains, arguing that DAn axons do not undergo the same Stromalin-dependent axonal pruning that occurs with γ MBn or that such pruning is transient and fails to persist into adulthood. stromalinRNAi was expressed in the γ MBn, and adult γ MBn morphology was examined using membrane-bound GFP staining but did not detect a pronounced morphological difference in these neurons. Presumably, a complete loss of function is required to detect the pronounced pruning defects observed previously. Moreover, impairing synaptic vesicle transport to axon terminals reversed the enhanced memory phenotype. Taken together, these data fail to support the hypothesis that developmental axonal pruning deficits of DAn lead to the enhancement in learning and memory scores in flies with Stromalin KD in these same neurons (Phan, 2018).

While this study focused on DAn, the data indicate that Stromalin's role in constraining synaptic vesicle pool size extends to other neurons of the Drosophila brain, since Syt:GFP increases were also detected with pan-neuronal KD and with KD in the cholinergic MB Kenyon cell neurons. The alteration of neurotransmitter release in a variety of neurons with Stromalin KD is likely to have a profound effect on a range of different behaviors, since mutations in genes affecting synaptic communication have been associated with many behavioral/cognitive, neurodevelopmental, neurodegenerative, and neuropsychiatric disorders. Similarly, it is predicted that broad expression of the unc104RNAi transgene would alter other behaviors and generally in ways opposite of stromalinRNAi with an appropriate level of expression. Thus, these transgenes offer valuable new tools for modulating SV content across neurons to probe effects on synaptic communication and behavioral processes. Interestingly, the stromalinRNAi effects were able to rescue the modest learning impairments caused by unc104RNAi expression in DAn, which suggests that increasing synaptic vesicle content may provide a potential symptomatic treatment for patients with KIF1A mutation (Phan, 2018).

Mutations in the highly conserved cohesin complex genes SMC1, SMC3, Rad21, and stromalin (STAG1/2 in mammals) are known to cause cohesinopathies, such as Cornelia de Lange Syndrome. The current observations prompt the important question of whether alterations in the synaptic vesicle pool and synaptic communication underlie some of the phenotypes associated with the cohesinopathies. The increased memory performance that was observed with Stromalin and SMC1 KD seems at odds with some phenotypes like intellectual disability found in patients. However, an increase in the SV pool across many different types of cells in the human brain resulting from a genomic mutation may produce a more complex and opposite phenotype for learning. Other behavioral phenotypes associated with cohesinopathies, including attention deficit disorder, hyperactivity, repetitive behaviors, and autistic behaviors, might also be explainable by altered synaptic vesicle pools and can interfere with learning and memory processes. Furthermore, the increased SV phenotype may also explain the susceptibility of individuals with cohesinopathies to seizures, since SV depletion following repeated neural stimulation is a common mechanism for synaptic depression, important for limiting synaptic hyperactivity that can otherwise lead to runaway network activity. Thus, cohesin complex gene mutations may attenuate SV depletion, thereby impairing normal synaptic depression and contributing to the development of seizures and behavioral dysfunction in humans (Phan, 2018).

Go to top

Electrical synapses between mushroom body neurons are critical for consolidated memory retrieval in Drosophila

Shyu, W. H., Lee, W. P., Chiang, M. H., Chang, C. C., Fu, T. F., Chiang, H. C., Wu, T. and Wu, C. L. (2019). PLoS Genet 15(5): e1008153. PubMed ID: 31071084

Electrical synapses between neurons, also known as gap junctions, are direct cell membrane channels between adjacent neurons. Gap junctions play a role in the synchronization of neuronal network activity; however, their involvement in cognition has not been well characterized. Three-hour olfactory associative memory in Drosophila has two components: consolidated anesthesia-resistant memory (ARM) and labile anesthesia-sensitive memory (ASM). This study shows that knockdown of the gap junction gene innexin5 (inx5) in mushroom body (MB) neurons disrupted ARM, while leaving ASM intact. Whole-mount brain immunohistochemistry indicated that INX5 protein was preferentially expressed in the somas, calyxes, and lobes regions of the MB neurons. Adult-stage-specific knockdown of inx5 in αβ neurons disrupted ARM, suggesting a specific requirement of INX5 in αβ neurons for ARM formation. Hyperpolarization of αβ neurons during memory retrieval by expressing an engineered halorhodopsin (eNpHR) also disrupted ARM. Administration of the gap junction blocker carbenoxolone (CBX) reduced the proportion of odor responsive alphabeta neurons to the training odor 3 hours after training. Finally, the α-branch-specific 3-hour ARM-specific memory trace was also diminished with CBX treatment and in inx5 knockdown flies. Altogether, these results suggest INX5 gap junction channels in αβ neurons for ARM retrieval and also provide a more detailed neuronal mechanism for consolidated memory in Drosophila (Shyu, 2019).

In fruit flies, two parallel MB circuits, containing αβ and α'β' neurons, are involved in ARM formation. Radish expression in αβ neurons is required for partial ARM, whereas octβ2R expression in MB α'β' neurons is required for the rest part of ARM, suggesting that two distinct cellular mechanisms regulate ARM in different MB neurons. The radish gene encodes a protein with a predicted cAMP-dependent protein kinase phosphorylation site, which can bind Rac1 to regulate the rearrangement of the cytoskeleton and affect synaptic structural morphology. The interaction of RADISH and BRUCHPILOT at the synaptic active zone has been proposed to regulate neurotransmitter release, and genetic knockdown of radish or bruchpilot in αβ neurons disrupts ARM. A recent study indicated that Drk-Drok signaling is essential for ARM formation in αβ neurons, and related to dynamic cytoskeletal changes. In addition, the dopamine type 2 (D2R) and serotonin (5HT1A) receptors in αβ neurons are also critical for ARM formation (Shyu, 2019).

The key finding of this study is that the gap junction protein INX5 in αβ neurons is critical for 3-hour ARM retrieval. This conclusion is supported by four independent lines of evidence. First, immunohistochemistry data indicated that INX5 was preferentially expressed in the MB calyxes and somas, and these INX5-positive signals were reduced in OK107-GAL4 > UAS-inx5RNAi flies. Second, adult-stage-specific knockdown of inx5 in αβ neurons impaired ARM. Third, eNpHR-mediated inhibition of action potential in αβ neurons during retrieval also impaired ARM. Forth, knockdown of inx5 in αβ neurons inhibited the training-induced cellular calcium responses in the MB α-lobe region 3 hours after odor/shock association (Shyu, 2019).

Previous studies have concluded that αβ neuronal activity is involved in 3-hour memory retrieval using shibirets to transiently block chemical synaptic transmissions via inhibiting neurotransmitter recycling. Three-hour memory is composed of ASM and ARM, each accounting for about half of the memory retention level. In a recent study, it was shown that the inhibition of neurotransmitter recycling in αβ neurons during memory retrieval disrupted 3-hour ARM. However, blocking neurotransmitter recycling in αβ neurons during memory acquisition and consolidation did not affect 3-hour ARM. The function of gap junctions in the electrical synapses is to coordinate the propagation of action potential in neuronal networks, and shibirets cannot block gap junction-mediated electrical synapses. This study therefore used eNpHR to transiently silence action potential in αβ neurons to confirm the requirement of αβ neuronal activity during the ARM formation process. The data showed an eNpHR-mediated hyperpolarization of αβ neurons during memory retrieval but not acquisition or consolidation, impaired ARM, suggesting that action potential in αβ neurons is required only for ARM retrieval. Brain immunostaining data showed that INX5 gap junction proteins are strongly expressed in the calyxes and somas of αβ neurons, and knockdown of inx5 gap junction gene in αβ neurons disrupted ARM. The expression of gap junction is critical for neuronal functions since it plays a role in the propagation of action potential between adjacent neurons. It is therefore concluded that the gap junction channels composed of INX5 in αβ neurons are critical for ARM retrieval. A recent study showed that the gap junction protein INX2 regulates calcium transmission across the follicle cells during Drosophila oogenesis. In addition, INX1/INX2 induces calcium oscillations in the glial cells of the blood-brain barrier (BBB), enabling signal amplification and synchronization across the BBB in fruit flies. Furthermore, the gap junction protein INX6 is important for promoting synchronous neuronal activity in the dorsal fan-shaped body (dFB) in the fly brain that is critical for the sleep switch. In mammals, most neuronal gap junctions in the brain are composed of Connexin-36 (Cx36) and are involved in synchronizing the hippocampal neuronal oscillatory patterns, which is required for emotional memories. Therefore, it is possible that gap junction channels composed of INX5 mediate neuronal activity amplification and synchronization across αβ neurons, boosting the synaptic output strength during ARM retrieval (Shyu, 2019).

By using the newly developed calcium indicator GCaMP6, the increased proportion was observed of training odor-responsive αβ neurons 3 hours after odor/shock association, and this phenomenon was abolished after treatment with gap junction blocker CBX). Furthermore, significant enhancement was observed of the training-induced cellular calcium response to the training odor in the MB α-lobe branch 3 hours after odor/shock association. According to the broad consensus of the field, the memory trace is supposed to be formed in the vertical lobe of the MBs by the activity contingency of MBs and dopaminergic Protocerebral Posterior Lateral 1 (PPL1) neurons, which represent odor and punitive shock, respectively. Therefore, it is possible that 3-hour memory trace back propagation of somas' activity occurs from the MB lobes during memory retrieval. In addition, the branch specific modifications via MB input neurons (e.g., Protocerebral Anterior Medial, PAM) may occur during memory retrieval, hence the memory trace was only observed in α-lobe branch but not the β-lobe of MBs. This training-induced 3-hour ARM-specific memory trace was eliminated by treatment with the gap junction blocker, CBX, during memory retrieval or by genetic knockdown of inx5 in αβ neurons. Although a significant 3-hour ARM-specific memory trace was also observed in α'β' neurons, this phenomenon was independent of the gap junction. From this, it is proposed that an unknown dynamic mechanism regulates the permeability of gap junction channels composed of INX5 in αβ neurons after training. Recently, cryoelectron microscopy revealed that the structure of C.elegans INX6 was highly similar to that of the vertebrate gap junction protein Connexin-26 (Cx26). Connexin properties, such as gating and assembly, can be regulated by phosphorylation. Additionally, the functions of Innexins or Connexins can also be regulated by changes in the intracellular pH and calcium levels. Establishing whether the properties of INX5 in the MBs are modified following conditioned training will provide insights into the neuronal mechanisms of ARM (Shyu, 2019).

Go to top

miR-92a Suppresses Mushroom Body-Dependent Memory Consolidation in Drosophila

Guven-Ozkan, T., Busto, G. U., Jung, J. Y., Drago, I. and Davis, R. L. (2020). eNeuro. PubMed ID: 32737186

MicroRNAs fine tune gene expression to regulate many aspects of nervous system physiology. This study shows that miR-92a suppresses memory consolidation that occurs in the αβ and γ mushroom body neurons of Drosophila, making miR-92a a memory suppressor microRNA. Bioinformatics analyses suggested that mRNAs encoding kinesin heavy chain 73 (Khc73), a protein that belongs to Kinesin-3 family of anterograde motor proteins, may be a functional target of miR-92a. Behavioral studies that employed expression of khc73 with and without its 3'UTR containing miR-92a target sites, luciferase assays in HEK cells with reporters containing wild-type and mutant target sequences in the miR-92a 3' UTR, and immunohistochemistry experiments involving Khc73 expression with and without the wild-type khc73 3'UTR all point to the conclusion that khc73 is a major target of miR-92a in its functional role as a microRNA memory suppressor gene (Guven-Ozkan, 2020).

Go to top

Inhibitory muscarinic acetylcholine receptors enhance aversive olfactory learning in adult Drosophila

Bielopolski, N., Amin, H., Apostolopoulou, A. A., Rozenfeld, E., Lerner, H., Huetteroth, W., Lin, A. C. and Parnas, M. (2019). Inhibitory muscarinic acetylcholine receptors enhance aversive olfactory learning in adult Drosophila. Elife 8. PubMed ID: 31215865

Olfactory associative learning in Drosophila is mediated by synaptic plasticity between the Kenyon cells of the mushroom body and their output neurons. Both Kenyon cells and their inputs from projection neurons are cholinergic, yet little is known about the physiological function of muscarinic acetylcholine receptors in learning in adult flies. This study shows that aversive olfactory learning in adult flies requires type A muscarinic acetylcholine receptors (mAChR-A), particularly in the gamma subtype of Kenyon cells. mAChR-A inhibits odor responses and is localized in Kenyon cell dendrites. Moreover, mAChR-A knockdown impairs the learning-associated depression of odor responses in a mushroom body output neuron. These results suggest that mAChR-A function in Kenyon cell dendrites is required for synaptic plasticity between Kenyon cells and their output neurons (Bielopolski, 2019).

Animals learn to modify their behavior based on past experience by changing connection strengths between neurons, and this synaptic plasticity is often regulated by metabotropic receptors. In particular, neurons commonly express both ionotropic and metabotropic receptors for the same neurotransmitter, where the two may mediate different functions (e.g., direct excitation/inhibition vs. synaptic plasticity). In mammals, where glutamate is the principal excitatory neurotransmitter, metabotropic glutamate receptors (mGluRs) have been widely implicated in synaptic plasticity and memory. Given the complexity of linking behavior to artificially induced plasticity in brain slices, it would be useful to study the role of metabotropic receptors in learning in a simpler genetic model system with a clearer behavioral readout of synaptic plasticity. One such system is Drosophila, where powerful genetic tools and well-defined anatomy have yielded a detailed understanding of the circuit and molecular mechanisms underlying associative memory. The principal excitatory neurotransmitter in Drosophila is acetylcholine, but, surprisingly, little is known about the function of metabotropic acetylcholine signaling in synaptic plasticity or neuromodulation in Drosophila. This study addresses this question using olfactory associative memory (Bielopolski, 2019).

Flies can learn to associate an odor (conditioned stimulus, CS) with a positive (sugar) or a negative (electric shock) unconditioned stimulus (US), so that they later approach 'rewarded' odors and avoid 'punished' odors. This association is thought to be formed in the presynaptic terminals of the ~2000 Kenyon cells (KCs) that make up the mushroom body (MB), the fly's olfactory memory center. These KCs are activated by odors via second-order olfactory neurons called projection neurons (PNs). Each odor elicits responses in a sparse subset of KCs so that odor identity is encoded in which KCs respond to each odor. When an odor (CS) is paired with reward/punishment (US), an odor-specific set of KCs is activated at the same time that dopaminergic neurons (DANs) release dopamine onto KC presynaptic terminals. The coincident activation causes long-term depression (LTD) of synapses from the odor-activated KCs onto mushroom body output neurons (MBONs) that lead to approach or avoidance behavior. In particular, training specifically depresses KC-MBON synapses of the 'wrong' valence (e.g. odor-punishment pairing depresses odor responses of MBONs that lead to approach behavior), because different pairs of 'matching' DANs/MBONs (e.g. punishment/approach, reward/avoidance) innervate distinct regions along KC axons (Bielopolski, 2019).

Both MB input (PNs) and output (KCs) are cholinergic, and KCs express both ionotropic (nicotinic) and metabotropic (muscarinic) acetylcholine receptors. The nicotinic receptors mediate fast excitatory synaptic currents, while the physiological function of the muscarinic receptors is unknown. Muscarinic acetylcholine receptors (mAChRs) are G-protein-coupled receptors; out of the three mAChRs in Drosophila (mAChR-A, mAChR-B and mAChR-C), mAChR-A (also called Dm1, mAcR-60C or mAChR) is the most closely homologous to mammalian mAChRs. Mammalian mAChRs are typically divided between 'M1-type' (M1/M3/M5), which signal via Gq and are generally excitatory, and 'M2-type' (M2/M4), which signal via Gi/o and are generally inhibitory. Drosophila mAChR-A seems to use 'M1-type' signaling: when heterologously expressed in Chinese hamster ovary (CHO) cells, it signals via Gq protein to activate phospholipase C, which produces inositol trisphosphate to release Ca2+ from internal stores (Bielopolski, 2019).

Recent work indicates that mAChR-A is required for aversive olfactory learning in Drosophila larvae, as knocking down mAChR-A expression in KCs impairs learning. However, it is unclear whether mAChR-A is involved in olfactory learning in adult Drosophila, given that mAChR-A is thought to signal through Gq, and in adult flies Gq signaling downstream of the dopamine receptor Damb promotes forgetting, not learning. Moreover, it is unknown how mAChR-A affects the activity or physiology of KCs, where it acts (at KC axons or dendrites or both), and how these effects contribute to olfactory learning (Bielopolski, 2019).

This study shows that mAChR-A is required in KCs for aversive olfactory learning in adult Drosophila. Surprisingly, genetic and pharmacological manipulations of mAChR-A suggest that mAChR-A is inhibitory and acts on KC dendrites. Moreover, mAChR-A knockdown impairs the learning-associated depression of odor responses in an MB output neuron, MB-MVP2, that is required for aversive memory retrieval. It is suggested that dendritically acting mAChR-A is required for synaptic depression between KCs and their outputs (Bielopolski, 2019).

This study shows that mAChR-A is required in γ KCs for aversive olfactory learning and short-term memory in adult Drosophila. Knocking down mAChR-A increases KC odor responses, while the mAChR-A agonist muscarine suppresses KC activity. Knocking down mAChR-A prevents aversive learning from reducing responses of the MB output neuron MB-MVP2 to the conditioned odor, suggesting that mAChR-A is required for the learning-related depression of KC-->MBON synapses (Bielopolski, 2019).

Why is mAChR-A only required for aversive learning in γ KCs, not αβ or α'β' KCs? Although the mAChR-A MiMIC gene trap agrees with single-cell transcriptome analysis that α'β' KCs express less mAChR-A than do γ and αβ KCs, transcriptome analysis indicates that α'β' KCs do express some mAChR-A. Moreover, γ and αβ KCs express similar levels of mAChR-A. It may be that the RNAi knockdown is less efficient at affecting the physiology of αβ and α'β' KCs than γ KCs, whether because the knockdown is less efficient at reducing protein levels, or because αβ and α'β' KCs have different intrinsic properties or a different function of mAChR-A such that 40% of normal mAChR-A levels is sufficient in αβ and α'β' KCs but not γ KCs. This interpretation is supported by the finding that mAChR-A RNAi knockdown significantly increases odor responses only in the γ lobe, not the αβ or α'β' lobes. Alternatively, γ, αβ and α'β' KCs are thought to be important mainly for short-term memory, long-term memory, and memory consolidation, respectively; as this study tested only short-term memory, mAChR-A may carry out the same function in all KCs, but only its role in γ KCs is required for short-term (as opposed to long-term) memory. Indeed, the key plasticity gene DopR1 is required in γ, not αβ or α'β'4 KCs, for short-term memory. It may be that mAChR-A is required in non-γ KC types for other forms of memory besides short-term aversive memory, such as appetitive conditioning or other phases of memory like long-term memory. The finding that mAChR-A is required in γ KCs for aversive short-term memory is consistent with the finding that mAChR-A knockdown in KCs disrupts training-induced depression of odor responses in MB-MVP2, an MBON postsynaptic to γ KCs required for aversive short-term memory. However, the latter finding does not rule out the possibility that other MBONs postsynaptic to non-γ KCs may also be affected by mAChR-A knockdown in KCs (Bielopolski, 2019).

mAChR-A seems to inhibit KC odor responses, because knocking down mAChR-A increases odor responses in the calyx and γ lobe, while activating mAChR-A with bath or local application of muscarine decreases KC odor responses. Some details differ between the genetic and pharmacological results. In particular, while mAChR-A knockdown mainly affects γ KCs, with other subtypes inconsistently affected, muscarine reduces responses in all KC subtypes. What explains these differences? mAChR-A might be weakly activated in physiological conditions, in which case gain of function would cause a stronger effect than loss of function. Similarly, pharmacological activation of mAChR-A is likely a more drastic manipulation than a 60% reduction of mAChR-A mRNA levels. Although network effects from muscarine application cannot be entirely ruled out, the effect of muscarine does not stem from PNs or APLand locally applied muscarine would have little effect on neurons outside the mushroom body (Bielopolski, 2019).

How does mAChR-A inhibit odor-evoked Ca2+ influx in KCs? Given that mAChR-A signals through Gq when expressed in CHO cells, that muscarinic Gq signaling normally increases excitability in mammals, and that pan-neuronal artificial activation of Gq signaling in Drosophila larvae increases overall excitability, it may be surprising that mAChR-A inhibits KCs. However, Gq signaling may exert different effects on different neurons in the fly brain, and some examples exist of inhibitory Gq signaling by mammalian mAChRs. M1/M3/M5 receptors acting via Gq can inhibit voltage-dependent Ca2+ channels, reduce voltage-gated Na +currents, or trigger surface transport of KCNQ channels, thus increasing inhibitory K+ currents. Drosophila mAChR-A may inhibit KCs through similar mechanisms (Bielopolski, 2019).

What is the source of ACh which activates mAChR-A and modulates odor responses? In the calyx, cholinergic PNs are certainly a major source of ACh. However, KCs themselves are cholinergic and release neurotransmitter in both the calyx and lobes. KCs form synapses on each other in the calyx, possibly allowing mAChR-A to mediate lateral inhibition, in conjunction with the lateral inhibition provided by the GABAergic APL neuron (Bielopolski, 2019).

What function does mAChR-A serve in learning and memory? The results indicate that mAChR-A knockdown prevents the learning-associated weakening of KC-MBON synapses, in particular for MBON-γ1pedc>α/β, aka MB-MVP2. One potential explanation is that the increased odor-evoked Ca2+ influx observed in knockdown flies increases synaptic release, which overrides the learning-associated synaptic depression. However, increased odor-evoked Ca2+ influx per se is unlikely on its own to straightforwardly explain a learning defect, because other genetic manipulations that increase odor-evoked Ca2+ influx in KCs either have no effect on, or even improve, olfactory learning. For example, knocking down GABA synthesis in the inhibitory APL neuron increases odor-evoked Ca2+ influx in KCs and improves olfactory learning (Bielopolski, 2019).

The most intuitive explanation would be that mAChR-A acts at KC synaptic terminals in KC axons to help depress KC-MBON synapses. Yet overexpressed mAChR-A localizes to KC dendrites, not axons, and functionally rescues mAChR-A hypomorphic mutants, showing that dendritic mAChR-A suffices for its function in learning and memory. Does this show that mAChR-A has no role in KC axons? The inability to detect GFP expressed from the mAChR-A MiMIC gene trap suggests that normally there may only be a small amount of mAChR-A in KCs. It may be that with mAChR-A-FLAG overexpression, the correct (undetectable) amount of mAChR-A is trafficked to and functions in axons, but due to a bottleneck in axonal transport, the excess tagged mAChR-A is trapped in KC dendrites. While the results do not rule out this possibility, a general bottleneck in axonal transport seems unlikely as many overexpressed proteins are localized to KC axons. It is more parsimonious to take the dendritic localization of mAChR-A-FLAG at face value and infer that mAChR-A functions in KC dendrites (Bielopolski, 2019).

How can mAChR-A in KC dendrites affect synaptic plasticity in KC axons? mAChR-A signaling might change the shape or duration of KC action potentials, an effect that could potentially propagate to KC axon terminals. Such changes in the action potential waveform may not be detected by calcium imaging, but could potentially affect a 'coincidence detector' in KC axons that detects when odor (i.e. KC activity) coincides with reward/punishment (i.e., dopamine). This coincidence detector is generally believed to be the Ca2+-dependent adenylyl cyclase Rutabaga. Changing the waveform of KC action potentials could potentially affect local dynamics of Ca2+ influx near Rutabaga molecules. In addition, rutabaga mutations do not abolish learning (mutants have ~40-50% of normal learning scores), so there may be additional coincidence detection mechanisms affected by action potential waveforms. Testing this idea would require a better understanding of biochemical events underlying learning at KC synaptic terminals (Bielopolski, 2019).

Alternatively, mAChR-A's effects on synaptic plasticity may not occur acutely. Although purely developmental effects of mAChR-A were ruled out through adult-only RNAi expression, knocking out mAChR-A for several days in adulthood might still affect KC physiology in a not-entirely-acute way. Perhaps, as with other G-protein-coupled receptors, muscarinic receptors can affect gene expression -- if so, this could have wide-ranging effects on KC physiology: for example, action potential waveform, expression of key genes required for synaptic plasticity, etc. Another intriguing possibility is suggested by an apparent paradox: both mAChR-A and the dopamine receptor Damb signal through Gq, but mAChR-A promotes learning while Damb promotes forgetting. How can Gq mediate apparently opposite effects? Perhaps Gq signaling aids both learning and forgetting by generally rendering synapses more labile. Indeed, although damb mutants retain memories for longer than wildtype, their initial learning is slightly impaired; damb mutant larvae are also impaired in aversive olfactory learning. Although one study reports that knocking down Gq in KCs did not impair initial memory, the Gq knockdown may not have been strong enough; also, that study shocked flies with 90 V shocks, which also gives normal learning in mAChR-A knockdown flies (Bielopolski, 2019).

Such hypotheses posit that mAChR-A regulates synaptic plasticity 'competence' rather than participating directly in the plasticity mechanism itself. Why should synaptic plasticity competence be controlled by an activity-dependent mechanism? It is tempting to speculate that mAChR-A may allow a kind of metaplasticity in which exposure to odors (hence activation of mAChR-A in KCs) makes flies' learning mechanisms more sensitive. Indeed, mAChR-A is required for learning with moderate (50 V) shocks, not severe (90 V) shocks. Future studies may further clarify how muscarinic signaling contributes to olfactory learning (Bielopolski, 2019).

Go to top

A GABAergic feedback shapes dopaminergic input on the Drosophila mushroom body to promote appetitive long-term memory

Pavlowsky, A., Schor, J., Placais, P. Y. and Preat, T. (2018). Curr Biol. 28(11):1783-1793. Pubmed ID: 29779874

Memory consolidation is a crucial step for long-term memory (LTM) storage. However, a clear picture of how memory consolidation is regulated at the neuronal circuit level is still lacking. This study took advantage of the Drosophila olfactory memory center, the mushroom body (MB), to address this question in the context of appetitive LTM. The MB lobes, which are made by the fascicled axons of the MB intrinsic neurons, are organized into discrete anatomical modules, each covered by the terminals of a defined type of dopaminergic neuron (DAN) and the dendrites of a corresponding type of MB output neuron (MBON). An essential role has been revealed of one DAN, the MP1 neuron, in the formation of appetitive LTM. The MP1 neuron is anatomically matched to the GABAergic MBON MVP2, which has been attributed feedforward inhibitory functions recently. This study used behavior experiments and in vivo imaging to challenge the existence of MP1-MVP2 synapses and investigate their role in appetitive LTM consolidation. MP1 and MVP2 neurons form an anatomically and functionally recurrent circuit, which features a feedback inhibition that regulates consolidation of appetitive memory. This circuit involves two opposite type 1 and type 2 dopamine receptors (the type 1 DAMB and the type 2 dD2R) in MVP2 neurons and the metabotropic GABAB-R1 receptor in MP1 neurons. It is proposed that this dual-receptor feedback supports a bidirectional self-regulation of MP1 input to the MB. This mechanism displays striking similarities with the mammalian reward system, in which modulation of the dopaminergic signal is primarily assigned to inhibitory neurons (Pavlowsky, 2018).

Formation of a memory engram is a multi-step process, from encoding the relevant information to the final storage of memory traces. Describing the neuronal architecture and functions that underlie each step of this process is crucial to understanding memory ability. In Drosophila, a very fine knowledge is available of the anatomy of the mushroom body (MB), the major olfactory integrative brain center, as well as its input and output neurons. The mapping to these circuits of various functional modalities occurring at the different stages of memory encoding, storage, and recall is also quite advanced (Pavlowsky, 2018).

Drosophila MBs are paired structures including ~2,000 intrinsic neurons per brain hemisphere. These neurons receive dendritic input from the antennal lobes through projection neurons in the calyx area on the posterior part of the brain. Their axons form a fascicle, called a peduncle, that traverses the brain to the anterior part, where axons branch to form horizontal and vertical lobes according to three major branching patterns (α/β, α'/β' and γ). MB lobes are tiled by spatially segregated presynaptic projections from dopamine neurons (DANs), on the one hand, and dendrites of MB output neurons (MBONs), on the other hand. DANs and MBONs are matched to form defined anatomical compartments that are increasingly considered as independent functional units. On several of these compartments, it was shown that DAN activity can induce heterosynaptic plasticity at the MB/MBON synapse, which could be a cellular substrate of memory encoding (Pavlowsky, 2018).

In addition to this canonical anatomical motif of the DAN/MB intrinsic neurons/MBON triad, electron microscopy connectome reconstruction in the larval brain has evidenced recently that DANs have direct synaptic connections to their matched MBONs. In the adult, direct DAN-to-MBON synapses have also been observed in several compartments of the MB vertical lobes (Pavlowsky, 2018).

MB activity is regulated by a broad spectrum of neuromodulatory input, among which tonic dopamine signaling plays an important role in the regulation of memory persistence or expression. In particular, it has been shown that sustained rhythmic activity of the MP1 DAN, also named PPL1-γ1pedc and which innervates the γ1 module and the α/β peduncle, is crucial after conditioning to enable the consolidation of both aversive and appetitive long-term memory (LTM), the most stable memory forms that rely on de novo protein synthesis. The MP1 neuron is anatomically matched with the MVP2 neuron, a GABA-ergic MBON that shows a complex arborization. The MVP2 neuron, also named MBON-γ1pedc > α/β, possesses two dendritic domains on the γ1 and peduncle compartments. On the ipsilateral side, MVP2 has presynaptic projections on MB vertical and medial lobes and also targets brain areas outside MB where other MBONs project. In particular, MVP2 neurons mediates a feedforward inhibition of specific MBONs involved in aversive and appetitive memory retrieval. Interestingly, MVP2 neurons also send a presynaptic projection onto the contralateral peduncle, a place of MP1 presynaptic coverage. Hence, the anatomy of the MP1-MVP2 neurons is compatible with the existence of feedback circuitry. This study tested experimentally the existence of such a functional feedback in the context of appetitive LTM formation (Pavlowsky, 2018).

Appetitive memory results from the paired delivery of an odorant and a sugar to starved flies. Only one pairing is sufficient to form both short-term memory (STM) and LTM, but it was shown that these two memory phases stem from distinct properties of the reinforcing sugar: although the sweetness of the sugar is sufficient so that flies form appetitive STM, the formation of LTM requires that the conditioning is made with a caloric sugar. The nutritional value of the reinforcing sugar translates in the fly brain as a post-ingestive sustained rhythmic signaling from MP1 neurons that is necessary to consolidate LTM. At the cellular level, STM and LTM stem from parallel and independent memory traces located in distinct subsets of MB neurons; respectively, γ neurons and α/β neurons. Several MB output circuits have been involved in the retrieval of appetitive STM (MBON-γ2α'1), LTM (MBON-α3, MBON-α1), or both (M4/M6, also named MBON-γ5β'2a/MBON-β'2mp), providing as many candidate synaptic sites of memory encoding (Pavlowsky, 2018).

This work confirmed that post-training MP1 activity is required for LTM formation, but it was shown in addition that this activity must be temporally restricted. MP1 activity is self-regulated through an inhibitory feedback by MVP2 neurons. Immediately after conditioning, the oscillatory activity of MP1 is enhanced and MVP2 is inhibited. After about 30 min, MVP2 is activated, terminating the period of MP1 increased signaling, which, this study shows is a requirement for proper LTM formation. It is proposed that the bidirectional action of this feedback loop is based at the molecular level on the sequential involvement of two antagonist dopamine receptors, the type 1 DAMB and the type 2 dD2R on one side and the metabotropic GABAB-R1 receptor on the other side (Pavlowsky, 2018).

This work describes a functional inhibitory feedback from an MBON, the GABA-ergic MVP2 neuron, to the dopaminergic neuron of the same MB module, the MP1 neuron. Anatomical data from synaptic staining and electron microscopy, as well as the requirement of a specific GABA receptor in MP1 neurons for appetitive LTM, lead to the hypothesis of a direct connection between MVP2 and MP1 neurons, although alternative scenarios featuring plurisynaptic circuits involving additional GABAergic neurons cannot be ruled out at this stage. Using time-resolved manipulation of neuronal activity, it was shown that this feedback circuit is involved in the first hour after appetitive conditioning for LTM formation. It was already known, and confirmed in this study, that the activity of MP1 neurons, in the form of regular calcium oscillations, is necessary in the first 30-45 min after conditioning to build LTM. Strikingly, in the present work, it was shown that, after this initial time period, the activity of MP1 neurons is not merely dispensable but rather deleterious for LTM formation, since activating MP1 neurons from 0.5 hr to 1 hr after conditioning caused an LTM defect (Pavlowsky, 2018).

Conversely, it was found that, in that time interval where MP1 neuron activity is deleterious, MVP2 neurons need to be active for normal LTM performance. Imaging experiments showed that blocking MVP2 neurons increased the persistence of MP1 neuron oscillations, up to more than 1 hr post-conditioning. The same effect was observed when the GABAB-R1 receptor was knocked down in MP1 neurons. Interestingly, blocking MVP2 neurons or GABA-ergic signaling in MP1 neurons mostly affected the frequency and the regularity of MP1 calcium signals, without markedly increasing their amplitude. Hence MVP2 neurons seem to be involved in terminating the period of sustained oscillatory signaling from MP1 neurons rather than merely decreasing MP1 activity. However, in the first 0.5 hr after conditioning, MVP2 neuron activity is not simply dispensable but also deleterious for LTM. Since MVP2 neurons have an inhibitory effect on MP1 activity, it is likely that MVP2 neurons have to be inhibited to let MP1 oscillations occur. Strikingly, this study established that MVP2 neurons are modulated by dopamine signaling through two receptors: DAMB, a type 1 activating receptor; and dD2R, a type 2 inhibitory receptor. Although these two receptors have opposite downstream effects, both are required in MVP2 for normal LTM performance. Overall, the results evidence that the MP1-MVP2 feedback circuit is functionally designed to allow the onset of LTM-gating oscillations only on a precise time windows of about 0.5 hr after conditioning (Pavlowsky, 2018).

It is proposed that MP1 activity is self-regulated through a dual receptor mechanism that controls MVP2 feedback. Initially, the ongoing activity of MP1 neurons inhibits MVP2 neurons through the dD2 receptor, which allows for sustained MP1 activity. In a second step, DAMB is activated in MVP2 neurons to enable the inhibitory feedback that shuts off MP1 oscillations. This model unifies molecular data and the results obtained from time-resolved thermogenetic manipulation of neuronal activity; unfortunately, such temporality of receptor involvement cannot be tested with RNAi-based knockdown (Pavlowsky, 2018).

DAMB and dD2R are two G-protein-coupled dopamine receptors. Although dD2R is a clear homolog of mammalian D2 receptor, and is negatively coupled to cAMP synthesis, the molecular mechanisms downstream of DAMB appear to be more diverse. It was shown that DAMB activation can stimulate cAMP synthesis, similarly to the function of a type 1 receptor, likely through Gβγ-coupled signaling. Surprisingly, it was recently shown that DAMB-mediated dopamine signaling could transiently inhibit the spiking of sleep-promoting neurons through the same G-protein pathway. Additionally, it was shown that DAMB can also activate downstream calcium signaling from intracellular calcium stores. In the current model, MVP2 neurons need, at one point, to be activated to dampen MP1 oscillations, so activating functions of DAMB seem to be more relevant in the present environment. Interestingly, physiological measurements in a heterologous system showed that cAMP activation occurs within tens of minutes, while calcium activation occurs on much shorter timescales. The delayed requirement of MVP2 activity (starting ~30 min after conditioning) seems to be more consistent with an activation of the cAMP pathway. It would be helpful in the future to decipher the molecular mechanism downstream of DAMB involved in this feedback loop. The sequential activation of two distinct dopamine receptors could be due to different affinities for dopamine. Indeed, pharmacological studies show that D2R-like receptors have a higher affinity toward dopamine compared to the D1-like receptors in mammals. However, in the specific case of Drosophila D2R and DAMB, similar dopamine affinities for both receptors were reported (0.5 μM for D2R [52 and 0.1-1 μM for DAMB, although these are all obtained from in vitro preparations of cultured cells. There could be also be subtler differences of activation kinetics based both on the quantity and on the mode of dopamine release by MP1 neurons (Pavlowsky, 2018).

MP1 neurons and MVP2 neurons have been shown to play crucial roles in both aversive and appetitive memories. During aversive conditioning, MP1 neurons mediate the unconditioned stimulus, which is thought to involve dDA1 activation in MB neurons. In a recent report, it was shown that suppressing the activity of MVP2 neurons during an odor presentation leads to the formation of an aversive memory toward this odor. In light of this result, these authors proposed that the role of steady-state MVP2 activity is to prevent the formation of irrelevant memory from insignificant stimuli. Given the role of MP1 in the signaling of negative stimuli during aversive learning, this finding and its interpretation are fully consistent with the existence of an inhibitory feedback from MVP2 neurons to MP1 neurons, as reported in the present work. MP1 neurons are also central in the formation of LTM after conditioning. Tonic signaling through slow oscillations of MP1 neurons gates the formation of aversive LTM after spaced training. The same kind of sustained post-training signaling builds LTM after appetitive conditioning. Both in aversive and appetitive paradigms, this LTM-gating function involves DAMB signaling in MB neurons. After aversive spaced training, it was shown that DAMB activation triggers an upregulation of MB energy metabolism, which starts the consolidation of LTM. Finally, MP1 neurons also regulate the retrieval of appetitive STM. MP1 inhibition in starved flies, through suppressive dNPF signaling, allows integration of the appetitive motivational state with the expression of MB-encoded memory trace during retrieval to allow for the expression of appetitive STM. This involves enhanced feedforward inhibition from MVP2 neurons to the M4/M6 MBONs that mediate appetitive memory retrieval. The fact that MP1 inhibition goes along with enhanced MVP2 activity is consistent with the fact that baseline MP1 activity can drive an inhibition of MVP2 through dD2R, as is reported in this study. This may explain why a knockdown of dD2R in MVP2 neurons, by indiscriminately disturbing this MP1-MVP2 inhibitory link, would impair the odor-specific message carried by M4/M6 neurons for memory retrieval and cause an STM defect. All these findings illustrate how the sophistication of MP1 neuron involvement in memory is tightly linked to the diversity of receptors and neuronal targets that it can activate. A finer understanding of these processes calls for higher resolution physiological measurements to understand how the various dopamine receptors are sensitive to different modalities or kinetics of dopamine release (Pavlowsky, 2018).

Recently, it was shown that acquisition and consolidation of appetitive LTM also rely on a positive-feedback circuit involving the α1 MB compartment, dopaminergic PAM-α1, and glutamatergic MBON-α1 neurons (Ichinose, 2015). Thus, consolidation of appetitive memory involves two different recurrent circuits that share common features, such as the MBON's dual functions in consolidation and retrieval of memory. MP1 neurons are activated after a conditioning with a nutritious sugar, which is necessary for LTM formation. PAM-α1 neurons are activated during conditioning and probably mediate the coincidence detection between sugar intake and odor perception within MB neurons. The recurrent activity of the α1 compartment loop is also necessary for proper LTM formation, presumably to stabilize a nascent memory trace. Interestingly, the electron microscopy reconstruction of the adult MB vertical lobes recently showed that MVP2 neurons form direct synapses with MBONs in the α2 and α3 modules and, probably, in the α1 compartment as well. Therefore, the two feedback circuits may not be independent, and MVP2 neurons may also mediate a feedforward input from the MP1/MVP2 loop to the PAM-α1/MBON-α1 loop. The dD2R-mediated inhibition of MVP2 neurons by MP1 activity immediately after conditioning could, therefore, help in maintaining the recurrent activity in the α1 compartment (Pavlowsky, 2018).

In conclusion, this study shown here that a negative-feedback loop functions to control appetitive LTM formation, likely involving two antagonist dopaminergic receptors. This negative-feedback loop is strikingly similar to one recently described in the mammalian mesolimbic system in which feedback from inhibitory neurons prevents the over-activation of dopaminergic neurons. These two circuits have at least three common features: they rely on the metabotropic receptors DA1 and GABABR1; they comprise dopaminergic and inhibitory neurons, which are monosynaptically connected in mammals, and possibly also in Drosophila; and they are involved in the memory acquisition of motivationally relevant stimuli. These shared properties of negative-feedback loops highlight how similar strategies exist at both the network and molecular levels to regulate certain related behaviors across species (Pavlowsky, 2018).

Go to top

Mushroom body specific transcriptome analysis reveals dynamic regulation of learning and memory genes after acquisition of long-term courtship memory in Drosophila

Jones, S. G., Nixon, K. C. J., Chubak, M. C. and Kramer, J. M. (2018). G3 (Bethesda). PubMed ID: 30158319

The formation and recall of long-term memory (LTM) requires neuron activity-induced gene expression. The complex spatial and temporal dynamics of memory formation creates significant challenges in defining memory-relevant gene expression changes. The Drosophila mushroom body (MB) is a signaling hub in the insect brain that integrates sensory information to form memories across several different experimental memory paradigms. This study performed transcriptome analysis in the MB at two time points after the acquisition of LTM: 1 hour and 24 hours. The MB transcriptome was compared to biologically paired whole head (WH) transcriptomes. In both, more transcript level changes were identified at 1 hour after memory acquisition (WH = 322, MB = 302) than at 24 hours (WH = 23, MB = 20). WH samples showed downregulation of developmental genes and upregulation of sensory response genes. In contrast, MB samples showed vastly different changes in transcripts involved in biological processes that are specifically related to LTM. MB-downregulated genes were highly enriched for metabolic function. MB-upregulated genes were highly enriched for known learning and memory processes, including calcium-mediated neurotransmitter release and cAMP signalling. The neuron activity inducible genes Hr38 and sr were also specifically induced in the MB. These results highlight the importance of sampling time and cell type in capturing biologically relevant transcript level changes involved in learning and memory. The data suggests that MB cells transiently upregulate known memory-related pathways after memory acquisition (Jones, 2018).

Go to top

Wnt signaling is required for long-term memory formation

Tan, Y., Yu, D., Busto, G. U., Wilson, C. and Davis, R. L. (2013). Cell Rep 4: 1082-1089. PubMed ID: 24035392

Wnt signaling regulates synaptic plasticity and neurogenesis in the adult nervous system, suggesting a potential role in behavioral processes. This study probed the requirement for Wnt signaling during olfactory memory formation in Drosophila using an inducible RNAi approach. Interfering with β-catenin expression in adult mushroom body neurons specifically impairs long-term memory (LTM) without altering short-term memory. The impairment is reversible, being rescued by expression of a wild-type β-catenin transgene, and correlates with disruption of a cellular LTM trace. Inhibition of wingless, a Wnt ligand, and arrow, a Wnt coreceptor, also impairs LTM. Wingless expression in wild-type flies is transiently elevated in the brain after LTM conditioning. Thus, inhibiting three key components of the Wnt signaling pathway in adult mushroom bodies impairs LTM, indicating that this pathway mechanistically underlies this specific form of memory (Tan, 2013).

This study was prompted by a previous discovery that a casein kinase Iγ homolog (CkIγ), gilgamesh (gish), is required for STM in Drosophila. CkIgγmediated phosphorylation of the cytoplasmic tail of Lrp5/6 (Arr) is crucial for Wnt/β-catenin signaling (Davidson, 2005), and it was predicted that disruption of the Wnt signaling pathway would perturb STM. Surprisingly, however, it was found that knockdown of the four Wnt signaling components leaves STM intact. The likely explanation for this discrepancy is that Gish serves other important functions in STM formation besides its role in LTM through phosphorylation of the Arr receptor (Tan, 2013).

How does Wnt signaling in the MB neurons mediate the formation of LTM? Since the normal expression of β-catenin, Wg, and Arr is required in the set of MB neurons defined by P{MB-GeneSwitch}12-1, and Wg is a short-range ligand, a model is favored in which the Wnt ligand, Wg, participates in an autocrine fashion in the MB neurons. Spaced conditioning, which produces long-term behavioral memory, but not massed or single-cycle conditioning, leads to a transient increase in wg expression in the MB neurons, perhaps as a step downstream of Creb. The subsequent secretion of Wg by the MB neurons activates the Fz/Arr receptor, leading to the accumulation of β-catenin in the MB neurons. β-catenin, in turn, orchestrates transcriptional changes in the MB neurons that are required for LTM, as well as the breaking and remaking of cell contacts through N-cadherin function, which is necessary for the reorganization of synapses for LTM storage. Recently, ribonucleoprotein particles containing synaptic protein transcripts were shown to exit the nucleus through a nuclear envelope budding process in response to Wnt signaling at the Drosophila neuromuscular junction (Speese, 2012). Wnt-dependent nuclear budding could provide the initial step for transporting RNAs to synapses for local protein synthesis and LTM formation (Tan, 2013).

Go to top

Fasting launches CRTC to facilitate long-term memory formation in Drosophila

Hirano, Y., Masuda, T., Naganos, S., Matsuno, M., Ueno, K., Miyashita, T., Horiuchi, J. and Saitoe, M. (2013). Science 339: 443-446. PubMed ID: 23349290

Canonical aversive long-term memory (LTM) formation in Drosophila requires multiple spaced trainings, whereas appetitive LTM can be formed after a single training. Appetitive LTM requires fasting prior to training, which increases motivation for food intake. However, this study found that fasting facilitates LTM formation in general; aversive LTM formation also occurred after single-cycle training when mild fasting was applied before training. Both fasting-dependent LTM (fLTM) and spaced training-dependent LTM (spLTM) requires protein synthesis and cyclic adenosine monophosphate response element-binding protein (CREB) activity. However, spLTM requires CREB activity in two neural populations--mushroom body and dorsal-anterior-lateral (DAL) neurons--whereas fLTM required CREB activity only in mushroom body neurons. fLTM uses the CREB coactivator CREB-regulated transcription coactivator (CRTC), whereas spLTM uses the coactivator CBP. Thus, flies use distinct LTM machinery depending on their hunger state (Hirano, 2013).

In Drosophila, canonical aversive long-term memory (LTM), which is dependent on de novo gene expression and protein synthesis, is generated after multiple rounds of spaced training. In contrast, appetitive LTM can be formed by single-cycle training. Because both aversive and appetitive LTM require protein synthesis and activation of CREB, it is likely that both types of LTM are formed by similar mechanisms. Appetitive and aversive LTM are known to differ (i.e., octopamine is specifically involved in appetitive but not aversive memory formation). However, it remains unclear why single-cycle training is sufficient for appetitive but not aversive LTM formation. Appetitive LTM cannot form unless fasting precedes training. Although fasting increases motivation for food intake (a requirement for appetitive memory) it was suspected that fasting may activate a second, motivation-independent, memory mechanism that facilitates LTM formation after single-cycle training (Hirano, 2013).

Flies were deprived of food for various periods of time and then subjected to aversive single-cycle training. Fasting prior to training significantly enhanced 1-day memory, with a peak at 16 hours of fasting and a return to nonfasting levels at 20 to 24 hours of fasting. In contrast, 16 hours of fasting did not increase short-term memory (STM, measured 1 hour after training). In this protocol, flies were returned to food vials after training, raising a possibility that the perception of food as a reward after training may enhance the previous aversive memory. This possibility was tested by inserting refeeding periods between food deprivation and training. Although fasting followed by a 4-hour refeeding period failed to induce appetitive LTM, it significantly enhanced aversive 1-day memory; this finding suggests that enhancement of aversive memory occurs through a mechanism unrelated to increased motivation or perception of food as a reward. A 6-hour refeeding period was sufficient to prevent aversive memory enhancement. Continuous food deprivation after training suppressed aversive memory enhancement, which indicates that both fasting before training and feeding after training are required to enhance aversive memory (Hirano, 2013).

Administration of the protein synthesis inhibitor cycloheximide (CHX) abolished 1-day memory enhancement but had no effect on 1-hour memory, supporting the idea that memory enhancement consists of an increase of LTM. Memory remaining after CHX treatment is likely to be protein synthesis-independent, anesthesia-resistant memory (ARM). Fasting for 16 hours neither enhanced protein synthesis-independent memory nor canonical aversive LTM generated by spaced training (spLTM). Furthermore, fasting-dependent memory decayed within 4 days, and food deprivation did not enhance 4-day spLTM, indicating that fasting-dependent memory is physiologically different from spLTM (Hirano, 2013).

Fasting-dependent memory was blocked by acute, dose-dependent, expression of CREB2-b, a repressor isoform of CREB, in the mushroom bodies (MBs). Expression of the repressor from two copies of UAS-CREB2-b under control of the MB247-Switch (MBsw) GAL4 driver, which induces UAS transgene expression upon RU486 feeding, significantly suppressed fasting-dependent memory upon RU486 feeding, whereas expression from one copy of UAS-CREB2-b did not. Defects in LTM formation are highly correlated with CREB2-b amounts. Significantly higher MBsw-dependent expression of CREB proteins was found in flies carrying two copies of UAS-CREB2-b relative to flies carrying one copy. MBsw-dependent CREB2-b expression did not affect STM in either fed or food-deprived conditions. Because the aversive memory enhanced by fasting is mediated by protein synthesis and CREB, this memory is referred to as fasting-dependent LTM (fLTM). Similar to the results in aversive fLTM, MBsw-dependent CREB2-b expression also decreased appetitive LTM but not appetitive STM (Hirano, 2013).

A recent study (Chen, 2012) concluded that CREB activity in MB neurons is not required for spLTM. In that study, CREB2-b was expressed using the OK107 MB driver and GAL80ts was used to restrict CREB2-b expression to 30°C. However, this study found that the GAL80ts construct still inhibited expression of CREB considerably at 30°C. When higher amounts of CREB2-b were acutely expressed in MBs using MBsw, a significant decrease was observed in 1-day spLTM, indicating that CREB activity in the MBs is likely to be required for spLTM (Hirano, 2013).

Expression of CREB2-b in two dorsal-anterior-lateral (DAL) neurons impaired aversive spLTM. In contrast, expression of CREB2-b in DAL neurons did not affect aversive fLTM. Moreover, appetitive LTM was also not affected by expression of CREB2-b in DAL neurons. MBsw did not express GAL4 in DAL neurons (Hirano, 2013).

CREB requires coactivators, including CBP (CREB-binding protein), to activate transcription needed for LTM formation. Acute expression of an inverted repeat of CBP (CBP-IR) in MBs significantly impaired spLTM without affecting either STM or 1-day memory after multiple massed trainings, which do not lead to LTM formation. However, neither aversive fLTM nor appetitive LTM was impaired by CBP-IR expression, indicating that an alternative coactivator may be required for fasting-dependent memory (Hirano, 2013).

Recent studies demonstrate the involvement of a cAMP-regulated transcriptional coactivator (CRTC) in hippocampal plasticity. In metabolic tissues, phosphorylated CRTC is sequestered in the cytoplasm while dephosphorylated CRTC translocates to the nucleus to promote CREB-dependent gene expression. Fasting causes CRTC dephosphorylation and activation. In line with this, significant accumulation of hemagglutinin (HA)-tagged CRTC (CRTC-HA) was found within MB nuclei after 16 hours of food deprivation. Subcellular fractionation indicated that food deprivation causes CRTC-HA nuclear translocation without affecting total CRTC-HA amounts (Hirano, 2013).

To examine the role of CRTC in fLTM and spLTM, a CRTC inverted repeat (CRTC-IR) was acutely expressed using MBsw, and suppression of aversive fLTM was observed but no effect was seen on STM. CHX treatment did not further decrease 1-day aversive memory, and CRTC-IR expression from a second MB driver, OK107, also impaired fLTM formation. CRTC-IR expression from MBsw also impaired appetitive LTM without affecting appetitive STM. In contrast, CRTC-IR expression from MBsw did not impair spLTM. CRTC-IR expression in DAL neurons had no effect on either aversive fLTM or appetitive LTM. Consistent with these results showing lack of fLTM after 24-hour fasting, 1-day aversive memory after 24-hour fasting did not decrease upon CRTC-IR expression in MBs (Hirano, 2013).

To examine the effects of spaced training on fLTM and the effects of fasting on spLTM, fed or fasted flies expressing either CBP-IR or CRTC-IR were space-trained. When CBP-IR was expressed to impair spLTM, 1-day memory after spaced training was impaired in fed conditions but not in fasting conditions, which suggested that spaced training protocols do not block fLTM. When CRTC-IR was expressed to impair fLTM formation, 1-day memory after spaced training was not affected by fasting, which suggested that mild fasting does not impair spLTM formation (Hirano, 2013).

Is activation of CRTC sufficient to generate fLTM in the absence of fasting? HA-tagged constitutively active CRTC (CRTC-SA-HA) was expressed from MBsw, and its nuclear accumulation was observed in the absence of fasting. Acute expression of CRTC-SA-HA from MBsw increased 1-day aversive memory after single-cycle training in fed flies, and this increase was not further enhanced by fasting. In contrast, expression of control CRTC-HA did not alter the fasting requirement for memory enhancement. CRTC-SA-HA expression did not affect feeding itself, which suggested that the memory enhancement is not due to impaired feeding. Taken together, CRTC activity in MBs is necessary and sufficient to form fLTM. Similar to the effects of fasting, CRTC-SA-HA expression did not affect STM or 4-day spLTM (Hirano, 2013).

In mammalian metabolic tissues, CRTC is phosphorylated by insulin signaling, which is suppressed by fasting. CRTC phosphorylation is also regulated by insulin signaling in flies. To determine whether reduced insulin signaling activates CRTC and promotes fLTM formation, heterozygous mutants for chico, which encodes an adaptor protein required for insulin signaling, were tested. Although chico1 null mutants are semilethal and defective for olfactory learning, heterozygous chico1/+ mutants are viable and display normal learning (Hirano, 2013).

CRTC accumulated in MB nuclei in chico1/+ mutants in the absence of food deprivation. Under conditions where flies were fed, chico1/+ flies had significantly greater 1-day memory after single-cycle training relative to control flies, whereas 1-hour memory was unaffected. Enhanced 1-day memory in chico1/+ flies was not further enhanced by fasting. Because the chico1/+ mutation does not affect feeding itself, the memory enhancement would not seem to be attributable to impaired feeding. The increased 1-day memory in chico1/+ mutants was suppressed by CHX treatment and CRTC-IR expression using MBsw, which suggests that reduced insulin signaling mimics fLTM through activation of CRTC in MBs (Hirano, 2013).

Single-cycle training after mild fasting generates both appetitive and aversive LTM, and CRTC in the MBs plays a key role in both types of LTM. A CRTC-dependent LTM pathway is unlikely to be involved in increasing motivation required to form appetitive memory, because CRTC knockdown did not affect appetitive STM and because CRTC-SA expression was not sufficient to form appetitive LTM without prior fasting. Although mild 16-hour fasting induced aversive fLTM, longer 24-hour fasting impaired aversive fLTM but not appetitive LTM. Thus, although aversive and appetitive fLTM share mechanistic similarities, they may be regulated by different inputs controlling motivation and fasting time courses. Because nuclear translocation of CRTC was sustained even after 24 hours of food deprivation, prolonged fasting may suppress a CRTC-independent step in aversive fLTM formation. spLTM was not affected by 24-hour fasting prior to training, which suggests that the unknown inhibitory effect of 24-hour fasting does not occur after spaced training. Continuous food deprivation after training suppressed aversive fLTM. Another study has reported that continuous food-deprivation after spaced training suppresses spLTM as well (Hirano, 2013).

Suppression of aversive LTM by prolonged fasting may ensure that starving flies pursue available food, with less concern for safety. Although the biological importance of aversive fLTM in natural environments is currently unclear, the current results indicate that different physiological states may induce different types of LTM in flies (Hirano, 2013).

Go to top

Muscarinic ACh receptors contribute to aversive olfactory learning in Drosophila

Silva, B., Molina-Fernandez, C., Ugalde, M. B., Tognarelli, E. I., Angel, C. and Campusano, J. M. (2015). Neural Plast 2015: 658918. PubMed ID: 26380118

The most studied form of associative learning in Drosophila consists in pairing an odorant, the conditioned stimulus (CS), with an unconditioned stimulus (US). The timely arrival of the CS and US information to a specific Drosophila brain association region, the mushroom bodies (MB), can induce new olfactory memories. Thus, the MB is considered a coincidence detector. It has been shown that olfactory information is conveyed to the MB through cholinergic inputs that activate acetylcholine (ACh) receptors, while the US is encoded by biogenic amine (BA) systems. This study evaluates the proposition that, as in mammals, GPCR muscarinic ACh receptors (mAChRs) contribute to memory formation in Drosophila. The results show that pharmacological and genetic blockade of mAChRs in MB disrupts olfactory aversive memory in larvae. This effect is not explained by an alteration in the ability of animals to respond to odorants or to execute motor programs. These results show that mAChRs in MB contribute to generating olfactory memories in Drosophila (Silva, 2015).

Go to top

Visualization of a distributed synaptic memory code in the Drosophila brain

Bilz, F., Geurten, B. R. H., Hancock, C. E., Widmann, A. and Fiala, A. (2020). Neuron. PubMed ID: 32268119

During associative conditioning, animals learn which sensory cues are predictive for positive or negative conditions. Because sensory cues are encoded by distributed neurons, one has to monitor plasticity across many synapses to capture how learned information is encoded. This study analyzed synaptic boutons of Kenyon cells of the Drosophila mushroom body gamma lobe, a brain structure that mediates olfactory learning. A fluorescent Ca(2+) sensor was expressed in single Kenyon cells so that axonal boutons could be assigned to distinct cells and Ca(2+) could be measured across many animals. Learning induced directed synaptic plasticity in specific compartments along the axons. Moreover, it was shown that odor-evoked Ca(2+) dynamics across boutons decorrelate as a result of associative learning. Information theory indicates that learning renders the stimulus representation more distinct compared with naive stimuli. These data reveal that synaptic boutons rather than cells act as individually modifiable units, and coherence among them is a memory-encoding parameter (Bilz, 2020).

Deciphering how brain circuits, the neurons they consist of, and their synaptic connections acquire and encode learned information is a key task in modern neuroscience. Decades of research have led to current understanding of changes in synaptic transmission as a key neuronal substrate underlying learning and memory formation. Sensory stimuli that can be learned and memorized are encoded in the brain as neuronal activity that is sparsely distributed across ensembles of neurons and levels of processing. Physical changes in synaptic transmission underlying the encoding of a specific memory are, therefore, also distributed across many neurons and synapses. It is assumed that, during learning, synaptic connections between neuronal ensembles that are active during the perception of a stimulus become modified such that their combined activity pattern can be retrieved during memory recall, thereby instructing future behavioral action. The sparsely distributed nature of these memory traces (or engrams) makes it challenging to experimentally determine the rules by which many individual synaptic connections change. It is difficult to monitor plasticity in individual synapses at sufficient resolution and to observe many synapses comprehensively at the same time (Bilz, 2020).

This problem was addressed using the fruit fly Drosophila melanogaster, a key model organism for the analysis of neuronal substrates underlying learning and memory. The system is advantageous because it combines few but often genetically tractable neurons with a behavioral repertoire and neuronal complexity rich enough to allow for conceptual comparison with mammals. In Drosophila, classical olfactory conditioning is a widely used learning paradigm. In this training procedure, animals learn to avoid or approach a specific odor as a conditioned stimulus (CS) when it is temporally paired with a punishing or rewarding unconditional stimulus (US), such as an electric shock or sugar. Odors are detected by ~1,320 olfactory sensory neurons per hemisphere, located on the third antennal segments and maxillary palps, that project to the glomeruli of the antennal lobes, which are the structural and functional analogs of the vertebrate olfactory bulbs. Second-order olfactory projection neurons relay the processed odor information to the lateral horn and the mushroom body (MB) calyx; this structure forms the main sensory input region of the MB, which consists of approximately 2,000-2,500 intrinsic neurons (Kenyon cells [KCs]). At the projection neuron-to-KC synapses, odor information is transformed from highly combinatorial neuronal activity (dense code)-analogous to the situation in the mammalian olfactory bulb-to a nonstereotypic and sparsely distributed pattern of KC activity (sparse code)-similar to the situation in the anterior piriform cortex of mammals. The parallel bundles of KC axons collectively form the MB lobes. The site of coincidence between the CS and US and the synaptic circuitry that underlies associative olfactory learning and short-term memory is confined to the γ lobes, which contain approximately 650 KCs per hemisphere. The γ lobes are divided into five zonal compartments (γ1-γ5), each of which is defined by the dendritic innervations of only one or two MB output neurons (MBONs), whose dendritic trees integrate information across multiple KCs (Bilz, 2020).

MBONs show behavior-instructive properties: optogenetic activation of MBONs that innervate the γ1, γ2, or γ3 compartments induces behavioral attraction toward the light source, whereas optogenetic activation of MBONs that innervate the γ4 and γ5 compartments induces behavioral repulsion from light. In addition, the γ lobes also receive compartmentalized input from dopaminergic neurons (DANs), which exert punishing or rewarding US-signaling properties, similar to the function of DANs in mammals. Dopaminergic neurons that innervate the γ1 and γ2 compartments mediate punishment. In contrast, DANs that innervate the γ4 and γ5 compartments signal reward. Importantly, synaptic plasticity induced by optogenetic stimulation of DANs is presynaptically localized to KCs, but not postsynaptically localized to MBONs. Postsynaptic changes in odor-evoked MBON activity as a result of associative learning or optogenetic activation of DANs have been determined electrophysiologically and by using optical Ca2+ imaging. However, learning-induced presynaptic plasticity in axonal KC synapses has never been directly observed. The sparsely distributed and stochastic nature of odor representations and the density of axonal KC fibers make an analysis of plasticity at the presynaptic level extremely challenging. Thus, the rules by which KC presynapses change in order to convert a high-dimensional sensory code into a low-dimensional, behavior-instructing output remain unclear. To elucidate the changes that KC synapses undergo following associative learning, this study used Ca2+ imaging to visualize synaptic activity in Drosophila at the exact integration point between the CS and US signals (i.e., at the axonal KC boutons) (Bilz, 2020).

This paper quantitatively determined stimulus-evoked Ca2+ dynamics at axonal synaptic boutons within individual KCs and distributed across KCs of the MB. These data unexpectedly revealed that boutons along KC axons are not uniformly activated by odors. Rather, individual boutons, even of the same neurons, show individualized responsiveness. However, boutons located within the same compartments γ2-γ4 showed a more coherent odor tuning, implicating that these compartments act as functional units, the borders of which are likely determined by input neurons, such as DANs. In fact, compartmentalized modulation of subcellular cyclic AMP (cAMP) levels in KCs, driven by input DANs, has been shown. The γ5 compartment represented an exception in that the KC boutons showed less correlated odor-evoked activity, which might underlie the reported finding that γ5 MBONs do not reliably respond to odors. The variability in odor responsiveness between individual boutons of even the same KCs is reminiscent of the plasticity- and rutabaga-dependent individualization of MBONs, which can be interpreted as a reflection of an animal's individual olfactory prior experience. Because MBONs are the direct downstream targets of axonal KC synapses, and dopamine-dependent plasticity has been shown to be presynaptically localized to KC boutons, it is suggested that the variability between boutons might represent the source of individualization of MBON responses. The concept that learning-dependent plasticity is localized to axonal KC presynapses and not to MBON postsynapses is also corroborated by reports that reversibly blocking transmitter release from KCs during training only does not prevent the acquisition of an associative memory (Bilz, 2020).

Mechanistically, the phenomenon of functional individualization of synaptic boutons has been analyzed in motor neurons of the neuromuscular junctions in larval Drosophila. Here, octopamine acts as a neuromodulator that alters cAMP levels in synaptic boutons, similar to the action of dopamine in KCs. The distribution of the phosphodiesterase Dunce located at the peripheral rims of boutons restricts the diffusion of cAMP to individual boutons. Because Dunce is enriched in KCs and is required for proper olfactory learning, future research would be well placed to test whether Dunce mediates functional individualization of synaptic boutons in KCs also. Individualized modulation of synaptic boutons, rather than entire neurons, is also reminiscent of the situation in mammals. Here, dendritic spines have been shown to be selectively modified in the course of learning, implying that synapses and spines represent basal elements of learning and memory formation rather than entire neurons (Bilz, 2020).

To analyze how learning modifies the synaptically distributed odor representation, the animals were subjected to an aversive associative training procedure. Previous studies on MBONs have shown that pairing an odor with electric shock or artificial activation of punishment-mediating DANs leads to a depression of the activity of MBONs innervating the γ1 and γ2 compartments. The optogenetic activation of these MBONs confers behavioral attraction of animals toward the light source. The function of the γ3 compartment in associative learning is less clear. However, punitive electric shocks evoke activity in DANs innervating γ1, γ2, and γ3, in contrast to a rewarding sugar stimulus that activates DANs targeting γ4 and γ5. Therefore, the predominant hypothesis in light of the above findings is that the relative activity of compartment-specific MBONs encodes the learned valence of odor stimuli. In agreement with this hypothesis, this study shows that the median Ca2+ response amplitudes across the γ2 and γ3 compartments are reduced after aversive conditioning, homogeneous depression was not observed but bidirectional modulation was observed across these bouton populations. As a novel finding, a complementary effect for the CS− condition was observed; i.e., more boutons showed an increase in activity in the γ4 compartment. It is well established that, in the course of differential olfactory conditioning, the CS− is learned as indicative for the absence of punishment. The finding of opposite effects for CS+ and CS−, both in the direction of changes in Ca2+ activity and different γ lobe compartments (γ2 versus γ4), modifies the prevailing model of associative learning. However, it must also be noted that the current experiments differed from many previously reported approaches in two aspects. First, electric shock punishment was used as an US, in contrast to artificial activation of single DANs. Second, the animals were subjected to a training procedure while they were restrained under a microscope, allowing for a within-animal comparison of before and after training. This approach differs from the comparison of untrained animals with animals 1 h after associative training (Bilz, 2020).

The current findings relate to one aspect of classical conditioning: the learned stimulus representation embodies behavior-instructive properties. Specifically, in the case of aversive olfactory conditioning, it induces repulsion from the odor source. However, a second aspect of associative learning is the 'memory content.' The specific stimulus representation (i.e., odor identity) must be demarcated as outcome predictive from other representations that do not carry any learned valence. The sparse nature of KC odor tuning and the relatively large array of KCs ensures a high degree of odor specificity, although γ-type KCs are less selective compared with α/β and α'/β' KCs. However, odor discriminability of individual KCs cannot be complete; KCs respond in some instances to more than one odorant, such as both the trained CS+ and CS− conditions. This was also the case in some synaptic KC responses to the odorants measured in this study. The degree of overlap between neuronal stimulus representations determines the degree to which a learned response toward one stimulus is generalized to a second stimulus. However, through differential training, flies can learn to disentangle similar odorants and increase their ability to differentiate between them. This phenomenon cannot be explained by gross facilitation or depression of KC presynapses. This study found that association of an odor with punishment caused a strong decorrelation of odor-evoked Ca2+ activity across boutons of most γ lobe compartments, both between and within individual KCs, and independent from valence (aversive punishment and appetitive relief learning). Therefore, overlapping KC boutons can be part of both the CS+ and CS− representations, dependent on correlations with the remaining synapses that encode the respective stimuli. It is suggested that this could represent a mechanism by which synaptic interference between overlapping neuronal assemblies encoding for learned, aversive odors (CS+) and similar odors not predictive for punishment (CS−) may be circumvented. The finding that synaptic boutons show decorrelated activity in response to a trained stimulus suggests that the particular odor representation is also uniquely demarcated from other odor representations. Coding theory suggests that stimulus-specific information is transmitted most efficiently if the units encoding the stimulus are decorrelated, reducing redundancy. Exactly this is quantitatively shown in this study in the gain in entropy through associative training that was found for the synaptic odor representation. This principle is not confined to the MB or insect brains. It has, for example, recently been shown that, in the rat hippocampus, an increased variability in synaptic activity within populations of CA1 neurons as a result of learning increases the information coding ability of this brain region (Bilz, 2020).

Currently clear understanding is lacking of the physiological mechanisms that enable synaptic boutons to correlate or decorrelate Ca2+ activity both temporally and in amplitude, depending on whether the driving stimulus has been trained. Coherent neuronal activity, be it correlations between spike trains or between synaptic Ca2+ transients, is regarded as an emergent property of a neuronal circuit that typically depends on the balance between excitatory and inhibitory input. Recent connectomic studies based on electron microscopic 3D reconstructions of MBs of larval and adult Drosophila have uncovered remarkable microcircuitry complexity at the axonal KC bouton level. For example, groups of KCs are synaptically interconnected with each other, and KC-MBON connections are often confined to Rosetta-like structures indicative of cooperative action of KC outputs. The KC-KC and KC-MBON connections also show electrical coupling through gap junctions. Synchronous, correlated neuronal activity often depends on gap junctions, such as in the olfactory bulb of mammals, in the antennal lobe of insects, or in the mammalian cortex. Thus, it is tempting to speculate that modulation of gap junctions might potentially affect correlative activity of KC boutons as a result of associative training. Regardless of the exact physiological mechanism, the data suggest that correlations at the synaptic circuit level contribute to encoding learned information, thereby combining the concept of a synaptic (rather than cellular) distribution of a memory code with that of correlated neuronal activity as a coding parameter (Bilz, 2020)

Go to top

Presynaptic developmental plasticity allows robust sparse wiring of the Drosophila mushroom body

Elkahlah, N. A., Rogow, J. A., Ahmed, M. and Clowney, E. J. (2020). Elife 9. PubMed ID: 31913123

Presynaptic developmental plasticity allows robust sparse wiring of the Drosophila mushroom body

In order to represent complex stimuli, principle neurons of associative learning regions receive combinatorial sensory inputs. Density of combinatorial innervation is theorized to determine the number of distinct stimuli that can be represented and distinguished from one another, with sparse innervation thought to optimize the complexity of representations in networks of limited size. How the convergence of combinatorial inputs to principle neurons of associative brain regions is established during development is unknown. This study explored the developmental patterning of sparse olfactory inputs to Kenyon cells of the Drosophila melanogaster mushroom body. By manipulating the ratio between pre- and post-synaptic cells, it was found that postsynaptic Kenyon cells set convergence ratio: Kenyon cells produce fixed distributions of dendritic claws while presynaptic processes are plastic. Moreover, this study showed that sparse odor responses are preserved in mushroom bodies with reduced cellular repertoires, suggesting that developmental specification of convergence ratio allows functional robustness (Elkahlah, 2020).

The environmental stimuli animals encounter on a day-to-day basis are extraordinarily numerous. Olfactory systems have evolved to cope with this diversity by maximizing the chemicals that can be detected, through the amplification of chemosensory receptor gene families, and through combinatorial coding, which expands representation capacity from the number of receptors in the genome to the number of combinations among them. The arthropod mushroom body is a cerebellum-like associative learning structure with a well-understood role in representing sensory stimuli and associating sensory and contextual cues. While mushroom bodies of different insect species process information from a variety of sensory modalities, 90% of Kenyon cell inputs in Drosophila melanogaster are olfactory. The mushroom body of each hemisphere has ~2000 Kenyon cells (KCs), which are two synapses from the sensory periphery. Each olfactory receptor neuron in the antennae of adult flies expresses one or two of 75 olfactory receptor genes encoded in the genome. The axons of neurons expressing the same receptor converge on one of 54 glomeruli in the antennal lobe. Approximately 150 uniglomerular projection neurons (PNs) have dendrites in one of the 54 glomeruli and carry signals about distinct receptor channels to two regions of the protocerebrum, the lateral horn and the mushroom body calyx. PN inputs to the lateral horn are thought to underlie innate behaviors, while inputs to the mushroom body allow flexible learned association of odor stimuli with behavioral context (Elkahlah, 2020).

In the mushroom body calyx, the presynaptic sites of individual olfactory PNs cluster into multi-synaptic boutons, with PNs of different types (innervating different glomeruli) producing consistent, characteristic bouton numbers. Each PN makes 1-20 boutons, and each bouton is wrapped by claws of ~10 KCs, such that each PN sends output to between 10 and 200 of the 2000 KCs. KCs in turn have 3-10 (average of five) claws, which innervate boutons of various PNs. Each KC therefore receives innervation from only a minority of the 54 incoming sensory channels, and different individual KCs receive different and relatively unstructured combinations of inputs. The sets of inputs to individual cells vary across hemispheres and likely across individuals. Associative learning mechanisms operate at KC output synapses, in the mushroom body axonal lobes, to re-weight KC outputs depending on experience and shift animal behavior (Elkahlah, 2020).

The mushroom body is a simplified and experimentally tractable example of an expansion layer, in which a set of sensory inputs is mapped combinatorially onto a much larger set of postsynaptic cells, increasing the dimensionality of sensory representations. Like the diversification of antibodies by V(D)J recombination, the diversification of sensory input combinations across KCs is thought to allow them to represent arbitrary odors, regardless of evolutionary experience. Neurons of many other expansion layers receive similarly few, or sparse, sensory inputs. These include the cerebellum proper, the electric organ of mormyrid fish, the dorsal cochlear nucleus, and the hippocampus. Cerebellar granule cells have an average of four large, claw-shaped dendrites that are innervated by clustered mossy fiber presynaptic sites in mossy fiber rosettes. The similar convergence ratios in the cerebellum and mushroom body (4 or 5 sensory inputs, respectively, per expansion layer cell) are thought to maximize dimensionality of sensory representations by optimizing the tradeoff between stimulus representation, which is maximized when expansion layer neurons receive large combinations of inputs, and stimulus separation, which is maximized when expansion layer neurons receive few inputs. The number of sensory inputs received by expansion layer neurons is thus a crucial parameter in sensory coding. How the density of inputs to expansion layer neurons is developmentally programmed is not understood in any system (Elkahlah, 2020).

Innervation complexity more generally has been studied in the peripheral nervous system and in the developing mammalian cortex. In peripheral sensory neurons, most prominently those of the Drosophila larval body wall, cell-autonomous mechanisms profoundly influence dendritic complexity. However, sensory neurons do not need to coordinate their innervation with presynaptic partners. In the vertebrate peripheral nervous system, including the rabbit ciliary ganglion and vertebrate neuromuscular junction, postsynaptic neurons or muscles are thought to dictate wiring complexity. In contrast, in the developing cortex, extracellular signals including BDNF play a strong role in influencing dendritic complexity, suggesting that presynaptic cells and glia also influence connectivity density. Therefore, while mechanisms acting in both pre- and post-synaptic cells can influence innervation complexity, there is a need to directly compare how pre- and post-synaptic cells influence one another (Elkahlah, 2020).

This study sought to ask how convergence ratio is set in the mushroom body calyx. By bidirectionally varying the populations of pre- and post-synaptic cells, it was possible to make many different mushroom body aberrations. Across these conditions, a consistent pattern of compensations was found: the number of claws per KC remained largely constant, while the number of presynaptic boutons per olfactory PN varied bidirectionally and in response to changes of both the PN and KC populations. It is therefore concluded that in this circuit, connectivity density is set by aspects of KC specification and is accomplished by flexible innervation of the calyx by PNs (Elkahlah, 2020).

Cerebellum-like architectures are found repeatedly in phylogenetically distinct organisms, suggesting that they have broad evolutionary utility. Yet the developmental production of sparseness, a key wiring feature allowing high-dimensional sensory representations, is not understood. This study begins to investigate the development of sparse connectivity in the cerebellum-like associative learning center of insects, the mushroom body. By varying the ratio between presynaptic olfactory PNs and postsynaptic KCs, it was found that connectivity density is set by KCs: KC claw number changes little across a wide range of PN::KC ratios, KC number predicts PN bouton number, and PNs exhibit wide variety in bouton number to satisfy KC demands. This strategy for generating connectivity density would preserve sparseness in KC inputs across developmental time and upon evolutionary changes in cell repertoires and thus maintain the olfactory coding function of the mushroom body over a range of conditions (Elkahlah, 2020).

Implications for development: Different projection neuron to Kenyon cell ratios across developmental stages

For many animals, brain development continues while juveniles are already living freely--searching for food, navigating, and learning about the environment. Developmental transitions in brain structure are particularly stark in holometabolous insects, who build the brain twice. In D. mel, neurons generated during embryonic and early larval stages wire the larval nervous system. These circuits support larval behaviors, while neural stem cells continue to generate additional neurons that will be used to build the adult circuit. In keeping with this, the ratio between PNs and KCs in the larval olfactory circuit is starkly different from the adult: 21 embryonically-born PNs wire to early-born KCs to construct the larval mushroom body . Connections among these populations dissolve in early pupae, and are then re-wired during pupal development, joined by ~100 more larvally-born PNs, and >1000 more KCs per hemisphere that continue to be born until immediately before adult eclosion (Elkahlah, 2020).

The 21 PNs in the early larva connect to ~75 KCs, a 1:3 ratio, while in the adult, ~150 PNs connect to ~2000 KCs, a 1:10 ratio.. This study found that unlike cells in many other systems, including the vertebrate cerebellum, PNs and KCs did not rely on each other for survival signals. This may be due to the constantly changing ratio between these cell types across developmental time. Instead, setting connectivity density cell-autonomously in KCs could allow KCs to obtain the appropriate number of inputs at the different life stages of the animal, when cellular constituents are very different from one another. Similarly, while PN neurogenesis ceases well before PNs and KCs begin to contact one another in the pupa, it is estimated that ~10% of KCs are born after PN:KC synapsis has already initiated. Strict, cell-autonomous dendrite structuring and flexible PN bouton production could together ensure that late-born KCs obtain the inputs appropriate to support coding (Elkahlah, 2020).

Implications for coding: Balancing projection neuron representations across the calyx

Olfactory PNs of different types are specified in a predictable temporal order, have characteristic numbers of boutons, and overlap in their innervation of the calyx. Differences in bouton number across different PNs allow different odor channels to be differentially represented in the calyx and in KC odor responses. Several classes of PNs also differ in number between the sexes. While PNs changed their individual bouton repertoires in response to changes in cell repertoires, this study found that to some extent, the representation level of different PNs in the calyx was preserved. For example, this study shows the effect of reducing KC number on bouton production by the VM6 PN and 42D01 PNs. While each population of PNs reduced individual bouton number in this condition, they retained their typical relative representation. The VM6 PN reduced its boutons from 10 to 5, while the 42D01 PNs decreased their boutons from 4 to 2. Similarly, this study expanded the PN population by inducing ectopic PN neuroblast duplication. In these experiments, amplification of the ventrolateral clone was mainly observed. It was found that individual anterodorsal VM6 cells did not scale down their boutons when the ventrolateral PN clone expanded, but only when VM6 itself was duplicated. Again, this could maintain the relative wild type representations of different odor channels in the calyx. A recent analysis suggests that the spontaneous activity of different ORs correlates with number of boutons representing that odor channel in the calyx. One possible model for how PNs scale to KC numbers while maintaining their relative representations in the calyx is thus that KC number limits total bouton number across all PNs, while allocation of these boutons to individual PNs is determined by activity-based competition among PN types (Elkahlah, 2020).

Implications for coding: Maximizing dimensionality of odor representations

Qualitative aspects of sparse coding in the mushroom body appear robust to severe perturbations to the circuit. Alternative developmental compensatory mechanisms would be much less likely to preserve sparse coding. For example, this study increased the ratio of PNs to KCs in two ways, by increasing the number of PNs and by decreasing the number of KCs. In both cases, PNs dialed down their bouton number, making 25-50% of the boutons they make in wild type. This allowed the KCs to receive their typical number of inputs. If in contrast bouton number was rigid and claw number flexible, in these cases KCs would have expanded their claw production 2-4 fold to innervate all the incoming PN boutons. Individual KCs with for example 20 instead of 5 claws would receive input from ~40% of glomeruli, increasing the overlap in tuning across different KCs and degrading the ability of the network to distinguish different stimuli from one another (Elkahlah, 2020).

In two other cases, this study increased the ratio of KCs to PNs, by increasing the number of KCs and by decreasing the number of PNs. Again, KCs retained their typical claw number. If instead PNs had maintained a static production of boutons while KCs had adjusted their claw production, KCs would receive very few inputs. While increasing the number of inputs per KC is theorized to reduce dimensionality of odor responses by making different KCs more similar to one another, decreasing the number of inputs per KC is theorized to reduce dimensionality by reducing the number of different possible KC input combinations. That this sweet spot maximizing dimensionality, ~5 inputs per cell, is programmed into KC identity testifies to the evolutionary importance of maintaining connectivity density in associative brain centers that rely on combinatorial coding (Elkahlah, 2020).

Implications for evolution: The potential for mushroom body function despite perturbations

The olfactory receptors are the largest insect gene family and have been subject to frequent and extreme gains and losses in many clades. Similarly, brain centers devoted to learning are radically different across species, as exemplified by the diversity in KC repertoire across arthropods. In order to acquire a novel olfactory circuit, many different evolutionary changes are required: A new receptor must evolve, an OSN type that uniquely expresses the receptor needs to arise, that OSN needs to synapse onto PNs, and a new PN type and new glomerulus must arise. For these events to accrue over time, each individual change must be compatible with continued circuit function and olfactory behavior. While development of a dedicated circuit that assigns an innate meaning to a newly-detectable odor would require many further changes, the signal could add to fitness immediately through representation in the mushroom body (Elkahlah, 2020).

This study has described two mechanisms of developmental robustness that maintain coherent mushroom body wiring in the face of a broad range of phenotypic alterations. First, it was observed that olfactory PNs can adjust to gain and loss of PNs while maintaining the balance of odor channel representations in the calyx. Plastic development of PN presynaptic sites that makes room for additional players in the repertoire would allow immediate access of evolutionarily expanded PNs to the calyx and the use of their signals for olfactory learning, thus making time for the evolution of hardwired circuits for innate interpretations. Second, this study showed that developmental programs wiring the calyx can accommodate variation in KC number from at least 1/4 to 2-fold the endogenous complement. Again, this flexibility could support continued MB function on the path to the evolution of mushroom body innovations. Future experiments will ask how KC claw number is developmentally programmed, and what mechanisms operate in olfactory PNs to allow them to tailor bouton production to the KC repertoire (Elkahlah, 2020).

Go to top

Spaced training forms complementary long-term memories of opposite valence in Drosophila

Jacob, P. F. and Waddell, S. (2020). Neuron. PubMed ID: 32289250

Spaced training forms complementary long-term memories of opposite valence in Drosophila

Forming long-term memory (LTM) often requires repetitive experience spread over time. Studies in Drosophila suggest aversive olfactory LTM is optimal after spaced training, multiple trials of differential odor conditioning with rest intervals. Memory after spaced training is frequently compared to that after the same number of trials without intervals. This study shows that, after spaced training, flies acquire additional information and form an aversive memory for the shock-paired odor and a slowly emerging and more persistent 'safety-memory' for the explicitly unpaired odor. Safety-memory acquisition requires repetition, order, and spacing of the training trials and relies on triggering specific rewarding dopaminergic neurons. Co-existence of aversive and safety memories is evident as depression of odor-specific responses at different combinations of junctions in the mushroom body output network; combining two outputs appears to signal relative safety. Having complementary aversive and safety memories augments LTM performance after spaced training by making the odor preference more certain (Jacob, 2020).

The gain in memory performance obtained from spacing learning sessions has intrigued scientists for over a century. Early work using Drosophila demonstrated that spaced training produced protein-synthesis-dependent 'aversive LTM', whereas massed training did not. Many subsequent studies have compared memory after spaced training to that following massed training. This study found that flies learn additional safety information for the CS odor when subjected to spaced training. Parallel complementary CS+ aversive and CS approach memories therefore account for the discriminative odor preference observed 24 h after differential spaced training. In contrast, flies only form an avoidance memory for the shock-paired odor when they are mass trained. Surprisingly, radish mutant flies did not form CS+ aversive memory after spaced training, yet their CS memory appeared unaffected. In contrast, CXM feeding abolished CS memory, but CS+ memory was not significantly reduced. If previous operational definitions are used, these data suggest that CS memory is protein-synthesis-dependent LTM, whereas the CS+ component is ARM. It is therefore important to rethink the many prior studies that have assumed they were measuring only avoidance of CS+ after spaced training (Jacob, 2020).

Recording a timeline of performance after spaced training revealed that CS+ avoidance and CS approach memories have a very different dynamic. The CS+ avoidance memory was evident immediately after training, rapidly decayed over 24 h, and was absent at 4 days. In stark contrast, CS approach memory emerged slowly after training and lasted for at least 4 days-a trajectory reminiscent of that of long-term appetitive memory reinforced by nutritious sugar. The discovery that the processes underlying CS+ (ARM) and CS (LTM) memories have different timing, and different anatomical locations, gives the previously reported mechanistic differences an entirely new perspective. The current data suggest that, rather than occurring in the same neurons, ARM and LTM represent each of the odors employed in differential spaced training. They are therefore likely to be represented in unique populations of odor-activated KCs. In addition, different DANs reinforce CS+ ARM and CS LTM at different KC-MBON junctions (see Circuit diagrams of the mushroom body). It follows that, after spaced training, processing of CS+ ARM, which includes the mushroom body-enriched radish encoded Rap GAP, will occur in different KCs and at different KC locations and output synapses than do the molecular mechanisms that underlie protein-synthesis-dependent CS LTM (Jacob, 2020).

The valence of olfactory memories can be reversed from aversive to appetitive if the relative timing of odor and reinforcement is altered during training. If shock, or artificial DAN activation, is presented ~45 s before the odor, flies form an appetitive relief memory for that odor. Experiments with artificial DAN activation suggest that relief learning is represented by dopamine potentiating an MBON's response to the conditioned odor. If spaced training utilized the same relief-from-punishment mechanism as that in a previous study, CS approach memory would be coded as potentiation of the same connections as those coding CS+ avoidance as a depression. However, this study observed co-existence of aversive and approach memories at different places in the MBON network. The data instead indicate that CS approach is coded by specific appetitively reinforcing DANs that direct depression of KC outputs onto corresponding MBONs. This study also explicitly tested whether spaced relief training could form an equivalent long-term CS approach memory. These experiments demonstrated that the memory formed differs greatly from that formed after differential spaced training. Most importantly, memory after spaced relief training can be measured immediately but does not persist for 24 h. CS memory after spaced training emerges slowly and persists for at least 4 days. It is therefore proposed that CS approach after spaced training reflects a safety memory for the CS, rather than that the CS has been associated with the cessation of punishment. Relief and safety learning are also different in rodents. It is proposed that the reason massed training does not form CS approach memory is that it lacks a period of safety after each CS presentation (Jacob, 2020).

Aversive LTM performance, after spaced training, is largely considered to rely on αβ KCs and to be retrieved via α2sc (MB-V2) MBONs. However, others have indicated that the network properties are more distributed and that output from γ3, &gamma3β'1- and α3MBONs is required to retrieve aversive LTM. The current work suggests that there are different reasons why blocking these MBONs during testing impairs 24 h memory after spaced training. Consistent with prior work, this study recorded depressed responses to CS+ in α2sc-MBONs 24 h after spaced training. Depression of α2sc-MBON responses is therefore critical if flies are to express CS+ avoidance. This study also observed strong depression of MBON-α3 CS+ responses after spaced training. The role for α3 MBONs has been disputed. At this point it is not possible to reconcile differences between the studies, other than perhaps the number of training trials, strength of reinforcement, and relative hunger state of the flies. It is also noted that many recent studies use robots where flies remain in the same tube for the entire training session. In contrast, earlier studies and the current experiments utilized manual training where flies are transferred from the training chamber between trials. Nevertheless, the current data suggest that α2sc- and α3- MBONs house plasticity relevant for expression of CS+ aversive memory (Jacob, 2020).

It has been reported that MBON-β2mp and MBON-γ5β'2α (M4/6) are not required for LTM retrieval after spaced training. However, this study found that appropriately ordered CS+/CS spaced trials depressed responses to CS in β' 2mp-MBONs. In addition, this study found that PAM-b' 2mp DANs are required for the formation of CS approach memory. The current results therefore indicate a specific role for the β' 2mp subcompartment of the β' 2 MB zone and that MBON-β' 2mp plasticity is required to express CS approach memory (Jacob, 2020).

The negative sign of odor response plasticity of α2sc-, α3-, and β2mp-MBONs makes intuitive sense with the known valence of these pathways. Responses to CS+ in approach-directing α2sc- and α3- MBONs were depressed, which would favor odor avoidance. In contrast, depressing responses to CS to avoidance-directing β'2mp-MBONs should promote odor approach (Jacob, 2020).

This study also discovered roles for PAM-γ3 and PAM-β'1 DANs and recorded traces of both CS+ and CS memory in the corresponding γ3,γ3b'1-MBONs. MBON dendrites in the γ3 compartment showed a decreased response to the CS+, irrespective of the order of CS+ and CS in the training trials, consistent with the rules of forming aversive CS+ memory. In contrast, CS responses were decreased in the β'1 tuft of γ3β'1 MBON dendrites, but only if flies were trained with CS+ and then CS, in that order. Plasticity in β'1 of the γ3β'1 MBON therefore followed the order rule observed for conditioning CS approach behavior. Interestingly, recording in the axons of g3 and g3b01 MBONs suggested that CS+ and CS plasticity cancel each other out. Unfortunately, the split-GAL4 used for driving GCaMP expression in g3b01 MBONs also labels g3 MBONs. Therefore, although only the γ3β'1 MBONs have a dendrite in both γ3 and β'1 compartments, it is not possible at this stage to be certain that γ3β'1 MBONs alone integrate CS+ and CS memory traces (Jacob, 2020).

To decipher the relative role of γ3 and β'1 plasticity, this study individually blocked output from PAM-γ3 and PAM-β'1 DANs during training and tested the resulting memories. Behavioral observations after the PAM-γ3 block were particularly revealing. PAM-γ3 DANs respond to shock, and their forced activation reinforces aversive memories. However, blocking PAM-γ3 DANs during spaced training did not impair CS+ avoidance and instead impaired CS approach when flies were tested with CS versus a novel odor. A CS memory defect was also observed when the appetitively reinforcing PAM β'1 DANs were blocked during training, although this manipulation also impaired CS+ versus CS performance. Lastly, blocking the γ3, γ3β'1-MBONs during testing selectively impaired expression of CS but not CS+ memory. It is therefore proposed that γ3β'1 MBONs integrate the γ3 CS+ danger and β'1 CS safety plasticity to compute a relative safety signal. The importance of this is only obvious if the γ3,γ3β'1-MBONs are blocked during testing or remove aversive CS+ plasticity in γ3 and thereby reveal the behavioral consequence of unopposed CS plasticity in the β'1 region of γ3β'1 MBONs. Since MBON-γ3 and MBON-γ3β'1 are GABAergic, spaced training sequentially alters the level of CS+ and CS driven inhibition that is imposed on their downstream target neurons (Jacob, 2020).

The results of this study demonstrate that DANs reinforce the delayed recognition of safety. Formation of CS approach memory requires appetitively reinforcing PAM-β'2mp and PAM-β'1 DANs and, surprisingly, aversively reinforcing PAM-γ3 DANs. As noted above, PAM-γ3 DANs most likely provide an aversive teaching signal that directs CS+ plasticity in the γ3 region of MBON-γ3β'1 dendrites. Blocking output from PAM-β'2mp, PAM-β'1, or PAM-γ3 DANs, which are presumably responsible for each part of the LTM-correlated plasticity, reveals they are required for the formation of CS approach memories during training. Blocking most PAM DANs further localized an essential role during CS presentation in each spaced training trial, suggesting that safety-memory formation is driven by CS odor. However, safety-memory formation also requires that each CS+ exposure precedes each CS exposure in each training trial. Therefore, PAM DANs also have to somehow register a temporally locked negatively reinforced CS+ reference to be able to classify the following CS as safe. Lastly, repetition is a necessary element of triggering DANs to code safety. Imaging of the activity of appetitively reinforcing PAM-β2mp and PAM-β'1 DANs during and after training suggests they gradually acquire the capacity to reinforce CS approach memory across differential spaced-training trial repetitions. Both PAM-β'2mp and PAM-β'1 DANs exhibited an increased activation by CS odor, relative to the CS+, over consecutive training trials, and this difference was particularly clear when activity after the sixth training trial was compared to activity before training. In addition, the shock responsiveness of PAM-β2mp appeared to diminish over time. It is proposed that over repetitive trials the CS odor becomes the trigger that activates PAM-β2mp and PAM-β'1 DANs (Jacob, 2020).

It is conceivable that formation of long-term CS+ and CS memories is orchestrated by aversive reinforcement signals provided by the PPL1-γ1pedc (MP1) and PPL1-γ2α'1 (MV1) DANs in each shock-paired CS+ trial. PPL1-γ1pedc DANs code aversive learning by depressing odor-specific input to feedforward GABAergic g1pedc>a;αβ (MVP2) MBONs. Although MVP2 output is only required for the expression of short-term aversive memory, the plasticity remains for several hours. Each shock-reinforced odor trial therefore changes the state of the rest of the MBON and DAN network for subsequent exposures and reinforced trials. This has been proposed to release PPL1- α'2α2 DANs so they can reinforce LTM at the KC-MBON- α2sc junction. A similar release from inhibition of PPL1-α3, PAM-β'2mp, and PAM-β'1 DANs could account for spaced-training-driven plasticity at MBON-α3 and prime the PAM-β'2mp and PAM-β'1 to reinforce the CS memory (Jacob, 2020).

However, the data instead suggest that plasticity of the GABAergic γ3β'1 MBONs is essential for the formation of safety memory. Whereas blocking all PPL1 DANs abolished CS+ memory, CS memory was unaffected by this manipulation. In contrast, blocking shock-activated PAM-γ3 DANs during training selectively impaired the formation of CS memory. It is therefore proposed that spaced-training-evoked PAM-γ3 DAN activity cumulatively depresses CS+ driven activity of γ3β'1 MBONs, and this releases the PAM-β'1 and PAM- β2mp DANs from inhibition to reinforce CS memory. Such a model potentially explains the required relationship between CS+ and CS memories, the need for trial repetition, and the relative increase in the responses of these DANs to CS with each training trial. Although the results do not provide an explanation for the optimal 15 min ITI (or proposed period of safety), prior studies have suggested that protein-synthesis-dependent LTM formation requires the timing of consecutive spaced training trials to coincide with the peak of training-induced MAPK activity in KCs (Jacob, 2020).

Reinforcing PAM DANs have also been implicated in memory formation with sugar, water, and alcohol reward, with relative shock, with the absence of expected shock, and after courtship. In addition, they provide control of state-dependent memory expression and unlearned behavioral responses to volatile cues. In some cases, these processes clearly involve different DANs, whereas in others they appear to involve DANs that innervate the same MB compartments. More refined tools, connectomics, and experiments should help reveal the full extent of functional heterogeneity (Jacob, 2020).

Fly behavior has previously been shown to depend on the addition of supporting or conflicting experience. When differentially conditioned by the pairing of one odor with shock and the other with sugar, flies show additive initial performance compared to that observed if only one of the two odors is reinforced. This situation resembles that described in this study after spaced training except that the second odor is explicitly unpaired, and additive performance emerges from complementary LTM. With the benefit of retrospect, it makes intuitive sense that over repetitive spaced trials flies learn 'where the punishment is and where it is not.' These parallel memories make it easier for flies to distinguish between the two odors when tested together (Jacob, 2020).

In contrast, flies simultaneously form parallel competing memories when trained with bitter-tainted sugar, and their performance switches from aversion to approach over time, as dictated by the superior persistence of the nutrient-dependent sugar memory. A similar time-dependent behavioral transition is evident when flies are trained with alcohol reinforcement. A competition between memories of opposing valence also underlies the extinction of both appetitive and aversive memories. However, opposing extinction memories are sequentially formed and are reinforced by the absence of an expected outcome, rather than explicit pairing. In these cases, forming parallel memories reduces the certainty of odor choice (Jacob, 2020).

Together, these studies suggest that forming parallel memories in different places is a general MBON network feature that allows flies to summate experience over time to optimize the expression of learned behavior (Jacob, 2020).

Go to top

Abdelrahman, N. Y., Vasilaki, E. and Lin, A. C.
(2021). Compensatory variability in network parameters enhances memory performance in the Drosophila mushroom body. Proc Natl Acad Sci U S A 118(49). PubMed ID: 34845010

Neural circuits use homeostatic compensation to achieve consistent behavior despite variability in underlying intrinsic and network parameters. However, it remains unclear how compensation regulates variability across a population of the same type of neurons within an individual and what computational benefits might result from such compensation. These questions were addressed in the Drosophila mushroom body, the fly's olfactory memory center. In a computational model, this study showed that under sparse coding conditions, memory performance is degraded when the mushroom body's principal neurons, Kenyon cells (KCs), vary realistically in key parameters governing their excitability. However, memory performance is rescued while maintaining realistic variability if parameters compensate for each other to equalize KC average activity. Such compensation can be achieved through both activity-dependent and activity-independent mechanisms. Finally, it was shown that correlations predicted by the model's compensatory mechanisms appear in the Drosophila hemibrain connectome. These findings reveal compensatory variability in the mushroom body and describe its computational benefits for associative memory (Abdelrahman, 2021).

Go to top

Manoim, J. E., Davidson, A. M., Weiss, S., Hige, T. and Parnas, M.
(2022). Lateral axonal modulation is required for stimulus-specific olfactory conditioning in Drosophila. Curr Biol 32(20): 4438-4450. PubMed ID: 36130601

Effective and stimulus-specific learning is essential for animals' survival. Two major mechanisms are known to aid stimulus specificity of associative learning. One is accurate stimulus-specific representations in neurons. The second is a limited effective temporal window for the reinforcing signals to induce neuromodulation after sensory stimuli. However, these mechanisms are often imperfect in preventing unspecific associations; different sensory stimuli can be represented by overlapping populations of neurons, and more importantly, the reinforcing signals alone can induce neuromodulation even without coincident sensory-evoked neuronal activity. This paper reports a crucial neuromodulatory mechanism that counteracts both limitations and is thereby essential for stimulus specificity of learning. In Drosophila, olfactory signals are sparsely represented by cholinergic Kenyon cells (KCs), which receive dopaminergic reinforcing input. KCs were found to have numerous axo-axonic connections mediated by the muscarinic type-B receptor (mAChR-B). By using functional imaging and optogenetic approaches, it was shown that these axo-axonic connections suppress both odor-evoked calcium responses and dopamine-evoked cAMP signals in neighboring KCs. Strikingly, behavior experiments demonstrate that mAChR-B knockdown in KCs impairs olfactory learning by inducing undesired changes to the valence of an odor that was not associated with the reinforcer. Thus, this local neuromodulation acts in concert with sparse sensory representations and global dopaminergic modulation to achieve effective and accurate memory formation (Manoim, 2022).

This study showed that KC-KC axonal interaction is mediated by mAChR-B. This mAChR-B-mediated neuromodulation has dual roles: it decreases both odor-evoked Ca2+ elevation and DA-induced cAMP elevation. Thus, this neuromodulation suppresses both signals that are required for KC-MBON synaptic plasticity. In behavior experiments, it was demonstrated that mAChR-B knock-down (KD) in KCs impairs stimulus specificity of learning. This study reveals a novel form of local neuromodulation, which improves sensory discrimination during learning (Manoim, 2022).

This study identified the first biological functions of axo-axonic synapses between KCs. Olfactory coding in the insect MB is a well-established model system to study the circuit mechanisms and benefits of sparse sensory representations. The abundance of KC-KC synapses at the axons discovered by the EM connectome surprised the field at first because excitatory cholinergic interactions may ruin the very benefit of the sparse coding in olfactory learning. However, Ca2+ imaging demonstrated that the net effect of those cholinergic transmissions is, in fact, inhibitory. The lateral inhibition mediated by mAChR-B should further enhance, rather than ruin, the benefit of sparse coding and thereby improve the stimulus specificity of learning. Although the population of KCs that show reliable responses to a given odor is sparse (~5%), many more KCs are activated in a given odor presentation. This is because there is a larger population of unreliable responders, making up to ~15% of total KCs active in a given trial (Manoim, 2022).

Since those unreliable responders tend to show weaker Ca2+ responses than the reliable ones, it is reasonable to speculate that mAChR-B-mediated mutual inhibition would preferentially suppress unreliable responders, letting reliable responders win the lateral competition. Since even a single, 1-s odor-DAN activation pairing can induce robust KC-MBON synaptic plasticity, presence of unreliable responders can significantly compromise the synapse specificity of plasticity. Restricting Ca2+ responses to reliable responders should therefore greatly enhance the stimulus specificity of learning (Manoim, 2022).

To support thos finding, selective inhibition of Go signaling in KCs by expressing pertussis toxin (PTX) impairs aversive learning, and this effect was mapped to αβ and γ KCs, which were found to express mAChR-B most abundantly. Furthermore, expression of PTX disinhibits odor-evoked vesicular release in γ KCs, and PTX-induced learning defect was ameliorated by hyperpolarization or blocking synaptic output of γ KCs (Manoim, 2022).

It is argued that mAChR-B-mediated inhibitory communication between γ KCs contributes at least in part to those previous observations. Lateral communication through mAChR-B also suppresses cAMP signals in KCs, which counteracts Dop1R1-mediated DA action during associative conditioning. Since DA release in the MB likely takes a form of volume transmission, it cannot provide target specificity of modulation. Furthermore, although induction of LTD depends on coincident activity of KCs and DANs, elevation of cAMP can be triggered by DA application alone, although DA input followed by KC activity could induce opposite plasticity (i.e., potentiation) via another type of DA receptor (Manoim, 2022).

Thus, lateral inhibition of cAMP signals by Gi/o-coupled mAChR-B plays an essential role in the maintenance of target specificity of modulation. Taken together, dual actions of mAChR-B on local Ca2+ and cAMP signals at KC axons, where plasticity is supposed to take place, should directly contribute to synapse specificity of plasticity. If animals lack mAChR-B in KCs, axons of unreliable responders to CS+ would stay mildly active during conditioning. Furthermore, DA release on KCs causes some unchecked increase in cAMP in inactive and mildly active KCs. Consequently, some plasticity occurs in these KCs, even if to a lesser extent than in the KCs that are reliably and strongly activated by the CS+. Thus, absence of mAChR-B would minimally affect plasticity of KCs that are reliably activated by the CS+, assuming that those KCs are nearly maximally depressed by learning-related plasticity in the presence of mAChR-B. However, other KCs, which may include reliable responders to the CS−, will also undergo plasticity. This should result in unspecific association and that is exactly the type of learning defect observed in mAChR-B KD flies (Manoim, 2022).

The above model suggests that mAChR-B is required during memory acquisition. However, previous studies suggested that blocking KC synaptic output during memory acquisition does not affect aversive memory. How can one reconcile these two seemingly contradictory results? The experimental approach (i.e., RNAi KD of mAChR-B) precluded the ability to control the receptor function with high temporal specificity, and therefore, it was not possible to directly test whether mAChR-B is required during memory acquisition. Nevertheless, it is plausible that KC output affects memory acquisition via mAChR-B. Previous literature relied on temperature-sensitive Shibirets1 (shits1), which blocks synaptic release at the restrictive temperature, to demonstrate that KC output is not required during memory acquisition. However, it has been shown that substantial release is still maintained with shits1 even at the restrictive temperature (Manoim, 2022).

GPCRs are known to be activated at extremely low concentrations, ranging in the nM. On the other hand, nicotinic receptors operate at higher concentrations, often in the range of μM. Thus, it is possible that in the presence of shits1, there is some residual release from KCs at the restrictive temperature that is sufficient to activate mAChR-B but not the nicotinic receptors on downstream neurons. Thus, these results shed light on the role of KC output during memory acquisition, which may have been overlooked in previous studies (Manoim, 2022).

What may be the cellular mechanisms underlying the effects of mAChR-B on cAMP and Ca2+ level? mAChR-B was shown to be coupled to Gi/o, which is known to inhibit the cAMP synthase, adenylate cyclase, which is widely expressed in KCs (Manoim, 2022).

In addition, the Gβγ subunits have been demonstrated to be able to directly block voltage-gated Ca2+ channels. Gβγ can also directly open inward rectifying potassium channels that would oppose the changes in membrane potential required for the gating of voltage-gated Ca2+ channels, although these potassium channels are not broadly expressed in KCs. In this regard, it would be interesting to note that behavioral and physiological effects of mAChR-B KD were observed only whend KD was performed in γ KCs, although the results indicate that those receptors are also expressed in αβ KCs. This could be due to potential diversity in the intracellular signaling molecules among KC subtypes. Another possibility is that the efficiency of RNAi KD is somehow different between those KCs. It is also possible that the relatively lower number of KC-KC connections between αβ KCs may be insufficient to activate mAChR-B in the experimental contexts. Nevertheless, it is noted that a number of studies have demonstrated that γ KCs have a dominant role at the stage of acquisition of short-term memory, which is consistent with the model that proposes the critical role of mAChR-B during memory acquisition (Manoim, 2022).

Although the majority of studies on population-level sensory coding has focused on somatic Ca2+ or extracellular electrophysiological recordings, this study sheds light on the importance of local regulation of Ca2+ and other intracellular signals at the axons when it comes to stimulus specificity of learning. Are there other mechanisms that may be involved in reducing unspecific conditioning? One potential source of such mechanisms is the APL neuron, a single GABAergic neuron in the MB that is excited by KCs and provides feedback inhibition to KCs (Manoim, 2022).

Since activity of APL neuron contributes to sparse and decorrelated olfactory representations in KCs, it is possible that GABAergic input to KC axons also serves to prevent unspecific learning. Release of GABA onto KC axons is expected to have similar effects as the activation of mAChR-B. Specifically, the activation of the Gi/o-coupled GABA-B receptors that are widely expressed in KCs should have similar effects as activation of mAChR-B. However, in the current experiments, lateral inhibition induced by optogenetic activation of a subset of KCs was completely suppressed by mAChR-B KD, suggesting that APL neuron did not contribute to lateral suppression of Ca2+ response at least in the current experimental condition. This result is consistent with the prediction that individual KCs inhibit themselves via APL neuron more strongly than they inhibit the others due to the localized nature of the activity of APL neuron's neurites and the geometric arrangement of the ultrastructurally identified synapses (Manoim, 2022).

Nonetheless, whether APL neuron contributes to sparsening of axonal activity to prevent unspecific conditioning remains to be examined. In summary, the current study identifies functional roles of axo-axonic cholinergic interactions by uncovering previously unknown local neuromodulation that can enhance the stimulus specificity of learning and refines the DA-centric view of MB plasticity (Manoim, 2022).

Go to top

Zheng, Z., Li, F., Fisher, C., Ali, I. J., Sharifi, N., Calle-Schuler, S., Hsu, J., Masoodpanah, N., Kmecova, L., Kazimiers, T., Perlman, E., Nichols, M., Li, P. H., Jain, V. and Bock, D. D.
(2022).Structured sampling of olfactory input by the fly mushroom body. Curr Biol 32(15): 3334-3349. PubMed ID: 35797998

Associative memory formation and recall in the fruit fly Drosophila melanogaster is subserved by the mushroom body (MB). Upon arrival in the MB, sensory information undergoes a profound transformation from broadly tuned and stereotyped odorant responses in the olfactory projection neuron (PN) layer to narrowly tuned and nonstereotyped responses in the Kenyon cells (KCs). Theory and experiment suggest that this transformation is implemented by random connectivity between KCs and PNs. However, this hypothesis has been challenging to test, given the difficulty of mapping synaptic connections between large numbers of brain-spanning neurons. This study used a recent whole-brain electron microscopy volume of the adult fruit fly to map PN-to-KC connectivity at synaptic resolution. The PN-KC connectome revealed unexpected structure, with preponderantly food-responsive PN types converging at above-chance levels on downstream KCs. Axons of the overconvergent PN types tended to arborize near one another in the MB main calyx, making local KC dendrites more likely to receive input from those types. Overconvergent PN types preferentially co-arborize and connect with dendrites of αβ and α'β' KC subtypes. Computational simulation of the observed network showed degraded discrimination performance compared with a random network, except when all signal flowed through the overconvergent, primarily food-responsive PN types. Additional theory and experiment will be needed to fully characterize the impact of the observed non-random network structure on associative memory formation and recall (Zheng, 2022).

Go to top

Poppinga, H., Çoban, B., Meltzer, H., Mayseless, O., Widmann, A., Schuldiner, O. and Fiala, A.
(2022). Pruning deficits of the developing Drosophila mushroom body result in mild impairment in associative odour learning and cause hyperactivity. Open Biol 12(9): 220096. PubMed ID: 36128716

The principles of how brain circuits establish themselves during development are largely conserved across animal species. Connections made during embryonic development that are appropriate for an early life stage are frequently remodelled later in ontogeny via pruning and subsequent regrowth to generate adult-specific connectivity. The mushroom body of the fruit fly Drosophila melanogaster is a well-established model circuit for examining the cellular mechanisms underlying neurite remodelling. This central brain circuit integrates sensory information with learned and innate valences to adaptively instruct behavioural decisions. Thereby, the mushroom body organizes adaptive behaviour, such as associative learning. However, little is known about the specific aspects of behaviour that require mushroom body remodelling. This study used genetic interventions to prevent the intrinsic neurons of the larval mushroom body (γ-type Kenyon cells) from remodelling. It was asked to what degree remodelling deficits resulted in impaired behaviour. Deficits were found to cause hyperactivity and mild impairment in differential aversive olfactory learning, but not appetitive learning. Maintenance of circadian rhythm and sleep were not affected. It is concluded that neurite pruning and regrowth of γ-type Kenyon cells is not required for the establishment of circuits that mediate associative odour learning per se, but it does improve distinct learning tasks (Poppinga, 2022).

Go to top

Olivares, G. H., Nunez-Villegas, F., Candia, N., Orostica, K., Gonzalez-Ramirez, M. C., Vega-Macaya, F., Zuniga, N., Molina, C., Oliva, C., Mackay, T. F. C., Verdugo, R. A. and Olguin, P.
(2023). Early-life nutrition interacts with developmental genes to shape the brain and sleep behavior in Drosophila melanogaster. Sleep. PubMed ID: 36718043

The mechanisms by which the genotype interacts with nutrition during development to contribute to the variation of complex behaviors and brain morphology of adults are not well understood. This study used the Drosophila Genetic Reference Panel to identify genes and pathways underlying these interactions in sleep behavior and mushroom body morphology. Early-life nutritional restriction effects on sleep behavior and brain morphology was shown to depend on the genotype. Genes associated with sleep sensitivity to early-life nutrition were enriched for protein-protein interactions responsible for translation, endocytosis regulation, ubiquitination, lipid metabolism, and neural development. By manipulating the expression of candidate genes in the mushroom bodies and all neurons, it was confirmed that genes regulating neural development, translation and insulin signaling contribute to the variable response of sleep and brain morphology to early-life nutrition. The interaction between differential expression of candidate genes with nutritional restriction in early life resides in the mushroom bodies or other neurons, and these effects are sex specific. Natural variation in genes that control the systemic response to nutrition and brain development and function interact with early-life nutrition in different types of neurons to contribute to the variation of brain morphology and adult sleep behavior (Olivares, 2023).

Go to top

Sears, J. C. and Broadie, K.
(2022). Temporally and Spatially Localized PKA Activity within Learning and Memory Circuitry Regulated by Network Feedback. eNeuro 9(2). PubMed ID: 35301221

Dynamic functional connectivity within brain circuits requires coordination of intercellular signaling and intracellular signal transduction. Critical roles for cAMP-dependent protein kinase A (PKA) signaling are well established in the Drosophila mushroom body (MB) learning and memory circuitry, but local PKA activity within this well-mapped neuronal network is uncharacterized. This study use an in vivo PKA activity sensor (PKA-SPARK) to test spatiotemporal regulatory requirements in the MB axon lobes. Immature animals had little detectable PKA activity, whereas postcritical period adults showed high field-selective activation primarily in just 3/16 defined output regions. In addition to the age-dependent PKA activity in distinct α'/β' lobe nodes, females show sex-dependent elevation compared with males in these same restricted regions. Loss of neural cell body Fragile X mental retardation protein (FMRP) and Rugose [human Neurobeachin (NBEA)] suppresses localized PKA activity, whereas overexpression (OE) of MB lobe PKA-synergist Meng-Po (human SBK1) promotes PKA activity. Elevated Meng-Po subverts the PKA age-dependence, with elevated activity in immature animals, and spatial-restriction, with striking γ lobe activity. Testing circuit signaling requirements with temperature-sensitive shibire (human Dynamin) blockade, broadly expanded PKA activity was found within the MB lobes. Using transgenic tetanus toxin to block MB synaptic output, greatly heightened PKA activity was found in virtually all MB lobe fields, although the age-dependence is maintained. It is concluded spatiotemporally restricted PKA activity signaling within this well-mapped learning/memory circuit is age-dependent and sex-dependent, driven by FMRP-Rugose pathway activation, temporally promoted by Meng-Po kinase function, and restricted by output neurotransmission providing network feedback (Sears, 2022).

This study explored the spatiotemporal regulation of PKA activity within the MB lobes. PKA signaling initiates in early adulthood, with heightened activity in just 3/16 MB lobe output neuron fields (α'1, β'1, and β'2ap). In addition to age-dependence, this regional PKA signaling displays sex-dependence, with elevation in females over males. These findings were made possible with the PKA-SPARK biosensor; this fluorescent reporter uses motifs found in earlier PKA sensors, with pharmacological and genetic approaches promoting and preventing PKA activity verifying this new tool in cell culture and in vivo. While it is possible that the PKA-SPARK reporter is revealing only the strongest PKA activity, at the least α'1, β'1, and β'2ap connectivity regions have much higher PKA activity levels compared with the rest of the MB neuropil. In two disease models of intellectual disability and ASDs, from loss of either FMRP or Rugose/NBEA, PKA activity remains spatiotemporally restricted, but is dramatically reduced. There is surprisingly little change from overactivation of the FMRP→Rg control pathway, but PKA activity is profoundly altered by the PKA pathway Meng-Po kinase, with OE enhancing spatiotemporal PKA signaling and loss suppressing PKA activity. Network feedback downstream of KC neurotransmission strongly suppresses PKA activity, since blocking KC synaptic output with conditional shibirets or transgenic TNT induces widespread PKA activity signaling. Thus, localized PKA activity is highly regulated at the circuit level (Sears, 2022).

At a macro level, the α'1 and β'1 regions with the highest localized PKA activity levels have been linked to valence, i.e., whether local activation causes animals to approach or avoid a stimulus. The α'1 and β'2 fields exhibit opposing output valence (positive and negative, respectively), while the β'1 role appears less clear. However, higher PKA signaling can be activated in the γ3 region (via Meng-Po), which in combination with the β'1 region drives positive valence. Moreover, dopaminergic and serotonergic biosensor signals function through the β' region from a variety of external sense stimuli, to which the α'1/β'1 fields display high sensitivity, suggesting broad responsiveness. Very recent work shows high spontaneous activity in α'/β' restricted to young animals , suggesting α'/β' PKA signaling may be controlled by selective developmental activity. Moreover, inhibiting miR-92a in α/β and γ, but not α'/β', was shown to enhance memory, indicating another layer of lobe-selective circuit regulation. Note that the output neurons from these MB lobe regions are different (α'1 cholinergic, β'1 GABAergic, and β'2 glutamatergic), suggesting more complex integrative circuit functions. PKA signaling activation after the early-use critical period is consistent with sensory experience-dependent regulation. Importantly, MB sensory integration functions differ markedly between females and males, correlating with the report in this study of sex-dependent PKA signaling differences in females compared with males (Sears, 2022).

Two learning and memory proteins, FMRP and Rugose, are needed for full PKA activity in the α'1 and β'1 MB lobe output neuron fields. RNA-binding FMRP is a translational regulator known to facilitate PKA signaling, which is lost in the FXS, the commonest heritable cause of intellectual disability and ASD. Rugose/NBEA is a PKA-anchor that facilitates learning and memory, and is also associated with ASDs. Previous work has shown FMRP binds to rugose mRNA to drive KC expression. As predicted, disruption of this FMRP→Rg regulative pathway strongly impairs PKA activity in the MB lobes. In contrast, localized PKA signaling is dramatically strengthened by MB OE of the Meng-Po kinase, which induces early-onset PKA activity before adult sensory experience, spatially expands high PKA activity to the γ3 MBON field, and profoundly elevates PKA activity within all the normal MB lobe regions of heightened PKA signaling. Consistently, Meng-Po kinase OE also greatly improves learning and memory, via PKA phosphorylation, but additionally via signaling feedback synergy. Moreover, meng-po RNAi causes the opposite result of reducing localized PKA-SPARK puncta. Based on both loss and gain of function, it is suggested that Meng-Po enhances localized PKA activity, reflecting circuit level kinase regulation. Determining how Meng-Po-regulated PKA activity determines circuit excitability and regional balance will be a major subject of future research (Sears, 2022).

Two different KC synaptic output blocking methods dramatically expand PKA activity signaling in the MB. Both conditional shibirets and transgenic tetanus toxin tools block KC neurotransmission, but through quite different mechanisms. At 33°C, KC-targeted shibirets drives PKA activity expansion in the MB γ lobe. This change could indicate cross-compartment network interactions between the γ lobe and other MB regions. Increasing γ1 PKA activity is especially interesting, as γ1 toggles inhibition of other MB regions. The tetanus toxin protease blocks neurotransmission through eliminating SV exocytosis, and therefore provides a stronger and more selective means to silence KC synaptic output. Consistently, TNT animals show a more profound expansion of PKA activity throughout the MB lobes, albeit again affecting only spatial and not temporal patterning. Neither shibirets shibirets nor TNT blockade alters early PKA activity, suggesting induction of PKA signaling is determined primarily by later experience-driven activity. Localized PKA activity changes with KC output block implies active circuit balance; for example, weighing aversive versus attractive responses to sensory stimuli. The widespread PKA activity upregulation with KC output block leads to a hypothesis that enhancing KC neuron activity should result in elevated PKA signaling (Sears, 2022).

At the MB circuit level, multiple candidate synaptic pathways need to be explored for roles in local PKA activity regulation in different MB lobe output neuron fields. GABAergic inputs to dopaminergic neurons are one likely candidate, since GABA treatment has been shown to correct dfmr1 mutant circuit defects exacerbated by glutamate exposure. Moreover, GABAergic anterior paired lateral (APL) neurons broadly control MB activity through feedback to the KCs, and are most strongly activated by the α'/β' lobes. Recent work shows that treatment with a dopamine transport inhibitor also ameliorates rugose mutant social interaction and memory deficits. Another candidate is the Amnesiac neuropeptide from the serotonergic dorsal paired medial neurons, required for their normal development in the broad innervation of the MB lobes. In the context of these studies, de facto depression may feedback onto KCs to promote PKA activity signaling. The impact of upstream input onto the MB lobes is an important consideration, including how this circuitry combines with spontaneous MB activity and internal lobe circuitry to determine PKA signaling. Future research directions should attempt to dissect how these different layers of neuromodulation control localized PKA activity signaling within the MB lobe circuit, and between females and males, by manipulating input-specific neuronal activity in targeted transgenic studies (Sears, 2022).

In conclusion, this study reports that PKA activity signaling in the Drosophila brain MB learning and memory center is highly induced during early experiential adulthood, with selective upregulation in the α'1, β'1, and β'2ap MB lobe output neuron regions. Age-specific and sex-specific PKA signaling controlled within KCs and downstream of KC output shows that spatiotemporally restricted MB lobe PKA activity is regulated through a combination of both intracellular control and intercellular network-level mechanisms. Importantly, PKA signaling can be precociously promoted and spatial expanded though the activity of the PKA-synergist Meng-Po kinase. Moreover, KC neurotransmission inhibits localized PKA signaling within the MB circuit. Future studies will be aimed toward generating new genetic responder tools to test KC signaling with both neurotransmission output blockade and activity promotion of upstream and downstream MB circuit components, simultaneously and independently of KCs. PKA activity signaling will be tested with the manipulation of specific KC partners, by altering neurotransmission signaling in combination with postsynaptic neurotransmitter receptor mutants to determine network communication cues. Taken together with the current work, these ongoing studies will continue to expand understanding of circuit-level PKA signaling regulation in normal function and in neurologic disease model contexts (Sears, 2022).

Go to top

Baltruschat, L., Prisco, L., Ranft, P., Lauritzen, J. S., Fiala, A., Bock, D. D. and Tavosanis, G.
(2021). Circuit reorganization in the Drosophila mushroom body calyx accompanies memory consolidation. Cell Rep 34(11): 108871. PubMed ID: 33730583

The formation and consolidation of memories are complex phenomena involving synaptic plasticity, microcircuit reorganization, and the formation of multiple representations within distinct circuits. To gain insight into the structural aspects of memory consolidation, this study focused on the calyx of the Drosophila mushroom body. In this essential center, essential for olfactory learning, second- and third-order neurons connect through large synaptic microglomeruli, which this study dissected at the electron microscopy level. Focusing on microglomeruli that respond to a specific odor, it was revealed that appetitive long-term memory results in increased numbers of precisely those functional microglomeruli responding to the conditioned odor. Hindering memory consolidation by non-coincident presentation of odor and reward, by blocking protein synthesis, or by including memory mutants suppress these structural changes, revealing their tight correlation with the process of memory consolidation. Thus, olfactory long-term memory is associated with input-specific structural modifications in a high-order center of the fly brain (Baltruschat, 2021).

The capacity to use past experience to guide future action is a fundamental and conserved function of the nervous system. Associative memory formation, initiated by the coincident detection of a conditioned stimulus (CS; e.g., odor) and an unconditioned stimulus (US; e.g., sugar reward), leads to a short-lived memory (STM) trace within distinct circuits. Memories can be consolidated into long-term memories (LTMs) through processes that depend on de novo protein synthesis, require structural modifications within the involved neuronal circuits, and might lead to the recruitment of additional ones. Compared with modulation of existing connections, the reorganization of circuits affords the unique possibility of sampling for potential new partners. Nonetheless, only few examples of rewiring associated with learning have been established thus far (Baltruschat, 2021).

The formation and retrieval of olfactory-associative memories in Drosophila require the mushroom body (MB). Within the main MB input compartment, the calyx (MBC), second-order projection neurons (PNs), delivers olfactory information through cholinergic synapses to the intrinsic MB neurons, the Kenyon cells. In the MBC, large, olfactory PN boutons are enwrapped by the claw-like dendrite termini of ∼11 KCs on average, thereby forming characteristic synaptic complexes, the microglomeruli (MGs), which display functional and structural plasticity in adaptation and upon silencing. To start systematically addressing the mechanisms that support memory consolidation, this study sought to investigate the properties of identifiable synaptic MGs in the MB of the adult brain of Drosophila after the establishment of LTMs (Baltruschat, 2021).

Combining behavioral experiments with high-resolution microscopy and functional imaging, this study demonstrates that the consolidation of appetitive olfactory memories closely correlates with an increase in the number of MGs formed by the PNs that deliver the conditioned stimulus and their postsynaptic KC partners. These structural changes result in additional, functional synaptic connections. Thus, the circuit in the calyx of the fly MB reorganizes accompanying the consolidation of associative memories (Baltruschat, 2021).

This study reports input-specific reorganization of the adult MBC circuit associated with the formation of long-term, appetitive memory. By visualizing presynaptic markers in PNs and the KC postsynaptic densities, this study uncovered an increase in the number of PN boutons and, at the same time, reveal that these boutons are enveloped by KC postsynaptic profiles, suggesting that new MGs are formed during memory consolidation. These findings are particularly remarkable, given the high degree of complexity of the MG microcircuits revealed by EM reconstruction and including the dendrite claws of multiple KCs of distinct subtypes. The cellular mechanisms leading to the increased number of odor-specific complex MGs remain to be clarified, but they will require a tight coordination between pre- and postsynaptic partners. In this context, mutations in synaptic proteins or in proteins mediating cell-cell interactions, which specifically block LTM, will be of great interest (Baltruschat, 2021).

It is suggested that remodeling could be driven by intrinsic reactivation of KCs during the consolidation phase or by modulatory inputs into the calyx. In either case, a complex pattern of activation is expected, that might be difficult to reproduce in artificial settings. Although the present observations are limited for technical reasons to the specific case of cVA, the overall density of PN boutons in the MBC increases after appetitive long-term conditioning in honeybees, as well as in leaf-cutting ants after avoidance learning. Based on that and given that the olfactory pathway of cVA is not distinguishable from that of other odors, it is thus suggested that the findings might be generalizable. In comparison with those systems, however, genetic and functional identification of PN subsets were used to reveal that the structural modifications are specific and limited to the PNs conveying the conditioned odor. Importantly, in vivo functional imaging data support the view that the circuit reorganization leads to additional functional MGs responding to the conditioned odor. In addition, they demonstrate a specific change in functional response in the KC dendrites toward the trained odor because the calcium levels drop faster toward baseline after appetitive associative conditioning. The faster decay kinetics and more skewed response toward the onset of the stimulus could contribute to a more-efficient temporal summation of responses or refine the KC response and might be related to inhibitory modifications. An important open question is the effect of the increased number of responding MGs on the pattern of KC activation. KCs respond sparsely to odor input and require the coincident activation of multiples of their claws to produce an action potential. The data might underlie the addition of connections between the active PNs and a set of already-responding KCs, leading to facilitated response to the conditioned odor without changing the set of responding KCs. A recent publication, however, suggests an exciting alternative view. After aversive LTM establishment, the number of KCs responding to the conditioned odor is increased. If it is hypothesized that appetitive conditioning leads to a similar outcome, the data could provide anatomical and functional support to these findings. The pattern of KC response could, thus, be modulated by experience in adulthood and might represent a rich signifier of sensory stimulus and context. Reconstruction of an MG from EM serial sections derived from FAFB dataset (Baltruschat, 2021).

Go to top

Prisco, L., Deimel, S. H., Yeliseyeva, H., Fiala, A. and Tavosanis, G.
(2021). The anterior paired lateral neuron normalizes odour-evoked activity in the Drosophila mushroom body calyx. Elife 10. PubMed ID: 34964714

To identify and memorize discrete but similar environmental inputs, the brain needs to distinguish between subtle differences of activity patterns in defined neuronal populations. The Kenyon cells (KCs) of the Drosophila adult mushroom body (MB) respond sparsely to complex olfactory input, a property that is thought to support stimuli discrimination in the MB. To understand how this property emerges, the role was explored of the inhibitory anterior paired lateral (APL) neuron in the input circuit of the MB, the calyx. Within the calyx, presynaptic boutons of projection neurons (PNs) form large synaptic microglomeruli (MGs) with dendrites of postsynaptic KCs. Combining electron microscopy (EM) data analysis and in vivo calcium imaging, it was shown that APL, via inhibitory and reciprocal synapses targeting both PN boutons and KC dendrites, normalizes odour-evoked representations in MGs of the calyx. APL response scales with the PN input strength and is regionalized around PN input distribution. These data indicate that the formation of a sparse code by the KCs requires APL-driven normalization of their MG postsynaptic responses. This work provides experimental insights on how inhibition shapes sensory information representation in a higher brain centre, thereby supporting stimuli discrimination and allowing for efficient associative memory formation (Prisco, 2021).

While the importance of inhibition in reducing the overlap among stimuli representation has been postulated many decades ago and supported by more recent experimental evidence, the complete mechanism by which inhibition supports stimuli discrimination is not fully understood yet. This study shows that the inhibitory APL neuron, by participating in the structure of MGs of the Drosophila MB calyx, provides inhibition scaled to the PNs excitatory inputs to the calyx. As a result, the average strength and the distribution of postsynaptic responses in KC dendritic claws become more similar across different odour representations. It is suggested that this normalization of postsynaptic responses operated by APL is at the core of pattern separation in the MB (Prisco, 2021).

Pattern separation is obtained in the MB through the formation of a sparse response in the KC layer. The decoding of a sparse code, in general, increases the storage capacity of associative networks, thereby supporting learning and classification tasks. In fact, sparse neuronal representations are described in several organisms including mammals, songbirds, and insects. APL was reported to play a key role in maintaining KCs responses sparse, but the underlying mechanism was far from understood. KCs receive inputs from six to eight PNs on average and, due to KCs high firing threshold, require more than half of those inputs to be coactive to spike. The current data suggest that the APL neuron, by confining KC claws responses within a certain range of activation, ensures that KCs requirement of multiple coactive claws is respected even in the presence of highly variable input strengths. In other words, APL inhibition makes KC input integration dependent on the combinatorial pattern of inputs rather than on the strength of individual inputs. In support of this, blocking APL leads to an increased correspondence between input strength and KC response. Of notice, odour discrimination is achieved at multiple levels of the Drosophila olfactory pathway by different types of inhibitory neurons. Indeed, input gain control normalization has been described for GABAergic interneurons in the AL as well as for inhibitory iPNs at the lateral horn. Additionally, APL and its homolog GGN in the locust showed increased depolarization in response to increasing odour concentration. This, combined with the current findings, suggests that the normalization performed by APL might be acting not only across stimuli identities, but also among concentrations of the same stimulus (Prisco, 2021).

Structural and functional data point towards the involvement of APL in a feedforward loop from PN boutons to KC claws, as well as a closed feedback loop with PN boutons. An advantage of using recurrent circuits to provide inhibition is that such a system can deal with a wide range of input strength, as inhibition and excitation strengths are proportional. Indeed, EM analysis revealed both pre- and postsynaptic connection between APL and PN boutons, linearly proportional to each other, and the differences in the APL calcium influx in response to odours correlated to the variability measured in PNs. So far, APL has been mainly described as a feedback neuron for KCs. However, feedforward inhibitory neurons from the input population onto the next layer have been described in other neuronal networks performing pattern separation. For example, granule cells receive both feedforward and feedback inhibition from Golgi cells at the cerebellar cortex, which are driven by excitatory inputs from the mossy fibers and granule cells' axons, respectively. Moreover, it has recently been demonstrated that Golgi cells recruit scales with the mossy fibers input density, similarly to what was observed in the functional imaging experiments carried out in this study. Additionally, adaptive regulation of KCs sparseness by feedforward inhibition has already been theorized in realistic computational models of insect's MBs. Regarding KCs-to-APL connections, a positive linearity was found among pre- and postsynapses between these two cells, confirming the presence of a local feedback loop within KC dendrites and the APL at the calyx. Furthermore, it has been reported that α/β KCs receive more inhibitory synapses along their dendritic trees compared to γ and α'/β', where the majority of synapses received from the APL is localized on KC claws instead. As the ability of inhibitory synapses to shunt current from excitatory synapses depends on the spatial arrangement of the two inputs, it is speculated that the difference in APL synapses localization could contribute to some of the electrophysiological differences recorded among distinct KCs type. For example, α/βc KCs were found to have a higher input resistances and longer membrane time constants compared to α'/β' KCs, resulting in a sigmoidal current-spike frequency function rather than a linear one. Additionally, a difference in synapses distribution can also indicate that two inhibitory mechanisms coexist at the MB calyx, similarly to what has been shown in the cerebellum where Golgi cells are responsible for both tonic inhibition, controlling granule cells spike number, gain control, and phasic inhibition, limiting the duration of granule cells responses (Prisco, 2021).

Finally, volumetric calcium imaging showed that the APL inhibition is local within the MB calyx. In particular, a difference was found in the APL calcium transients when flies were stimulated with odours that activate PN subsets with segregated bouton distribution in the calyx. These data suggest that APL inhibition onto MGs can be imagined as a gradient that peaks at the MGs active during a given stimulus and attenuates with distance. Non-spiking interneurons in insects are typically large and characterized by complex neurite branching, an ideal structure to support local microcircuits. As a matter of fact, similar examples of localized APL response as described in the current study have been reported in the Drosophila MB as well as in the APL's homolog GGN in the locust. An advantage of having local microcircuits is that it allows a single neuron to mimic the activity of several inhibitory interneurons, as described in amacrine cells of both mammals. Additionally, a parallel local-global inhibition is suggested to expand the dynamic range of inputs able to activate KCs (Prisco, 2021).

An important open question is whether the APL inhibition onto MGs of the MB calyx is more of a presynaptic phenomenon, therefore acting on PN boutons output, or a postsynaptic one on KCs claws. Functional data reveal a clear impact of APL on the postsynaptic response in MGs, while the PN boutons display a broad range of activity levels. Accordingly, silencing of the GABAA receptor Rdl in KCs increased calcium responses in the MB, including the calyx, and reduced sparseness of odour representations . However, due to the presence of presynapses from APL to PN boutons, a presynaptic component of APL inhibition is certainly possible (Prisco, 2021).

One possible caveat to the hypothesis is given by the fact that reducing GABA synthesis in APL by RNAi has been found to improve olfactory learning. However, this could be explained by a low efficiency of RNAi in this case. Indeed, incomplete silencing via RNAi increases KC output without affecting sparseness, whereas blocking APL output via shibirets leads to large, overlapping odour representations and impaired olfactory discrimination (Prisco, 2021).

Taken together, this study provides novel insights on how feedforward inhibition via APL shapes the postsynaptic response to olfactory inputs in the MB calyx and contributes to maintaining odour-evoked KC activity sparse. In the future, it will be interesting to investigate the impact of APL on memory consolidation, which has been associated with structural plasticity in the calyx and with changes in the KC response (Prisco, 2021).

Go to top

Hamid, R., Sant, H. S. and Kulkarni, M. N.
(2021). Choline Transporter regulates olfactory habituation via a neuronal triad of excitatory, inhibitory and mushroom body neurons. PLoS Genet 17(12): e1009938. PubMed ID: 34914708

Choline is an essential component of Acetylcholine (ACh) biosynthesis pathway which requires high-affinity Choline transporter (ChT) for its uptake into the presynaptic terminals of cholinergic neurons. A previous study had reported a predominant expression of ChT in memory processing and storing region of the Drosophila brain called mushroom bodies (MBs). It is unknown how ChT contributes to the functional principles of MB operation. This study demonstrated the role of ChT in Habituation, a non-associative form of learning. Odour driven habituation traces are laid down in ChT dependent manner in antennal lobes (AL), projection neurons (PNs), and MBs. Reduced habituation due to knock-down of ChT in MBs causes hypersensitivity towards odour, suggesting that ChT also regulates incoming stimulus suppression. Importantly, it was show for the first time that ChT is not unique to cholinergic neurons but is also required in inhibitory GABAergic neurons to drive habituation behaviour. These results support a model in which ChT regulates both habituation and incoming stimuli through multiple circuit loci via an interplay between excitatory and inhibitory neurons. Strikingly, the lack of ChT in MBs shows characteristics similar to the major reported features of Autism spectrum disorders (ASD), including attenuated habituation, sensory hypersensitivity as well as defective GABAergic signalling. These data establish the role of ChT in habituation and suggest that its dysfunction may contribute to neuropsychiatric disorders like ASD (Hamid, 2021).

Go to top

Haussmann, I. U., Wu, Y., Nallasivan, M. P., Archer, N., Bodi, Z., Hebenstreit, D., Waddell, S., Fray, R. and Soller, M.
(2022). CMTr cap-adjacent 2'-O-ribose mRNA methyltransferases are required for reward learning and mRNA localization to synapses. Nat Commun 13(1): 1209. PubMed ID: 35260552

Cap-adjacent nucleotides of animal, protist and viral mRNAs can be O-methylated at the 2' position of the ribose (cOMe). The functions of cOMe in animals, however, remain largely unknown. This study shows that the two cap methyltransferases (CMTr1 and CMTr2) of Drosophila can methylate the ribose of the first nucleotide in mRNA. Double-mutant flies lack cOMe but are viable. Consistent with prominent neuronal expression, they have a reward learning defect that can be rescued by conditional expression in mushroom body neurons before training. Among CMTr targets are cell adhesion and signaling molecules. Many are relevant for learning, and are also targets of Fragile X Mental Retardation Protein (FMRP). Like FMRP, cOMe is required for localization of untranslated mRNAs to synapses and enhances binding of the cap binding complex in the nucleus. Hence, thie study reveals a mechanism to co-transcriptionally prime mRNAs by cOMe for localized protein synthesis at synapses (Haussmann, 2022).

Go to top

Abdelrahman, N. Y., Vasilaki, E. and Lin, A. C.
(2021). Compensatory variability in network parameters enhances memory performance in the Drosophila mushroom body. Proc Natl Acad Sci U S A 118(49). PubMed ID: 34845010

Neural circuits use homeostatic compensation to achieve consistent behavior despite variability in underlying intrinsic and network parameters. However, it remains unclear how compensation regulates variability across a population of the same type of neurons within an individual and what computational benefits might result from such compensation. These questions were addressed in the Drosophila mushroom body, the fly's olfactory memory center. In a computational model, this study showed that under sparse coding conditions, memory performance is degraded when the mushroom body's principal neurons, Kenyon cells (KCs), vary realistically in key parameters governing their excitability. However, memory performance is rescued while maintaining realistic variability if parameters compensate for each other to equalize KC average activity. Such compensation can be achieved through both activity-dependent and activity-independent mechanisms. Finally, it was shown that correlations predicted by the model's compensatory mechanisms appear in the Drosophila hemibrain connectome. These findings reveal compensatory variability in the mushroom body and describe its computational benefits for associative memory (Abdelrahman, 2021).

This study examined under what conditions interneuronal variability would improve vs. impair associative memory. Using a computational model of the fly mushroom body, it was shown that under sparse coding conditions, associative memory performance is reduced by experimentally realistic variability among KCs in parameters that control neuronal excitability (spiking threshold and the number/strength of excitatory inputs). These deficits arise from unequal average activity levels among KCs. However, memory performance can be rescued by using variability along one parameter to compensate for variability along other parameters, thereby equalizing average activity among KCs. These compensatory models predicted that certain KC features would be correlated with each other, and these predictions were borne out in the hemibrain connectome. In short, this study showed 1) the computational benefits of compensatory variation, 2) multiple mechanisms by which such compensation can occur, and 3) anatomical evidence that such compensation does, in fact, occur. Note that when 'equalizing KC activity,' is said, it does. not mean that all KCs should respond the same to a given odor. Rather, in each responding uniquely to different odors (due to their unique combinations of inputs from different PNs), they should keep their average activity levels the same. That is, while KCs' odor responses should be heterogeneous, their average activity should be homogeneous (Abdelrahman, 2021).

How robust are the connectome analyses of this study? Correlations were found between anatomical proxies for the physiological properties predicted to be correlated in the models (i.e., KCs receiving excitation from more PNs should have weaker excitatory inputs, while KCs receiving more overall excitation should also receive more inhibition). In particular, the number of synapses per connection was measured as a proxy for the strength of a connection. This proxy seems valid based on matching anatomical and electrophysiological data. However, other factors affecting synaptic strength (receptor expression, posttranslational modification of receptors, presynaptic vesicle release, input resistance, etc.) would not be visible in the connectome. Of course, such factors could further enable compensatory variability. It is also worth noting that the connectome data are from only one individual (Abdelrahman, 2021).

The distance between PN-KC synapses and the peduncle is used as a proxy for the passive decay of synaptic currents as they travel to the spike initiation zone. In the absence of detailed compartmental models of KCs, it is hard to predict exactly how much increased distance would reduce the effective strength of synaptic inputs, but it is plausible to assume that signals decay monotonically with distance. Note that calcium signals are often entirely restricted to one dendritic claw. Another caveat is that the posterior boundary of the peduncle is only an estimate (although a plausible one) of the location of the spike initiation zone. However, inaccurate locations should only produce fictitious correlations if the error is correlated with the number of PN-KC synapses per KC (and only in αβ−c and γ-main KCs, not other KCs), which seems unlikely (Abdelrahman, 2021).

This work is consistent with prior work, both theoretical and experimental, showing that compensatory variability can maintain consistent network behavior. However, this study analyzed the computational benefits of equalizing activity levels across neurons in a population (as opposed to across individual animals or over time). A recent preprint showed that equalizing response probabilities among KCs reduces memory generalization, but the current showed that equalizing average activity outperforms equalizing response probabilities. Another model of the mushroom body used compensatory inhibition, in which the strength of inhibition onto each KC was proportional to its average excitation, similar to the inhibitory plasticity model. However, the previous work did not analyze the specific benefits from the compensatory variation; it also set the inhibition strong enough that average net excitation was zero, whereas this study shows that when inhibition is constrained to be only strong enough to reduce KC activity by approximately half (consistent with experimental data), inhibition alone cannot realistically equalize KC activity. In addition, there is experimental support for the current models' predictions that KCs with more PN inputs would have weaker excitatory inputs; when predicting whether calcium influxes in individual claws would add up to cause a suprathreshold response in the whole KC, the most accurate prediction came from dividing the sum of claw responses by the log of the number of claws. However, the functional benefits of this result only become clear with computational models. Finally, the larval mushroom body shows a similar relationship between number and strength of PN-KC connections; the more PN inputs a KC has, the fewer synapses per PN-KC connection; however, again, the larval work did not analyze the computational benefits of this correlation (Abdelrahman, 2021).

This study modeled two forms of compensation: direct correlations between neuronal parameters and activity-dependent homeostasis. Both forms improve performance and predict observed correlations in the connectome. Certainly, activity-dependent mechanisms are plausible as KCs regulate their own activity homeostatically in response to perturbations in activity. Indeed, different KC subtypes use different combinations of mechanisms for homeostatic plasticity, consistent with the different correlations observed in the connectome for different KC subtypes. The activity-dependent models lend themselves to straightforward biological interpretations. Excitatory or inhibitory synaptic weights could be tuned by activity-dependent regulation of the number of synapses per connection or expression/localization of receptors or other postsynaptic machinery. Spiking thresholds could be tuned by altering voltage-gated ion conductances or moving/resizing the spike initiation zone. Such homeostatic plasticity would be akin to the sensory gain control implemented by feedback inhibition but on a slower timescale (Abdelrahman, 2021).

On the other hand, KCs are not infinitely flexible in homeostatic regulation; for example, complete blockade of inhibition causes the same increase in KC activity regardless of whether the blockade is acute (16 to 24 h) or constitutive (throughout life). This apparent lack of activity-dependent down-regulation of excitation suggests that activity-independent mechanisms might contribute to compensatory variation in KCs, as occurs for ion conductances in lobster stomatogastric ganglion neurons. For example, the inverse correlation of w and N arises from the fact that the number of PN-KC synapses per KC increases only sublinearly with increasing numbers of claws (i.e., PN inputs). Perhaps a metabolic or gene regulatory constraint prevents claws from recruiting postsynaptic machinery in linear proportion to their number. (Interestingly, this suppression is stronger in larvae, where the number of PN-KC synapses per KC is actually constant relative to the number of claws). Meanwhile, the correlation between the number of inhibitory synapses and the number of excitatory synapses might be explained if excitatory and inhibitory synapses share bottleneck synaptogenesis regulators on the postsynaptic side. Although activity-dependent compensation produced superior performance in the current model compared with activity-independent compensation thanks to its more effective equalization of KC average activity (most likely because it takes into account the unequal activity of different PNs), activity-dependent mechanisms suffered when the model network switched to a novel odor environment. Given that it is desirable for even a newly eclosed fly to learn well and for flies to learn to discriminate arbitrary novel odors, activity-independent mechanisms for compensatory variation may be more effective in nature (Abdelrahman, 2021).

Compensatory variability to equalize activity across neurons could also occur in other systems. The vertebrate cerebellum has an analogous architecture to the insect mushroom body; cerebellar granule cells are strikingly similar to KCs in their circuit anatomy, proposed role in 'expansion recoding' for improved memory, and even signaling pathways for synaptic plasticity . Whereas cortical neurons' average spontaneous firing rates vary over several orders of magnitude, granule cells are, like KCs, mostly silent at rest, and it is plausible that their average activity levels might be similar (while maintaining distinct responses to different stimuli). Granule cell input synapses undergo homeostatic plasticity, while compartmental models suggest that differences in granule cells' dendritic morphology would affect their activity levels, an effect attenuated by inhibition, raising the possibility that granule cells may also modulate interneuronal variability through activity-dependent mechanisms. Future experiments may test whether compensatory variability occurs in, and improves the function of, the cerebellum or other brain circuits. Finally, activity-dependent compensation may provide useful techniques for machine learning. For example, this study found that performance of a reservoir computing network could be improved if thresholds of individual neurons are initialized to achieve a particular activity probability given the distribution of input activities (Abdelrahman, 2021).

Go to top

Hidalgo, S., Fuenzalida-Uribe, N., Molina-Mateo, D., Escobar, A. P., Oliva, C., Espana, R. A., Andres, M. E. and Campusano, J. M.
(2020). Study of the release of endogenous amines in Drosophila brain in vivo in response to stimuli linked to aversive olfactory conditioning. J Neurochem. PubMed ID: 32596813

A highly challenging question in neuroscience is to understand how aminergic neural circuits contribute to the planning and execution of behaviors, including the generation of olfactory memories. In this regard, electrophysiological techniques like Local Field Potential or imaging methods have been used to answer questions relevant to cell and circuit physiology in different animal models, such as the fly Drosophila melanogaster. However, these techniques do not provide information on the neurochemical identity of the circuits of interest. Different approaches including fast scan cyclic voltammetry, allow researchers to identify and quantify in a timely fashion the release of endogenous neuroactive molecules, but have been only used in in vitro Drosophila brain preparations. This study report a procedure to record for the first time the release of endogenous amines -dopamine, serotonin and octopamine- in adult fly brain in vivo, by fast scan cyclic voltammetry. As a proof of principle, recordings were carried out in the calyx region of the Mushroom Bodies, the brain area mainly associated to the generation of olfactory memories in flies. By using principal component regression in normalized training sets for in vivo recordings, an increase in was detected octopamine and serotonin levels in response to electric shock and olfactory cues respectively. This new approach allows the study of dynamic changes in amine neurotransmission that underlie complex behaviors in Drosophila and sheds new light on the contribution of these amines to olfactory processing in this animal model (Hidalgo, 2020).

Go to top

Ganguly, A., Qi, C., Bajaj, J. and Lee, D. (2020)
(2020). Serotonin receptor 5-HT7 in Drosophila mushroom body neurons mediates larval appetitive olfactory learning. Sci Rep 10(1): 21267. PubMed ID: 33277559

Serotonin (5-HT) and dopamine are critical neuromodulators known to regulate a range of behaviors in invertebrates and mammals, such as learning and memory. Effects of both serotonin and dopamine are mediated largely through their downstream G-protein coupled receptors through cAMP-PKA signaling. While the role of dopamine in olfactory learning in Drosophila is well described, the function of serotonin and its downstream receptors on Drosophila olfactory learning remain largely unexplored. This study showed that the output of serotonergic neurons, possibly through points of synaptic contacts on the mushroom body (MB), is essential for training during olfactory associative learning in Drosophila larvae. Additionally, it was demonstrated that the regulation of olfactory associative learning by serotonin is mediated by its downstream receptor (d5-HT7) in a cAMP-dependent manner. d5-HT7 expression specifically in the MB, an anatomical structure essential for olfactory learning in Drosophila, is critical for olfactory associative learning. Importantly this work shows that spatio-temporal restriction of d5-HT7 expression to the MB is sufficient to rescue olfactory learning deficits in a d5-HT7 null larvae. In summary, these results establish a critical, and previously unknown, role of d5-HT7 in olfactory learning (Ganguly, 2020).

Go to top

>Zeng, J., Li, X., Zhang, R., Lv, M., Wang, Y., Tan, K., Xia, X., Wan, J., Jing, M., Zhang, X., Li, Y., Yang, Y., Wang, L., Chu, J., Li, Y. and Li, Y.
(2023). Local 5-HT signaling bi-directionally regulates the coincidence time window for associative learning. Neuron. PubMed ID: 36706757

Local 5-HT signaling bi-directionally regulates the coincidence time window for associative learning

The coincidence between conditioned stimulus (CS) and unconditioned stimulus (US) is essential for associative learning; however, the mechanism regulating the duration of this temporal window remains unclear. This study found that serotonin (5-HT) bi-directionally regulates the coincidence time window of olfactory learning in Drosophila and affects synaptic plasticity of Kenyon cells (KCs) in the mushroom body (MB). Utilizing GPCR-activation-based (GRAB) neurotransmitter sensors, this study found that Kenyon cell (KC)-released acetylcholine (ACh) activates a serotonergic dorsal paired medial (DPM) neuron, which in turn provides inhibitory feedback to KCs. Physiological stimuli induce spatially heterogeneous 5-HT signals, which proportionally gate the intrinsic coincidence time windows of different MB compartments. Artificially reducing or increasing the DPM neuron-released 5-HT shortens or prolongs the coincidence window, respectively. In a sequential trace conditioning paradigm, this serotonergic neuromodulation helps to bridge the CS-US temporal gap. Altogether, this study reports a model circuitry for perceiving the temporal coincidence and determining the causal relationship between environmental events (Zeng, 2023).

Go to top

Jiang, L. and Litwin-Kumar, A.
(2021). Models of heterogeneous dopamine signaling in an insect learning and memory center. PLoS Comput Biol 17(8): e1009205. PubMed ID: 34375329

The Drosophila mushroom body exhibits dopamine dependent synaptic plasticity that underlies the acquisition of associative memories. Recordings of dopamine neurons in this system have identified signals related to external reinforcement such as reward and punishment. However, other factors including locomotion, novelty, reward expectation, and internal state have also recently been shown to modulate dopamine neurons. This heterogeneity is at odds with typical modeling approaches in which these neurons are assumed to encode a global, scalar error signal. How is dopamine dependent plasticity coordinated in the presence of such heterogeneity? This study developed a modeling approach that infers a pattern of dopamine activity sufficient to solve defined behavioral tasks, given architectural constraints informed by knowledge of mushroom body circuitry. Model dopamine neurons exhibit diverse tuning to task parameters while nonetheless producing coherent learned behaviors. Notably, reward prediction error emerges as a mode of population activity distributed across these neurons. These results provide a mechanistic framework that accounts for the heterogeneity of dopamine activity during learning and behavior (Jiang, 2021).

Go to top

Zhao, B., Sun, J., Li, Q. and Zhong, Y.
(2021). Differential conditioning produces merged long-term memory in Drosophila. Elife 10. PubMed ID: 34279222

Multiple spaced trials of aversive differential conditioning can produce two independent long-term memories (LTMs) of opposite valence. One is an aversive memory for avoiding the conditioned stimulus (CS+), and the other is a safety memory for approaching the non-conditioned stimulus (CS-). This study shows that a single trial of aversive differential conditioning yields one merged LTM (mLTM) for avoiding both CS+ and CS-. Such mLTM can be detected after sequential exposures to the shock-paired CS+ and unpaired CS-, and be retrieved by either CS+ or CS-. The formation of mLTM relies on triggering aversive-reinforcing dopaminergic neurons and subsequent new protein synthesis. Expressing mLTM involves αβ Kenyon cells and corresponding approach-directing mushroom body output neurons (MBONs), in which similar-amplitude long-term depression of responses to CS+ and CS- seems to signal the mLTM. These results suggest that animals can develop distinct strategies for occasional and repeated threatening experiences (Zhao, 2021).

To survive in a complex environment, animals need to learn from threatening experiences to avoid potential dangers. From invertebrates to humans, aversive differential conditioning is widely used to study memories produced by threatening experiences. After repetitive spaced trials of conditioning, animals form two complementary long-term memories (LTMs) of opposite valence, including the aversive memory to the conditioned stimulus (CS+) and the rewarding memory to the non-conditioned stimulus (CS-). Such complementary LTMs result in enhanced long-lasting discrimination between CS+ and CS- through guiding avoidance to CS+ and approach to CS-. However, it remains unclear whether and how occasional threatening experiences, such as single-trial conditionings, would induce long-lasting changes in future escape behavior (Zhao, 2021).

From invertebrates to humans, experience-dependent long-lasting behavioral modifications mainly rely on the formation of LTMs. In Drosophila, there are at least two categories of aversive olfactory LTMs that last for more than 7 days. One is the spaced training-induced LTM that can be observed after repetitive spaced training with inter-trial rests (multiple trials with a 15 min rest interval between each), but not after either single-trial training or repetitive massed training without interval. Forming such aversive LTM requires new protein synthesis and the paired posterior lateral 1 (PPL1) cluster of dopaminergic neurons (DANs) to depress the connection between odor-activated Kenyon cells (KCs) in the mushroom body (MB) αβ lobe and downstream α2sc (MB-V2) MB output neurons (MBONs). The other is a recently reported context-dependent LTM that forms after single-trial training, which does not require protein synthesis-dependent consolidation. The expression of context-dependent LTM relies on multisensory integration in the lateral horn and is not affected by blocking KCs. However, all these observations derived from the same design principle that evaluates memory performance through testing the discrimination between CS+ and CS-. Thus, direct responses to CS+ and CS- have been largely overlooked (Zhao, 2021).

The current study introduced a third-odor test, in which flies were given a choice between either CS+ and a novel odor, or CS- and a novel odor. It was therefore identified that the single-trial differential conditioning produces a merged LTM (mLTM) guiding avoidances of both CS+ and CS- for several days after training. The encoding and expression of such mLTM involve new protein synthesis, PPL1 DANs, αβ KCs, and α2sc MBONs. These findings suggest that animals utilize distinct escape strategies for facing occasional and repeated dangers (Zhao, 2021).

In the current study, the use of third-odor test leads to a conclusion that single-trial training produces an mLTM for guiding flies to avoid both CS+ and CS- for more than 7 days. Three categories of evidence in support of this conclusion are outlined below (Zhao, 2021)

First, throughout this study, the amplitudes of long-term avoidances of CS+ and CS- are always at a similar level under various conditions, including pharmacological treatment, cold-shock treatment, odor re-exposure, paradigm alteration, and neural circuitry manipulations. Second, re-exposure to either one of CS+ and CS- alone can extinguish both CS+ avoidance and CS- avoidance. Third, the long-term avoidances of CS+ and CS- can be recorded as the depression of odor-evoked responses in the same α2sc MBONs, meanwhile, CS+ avoidance and CS- avoidance both involve the same PPL1 DANs, αβ KCs, and α2sc MBONs. Thus, CS+ avoidance and CS- avoidance derive from the same aversive mLTM, instead of two parallel LTMs of the same valence. The significance of these findings is further discussed below (Zhao, 2021).

Combining with a recent report that uses a similar third-odor test to dissect LTMs induced by multi-trial spaced training, it is concluded that spaced multi-trial aversive differential conditioning produces two independent LTMs of opposite valence for avoiding CS+ and approaching CS-, whereas single-trial aversive differential conditioning yields one mLTM that guides avoidances of both CS+ and CS-. Thus, animals can develop distinct escape strategies for different categories of dangers. When the same dangerous situation has been experienced repeatedly, animals would remember the detailed information to guide behavior in the next similar situation. However, when the dangerous event has only been experienced occasionally, animals would choose to avoid all potentially dangerous cues as a more reserved survival strategy (Zhao, 2021).

Moreover, the differences between single-trial training-induced mLTM and multi-trial training-induced complementary LTMs lead to the question of how these differences are induced by different training sessions. It has been reported that multi-trial spaced training induces depressed responses to CS+ in α2sc MBONs and α3 MBONs are required for aversive LTM to CS+, whereas the modulated responses to CS- in β'2mp MBONs and γ3, γ3β'1 MBONs appears to be responsible for the safety memory to CS-. In contrast, this study found that single-trial training is sufficient to induce the depressed responses to both CS+ and CS- in α2sc MBONs. Therefore, the results suggest a lower threshold and specificity of the plasticity between KCs-α2sc MBONs, compared to KCs-α3 MBONs, KCs-β'2mp MBONs, and KCs-γ3, γ3β'1 MBONs connections. Consequently, changing these synaptic connections requires involving more training sessions (Zhao, 2021).

Go to top

Turrel, O., Rabah, Y., Placais, P. Y., Goguel, V. and Preat, T.
(2020). Drosophila middle-term memory: Amnesiac is required for PKA activation in the mushroom bodies, a function modulated by Neprilysin 1. J Neurosci. PubMed ID: 32303647

In Drosophila, the mushroom bodies (MB) constitute the central brain structure for olfactory associative memory. As in mammals, the cAMP/PKA pathway plays a key role in memory formation. In the MB, Rutabaga adenylate cyclase acts as a coincidence detector during associative conditioning to integrate calcium influx resulting from acetylcholine stimulation and G protein activation resulting from dopaminergic stimulation. Amnesiac encodes a secreted neuropeptide required in the MB for two phases of aversive olfactory memory. Previous sequence analysis has revealed strong homology with the mammalian pituitary adenylate cyclase-activating peptide (PACAP). This study examined whether amnesiac is involved in cAMP/PKA dynamics in response to dopamine and acetylcholine co-stimulation in living flies. Experiments were carried out with both sexes, or with either sex. The data show that amnesiac is necessary for the PKA activation process that results from coincidence detection in the MB. Since PACAP peptide is cleaved by the human membrane neprilysin hNEP, an interaction was sought between Amnesiac and Neprilysin 1 (Nep1), a fly neprilysin involved in memory. When Nep1 expression is acutely knocked down in adult MB, memory deficits displayed by amn hypomorphic mutants are rescued. Consistently, Nep1 inhibition also restores normal PKA activation in amn mutant flies. Taken together the results suggest that Nep1 targets Amnesiac degradation in order to terminate its signaling function. This work thus highlights a key role for Amnesiac in establishing within the MB the PKA dynamics that sustain middle-term memory formation, a function modulated by Nep1 (Turrel, 2020).

Associative learning, which temporally pairs a conditioned stimulus (CS) to an unconditioned stimulus (US), is a powerful way of acquiring adaptive behavior. At the molecular and cellular levels, the association between CS and US is mediated by coincidence detection mechanisms that reflect the superadditive activation of a molecular pathway in the presence of both stimuli. One of the major coincidence detectors is the cAMP/PKA pathway, which depends on Type-I adenylate cyclases stimulated by both calcium/calmodulin, via acetylcholine signaling representing the CS, and G-protein coupled to dopamine metabotropic receptors activated by dopaminergic neurons encoding the US (Turrel, 2020).

In Drosophila, the mushroom bodies (MB) constitute the central integrative brain structure for olfactory memory. The MB are composed of 4000 intrinsic neurons called Kenyon cells (KC), and classed into three subtypes whose axons form two vertical (a and a9) and three medial (b, b9, and g) lobes. Using a classical conditioning paradigm in which an odorant (CS) was paired to electric shocks (US), a previous study revealed flies are capable of forming six discrete memory phases reflected at the neural network level. Among these phases are middle-term memory (MTM) and long-term memory (LTM), which are both encoded in a/b KC. As in mammals, the fly cAMP/PKA pathway plays a key role in associative memory, wherein the adenylate cyclase Rutabaga (Rut) acts as a coincidence detector in the MB to associate the CS and US pathways (Turrel, 2020).

The amnesiac Drosophila mutant (amn) was isolated in a memory defect behavioral screen. As with other components of the cAMP/PKA pathway involved in Drosophila memory, amn is expressed in the MB. It was recently showen that amn expression in the MB is specifically required for MTM and LTM. amn encodes a neuropeptide precursor with a signal sequence. Sequence analyses suggest the existence of three peptides, with one of them homologous to mammalian pituitary adenylate cyclase-activating peptide (PACAP). PACAP is widely expressed throughout the brain, acting as a neuromodulator or neurotrophic factor through activation of G-protein-linked receptors to regulate a variety of physiological processes through stimulation of cAMP production. Furthermore, PACAP may exert a role in learning and memory (Turrel, 2020).

After its release, a neurotransmitter's action is terminated either by diffusion, re-uptake by the presynaptic neuron, or enzymatic degradation. In contrast, neuropeptide signaling is exclusively terminated by enzymatic degradation. Possible enzyme candidates include neprilysins, type 1 metalloproteinases whose main function is the degradation of signaling peptides at the cell surface. Indeed, the human neprilysin hNEP is capable of cleaving a PACAP neuropeptide. Drosophila express four neprilysins that are all required for MTM and LTM, establishing that neuropeptide degradation is a central process for memory formation. Among the four neprilysins, Neprilysin 1 (Nep1) is the only one whose expression is required for MTM in the MBx (Turrel, 2020).

This study aimed to confirm whether AMN intervenes in memory by modulating cAMP concentration, as suggested by its sequence homology. For this, PKA dynamics were analyzed in the MB vertical lobes. The results show that amn mutant brains fail to display PKA activation in the a lobe in response to co-application of dopamine and acetylcholine. Whether Nep is involved in terminating AMN action was examined. Using RNAi, it was shown that Nep1 knock-down restores normal MTM and normal PKA dynamics in amn mutants, establishing a functional interaction between Nep1 and AMN in the MB (Turrel, 2020).

Previous work has shown that AMN expression is required in the MB for Drosophila memory. This study established that AMN expression in the MB is necessary for the synergistic activation of PKA observed on co-stimulation by dopamine and acetylcholine in the a lobe, a process that is thought to mimic the coincidence detection event underlying memory formation. Furthermore, the data demonstrate a functional interaction between AMN and Nep1, suggesting that Nep1 targets AMN degradation, thereby terminating AMN signaling (Turrel, 2020).

Six different aversive memory phases that are spatially segregated have been described in Drosophila. Their formation relies on distinct neuronal circuits, as well as distinct molecular and cellular mechanisms. rut mutants are impaired in specific memory phases, including short-term memory (STM), encoded in g KC, and MTM encoded in a/b KC. It was previously shown that Rut expression restricted to g KC is sufficient to restore STM, but not MTM, in rut mutant flies. It is thus likely that distinct Rut-mediated coincidence detection events occur in parallel in g and a/b KC, resulting in STM and MTM formation, respectively. Interestingly, mutants expressing a reduced amn level display normal STM. Thus, AMN is most likely not required for the coincidence detection process that leads to STM, a process that remains to be identified. Using in vivo imaging, previous work showed that coapplication of dopamine and acetylcholine induces a strong synergistic PKA response, which is Rut dependent and occurs specifically in MB vertical lobes. This study shows that this coincident PKA activation in the a lobe is abolished in amn mutants, while neither calcium signaling nor cAMP signaling following dopaminergic stimulation alone are altered. It is proposed that PKA activation mimics the coincidence detection event that occurs in a/b KC during MTM formation, and that AMN intervenes in this process by enabling a sustained Rut-mediated PKA activation in the MB a lobe (Turrel, 2020).

AMN might thus act at a step that ranges from the initial coincidence detection event that provokes Rut activation, to the final level of PKA activation. This is consistent with previous reports that AMN and DC0, the fly PKA catalytic subunit, act in a common pathway, and that AMN function is upstream of DC0 function. If AMN plays a role posterior to the coincidence detection event, it could be involved in an increase in cAMP concentration through the inhibition of phosphodiesterases that degrade cAMP. Indeed, dopamine receptors positively coupled to adenylate cyclases are equally distributed in all MB lobes as are DC0 and Rut, whereas 100 mM dopamine only induces a PKA response in the a lobe. This spatial control is achieved by the cAMP-specific phosphodiesterase Dunce (Dnc) which preferentially degrades cAMP in the b and g lobes, thus restricting high dopamine-induced PKA activation to the a lobe. AMN could thus be involved in the specific inhibition of Dnc in the a lobe (Turrel, 2020).

One attractive alternative hypothesis is that AMN action could take place at the level of Rut activation itself. Indeed, the fact that one of the AMN peptides is homologous to PACAP suggests that AMN might play a role in activating the adenylate cyclase Rut through G-protein-coupled receptors. This hypothesis fits with sequence prediction , and is supported by studies showing that AMN is functionally related to human PACAP. It was initially reported that Rut is activated by the application of human PACAP-38, and later shown that bath application of PACAP-38 rescues L-type current deficiency in amnX8 larval muscle fibers . Such rescue is abolished by application of an antagonist to Type-I PACAP-receptor as well as by application of an inhibitor of AC (Turrel, 2020).

Although STM and MTM both rely on the cAMP/PKA pathway, not only these processes occur in separate KC, but while STM is instantaneously acquired, MTM is acquired in a dynamic fashion following a two-step mechanism. It is proposed that AMN function is specifically required in the incremental build-up of MTM by boosting Rut activation following the initial event of coincidence detection, namely the first CS/US association of the training protocol. In this model, this first association results in an initial moderate level of Rut activation, followed by a moderate level of PKA activation (Turrel, 2020).

This moderate level of PKA activation does not mediate MTM formation and is below detection threshold. It is hypothesized that this initial increase in PKA activity, directly or indirectly, triggers the second step of the process, namely AMN secretion, and thus generate an activation loop whereby AMN activates Rut, hence creating a much higher level of Rut activation and subsequent high levels of PKA activation that is observable with the AKAR2 probe. MTM formation would rely on an AMN-dependent PKA-activation loop terminated on AMN degradation by Nep1 (Turrel, 2020).

One previous study has indicated that human PACAP is a substrate for hNEP (Gourlet, 1997), and this present work in Drosophila describes a functional interaction between AMN and Nep1. Importantly, whereas Nep1 knock-down rescues the amn mutant memory phenotype in a genetic context where the AMN level is reduced to ~50% versus wild-type flies (heterozygous for the amn null allele), it fails to do so in a genetic context where AMN is absent (i.e., in flies hemizygous for the amn null allele). Namely, the memory rescue observed on Nep1 inhibition is dependent on the presence of AMN, suggesting that this latter is targeted by Nep1. While a biochemical confirmation of this hypothesis would be welcome, it is technically difficult to achieve. Specifically, not only are AMN antibodies not available, but amn mRNA is expressed at very low levels, indicating that AMN peptide may be very scarce (Turrel, 2020).

The observation that the AMN peptide may be targeted by Nep1 is in agreement with a neuromodulatory function. Once released, a signaling molecule must be removed from its site of action to prevent continued stimulation, and to allow new signals to propagate. If neurotransmitter's action is terminated either by diffusion, re-uptake by the presynaptic neuron, or enzymatic degradation, signaling neuropeptides are specifically removed by degradation. The intensity and duration of neuropeptide-mediated signals are thus controlled via the cleavage of these neuropeptides by peptidases like neprilysins. Despite a few exceptions, neprilysins occur as integral membrane endopeptidases whose catalytic site faces the extracellular compartment. It is hypothesized that on conditioning, AMN is secreted by the KC to participate in Rut activation via G-protein-coupled receptors, and is ultimately removed from the extracellular compartment by Nep1 anchored at the KC membrane. Importantly, AMN expression in the MB restores normal PKA dynamics in amn null mutant flies, suggesting that the AMN peptide secreted by the MB on conditioning should act in an autocrine-like way to sustain Rut activity in the a/b neurons. Interestingly, the effects of neuropeptide transmitters are very diverse and often long-lived, which fits well with the specific involvement of AMN peptide in non-immediate memory phases via sustained PKA activation (Turrel, 2020).

Up to date, fly neprilysins have been involved in several behaviors: in the control of circadian rhythms, via hydrolysis of the pigment dispersing factor neurotransmitter, and in the control of food intake via cleavage of insulin-like regulatory peptides. In the latter study, it was shown that both Neprilysin 4 knock-down and overexpression in the larval CNS cause reduced food intake (Hallier, 2016). In a similar way, this study shows that both Nep1 knock-down and overexpression in a/b KC impairs MTM, consistent with the need for a proper control of AMN levels. It is suggested that Nep1 overexpression results in amn loss of function, whereas Nep1 knock-down causes the prolongation of AMN action, thus generating a prolonged activation of the cAMP/PKA pathway, a process deleterious for memory. This is in agreement with a previous study demonstrating that overexpressing DC0 in the MB impairs MTM (Turrel, 2020).

In conclusion, this study reports an acute role for AMN in memory formation via the PKA pathway in the a/b MB neurons, a function modulated by Nep1. These results thus support a role for AMN as an activating adenylate cyclase peptide, much like the role of PACAP, bringing clarity to the role PACAP may play in memory consolidation in mammals (Turrel, 2020).

Go to top

/dd>
Availability of food determines the need for sleep in memory consolidation

Chouhan, N. S., Griffith, L. C., Haynes, P. and Sehgal, A. (2021). Nature 589(7843): 582-585. PubMed ID: 33268891

Sleep remains a major mystery of biology, with little understood about its basic function. One of the most commonly proposed functions of sleep is the consolidation of memory. However, as conditions such as starvation require the organism to be awake and active, the ability to switch to a memory consolidation mechanism that is not contingent on sleep may confer an evolutionary advantage. This study identified an adaptive circuit-based mechanism that enables Drosophila to form sleep-dependent and sleep-independent memory. Flies fed after appetitive conditioning needed increased sleep for memory consolidation, but flies starved after training did not require sleep to form memories. Memory in fed flies is mediated by the anterior-posterior α'/β' neurons of the mushroom body, while memory under starvation is mediated by medial α'/β' neurons. Sleep-dependent and sleep-independent memory rely on distinct dopaminergic neurons and corresponding mushroom body output neurons. However, sleep and memory are coupled such that mushroom body neurons required for sleep-dependent memory also promote sleep. Flies lacking Neuropeptide F display sleep-dependent memory even when starved, suggesting that circuit selection is determined by hunger. This plasticity in memory circuits enables flies to retain essential information in changing environments (Chouhan, 2021)

Go to top

/dd>
Learning with reinforcement prediction errors in a model of the Drosophila mushroom body

Bennett, J. E. M., Philippides, A. and Nowotny, T. (2021). Nat Commun 12(1): 2569. PubMed ID: 33963189

Effective decision making in a changing environment demands that accurate predictions are learned about decision outcomes. In Drosophila, such learning is orchestrated in part by the mushroom body, where dopamine neurons signal reinforcing stimuli to modulate plasticity presynaptic to mushroom body output neurons. Building on previous mushroom body models, in which dopamine neurons signal absolute reinforcement, it is proposed instead that dopamine neurons signal reinforcement prediction errors by utilising feedback reinforcement predictions from output neurons. Plasticity rules were formulated that minimise prediction errors, verify that output neurons learn accurate reinforcement predictions in simulations, and postulate connectivity that explains more physiological observations than an experimentally constrained model. The constrained and augmented models reproduce a broad range of conditioning and blocking experiments, and this study demonstrated that the absence of blocking does not imply the absence of prediction error dependent learning. These results provide five predictions that can be tested using established experimental methods (Bennett, 2021).

Effective decision making benefits from an organism's ability to accurately predict the rewarding and punishing outcomes of each decision, so that it can meaningfully compare the available options and act to bring about the greatest reward. In many scenarios, an organism must learn to associate the valence of each outcome with the sensory cues predicting it. A broadly successful theory of reinforcement learning is the delta rule, whereby reinforcement predictions (RPs) are updated in proportion to reinforcement prediction errors (RPEs): the difference between predicted and received reinforcements. RPEs are more effective as a learning signal than absolute reinforcement signals because RPEs diminish as the prediction becomes more accurate, adding stability to the learning process. In mammals, RPEs related to rewards are signalled by dopamine neurons (DANs) in the ventral tegmental area and substantia nigra, enabling the brain to implement approximations to the delta rule. In Drosophila melanogaster, DANs that project to the mushroom body (MB) (see Valence-specific model of the mushroom body) provide both reward and punishment modulated signals that are required for associative learning5. However, to date, MB DAN activity is typically interpreted as signalling absolute reinforcements (either positive or negative) for two reasons: (1) a lack of direct evidence for RPE signals in DANs, and (2) limited evidence in insects for the blocking phenomenon, in which conditioning of one stimulus can be impaired if it is presented alongside a previously conditioned stimulus, an effect that is indicative of RPE-dependent learning. This study has incorporated anatomical and functional data from recent experiments into a computational model of the MB, in which MB DANs do compute RPEs. The model provides a circuit-level description for delta rule learning in the MB, which is use to demonstrate why the absence of blocking does not necessarily imply the absence of RPEs (Bennett, 2021).

The MB is organised into lateral and medial lobes of neuropil in which sensory encoding Kenyon cells (KCs) innervate the dendrites of MB output neurons (MBONs) that modulate behaviour. Consistent with its role in associative learning, DAN signals modulate MBON activity via synaptic plasticity at KC -> MBON synapses. Current models of MB function posit that the MB lobes encode either positive or negative valences of reinforcement signals and actions. Most DANs in the protocerebral anterior medial (PAM) cluster (called D+ in the model presented in this study) are activated by rewards, or positive reinforcement (R+), and their activation results in depression at synapses between coactive KCs (K) and MBONs that are thought to induce avoidance behaviours (M−). DANs in the protocerebral posterior lateral 1 (PPL1) cluster (D−) are activated by punishments, i.e., negative reinforcement (R−), and their activation results in depression at synapses between coactive KCs and MBONs that induce approach behaviours (M+). A fly can therefore learn to approach rewarding cues or avoid punishing cues as a result of synaptic depression at KC inputs to avoidance or approach MBONs, respectively (Bennett, 2021).

To date, there is only indirect evidence for RPE signals in MB DANs. DAN activity is modulated by feedforward reinforcement signals, but some DANs also receive excitatory feedback from MBONs, and it is likely this extends to all MBONs whose axons are proximal to DAN dendrites. The difference between approach and avoidance MBON firing rates is intrepeted as a RP that motivates behaviour, consistent with the observation that behavioural valence scales with the difference between approach and avoidance MBON firing rates. As such, DANs that integrate feedforward reinforcement signals and feedback RPs from MBONs are primed to signal RPEs for learning. These latter two features have yet to be incorporated in computational models of the MB (Bennett, 2021).

This study incorporate the experimental data described above to formulate a reduced computational model of the MB circuitry, demonstrate how DANs may compute RPEs, derive a plasticity rule for KC -> MBON synapses that minimises RPEs, and verify in simulations that the MB model learns accurate RPs. A limitation to the model was identified that imposes an upper bound on RP magnitudes, and demonstrate how putative connections between DANs, KCs and MBONs help circumvent this limitation. Introducing these additional connections yields testable predictions for future experiments as well as explaining a broader range of existing experimental observations that connect DAN and MBON stimulus responses to learning. Lastly, this study shows that both incarnations of the model—-with and without additional connections—-capture a wide range of observations from classical conditioning and blocking experiments in Drosophila. Different behavioural outcomes in the two models for specific experiments provide further strong experimental predictions (Bennett, 2021).

Successful decision making relies on the ability to accurately predict, and thus reliably compare, the outcomes of choices that are available to an agent. The delta rule, as developed by Rescorla and Wagner (1972) [ A theory of Pavlovian conditioning: variantions in the effectiveness of reinforcement and nonreinforcement. in Classical conditioning II: current research and theory, 64-99 (eds Black, A. H. & Prokasy, W. F.) (Appleton-Century-Crofts, 1972)], updates beliefs in proportion to a prediction error, providing a method to learn accurate and stable predictions. This work investigated the hypothesis that, in Drosophila melanogaster, the MB implements the delta rule. It is posited that approach and avoidance MBONs together encode RPs, and that feedback from MBONs to DANs, if subtracted from feedforward reinforcement signals, endows DANs with the ability to compute RPEs, which are used to modulate synaptic plasticity. A plasticity rule was formulated that minimises RPEs, and the effectiveness of the rule was verified in simulations of MAFC tasks. This study demonstrated how the established valence-specific circuitry of the MB restricted the learned RPs to within a given range, and postulated cross-compartmental connections, from MBONs to DANs, that could overcome this restriction. Such cross-compartmental connections are found in Drosophila larvae, but their functional relevance is unknown. Two MB models are presented that yield RPEs in DAN activity and that learn accurate RPs: (1) the VSλ model, in which plasticity incorporates a constant source of synaptic potentiation; (2) the MV model, in which mixed-valence connectivity between DANs, MBONs and KC -> MBON synapses is proposed. Both the VSλ and the MV models receive equally good support from behavioural experiments in which different genetic interventions impaired learning, while the MV model provides a mechanistic account for a greater variety of physiological changes that occur in individual neurons after learning. It is plausible, and can be beneficial, for both the VS&lambda and MV models to operate in parallel in the MB, as separately learning positive and negative aspects of decision outcomes, if they arise from independent sources, is important for context-dependent modulation of behaviour. Such learning has been proposed for the mammalian basal ganglia. This study also demonstrated why the absence of strong blocking effects in insect experiments does not necessarily imply that insects do not utilise RPEs for learning (Bennett, 2021).

The models yield predictions that can be tested using established experimental protocols. Below, which model supports each prediction is specified (Bennett, 2021).

Responses in single DANs too the unconditioned stimulus (US), when paired with a CS+, should decay towards a baseline over successive CS ± US pairings, as a result of the learned changes in MBON firing rates. Only one previous study has measured DAN responses throughout several CS-US pairings in Drosophila. Consistent with DAN responses in the current model, previous work has shown decaying responses in DANs in the γ- and β'-lobes during paired CS+ and US stimulation. However, they reported similar decaying responses when the CS+ and US were unpaired (separated by 90 s) that were not significantly different from the paired condition. The paper concluded that DANs do not exhibit RPEs, and that the decaying DAN responses were a result of non-associative plasticity. An alternative interpretation is that a 90 s gap between CS+ and US does not induce DAN responses that are significantly different from the paired condition, and that additional processes prevent the behavioural expression of learning. Ultimately, the evidence for either effect is insufficient. Furthermore, Dylla observed increased CS+ responses in DANs after training. Conversely, after training in these models, when the US was set to zero, DAN responses to the CS+ decreased. Interpreting post-training activity in DANs as responses to the CS+ alone, or alternatively as responses to an omitted US, are equally valid in the current model because the CS+ and US always occurred together. Resolving time within trials in the models would allow better addressing of this conflict with experiments. The Dylla results are, however, consistent with the temporal difference (TD) learning rule (as are studies on second order conditioning in Drosophila), of which the Rescorla-Wagner rule used in this work is a simplified case (Bennett, 2021).

DANs of both valence modulate plasticity at MBONs of a single valence. In contrast, anatomical and functional experimental data suggest that, in each MB compartment, the DANs and MBONs have opposite valences. However, the GAL4 lines used to label DANs in the PAM cluster often include as many as 20-30 cells each, and it has not yet been determined whether all labelled DANs exhibit the same valence preference. Similarly, the valence encoded by MBONs is not always obvious. It is not clear whether optogenetically activated MBONs biased flies to approach the light stimulus, or to exhibit no-go behaviour that kept them within the light. In larval Drosophila, there are several examples of cross-compartmental DANs and MBONs, but a full account of the valence encoded by these neurons is yet to be provided. In adult Drosophila, γ1-pedc MBONs deliver cross-compartmental inhibition, such that M4/6 MBONs are effectively modulated by both aversive PPL1-γ1-pedc DANs and appetitive PAM DANs (Bennett, 2021).

This is not the first publication to present a MB model that makes effective decisions after learning about multiple reinforced cues. However, these models utilise absolute reinforcement signals, as well as bounded synapses that cannot strengthen indefinitely with continued reinforcements. Thus, given enough training, these models would not differentiate between two cues that were associated with reinforcements of the same sign, but different magnitudes. Carefully designed mechanisms are therefore required to promote stability as well as differentiability of same sign, different magnitude reinforcements. The model builds upon these studies by incorporating feedback from MBONs to DANs. This allows KC -> MBON synapses to accurately encode the reinforcement magnitude and sign with stable fixed points that are reached when the RPE signalled by DANs decays to zero. Alternative mechanisms that may promote stability and differentiability are forgetting (e.g., by synaptic weight decay), or adaptation in DAN responses. Exploring these possibilities in a MB model for comparison with the RPE hypothesis is well worth while, but goes beyond the scope of this work (Bennett, 2021).

Central to this work is the assumption that the MB has only a single objective: to minimise the RPE. In reality, an organism must satisfy multiple objectives that may be mutually opposed. In Drosophila, anatomically segregated DANs in the γ-lobe encode water rewards, sugar rewards, and motor activity, suggesting that Drosophila do indeed learn to satisfy multiple objectives. Multi-objective optimisation is a challenging problem, and goes beyond the scope of this work. Nevertheless, for many objectives, the principle that accurate predictions aid decision making, which forms the basis of this work, still applies (Bennett, 2021).

For simplicity, the simulations compress all events within a trial to a single point in time, and are therefore unable to address some time-dependent features of learning. For example, activating DANs either before or after cue exposure can induce memories with opposite valences; in locusts, the relative timing of KC and MBON spikes is important, though not necessarily in Drosophila. Nor has this study addressed the credit assignment problem: how to associate a cue with reinforcement when they do not occur simultaneously. A candidate solution is TD learning, whereby reinforcement information is back-propagated in time to all cues that predict it. While DAN responses in the MB hint at TD learning, it is not yet clear how the MB circuity could implement it. An alternative solution is an eligibility trace, which enables synaptic weights to be updated upon reinforcement even after presynaptic activity has ceased (Bennett, 2021).

Lastly, this work addresses memory acquisition, but not memory consolidation, which is supported by distinct circuits within the MB. Incorporating memory stabilising mechanisms may help to better align simulations of genetic interventions with fly behaviour in conditioning experiments (Bennett, 2021).

By incorporating the fact that KC responses to compound stimuli are non-linear combinations of their responses to the components, the model described in this paper was used to demonstrate why the lack of evidence for blocking in insects cannot be taken as evidence against RPE-dependent learning in insects. The model provides a neural circuit instantiation of similar arguments in the literature, whereby variable degrees of blocking can be explained if the brain utilises representations of stimulus configurations, or latent causes, which allow learned associations to be generalised between a compound stimulus and its individual elements by varying amounts. The effects of such configural representations on blocking are more likely when the component stimuli are similar, for example, if they engage the same sensory modality, as was the case in previous studies. By using component stimuli that do engage different sensory modalities, experiments with locusts have indeed uncovered strong blocking effects (Bennett, 2021).

This study has developed a model of the MB that goes beyond previous models by incorporating feedback from MBONs to DANs, and has shown how such a MB circuit can learn accurate RPs through DAN mediated RPE signals. The model provides a basis for understanding a broad range of behavioural experiments, and reveals limitations to learning given the anatomical data currently available from the MB. Those limitations may be overcome with additional connectivity between DANs, MBONs and KCs, which provide five strong predictions from this work (Bennett, 2021).

Go to top

Social facilitation of long-lasting memory is mediated by CO(2) in Drosophila

Muria, A., Musso, P. Y., Durrieu, M., Portugal, F. R., Ronsin, B., Gordon, M. D., Jeanson, R. and Isabel, G. (2021). Curr Biol. PubMed ID: 33740428

How social interactions influence cognition is a fundamental question, yet rarely addressed at the neurobiological level. It is well established that the presence of conspecifics affects learning and memory performance, but the neural basis of this process has only recently begun to be investigated. In the fruit fly Drosophila melanogaster, the presence of other flies improves retrieval of a long-lasting olfactory memory. This study demonstrates that this is a composite memory composed of two distinct elements. One is an individual memory that depends on outputs from the α'β' Kenyon cells (KCs) of the mushroom bodies (MBs), the memory center in the insect brain. The other is a group memory requiring output from the αβ KCs, a distinct sub-part of the MBs. Social facilitation of memory increases with group size and is triggered by CO(2) released by group members. Among the different known neurons carrying CO(2) information in the brain, this study established that the bilateral ventral projection neuron (biVPN), which projects onto the MBs, is necessary for social facilitation. Moreover, it was demonstrated that CO(2)-evoked memory engages a serotoninergic pathway involving the dorsal-paired medial (DPM) neurons, revealing a new role for this pair of serotonergic neurons. Overall, this study identified both the sensorial cue and the neural circuit (biVPN>αβ>DPM>αβ) governing social facilitation of memory in flies. This study provides demonstration that being in a group recruits the expression of a cryptic memory and that variations in CO(2) concentration can affect cognitive processes in insects (Maria, 2021).

The ability of an individual to form distinct memories and refer to past experiences contributes to the survival of many species. Sensory stimuli from the environment are processed and integrated during memory formation and retrieval, sometimes impacting animal physiology over the very long term. In so-called social species, conspecifics are part of each individual's environment and constitute an important source of information that can lead to social learning. Although social learning has been widely examined in the literature, the influence of social context on memory retrieval has been poorly addressed, as most memory protocols are carried out on isolated individuals. This is not the case for the fruit fly Drosophila melanogaster, for which memory studies are generally carried out on groups and thus measure memory expression in a social context (Maria, 2021).

Despite a small brain of about 100,000 neurons, Drosophila can learn to associate and memorize different stimuli. A protocol leading to a measurable aversive olfactory memory is widely used in the literature. When exposed to one odor (conditioned stimulus plus; CS+) associated with electric shocks versus another odor (conditioned stimulus minus; CS-) without electric shock, flies learn the association between the CS+ odor and electrical shocks and form an aversive associative olfactory memory. Memory is then scored using a T maze offering a choice between two compartments enriched in the previously negatively reinforced CS+ odor versus the non-reinforced CS- odor. Memory is thus revealed by a selective avoidance of the CS+. After a single training protocol, this memory is short lasting. However, repeated training cycles generate a long-lasting memory that is measurable at least 24 h after training. Multiple training cycles without any resting period (i.e., massed training) form a consolidated memory that persists for at least 24 h and is independent of de novo protein synthesis. So far, this form of consolidated memory has been characterized as anesthesia-resistant memory (ARM) because it is resistant to a cold-shock anesthesia. Interestingly, memory after massed training is socially facilitated, as flies tested in groups perform better than individuals tested alone, which is not the case for short-lasting memory. After massed training, only flies that express ARM are influenced by the social context during memory retrieval, which implies that ARM formed after massed conditioning is required to reveal this socially facilitated memory (hereafter SFM). Another form of consolidated memory can be generated by multiple training cycles performed with a 15-min resting period between each cycle (i.e., spaced training), which leads to a robust memory dependent at least partly on de novo protein synthesis and defined as long-term memory (LTM). Recent work proposed that spaced training leads to a dual memory composed of a safety memory for the CS-, identified as the de novo protein synthesis LTM, and an aversive memory for the CS+, which displays similarities with ARM generated by a massed training.9 Unlike memory generated after massed conditioning, individual memory (i.e., memory performance of a fly tested alone) is much higher and not sensitive to the social context. The lack of influence of the social context after spaced training could be explained by the high individual memory, which would have reached a ceiling effect. Alternatively, the ARM generated by spaced conditioning might be different from that formed by massed training and not be subject to SFM or, although sharing similarities with ARM, the CS+ memory measured after spaced training might not be ARM as formally described in other studies. In any case, only memory formed after massed training is predisposed to SFM, for which memory performance increases in a social context. Although social facilitation of memory retrieval has been reported in humans, the increased memory performance of Drosophila tested in groups constitutes the first example of this phenomenon in invertebrates. Understanding the mechanisms underlying SFM could lead to insight into how social interactions influence cognition (Maria, 2021).

This study has shown that CO2 can act as a facilitating cue leading to an improvement in memory retrieval. Moreover, it was demonstrated that such improvement relies on the expression of ARM formed after a massed training, which is expressed distinctly from individual memory, and the neural network supporting the expression of this additional CO2-sensitive memory was identified. Memory retrieval within a group relies on the recruitment of a second neural network in addition to the one required when flies are tested alone. SFM is not a simple improvement of the expression of an individual memory but constitutes a memory expression in its own right. Therefore, the memory revealed in a social context is actually a composite memory consisting of two previously encoded memories, ASM and ARM, whose expression relies on distinct neural structures. Expression of these memories is indeed independent and additive given that the inhibition of one memory during the retrieval phase does not impair the expression of the other. Thus, this work has provided evidence that ASM is the memory expressed when flies are tested individually and is independent of CO2, whereas SFM has been characterized as the additional expression of ARM in a social context (Maria, 2021).

The predictability of an unconditioned stimulus (US) by an originally neutral stimulus becomes higher upon repetition of the stimulus pairing over extended periods. In Drosophila, two types of aversive long-lasting memories have been characterized. On the one hand, the composite memory described in the present study arises after massed training and is independent of protein synthesis. On the other hand, another form of consolidated memory occurs after spaced training and is dependent on de novo protein synthesis (LTM). Recently, this consolidated memory has been defined as the addition of LTM and ARM, an aversive memory independent of protein synthesis. ARM potentially generated by spaced training and the socially facilitated ARM generated by massed training would involve distinct molecular processes, as suggested by the distinct pathways recruited by spaced and massed trainings. Indeed, serotonin synthesis inhibitor para-chlorophenylalanine (pCPA) treatment, the Drk mutation, or the biVPN blockade (this study) impairs the memory formed after massed training but not the memory generated by spaced training. Like ARM measured after massed conditioning, the CS+ memory measured after spaced training is Radish dependent, which led to its characterization as ARM. However, the memory generated by spaced conditioning does not seem to share the other ARM characteristics detailed above and it should be considered that this CS+ memory would not be ARM in the classical sense, as supported by other studies. In any case, memory formed after spaced training is the most stable form of memory reported in Drosophila and can last up to 7 days post-training. It enables high individual retrieval performances but requires, at least in part, de novo protein synthesis (LTM) involving metabolically costly processes, which can occur at the expense of an animal's fitness under stressful conditions. Similar to aversive LTM formed after spaced training, long-lasting appetitive memory depends on de novo protein synthesis. Interestingly, neither aversive nor appetitive memory dependent on protein synthesis is socially facilitated. The SFM mechanism, purely independent of protein synthesis, would then allow flies to behave appropriately while reducing the costs of learning. Surprisingly, social context does not influence the formation of SFM but rather only its retrieval. This suggests that CO2 possibly released by flies during training does not foster individual learning. this would indicate that the training procedure used in this study generated sufficiently high levels of learning for the influence of the social context to become negligible. Because CO2 is not necessary for the retrieval of memory formed after aversive spaced training, it is concluded that CO2 does not play a general role as a memory enhancer. This aspect deserves further investigation (Maria, 2021).

Besides Drosophila, an influence of the social context on memory retrieval has been highlighted in humans, first addressed by Kenneth Spence in 1956 and summarized by the Drive theory. According to this theory, an individual's performance is potentiated by the presence of other individuals provided that the task performed has been correctly learned beforehand. Social facilitation of memory in Drosophila is consistent with this theory. Yet, because the studies in humans have focused only on short-term restitution, the influence of social context on long-lasting retrieval evinced in the current work remains to be addressed in other taxa, such as rodents or insects. Memory tests are typically conducted on individuals because the characterization of memory refers to an individual's acquisition, storage, and retrieval of information. Yet, in light of the current findings, it would be interesting to determine to what extent social context affects memory retrieval in other animal species (Maria, 2021).

This study showed that CO2 recruits additional circuits leading to the socially facilitated ARM expression. Flies emit and process more CO2 in a group, possibly integrating CO2 as a marker of stress. Therefore, CO2 can be conceived as a stress cue enhancing a fly's attention, changing its representation of the environment, and mediating the expression of an additive memory. Indeed, this study has provided evidence that exposure to CO2 alters the CS- response in DPM neurons, which could stimulate flies' awareness to the CS+ memory trace by inhibiting the responses to the irrelevant CS- stimulus. In vertebrates, moderate stress can promote aversive long-lasting memory. Although memory mechanisms described for vertebrates differ from those in the current model, the benefits of moderate stress on memory seem to be common across species (Maria, 2021).

So far, the role of CO2 in insect behavior has been mostly limited to naive avoidance and attraction. This study reveals an important role for CO2 as a facilitator of olfactory memory. In natural environments, CO2 is a ubiquitous cue, including within the nest of eusocial insects such as ants, termites, or bees8 that can be potentially significant and attractive. It is an attractive cue for insects at food sources and oviposition sites and also plays a key role in host detection for hematophagous insects such as tsetse flies or mosquitoes. Olfactory learning plays a significant role in host preference and disease transmission in blood-feeding insects. Thus, exploring the impact of CO2 on memory processes in these insects would be interesting to develop and improve control strategies to reduce the risk of disease transmission. These findings suggest that CO2 may have an unsuspected impact on the cognition of a broad spectrum of insect species (Maria, 2021).

Go to top

Prior experience conditionally inhibits the expression of new learning in Drosophila

Jacob, P. F., Vargas-Gutierrez, P., Okray, Z., Vietti-Michelina, S., Felsenberg, J. and Waddell, S. (2021). Curr Biol. PubMed ID: 34146482

Prior experience of a stimulus can inhibit subsequent acquisition or expression of a learned association of that stimulus. However, the neuronal manifestations of this learning effect, named latent inhibition (LI), are poorly understood. This study shows that prior odor exposure can produce context-dependent LI of later appetitive olfactory memory performance in Drosophila. Odor pre-exposure forms a short-lived aversive memory whose lone expression lacks context-dependence. Acquisition of odor pre-exposure memory requires aversively reinforcing dopaminergic neurons that innervate two mushroom body compartments-one group of which exhibits increasing activity with successive odor experience. Odor-specific responses of the corresponding mushroom body output neurons are suppressed, and their output is necessary for expression of both pre-exposure memory and LI of appetitive memory. Therefore, odor pre-exposure attaches negative valence to the odor itself, and LI of appetitive memory results from a temporary and context-dependent retrieval deficit imposed by competition with the parallel short-lived aversive memory (Jacob, 2021).

Keeping track of life experience allows animals to benefit from all of their prior knowledge when learning new information and using their memory to direct behavior. Although the subject of great early debate among learning theorists, it is now accepted that learning occurs even without explicit rewards or punishment. Classic experiments showed that rats given the prior opportunity to roam in an empty maze performed better when they were later trained with rewards presented in specific locations (Jacob, 2021).

Becoming familiar with the maze without explicit reinforcement and an obvious initial change in the animal's behavior was called 'latent learning.' Attempts to replicate a facilitating effect of latent learning using classical conditioning led to an unexpected observation. Pre-exposing animals to a stimulus instead often inhibited the ability of the animal to learn using that stimulus--a phenomenon given the name 'latent inhibition' (LI) (Jacob, 2021).

LI has been heavily studied for the last 50 years, and two alternative theories have been proposed to account for the inhibitory effect of stimulus pre-exposure. In the acquisition (A) model, subsequent learning is considered to be impaired because pre-exposure alters the capacity for the stimulus to enter into new associations. In contrast, in the retrieval (R) model, learning is still believed to occur but memory expression is impaired. A strong argument in favor of the R model is the observation that LI often appears to be limited in time, leading to expression of the subsequent learning undergoing 'spontaneous recovery.' Importantly, both theories of latent inhibition assume that something is learned during pre-exposure such as primitive properties of the stimulus including its specific identity, intensity (e.g., concentration), and salience. In addition, LI is often sensitive to the consistency of the context within which the animal is pre-exposed, taught, and tested for memory expression. This led to the proposal that first learning an association between the stimulus and its context makes it difficult for the animal to subsequently associate the stimulus with reinforcement during training. Studying olfactory learning in the relatively small brain of Drosophila has potential to define how LI can operate and reveal an underlying neuronal circuit mechanism. Several earlier studies in both adult flies and larvae demonstrated that repeated exposure to an odor can alter its apparent valence to the fly, either making the fly avoid it more or become unresponsive to it. Although an A model for LI has been reported with appetitive conditioning in the honeybee, a prior study in adult Drosophila did not observe any effect on aversive conditioning following a single odor pre-exposure (Jacob, 2021).

Associative olfactory learning in Drosophila relies on the neuronal circuitry of the mushroom body (MB). Individual odors are represented as activity in sparse and largely non-overlapping subpopulations of the ∼4,000 intrinsic neurons called Kenyon cells (KCs). Positive or negative valence can be assigned to these odor representations by anatomically discrete dopaminergic neurons (DANs) which, via dopamine receptor-directed cyclic AMP (cAMP)-dependent plasticity, modulate the efficacy of KC output synapses onto different downstream mushroom body output neurons (MBONs), whose dendrites occupy the same MB compartment. Aversive learning depresses KC synapses onto MBONs whose activation favors approach, whereas appetitive learning reduces odor-drive to MBONs favoring avoidance. By establishing a skew in the valence of the odor-driven MBON network, learned information subsequently directs either odor avoidance or attraction behavior (Jacob, 2021).

A number of studies indicate that discrete experience is represented as plasticity of different combinations of KC-MBON connections, directed by the engagement of unique combinations of DANs. For example, different types of DANs have been implicated in coding memories for specific rewards (e.g., water, the sweet taste and nutrient value of sugars, the absence of expected shock, and the delayed recognition of safety) (Jacob, 2021).

In contrast, the same PPL1 DANs appear to be required to code aversive memories for electric shock, bitter taste, and heat, although imaging suggests they are activated by temperature decreases and to noxious heat (Jacob, 2021).

By forming and storing conflicting and complementary memories in different places, the fly can more effectively direct its behavior to reflect a history of experience. This study shows that prior odor exposure can temporarily inhibit memory performance after subsequent appetitive learning in Drosophila. This inhibitory effect is sensitive to a change of context across the pre-exposure, training, and testing periods, consistent with it being a form of LI. Odor pre-exposure forms a short-lived odor-specific aversive memory, whose acquisition requires the γ2α'1 and α3 DANs, the latter of which become sensitized to consecutive odor presentation. As a consequence, aversive memory is apparent as a decrease in the odor-evoked response of the corresponding approach-directing γ2α'1 and α3 MBONs. Blocking the α3 MBONs impairs the expression of the aversive odor pre-exposure memory and abolishes LI of appetitive memory. The short-lived presence of a parallel and differently located odor-specific aversive memory therefore temporarily inhibits the retrieval of a subsequently formed appetitive memory for that same odor. These data provide evidence for a context-dependent R model of latent inhibition in Drosophila (Jacob, 2021).

This study has demonstrated a form of latent inhibition (LI) in Drosophila and has identified an underlying neuronal mechanism. Repeated odor presentation can form a labile self-reinforced aversive memory for that odor, that can temporarily compete with the expression of a newly acquired appetitive memory for that same odor. During memory testing, the conditioned odor should therefore activate both the memory of the pre-exposure (odor-self) and that of the appetitive conditioning (odor-sugar). Importantly, the aversive pre-exposure memory is labile, which means that the LI effect is transient. As a result, the appetitive memory performance exhibits 'spontaneous recovery.' These results demonstrate that an R model underlies this form of LI in the fly. An A model is not supported because flies acquire an associative reward memory for the odor after pre-exposures of that odor. Instead, the expression of the learned approach performance is impeded by the co-expression of a competing aversive pre-exposure memory. In further support of this R model, the same pre-exposure regimen caused facilitation of a subsequently acquired aversive olfactory memory. In this instance, the pre-exposure memory adds to the new aversive associative memory, rather than competes with an appetitive memory. Last, pre-exposure to the to be non-reinforced odor (the CS−) enhanced subsequent appetitive memory performance but inhibited aversive memory performance-a logical expectation of the CS− acquiring negative valence during pre-exposure (Jacob, 2021).

A defining feature of LI is sensitivity to the consistency of the context in which the pre-exposure, learning, and testing are carried out. Changing between the clear and paper-lined tubes did not impair LI, suggesting that the flies likely consider these to be a similar context. However, if odor pre-exposure, learning, and testing were performed in different contexts (i.e., a copper grid-lined versus a paper-lined or clear tube) LI was abolished. Most strikingly, LI could be restored if copper grid tubes were used to provide the same context during pre-exposure and testing. In line with prior theories and studies of LI in other animals, these results suggest that flies learn an association between the odor and the context in which it is experienced during the non-reinforced pre-exposure. As a result, the pre-exposure memory gains context-dependence, and experiments show it is not retrieved and therefore does not interfere with the newer appetitive memory if the context is different when memory is tested. The failure to retrieve the pre-exposure memory in a different context manifests as a loss of LI-the appetitive memory is fully expressed. This study therefore reveals that the context-dependency of LI results from the ability (correct context, LI evident) or inability (wrong context, no LI) to retrieve the pre-exposure memory. In addition, context only plays a role in the expression of the pre-exposure memory when it is in conflict with a subsequently acquired appetitive memory. Further work will be required to define what the flies recognize as a 'change of context.' There are many possibilities including background odors, tube/paper texture, relative luminance, and other flies in the group (Jacob, 2021).

This study found that the odor-driven activity of γ2α'1 and α3 DANs increased with repeated odor pre-exposure and that they were required for the formation of the odor pre-exposure memory. In addition, the odor-specific responses of the corresponding MBONs were depressed following pre-exposure. It was therefore concluded that ramping odor-driven DAN activity assigns negative value to the odor itself by depressing odor-specific KC connections onto the γ2α'1 and α3 MBONs. In support of this model, repeated pre-exposure of flies to a lower and less innately aversive odor concentration did not increase the activity of the α3 DANs or form an aversive pre-exposure memory. Importantly, reduced odor activation of the approach-directing γ2α'1 and α3 MBONs is sufficient to account for the aversive nature of pre-exposure memory. Moreover, both expression of pre-exposure memory and LI are abolished if the α3 MBONs are blocked during testing, confirming the model that LI is produced by the expression of the aversive pre-exposure memory competing with that of the associative reward memory believed to be represented as depression of conditioned odor responses of γ5, β'2, and α1 MBONs. Several prior Drosophila studies have documented changes in odor-driven behavior following different regimens of odor exposure, many of which employed longer durations or more trials than those employed here, and that produced shorter-lived inhibitory effects (Jacob, 2021).

One of these studies described odor-driven activity of the PPL1-α'3 DANs, and subsequent depression of odor-specific responses of the α'3 MBONs to underlie how flies can become familiar with an odor following repeated short exposures. In contrast, this study shows that two longer and spaced odor exposures produce an aversive memory that manifests as plasticity of γ2α'1 and α3 DANs and MBONs. Moreover, whereas this study shows a retrieval defect underlies LI of appetitive memory, a reduced attention/familiarity/habituation to the odor following pre-exposures would be expected to result in a subsequent acquisition defect (and A model), likely of both appetitive and aversive learning. It will nevertheless be important to understand how these different types of olfactory experience, and their supporting plasticity mechanisms, are represented and combined in the brain (Jacob, 2021).

LI has often been compared to memory extinction, and the current work in the fly shows that very similar neuronal mechanisms and learning models account for both of these phenomena. Pre-exposure learning in Drosophila appears to follow similar rules to extinction learning following aversive olfactory conditioning: 2 spaced trials with 15-min ITI are more efficient than massed training with 1-min ITI, and in both cases, a resulting parallel opposing odor-nothing memory inhibits the retrieval/expression of the odor-punishment or odor-reward memory. The obvious difference is that the interfering non-reinforced odor memory is formed before learning for LI and after learning for extinction (Jacob, 2021).

These studies of learning, extinction, and LI suggest that flies acquire and store all of their experience (rewarded and unrewarded, punished and unpunished) as parallel memory traces. As a result, when evoked by an appropriate cue, the relevant experiences are compared/combined at the time of retrieval to determine the most fitting behavioral outcome. Such a model is reminiscent of the comparator hypothesis, devised mostly from experiments in rodents. Because recent studies suggest similar processes underlie extinction of fear in flies, rodents, and humans, it seems likely that the form and mechanism of LI described in this study will also be relevant across phyla (Jacob, 2021).

Go to top

Synthesis of conserved odor object representations in a random, divergent-convergent network

Endo, K., Tsuchimoto, Y. and Kazama, H. (2020). Neuron. PubMed ID: 32814018

Animals are capable of recognizing mixtures and groups of odors as a unitary object. However, how odor object representations are generated in the brain remains elusive. This study investigated sensory transformation between the primary olfactory center and its downstream region, the mushroom body (MB), in Drosophila and shows that clustered representations for mixtures and groups of odors emerge in the MB at the population and single-cell levels. Decoding analyses demonstrate that neurons selective for mixtures and groups enhance odor generalization. Responses of these neurons and those selective for individual odors all emerge in an experimentally well-constrained model implementing divergent-convergent, random connectivity between the primary center and the MB. Furthermore, this study found that relative odor representations are conserved across animals despite this random connectivity. These results show that the generation of distinct representations for individual odors and groups and mixtures of odors in the MB can be understood in a unified computational and mechanistic framework (Endo, 2020).

Recognizing objects in a complex environment is a hallmark of sensory systems. The olfactory system is not an exception, as, for example, it has an ability to recognize mixtures of odorants emitted from a single source as a unitary object despite slight differences in composition or ratio of components, alongside the ability to detect individual component odorants. It also recognizes different odors as a group if they share some physicochemical features that give rise to a particular perceptual quality, such as 'minty' or 'woody'. The ability to recognize mixtures and groups of odors as objects allows animals to generalize the ethological meaning of the previously experienced odors to newly encountered ones, either to mixtures of different mixing ratios or to other odors of the same group (Endo, 2020).

Previous studies have implicated that odor object representations are generated in the olfactory association area, the piriform cortex in mammals and the mushroom body (MB) in insects. Anatomically, piriform and MB neurons receive divergent and convergent inputs from output neurons in a subset of glomeruli in the olfactory bulb and the antennal lobe (AL), respectively and thus are in a position to combine odor features encoded separately in glomeruli to form odor object representations. Indeed, the rodent piriform cortex represents mixtures as not a mere sum of individual components and does not exhibit significant cross-habituation between mixture and components. Moreover, the human posterior piriform cortex clusters chemically and structurally diverse odorants into groups with distinct perceptual qualities in its neural response space. However, whether and how these odor object representations first emerge in the cortex remains largely unknown, as responses are seldom compared with those in the earlier layers of the circuit under the same experimental condition (Endo, 2020).

More tractable olfactory systems of insects have contributed to the understanding of mechanisms underlying olfactory computations in the association area. Electrophysiological studies have examined the transformation between projection neurons (PNs) in the AL and Kenyon cells (KCs) in the MB, with an emphasis on temporal processing, and imaging studies have examined the spatial aspect of KC odor responses at cellular and subcellular resolution. These studies have characterized the sparse and decorrelated nature of odor responses, which is proposed to be beneficial for discrimination between individual odors. However, mixture- and group-specific odor representations have yet to be uncovered, likely because the modest fraction of KCs (<5%) previously recorded in each brain may not have been sufficient to capture the full aspects of odor representations, especially given that odors are sparsely encoded without apparent spatial organization by physiologically unidentifiable KCs (Endo, 2020).

Using brain-region-wide calcium imaging, this study investigated the transformation of olfactory representations in the Drosophila MB by recording from PNs in most of the AL glomeruli and all ∼2,000 KCs in the MB and found that besides distinct representations for individual odors, clustered representations for mixtures and groups of odors emerge in the MB at both the population and single-cell levels. A classification approach showed that these clustered representations are useful for generalization of odors. Model KCs implementing the experimentally reported mechanisms recapitulated these responses. Strikingly, this study further found that the correlational structure of population odor representations is conserved across animals in the MB, even though KCs sample PNs arbitrarily in general and stereotypy is absent in the responses of genetically tagged, individual KCs . This preservation of relative odor representations by random projections was proposed by multiple theoretical studies, and therefore, the results provide concrete evidence in favor of this prediction. In sum, random, divergent-convergent circuitry in the MB synthesizes conserved, odor object representations to support olfactory generalization (Endo, 2020).

Because previous studies focused on the sparse and decorrelated nature of odor information in the MB, partial sampling of KCs sufficed to draw the conclusion. In contrast, this study aimed at evaluating the ability of the MB to represent various groups and mixtures of odors as distinct objects by recording from virtually all the KCs. This was accomplished using volumetric Ca2+ imaging, which provides a wide spatial coverage in exchange for fine temporal resolution. Importantly, odor responses were also recorded from PNs in most of the AL glomeruli under identical experimental conditions. This dataset therefore allowed comprehensive analysis of the transformation of spatial codes between two directly connected layers of a central circuit at the granularity of single glomeruli/cells (Endo, 2020).

By comparing the odor responses in the AL and MB, the MB was found to represent not only individual monomolecular odors but also groups of odors more distinctly. Accordingly, the ability to generalize odors was higher in the MB than in the AL. These population odor representations were built from individual KCs that respond selectively to multiple or single odors of a given group. This classification analysis indicated that such multi-selective KCs are the key contributors to odor generalization (Endo, 2020).

It was noticed that clustered representations are not restricted to groups of odors that share obvious physicochemical features like alcohols or benzenoids. Generalization accuracy was high for many groups consisting of arbitrarily chosen odors and was only weakly correlated to the mean physicochemical distance between odors belonging to a given group. This may reflect that even though the physicochemical distance is calculated from a large number of molecular descriptors, it cannot be a measure of similarity of all the features of odors detected by the olfactory receptors. If odors share some elemental features that activate a common set of glomeruli, then KCs connected to these glomeruli will be recruited and represent the odors as a group along one of the dimensions of the MB coding space. This was exemplified by the representations of 2,3-butanedione and ethyl butyrate in the MB principle component (PC) space. Although these two odors are not entirely similar, as they are at the opposite ends along PC3, they are at the same end of PC2 and thus are represented as a group that is linearly separable from the other odors. Therefore, the high dimensionality of the MB coding space constructed from independent KCs, each of which represents a specific combination of glomerular inputs, allows various groups of odors to be represented distinctly in this brain region (Endo, 2020).

This study also found that clustered representations of odor mixtures emerge in the MB. Animals can perceive mixtures as an odor object, and insects exhibit behavioral responses to mixtures that are dependent on the MB and are similarly explained by configural rather than elemental processing. This study has revealed mixture representations in the MB, which are invariant to mixing ratios and at the same time distinct from component representations, a neural correlate for configural processing of mixtures. At the level of individual cells, KCs were found that respond specifically and invariantly to mixtures at diverse mixing ratios (Endo, 2020).

In parallel, there were KCs responding exclusively and invariantly to odors containing one of the components. These are reminiscent of the KCs in the locust MB that signaled the presence of a mixture component (Shen et al., 2013) and may underlie the animals' ability to identify components in a mixture (Smith, 1998; Reinhard et al., 2010; Rokni et al., 2014; Schneider et al., 2016). Therefore, representations in support of elemental as well as configural processing coexist in the MB. This does not represent a paradox, because individual KCs sample different sets of glomeruli and represent mixtures and components independently along different dimensions of the MB coding space (Endo, 2020).

How, then, are responses of these KCs generated in the MB? Although the contribution of as yet-unreported factors is anticipated, the simulation approach demonstrated that weighted summation of inputs from a random subset of PNs with global inhibition and nonlinear thresholding, all of which are experimentally reported mechanisms, are sufficient to broadly recapitulate the actual odor responses of KCs. Notably, responses of single-selective, multi-selective (group-selective), and mixture-selective KCs all emerged from the simulation, suggesting that generation of representations for individual odors and groups and mixtures of odors can be understood in a unified computational and mechanistic framework (Endo, 2020).

The circuitry of the insect MB has been argued to resemble that of the vertebrate cerebellar cortex, where divergent-convergent feedforward connections of mossy fibers onto granule cells are considered to function as a pattern separator. Based on the results of simulation, a view is proposed that the divergent-convergent connections rather act as a pattern generator, arbitrarily combining glomerular inputs and synthesizing them into diverse clustered representations not only for individual odors but also for groups of odors in the MB (Endo, 2020).

This view might also be instructive in explaining as to why olfactory receptors first decompose information about odors into specific physicochemical features and represent them in distributed glomeruli. It has been considered that decomposition is beneficial for the olfactory system to implement combinatorial coding and therefore increase its coding capacity. The results of this study suggest that decomposition is also beneficial for decoding; it allows KCs to integrate elemental olfactory features in various ways via random connectivity to generate a variety of odor object representations for later flexible association (Endo, 2020).

The MB has been seen to represent a variety of odor groups and mixtures along different dimensions of a coding space. The valence of some of these odor groups may initially be neutral, in which case the clustered representations in the MB themselves would be of little use to the animal's action selection. Therefore, an important subsequent task for the neural system is to select a relevant odor group and assign an ethological meaning to it based on the animal's experience. This study has shown that this can be conceptually accomplished by a linear decoder that is trained to classify odor groups and mixtures from others (Endo, 2020).

The likely biological decoders are MB output neurons (MBONs), each receiving synaptic inputs from a large number of KCs. Synaptic inputs that are specifically recruited by a conditioned odor will undergo a long-term plastic change following an association with an unconditioned stimulus. Because this plastic change in synaptic weights is equivalent to fitting the weights for individual KCs in the decoding analysis, in principle, any odor object representations in the MB can be flexibly and accurately linked to behavior following an olfactory association (Endo, 2020).

Although previous studies have provided some hints, little was known about the stereotypy of odor representations. This study has shown at cellular resolution that the relative odor representations in the MB are highly conserved across animals at the population level, even though no such stereotypy was discernible at the single-cell level (Endo, 2020).

One may expect that the stereotypy of odor responses in the AL will be lost in the MB due to the overall random connectivity between glomeruli and KCs. However, this random projection likely is the very mechanism for preserving stereotypy, because, theoretically, it can embed data from one Euclidean space into another of different dimensions without substantially altering the distance between data points. Multiple theoretical studies also indicate that relative odor representations are preserved by random projections. Therefore, although odor representations become more distinct through the AL-to-MB transformation, their correlational structure is maintained, providing a neural basis for individual animals to express shared behavioral responses to the same set of odors following discriminative olfactory conditioning (Endo, 2020).

Go to top

Metabolic learning and memory formation by the brain influence systemic metabolic homeostasis

Zhang, Y., Liu, G., Yan, J., Zhang, Y., Li, B. and Cai, D. (2015). Nat Commun 6: 6704. PubMed ID: 25848677

Metabolic homeostasis is regulated by the brain, but whether this regulation involves learning and memory of metabolic information remains unexplored. This study use a calorie-based, taste-independent learning/memory paradigm to show that Drosophila form metabolic memories that help in balancing food choice with caloric intake; however, this metabolic learning or memory is lost under chronic high-calorie feeding. Loss of individual learning/memory-regulating genes causes a metabolic learning defect, leading to elevated trehalose and lipid levels. Importantly, this function of metabolic learning requires not only the mushroom body but also the hypothalamus-like pars intercerebralis, while NF-κB activation in the pars intercerebralis mimics chronic overnutrition in that it causes metabolic learning impairment and disorders. Finally, this study evaluated this concept of metabolic learning/memory in mice, suggesting that the hypothalamus is involved in a form of nutritional learning and memory, which is critical for determining resistance or susceptibility to obesity. In conclusion, these data indicate that the brain, and potentially the hypothalamus, direct metabolic learning and the formation of memories, which contribute to the control of systemic metabolic homeostasis (Zhang, 2015).

This work consists of a series of studies in Drosophila: these animals were found to temporarily develop metabolic learning to balance food choice with caloric intake. In Drosophila research, sugar has often been used for studying the appetitive reward value of food taste. Of interest, recent research has suggested that fruit flies can distinguish caloric values from the taste property of food. Using tasteless sorbitol as a carbohydrate source to generate an environmental condition that contained NC versus HC food, this study revealed that Drosophila can develop a form of metabolic learning and memory independently of taste, by which flies are guided to have a preference for normal caloric environment rather than high-caloric environment. However, this form of metabolic memory does not seem robust, as it is vulnerably diminished under genetic or environmental influences. It is postulated that this vulnerability to overnutrition is particularly prominent for mammals (such as C57BL/6J mice), and overnutritional reward-induced excess in caloric intake can quickly become dominant. This effect can be consistently induced in Drosophila when learning/memory-regulating genes are inhibited in the brain or the hypothalamus-like PI region. It was observed that each of these genetic disruptions led to impaired metabolic learning, resulting in increased caloric intake and, on a chronic basis, the development of lipid excess and diabetes-like phenotype. Indeed, it has been documented that chronic high-sugar feeding is sufficient to cause insulin resistance, obesity and diabetes in Drosophila. It is yet unclear whether this metabolic learning can induce an appetitive memory of normal caloric environment or an aversive memory of high-caloric environment. Regardless, the findings in this work have provoked a stimulating question, that is, whether this form of metabolic learning and memory is present in the mammals and, if so, whether this mechanism can be consolidated to improve the control of metabolic physiology and prevent against diseases. These mouse studies may provide an initial support to this concept and strategy, but clearly, in-depth future research is much needed (Zhang, 2015).

In light of the underlying neural basis for this metabolic learning, this study indicates that multiple brain regions are required, including the hypothalamus-like PI region in addition to the MB (equivalent to the hippocampus in mammals), which is classically needed for learning and memory formation. Anatomically, the PI region is located in the unpaired anteromedial domain of the protocerebral cortex, which is near the calyces of the MB and the dorsal part of the central complex (another brain region for regulating learning and memory). Functionally, the PI region has been demonstrated to coordinate with the MB in regulating various physiological activities in Drosophila. Thus, it is very possible that some PI neurons present nutritional information to the MB and thus induce metabolic learning and memory formation. However, the underlying detailed mechanism is still unknown, especially if this process involves a role of dilps, which represent the prototypical neuropeptides produced by the PI neurons. Considering that the PI region in flies is equivalent to the mammalian hypothalamus, this study was extended to mouse models by comparatively analysing A/J versus C57BL/6J mice—which are known to have different diet preference as well as different susceptibilities to obesity development. While A/J mice showed a learning process of distinguishing NC versus HC food, C57BL/6J mice failed to do so. It is particularly notable that this difference of learning and memory between these two strains is associated with differential expression profiles of learning/memory genes in the hypothalamus rather than the hippocampus. This finding, in conjunction with the Drosophila study, highlights a potential that the hypothalamus has a unique role in mediating metabolic learning and memory formation. Although the mouse experiments cannot exclude the impacts from the taste/smell properties of the studied food, the results demonstrated that there is a form of nutritional memory, which seems dissociable from the memory of overnutritional reward. These initial observations in mice lend an agreement with findings in Drosophila, suggesting that the brain and potentially the hypothalamus can link nutritional environment to a form of metabolic learning and memory homeostasis (Zhang, 2015).

From a disease perspective, metabolic learning in Drosophila is impaired under chronic overnutrition, and the mouse study was in line with this understanding. This response to overnutrition is useful when famine is outstanding; however, it is a dilemma when metabolic disease is of concern, much like the scenario pertaining to leptin resistance under chronic overnutrition, whereas an increase in leptin sensitivity is demanded to reduce obesity. Recently, it was established that NF-κB-dependent hypothalamic inflammation links chronic overnutrition to the central dysregulation of metabolic balance. This study showed that activation of the NF-κB pathway in the PI region weakened the function of metabolic learning and, conversely, NF-κB inhibition in this region provided a protective effect against chronic overnutrition-impaired metabolic learning. These findings are in alignment with the literature, for example, pan-neuronal NF-κB inhibition was shown to improve activity-dependent synaptic signalling and cognitive function including learning and memory formation, and persistent NF-κB activation inhibits neuronal survival and the function of learning and memory formation. Hence, overnutrition-induced neural NF-κB activation has a negative impact on metabolic learning and memory formation in regulating metabolic homeostasis homeostasis (Zhang, 2015).

To summarize the findings in this work, a series of behavioural studies was performed revealing that Drosophila have a form of metabolic learning and memory, through which the flies are directed to balance food choice with caloric intake in relevant environments. Several learning/memory-regulating genes including rut, dnc and tequila are involved in this function, and brain regions including the PI in addition to the MB are required to induce this mechanism. On the other hand, metabolic learning is impaired under chronic overnutrition through NF-κB activation, leading to excess exposure to calorie-enriched environment, which causes metabolic disorders. Overall, metabolic learning and memory formation by the brain and potentially the hypothalamus play a role in controlling metabolic homeostasis homeostasis (Zhang, 2015).

Go to top

Spatio-temporal in vivo recording of dCREB2 dynamics in Drosophila long-term memory processing

Zhang, J., Tanenhaus, A. K., Davis, J. C., Hanlon, B. M. and Yin, J. C. (2014). Neurobiol Learn Mem 118C: 80-88. PubMed ID: 25460038

CREB (cAMP response element-binding protein) is an evolutionarily conserved transcription factor, playing key roles in synaptic plasticity, intrinsic excitability and long-term memory (LTM) formation. The Drosophila homologue of mammalian CREB, dCREB2, is also important for LTM. However, the spatio-temporal nature of dCREB2 activity during memory consolidation is poorly understood. Using an in vivo reporter system, this study examined dCREB2 activity continuously in specific brain regions during LTM processing. Two brain regions that have been shown to be important for Drosophila LTM are the ellipsoid body (EB) and the mushroom body (MB). dCREB2 reporter activity is persistently elevated in EB R2/R4m neurons, but not neighboring R3/R4d neurons, following LTM-inducing training. In multiple subsets of MB neurons, dCREB2 reporter activity is suppressed immediately following LTM-specific training, and elevated during late windows. In addition, heterogeneous responses were observed across different subsets of neurons in MB αβ lobe during LTM processing. All of these changes suggest that dCREB2 functions in both the EB and MB for LTM formation, and that this activity contributes to the process of systems consolidation (Zhang, 2014).

Go to top

Tip60 HAT action mediates environmental enrichment induced cognitive restoration

Xu, S., Panikker, P., Iqbal, S. and Elefant, F. (2016). PLoS One 11: e0159623. PubMed ID: 27454757

Environmental enrichment (EE) conditions have beneficial effects for reinstating cognitive ability in neuropathological disorders like Alzheimer's disease (AD). While EE benefits involve epigenetic gene control mechanisms that comprise histone acetylation, the histone acetyltransferases (HATs) involved remain largely unknown. This study examine a role for Tip60 HAT action in mediating activity- dependent beneficial neuroadaptations to EE using the Drosophila CNS mushroom body (MB) as a well-characterized cognition model. Flies raised under EE conditions were shown to display enhanced MB axonal outgrowth, synaptic marker protein production, histone acetylation induction and transcriptional activation of cognition linked genes when compared to their genotypically identical siblings raised under isolated conditions. Further, these beneficial changes are impaired in both Tip60 HAT mutant flies and APP neurodegenerative flies. While EE conditions provide some beneficial neuroadaptive changes in the APP neurodegenerative fly MB, such positive changes are significantly enhanced by increasing MB Tip60 HAT levels. These results implicate Tip60 as a critical mediator of EE-induced benefits, and provide broad insights into synergistic behavioral and epigenetic based therapeutic approaches for treatment of cognitive disorder (Xu, 2016).

Go to top

Drosophila neprilysins are involved in middle-term and long-term memory

Turrel, O., Lampin-Saint-Amaux, A., Préat, T. and Goguel, V. (2016). J Neurosci 36: 9535-9546. PubMed ID: 2762970

Neprilysins are type II metalloproteinases known to degrade and inactivate a number of small peptides. Neprilysins in particular are the major amyloid-β peptide-degrading enzymes. In mouse models of Alzheimer's disease, neprilysin overexpression improves learning and memory deficits, whereas neprilysin deficiency aggravates the behavioral phenotypes. However, whether these enzymes are involved in memory in nonpathological conditions is an open question. Drosophila melanogaster is a well suited model system with which to address this issue. Several memory phases have been characterized in this organism and the neuronal circuits involved are well described. The fly genome contains five neprilysin-encoding genes, four of which are expressed in the adult (see Neprilysin 4). Using conditional RNA interference, this study shows that all four neprilysins are involved in middle-term and long-term memory. Strikingly, all four are required in a single pair of neurons, the dorsal paired medial (DPM) neurons that broadly innervate the mushroom bodies (MBs), the center of olfactory memory. Neprilysins are also required in the MB, reflecting the functional relationship between the DPM neurons and the MB, a circuit believed to stabilize memories. Together, these data establish a role for neprilysins in two specific memory phases and further show that DPM neurons play a critical role in the proper targeting of neuropeptides involved in these processes (Turrell, 2016).

Research on neprilysins has essentially focused on their role as the main Aβ-degrading enzymes in pathological situations and as biomarkers in heart failure. Using Drosophila this study has established that neprilysins are involved in specific types of memory. Disrupting the expression of any neprilysin impairs MTM and LTM, revealing that one or several neuropeptides need to be targeted to enable proper memory formation. Interestingly, all four neprilysins expressed in the fly are required in the MB and also in DPM neurons, a pair of large neurons that broadly innervates the MB and are involved in memory consolidation (Turrel, 2016).

Neprilysins have been described extensively as proteases acting on substrates of no more than 50 residues, except Drosophila Nep4, which is involved in muscle integrity independently of its catalytic activity. There is a consensus that neprilysins function by turning off neuropeptide signals at the synapse. In addition, there is evidence to suggest that neprilysin processing could lead to the activation of neuromodulators. Therefore, in addition to their role in Aβ degradation, neprilysins also inactivate a large number of peptides and are thus equally involved in a large number of processes (Turrel, 2016).

In Drosophila, several small peptides have been linked to olfactory memory. Short neuropeptide F (sNPF) is highly expressed in the MB and has been described as a functional neuromodulator of appetitive memory. Drosophila neuropeptide F (dNPF) has been shown to provide a motivational switch in the MB that controls appetitive memory output. Interestingly, dNPF is an ortholog of mammalian NPY, a peptide identified as an hNEP substrate. hNEP can process NPY in transgenic mice to produce neuroactive fragments. Because components of dNPF/NPY signaling are conserved at both the functional and molecular levels, it is possible that dNPF is targeted by neprilysins. It remains to be determined whether such peptides are involved in aversive memory and, conversely, whether neprilysins are involved in appetitive memory (Turrel, 2016).

Although all four Drosophila neprilysins are involved in identical memory phases, they exhibit distinct features in terms of the neuronal circuits involved. Only Nep1 inhibition in α/β MB and DPM neurons alters both MTM and LTM. One hypothesis is that Nep1 expressed in DPM and MB neurons plays the same role, targeting a single substrate at synapses connecting the two structures. If so, the lifetime of such a substrate would need to be restricted strictly to limit its effect (Turrel, 2016).

Like the other neprilysins, Nep2 is involved in MTM and LTM, but it exhibits a peculiar characteristic: Nep2 inhibition in DPM neurons leads to MTM disruption, whereas it does not alter LTM. Although it cannot be ruled out that Nep2 expression in DPM neurons is required for LTM, but that its silencing does not reach a level critical for this process, the data suggest that Nep2 expression in DPM neurons is not required for LTM formation. It is noteworthy that neprilysins are synthesized as type II integral membrane proteins, whereas Nep2 is a soluble secreted endopeptidase. Whether an endopeptidase is tethered or fully secreted will have important implications in terms of field of activity and enzyme concentration at the membrane surface. It is possible that Nep2 secreted from either DPM neurons or another structure in the vicinity, such as MB neurons, is able to play an identical role. Therefore, Nep2 reduction in either neuronal structure would not be sufficient to affect LTM. In contrast, Nep2 expression in such a structure would not be able to compensate Nep2 silencing in DPM neurons for MTM, pointing to a distinct requirement for MTM and LTM formation. Nep2 might be required at a distinct concentration and/or localization for MTM and LTM or it may target distinct substrates for these two processes (Turrel, 2016).

The data reveal functional redundancy among Nep2, Nep3, and Nep4 for LTM formation in the α/β neurons. Namely, concomitant silencing of Nep2 + Nep3 or Nep3 + Nep4 leads to altered LTM. It is possible that several neprilysins target a single neuropeptide. However, because concomitant silencing of Nep2 + Nep4 does not affect LTM, it seems more likely that several targets are involved in LTM (Turrel, 2016).

The memory phenotypes observed after each neprilysin reduction are reminiscent of the pattern in APPL mutants. Indeed, it was shown previously that expression of APPL, the APP fly ortholog, is required in the MB for MTM and LTM formation, but not for learning and ARM (Goguel, 2011; Bourdet, 2015). An attractive hypothesis is that Aβ peptide derived from physiological processing of APPL might play a role in memory and act as a substrate for one of the neprilysin peptidases. Nep2 would be a good candidate because several studies have shown that it is capable of degrading human Aβ. Supporting this hypothesis, several reports in mammals have implicated low physiological concentrations of Aβ peptide in memory formation (Turrel, 2016).

The memory phenotypes observed in this study are equally reminiscent of the pattern displayed by amnesiac (amn) mutants. The amn gene isolated through behavioral screening for memory mutants was later shown to encode a predicted neuropeptide precursor. Although the mature products of the amn gene have not been identified, sequence analyses suggested the existence of three potential peptides. One of them is homologous to mammalian pituitary adenylate cyclase-activating peptide (PACAP), a neuromodulator and neurotransmitter that regulates a variety of physiological processes through stimulation of cAMP production. In vitro studies have shown that hNEP can degrade PACAP and the analysis of the biological properties of the resulting fragments found that PACAP degradation by hNEP produces active metabolites selective for a particular receptor subtype. One of the major sites of PACAP cleavage by hNEP is conserved in the AMN peptide. It was shown that AMN is highly expressed in DPM neurons, where the four neprilysins are required for MTM. DPM output is required during the consolidation phase for MTM and it was suggested that DPM might release the AMN modulatory neuropeptide that alters the physiology of MB neurons to help stabilize or consolidate odor memories. The fact that neuropeptides are often coreleased with classical neurotransmitters, but generally have slower and longer-lasting postsynaptic effects, has prompted the hypothesis that AMN peptides may be released at the MB to produce relatively long-lasting, physiological changes. Given this context, it is tempting to speculate that AMN might be one of the Drosophila neprilysin's targets (Turrel, 2016).

Both the axons and dendrites of DPM are evenly distributed in different lobes of the MB, and it has been suggested that DPM neurons are presynaptic and postsynaptic to the MB neurons and are recurrent feedback neurons. Because neprilysins are necessary in the DPM, and also in the α/β neurons of the MB where MTM and LTM are stored, these proteins could be involved in maintaining a loop between the DPM and MB lobes by restricting the lifetime of neuromodulators. The DPM-α/β neurons circuit has been shown recently to also modulate egg-laying decision via the AMN neuropeptide. It would be interesting to learn whether neprilysins are involved in this process or if their function is restricted to memory formation (Turrel, 2016).

Despite the importance of the MB for olfactory memory, a functional neurotransmitter or coreleased peptidic neuromodulators produced by MB-intrinsic cells has long remained elusive. It was shown recently that acetylcholine is a Kenyon cell transmitter. The fact that several neprilysins are required for MTM and LTM suggests the involvement of at least one neuropeptide. It remains to be determined whether neprilysin targets are released from the DPM and/or MB and whether identical or distinct neuropeptide substrates support MTM and LTM processes. The sum of the work reported in this study highlights the critical role of the DPM in inactivating and/or processing neuropeptides involved in memory processes connected to the MB (Turrel, 2016).

Go to top

Mushroom body subsets encode CREB2-dependent water-reward long-term memory in Drosophila

Lee, W. P., Chiang, M. H., Chang, L. Y., Lee, J. Y., Tsai, Y. L., Chiu, T. H., Chiang, H. C., Fu, T. F., Wu, T. and Wu, C. L. (2020). PLoS Genet 16(8): e1008963. PubMed ID: 32780743

Long-term memory (LTM) formation depends on the conversed cAMP response element-binding protein (CREB)-dependent gene transcription followed by de novo protein synthesis. Thirsty fruit flies can be trained to associate an odor with water reward to form water-reward LTM (wLTM), which can last for over 24 hours without a significant decline. The role of de novo protein synthesis and CREB-regulated gene expression changes in neural circuits that contribute to wLTM remains unclear. This study shows that acute inhibition of protein synthesis in the mushroom body (MB) αβ or γ neurons during memory formation using a cold-sensitive ribosome-inactivating toxin disrupts wLTM. Furthermore, adult stage-specific expression of dCREB2b in αβ or γ neurons also disrupts wLTM. The MB αβ and γ neurons can be further classified into five different neuronal subsets including αβ core, αβ surface, αβ posterior, γ main, and γ dorsal. This stuyd observed that the neurotransmission from αβ surface and γ dorsal neuron subsets is required for wLTM retrieval, whereas the αβ core, αβ posterior, and γ main are dispensable. Adult stage-specific expression of dCREB2b in αβ surface and γ dorsal neurons inhibits wLTM formation. In vivo calcium imaging revealed that αβ surface and γ dorsal neurons form wLTM traces with different dynamic properties, and these memory traces are abolished by dCREB2b expression. These results suggest that a small population of neurons within the MB circuits support long-term storage of water-reward memory in Drosophila (Lee, 2020).

CREB-dependent gene transcription is critical for memory formation, especially LTM, in both vertebrates and invertebrates. Several previous studies in Drosophila suggest that LTM formation requires CREB-dependent gene transcription followed by de novo protein synthesis. Moreover, it has been reported that CREB2 activity in both αβ and α'β' neurons is critical for appetitive LTM produced by sugar-reward conditioning. This study observed that wLTM formation requires de novo protein synthesis in the αβ and γ neurons. Moreover, adult stage-specific expression of the CREB2 repressor (dCREB2b) in αβ or γ neurons disrupts wLTM, whereas CREB2 activity in α'β' neurons is dispensable. Food or water deprivation is necessary to induce the motivational drive in flies to form sugar- or water-reward LTM since different motivational drives are critical for distinct memories. It has been shown that individual internal motivational inputs for sugar and water are delivered via distinct MB input neurons, which finally induce sugar- or water-reward LTM in different MB neuron subsets. Previous studies together with the current findings suggest that CREB2 activity is required for both sugar- and water-reward LTMs, however, these LTMs are processed in different MB circuits (Lee, 2020).

Neurotransmission from αβ neurons is required for both shock-punitive and sugar-reward LTMs retrieval, whereas neurotransmission from γ neurons is dispensable for retrieval of both shock-punitive and sugar-reward LTMs. However, a previous study showed that the neurotransmission from αβ and γ neurons is required for wLTM retrieval suggesting that wLTM is different from the other types of olfactory associative LTMs at MB circuit levels. The αβ and γ neurons are classified into αβ core, αβ surface, αβ posterior, γ main, and γ dorsal neuron subsets according to the morphology of their axons. It has been shown that neurotransmission from the combination of MB αβ surface and αβ posterior subsets is necessary for the retrieval of both shock-punitive and sugar-rewarded LTMs. Another study suggests that the neurotransmission from αβ surface and αβ core subsets is required for the retrieval of sugar-reward LTM. These results imply that several different subsets of αβ neurons participate in the retrieval of Drosophila sugar-reward LTM. This study showed that neurotransmission only from αβ surface is necessary for wLTM retrieval, whereas the αβ core and αβ posterior subdivisions are dispensable. Contrary to the sugar reward conditioning in which γ neurons are dispensable for the retrieval of sugar-reward LTM, wLTM retrieval requires neurotransmission from γ dorsal but not from γ main neuron subset. Taken together, these results imply that γ dorsal neuronal activity is specifically required for wLTM retrieval but not for sugar-reward LTM. Adult stage-specific expression of dCREB2b or blocking de novo protein synthesis in αβ surface and γ dorsal neurons disrupts wLTM, further suggesting the crucial role of αβ surface and γ dorsal neurons in Drosophila wLTM process (Lee, 2020).

A previous study suggests that PAM-β'1 neurons convey the water-rewarding event as the US signal to the MB β' lobes, and the neurotransmission in α'β' neurons is required for wLTM consolidation. How the wLTM is transferred from α'β' neurons and finally stored in αβ surface and γ dorsal neurons through system consolidation is still unclear. In both shock-punitive and sugar-reward LTMs, the neurotransmission in α'β', γ, and, αβ neurons is required for at least 3 hours after conditioning, but the expression of 24-hour shock-punitive or sugar-rewarded memories only requires neurotransmission in αβ neurons. In addition, the expression of DopR1, a D1-like dopamine receptor, in the γ neurons is sufficient to fully support the shock-punitive STM and LTM in DopR1 mutant background (dumb2),hβ' neurons [41]. Another possibility is that the activity from α'β' neurons is transmitted to αβ surface and γ dorsal neurons via the relevant α'β' MBONs and their downstream neurons. Therefore, it is noteworthy to test the physiological roles of DPM neurons and α'β' MBONs during consolidation phase of wLTM (Lee, 2020).

A significant increase in cellular calcium response to training odor in the αβ surface neurons, but not in other αβ neuron subsets, was observed 24-hour after water-reward conditioning. These results are consistent with the behavioral study showing that neurotransmission from αβ surface neurons is required for wLTM retrieval, whereas the αβ core and αβ posterior neuron subsets are dispensable. This training-induced increased calcium response was abolished in water-sated or dCREB2b expressing flies. A previous study showed that the fly forms α-lobe branch-specific aversive LTM trace 24-hour after odor/shock conditioning. In a recent study, the α-lobe branch-specific aversive anesthesia-resistant memory (ARM) trace 3-hour after odor/shock conditioning was also observed. Intriguingly, it was found that both α- and β- lobes of the surface neurons show increased calcium response to training odor 24-hour after odor/water conditioning, suggesting that the wLTM trace is not specific to the α-lobe branch (Lee, 2020).

An increased GCaMP response to training odor in MB γ neurons is observed at 24-hour after ten sessions of spaced odor/shock training, and this increased calcium response is abolished by expressing dCREB2b in γ neurons throughout the fly development. This study found that γ main neuron subset shows an evoked calcium response to both odors, but no further increased calcium response to the training odor at 24-hour after water-reward conditioning was observed. A previous study suggests that γ dorsal neurons respond to visual stimuli, which is required for visual, but not for aversive olfactory memory in Drosophila. However, this study observed a decreased calcium response to odor stimuli in the γ dorsal lobe, which is consistent with the electrophysiological study showing slow inhibitory responses to odor stimuli in the γ dorsal neurons. Interestingly, a further decrease was observed in calcium responses to the training odor in the γ dorsal neurons 24-hour after water-reward conditioning. This training-induced additional decrease in the calcium response is abolished in the water-sated or acutely dCREB2b expressing flies, suggesting a type of wLTM trace in the γ dorsal neurons different from the αβ surface neurons. Since blocking neurotransmission from the γ dorsal neuron subset during memory retrieval disrupts wLTM, why the γ dorsal neurons show additionally suppressed calcium response to training odor, needs to be answered. One possible explanation is that odor/water association alters the olfactory response of MB neurons to the training odor, and this change can be represented by an increased or decreased calcium response as compared to the response of the non-training odor (memory traces). The training-induced differences in odor responsive levels in the MB allow the flies to distinguish two odors by increasing the contrast and perform appropriate behavioral output during testing. However, shits abolishes the increased or decreased training-odor responses thereby eliminating the contrast between odors, and consequently, the flies could not distinguish two odors and make appropriate behavioral output during testing (Lee, 2020).

In conclusion, this study has shown that αβ surface and γ dorsal neuron subsets regulate Drosophila wLTM. Blocking neurotransmission from αβ surface or γ dorsal neurons only abolishes wLTM retrieval but does not affect the olfactory acuity or water preference in thirsty flies. Further, adult stage-specific expression of dCREB2b or blocking de novo protein synthesis in αβ surface and γ dorsal neurons disrupts wLTM. Different dynamics of cellular wLTM traces are formed in the αβ surface and γ dorsal neurons, which are blocked by dCREB2b expression. Taken together, these results reveal a small population of MB neurons that encode wLTM in the brain and provide a broader view of the olfactory memory process in fruit flies (Lee, 2020).

Go to top

Temporal integration of cholinergic and GABAergic inputs in isolated insect mushroom body neurons exposes pairing-specific signal processing

Raccuglia, D. and Mueller, U. (2014). J Neurosci 34: 16086-16092. PubMed ID: 25429149

GABAergic modulation of neuronal activity plays a crucial role in physiological processes including learning and memory in both insects and mammals. During olfactory learning in honeybees (Apis mellifera) and Drosophila melanogaster the temporal relation between excitatory cholinergic and inhibitory GABAergic inputs critically affects learning. However, the cellular mechanisms of temporal integration of these antagonistic inputs are unknown. To address this question, this study used calcium imaging of isolated honeybee and Drosophila Kenyon cells (KCs), which are targets of cholinergic and GABAergic inputs during olfactory learning. In the whole population of honeybee KCs, pairing of acetylcholine (ACh) and γ-aminobutyric acid (GABA) was found to reduce the ACh-induced calcium influx, and depending on their temporal sequence, induces different forms of neuronal plasticity. After ACh-GABA pairing the calcium influx of a subsequent excitatory stimulus is increased, while GABA-ACh pairing affects the decay time leading to elevated calcium levels during the late phase of a subsequent excitatory stimulus. In an exactly defined subset of Drosophila KCs implicated in learning similar pairing-specific differences were found. Specifically the GABA-ACh pairing splits the KCs in two functional subgroups: one is only weakly inhibited by GABA and shows no neuronal plasticity and the other subgroup is strongly inhibited by GABA and shows elevated calcium levels during the late phase of a subsequent excitatory stimulus. These findings provide evidence that insect KCs are capable of contributing to temporal processing of cholinergic and GABAergic inputs, which provides a neuronal mechanism of the differential temporal role of GABAergic inhibition during learning (Raccuglia, 2014).

Go to top

Debra, a protein mediating lysosomal degradation, is required for long-term memory in Drosophila

Kottler, B., Lampin-Saint-Amaux, A., Comas, D., Preat, T. and Goguel, V. (2011). PLoS One 6: e25902. PubMed ID: 21991383

A central goal of neuroscience is to understand how neural circuits encode memory and guide behavior changes. Many of the molecular mechanisms underlying memory are conserved from flies to mammals, and Drosophila has been used extensively to study memory processes. To identify new genes involved in long-term memory, Drosophila enhancer-trap P(Gal4) lines were screened showing Gal4 expression in the mushroom bodies, a specialized brain structure involved in olfactory memory. This screening led to the isolation of a memory mutant that carries a P-element insertion in the debra locus. debra encodes a protein involved in the Hedgehog signaling pathway as a mediator of protein degradation by the lysosome. To study debra's role in memory, debra overexpression, as well as debra silencing mediated by RNA interference, were achieved. Experiments conducted with a conditional driver that allowed transgene expression to be resticted in the adult mushroom bodies led to a long-term memory defect. Several conclusions can be drawn from these results: (1) debra levels must be precisely regulated to support normal long-term memory, (2) the role of debra in this process is physiological rather than developmental, and (3) debra is specifically required for long-term memory, as it is dispensable for earlier memory phases. Drosophila long-term memory is the only long-lasting memory phase whose formation requires de novo protein synthesis, a process underlying synaptic plasticity. It has been shown in several organisms that regulation of proteins at synapses occurs not only at translation level of but also via protein degradation, acting in remodeling synapses. This work gives further support to a role of protein degradation in long-term memory, and suggests that the lysosome plays a role in this process (Kottler, 2011).

Drosophila melanogaster constitutes a useful model to study the molecular basis underlying memory processes. Its brain, despite its small size, is highly organized and exhibits specialized structures. Furthermore, many of the mechanisms inherent in memory are conserved from flies to mammals. Studies in Drosophila combine the use of powerful genetic tools together with the possibility of analyzing a large repertoire of behaviors. The genetic basis of olfactory learning and memory has been studied for more than 30 years in Drosophila, providing insights into some of the genes involved in short-term and long-term memory formation (Kottler, 2011).

Aversive olfactory memory studies generally rely on classical conditioning of an odor-avoidance response. In this paradigm, groups of flies are successively exposed to two distinct odors, only one of which is accompanied by electric shocks. Memory scores are determined by placing the flies in the center of a T-maze where they are simultaneously exposed to the two odors during one minut. Depending on the training protocol, different types of memory can be measured. Short-term memory (STM) and anaesthesia-resistant memory (ARM) are formed after one cycle of training. STM is a labile memory phase sensitive to cold shock anaesthesia that lasts for a few hours. In contrast, ARM is a consolidated form of memory resistant to cold shock that can last for days. Long-term memory (LTM) is also a form of consolidated memory, but unlike ARM, its formation is sensitive to an inhibitor of cytoplasmic protein synthesis, indicating that de novo protein synthesis is required. LTM is generated after spaced-conditioning consisting of repeated training sessions, each separated by a rest period. LTM is generally thought to occur through changes in synaptic efficacy produced by a restructuring of synapses (Kottler, 2011).

The requirement for de novo gene expression during LTM formation has been widely observed in a number of different model systems. The cAMP response element-binding protein is an LTM-specific regulator of gene expression in Drosophila and in other species. Several other transcription regulators are required for proper LTM including Adf-1 and Stat92E in Drosophila, and CCAAT/enhancer-binding protein, Zif-268, AP-1, and NF-kB in mammals. The Notch signaling receptor has also been implicated in LTM. In addition to transcription, local control of translation, and proteases are as well involved in Drosophila LTM. Crammer, a protein required for LTM, has been shown to inhibit Cathepsin L, a protease that could be involved in lysosome function (Kottler, 2011 and references tgerein).

A large collection of evidence indicates that mushroom bodies (MBs) play a pivotal role in olfactory memory. The MBs form a bilaterally symmetrical structure in the central brain and consist of approximately 4,000 neurons called Kenyon cells. Three types of Kenyon cells (α/β, α'/β', and γ) project their axons ventrally to form the peduncle that splits into five lobes, two vertical (α and α') and three median (β, β', and γ). The lobes are assumed to be the synaptic output region of the MBs. In addition, neurons of the lobes are targeted by multiple inputs (Kottler, 2011).

Many genes required for LTM have been shown to be expressed in the MBs, prompting this study to analyze enhancer-trap P(Gal4) lines showing Gal4 expression in the MBs to characterize new LTM mutants. This report identified debra, a gene involved in protein degradation by the lysosome, as being specifically required for LTM (Kottler, 2011).

An enhancer-trap P(Gal4) inserted nearby the dbr gene lead to Gal4-dependent expression in the MBs, a major center of olfactory memory. The MB247 driver used to affect dbr levels in this study leads to a specific expression in the MB α/β and γ neurons, consistent with additional reports showing that these neurons are involved in aversive olfactory LTM (Kottler, 2011).

Several reports have shown that dbr is involved in various developmental processes. Importantly, the use of conditional silencing in this study reveals that the LTM-specific impairment observed is not caused by a developmental defect, demonstrating that dbr is physiologically involved in LTM processing (Kottler, 2011).

Dbr does not exhibit any obvious homology with known proteins, and its molecular function is unknown. Dbr has been shown to interact with the F-box protein Slimb, an ubiquitin ligase (Dai, 2003). In cooperation with Slimb, Dbr induces the polyubiquitination of phosphorylated Ci-155, a transcription factor that mediates Hedgehog signaling. Interestingly, similar to Dbr, Slimb has been implicated in LTM formation, thus pointing to a role for ubiquitination in LTM processing. These observations are reminiscent of a previous study showing that the highly conserved ubiquitin ligase Neuralized (Neur) is involved in LTM. Neur is expressed in the adult MB α/β neurons and is a limiting factor for LTM formation: loss of one copy of neur gene results in significant LTM impairment whereas Neur overexpression results in a dose-dependent enhancement of LTM. In contrast, both dbr silencing and dbr overexpression in the adult MBs generate a LTM defect, showing that dbr levels must be precisely regulated to support normal LTM, a situation similar to previous reports describing LTM-specific mutants (Kottler, 2011).

Interestingly, dbr is specifically required for LTM since it is dispensable for earlier memory phases. LTM is the only form of memory that relies on de novo protein synthesis, a process thought to underlie synaptic plasticity. Since proteins are the molecular actors that mediate signal transduction, protein synthesis as well as protein degradation must be important for plasticity and memory. Indeed, regulated proteolysis plays a critical role in the remodeling of synapses. Regulated proteolysis is achieved by two major systems in eukaryotic cells: the proteasome and the lysosome. The lysosome degrades most membrane and endocytosed proteins. Owing to their large surface-to-volume ratio, the degradation of membrane proteins such as receptors by the endocytic/lysosomal pathway must be especially efficient and tightly regulated in neurons. Whereas several studies have implicated the proteasome in LTM in Aplysia, in the crab and in mammals, less is known about the implication of the lysosome in this process. It has been suggested that Neur is implicated in both the proteasome and the lysosome degradation pathways. Dbr is involved in protein degradation, and has been characterized as a component of the multivesicular bodies (MVB), an actor of the lysosome pathway (Dai 2003). Ubiquitinated receptors undergo endocytosis and become incorporated into endosomes that are in turn sequestered into MVB. Subsequently, the MVB membrane becomes continuous with lysosomes leading to degradation of the receptor. Although it cannot be ruled out that dbr could be implicated in LTM via another pathway, it is suggested that its function in LTM takes place through the lysosomal protein degradation pathway (Kottler, 2011).

Go to top

Imp promotes axonal remodeling by regulating profilin mRNA during brain development

Medioni, C., Ramialison, M., Ephrussi, A. and Besse, F. (2014). Curr Biol 24(7): 793-800. PubMed ID: 24656828

Neuronal remodeling is essential for the refinement of neuronal circuits in response to developmental cues. Although this process involves pruning or retraction of axonal projections followed by axonal regrowth and branching, how these steps are controlled is poorly understood. Drosophila mushroom body (MB) γ neurons provide a paradigm for the study of neuronal remodeling, as their larval axonal branches are pruned during metamorphosis and re-extend to form adult-specific branches. This study identified the RNA binding protein Imp as a key regulator of axonal remodeling. Imp is the sole fly member of a conserved family of proteins that bind target mRNAs to promote their subcellular targeting. Whereas Imp is dispensable for the initial growth of MB γ neuron axons, it is required for the regrowth and ramification of axonal branches that have undergone pruning. Furthermore, Imp is actively transported to axons undergoing developmental remodeling. Finally, it was demonstrated that profilin mRNA is a direct and functional target of Imp that localizes to axons and controls axonal regrowth. This study reveals that mRNA localization machineries are actively recruited to axons upon remodeling and suggests a role of mRNA transport in developmentally programmed rewiring of neuronal circuits during brain maturation (Medroni, 2014).

In cultured vertebrate neurons, ZBP1 mediates the transport of β-actin mRNA to axons, a process required for the chemiotropic response of growth cones to guidance cues. Whether these observations reflect a general requirement for ZBP1 and axonal mRNA transport during brain development has remained unclear. This study found that Imp, the Drosophila ZBP1 ortholog, accumulates in the cell bodies of a large number of neural cells in adult brain. Strikingly, Imp was additionally observed in the axonal compartment of a subpopulation of mushroom body (MB) neurons. MBs are composed of three main neuronal types (αβ, α'β', and γ) with specific axonal projection patterns and developmental programs. α'β' and αβ neurons are generated during late larval stage and early metamorphosis and are maintained until adulthood. γ neurons are born during late embryogenesis and early larval stages and undergo extensive remodeling during metamorphosis. Imp was found to be enriched in adult γ neuron axons where it colocalized with FasciclinII, but it was not detected in the axons of nonremodeling MB neurons (αβ and α'β' neurons). To test whether Imp is expressed in αβ and α'β' neurons, brains expressing GFP in γ and αβ-core neurons were labelled with antibodies against Imp and Trio, a protein specifically expressed in adult α'β' and γ neurons. Imp was not detected in αβ-core neurons but accumulated in the cell bodies of both α'β' and γ neurons. Thus, both the expression and subcellular distribution of Imp are tightly regulated in Drosophila MB neurons (Medroni, 2014).

To investigate whether Imp axonal translocation is developmentally regulated, the distribution of Imp within γ neurons was examined at different stages. In third-instar larvae, Imp accumulated exclusively in the cell bodies and was not observed in axons. During metamorphosis (pupariation), MB γ neurons first prune the distal part of their axons and then re-extend a medial branch to establish adult-specific projections. Six hours after puparium formation (APF), Imp was weakly detected in γ neuron axons. Such an axonal accumulation of Imp was visible at the time larval γ neurons have completed the pruning of their axonal processes (18 hr APF). During the subsequent intensive growth phase, Imp was enriched at the tip of axons, where it accumulated in particles. Thus, the translocation of Imp to axons is developmentally controlled, and correlates with axonal remodeling (Medroni, 2014).

To test whether Imp is required for γ axon developmental remodeling, the morphology was analyzed of adult homozygous mutant neurons generated using the MARCM (mosaic analysis with a repressible cell marker) system. Clones in which the entire progeny of a neuroblast was mutant exhibited a reduced number of cells and an altered morphology. Although wild-type adult γ axons typically span the entire medial lobe, a mixture of elongated and nonelongated axons was observed upon imp inactivation. To better visualize the morphology of mutant neurons, single labeled neurons were analyzed. Wild-type adult γ neurons extend one main axonal process that reaches the extremity of the MB medial lobe. Several secondary branches typically form along this main axonal process. In contrast, about 50% of imp γ axons failed to reach the end of the medial lobe. These defects did not result from axon retraction, as the proportion of defective axons did not increase with age. Interestingly, mutant axons of normal length but lost directionality were observed, suggesting that imp may be required for the response of γ axons to guidance cues during metamorphosis. imp mutant neurons also exhibited an overall decrease in the complexity of axonal arborization patterns characterized by a reduced number of terminal branches. Both phenotypes were significantly suppressed upon expression of a wild-type copy of Imp in γ neurons, revealing that imp acts cell autonomously to control axonal regrowth and branching (Medroni, 2014).

To determine whether Imp function in axonal growth correlates with its accumulation in axons, the requirement for Imp was investigated in two neuronal cell types where it is exclusively detected in cell bodies: larval γ neurons and α'β' neurons. Both single larval γ neurons and single adult α'β' neurons mutant for imp projected their axons normally. Furthermore, larval γ neuron neuroblast clones exhibited a normal morphology, confirming that imp is not necessary for initial axon growth. These results show that Imp is specifically required for the growth and branching of remodeling γ axons and suggest that its translocation to axons is critical for this function (Medroni, 2014).

To address whether Imp is transported actively to the axons of regrowing γ neurons, a live-imaging protocol was developed using cultured pupal brains expressing functional GFP-Imp fusions specifically in γ neurons). The culture conditions supported efficient axonal growth, as MB neurons from cultured brains grew similarly to their counterparts developing inside the pupa. Fast confocal imaging of axon bundles revealed that GFP-Imp fusions accumulated in particles undergoing bidirectional movement. In contrast, no particles could be detected upon expression of GFP alone. Motile GFP-Imp particles were distributed into three classes: particles with a strong net anterograde (56%) or retrograde (36%) movement and particles with little net bias (8%). Individually tracked particle trajectories were broken into segments to calculate velocities. Segmental velocities distributed over a wide range, with mean anterograde and retrograde segmental velocities of 0.98 ± 0.05 microm/s and 0.73 ± 0.03 microm/s, respectively. Furthermore, curves matching the graph of a quadratic function were obtained upon plotting of the mean square displacement (MSD) values over time, indicating that GFP-Imp particles undergo directed transport rather than diffusion. To assess the role of microtubules (MTs) in this process, brains were treated with colchicine. This treatment abolished MT dynamics, as revealed by the loss of EB1-GFP comets characteristic of growing MT plus ends. Strikingly, motile GFP-Imp particles were no longer observed under these conditions. These results demonstrate that Imp is a component of particles undergoing active MT-dependent transport during midpupariation, consistent with a role of Imp in the transport of selected mRNAs to regrowing γ axons (Medroni, 2014).

Previous in vitro studies have revealed that the axons of immature neurons are enriched in mRNAs encoding regulators of the actin cytoskeleton that play critical roles in axonal growth and guidance. To identify Imp mRNA targets, an immunoprecipitation RT-PCR-based screen was performed for mRNAs encoding actin regulators. Imp was found to selectively associate with chickadee (chic) mRNA, which encodes the G-actin binding protein Profilin. As revealed by affinity pull-down assays, endogenous Imp associated with the chic 3' untranslated region (UTR), but not with the chic coding sequence. To test whether Imp can interact with chic mRNA directly, the binding of recombinant MBP-Imp to the chic 3' UTR was analyzed in electrophoretic mobility shift assays. Retarded complexes formed in the presence of the chic 3' UTR, but not in the presence of a nonrelated RNA (y14). Furthermore, no significant interaction was observed when other MBP-tagged proteins were used. Notably, two discrete complexes were detected in the presence of low amounts of Imp, whereas higher-order complexes were formed with increasing amounts of Imp. Formation of these complexes was outcompeted by the addition of nonlabeled RNAs corresponding to the chic 3' UTR, but not to the chic coding sequence. Altogether, these results show that Imp associates with chic mRNA in vivo and that it can bind directly and specifically to the chic 3' UTR (Medroni, 2014).

To test whether chic mRNA localizes to the neurites of regrowing γ neurons, in situ hybridization was performed on pupal and adult brains. The poor signal-to-noise ratio obtained with this method at these stages, combined with the relatively low levels of axonally localized mRNAs, did not allow chic transcripts or reporters to be unambiguously detected in axons. Thus chic reporter constructs expressed under the control of the γ-specific 201Y-Gal4 driver was used and fluorescent in situ hybridizations was used on dissociated neurons extracted from 24 hr APF pupae and cultured for 3-4 days. chic reporter mRNAs could be observed in the neurites of γ neurons at a significantly higher frequency than control gfp mRNAs. Furthermore, chic mRNA and Imp colocalized in developing neurites, consistent with their association within mRNA transport complexes (Medroni, 2014).

To test whether the region of chic bound by Imp is required for chic mRNA localization to developing neurites, the distribution of reporters containing both the chic coding sequence and 3' UTR was compared with that of reporters lacking the chic 3' UTR. Transcripts with the chic 3' UTR localized more efficiently than transcripts lacking it, suggesting that Imp binding to the 3' UTR promotes chic axonal targeting. To exclude an effect of Imp on chic mRNA stability, the levels of chic transcripts were analyzed in cultured S2R+ cells. No significant differences in chic mRNA and Chic protein levels could be observed upon imp inactivation in these conditions (Medroni, 2014).

To functionally test the importance of chic regulation in vivo, the phenotypes associated with chic downregulation were examined. Consistent with described roles of Profilin in regulating F-actin polymerization and axonal pathfinding, it was observed that chic mutant γ neurons fail to properly extend their axons. More importantly, overexpression of chic significantly rescued the axonal growth defects observed in imp mutant neurons. Similar results were obtained with two independent UAS-chic transgenes, but not with overexpression of another regulator of F-actin polymerization (enabled). These results suggest that imp controls axonal remodeling by regulating chic expression in vivo and reveal that forced accumulation of Chic protein in axons can partially compensate for the loss of imp function (Medroni, 2014).

In conclusion, the finding that Drosophila Imp is required for γ axon regrowth but is dispensable for initial axonal growth suggests a novel and specific function of axonal mRNA targeting in developmental remodeling of the brain. Furthermore, these results highlight mechanistic similarities between developmental axonal regrowth and postinjury axonal regeneration, a process known to depend on axonal mRNA transport. Finally, this study uncovers that the translocation of Imp to γ axons is tightly linked to their developmental remodeling program. This reveals that mRNA transport machineries are subject to precise spatiotemporal regulation and may be specifically recruited in the context of developmental rewiring of the brain. It will now be interesting to identify the signals controlling the localization and the activity of mRNA transport machineries during this process (Medroni, 2014).

Go to top

Axonal chemokine-like orion induces astrocyte infiltration and engulfment during mushroom body neuronal remodeling

Boulanger, A., Thinat, C., Zuchner, S., Fradkin, L. G., Lortat-Jacob, H. and Dura, J. M. (2021). Nat Commun 12(1): 1849. PubMed ID: 33758182

The remodeling of neurons is a conserved fundamental mechanism underlying nervous system maturation and function. Astrocytes can clear neuronal debris and they have an active role in neuronal remodeling. Developmental axon pruning of Drosophila memory center neurons occurs via a degenerative process mediated by infiltrating astrocytes. However, how astrocytes are recruited to the axons during brain development is unclear. Using an unbiased screen, the gene requirement of orion/CG2206, encoding for a chemokine-like protein, was identified in the developing mushroom bodies. Functional analysis shows that Orion is necessary for both axonal pruning and removal of axonal debris. Orion performs its functions extracellularly and bears some features common to chemokines, a family of chemoattractant cytokines. It is proposed that Orion is a neuronal signal that elicits astrocyte infiltration and astrocyte-driven axonal engulfment required during neuronal remodeling in the Drosophila developing brain (Boulanger, 2021).

Neuronal remodeling is a widely used developmental mechanism, across the animal kingdom, to refine dendrite and axon targeting necessary for the maturation of neural circuits. Importantly, similar molecular and cellular events can occur during neurodevelopmental disorders or after nervous system injury. A key role for glial cells in synaptic pruning and critical signaling pathways between glia and neurons has been identified. In Drosophila, the mushroom body (MB), a brain memory center, is remodeled at metamorphosis and MB γ neuron pruning occurs by a degenerative mechanism. Astrocytes surrounding the MB have an active role in the process: blocking their infiltration into the MBs prevents remodeling. MB γ neuron remodeling relies on two processes: axon fragmentation and the subsequent clearance of axonal debris. Importantly, it has been shown that astrocytes are involved in these two processes and that these two processes can be decoupled. Altering the ecdysone signaling in astrocytes, during metamorphosis, results both in a partial axon pruning defect, visualized as either some individual larval axons or as thin bundles of intact larval axons remaining in the adults, and also in a strong defect in clearance of debris, visualized by the presence of clusters of axonal debris. Astrocytes have only a minor role in axon severing as evidenced by the observation that most of the MB γ axons are correctly pruned when ecdysone signaling is altered in these cells. When astrocyte function is blocked, the γ axon-intrinsic fragmentation process remains functional and the majority of axons degenerate (Boulanger, 2021).

It has been widely proposed that a 'find-me/eat-me' signal emanating from the degenerating γ neurons is necessary for astrocyte infiltration and engulfment of the degenerated larval axons. However, the nature of this glial recruitment signal is unclear (Boulanger, 2021).

This study has identified a gene (orion), not previously described, by screening for viable ethyl methanesulfonate (EMS)-induced mutations and not for lethal mutations in MB clones as was done previously. This allowed the identification of genes involved in glial cell function by directly screening for defects in MB axon pruning. It was found that orion1, a viable X-chromosome mutation, is necessary for both the pruning of some γ axons and removal of the resulting debris. orion is secreted from the neurons, remains near the axon membranes where it associates with infiltrating astrocytes, and is necessary for astrocyte infiltration into the γ bundle. This implies a role for an as-yet-undefined orion receptor on the surface of the astrocytes. orion bears some chemokine features, for example, a CX3C motif, three glycosaminoglycan (GAG) binding consensus sequences that are required for its function. Altogether, these results identify a neuron-secreted extracellular messenger, which is likely to be the long-searched-for signal responsible for astrocyte infiltration and engulfment of the degenerated larval axons and demonstrate its involvement for neuronal remodeling (Boulanger, 2021).

Adult orion1 individuals showed a clear and highly penetrant MB axon pruning phenotype as revealed by the presence of some adult unpruned vertical γ axons as well as the strong presence of debris (100% of mutant MBs). Astrocytes, visualized with alrm-GAL4, are the major glial subtype responsible for the clearance of the MB axon debris. The presence of γ axon debris is a landmark of defective astrocyte function, as has been described, and is also further shown in this study. The unpruned axon phenotype was particularly apparent during metamorphosis. At 24 h after puparium formation (APF), although γ axon branches were nearly completely absent in the wild-type control, they persisted in the orion1 mutant brains, where a significant accumulation of debris was also observed. The number of unpruned axons at this stage is lower in orion1 than in Hr39C13 where the γ axon-intrinsic process of pruning is blocked. In addition, the MB dendrite pruning was clearly affected in orion1 individuals (Boulanger, 2021).

The orion1 EMS mutation was localized by standard duplication and deficiency mapping as well as by whole-genome sequencing. The orion gene (CG2206) encodes two putatively secreted proteins: Orion-A [664 amino acid (a.a.)] and Orion-B (646 a.a.), whose messenger RNAs (mRNAs) arise from two different promoters. These two proteins differ in their N-terminal domains and are identical in the remainder of their sequences. The EMS mutation is a G to C nucleotide change inducing the substitution of the glycine (at position 629 for Orion-A and 611 for Orion-B) into an aspartic acid. The mutation lies in the common shared part and therefore affects both Orion-A and -B functions. Both isoforms display a signal peptide at their N termini, suggesting that they are secreted. Interestingly, a CX3C chemokine signature is present in the orion common region. Chemokines are a family of chemoattractant cytokines, characterized by a CC, CXC, or CX3C motif, promoting the directional migration of cells within different tissues. Mammalian CX3CL1 (also known as fractalkine) is involved in, among other contexts, neuron-glia communication. Mammalian Fractalkines display conserved intramolecular disulfide bonds that appear to be conserved with respect to their distance from the CX3C motif present in both orion isoforms. Fractalkine and its receptor, CX3CR1, have been recently shown to be required for post-trauma cortical brain neuron microglia-mediated remodeling in a mouse whisker lesioning paradigm. This study observed that the change of the CX3C motif into CX4C or AX3C blocked the orion function necessary for the MB pruning. Similarly, the removal of the signal peptide also prevented pruning. These two results indicate that the orion isoforms likely act as secreted chemokine-like molecules. Three CRISPR/Cas9-mediated mutations in the orion gene were generated that either delete the common part (orionΔC), the A-specific part (orionΔA), or the B-specific part (orionΔB). Noticeably, orionΔC displayed the same MB pruning phenotype as orion1, which is also the same in orion1/Deficiency females, indicating that orion1 and orionΔC are likely null alleles for this phenotype. In contrast, orionΔA and orionΔB have no MB phenotype by themselves, indicating the likelihood of functional redundancy between the two proteins in the pruning process (Boulanger, 2021).

Using the GAL4/UAS system, this study found that expression of wild-type orion in the orion1 MB γ neurons (201Y-GAL4) fully rescued the MB mutant phenotype (100% of wild-type MBs n = 387), although wild-type orion expression in the astrocytes (alrm-GAL4) did not rescue. repo-GAL4 could not be used because of lethality when combined with UAS-orion. This supports the hypothesis that orion is produced by axons and, although necessary for astrocyte infiltration, not by astrocytes. Both UAS-orion-A and UAS-orion-B rescued the orion1 pruning phenotype indicating again a likely functional redundancy between the two proteins at least in the pruning process. Complementary to the rescue results, this study found that the expression of an orion-targeting RNA interference (RNAi) in the MBs produced unpruned axons similar to that in orion1, although the debris is not apparent likely due to an incomplete inactivation of the gene expression by the RNAi. The expression of the same RNAi in the glia had no effect. Using the mosaic analysis with a repressible cell marker (MARCM), it was found that orion1 homozygous mutant neuroblast clones of γ neurons, in orion1/+ phenotypically wild-type individuals, were normally pruned. Therefore, orion1 is a non-cell-autonomous mutation that is expected if the orion proteins are secreted. orion proteins secreted by the surrounding wild-type axons rescue the pruning defects in the orion mutant clones (Boulanger, 2021).

From genetic data, orion expression is expected in the γ neurons. The fine temporal transcriptional landscape of MB γ neurons was recently described and a corresponding resource is freely accessible. Noteworthy, orion is transcribed at 0 h APF and dramatically decreases at 9 h APF with a peak at 3 h APF. The nuclear receptor EcR-B1 and its target Sox14 are key transcriptional factors required for MB neuronal remodeling. orion was found to be a likely transcriptional target of EcR-B1 and Sox14. This is also consistent with earlier microarray analysis observations. Noticeably, forced expression of UAS-EcR-B1 in the MBs did not rescue the orion mutant phenotype and EcR-B1 expression, in the MB nuclei, and is not altered in orion1 individuals. Furthermore, the unpruned axon phenotype produced by orion RNAi is rescued by forced expression of EcR-B1 in the MBs. Therefore, the genetic interaction analyses support orion being downstream of EcR-B1 (Boulanger, 2021).

Further molecular and cellular work focused on Orion-B alone since a functional redundancy between the two isoforms was apparent. The Orion-B protein was expressed in the γ neurons using an UAS-orion-B-Myc insert and the 201Y-GAL4 driver. Orion-B was present along the MB lobes and extracellularly present as visualized by anti-Myc staining. Indeed, anti-Myc staining was particularly strong at the tip of the lobes indicating the presence of extracellular Orion-B. Synaptic terminals are condensed in the γ axon varicosities that disappear progressively during remodeling and hole-like structures corresponding to the vestiges of disappeared varicosities can be observed at 6 h APF. The presence of Myc-labeled Orion-B was noticed inside these hole-like structures. The secretion of the Orion proteins should be under the control of their signal peptide and, therefore, Orion proteins lacking their signal peptide (ΔSP) should not show this 'extracellular' phenotype. When UAS-orion-B-Myc-ΔSP was expressed, Orion-B was not observed outside the axons or in the hole-like structures. The possibility that this 'extracellular' phenotype was due to some peculiarities of the Myc labeling was excluded by using a UAS-drl-Myc construct. Drl is a membrane-bound receptor tyrosine kinase and Drl-Myc staining, unlike Orion-B, was not observed outside the axons or in the hole-like structures. In addition, the presence of Myc-labeled Orion-B protein not associated with green fluorescent protein (GFP)-labeled axon membranes can be observed outside the γ axon bundle in 3D reconstructing images. Nevertheless, these signals are possibly located inside the glial compartments and not as freely diffusing orion protein. Finally, supporting the hypothesis that orion acts as a secreted protein, it has been reported to be present in biochemically purified exosomes, indicating that it may act on the glia via its presence on or in exosomes (Boulanger, 2021).

Since glial cells are likely directly involved in the orion1 pruning phenotype, their behavior early during the pruning process was examined. At 6 h APF the axon pruning process starts and is complete by 24 h APF, but the presence of glial cells in the vicinity of the wild-type γ lobes is already clearly apparent at 6 h APF9. Glial cells visualized by a membrane-targeted GFP (UAS-mGFP) under the control of repo-GAL4 were examined, and the γ axons were co-stained with anti-Fas2. At 6 h APF, a striking difference was noted between wild-type and orion1 brains. Unlike in the wild-type control, there is essentially no glial cell invasion of the γ bundle in the mutant. Interestingly, glial infiltration as well as engulfment of the degenerated larval axons was not observed in orion1 neither at 12 h APF nor at 24 h APF, suggesting that glial cells never infiltrate MBs in mutant individuals. The possibility that this lack of glial cell activity was due to a lower number of astrocytes in mutant versus wild-type brains was ruled out (Boulanger, 2021).

The proximity was examined between MB Orion-Myc and astrocytes, as inferred from the shape of the glial cells, labeled with the anti-Drpr antibody at 6 h APF. The distribution was examined along the vertical γ lobes (60 μm of distance) of Orion-B-Myc (wild-type protein) and of Orion-B-ΔSP-Myc (not secreted), in an otherwise wild-type background. Quantification was performed only from images where an astrocyte sat on the top of the vertical lobe. A peak of Orion-Myc localization was always found in the axonal region close to the astrocyte (<7 μm) when wild-type Orion-B-Myc was quantified. However, this was not the case (n = 9) when Orion-B-ΔSP-Myc was quantified. This strongly suggests that astrocytic processes may be 'attracted' by secreted orion (Boulanger, 2021).

Moreover, it was observed that extracellularly present orion stays close to axon membranes. Protein (in particular chemokine) localization to membranes is often mediated by GAGs, a family of highly anionic polysaccharides that occurs both at the cell surface and within the extracellular matrix. GAGs, to which all chemokines bind, ensure that these signaling proteins are presented at the correct site and time in order to mediate their functions. Three consensus sequences for GAG linkage were identified in the common part of orion. These sequences were mutated individually, and the mutant proteins were examined for their ability to rescue the orion1 pruning deficit in vivo. The three GAG sites are required for full orion function, although mutating only GAG3 produced a strong mutant phenotype (Boulanger, 2021).

These findings imply a role for an as-yet-undefined orion receptor on the surface of the glial cells. The glial receptor draper (drpr) seemed an obvious candidate, although Drpr ligands unrelated to orion have been identified. The MB remodeling phenotypes in orion1 and drprΔ5 are, however, different with orion mutant phenotype being stronger than the drpr one. The use of an UAS-mGFP driven by 201Y-GAL4, instead of anti-Fas2, where the labeling of αβ axons often masks individual unpruned γ axons, allowed occasionally observation of unpruned axons in drprΔ5 1-week-old post-eclosion brains in addition to uncleared debris. This indicates a certain degree of previously undescribed unpruned axon persistence in the mutant background. Nevertheless, only orion mutant displayed a 100% penetrant phenotype of both unpruned axons and debris (strong category) in adult flies, which are still present in old flies. On the contrary, the weaker drpr mutant phenotype strongly decreases throughout adulthood. This suggests that Drpr is not an, or at least not the sole, orion receptor (Boulanger, 2021).

Independently of the possible role of Drpr as an orion receptor, it was of interest to test if orion could activate the drpr signaling pathway as it is the case for neuron-derived injury released factors and Spätzle ligands, which bind to glial insulin-like receptors and Toll-6, respectively, upregulating in turn the expression of drpr in phagocytic glia. These ligands are necessary for axonal debris elimination and act as a find-me/eat-me signal in injury and apoptosis, as orion is doing for MB pruning. The data indicate that orion does not modify either Drpr expression nor the level of the drpr transcriptional activator STAT92E in astrocytes. Consequently, orion does not seem to induce the Drpr signaling pathway in astrocytes (Boulanger, 2021).

This study has uncovered a neuronally secreted chemokine-like protein acting as a 'find-me/eat-me' signal involved in the neuron-glia crosstalk required for axon pruning during developmental neuron remodeling. Chemokine-like signaling in insects was not described previously and, furthermore, the results point to an unexpected conservation of chemokine CX3C signaling in the modulation of neural circuits. Thus, it is possible that chemokine involvement in neuron/glial cell interaction is an evolutionarily ancient mechanism (Boulanger, 2021).

Go to top

Combining developmental and perturbation-seq uncovers transcriptional nodules orchestrating neuronal remodeling

Alyagor, I., Berkun, V., Keren-Shaul, H., Marmor-Kollet, N., David, E., Mayseless, O., Issman-Zecharya, N., Amit, I. and Schuldiner, O. (2018). Dev Cell 47(1): 38-52. PubMed ID: 30300589

Developmental neuronal remodeling is an evolutionarily conserved mechanism required for precise wiring of nervous systems. Despite its fundamental role in neurodevelopment and proposed contribution to various neuropsychiatric disorders, the underlying mechanisms are largely unknown. This study uncovered the fine temporal transcriptional landscape of Drosophila mushroom body gamma neurons undergoing stereotypical remodeling (see Hierarchical TF Networks Regulate Axon Pruning). The data reveal rapid and dramatic changes in the transcriptional landscape during development. Focusing on DNA binding proteins, eleven were identified that are required for remodeling. Furthermore, developing gamma neurons perturbed for three key transcription factors required for pruning were sequenced. A hierarchical network is described featuring positive and negative feedback loops. Superimposing the perturbation-seq on the developmental expression atlas highlights a framework of transcriptional modules that together drive remodeling. Overall, this study provides a broad and detailed molecular insight into the complex regulatory dynamics of developmental remodeling and thus offers a pipeline to dissect developmental processes via RNA profiling (Alyagor, 2018).

Go to top

Drosophila Hrp48 is required for mushroom body axon growth, branching and guidance

Bruckert, H., Marchetti, G., Ramialison, M. and Besse, F. (2015). PLoS One 10: e0136610. PubMed ID: 26313745

RNA binding proteins assemble on mRNAs to control every single step of their life cycle, from nuclear splicing to cytoplasmic localization, stabilization or translation. This study investigated the role of Drosophila Hrp48, a fly homologue of mammalian hnRNP A2/B1, during central nervous system development. Using a combination of mutant conditions, hrp48 was shown to be required for the formation, growth and guidance of axonal branches in Mushroom Body neurons. Furthermore, hrp48 inactivation induces an overextension of Mushroom Body dorsal axonal branches, with a significantly higher penetrance in females than in males. Finally, as demonstrated by immunolocalization studies, Hrp48 is confined to Mushroom Body neuron cell bodies, where it accumulates in the cytoplasm from larval stages to adulthood. Altogether, these data provide evidence for a crucial in vivo role of the hnRNP Hrp48 in multiple aspects of axon guidance and branching during nervous system development. They also indicate cryptic sex differences in the development of sexually non-dimorphic neuronal structures (Bruckert, 2015).

Go to top

The formin DAAM functions as molecular effector of the planar cell polarity pathway during axonal development in Drosophila

Gombos, R., Migh, E., Antal, O., Mukherjee, A., Jenny, A. and Mihaly, J. (2015). J Neurosci 35: 10154-10167. PubMed ID: 26180192

Recent studies established that the planar cell polarity (PCP) pathway is critical for various aspects of nervous system development and function, including axonal guidance. Although it seems clear that PCP signaling regulates actin dynamics, the mechanisms through which this occurs remain elusive. This study established a functional link between the PCP system and one specific actin regulator, the formin DAAM, which has previously been shown to be required for embryonic axonal morphogenesis and filopodia formation in the growth cone. DAAM also plays a pivotal role during axonal growth and guidance in the adult Drosophila mushroom body, a brain center for learning and memory. By using a combination of genetic and biochemical assays, it was demonstrated that Wnt5 and the PCP signaling proteins Frizzled, Strabismus, and Dishevelled act in concert with the small GTPase Rac1 to activate the actin assembly functions of DAAM essential for correct targeting of mushroom body axons. Collectively, these data suggest that DAAM is used as a major molecular effector of the PCP guidance pathway. By uncovering a signaling system from the Wnt5 guidance cue to an actin assembly factor, it is proposed that the Wnt5/PCP navigation system is linked by DAAM to the regulation of the growth cone actin cytoskeleton, and thereby growth cone behavior, in a direct way (Gombos, 2015).

This study has shown that DAAM plays an important role in the regulation of axonal growth and guidance of the Drosophila MB neurons. Several lines of evidence suggest that DAAM acts in concert with Wnt5 and the core PCP proteins to ensure correct targeting of the KC axons. DAAM functions downstream of Dsh and Rac1, and its ability to promote actin assembly is absolutely required for neural development in the MB. These data suggest a simple model in which axon guidance cues, such as Wnt5, signal through the PCP pathway to activate DAAM to control actin filament formation in the neuronal growth cone. Thus, PCP signaling appears to be linked to cytoskeleton regulation in a direct way, and these results provide compelling experimental evidence suggesting that, at least in neuronal cells, the major cellular target of PCP signaling is the actin cytoskeleton (Gombos, 2015).

Formins are highly potent actin assembly factors that are under tight regulation in vivo. The major mechanism of controlling the activity of the Diaphanous-related formin (DRF) subfamily involves an intramolecular autoinhibitory interaction between the N-terminal diaphanous inhibitory domain (DID) and the C-terminal Diaphanous autoinhibitory domain (DAD). This inhibition can be relieved upon binding of an activated Rho family GTPase that interacts with the GBD (GTP-ase binding domain)/DID region and also by proteins that bind to the DAD domain. Consistently, this study found that the Rac1 GTPase and the DAD domain binding Dsh protein both play role in DAAM activation in MB neurons. With this regard, it is notable that, despite that dsh1 is considered a PCP-null allele, the DAAMEx1, dsh1 double hemizygous mutants exhibit a stronger MB phenotype than dsh1 mutants alone, suggesting that DAAM must receive Dsh-independent regulatory inputs for which Rac1 is a prime candidate. Although previous work indicated that Rho GTPases might function downstream of Dsh in a linear pathway), the data suggest that Dsh and Rac1 act in parallel pathways in the MB. As the impairment of GTPase binding severely, but not completely, abolishes DAAM activity, it is concluded that Rac1 is likely to have a stronger contribution to DAAM activation in vivo; nonetheless, the simultaneous binding of Dsh appears to be required for full activation (Gombos, 2015).

Presumably, the most remarkable feature of the PCP system relies in its ability to create subcellular asymmetries. Therefore, it is a tempting idea that, upon guidance signaling, the PCP proteins are involved in the generation of molecular asymmetries within axonal growth cones, yet recent attempts failed to reveal such polarized distributions in MB neurons. Interestingly, however, it was shown that Fz and Vang display a differential requirement during development of the MBs, with Fz predominantly acting in the dorsal lobes and Vang predominantly acting in the medial lobes (MLs). This study found that, in contrast to Fz and Vang, DAAM plays a crucial role in both lobes of the MBs. Additionally, it was demonstrated that Fz promotes the formation of membrane-associated Dsh-DAAM complexes in S2 cells. This result, together with genetic data, suggests that DAAM acts as the downstream effector of a Fz/Dsh module, which is required for the correct growth and guidance of the dorsal MB axon branches (Gombos, 2015).

In addition to their potential connection to Fz signaling in the dorsal lobe, DAAM and Dsh were linked to Vang- and Wnt5-dependent ML development as well. Wnt5 and Vang have an identical effect on ML development when overexpressed, and this GOF phenotype can be suppressed by the same set of mutations (DAAM, dsh, Rac1). In particular, the putative PCP-null dsh1 allele and heterozygosity for Rac1 cause an almost equally strong, yet partial, suppression with regard to the ML fusion phenotype. This is best explained by assuming that Wnt5 and Vang signal both in a Dsh-dependent and in a Dsh-independent, but Rac-dependent, manner. With regard to DAAM, this study has shown that DAAM nearly completely suppresses the GOF of Wnt5 and Vang, and Dsh and Rac1 both contribute to DAAM activation. Collectively, these data suggest a model in which Wnt5 and Vang promote β lobe extension by signaling to Dsh and Rac1 that will activate DAAM in parallel to each other. The colocalization of Vang and DAAM, observed in S2 cells, indicates that they may bind each other directly, which would be in good accordance with genetic data suggesting a close functional link between DAAM and Vang during β lobe development. However, formins are not known to bind Vang proteins; therefore, an indirect interaction, mediated by Rac1, which has recently been shown to be bound and redistributed by Vangl2 in epithelial cell lines, appears a more likely possibility (Gombos, 2015).

As discussed above, and contrary to Vang, Fz does not appear to be required for ML development, or if anything, it might play an opposite role, as loss of fz leads to ML fusion in 16.1% of the lobes. This is a surprising observation at first glance as Wnt proteins are thought to activate members of the Fz receptor family, but former analysis of Wnt5 signaling during MB development also failed to reveal a Fz requirement in the β lobes. Instead, Wnt5 has been linked to other type of Wnt receptors, the Ryk/Derailed atypical tyrosine kinase receptors, which are known to be involved in axonal guidance in flies and vertebrates. In light of these results, it will be of future interest to analyze the Wnt5-Vang connection in the MB in more details and identify the Wnt5 receptor in this context (Gombos, 2015).

Consistent with the lack of lobe-specific requirement for dsh and DAAM, the current studies revealed that Dsh, DAAM, and Rac1 are used as common effector elements of a dorsal lobe-specific Fz-dependent signal and a Vang-dependent ML-specific signal. It follows that Dsh and DAAM are likely to take part in two types of PCP complexes. Although, in vitro, Dsh has the ability to interact with both Fz and Vang, the conclusion that Dsh functions downstream of Vang in the β lobes is markedly different from the classical PCP regulatory context in which the Fz/Dsh and Vang/Pk complexes have opposing effects. Thus, this result, together with the Wnt5-Vang data, substantiates the earlier findings that the PCP system operates at least partly differently in neurons than during tissue polarity signaling (Gombos, 2015).

During PCP signaling, the vertebrate DAAM orthologs control convergence and extension movements, polarized cell movements during vertebrate gastrulatio. In contrast, DAAM is dispensable for classical planar polarity establishment in flies, suggesting that the tissue polarity function of DAAM might be restricted only to vertebrates. Despite the lack of direct function in establishing tissue polarization, this study provides evidence that DAAM is linked to the PCP pathway in another important regulatory context, notably directed neuronal development in the adult brain. Consistent with the results, recent studies revealed that PCP signaling and DAAM regulate neural development in planarians and in Xenopus embryos. Given that the vertebrate PCP proteins are known to be involved in multiple aspects of CNS development, and the vertebrate DAAM orthologs are strongly expressed in the CNS, it is conceivable that the PCP/DAAM module represents a highly conserved regulatory system that is used to regulate various aspects of neuronal development throughout evolution (Gombos, 2015).

Go to top

Nuclear Transcriptomes of the Seven Neuronal Cell Types That Constitute the Drosophila Mushroom Bodies

Shih, M. M., Davis, F. P., Henry, G. L. and Dubnau, J. (2018). G3 (Bethesda) 9(1):81-94. PubMed ID: 30397017

The insect mushroom body (MB) is a conserved brain structure that plays key roles in a diverse array of behaviors. The Drosophila melanogaster MB is the primary invertebrate model of neural circuits related to memory formation and storage, and its development, morphology, wiring, and function has been extensively studied. MBs consist of intrinsic Kenyon Cells that are divided into three major neuron classes (the γ, α'/β' and α/β) and 7 cell subtypes (γd, γm, α''/β'ap, α''/β'm, α'/βp, α'/βs and α/βc) based on their birth order, morphology, and connectivity. These subtypes play distinct roles in memory processing, however the underlying transcriptional differences are unknown. This study used RNA sequencing (RNA-seq) to profile the nuclear transcriptomes of each MB neuronal cell subtypes. 350 MB class- or subtype-specific genes were identified, including the widely used α/β class marker Fas2 and the α'/β' class marker trio. Immunostaining corroborates the RNA-seq measurements at the protein level for several cases. Importantly, the data provide a full accounting of the neurotransmitter receptors, transporters, neurotransmitter biosynthetic enzymes, neuropeptides, and neuropeptide receptors expressed within each of these cell types. This high-quality, cell type-level transcriptome catalog for the Drosophila MB provides a valuable resource for the fly neuroscience community (Shih, 2018).

The results establish a high-quality, neuronal cell type-level transcriptome for Drosophila MB. 350 differentially expressed genes were identified that includes most of the previously reported MB lobe (class specific) markers and many novel class-specific or cell subtype-specific profiles of expression. In addition to the subtype level resolution of the experimental design, the TAPIN approach that was used also offers several advantages and technical differences with these prior approaches. First, because TAPIN is compatible with flash frozen tissue as the input, the method introduces minimal disturbance to the endogenous transcriptome as compared to more lengthy procedures for purification of neurons for expression profiling. Second, it may be relevant that TAPIN explicitly profiles nuclear RNAs, likely enriching for actively transcribed/nascent transcripts vs. abundant ones that are stably maintained in the cytoplasm. Thus, it would be attractive to apply this method to profile transcriptional response to behavioral perturbations (Shih, 2018).

Several previous studies have used genome-wide methods to profile expression in the Drosophila MB. One study used a microarray-based approach to profile expression of each of the three major classes of MB KCs and compared these profiles with expression in the rest of the brain. That study first focused on the expression of transposons, and subsequently used the same transcriptome dataset to discover that MB KCs are cholinergic based on expression of biosynthetic enzymes. Another used an RNA-seq-based approach to profile expression in relatively small pools of physically isolated α/β and γ class neurons to search for memory-related changes in gene expression. Other studies have used droplet-based single cell sequencing to profile the Drosophila brain, and by clustering the single cells they were able to identify the three MB classes, but not the further sub-division into neuronal subtypes (Shih, 2018).

Although this study used a different profiling method and resolved transcriptomes at the cell subtype rather than class level, the findings are broadly compatible with prior reports. The current dataset reveals strong expression of both ChAT and VAChT, consistent with the conclusion that MBs are cholinergic. These findings further support the conclusion that all of the individual MB KC subtypes are cholinergic. The datasets also are consistent in the expression of known class-specific markers. One notable difference is that a previous study reported high levels of expression of the 5-HT1B receptor in both α/β and γ classes of KCs, and another study also observed 5-HT1B expression in α/β and γ KCs single cell clusters. In contrast, this study saw no evidence for expression of this receptor in TAPIN-seq profiles. This difference could reflect methodology: the previous studies measured 5-HT1B receptor transcripts in the cytoplasm while this study measured the levels that are actively transcribed or present in the nucleus. This technical difference could be especially relevant for neurotransmitter receptors, some of which can be translated locally at dendrites (Shih, 2018).

The current dataset is the first to profile expression in this brain region at neuronal subtype resolution. This level of resolution is critical given the wealth of data on the functional differences of each MB KC subtype in Drosophila behaviors. Tbis dataset provides a full accounting within each of the MB KC subtypes of the profiles of expression of the cellular machinery to produce and receive neurotransmission, including small molecule transmitters and their receptors, neuropeptides and neuropeptide receptors and subunits of gap junctions. It is noteworthy that the TAPIN expression dataset supports the conclusions that all the adult MB KC subtypes are cholinergic, and that none of the subtypes express genes that would suggest the co-release of GABA, dopamine, glutamate, or serotonin. On the other hand, expression was detected of a spectrum of neuropeptides and their receptors. This observation is consistent with the hypothesis that MB KCs may co-release both acetylcholine and several neuropeptides (Shih, 2018).

In addition to these findings with regards to the inputs and outputs, this study identified 350 differentially expressed genes including many that distinguish MB KC classes or even individual cell subtypes. MB α/β' subtype showed 21 enriched genes and 11 depleted genes, contrasting with two other subtypes in the α/β class and two other classes. This uniqueness is supported by its unique odor responses and connectivity. Despite the limitation in the methodology, this study still identified distinct sets of enriched/depleted genes, indicating the differences between two subtypes in MB γ or α'/β' classes (Shih, 2018).

This dataset provides a valuable resource for the fly neuroscience community to conduct functional studies. For example, the data provide a list of previously unknown class specific and sub-type specific transcripts, whose impact on the functional differences between these neurons are not known. An arsenal of genetic tools to manipulate any gene's function within each of these cell subtypes already exists. In addition to olfactory associative memory, MBs also play fundamental roles in other forms of memory including visual and gustatory, temperature preference, courtship behaviors, stress response, food-seeking, sleep and responses to ethanol. This dataset will facilitate the discovery of neural mechanisms for each of these conserved behaviors (Shih, 2018).

Go to top

Autophagy within the mushroom body protects from synapse aging in a non-cell autonomous manner

Bhukel, A., Beuschel, C. B., Maglione, M., Lehmann, M., Juhasz, G., Madeo, F. and Sigrist, S. J. (2019). Nat Commun 10(1): 1318. PubMed ID: 30899013

Macroautophagy is an evolutionarily conserved cellular maintenance program, meant to protect the brain from premature aging and neurodegeneration. How neuronal autophagy, usually loosing efficacy with age, intersects with neuronal processes mediating brain maintenance remains to be explored. This study shows that impairing autophagy in the Drosophila learning center (mushroom body, MB) but not in other brain regions triggered changes normally restricted to aged brains: impaired associative olfactory memory as well as a brain-wide ultrastructural increase of presynaptic active zones (metaplasticity), a state non-compatible with memory formation. Mechanistically, decreasing autophagy within the MBs reduced expression of an NPY-family neuropeptide, and interfering with autocrine NPY signaling of the MBs provoked similar brain-wide metaplastic changes. The results in an exemplary fashion show that autophagy-regulated signaling emanating from a higher brain integration center can execute high-level control over other brain regions to steer life-strategy decisions such as whether or not to form memories (Bhukel, 2019).

The maintenance of neuronal homeostasis is severely threatened by aging. The strictly postnatal character of deficits observed after KD of core autophagy machinery triggered the hope that autophagy might have a specific relation to the aging process. The last few years have indeed seen an accumulation of evidences that the efficiency of autophagic clearance in neurons declines with age on organismal level. Hence, rejuvenating autophagy in aging neurons is considered a promising strategy to restore cognitive performance. Successfully exploring this direction will, however, depend on deepening insights at the intersection of autophagy, the relevant neuronal sub-cellular compartments, importantly synaptic specializations, and relevant neuron populations/brain regions (Bhukel, 2019).

The endogenous polyamine spermidine has prominent cardio-protective and neuro-protective effects and recent work finds spermidine restoration to counteract otherwise deteriorating health in aging mice in an autophagy-dependent manner. In Drosophila, restoring spermidine specifically suppressed age-induced decay in their ability to form olfactory memories, again in an autophagy-dependent manner. Concomitantly, in the aged Drosophila brain, previous work found a brain-wide, age-induced upshift in the ultrastructural size (EM: larger T-bars; STED: increased diameter of BRP scaffold) of presynaptic AZs (metaplasticity). Two findings causally linked this upshift to decreased olfactory memory performance. First, when continuously fed with spermidine, flies of 30 days of age (normally suffering from a complete loss of age-sensitive component of memory) were largely protected from these changes. Secondly, genetically provoking this up-shift eliminated the normally age-sensitive memory component in young animals already. An upshift in the AZ size should increase synaptic strength, evident in increased SV release in response to natural odors observed in aged but not aged-spermidine-fed flies. Presynaptic plasticity is crucial for forming memory traces in Drosophila. Previous work thus suggests that this presynaptic metaplasticity shifts the operational range of synapses in a way that they become unable to execute the plastic changes faithfully in response to conditioning stimuli (Bhukel, 2019).

This study further addressed the relation between defective autophagy, presynaptic ultrastructure and plasticity and olfactory memory formation. Autophagosome biogenesis is very dominant close to presynaptic specializations in distal axons in compartmentalized fashion and efficient macro-autophagy is essential for neuronal homeostasis and survival. Retrograde transport of autophagosomes might play a role in broader neuronal signaling processes, promoting neuronal complexity and preventing neurodegeneration. Surprisingly, however, the data do not favor a direct substrate relationship between AZ proteins and autophagy. Instead, evidence was found for a seemingly non-cell autonomous relation between brain-wide synapse organization and the autophagic status of the mere MB. After genetic impairment of autophagy (via atg5 or atg9 KD) using two different MB-specific Gal4-driver lines, the presynaptic metaplasticity was observed across the Drosophila olfactory system and beyond. While the autophagic arrest (p62 staining) was largely limited to the expression domain of these drivers, the synapses were pushed towards a state of metaplasticity. Since the ultrastructural size of AZs and the per AZ BRP levels increased equally in aged and MB-autophagy-challenged animals, it is concluded that the autophagic status of the MB neuron population executes a signaling process, which can control the per AZ amounts of BRP and other AZ proteins. Further studies are warranted to dissect the nature of these signaling processes (Bhukel, 2019).

Notably, accumulating evidences support the important role of neuropeptide Y (NPY) in aging and lifespan determination. NPY levels decrease with age in mice and re-substituting NPY is able to counteract age-induced changes of the brain at several levels. A cross-talk between autophagy and NPY in regulating the feeding behavior has been demonstrated in mice (Bhukel, 2019).

This study found that transcript expression level of an NPY family member (sNPF) are controlled by autophagy within the MBs. snpf hypomorph allele mimicking the MB reduction of sNPF of the MB-specific autophagy KD situations as well as the sNPF expression in aged animals. In this hypomorph allele a similar up regulation was observed in BRP Nc82 signal. KD of the snpfr using an MB-specific driver drove the brain-wide metaplastic change even stronger than the sNPF hypomorph (obviously only partially affecting the sNPF-specific signaling). This scenario in ultrastructural detail resembled both the age-induced and MB-specific autophagy-KD-induced metaplasticity phenotypes. These results, therefore, support the essential role of MB in integrating the metabolic state of Drosophila in an autocrine fashion to modulate the presynaptic release scaffold state throughout the fly brain. The mechanistic basis of this exciting regulation warrants further investigation. Interestingly, elevated cAMP signaling is generally driving plasticity in Drosophila neurons, while sNPF signaling is meant to reduce cAMP and thus potentially might be able to reset plastic changes such as increased BRP levels. In apparent contradiction to sNPF signaling directly widely controlling metaplasticity is the finding that MB-specific KD of the sNPFR sufficed to increase BRP levels. At this moment, it can only be speculated as to why KD of sNPF-receptor also results in extended metaplastic changes. Potentially, sNPF-receptor signaling within the MB might be important to control sNPF secretion in a physiological manner via a quasi-autocrine mechanism (Bhukel, 2019).

Intriguingly, the metaplastic state characterized both aged and MB-specific autophagy KD animals, and in both cases provoked a specific loss of the ASM component of memory. Notably, olfactory MTM measured in this study, are considered to be the direct precursor of olfactory LTM, which in turn have been shown to be energetically costly. Notably, autophagy and NPY signaling are prime candidate mechanisms for the therapy of age-induced cognitive processes (Bhukel, 2019).

Recent research has uncovered several examples connecting autophagy and hormonal-type regulations interacting between organ systems in non-cell autonomous regimes. For instance, Atg18 acts non-cell autonomously both in neurons and in intestines to firstly, maintain the wild-type lifespan of C.elegans and secondly, to respond to the dietary restriction and DAF-2 longevity signals. Atg18 in chemosensory neurons and intestines acts in parallel and converges on unidentified neurons that secrete neuropeptides to mediate the influence of Daf-2 on C.elegans lifespan through the transcription factor DAF-16/FOXO in response to reduced IGF signaling. In Drosophila, neuronal up-regulation of AMPK induces autophagy, via up-regulation of Atg1 non-cell autonomously in intestines and slows intestinal aging and vice versa. Moreover, up-regulation of Atg1 in neurons extends lifespan and maintains intestinal homeostasis during aging and these inter-tissue effects of AMPK/Atg1 were linked to altered insulin-like signaling. On the contrary, this study found the insulin producing cells (IPCs) themselves to not mediate the observed metaplastic state, as neither the KD of atg9 nor the KD of snpfr in Pars intercerebralis had any impact on the synaptic status of these flies (Bhukel, 2019).

Autophagy regulation is tightly connected to cellular energetics, nutrient recycling, and the maintenance of cellular energy status. The fruit fly can evaluate its metabolic state by integrating hunger and satiety signals at the very KC-to-MBON synapses in MB under control of dopaminergic neurons to control hunger-driven food-seeking behavior. At the same time, long-term memory encoding necessitates an increase in MB energy flux with dopamine signaling mediating this energy switch in the MB. In line with these findings, this study now provides a modeling basis to study these delicate relations in an exemplary fashion. Taken together, these data suggest that MB integrates the metabolic state of the flies via cross talk between autophagy and sNPF signaling with the decision whether to form memories or not and a block in this cross talk with aging gives rise to synaptic metaplasticity which initiates the age-induced memory impairment in Drosophila. It is tempting to speculate that the MB executes hierarchically, a high-level control integrating the metabolic and caloric situation with a life-strategy decision of whether or not to form mid-term memories (Bhukel, 2019).

Go to top

A Syndromic neurodevelopmental disorder caused by mutations in SMARCD1, a core SWI/SNF subunit needed for context-dependent neuronal gene regulation in flies

Nixon, K. C. J., Rousseau, J., Stone, M. H., Sarikahya, M., Ehresmann, S., Mizuno, S., Matsumoto, N., Miyake, N., Baralle, D., McKee, S., Izumi, K., Ritter, A. L., Heide, S., Heron, D., Depienne, C., Titheradge, H., Kramer, J. M. and Campeau, P. M. (2019). Am J Hum Genet 104(4): 596-610. PubMed ID: 30879640

Mutations in several genes encoding components of the SWI/SNF chromatin remodeling complex cause neurodevelopmental disorders (NDDs). This paper reports on five individuals with mutations in SMARCD1; the individuals present with developmental delay, intellectual disability, hypotonia, feeding difficulties, and small hands and feet. Trio exome sequencing proved the mutations to be de novo in four of the five individuals. Mutations in other SWI/SNF components cause Coffin-Siris syndrome, Nicolaides-Baraitser syndrome, or other syndromic and non-syndromic NDDs. Although the individuals presented in this study have dysmorphisms and some clinical overlap with these syndromes, they lack their typical facial dysmorphisms. To gain insight into the function of SMARCD1 in neurons, the Drosophila ortholog Bap60 was investigated in postmitotic memory-forming neurons of the adult Drosophila mushroom body (MB). Targeted knockdown of Bap60 in the MB of adult flies causes defects in long-term memory. Mushroom-body-specific transcriptome analysis revealed that Bap60 is required for context-dependent expression of genes involved in neuron function and development in juvenile flies when synaptic connections are actively being formed in response to experience. Taken together, this study identified an NDD caused by SMARCD1 mutations and establish a role for the SMARCD1 ortholog Bap60 in the regulation of neurodevelopmental genes during a critical time window of juvenile adult brain development when neuronal circuits that are required for learning and memory are formed (Nixon, 2019).

Go to top

Modulators of hormonal response regulate temporal fate specification in the Drosophila brain

Marchetti, G. and Tavosanis, G. (2019). PLoS Genet 15(12): e1008491. PubMed ID: 31809495

How a progenitor sequentially produces neurons of different fates and the impact of extrinsic signals conveying information about developmental progress or environmental conditions on this process represent key, but elusive questions. Each of the four progenitors of the Drosophila mushroom body (MB) sequentially gives rise to the MB neuron subtypes. The temporal fate determination pattern of MB neurons can be influenced by extrinsic cues, conveyed by the steroid hormone ecdysone. This study shows that the activation of Transforming Growth Factor-beta (TGF-beta) signalling via glial-derived Myoglianin regulates the fate transition between the early-born alpha'beta' and the pioneer alphabeta MB neurons by promoting the expression of the ecdysone receptor β1 isoform (EcR-β1). While TGF-beta signalling is required in MB neuronal progenitors to promote the expression of EcR-β1, ecdysone signalling acts postmitotically to consolidate the alpha'beta' MB fate. Indeed, it is proposed that if these signalling cascades are impaired alpha'beta' neurons lose their fate and convert to pioneer alphabeta. Conversely, an intrinsic signal conducted by the zinc finger transcription factor Kruppel-homolog 1 (Kr-h1) antagonises TGF-beta signalling and acts as negative regulator of the response mediated by ecdysone in promoting alpha'beta' MB neuron fate consolidation. Taken together, the consolidation of alpha'beta' MB neuron fate requires the response of progenitors to local signalling to enable postmitotic neurons to sense a systemic signal (Marchetti, 2019).

This study reveals a fundamental role for Myo-mediated TGF-β signalling in regulating fate specification of MB neurons. This signalling is initiated in the neuronal progenitors and it is proposed that it is necessary to consolidate the identity of newly born neurons by enabling them to sense and integrate the ecdysone hormonal signal. As modulator of this consolidation fate program, the factor Kr-h1 negatively regulates ecdysone signalling response and antagonises the TGF-β pathway (Marchetti, 2019).

Evidence derived from vertebrate models indicates that the temporal competence of neuronal precursors to generate different neuronal subtypes is governed by the combination of cell-intrinsic programs and extrinsic cues. In contrast, fate determination in the Drosophila nervous system appeared to be mainly determined by intrinsic cascades. Only recently, first reports started indicating that extrinsic factors can modulate fate decisions in the nervous system of the fly. Thus, fate decisions in the fly nervous system might follow principles that are more relatable to the ones utilised in vertebrate lineages than previously expected. Along these lines, the current data revealed a central role of TGF-β signalling in temporal fate specification during MB development. In the rodent hindbrain, midbrain and spinal cord, TGF-β signalling constrains the neural progenitor potency to promote fate transition from early to late born cell types, acting as a temporal switch signal regulating the expression of intrinsic identity factors in young progenitors. These similarities suggest that TGF-β might represent an evolutionary conserved extrinsic signal modulating temporal fate specification (Marchetti, 2019).

The present data suggest that TGF-β signalling links the temporal neuronal fate program to developmental progression. Re-examination of the EcR-β1 expression in dSmad21 mutant MB clones at late larval stages revealed a 12 hours delay in the onset of EcR-β1 expression leading to inability of MB neurons to respond to the prepupal ecdysone peak. Thus, TGF-β signalling might help to synchronize the production of distinct MB neuron subtypes coordinating diverse developmental programs. Accordingly, this study found that the glial Myo ligand mediates the TGF-β-dependent MB fate transition. Given that the prepupal ecdsyone peak is triggered after the larva reaches the critical weight point, it was hypothesise that glia serve as nutrition sensors in the brain during larval development and could be coordinating developmental timing of the fate specification program (Marchetti, 2019).

Although α'β' neurons are born during the larval stage, based on their immature dendrites and axons, and on the absence of functional response in appetitive olfactory learning behaviour, it appears that they are not fully differentiated at the end of larval life. Therefore, the initial state of these immature α'β' neurons could be labile. Their immature neurite trajectories might possess a certain degree of morphological plasticity, since at early pupal stages the axonal lobes are primarily made of α'β' axons, after γ axons have completely pruned. Indeed, the data provide strong support for the presence of an active consolidation signal required to maintain α'β' fate at adult stage. In fact, after impairment of TGF-β signalling, neurons born in the time window corresponding to the production phase of α'β' displayed the expected axonal pattern for α'β' neurons and expressed an α'β' marker before metamorphosis. Taken these data together, the alternative hypothesis that TGF-β signalling could be involved in the initial specification of α'β' MB neurons at mid-third instar appears much more unlikely. Notably, studies on fate specification in vertebrate systems have described a postmitotic fate consolidation event for developing motor and cortical neurons. In particular, the homeobox gene HB9 has an essential function in maintaining the fate of the motor neurons by actively suppressing the alternative V5 interneuron genetic program. Indeed, mice lacking HB9 function showed a normal number of motor neurons that acquired, though, molecular features of V5 interneurons. Interestingly, in absence of HB9 motor neurons are initially specified and they retain their characteristic axonal projection. Similarly, the expression of the retinoic acid receptor (RAR) is required to maintain the fate of layer V-III cortical neurons, and when the expression of RAR is abolished these neurons acquire the identity of layer II cortical neurons. These similarities in fate consolidation programs might reflect a common strategy in both invertebrates and vertebrates to first specify and then refine neuronal fate, according to the appropriate context (Marchetti, 2019).

Recently, RNA profiling analysis of MB neurons at different developmental time points uncovered a complex feedback regulation network that governs EcR expression. This combination of positive as well as negative feedback loops is required to coordinate EcR expression levels and its temporal regulation during brain development. FISH analysis suggested that TGF-β signalling promotes the transcription of the EcR-β1 gene in MB neurons at late wandering larval stage. Although detectable EcR-β1 protein is restricted to postmitotic MB neurons, genetic data revealed that TGF-β signalling is necessary in the MB progenitors to allow the expression of EcR-β1. This evidence raises the possibility that TGF-β signalling promotes the transcription of EcR gene in neuronal progenitors and potentially post-transcriptional mechanisms are involved to narrow down the translation of the EcR-β1 receptor only postmitotically. However, the data are against this hypothesis, since expression of EcR-β1 specifically in MB progenitors did not rescue the TGF-β signalling-dependent fate defects. Moreover, given that TGF-β signalling is required to consolidate the fate of the larval-born α'β' neurons at the end of larval stage, suggests that the TGF-β pathway regulates a consolidation fate process independently of cell division. In this scenario, the expression of EcR-β1 in the newly born neurons could be promoted via a cell-to-cell communication signalling cascade initiated in neuronal progenitors by the activity of TGF-β signalling. Examples of this type of signal transmission are represented by the juxtacrine signalling mediated by Notch, Semaphorin or Ephrin pathways. In particular, the intercellular interaction between Notch and its ligand Delta in neighbouring cells is fundamental to direct cell fate decisions (Marchetti, 2019).

In addition to an upstream regulation of ecdysone signalling, this study uncovered the intrinsic factor Kr-h1 as a downstream modulator of the ecdysone-dependent fate consolidation program. Interestingly, the transition from larval stage to metamorphosis is regulated by the balance of the two major hormones, the juvenile hormone (JH) and ecdysone. JH prevents metamorphosis by the induction of the transcription factor Kr-h1 within the ring gland, which in turn suppresses the up-regulation of the ecdysone-dependent metamorphic genes E93 and Broad Complex. The TGF-β/Activin pathway contributes to decreasing Kr-h1 expression via E93 allowing the beginning of metamorphosis. Along these lines, the antagonism between ecdysone and JH through Kr-h1 could potentially regulate the MB temporal fate cascade at the onset of metamorphosis (Marchetti, 2019).

In conclusion, this work shed light on the intrinsic and extrinsic mechanisms regulating the consolidation of the terminal fate. Understanding these processes will help gain insights into their dysregulation in neurodevelopmental disorders and into their role in stem cell reprogramming (Marchetti, 2019).

Go to top

Juvenile hormone drives the maturation of spontaneous mushroom body neural activity and learned behavior

Leinwand, S. G. and Scott, K. (2021). Neuron. PubMed ID: 33915110

Leinwand, S. G. and Scott, K. (2021). Neuron. PubMed ID: 33915110

Juvenile hormone drives the maturation of spontaneous mushroom body neural activity and learned behavior

Mature behaviors emerge from neural circuits sculpted by genetic programs and spontaneous and evoked neural activity. However, how neural activity is refined to drive maturation of learned behavior remains poorly understood. This study explored how transient hormonal signaling coordinates a neural activity state transition and maturation of associative learning. Spontaneous, asynchronous activity was identified in a Drosophila learning and memory brain region, the mushroom body. This activity declines significantly over the first week of adulthood. Moreover, this activity is generated cell-autonomously via Cacophony voltage-gated calcium channels in a single cell type, α'/β' Kenyon cells. Juvenile hormone, a crucial developmental regulator, acts transiently in α'/β' Kenyon cells during a young adult sensitive period to downregulate spontaneous activity and enable subsequent enhanced learning. Hormone signaling in young animals therefore controls a neural activity state transition and is required for improved associative learning, providing insight into the maturation of circuits and behavior (Leinwand, 2021).

Genetic programs and experience in the form of neural activity refine neural circuits, sculpting cognitive function over time. Activity state transitions in neural circuits are widespread during normal development. Achieving the mature activity state is correlated with the emergence of adult behavioral outputs. For example, periodic waves of spontaneous neural activity occur throughout immature visual, somatosensory, and motor brain regions in perinatal critical periods, before distinct, less correlated sensory-evoked or locomotion-related activity patterns emerge in older animals. Periodic bursts of spontaneous activity also occur in the hippocampus, specifically in early mammalian post-natal development, in a brief period prior to development of robust long-term potentiation. Although temporal evolution of spontaneous neural activity patterns is prevalent in developing circuits, the molecular mechanisms that control the timing of neural activity state transitions and coordinate maturation of behavioral outputs in young animals are largely unknown (Leinwand, 2021).

Hormones regulate multiple aspects of the maturation of the nervous system. Systemic hormonal signaling controls neural differentiation, remodeling, physiology, and other key events for the refinement of neural circuits. For example, sex steroid hormones act transient in a critical prenatal window to regulate the development of neural circuits for sexually dimorphic behaviors, producing enduring changes in the brain. Furthermore, many hormone receptors directly alter transcription and consequently have direct or indirect effects on ion channels, synapses, and neurotransmission. Gonadal hormone signaling accelerates the maturation of inhibitory neurotransmission in cortical circuits, with correlated effects on behavior. Moreover, thyroid hormones regulate synaptic transmission in the hippocampus in young animals, with clear implications for memory. Juvenile hormone (JH) is an insect hormone with functional similarities to mammalian thyroid hormones. JH circulates widely and acts on diverse neural circuits in young animals to regulate metamorphosis, reproduction, and courtship. Across species, hormonal signaling is therefore well poised to coordinate key transitions in the maturation of the nervous system and behavior at particular stages of animal development (Leinwand, 2021).

A mechanistic understanding of how the nervous system achieves activity state transitions will provide insight into the origins of mature behaviors. Critically, evaluating the role of hormones in neural activity and behavior maturation requires isolating their effects on specific cells in known circuits and at particular developmental times. The fruit fly Drosophila melanogaster system offers powerful genetic tools to causally link in vivo neural activity with behavior and to manipulate gene expression, with single-cell resolution. Recently, high levels of spontaneous neural activity were observed in the developing Drosophila visual system, illustrating that activity maturation occurs in invertebrate systems, as well as vertebrates. In examining activity in the adult fly brain, this study observed high levels of activity in the Drosophila mushroom body (MB) brain region that declined rapidly with age. The MB is critical for learned behavior. Physiological, molecular, behavioral, and anatomical studies, including a complete connectome, have provided a uniquely rich understanding of the neuronal architecture and function of the MB. Discovery of an immature to mature activity state transition in this well-described system offers an entry point to rigorously examine how neural activity in young animals drives refinement and maturation of behavior (Leinwand, 2021).

This study employed in vivo functional imaging and powerful genetic tools to describe a high spontaneous activity state in the Drosophila MB learning and memory brain center of young animals and its crucial role in the maturation of learned behavior. Spontaneous, asynchronous activity was identified specifically in one MB cell type, the α'/β' Kenyon cells (KCs), in young animals, that unexpectedly declines over the first week of adulthood. Cacophony (Cac) voltage-gated calcium channels mediate this young animal spontaneous activity. JH, a crucial regulator of insect development similar to vertebrate thyroid hormones, signaling specifically in α'/β' KCs during a sensitive period in early adulthood coordinates the maturation of neural activity states and is required for mature associative learning (Leinwand, 2021).

This study shows that JH acts on the α'/β' KCs of young adults to downregulate spontaneous activity and enhance associative learning in older animals. Specifically, it was found that α'/β' KCs exhibit sensorimotor-independent, TTX-insensitive asynchronous activity in young animals that is mediated by Cac voltage-gated calcium channels. JH signaling in a young animal sensitive period, when the titer of JH circulating is high, is required to achieve the mature, low KC activity state and enhance learned behavior. The discovery that a hormone triggers a neural activity state transition essential for robust learning provides a model for mechanistically probing the maturation of learning circuits and behavior (Leinwand, 2021).

Many animals are born with immature learning capabilities. In mammals, periodic giant depolarizing potentials occur in the immature hippocampus. Because multiple hippocampus-dependent learned behaviors are poor at the time of the giant depolarizing potentials and mature slowly over the first post-natal month, a correlation between this pattern of spontaneous activity and learning is apparent. However, causal links between hippocampal activity patterns and maturation of learned behavioral outputs are lacking, despite their profound implications for the plasticity to form new associations throughout adulthood. These studies demonstrate that Drosophila associative learning improves over the first week of adulthood and that appropriately regulated activity state transitions in higher-order brain regions are necessary for this learning maturation (Leinwand, 2021).

These studies reveal that the spontaneous activity generated in young α'/β' KCs is critical for honing the neural circuits that subsequently produce mature learning. Notably, the activity in young KCs is asynchronous and unpatterned, unlike the propagating waves of activity in immature visual and somatosensory areas or the rhythmic alternations in motor regions. In contrast with these sensorimotor systems, the MB and many higher-order brain regions are not topographically organized, and neighboring neurons do not respond to similar stimulus features. Instead, connectivity between KCs and their presynaptic partners is stochastic, and sensory-evoked responses are sparse and unpatterned. It is proposed that transient unpatterned activity in young animals is a necessary precursor to the unordered and spatially distributed sensory-evoked responses seen in adults, providing a substrate for subsequent adult experiences (Leinwand, 2021).

This study describes age-dependent spontaneous activity that is restricted to a single cell type within the learning circuit, the α'/β' KCs. Although multi-parallel and distributed processing in MB circuit modules gives rise to associative learning, the precise role of α'/β' KCs in learned behavior remains less well understood than other MB cell types. Sparse activity in α'/β' KCs may encode sensory information and information about reward or punishment. Behaviorally, α'/β' KCs are required for the acquisition and consolidation of appetitive and aversive olfactory and gustatory associative memories. Because KC activity coincident with salient stimuli are key elements to form associative memories, the poor learning performance of young animals and of older animals manipulated to aberrantly retain high levels of α'/β' KC activity is unexpected. The results suggest that high α'/β' KC activity is a necessary feature of immature circuits but may acutely interfere with robust learning. Activity state transitions in young animals may refine responses to conditioned stimuli in mature animals. Whether high α'/β' KC activity in young animals organizes or is permissive for the subsequent role of α'/β' KCs in memory acquisition and consolidation remains to be investigated (Leinwand, 2021).

Among the MB cell types, only α'/β' KCs undergo a high- to low-activity state transition. α'/β' KCs are not uniquely able to directly transduce JH, because the JH receptors Met and Gce are highly expressed throughout the MB. Cac voltage-gated calcium channels are also highly expressed in the entire MB. Nevertheless, α'/β' KCs were found to have the lowest firing threshold and, correspondingly, the highest rate of baseline and odor sensory-evoked spiking of the three KC classes. Although these physiological properties were not studied in the context of age, the current studies reveal a change in baseline activity states in the first week of adulthood. It is therefore speculated that α'/β' KCs are intrinsically more excitable due to a unique gene expression profile. Specific ligand- or voltage-gated ion channels or ion pumps may display α'/β' KC-biased expression and may undergo changes in expression in early adulthood that directly contribute to cellular excitability. α'/β' KCs may have distinct plasticity rules that derive from these age-dependent gene expression and excitability changes (Leinwand, 2021).

Hormone signaling regulates α'/β' KC physiology with age. Although Cac channels mediate young α'/β' KC spontaneous activity and JH signaling controls the neural activity state transition, a direct JH-to-Cac channel connection is unlikely. Cac mRNA expression in α'/β' KCs does not change over the first week of adulthood, consistent with the absence of evidence that Met and Gce hormone receptors directly target Cac channels. It is therefore hypothesized that JH may indirectly influence Cac channel function. The finding that the high activity retained in α'/β' KCs in mature flies with Met and Gce receptors knocked down was sensitive to the voltage-gated calcium channel antagonist PLTX supports an indirect link between JH and these channels. Thus, it is speculated that JH signaling normally produces transcriptional changes in young animals that influence the overall physiology and resting membrane potential of α'/β' KCs. These changes in α'/β' KC membrane potential may then reduce Cac channel opening and calcium flux over the first week of adulthood. Future investigation of how the direct targets of JH signaling ultimately influence the membrane potential and Cac function will provide new insights into the underlying circuit maturation mechanisms (Leinwand, 2021).

This study found that transient hormonal signaling is critically necessary to impart stable changes in neural activity and learned behavior. JH, acting on KCs of young animals, coordinates the decrease in spontaneous activity and the maturation of adult learned behavior. When JH signaling is disrupted transiently in α'/β' KCs during a sensitive period in young animals, older animals retain high levels of spontaneous KC activity and poor learned behavior, mimicking the activity and behavior of young animals. Thus, hormone signaling is essential for learning circuits to transition from an immature to a mature state capable of robust learning. The JH receptors Met and Gce, like many hormone receptors, can directly alter transcription. Gene expression changes downstream of these hormone receptors likely directly or indirectly modulate ion channels, synapses, and neurotransmission, thereby sculpting learning circuits. It is speculated that structural refinement of learning circuits underlies the maturation of learned behavior; therefore, further investigation of hormone-triggered molecular changes affecting neurotransmission may provide new entry points for investigating these fundamental age-dependent processes. Together, these studies provide insight into the maturation of activity states and learned behaviors and a platform to examine how hormonally evoked cellular changes enhance the acquisition and maintenance of learned associations (Leinwand, 2021).

Go to top

Extrinsic activin signaling cooperates with an intrinsic temporal program to increase mushroom body neuronal diversity

Rossi, A. M. and Desplan, C. (2020). Elife 9. PubMed ID: 32628110

Temporal patterning of neural progenitors leads to the sequential production of diverse neurons. To understand how extrinsic cues influence intrinsic temporal programs, Drosophila mushroom body progenitors (neuroblasts) were studied that sequentially produce only three neuronal types: γ, then α'β', followed by αβ. Opposing gradients of two RNA-binding proteins Imp and Syp comprise the intrinsic temporal program. Extrinsic activin signaling regulates the production of α'β' neurons but whether it affects the intrinsic temporal program was not known. This study shows that the activin ligand Myoglianin from glia regulates the temporal factor Imp in mushroom body neuroblasts. Neuroblasts missing the activin receptor Baboon have a delayed intrinsic program as Imp is higher than normal during the α'β' temporal window, causing the loss of α'β' neurons, a decrease in αβ neurons, and a likely increase in γ neurons, without affecting the overall number of neurons produced. These results illustrate that an extrinsic cue modifies an intrinsic temporal program to increase neuronal diversity (Rossi, 2020).

The building of intricate neural networks during development is controlled by highly coordinated patterning programs that regulate the generation of different neuronal types in the correct number, place and time. The sequential production of different neuronal types from individual progenitors, i.e. temporal patterning, is a conserved feature of neurogenesis. For instance, individual radial glia progenitors in the vertebrate cortex sequentially give rise to neurons that occupy the different cortical layers in an inside-out manner. In Drosophila, neural progenitors (called neuroblasts) also give rise to different neuronal types sequentially. For example, projection neurons in the antennal lobe are born in a stereotyped temporal order and innervate specific glomeruli. In both of these examples, individual progenitors age concomitantly with the developing animal (e.g., from embryonic stages 11-17 in mouse and from the first larval stage (L1) to the end of the final larva stage (L3) in Drosophila). Thus, these progenitors are exposed to changing environments that could alter their neuronal output. Indeed, classic heterochronic transplantation experiments demonstrated that young cortical progenitors placed in an old host environment alter their output to match the host environment and produce upper-layer neurons (Rossi, 2020).

The adult Drosophila central brain is built from ~100 neuroblasts that divide continuously from L1 to L3. Each asymmetric division regenerates the neuroblast and produces an intermediate progenitor called ganglion mother cell (GMC) that divides only once, typically producing two different cell types. Thus, during larval life central brain neuroblasts divide 50-60 times, sequentially producing many different neuronal types. All central brain neuroblasts progress through opposing temporal gradients of two RNA-binding proteins as they age: IGF-II mRNA binding protein (Imp) when they are young and Syncrip (Syp) when they are old. Loss of Imp or Syp in antennal lobe or Type II neuroblasts affects the ratio of young to old neuronal types. Imp and Syp also affect neuroblast lifespan. Thus, a single temporal program can affect both the diversity of neuronal types produced and their numbers (Rossi, 2020).

Since central brain neuroblasts produce different neuronal types through developmental time, roles for extrinsic cues have recently garnered attention. Ecdysone triggers all the major developmental transitions including progression into the different larval stages and entry in pupation. The majority of central brain neuroblasts are not responsive to ecdysone until mid-larval life when they begin to express the Ecdysone Receptor (EcR). Expressing a dominant-negative version of EcR (EcR-DN) in Type II neuroblasts delays the Imp to Syp transition that normally occurs ~60 hr after larval hatching (ALH). This leads to many more cells that express the early-born marker gene Repo and fewer cells that express the late-born marker gene Bsh (Rossi, 2020).

To further understand how extrinsic signals contribute to temporal patterning, Drosophila mushroom body neuroblasts were studied because of the deep understanding of their development. The mushroom body is comprised of ~2000 neurons (Kenyon cells) that belong to only three main neuronal types that have unique morphologies and play distinct roles in learning and memory. They receive input mainly from ~200 projection neurons that each relays odor information from olfactory receptor neurons. Each projection neuron connects to a random subset of Kenyon cells and each Kenyon cell receives input from ~7 different projection neurons. This connectivity pattern requires a large number of mushroom body neurons (~2,000) to represent complex odors. To produce this very large number of neurons, mushroom body development is unique in many respects. Mushroom body neurons are born from four identical neuroblasts that divide continuously (unlike any other neuroblast) from the late embryonic stages until the end of pupation (~9 days for ~250 divisions each). Furthermore, the two neurons born from each mushroom body GMC are identical. The neuronal simplicity of the adult mushroom body makes it ideal to study how extrinsic cues might affect diversity since the loss of any single neuronal type is obvious given that each is represented hundreds of times (Rossi, 2020).

The three main neuronal types that make up the adult mushroom body are produced sequentially during neurogenesis: first γ, followed by α'β', and then αβ neurons (see α'β' neurons are not generated from babo mutant neuroblasts), representing the simplest lineage in the central brain. The γ temporal window extends from L1 (the first larval stage) until mid-L3 (the final larval stage) when animals attain critical weight and are committed to metamorphosis; the α'β' window from mid-L3 to the beginning of pupation, and the αβ window from pupation until eclosion (the end of development). Like all other central brain neuroblasts Imp and Syp are expressed by mushroom body neuroblasts, but in much shallower gradients through time, which accounts for their extended lifespan. Imp and Syp are inherited by newborn neurons where they instruct temporal identity. Imp positively and Syp negatively regulate the translation of chronologically inappropriate morphogenesis (chinmo), a gene encoding a transcription factor that acts as a temporal morphogen in neurons. The first-born γ neurons are produced for the first ~85 cell divisions, when Imp levels in neuroblasts, and thus Chinmo in neurons, are high. α'β' neurons are produced for the next ~40 divisions, when Imp and Syp are at similar low levels that translate into lower Chinmo levels in neurons. Low Chinmo then regulates the expression in neurons of maternal gene required for meiosis (mamo), which encodes a transcription factor that specifies the α'β' fate and whose mRNA is stabilized by Syp. αβ neurons are generated for the final ~125 neuroblast divisions, when Syp levels are high, Imp is absent in neuroblasts, and thus Chinmo and Mamo are no longer expressed in neurons (Rossi, 2020).

Extrinsic cues are known to have important roles in regulating neuronal differentiation during mushroom body neurogenesis. The ecdysone peak that controls entry into pupation regulates γ neuron axonal remodeling. Ecdysone was also proposed to be required for the final differentiation of α'β' neurons. EcR expression in γ neurons is timed by activin signaling, a member of the TGFβ family, from local glia. Activin signaling from glia is also required for the α'β' fate: Knocking-down the activin pathway receptor Baboon (Babo) leads to the loss of α'β' neurons. It was proposed that activin signaling in mushroom body neuroblasts regulates the expression of EcR in prospective α'β' neurons and that when the activin pathway is inhibited, it leads to the transformation of α'β' neurons into later-born pioneer-αβ neurons (a subclass of the αβ class) (Rossi, 2020).

Although there is strong evidence that extrinsic cues have important functions in neuronal patterning in the Drosophila central brain, it remains unknown how extrinsic temporal cues interface with the Imp and Syp intrinsic temporal program to regulate neuronal specification. This question was addressed using the developing mushroom bodies. Activin signaling from glia was shown to be required for α'β' specification. However, this study also showed that activin signaling lowers the levels of the intrinsic factor Imp in mushroom body neuroblasts to define the mid-α'β' temporal identity window. Removing the activin receptor Babo in mutant clones leads to the loss of α'β' neurons, to fewer last-born αβ neurons, and to the likely generation of additional first-born γ neurons without affecting overall clone size. This appears to be caused by a delayed decrease in Imp levels, although the intrinsic temporal clock still progresses even in the absence of activin signaling. This study also demonstrated that ecdysone signaling is not necessary for the specification of α'β' neurons, although it might still be involved in later α'β' differentiation. These results provide a model for how intrinsic and extrinsic temporal programs operate within individual progenitors to regulate neuronal specification (Rossi, 2020).

Mushroom body neurogenesis is unique and programmed to generate many copies of a few neuronal types. During the early stages of mushroom body development, high Imp levels in mushroom body neuroblasts are inherited by newborn neurons and translated into high Chinmo levels to specify γ identity. As in other central brain neuroblasts, as development proceeds, inhibitory interactions between Imp and Syp help create a slow decrease of Imp and a corresponding increase of Syp. However, at the end of the γ temporal window (mid-L3), activin signaling from glia acts to rapidly reduce Imp levels in mushroom body neuroblasts without significantly affecting Syp, establishing a period of low Imp (and thus low Chinmo in neurons) and also low Syp. This is required for activating effector genes in prospective α'β' neurons, including Mamo, whose translation is promoted by Syp (Liu, 2019). The production of αβ identity begins when Imp is further decreased and Syp levels are high during pupation (see Model of how activin signaling defines the α'β' temporal identity window.). Low Chinmo in αβ neurons is also partly regulated by ecdysone signaling through the activation of Let-7-C, which targets chinmo for degradation. Based on this model, α'β' neurons could not be rescued by knocking-down Imp in babo clones, since low Imp is required for α'β' specification while the knockdown reduces its level below this requirement. It would be expected to rescue α'β' neurons if Imp levels were specifically reduced to the appropriate levels at L3. However, reducing Imp levels might not be the only function of activin signaling, which may explain why α'β' neurons are not simply made earlier (e.g., during L1-L2) when Imp is knocked-down (Rossi, 2020).

In babo mutant clones, it is speculated that additional γ neurons are produced at the expense of α'β' neurons since Imp levels in neuroblasts (as well as Chinmo in neurons) are higher for a longer time during development; There was also a significant decrease in the total number of αβ neurons in babo mutant clones that contrasts with a previous report that instead concluded that additional pioneer-αβ neurons are produced. It is believed that there is both an increase in the number of γ neurons and of the pioneer-αβ neuron subclass because pioneer-αβ neurons are the first of the αβ class to be specified (when Imp is still present at very low levels) during pupation. It is speculated that pioneer-αβ neurons are produced during the extended low Imp window that was detected during pupation in babo clones. However, this does not leave the time for the remaining population of αβ neurons to be formed, which explains why their number is reduced (Rossi, 2020).

This study has focused on the three main classes of mushroom body neurons although at least seven subtypes exist: 2 γ, 2 α'β' and 3 αβ. The subtypes are specified sequentially suggesting that each of the three broad mushroom body temporal windows can be subdivided further, either by fine-scale reading of the changing Imp and Syp gradients, by additional extrinsic cues, or perhaps by a tTF series as in other neuroblasts (Rossi, 2020).

Postembryonic central brain neuroblasts are long-lived and divide on average ~50 times. Unlike in other regions of the developing Drosophila brain, rapidly progressing series of tTFs have not yet been described in these neuroblasts. Instead, they express Imp and Syp in opposing temporal gradients. Conceptually, how Imp and Syp gradients translate into different neuronal identities through time has been compared to how morphogen gradients pattern tissues in space. During patterning of the anterior-posterior axis of the Drosophila embryo, the anterior gradient of the Bicoid morphogen and the posterior Nanos gradient are converted into discrete spatial domains that define cell fates. Since gradients contain unlimited information, differences in Imp and Syp levels through time could translate into different neuronal types. Another intriguing possibility is that tTF series could act downstream of Imp and Syp, similarly to how the gap genes in the Drosophila embryo act downstream of the anterior-posterior morphogens. This study has shown that another possibility is that temporal extrinsic cues can be incorporated by individual progenitors to increase neuronal diversity. In mushroom body neuroblasts activin signaling acts directly on the intrinsic program, effectively converting two broad temporal windows into three to help define an additional neuronal type. It is proposed that subdividing the broad Imp and Syp temporal windows by extrinsic cues may be a simple way to increase neuronal diversity in other central brain neuroblasts (Rossi, 2020).

This study has also shown that activin signaling times the Imp to Syp transition for mushroom body neuroblasts, similar to the function of ecdysone for other central brain neuroblasts. In both cases however, the switch still occurs, indicating that a separate independent clock continues to tick. This role for extrinsic cues during Drosophila neurogenesis is reminiscent of their roles on individual vertebrate progenitors. For example, hindbrain neural stem cells progressively produce motor neurons followed by serotonergic neurons before switching to producing glia. The motor neuron to serotonergic neuron switch is fine-tuned by TGFβ signaling. It would be interesting to determine if hindbrain neuronal subtypes are lost in TGFβ mutants, similar to how α'β' identity is lost in the mushroom bodies in babo mutants (Rossi, 2020).

The specification of α'β' neurons begins at mid-L3 with the onset of Mamo expression. In contrast, high levels of EcR are detected in mature mushroom body neurons starting at late L3. At this stage, both γ and α'β' neurons already exist and new α'β' neurons are still being generated. Thus, Mamo expression precedes EcR expression. These non-overlapping expression patterns suggest that ecdysone signaling does not regulate Mamo and therefore cannot control the specification of α'β' neurons. Furthermore, expression of UAS-EcR-RNAi or mutants for usp do not lead to the loss of α'β' neurons. It is noted that usp results contradict the loss of α'β' neuron in usp clones. However, α'β' neurons were seen in these clones based on the morphology of these neurons but the remodeling defect of γ neurons makes α'β' neurons difficult to identify. Nevertheless, ecdysone might still function later during α'β' differentiation, particularly during pupation when all mushroom body neurons express EcR (Rossi, 2020).

This study and that of Marchetti both show that expression of UAS-EcR-DN leads to the loss of α'β' neurons by acting in mushroom body neurons but not in neuroblasts. However, EcR must be first be expressed in the target cells of interest in order to make any conclusions about ecdysone function using UAS-EcR-DN. Since this study could not detect EcR protein in Mamo+ cells at L3, but expressing UAS-EcR-DN inhibits Mamo in those cells, it is concluded that EcR-DN artifactually represses Mamo and leads to the loss of α'β' neurons. This explains why expressing UAS-EcR-B1 does not rescue α'β' neurons in babo clones. However, Marchetti did rescue babo-RNAi by expressing EcR (Marchetti, 2019). This is likely because the current experiments were performed using babo MARCM clones in which the loss of α'β' neurons is much more severe than with babo-RNAi used in their experiments. Indeed, when attempts were made to eliminate α'β' neurons using a validated UAS-babo-RNAi construct, γ neurons did not remodel but there was only a minor (but significant) decrease in the number of α'β' neurons. This indicates that knocking-down babo with mb-Gal4 that is only weakly expressed in neuroblasts and newborn neurons is not strong enough to inhibit α'β' specification. Thus, it is speculated that the LexA line used by Marchetti (GMR26E01-LexA) may not be a reliable reporter for α'β' neurons upon babo knockdown, and that it might be ecdysone sensitive later in α'β' differentiation. Since EcR expression in all mushroom body neurons at L3 may be dependent on activin signaling directly in neurons, as it is in γ neurons for remodeling, expressing UAS-EcR-B1 together with UAS-babo-RNAi using OK107-Gal4 might both reduce the effectiveness of the RNAi while also allowing for the re-expression of GMR26E01-LexA (Rossi, 2020).

Glia are a source of the activin ligand myo, which is temporally expressed in brain glia starting at L3 to initiate the remodeling of mushroom body γ neurons and α'β' specification . However, knocking-down Myo from glia is not as severe as removing Babo from mushroom body neuroblasts. This might be due to incomplete knockdown of myo or to other sources of Myo, potentially from neurons. For example, in the vertebrate cortex, old neurons signal back to young neurons to control their numbers. It is also possible the Babo is activated by other activin ligands, including Activin and Dawdle. An intriguing hypothesis is that the temporal expression of myo in glia beginning at mid-L3 is induced by the attainment of critical weight and rising ecdysone levels. It would be interesting to determine whether blocking ecdysone signaling in glia leads to the loss of α'β' specification, similar to how blocking ecdysone reception in astrocytes prevents γ neuron remodeling (Rossi, 2020).

It is well established that extrinsic cues play important roles during vertebrate neurogenesis, either by regulating temporal competence of neural stem cells or by controlling the timing of temporal identity transitions. Competence changes mediated by extrinsic cues were demonstrated in classic heterochronic transplantation studies that showed that young donor progenitors produce old neuronal types when placed in older host brains. Recent studies show that the reverse is also true when old progenitors are placed in a young environment (Rossi, 2020).

Mechanisms of intrinsic temporal patterning are also conserved. For example, vertebrate retinal progenitor cells use an intrinsic tTF cascade to bias young, middle, and old retinal fates. Two of the factors (Ikaros and Casz1) used for intrinsic temporal patterning are orthologs to the Drosophila tTFs Hb and Cas. tTF series might also exist in cortical radial glia progenitors and even in the spinal cord. Recent results also show the importance of post-transcriptional regulation in defining either young or old cortical fates, which can be compared to the use of post-transcriptional regulators that are a hallmark of neuronal temporal patterning in Drosophila central brain neuroblasts. These studies highlight that the mechanisms driving the diversification of neuronal types are conserved (Rossi, 2020).

Go to top

Drosophila Graf regulates mushroom body beta-axon extension and olfactory long-term memory

Kim, S., Kim, J., Park, S., Park, J. J. and Lee, S. (2021). Mol Brain 14(1): 73. PubMed ID: 33892766

Loss-of-function mutations in the human oligophrenin-1 (OPHN1) gene cause intellectual disability, a prevailing neurodevelopmental condition. However, the role OPHN1 plays during neuronal development is not well understood. This study investigated the role of the Drosophila OPHN1 ortholog Graf in the development of the mushroom body (MB), a key brain structure for learning and memory in insects. Loss of Graf causes abnormal crossing of the MB β lobe over the brain midline during metamorphosis. This defect in Graf mutants is rescued by MB-specific expression of Graf and OPHN1. Furthermore, MB α/β neuron-specific RNA interference experiments and mosaic analyses indicate that Graf acts via a cell-autonomous mechanism. Consistent with the negative regulation of epidermal growth factor receptor (EGFR)-mitogen-activated protein kinase (MAPK) signaling by Graf, activation of this pathway is required for the β-lobe midline-crossing phenotype of Graf mutants. Finally, Graf mutants have impaired olfactory long-term memory. These findings reveal a role for Graf in MB axon development and suggest potential neurodevelopmental functions of human OPHN1 (Kim, 2021).

Go to top

The F-Box Protein CG5003 Regulates Axon Pruning and the Integrity of the Drosophila Mushroom Body

Yang, M., Guo, Y., Wang, S., Chen, C., Chang, Y. H. and Ho, M. S. (2021). . Front Mol Neurosci 14: 634784. PubMed ID: 33716667

Protein homeostasis serves as an important step in regulating diverse cellular processes underlying the function and development of the nervous system. In particular, the ubiquitination proteasome system (UPS), a universal pathway mediating protein degradation, contributes to the development of numerous synaptic structures, including the Drosophila olfactory-associative learning center mushroom body (MB), thereby affecting associated function. This study describes the function of a newly characterized Drosophila F-box protein CG5003, an adaptor for the RING-domain type E3 ligase (SCF complex), in MB development. Lacking CG5003 ubiquitously causes MB γ axon pruning defects and selective CG5003 expression in pan-neurons leads to both γ axon and α/β lobe abnormalities. Interestingly, change in CG5003 expression in MB neurons does not cause any abnormalities in axons, suggesting that CG5003 functions in cells extrinsic to MB to regulate its development. Mass spectrum analysis indicates that silencing CG5003 expression in all neurons affects expression levels of proteins in the cell and structural morphogenesis, transcription regulator activity, and catalytic activity. These findings reinforce the importance of UPS and identify a new factor in regulating neuronal development as exemplified by the synaptic structure of the MB (Yang, 2021).

Go to top

Polycomb group genes are required for neuronal pruning in Drosophila

Bu, S., Lau, S. S. Y., Yong, W. L., Zhang, H., Thiagarajan, S., Bashirullah, A. and Yu, F. (2023). BMC Biol 21(1): 33. PubMed ID: 36793038

Pruning that selectively eliminates unnecessary or incorrect neurites is required for proper wiring of the mature nervous system. During Drosophila metamorphosis, dendritic arbourization sensory neurons (ddaCs) and mushroom body (MB) γ neurons can selectively prune their larval dendrites and/or axons in response to the steroid hormone ecdysone. An ecdysone-induced transcriptional cascade plays a key role in initiating neuronal pruning. However, how downstream components of ecdysone signalling are induced remains not entirely understood. This study identified that Scm, a component of Polycomb group (PcG) complexes, is required for dendrite pruning of ddaC neurons. Two PcG complexes, PRC1 and PRC2, are important for dendrite pruning. Interestingly, depletion of PRC1 strongly enhances ectopic expression of Abdominal B (Abd-B) and Sex combs reduced, whereas loss of PRC2 causes mild upregulation of Ultrabithorax and Abdominal A in ddaC neurons. Among these Hox genes, overexpression of Abd-B causes the most severe pruning defects, suggesting its dominant effect. Knockdown of the core PRC1 component Polyhomeotic (Ph) or Abd-B overexpression selectively downregulates Mical expression, thereby inhibiting ecdysone signalling. Finally, Ph is also required for axon pruning and Abd-B silencing in MB γ neurons, indicating a conserved function of PRC1 in two types of pruning. This study demonstrates important roles of PcG and Hox genes in regulating ecdysone signalling and neuronal pruning in Drosophila. Moreover, these findings suggest a non-canonical and PRC2-independent role of PRC1 in Hox gene silencing during neuronal pruning (Bu, 2023).

Go to top

A KDM5-Prospero transcriptional axis functions during early neurodevelopment to regulate mushroom body formation

Hatch, H. A. M., Belalcazar, H. M., Marshall, O. J. and Secombe, J. (2021). Elife 10. PubMed ID: 33729157

Mutations in the lysine demethylase 5 (KDM5) family of transcriptional regulators are associated with intellectual disability, yet little is known regarding their spatiotemporal requirements or neurodevelopmental contributions. Utilizing the mushroom body (MB), a major learning and memory center within the Drosophila brain, this study demonstrates that KDM5 is required within ganglion mother cells and immature neurons for proper axogenesis. Moreover, the mechanism by which KDM5 functions in this context is independent of its canonical histone demethylase activity. Using in vivo transcriptional and binding analyses, a network of genes directly regulated by KDM5 was identified that are critical modulators of neurodevelopment. KDM5 directly regulates the expression of prospero, a transcription factor that this study demonstrates to be essential for MB morphogenesis. Prospero functions downstream of KDM5 and binds to approximately half of KDM5-regulated genes. Together, these data provide evidence for a KDM5-Prospero transcriptional axis that is essential for proper MB development (Hatch, 2021).

Magnesium efflux from Drosophila Kenyon Cells is critical for normal and diet-enhanced long-term memory
Wu, Y., Funato, Y., Meschi, E., Jovanoski, K., Miki, H. and Waddell, S. (2020). Elife 9. PubMed ID: 33242000

Dietary magnesium (Mg(2+)) supplementation can enhance memory in young and aged rats. Memory-enhancing capacity was largely ascribed to increases in hippocampal synaptic density and elevated expression of the NR2B subunit of the NMDA-type glutamate receptor. This study shows that Mg(2+) feeding also enhances long-term memory in Drosophila. Normal and Mg(2+) enhanced fly memory appears independent of NMDA receptors in the mushroom body and instead requires expression of a conserved CNNM-type Mg(2+)-efflux transporter encoded by the unextended (uex) gene. UEX contains a putative cyclic nucleotide-binding homology domain and its mutation separates a vital role for uex from a function in memory. Moreover, UEX localization in mushroom body Kenyon Cells is altered in memory defective flies harboring mutations in cAMP-related genes. Functional imaging suggests that UEX-dependent efflux is required for slow rhythmic maintenance of Kenyon Cell Mg(2+). It is proposed that regulated neuronal Mg(2+) efflux is critical for normal and Mg(2+) enhanced memory (Wu, 2020).

Magnesium (Mg2+) plays a critical role in cellular metabolism and is considered to be an essential co-factor for more than 350 enzymes. As a result, alterations of Mg2+ homeostasis are associated with a broad range of clinical conditions, including those affecting the nervous system, such as glaucoma, Parkinson's disease, Alzheimer's disease, anxiety, depression, and intellectual disability (Wu, 2020).

Perhaps surprisingly, increasing brain Mg2+ through diet can enhance neuronal plasticity and memory performance of young and aged rodents, measured in a variety of behavioral tasks. In addition, elevated Mg2+ reduced cognitive deficits in a mouse model of Alzheimer's disease and enhanced the extinction of fear memories. These apparently beneficial effects have led to the proposal that dietary Mg2+ may have therapeutic value for patients with a variety of memory-related (Wu, 2020).

Despite the large number of potential sites of Mg2+ action in the brain, the memory-enhancing property in rodents has largely been attributed to increases in hippocampal synaptic density and the activity of N-methyl-D-aspartate glutamate receptors (NMDARs). Extracellular Mg2+ blocks the channel pore of the NMDAR (see Drosophila NMDA receptors) and thereby inhibits the passage of other ions. Importantly, prior neuronal depolarization, driven by other transmitter receptors, is required to release the Mg2+ block on the NMDAR and permit glutamate-gated Ca2+ influx. The NMDAR therefore plays an important role in neuronal plasticity as a potential Hebbian coincidence detector. Acute elevation of extracellular Mg2+ concentration ([Mg2+]e) within the physiological range (0.8-1.2 mM) can antagonize induction of NMDAR-dependent long-term potentiation. In contrast, increasing [Mg2+]e for several hours in neuronal cultures leads to enhancement of NMDAR mediated currents and facilitation of the expression of LTP. The enhancing effects of increased [Mg2+]e were also observed in vivo in the brain of rats fed with Mg2+-L-threonate. Hippocampal neuronal circuits undergo homeostatic plasticity to accommodate the increased [Mg2+]e by upregulating expression of NR2B subunit containing NMDARs. The higher density of hippocampal synapses with NR2B containing NMDARs are believed to compensate for the chronic increase in [Mg2+]e by enhancing NMDAR currents during burst firing. In support of this model, mice that are genetically engineered to overexpress NR2B exhibit enhanced hippocampal LTP and behavioral memory (Wu, 2020).

Olfactory memory in Drosophila involves a heterosynaptic mechanism driven by reinforcing dopaminergic neurons, which results in presynaptic depression of cholinergic connections between odor-activated mushroom body (MB) Kenyon cells (KCs) and downstream mushroom body output neurons (MBONs). In addition, olfactory information is conveyed to KCs by cholinergic transmission from olfactory projection neurons. Although it is conceivable that glutamate is delivered to the MB network via an as yet to be identified route, there is currently no obvious location for NMDAR-dependent plasticity in the known architecture of the cholinergic input or output layers. The fly therefore provides a potential model to investigate other mechanisms through which dietary Mg2+ might enhance memory (Wu, 2020).

The reinforcing effects of dopamine depend on the Dop1R D1-type dopamine receptor, which is positively coupled with cAMP production. Moreover, early studies in Drosophila identified the dunce and rutabaga encoded cAMP phosphodiesterase and type I Ca2+-stimulated adenylate cyclase, respectively, to be essential for olfactory memory. Studies in mammalian cells have shown that hormones or agents that increase cellular cAMP level often elicit a significant Na+-dependent extrusion of Mg2+ into the extracellular space. However, it is unclear whether Mg2+ extrusion plays any role in memory processing (Wu, 2020).

This study demonstrates that Drosophila long-term memory (LTM) can be enhanced with dietary Mg2+ supplementation. The unextended (uex) gene, which encodes a functional fly ortholog of the mammalian Cyclin M2 Mg2+-efflux transporter (CNNM) proteins, is critical for the memory enhancing property of Mg2+. UEX function in MB KCs is required for LTM and functional restoration of uex reveals the MB to be the key site of Mg2+-dependent memory enhancement. Chronically changing cAMP metabolism by introducing mutations in the dnc or rut genes alters the cellular localization of UEX. Moreover, mutating the conserved cyclic nucleotide-binding homology (CNBH) domain in UEX uncouples an essential role for uex from its function in memory. UEX-driven Mg2+ efflux is required for slow rhythmic maintenance of KC Mg2+ levels suggesting a potential role for Mg2+ flux in memory processing (Wu, 2020).

This study observed an enhancement of olfactory LTM performance when flies were fed for 4 days before training with food supplemented with 80 mM [Mg2+]. This result resembles that reported in rats, although longer periods of feeding were required to raise brain [Mg2+] to memory-enhancing levels. A difference in optimal feeding time may reflect the size of the animal and perhaps the greater bioavailability of dietary Mg2+ in Drosophila. Whereas Mg2+-L-threonate (MgT) was a more effective means of delivering Mg2+ than magnesium chloride in rats, a similar enhancement of memory performance was observed when flies were fed with magnesium chloride, magnesium sulfate, or MgT (Wu, 2020).

Elevating [Mg2+]e in the rat brain leads to a compensatory upregulation of expression of the NR2B subunit of the NMDAR and therefore an increase in the proportion of postsynaptic NR2B-containing NMDARs. This class of NMDARs have a longer opening time suggesting that this switch in subunit composition represents a homeostatic plasticity mechanism to accommodate for the increased NMDAR block imposed by increasing [Mg2+]e. Moreover, overexpression of NR2B in the mouse forebrain can enhance synaptic facilitation and learning and memory performance, supporting an increase in NR2B being an important factor in Mg2+-enhanced memory. However, even in the original in vitro study of Mg2+-enhanced synaptic plasticity, it was noted that NMDAR currents were insufficient to fully explain the observed changes (Wu, 2020).

NMDAR subunit loss-of-function studies in the Drosophila KCs did not impair regular or Mg2+-enhanced memory. Furthermore, no obvious change was detected in the levels of brain-wide expression of glutamate receptor subunits in Mg2+-fed flies. Although NMDAR activity has previously been implicated in Drosophila olfactory memory, the effects were mostly ascribed to function outside the MB. In addition, overexpressing Nmdar1 in all neurons, or specifically in all KCs, did not alter STM or LTM. Ectopic overexpression in the MB of an NMDARN631Q version, which cannot be blocked by Mg2+, impaired LTM. However, this mutation permits ligand-gated Ca2+ entry, without the need for correlated neuronal depolarization, which may perturb KC function in unexpected ways. It is perhaps most noteworthy that learning-relevant synaptic depression in the MB can be driven by dopaminergic teaching signals delivered to cholinergic output synapses from odor-responsive KCs to specific MBONs. It is conceivable that KCs receive glutamate, from a source yet to be identified, but there is currently no obvious place in the MB network for NMDAR-dependent plasticity. Evidence therefore suggests that normal and Mg2+-enhanced Drosophila LTM is independent of NMDAR signaling in KCs. In addition, MagFRET measurements indicate that Mg2+ feeding also increases the [Mg2+]i of αβ KCs by approximately 50 μM (Wu, 2020).

This study identified a role for uex, the single fly ortholog of the evolutionarily conserved family of CNNM-type Mg2+ efflux transporters. There are four distinct CNNM genes in mice and humans, five in C. elegans, and two in zebrafish. The uex locus produces four alternatively spliced mRNA transcripts, but all encode the same 834 aa protein. The precise role of CNNM proteins in Mg2+ transport is somewhat contentious. Some propose that CNNM proteins are direct Mg2+ transporters, whereas others favor that they function as sensors of intracellular Mg2+ concentration [Mg2+]i and/or regulators of other Mg2+ transporters. This study found that ectopic expression of Drosophila UEX enhances Mg2+ efflux in HEK293 cells and that endogenous UEX limits [Mg2+]i in αβ KCs in the fly brain. Therefore, if UEX is not itself a Mg2+ transporter, it must be able to interact effectively with human Mg2+ efflux transporters and to influence Mg2+ extrusion in Drosophila. Since UEX is the only CNNM protein in the fly, it may serve all the roles of the four individual mammalian CNNMs. However, the ability of mouse CNNM2 to restore memory capacity to uex mutant flies suggests that the memory-relevant UEX function can be substituted by that of CNNM2 (Wu, 2020).

Interestingly, none of the disease-relevant variants of CNNM2 were able to complement the memory defect of uex mutant flies. The CNNM2 T568I variant substitutes a single amino acid in the second CBS domain. The oncogenic protein tyrosine phosphatases of the PRL (phosphatase of regenerating liver) family bind to the CBS domains of CNNM2 and CNNM3 and can inhibit their Mg2+ transport function. It will therefore be of interest to test the role of the UEX CBS domains and whether fly PRL-1 regulates UEX activity (Wu, 2020).

RNA-seq analysis reveals that uex is strongly expressed in the larval and adult fly digestive tract and nervous systems, as well as the ovaries suggesting that many uex mutations will be pleiotropic. The uexΔ allele, which deletes 272 amino acids (including part of the second CBS and the entire CNBH domain) from the UEX C-terminus, results in developmental lethality when homozygous, demonstrating that uex is an essential gene. Mammalian CNNM4 is localized to the basolateral membrane of intestinal epithelial cells. There it is believed to function in transcellular Mg2+ transport by exchanging intracellular Mg2+ for extracellular Na+ following apical entry through TRPM7 channels. Lethality in Drosophila could therefore arise from an inability to absorb sufficient Mg2+ through the larval gut. However, neuronally restricted expression of uexRNAi with elav-GAL4 also results in larval lethality, suggesting UEX has an additional role in early development of the nervous system, like CNNM2 in humans and zebrafish. Perhaps surprisingly, flies carrying homozygous or trans-heterozygous combinations of several hypomorphic uex alleles have defective appetitive and aversive memory performance, yet they seem otherwise unaffected (Wu, 2020).

Genetically engineering the uex locus to add a C-terminal HA tag to the UEX protein allowed localization of its expression in the brain. Labeling is particularly prominent in all major classes of KCs. Restricting knockdown of uex expression to all αβ KCs of adult flies, or even just the αβc subset reproduced the LTM defect. The LTM impairment was evident if uexRNAi expression in αβ neurons was restricted to adult flies, suggesting UEX has a more sustained role in neuronal physiology. In contrast, knocking down uex expression in either the αβs or α'β' neurons did not impair LTM. Activity of α'β' neurons is required after training to consolidate appetitive LTM, whereas αβc and αβs KC output, together and separately, is required for its expression. Therefore, observing normal LTM performance in flies with uex loss-of-function in αβs and α'β' neurons argues against a general deficiency of αβ neuronal function when manipulating uex (Wu, 2020).

Dietary Mg2+ could not enhance the defective LTM performance of flies that were constitutively uex mutant, or harbored αβ KC-restricted uex loss-of-function. However, expressing uex in the αβ KCs of uex mutant flies restored the ability of Mg2+ to enhance performance. Therefore, the αβ KCs are the cellular locus for Mg2+-enhanced memory in the fly (Wu, 2020).

It perhaps seems counterintuitive that UEX-directed magnesium efflux is required in KCs to support the memory-enhancing effects of Mg2+ feeding, when dietary Mg2+ elevates KC [Mg2+]i. At this stage, it can only be speculated as to why this is the case. It is assumed that the brain and αβ KCs, in particular, have to adapt in a balanced way to the higher levels of intracellular and extracellular Mg2+ that result from dietary supplementation. Live-imaging of KC [Mg2+]i in wild-type and uex mutant brains suggests that UEX-directed efflux is likely to be an essential factor in the active, and perhaps stimulus-evoked, homeostatic maintenance of these elevated levels (Wu, 2020).

A number of mammalian cell-types extrude Mg2+ in a cAMP-dependent manner, a few minutes after being exposed to β-adrenergic stimulation. The presence of a CNBH domain suggests that UEX and CNNMs could be directly regulated by cAMP. The importance of the CNBH was tested by introducing an R622K amino acid substitution that should block cAMP binding in the UEX CNBH. This subtle mutation abolished the ability of the uexR622K transgene to restore LTM performance to uex mutant flies. CRISPR was used to mutate the CNBH in the native uex locus. Although deleting the CNBH from CNNM4 abolished Mg2+ efflux activity, flies homozygous for the uexT626NRR lesion were viable, demonstrating that they retain a sufficient level of UEX function. However, these flies exhibited impaired immediate and long-term memory. In addition, the performance of uexT626NRR flies could not be enhanced by Mg2+ feeding. These data demonstrate that an intact CNBH is a critical element of memory-relevant UEX function. Binding of clathrin adaptor proteins to the CNNM4 CNBH has been implicated in basolateral targeting, suggesting that uexT626NRR might be inappropriately localized in KCs. Furthermore, KC expression of the CNNM2 E122K mutant variant, which retains residual function but has a trafficking defect, did not restore the uex LTM defect (Wu, 2020).

Although it has been questioned whether the CNNM2/3 CNBH domains bind cyclic nucleotides, this study found that FSK evoked an increase in αβ KC [Mg2+]i that was sensitive to uex mutation, and that UEX::HA was mislocalized in rut2080 adenylate cyclase and dnc1 phosphodiesterase learning defective mutant flies. Whereas UEX::HA label was evenly distributed in γ, αβc, and αβs KCs in wild-type flies, UEX::HA label was diminished in the γ and αβs KCs and was stronger in αβc neurons in rut2080 and dnc1 mutants. The chronic manipulations of cAMP in the mutants are therefore consistent with cAMP impacting UEX localization, perhaps by interacting with the CNBH. In addition, altered UEX localization may contribute to the memory defects of rut2080 and dnc1 flies (Wu, 2020).

The physiological data using Magnesium Green in mammalian cell culture and the genetically encoded MagIC reporter in αβ KCs demonstrate that fly UEX facilitates Mg2+ efflux. Stimulating the fly brain with FSK evoked a greater increase in αβ KC [Mg2+]i in uex mutant brains than in wild-type controls which provides the first evidence that UEX limits a rise in [Mg2+]i in Drosophila KCs. MagIC recordings also revealed a slow oscillation (centered around 0.015 Hz, approximately once a minute) of αβ KC [Mg2+]i that was dependent on UEX. The physiological function of this [Mg2+]i fluctuation is not yet understood although it likely reflects a homeostatic systems-level property of the cells. Biochemical oscillatory activity plays a crucial role in many aspects of cellular physiology. Most notably, circadian timed fluctuation of [Mg2+]i links dynamic cellular energy metabolism to clock-controlled translation through the Mg2+ sensitive mTOR (mechanistic target of rapamycin) pathway. It is therefore possible that slow Mg2+ oscillations could unite roles for cAMP, UEX, energy flux, and mTOR-dependent translation underlying LTM-relevant synaptic plasticity (Wu, 2020).

Go to top

Transneuronal Dpr12/DIP-delta interactions facilitate compartmentalized dopaminergic innervation of Drosophila mushroom body axons

Bornstein, B., Meltzer, H., Adler, R., Alyagor, I., Berkun, V., Cummings, G., Reh, F., Keren-Shaul, H., David, E., Riemensperger, T. and Schuldiner, O. (2021). Embo J: e105763. PubMed ID: 33847376

Transneuronal Dpr12/DIP-delta interactions facilitate compartmentalized dopaminergic innervation of Drosophila mushroom body axons

The mechanisms controlling wiring of neuronal networks are not completely understood. The stereotypic architecture of the Drosophila mushroom body (MB) offers a unique system to study circuit assembly. The adult medial MB γ-lobe is comprised of a long bundle of axons that wire with specific modulatory and output neurons in a tiled manner, defining five distinct zones. The immunoglobulin superfamily protein Dpr12 is cell-autonomously required in γ-neurons for their developmental regrowth into the distal γ4/5 zones, where both Dpr12 and its interacting protein, DIP-δ, are enriched. DIP-δ functions in a subset of dopaminergic neurons that wire with γ-neurons within the γ4/5 zone. During metamorphosis, these dopaminergic projections arrive to the γ4/5 zone prior to γ-axons, suggesting that γ-axons extend through a prepatterned region. Thus, Dpr12/DIP-γ transneuronal interaction is required for γ4/5 zone formation. This study sheds light onto molecular and cellular mechanisms underlying circuit formation within subcellular resolution (Bornstein, 2021).

Understanding of the development of complex neural circuits remains largely unknown. Specifically, how long axons can make en passant synapses with different partners in a stereotypic manner is not well understood. The unique development and morphology of the Drosophila MB γ-lobe, combined with the comprehensive genetic power of the fly, offer an excellent opportunity to dissect mechanisms required for wiring of complex neural networks, and specifically mechanisms that drive zonation within axonal bundles to allow for stereotypic localized innervation by distinct populations of neurons. This study has identified a molecular mechanism that mediates neuron-neuron interactions which subsequently promote the formation of stereotypic circuits that define subcellular axonal zones (Bornstein, 2021).

The adult γ-lobe is divided into zones (also known as compartments) due to specific and localized innervations by extrinsic MB neurons including MBONs and DANs. This study shows that the interaction between two IgSF proteins, Dpr12 on γ-KCs and DIP-δ on PAM-DANs, underlies the formation of the MB γ4/5 zones. Within each zone, input from DANs can modify synaptic strength between the KC and MBON to provide specific valence to sensory information. Based on the results presented in this study, it is speculated that various specific combinations of adhesion molecules may mediate target recognition events that occur between predefined synaptic pairs in other MB zones as well. γ-neurons express a broad spectrum of IgSFs in tight temporal regulation, highlighting their potential role in circuit formation. However, many adhesion molecules, including Dpr/DIPs, can form promiscuous interactions, making their analyses challenging. Future studies could use CRISPR/Cas9 technology to generate multi-gene mutations to further explore the adhesion code required for zone/compartment formation (Bornstein, 2021).

This study used the interaction between Dpr12 and DIP-δ to study the development of the γ4/5 zones. Developmental analyses have concluded that DIP-δ-expressing PAM-DANs arrive to the region of the γ4/5 zones before γ-axons. Interestingly, DIP-δ localization experiments suggest that in dpr12 mutant animals, PAM-DANs arrive to the right place (the future γ4/5 zones) during larval development, maintain their processes at least until 48 h APF, but eventually (at a yet unknown time point) eliminate or remodel their γ4/5 innervations, while maintaining and even strengthening/broadening other connections in this vicinity. Therefore, it is attractive to speculate that γ-axons extend into a prepatterned lobe. More studies comparing the development of other compartment-specific DANs as well as MBONs are however required (Bornstein, 2021).

This study demonstrates that Dpr12 is cell-autonomously required in γ-KCs, while DIP-δ is required in PAM-DANs for the formation of the γ4/5 zones. This is the first case in which a Dpr molecule was shown to be cell-autonomously required for correct wiring. However, the precise molecular mechanism by which the Dpr12-DIP-δ interaction mediates formation of the γ4/5 zones, or, in fact, how any wiring by Dpr-DIPs is achieved, is yet to be determined. The robust phenotype associated with loss of the Dpr12-DIP-δ interaction offers an excellent opportunity to delve into the mechanistic basis, which could potentially shed light on similar mechanisms in the visual system and the NMJ. Further research should focus on several critical questions that remain unresolved: (1) Why do the γ-axons stop prematurely? That γ-axons stall at the γ3-γ4 junction when the Dpr12-DIP-δ interaction is perturbed - which at least in principle is expected to be of adhesive nature - is not clear. One possibility is that axon growth into the γ4/5 zones depends on Dpr12-DIP-δ interaction either because they overcome a yet undiscovered inhibitory signal, or because they are positively required for the progression of the growth cone. Alternatively, Dpr12-DIP-δ interaction could be important for the stabilization of the connections between γ-axons and PAM-DAN processes to result in the formation of the γ4/5 zones. At 48 h APF, the large majority of dpr12 mutant γ-axons do not innervate the γ4/5 zones, arguing against the stability hypothesis; (2) What are the signaling pathways that mediate Dpr/DIP targeting recognition? None of the Dprs or DIPs contain a large intracellular domain that is capable of signaling. Identifying the potential co-receptor/s is a critical step in gaining a mechanistic understanding of axon targeting whether in the visual, motor or MB circuits. The results that DIP-α can replace DIP-δ suggest that signaling may be conserved between different Dpr-DIP pairs; (3) What is the significance of the GPI anchor? Many of the Dprs and DIPs are predicted to be GPI-anchored proteins, suggesting that they can be cleaved to create a secreted soluble form. Whether this is an important step in targeting has not yet been investigated. Interestingly, the vertebrate homologs of the DIPs, the IgLON subfamily, are GPI-anchored proteins that were shown to be cut by metalloproteinases to promote axonal outgrowth (Bornstein, 2021).

Expression patterns of Dpr and DIP molecules in the NMJ and visual system suggested a model where these molecules instruct target cell specificity. Recent loss-of-function experiments strengthened this target specificity hypothesis, as the DIP-α-Dpr10 interaction was shown to be important for motoneuron innervation of specific larval and adult muscles, and DIP-α-Dpr10/Dpr6 interactions for specific layer targeting in the visual system. The current results suggest that mechanisms used to target axons and dendrites to specific cell types or layers may be further implicated to orchestrate the wiring of long axons to different pre- and postsynaptic partners along their route and thus the formation of axonal zones (Bornstein, 2021).

This paper has described the interaction between two IgSF proteins mediates transneuronal communication that is required for proper wiring within specific zones of the Drosophila MB. The anatomical organization of the MB suggests that these interactions may provide target specificity for the long KC axon, while it forms en passant synapses with different targets along its length. While the existence of such wiring architecture is known from invertebrates such as Drosophila and C. elegans, long axons making distinct yet stereotypic en passant connections are not widely described in vertebrates. Given the existence of long axons, that travel through dense neuropil structures, such as mossy fibers in the hippocampus, cholinergic axons in the basal forebrain, and parallel fibers in the cerebellar cortex, it is posited that this type of connectivity exists in vertebrates but has not yet been described in detail due to technological limitations that are likely to be resolved soon. Pairwise IgSF-mediated molecular interactions are conserved in vertebrates and invertebrates, implying similar mechanisms to dictate axon and dendrite targeting of subcellular neurite zones in other organisms (Bornstein, 2021).

Go to top

Hierarchical architecture of dopaminergic circuits enables second-order conditioning in Drosophila

Yamada, D., Bushey, D., Li, F., Hibbard, K. L., Sammons, M., Funke, J., Litwin-Kumar, A., Hige, T. and Aso, Y. (2023). Elife 12. PubMed ID: 36692262

Dopaminergic neurons with distinct projection patterns and physiological properties compose memory subsystems in a brain. However, it is poorly understood whether or how they interact during complex learning. This study identified a feedforward circuit formed between dopamine subsystems and showed that it is essential for second-order conditioning, an ethologically important form of higher-order associative learning. The Drosophila mushroom body comprises a series of dopaminergic compartments, each of which exhibits distinct memory dynamics. A slow and stable memory compartment can serve as an effective 'teacher' by instructing other faster and transient memory compartments via a single key interneuron, which was identified by connectome analysis and neurotransmitter prediction. This excitatory interneuron acquires enhanced response to reward-predicting odor after first-order conditioning and, upon activation, evokes dopamine release in the 'student' compartments. These hierarchical connections between dopamine subsystems explain distinct properties of first- and second-order memory long known by behavioral psychologists (Yamada, 2023).

Go to top

Mushroom body input connections form independently of sensory activity in Drosophila melanogaster

Hayashi, T. T., MacKenzie, A. J., Ganguly, I., Ellis, K. E., Smihula, H. M., Jacob, M. S., Litwin-Kumar, A. and Caron, S. J. C. (2022). Curr Biol. PubMed ID: 35977547

Associative brain centers, such as the insect mushroom body, need to represent sensory information in an efficient manner. In Drosophila melanogaster, the Kenyon cells of the mushroom body integrate inputs from a random set of olfactory projection neurons, but some projection neurons-namely those activated by a few ethologically meaningful odors-connect to Kenyon cells more frequently than others. This biased and random connectivity pattern is conceivably advantageous, as it enables the mushroom body to represent a large number of odors as unique activity patterns while prioritizing the representation of a few specific odors. How this connectivity pattern is established remains largely unknown. This study tested whether the mechanisms patterning the connections between Kenyon cells and projection neurons depend on sensory activity or whether they are hardwired. A large number of mushroom body input connections were mapped in partially anosmic flies-flies lacking the obligate odorant co-receptor Orco-and in wild-type flies. Statistical analyses of these datasets reveal that the random and biased connectivity pattern observed between Kenyon cells and projection neurons forms normally in the absence of most olfactory sensory activity. This finding supports the idea that even comparatively subtle, population-level patterns of neuronal connectivity can be encoded by fixed genetic programs and are likely to be the result of evolved prioritization of ecologically and ethologically salient stimuli (Hayashi, 2022).

Go to top

A circular RNA Edis-Relish-castor axis regulates neuronal development in Drosophila

Liu, W., Liang, W., Xiong, X. P., Li, J. L. and Zhou, R. (2022). A circular RNA Edis-Relish-castor axis regulates neuronal development in Drosophila. PLoS Genet 18(10): e1010433. PubMed ID: 36301831

Circular RNAs (circRNAs) are a new group of noncoding/regulatory RNAs that are particularly abundant in the nervous system, however, their physiological functions are underexplored. This study reports that the brain-enriched circular RNA Edis (Ect4-derived immune suppressor) plays an essential role in neuronal development in Drosophila. Depletion of Edis in vivo causes defects in axonal projection patterns of mushroom body (MB) neurons in the brain, as well as impaired locomotor activity and shortened lifespan of adult flies. In addition, it was found that the castor gene, which encodes a transcription factor involved in neurodevelopment, is upregulated in Edis knockdown neurons. Notably, castor overexpression phenocopies Edis knockdown, and reducing castor levels suppresses the neurodevelopmental phenotypes in Edis-depleted neurons. Furthermore, chromatin immunoprecipitation analysis reveals that the transcription factor Relish, which plays a key role in regulating innate immunity signaling, occupies a pair of sites at the castor promoter, and that both sites are required for optimal castor gene activation by either immune challenge or Edis depletion. Lastly, Relish mutation and/or depletion can rescue both the castor gene hyperactivation phenotype and neuronal defects in Edis knockdown animals. It is concluded that the circular RNA Edis acts through Relish and castor to regulate neuronal development (Liu, 2022).

Go to top

Self-avoidance alone does not explain the function of Dscam1 in mushroom body axonal wiring

Dong, H., Guo, P., Zhang, J., Wu, L., Fu, Y., Li, L., Zhu, Y., Du, Y., Shi, J., Zhang, S., Li, G., Xu, B., Bian, L., Zhu, X., You, W., Shi, F., Yang, X., Huang, J. and Jin, Y. (2022). Curr Biol. PubMed ID: 35659864

Alternative splicing of Drosophila Dscam1 into 38,016 isoforms provides neurons with a unique molecular code for self-recognition and self-avoidance. A canonical model suggests that the homophilic binding of identical Dscam1 isoforms on the sister branches of mushroom body (MB) axons supports segregation with high fidelity, even when only a single isoform is expressed. This study generated a series of mutant flies with a single exon 4, 6, or 9 variant, encoding 1,584, 396, or 576 potential isoforms, respectively. Surprisingly, most of the mutants in the latter two groups exhibited obvious defects in the growth, branching, and segregation of MB axonal sister branches. This demonstrates that the repertoires of 396 and 576 Dscam1 isoforms were not sufficient for the normal patterning of axonal branches. Moreover, reducing Dscam1 levels largely reversed the defects caused by reduced isoform diversity, suggesting a functional link between Dscam1 expression levels and isoform diversity. Taken together, these results indicate that canonical self-avoidance alone does not explain the function of Dscam1 in MB axonal wiring (Dong, 2022).

Go to top

Hacking brain development to test models of sensory coding

Ahmed, M., Rajagopalan, A. E., Pan, Y., Li, Y., Williams, D. L., Pedersen, E. A., Thakral, M., Previero, A., Close, K. C., Christoforou, C. P., Cai, D., Turner, G. C. and Clowney, E. J. (2023). bioRxiv. PubMed ID: 36747712

Animals can discriminate myriad sensory stimuli but can also generalize from learned experience. One can probably distinguish the favorite teas of your colleagues while still recognizing that all tea pales in comparison to coffee. Tradeoffs between detection, discrimination, and generalization are inherent at every layer of sensory processing. During development, specific quantitative parameters are wired into perceptual circuits and set the playing field on which plasticity mechanisms play out. A primary goal of systems neuroscience is to understand how material properties of a circuit define the logical operations- computations--that it makes, and what good these computations are for survival. A cardinal method in biology-and the mechanism of evolution--is to change a unit or variable within a system and ask how this affects organismal function. This study used of knowledge of developmental wiring mechanisms to modify hard-wired circuit parameters in the Drosophila melanogaster mushroom body and assess the functional and behavioral consequences. By altering the number of expansion layer neurons (Kenyon cells) and their dendritic complexity, it was found that input number, but not cell number, tunes odor selectivity. Simple odor discrimination performance is maintained when Kenyon cell number is reduced and augmented by Kenyon cell expansion (Ahmed, 2023).

Go to top

36103878

Animals use prior experience to assign absolute (good or bad) and relative (better or worse) value to new experience. These learned values guide appropriate later decision making. Even though understanding of how the valuation system computes absolute value is relatively advanced, the mechanistic underpinnings of relative valuation are unclear. Thid study uncovered mechanisms of absolute and relative aversive valuation in Drosophila. Three types of punishment-sensitive dopaminergic neurons (DANs) respond differently to electric shock intensity. During learning, these punishment-sensitive DANs drive intensity-scaled plasticity at their respective mushroom body output neuron (MBON) connections to code absolute aversive value. In contrast, by comparing the absolute value of current and previous aversive experiences, the MBON-DAN network can code relative aversive value by using specific punishment-sensitive DANs and recruiting a specific subtype of reward-coding DANs. Behavioral and physiological experiments revealed that a specific subtype of reward-coding DAN assigns a "better than" value to the lesser of the two aversive experiences. This study therefore highlights how appetitive-aversive system interactions within the MB network can code and compare sequential aversive experiences to learn relative aversive value (Villar, 2022).

Anatomically discrete dopaminergic neurons (DANs) in Drosophila and mice provide either positive or negative teaching signals). In flies, these different DANs project to unique compartments of the mushroom body (MB), a central brain structure essential for olfactory learning and memory as well as several goal-directed behaviors. DANs from the protocerebral posterior lateral 1 (PPL1) cluster projecting to the vertical and proximal horizontal lobes of the MB relay punishment and signal negative value during learning.Many DANs from the protocerebral anterior medial (PAM) cluster projecting to the horizontal lobes of the MB assign positive reward value during learning. Sparse activation within the ~4,000 MB intrinsic Kenyon cells (KCs), which indirectly receive odorant information from sensory neurons in the periphery, provides the specificity of olfactory memories. KCs synapse onto mushroom body output neurons (MBONs), which project into downstream structures to drive (for most of them) approach or avoidance behavior. Some MBONs are synaptically interconnected, providing cross-excitation or -inhibition between MB compartments. Lastly, many MBONs make feedback or feedforward synapses outside the MB onto DAN dendrites. During olfactory associative learning, specific DANs releasing dopamine in individual MB compartments depress synaptic strengths between sparse odor-activated KCs and the MBONs whose dendrites reside within the relevant compartments. As a result, learning-induced plasticity within different MB compartments reconfigures the MBON ensemble output signal to promote either learned approach or avoidance behavior (Villar, 2022).

Olfactory aversive learning reduces odor drive of approach-directing MBONs, hence tilting the MBON network toward promoting odor avoidance. By contrast, appetitive learning reduces odor drive of avoidance-directing MBONs, leaving the network in a configuration promoting odor approach (Villar, 2022).

Flies can perceive, learn, and compare differences in the intensity of punishment and adapt their behavior accordingly. This study combined genetic interventions with behavioral analyses, anatomical characterization, and in vivo two-photon calcium imaging to investigate the detailed circuit requirements that allow flies to write and compare olfactory aversive memories of different intensities during learning to promote appropriate value-based choices. Aversive PPL1 DANs show differential responses to electric shock punishment of varying intensity. As a result, the intensity of shock reinforcement correlates with the magnitude of learning-driven plasticity at the corresponding KC to MBON junctions. Using a specific behavioral paradigm in which flies associate three odors with 0, 60, and 30 V punishment, respectively, this study identified the circuits involved in coding relative aversive value. Loss-of-function screening revealed a role for specific aversive DANs, in addition to the rewarding PAM-β'2aγ5n DANs, to learn relative aversive value. Recording from PAM-β'2aγ5n DANs during learning revealed these neurons to signal relative aversive value by increasing their responsiveness when the odor-low shock association is better than a previous odor-high shock association. Recording from three MBONs presynaptic to PAM-β'2aγ5n DANs revealed a positive difference in the odor responses of MBON-γ2α'1 between current and previous aversive experiences. This increased responsiveness of cholinergic MBON-γ2α'1 likely provides excitatory input necessary to drive the PAM-β'2aγ5n DANs 'better than' value signal for the less aversive experience, and they thereby learn relative aversive value. In support of this model, optogenetic activation of PAM-β'2aγ5n reward DANs, during learning, assigns a relative “better than” value to one of two identical odor-punishment associations.

This study addresses how animals assign absolute aversive value during learning and how they compare and ascribe relative aversive value information to consecutive negative experiences for them to make appropriate value-based decisions afterwards. Using the fruit fly Drosophila permitted a cellular resolution view of how the interaction between the appetitive and aversive DAN systems, within the MB network, is at the heart of the mechanistic underpinnings that compute a relative aversive value teaching signal. This work also indicates that coding of a relative 'worse than' aversive value likely involves different circuit mechanisms to those for 'better than' but that there may be some overlap (Villar, 2022).

PPL1 DANs reinforce a range of aversive memories with differing strength and persistence. The current data provide new insight into the functional diversity of these anatomically discrete DANs. This study found that individual aversively reinforcing PPL1-γ1pedc, PPL1-γ2α'1, and PPL1-α2α'2 DANs exhibit different intensity response profiles when flies were exposed to a series of shock voltages (Villar, 2022).

Importantly, the strength of their responses to electric shocks strongly correlated with the magnitude of plasticity of the odor-evoked responsiveness of their corresponding MBONs after differential conditioning. These results indicate that absolute aversive value is assigned to odors in different ways in the γ1pedc, γ2α'1, and α2α'2 MB compartments, consistent with the conclusion of a prior study that artificially activated individual DANs. Of note, this study did not observe significant shock responses in PPL1-α2α'2 DANs. These results are also in accordance with the absence of odor-evoked changes in the corresponding MBON-α2sc immediately after training (Figures 2G and 2H) and a lack of reinforcing properties when pairing artificial activation of PPL1-α2α'2 DANs with an odor. In addition, learning-dependent depression of odor responses in MBON-α2sc has been reported to be most relevant for expression of later forms of memory (Villar, 2022).

It was observed that the stronger the aversive experience, the greater the PPL1-γ1pedc DAN-driven depression of the CS+-evoked response of MBON-γ1ped>>αβ. Feedforward GABAergic inhibition from MBON-γ1ped>αβ to the primary axon of MBON-γ5β'2a is therefore reduced in a graded manner by aversive conditioning. MBON-γ5β'2a should therefore display a proportional increase in its CS+-evoked response to drive learned avoidance behavior (Villar, 2022).

These experiments uncovered a very different effect of absolute aversive conditioning at the MBON-γ2α'1 junction. Although the PPL1-γ2α'1 DANs were significantly triggered by shocks ≥30 V, their responses were comparable at all voltages between 30 and 90 V. Moreover, aversive conditioning did not significantly depress the CS+ responses of MBON-γ2α'1. Instead, it was observed that the responses to the CS- odor were specifically increased, and the CS- CS+ differential responses were correlated with the intensity of the shocks applied. The data therefore suggest that any odor that follows the CS+ with ≥45 V presentation during training gains the capacity to drive more activity in the cholinergic MBON-γ2α'1. In addition, the recordings indicate that the more aversive the first experience is, the stronger the cholinergic MBON-γ2α'1 activity will be to the subsequent 'better than' experience. These data reveal a key role for MBON-γ2α'1 in coding relative aversive value (Villar, 2022).

This study found that output from PAM-β'2aγ5n DANs was critical during the odor Z + 30 V presentation for relative aversive learning. These DANs receive direct excitatory cholinergic input from MBON-γ2α'1, and it is proposed that the strength of this excitation is key for the flies to assign a 'better than' reward value to the lesser of the two aversive experiences. As mentioned above, when odor Y is paired with 60 V shock in a differential conditioning assay, the CS- responses of MBON-γ2α'1 become elevated. This means that when Y + 60 V is followed by Z + 30 V, the Z odor will more strongly drive MBON-γ2α'1 and as a result will activate the PAM-β'2aγ5n DANs. In effect, it is hypothesized that any odor that follows a Y + 60 V experience is predisposed to be judged as 'better than,' unless it is itself accompanied by 60 V or a greater voltage. Analyses that subtracted MBON-γ2α'1 odor-evoked responses are entirely consistent with this model. Odor-driven activity of MBON-γ2α'1 is greater during the first period of the following Z + 30 V experience than during the same period (just after the first shock) of the prior Y + 60 V experience. Critically, this is also the time period during which an elevation of PAM-β'2aγ5n DAN activity was observed. It is speculated that the first of the 30 V shocks somehow further releases the PAM-β'2aγ5n DAN activity to be fully driven by MBON-γ2α'1, perhaps as a release of feedforward inhibition in the MBON-γ1ped>>αβ to MBON-γ5β'2a to PAM-β'2aγ5n DAN pathway (Villar, 2022).

The results and proposed models of PAM-β'2aγ5n DANs providing a 'better than' reward signal are in accordance with previous reports that PAM-β'2aγ5n activation provides appetitive reinforcement (Villar, 2022).

Are there limits to comparable aversive memories? Individual PPL1-DAN subtypes have different thresholds for activation, and intensity-dependent plasticity in their corresponding MBON junctions have similar thresholds. It is noted that these thresholds seem reflected in the range of comparisons that flies can make in a relative choice between different aversive memories, which point toward a threshold and a difference between voltages of 30 V as being optimal to efficiently estimate a relative difference. In the recordings, 30 V was the threshold for observing shock-evoked responses in PPL1-γ1pedc and PPL1-γ2α'1, but it did not trigger PPL1-α2α'2. In addition, 30 V produced significant plasticity of MBON-γ1ped>>αβ odor responses, but plasticity was not evident in MBON-γ2α'1 responses until 45 V. Thus, perhaps every odor paired with a voltage of ≤30 V is considered to be 'not so bad,' because it only depresses the GABAergic MBON-γ1ped>αβ responses and not the cholinergic MBON-γ2α'1 responses, thereby leaving CS+ odor-driven excitation of PAM-β'2aγ5n DANs from these MBONs. Although flies can differentiate between stronger aversive memories such as 90 versus 60 V, their relative choice performances are less good than 60 versus 30 V (Villar, 2022).

While significant shock responses were not observed for PPL1-α2α'2 DANs, a role was found for these neurons during learning of relative aversive value (Figure 4). MBON-γ1ped>αβ is GABAergic and is connected to PPL1-α2α'2 DANs. It is therefore possible that repeated pairing of odor Y + 60 V electric shocks (or anything above their threshold) during relative training induces enough CS+-evoked depression at MBON-γ1ped>αβ to release inhibition in PPL1-α2α'2 DANs while pairing the odor Z with 30 V shocks. The resulting plasticity in MBON-α2sc could explain the requirement of αβ surface and αβ core KCs during a relative choice between Y60 versus Z30 (Villar, 2022).

These results show that learning a relative 'better than' aversive value requires an interplay between aversively reinforcing PPL1 DANs modulating KC-MBON connections, which provide feedforward and recurrent feedback input that determines the activity of specific subtypes of rewarding PAM DANs. These results support long-held models in both vertebrates and invertebrates, suggesting that learning requires critical interactions between appetitive and aversive reinforcement systems. In the fly, and likely also in mammals, this process relies on opposing populations of DANs providing predictive signals needed to compare current and previous experience to assign (and update) both absolute and relative value to stimuli during learning. For instance, aversive memory extinction and reversal learning require the reward system in both vertebrates and invertebrates (Villar, 2022).

In all these cases, stimuli that represent the absence of a punishment are rewarded. In humans, the ventral striatum, targeted by numerous DA inputs from the ventral tegmental area (VTA) providing rewarding information, is essential to compare aversive experiences of different intensities. In the orbitofrontal cortex, relative coding of aversive (but also appetitive) experiences seem to require overlapping neuronal ensembles to select a preferred option and promote appropriate economical decisions in a specific spatial and temporal context. In the dopaminergic system, reward is also computed in a relative manner to broadcast value signals in different brain regions (Villar, 2022).

These DANs from the VTA and substantia nigra compute a prediction error to signal positive, but also negative, value. A similar value prediction error calculation has not yet been demonstrated experimentally in the fly. Instead, results from several studies in the fly suggest that errors are registered in the MB network by the action of DANs that signal the opposing value. The experiments carried out in this study suggest that a similar interplay between opposing populations of DANs, and plasticity at different MBON junctions in the MB network, permits computation of relative aversive value (or difference) between a prior and a new aversive experience. Combined with previous work and current computational models, the data provide key features of how the appetitive-aversive system interactions in the MB network using heterogeneous DANs can compare previous and current experience to 'pre-compute' a relative value during learning that facilitates future value-based decisions (Villar, 2022).

Go to top

Traip controls mushroom body size by suppressing mitotic defects

O'Neill, R. S. and Rusan, N. M. (2022). Development 149(7). PubMed ID: 35297981

Microcephaly is a failure to develop proper brain size and neuron number. Mutations in diverse genes are linked to microcephaly, including several with DNA damage repair (DDR) functions; however, it is not well understood how these DDR gene mutations limit brain size. One such gene is TRAIP, which has multiple functions in DDR. This study characterized the Drosophila TRAIP homolog nopo, hereafter traip, and found that traip mutants (traip-) have a brain-specific defect in the mushroom body (MB). traip- MBs were smaller and contained fewer neurons, but no neurodegeneration, consistent with human primary microcephaly. Reduced neuron numbers in traip- were explained by premature loss of MB neuroblasts (MB-NBs), in part via caspase-dependent cell death. Many traip- MB-NBs had prominent chromosome bridges in anaphase, along with polyploidy, aneuploidy or micronuclei. Traip localization during mitosis is sufficient for MB development, suggesting that Traip can repair chromosome bridges during mitosis if necessary. The results suggest that proper brain size is ensured by the recently described role for TRAIP in unloading stalled replication forks in mitosis, which suppresses DNA bridges and premature neural stem cell loss to promote proper neuron number (O'Neill, 2022).

This study in Drosophila shows that traip- shares several characteristics with human microcephaly mutants. First, the traip- phenotype is highly brain specific, with body defects being rare. Second, the traip- MB phenotype is developmental rather than neurodegenerative, reflecting a primary rather than secondary microcephaly-like disorder. Finally, as with many human microcephaly genes, traip functions to promote NPC proliferation and survival. Thus, traip- represents a powerful new disease model for understanding the etiological mechanisms underlying microcephaly (O'Neill, 2022).

TDespite their ubiquitous expression, mutations in microcephaly genes primarily affect the cerebral cortex in humans. Similarly, both traip and the DDR microcephaly gene MCPH1 are ubiquitously expressed in Drosophila, yet the MB is the only adult structure affected in their mutants. Although many tissues can make up for lost cells via compensatory proliferation, no such process appears to exist for replacing lost NPCs. Additionally, whereas most NBs have a limited window of proliferation, MB-NBs divide continuously from embryogenesis into late pupal stages, potentially allowing more accumulation of rare or small effects over many cell cycles. Thus, it is speculated that mutations in microcephaly genes likely affect all CB-NBs to some degree, but the MB-NBs are especially sensitive to these mutations as a consequence of their relatively prolonged period of proliferation. It is speculated that a similar explanation, including a prolonged period of rapid proliferation and lack of compensatory proliferation, may account for the sensitivity of the human cortex to microcephaly gene mutation (O'Neill, 2022).

This work provides the first link between a known function of Traip and proper brain development. Interphase nuclear localization is not required for Traip function, suggesting that Traip interphase functions are dispensable for MB-NB survival under normal conditions. Instead, it was discovered the presence of mitotic DNA bridges, sensitivity to inter-strand crosslinking agents, and RING domain dependence, consistent with the well-established role of TRAIP in unloading stalled forks to initiate repair. Furthermore, GFP::TraipΔNLS rescue experiments suggest either that Traip primarily performs this unloading function during mitosis, or else that Traip normally functions during interphase but is able to unload stalled forks during mitosis if necessary. Alternatively, it cannot be ruled out that there may be residual GFP::TraipΔNLS in the nucleus that allows continued function during interphase, or else that nuclear localization of Traip is not required for an interphase function. It is surmised that traip- MB-NBs have stalled replication machinery that remains loaded throughout mitosis, preventing mitotic DNA synthesis repair and proper sister chromatid segregation. As anaphase proceeds, attached sister chromatids are pulled to opposite poles and they form UFBs as the under-replicated DNA is stretched out between them. These bridges could be physically broken, leading to chromosome fragmentation, generating aneuploidy or micronuclei and causing nuclear deformations in daughter cells. Alternatively, persistence of DNA bridges at the cytokinetic furrow could induce mitotic exit and furrow regression, leading to multiple nuclei or polyploidy, which likely prevent further proliferation (O'Neill, 2022).

Under normal conditions, MB-NBs are lost at the end of pupal development via caspase-dependent apoptosis. This study found that traip- MB-NBs are lost prematurely, in part via caspase-dependent cell death, and thus fail to generate proper KC numbers. However, the caspase-inhibition experiments did not fully suppress traip- MB phenotypes, suggesting that additional redundant mechanisms may play a role in MB-NB loss. For example, when caspase-dependent apoptosis is inhibited, MB-NBs are primarily lost via autophagy. Alternatively, the irregular, crenellated nuclear envelope morphology of some traip- MB-NBs ( could point to non-apoptotic cell death pathways. Finally, aneuploidy-induced cell cycle exit in traip- MB-NBs could lead to loss via premature differentiation . Furthermore, it is likely that loss of KCs and/or GMCs also contributes some to traip- MB size defects (O'Neill, 2022).

TDNA bridge-induced defects likely feed into premature cell loss, but further work is required to dissect the pathways connecting them. In Drosophila, polyploid NBs can accumulate significant DNA damage as they enter mitosis, and chromosome breakage during mitosis in traip- could induce death through DNA-damage signaling. Drosophila embryos laid by traip- mothers do not survive, with extensive chromosome bridging and Chk2-dependent cell death, suggesting that DNA damage accumulation leads to cell loss in the rapidly dividing cells of the early embryo. In mammalian NPCs, polyploidy and binucleation can cause G1 arrest and apoptosis. In Drosophila, neurons can become polyploid in response to DNA damage, and NBs can become massively polyploid in some mutants , suggesting that, even though polyploidy may be better tolerated in flies, polyploid NBs are unlikely to complete additional mitoses successfully. The existence is inferred of traip- aneuploid MB-NBs, which produce a wide range of daughter KC numbers, suggesting that traip- generates some aneuploidies that are well tolerated and others that are highly lethal. Similarly, one recent study found that, although many karyotypes are permitted in NBs, loss of both copies of any of the three major Drosophila chromosomes resulted in a failure to proliferate and likely elimination. This parallels the situation in mammals, in which aneuploid NPCs and neurons are common, but also sensitive to G1 arrest, cell cycle exit, and apoptosis. Thus, both polyploidy and aneuploidy could stop further proliferation in traip- MB-NBs by preventing proper mitosis or inducing G1 arrest and cell cycle exit, eventually triggering cell loss via various mechanisms (O'Neill, 2022).

This study identified centrosome, spindle and cytokinetic furrow localizations for Traip that are important for function. One possibility is that the dynamic movement of Traip on the mitotic spindle and cytokinetic furrow promotes encounters with unresolved DNA bridges. GFP::Traip was never observed on bridges. However, as a single TRAIP protein is probably sufficient to unload each replisome, fluorescence detection may be unlikely. Interestingly, centrosome localization is a common aspect of microcephaly-linked proteins, including MCPH1, which also functions in DDR. Similar to Traip, MCPH1 has mitotic functions required for proper chromosome segregation, and mutations in MCPH1 lead to lagging chromosomes, DNA bridges and micronuclei. Mutations in microcephaly genes with centrosome-associated functions, such as CEP135 and CDK5RAP2, cause dysregulation of centrosome numbers, which also lead to chromosome segregation errors and aneuploidy . Thus, mitotic roles, ensuring proper chromosome segregation, and suppressing aneuploidy are common features of microcephaly-linked proteins. Future work seeking to understand these shared defects better may reveal a deeper etiological connection across microcephaly disorders (O'Neill, 2022).

Go to top

The connectome of the adult Drosophila mushroom body provides insights into function

Li, F., Lindsey, J. W., Marin, E. C., Otto, N., Dreher, M., Dempsey, G., Stark, I., Bates, A. S., Pleijzier, M. W., Schlegel, P., Nern, A., Takemura, S. Y., Eckstein, N., Yang, T., Francis, A., Braun, A., Parekh, R., Costa, M., Scheffer, L. K., Aso, Y., Jefferis, G. S., Abbott, L. F., Litwin-Kumar, A., Waddell, S. and Rubin, G. M. (2020). Elife 9. PubMed ID: 33315010

Making inferences about the computations performed by neuronal circuits from synapse-level connectivity maps is an emerging opportunity in neuroscience. The mushroom body (MB) is well positioned for developing and testing such an approach due to its conserved neuronal architecture, recently completed dense connectome, and extensive prior experimental studies of its roles in learning, memory and activity regulation. This study identified new components of the MB circuit in Drosophila, including extensive visual input and MB output neurons (MBONs) with direct connections to descending neurons. Unexpected structure was found in sensory inputs, in the transfer of information about different sensory modalities to MBONs, and in the modulation of that transfer by dopaminergic neurons (DANs). This study provides insights into the circuitry used to integrate MB outputs, connectivity between the MB and the central complex and inputs to DANs, including feedback from MBONs. The results provide a foundation for further theoretical and experimental work (Li, 2020).

Understanding how memories of past events are formed and then used to influence ongoing behavior are key challenges in neuroscience. It is generally accepted that parallel changes in connection strength across multiple circuits underlie the formation of a memory and that these changes are integrated to produce net changes in behavior. Animals learn to predict the value of sensory cues based on temporal correlations with reward or punishment. Such associative learning entails lasting changes in connections between neurons. It is now clear that different parts of the brain process and store different aspects of the information learned in a single event. In both flies and mammals, dopaminergic neurons play a key role in conveying information about whether an event has a positive or negative valence, and there are compelling parallels between the molecular diversity of dopaminergic cell types across these evolutionarily distant animals. However, there is only a limited understanding of how information about the outside world or internal brain state reaches different dopaminergic populations. Nor is the nature of the information that is stored in each parallel memory system understood, nor how these parallel memories interact to guide coherent behavior. It is believed such processes are governed by general and evolutionarily-conserved principles. In particular, it is believed the circuit logic that allows a brain to balance the competing demands of rapid learning with long-term stability of memory are likely to be the same in flies and mammals. Developing a comprehensive understanding of these circuits at the resolution of individual neurons and synapses will require the synergistic application of a variety of experimental methods together with theory and modeling. Many of the required methods are well developed in Drosophila, where the circuits underlying learning and memory are less complex than in mammals, and where detailed anatomical knowledge of the relevant circuits, which are believed will be essential, has just now become available. This study provide analysis of the complete connectome of a circuit involved in parallel processing of associative memories in adult fruit flies. The core architecture of this circuit is strikingly similar to that of the vertebrate cerebellum (see The shared circuit architecture of the mushroom body and the cerebellum) (Li, 2020).

The MB is the major site of associative learning in insects, and species that perform more complex behavioral tasks tend to have larger MBs. In the MB of each brain hemisphere, sensory stimuli are represented by the sparse activity of ~2000 Kenyon cells (KCs) whose dendrites form a structure called the MB calyx and whose parallel axonal fibers form the lobes of the MB (see Anatomy of the adult Drosophila MB) (Li, 2020).

The major sensory inputs to the Drosophila MB are olfactory, delivered by ~150 projection neurons (PNs) from the antennal lobe to the dendrites of the KCs in the MB calyx. KCs each receive input from an average of six PNs. For a KC to fire a spike, several of its PN inputs need to be simultaneously activated. This requirement, together with global feedback inhibition, ensures a sparse representation where only a small percentage of KCs are activated by an odor. The MB has a three layered divergent-convergent architecture in which the coherent information represented by olfactory PNs is expanded and decorrelated when delivered to the KCs. But the degree to which the structure of the sensory input representation is maintained by the KCs has been debated. This issue is explored, taking advantage of a nearly comprehensive dataset of KC inputs and outputs (Li, 2020).

While best studied for its role in olfactory associative learning, the MB also receives inputs from several other sensory modalities. A subset of projection neurons from the antennal lobe delivers information about temperature and humidity in both the larva. Taste conditioning also requires the MB and is believed to depend on specific KC populations, although the relevant inputs to these KCs have not yet been reported. This study identified one likely path for gustatory input to the MB (Li, 2020).

Drosophila MBs are also known to be able to form memories based on visual cues. Until a few years ago, it was thought that visual input reached the Drosophila MBs using only indirect, multisynaptic pathways as direct visual input from the optic lobes to the MBs, well known in Hymenoptera, had not been observed in any dipteran insect. A previous study identified two types of visual projection neurons (VPNs) connecting the optic lobes and the MB and additional connections have been observed recently by light microscopy. This study found that visual input was much more extensive than previously appreciated, with about 8% of KCs receiving predominantly visual input, and present in this study is a detailed description of neuronal pathways connecting the optic lobe and the MB. Visual sensory input appears to be segregated into distinct KC populations in both the larva and the adult, as is the case in honeybees. This study found two classes of KCs that receive predominantly visual sensory input, as well as MBONs that get the majority of their input from these segregated KC populations (Li, 2020).

MBONs provide the convergence element of the MB's three layer divergent-convergent circuit architecture. Previous work has identified 22 types of MBONs whose dendrites receive input from specific axonal segments of the KCs. The outputs of the MBONs drive learned behaviors. Approximately 20 types of dopaminergic neurons (DANs) innervate corresponding regions along the KC axons and are required for associative olfactory conditioning. Specifically, the presynaptic arbors of the DANs and postsynaptic dendrites of the MBONs overlap in distinct zones along the KC axons, defining the 15 compartmental units of the MB lobes. A large body of evidence indicates that these anatomically defined compartments of the MB are the units of associative learning (Li, 2020).

The DANs innervating different MBON compartments appear to play distinct roles in signaling reward vs. punishment, novelty vs. familiarity, the presence of olfactory cues and the activity state of the fly. These differences between DAN cell types presumably reflect in large part the nature of the inputs that each DAN receives, but knowledge of these inputs is just emerging and is far from comprehensive. DANs adjust synaptic weights between KCs and MBONs with cell type-specific rules and, in at least some cases, these differences arise from the effects of co-transmitters. In general, a causal association of KC responses with the activation of a DAN in a compartment results in depression of the synapses from the active KCs onto MBONs innervating that compartment. Different MB compartments are known to store and update non-redundant information as an animal experiences a series of learning events. In rodent and primate brains, recent studies have revealed that dopaminergic neurons are also molecularly diverse and encode prediction errors and other information based on cell type-specific rules (Li, 2020).

MBONs convey information about learned associations to the rest of the brain. Activation of individual MBONs can cause behavioral attraction or repulsion, according to the compartment in which their dendrites arborize. The combined output of multiple MBONs is likely to be integrated in downstream networks, but it is not understood how memories stored in multiple MB compartments alter these integrated signals to guide coherent and appropriate behaviors. Prior anatomical studies implied the existence of multiple layers of interneurons between MBONs and descending motor pathways. What is the nature of information processing in those layers? Anatomical studies using light microscopy provided the first hints. MBONs from different compartments send their outputs to the same brain regions, suggesting that they might converge on shared downstream targets. DANs often project to these same brain areas, raising the possibility of direct interaction between MBONs and DANs. The functional significance of such interactions has just begun to be investigated, and studies of the Drosophila larva, where a connectome of a numerically less complex MB is available, are providing valuable insights (Li, 2020).

The recently determined connectome of a portion of an adult female fly brain (hemibrain) provides connectivity data for ~22,500 neurons. Among them, ~2600 neurons have axons or dendrites in the MB, while ~1500 neurons are directly downstream of MBONs (using a threshold of 10 synapses from each MBON to each downstream target) and ~3200 are upstream of MB dopaminergic neurons (using a threshold of five synapses from each upstream neuron to each DAN). Thus this study will consider approximately one-third of the neurons in the central brain in this analysis of the MB ensemble (Li, 2020).

Throughout the paper synaptic thresholds were set in order to focus the descriptions and analyses on the most strongly connected neurons. In the above analysis, a higher threshold for MBON connections was chosen to downstream targets than for DAN inputs because the typical MBON has many more output synapses than a DAN has input synapses. At a thresholds of five synapses, DANs have a median of 31 different input neurons, but if the threshold to 10 synapses were increased this would decrease to only six different neurons. In contrast, at the threshold of 10 synapses, MBONs are connected to a median of 90 downstream neurons. There were some limitations resulting from not having a wiring diagram of the full central nervous system, as complete connectivity information for neurons with processes that extended outside the hemibrain volume was lacking. It was generally possible to mitigate these limitations by identifying the corresponding neurons in other EM or light microscopic datasets when the missing information was important for the analyses. Thus the hemibrain dataset was able to support a nearly comprehensive examination of the full neural network underlying the MB ensemble (Li, 2020).

Studies of the larval MB are providing parallel information on the structure and function of an MB with most of the same cell types, albeit fewer copies of each. The microcircuits inside three MB compartments in the adult were previously described and this study reports that the overall organization of these three compartments is conserved in a second individual of a different gender. More importantly, this study extends the analysis of microcircuits within the MB lobes to all 15 compartments, revealing additional aspects of spatial organization within individual compartments (Li, 2020).

The current study led to discovery of new morphological subtypes of KCs and to determine the sensory inputs delivered to the dendrites of each of the ~2000 KC. Considerable structure was found in the organization of those inputs and unexpectedly high levels of visual input, which was the majority sensory input for two classes of KCs. This segregation of distinct sensory representations into channels is maintained across the MB, such that MBONs, by sampling from different KCs, have access to different sensory modalities and representations. A new class of 'atypical' MBONs was discovered, consisting of 14 cell types, that have part of their dendritic arbors outside the MB lobes, allowing them to integrate input from KCs with other information; at least five of them make strong, direct synaptic contact onto descending neurons that drive motor action. How MBONs from different compartments interact with each other to potentially integrate and transform the signals passed from the MB to the rest of the brain is described, revealing a number of circuit motifs including multi-layered MBON-to-MBON feedforward networks and extensive convergence both onto common targets and onto each other through axo-axonal connections. Finally, this study analyzed the inputs to all 158 DANs that innervate the MB. Extensive direct feedback was found from MBONs to the dendrites of DANs, providing a mechanism of communication within and between MB compartments. Groups of DANs were found that share common inputs, providing mechanistic insights into the distributed parallel processing of aversive and appetitive reinforcement and other experimental observations (Li, 2020).

The hemibrain dataset contains the most comprehensive survey of the cell types and connectivity of MB neurons available to date. This connectome allowed probing in fine detail the circuitry underlying canonically proposed functions of the MB, including the representation of olfactory information by KCs, computation of valence by MBONs, and reinforcement of associations by DANs. Patterns in the input to DANs, MBON-to-DAN and MBON-to-MBON connectivity were found that suggest how associative learning in the MB can affect both the acquisition of new information through learning and the expression of previously learned responses. The connectome also reveals circuitry that supports non-canonical MB functions, including selective structure in non-olfactory pathways, a network of atypical MBONs, extensive heterogeneity in DAN inputs, and connections to central brain areas involved in navigation and movement (Li, 2020).

This analysis of the hemibrain connectome relied heavily on an extensive catalog of previously identified and genetically isolated cell types and on decades of study illuminating the link between MB physiology and fly behavior. It is worth emphasizing the interdependency of anatomy, physiology and behavior at the beginnning the post-connectomic era in fly research. Some of the neurons described that appear similarly connected may turn out to have diverse functions due to different physiology and, conversely, neurons that are morphologically distinct may turn out to have similar functions. In addition, a set of inputs with high synapse counts might appear, at the connectome level, to represent a major pathway for activating a particular neuron, but this will not be true if these inputs rarely fire at the same time. Likewise, a set of highly correlated inputs can be effective even if their individual synapse counts are modest. Lack of knowledge about correlated activity is probably the most significant uncertainty when attempting to map synapse counts onto circuit function, likely larger than possible errors in synapse identification and the effects of imposing various thresholds (see below). Finally, the connectome does not reveal gap junction connections or identify more distant non-synaptic modulation. These caveats should be kept in mind when interpreting connectome data (Li, 2020).

A number of these studies involved imposing a cutoff on the number of synapses required to include a particular connection in the analysis. The intent of these thresholds is to focus the analysis on what are likely to be the strongest inputs and outputs. Whether this cutoff should be based on a fixed number of synapses or on a percentage of total synapse counts is open to debate as, of course, is the actual value that the cutoff should take. Neurons differ widely in their numbers of inputs and outputs and these differences need to be taken into account when choosing thresholds. That is why information is presented about percentage of total inputs and outputs as well as synapse numbers. An approach to thresholding that takes such information into account has been used in a previous stdu; in the case of MBON to CX connections explored both here and in the previous study, differences in the connectivity map obtained with different thresholding methods were limited to the weakest connections (Li, 2020).

Most neurons make a large number of weak connections to other neurons, often involving just a single synapse. Some cases of low synapse number may be the result of incomplete reconstruction of neuronal arbors. In the MB lobes, an extensive effort was made to fully reconstruct all neurites and >80% of all computationally predicted synapses were assigned to identified neurons. Likewise, the arbors of MBONs and DANs outside the MB were extensively proofread. However, in other brain regions, where MBON outputs and DAN inputs lie, reconstruction was often much less complete. In such regions, it is difficult to estimate the extent to which the mapped synapses are representative of the full set of connections. A previous study reports an analysis of connectivity determined at two stages of reconstruction in the CX and, while the number of assigned synapses increased with additional proofreading, there was little difference in the connectivity maps. Errors in either synapse prediction or developmental wiring are also likely to produce some false connections represented by only one or two synapses. Of course, real, but sparse, connections might still be impactful if they fire concurrently, but this is not something that can be judged from the available information. Therefore, it seemed reasonable to ignore weak connections in this analyses. Thresholds must obviously be chosen with care and the effects of any particular cutoff value on results and conclusions should be assessed, preferably in conjunction with experimental data (Li, 2020).

Fly research has been greatly facilitated by the development of numerous fly lines that provide cell-type specific genetic access. This analysis has revealed, particularly in the case of DANs, subtypes within groups of neurons that would previously have been considered a single type. Thus, the higher resolution view of cell types that connectomics provides points out the need to develop driver lines or other experimental methods for more fine-tuned genetic access (Li, 2020).

The cell type constituents and circuit motifs of the MB in the adult fly have many similarities with its precursor at the larval stage of development. Both the larval and adult MBs support associative learning and, in both, PNs from the antennal lobe that convey olfactory information provide the majority of the sensory input, complemented by thermal, gustatory and visual sensory information that is segregated into distinct KC populations. However, the multi-layered organization of non-olfactory inputs in the main and accessory calyces (including integration of diverse input sources by LVINs) suggests that the KC representation in the adult is more highly enriched and specialized for non-olfactory sensory features. It is worth noting that the earliest born types of each of the three main adult KC classes (KCγs, KCγd, KCγt, KCα'β'ap1, KCαβp) appear to be specialized for non-olfactory sensory cues, and in most cases their dendrites lie in the accessory calyces (Li, 2020).

In the first instar larval MB, the only larval stage for which a connectome is available, there are roughly 70 mature KCs, which is increased nearly 30-fold in the adult. This enrichment likely increases odor discrimination and olfactory memory capacity. The larval MB has only eight compartments in its horizontal and vertical lobes. Although the increase in the adult to 15 compartments is only about a factor of two, the extent of their DAN modulation is greatly expanded. The larva has only seven confirmed DANs and five additional cells of unknown neurotransmitter thought to provide modulatory input, a factor of 15-fold less than the adult. Whereas each larval KC innervates all eight compartments, individual adult KCs innervate only five out of 15 compartments. Therefore the DANs in the larva are capable of modulating all KCs, whereas in the adult, DANs in different compartments modulate specific subtypes of KCs. The expansion of the number of DANs within many compartments in the adult MB, and subcompartmental targeting of individual DANs within a compartment, further increases the difference in granularity of DAN modulation between the larva and adult (Li, 2020).

MBONs feedback onto DANs and converge onto common downstream targets in both the larva and adult, implying some shared computational strategies. However, the greatly increased DAN complexity in the adult fly and the presence of subcompartmental organization of DAN axon targets not present in the larva suggest a substantial increase in the specificity of the learning signals involved in memory formation and raises the possibility of modality-specific learning signals to complement the multimodal KC representation (Li, 2020).

Previous theoretical work has emphasized the advantages of mixing sensory input to the KCs so that they provide a high-dimensional representation from which MBONs, guided by DAN modulation, can derive associations between sensory input and stimulus-valence. The connectome data modifies this viewpoint in two ways. First, although olfactory input to the KCs is highly mixed, various structural features reduce the dimensionality of the KC odor representation. A recent analysis of EM data from an adult Drosophila MB identified groups of PNs thought to represent food odors that are preferentially sampled by certain KCs. Consistent with this observation, the current analysis revealed subtype-specific biases in PN sampling by KCs, including an overrepresentation of specific glomeruli by α/β and α'/β' KCs. These biases, as well as other structural features, appear to arise from the stereotyped arrangement of PN axons within the calyx and their local sampling by KCs. This may reflect a developmental strategy by which the KC representation is organized to preferentially represent particular PN combinations. Further analyses of KC connectomes across hemispheres and animals, as well as experimental studies, will help evaluate the impact of this structure (Li, 2020).

The hemibrain connectome also revealed a second, more dramatic, structural feature of sensory input to the MB: non-olfactory input streams corresponding to visual and thermo-hygro sensation are strongly segregated. This organization may reflect the nature of stimulus-valence associations experienced by flies. In purely olfactory learning, when valence is associated with particular sets of olfactory receptors, MBONs need to be able to sample those combinations to successfully identify the stimulus. This requires that KCs mix input from multiple glomeruli within the olfactory stream. However, KCs that mix across streams corresponding to different sensory modalities may not be necessary if each modality can be used separately to identify valence. For example, either the visual appearance of an object or its odor may individually be sufficient to identify it as a food source. This study asked whether there is an advantage in having separate modality-specific sensory pathways, as seen in the connectome data, when valences can be decomposed in this way (Li, 2020).

To address this question a model is considered in which KC input is divided into two groups, visual and olfactory. In one version of the model ('shared KC population'), all KCs receive sparse random input from both types of PNs, corresponding to a high degree of mixing across modalities. In the other ('separate KC populations'), half of the KCs receive sparse random input exclusively from visual PNs and half only from olfactory PNs, corresponding to no cross-modal mixing. Two tasks were defined that differ in the way valences are defined. In the first (factorizable) version of the task, each olfactory stimulus and each visual stimulus is assigned a positive or neutral subvalence, and the net valence of the combined stimulus is positive if either of these components is positive (olfactory OR visual). In the second (unfactorizable) version, a valence is randomly assigned independently to each olfactory+visual stimulus pair. The ability of a model MBON, acting as a linear readout, was evaluated to determine valence from these two different KC configurations. Separate modality-specific KC populations are indeed beneficial when the valence can be identified from either one modality or the other. Dedicating different KC subtypes to distinct sensory modalities allows the predictive value of each modality to be learned separately. This result suggests that the divisions into KCs specialized for visual, thermo/hygro and olfactory signals may reflect how natural stimuli of different modalities are predictive of valence (Li, 2020).

On the basis of light-level studies, DAN modulation of KC-to-MBON synapses has been considered to operate at the resolution of MB compartments. However, taken with other recent studies, the morphology and connectivity data indicate that functionally distinct PAM-DAN subtypes operate within a MB compartment. DAN subtypes receive different inputs and likely modulate different KC-MBON synapses within a compartment (Li, 2020).

A prior analysis showed that PAM01-fb DANs were required to reinforce the absence of expected shock during aversive memory extinction, whereas a different set of γ5 DANs were needed for learning with sugar reinforcement. It is conceivable that another γ5 DAN subtype is required for male flies to learn courtship rejection. The connectome data revealed DAN subtypes in every compartment that is innervated by PAM DANs (Li, 2020).

The γ3 compartment provides an interesting example of subcompartmental targeting of modulation by DANs and of KC input onto MBONs. There are two subtypes of PAM DANs and three types of MBONs in the γ3 compartment. MBON09 and MBON30 primarily receive olfactory information from γm KCs whereas MBON33 primarily receives visual information from γd KCs. PAM12-dd and PAM12-md DANs appear to modulate KC inputs to MBON09/MBON30 or MBON33, respectively. Although existing driver lines do not separate the dd and md subtypes, PAM-γ3 DAN activity is suppressed by sugar and activated by electric shock. This study found that PAM12-dd DANs share input with PPL1 DANs conveying punishment signals, while PAM12-md DANs are co-wired with PAM08 (γ4) DANs conveying reward signals. Thus, synaptic transmission from two sets of modality-specific KCs to different MBONs can be independently modulated by DANs signaling different valences, all within a single compartment (Li, 2020).

To explore the implications of DAN modulation that is specific to sensory modality,the model presented above was expanded by including two DANs, one conveying the visual component of valence and the other the olfactory component. KCs were divided into visual and olfactory modalities, and two configurations for DAN modulation, one, 'shared reward signal', that is compartment-wide and non-specific, and the other, 'separate reward signals' were considered, in which each DAN only induces plasticity onto KC synapses matching its own modality. This latter case models a set of DANs that affect synapses from visual KCs onto the MBON and another set that affect olfactory synapses (alternatively, it could model two MBONs in different compartments that are modulated independently and converge onto a common target). This study found that, when KCs are divided into separate populations and separate modalities can be used to identify valence, learning is more efficient if the pathways are modulated individually (Li, 2020).

This analysis revealed that DANs receive very heterogeneous inputs but, nonetheless, some DANs both within and across compartments often share common input. This combination of heterogeneity and commonality provides many ways of functionally combining different DAN subtypes. For example, it is expected that this combination allows DANs to encode many different combinations of stimuli, actions and events in a state-dependent manner and to transmit this information to specific loci within the MB network (Li, 2020).

In addition to heterogeneous inputs from a variety of brain regions, the DAN network receives a complex arrangement of within and across-compartment monosynaptic input from a variety of MBONs, using both excitatory and inhibitory neurotransmitters. It was found that nearly all MB compartments contain at least one direct within-compartment MBON-DAN feedback connection. MBON feedback onto these DAN subtypes allows previously learned associations that modify MBON activity to affect future learning. MBONs that feedback onto the same DANs that modulate them could, if the result of learning is reduced DAN activity, prevent excess plasticity for an already learned association. In cases where MBON activity excites the DAN, the self-feedback motif could assure that learning does not stop until the MBON has been completely silenced. More generally, MBON inputs to DANs imply that dopaminergic signals themselves reflect learned knowledge and the actions it generates. This could, in turn, allow MBON modulation of DAN activity to support a number of learning paradigms beyond pure classical conditioning, including extinction, second-order conditioning, operant conditioning and reinforcement learning (Li, 2020).

Flies can perform second-order conditioning, in which a stimulus that comes to be associated with reinforcement may itself act as a pseudo-reinforcement when associated with other stimuli. This computational motif of learning the value of sensory states and using the inferred value as a surrogate reinforcement to guide behavioral learning is the core principle behind a class of machine learning techniques known as actor-critic algorithms. These algorithms consist of two modules: the 'actor', whose job is to map sensory inputs to behavioral outputs and the 'critic', whose job is to map sensory inputs to their inferred values and provide these values as a learning signal for the actor. In the mammalian basal ganglia, the dorsal and ventral striatum, the latter of which strongly influences the activity of dopamine neurons in areas including VTA, have been proposed to represent actor and critic modules, respectively. In the MB, this perspective suggests a possible additional 'critic' function for some MBONs beyond their known 'actor' role in directly driving behaviors. Consistent with this view, activation of individual MBONs can excite DANs in other compartments. This study found strong direct monosynaptic connections between some MBONs and DANs in other compartments. Functional studies will be needed to determine which MBONs, if any, participate in an actor-critic arrangement, and which circuit mechanisms-for example, release from inhibition, or reduction of excitation-are at work (Li, 2020).

Another potential role of cross-compartment MBON-DAN feedback is to gate the learning of certain associations so that the learning is contingent on other associations having already been formed. Such a mechanism could support forms of memory consolidation in which long-term memories are only stored after repeated exposure to a stimulus and an associated reward or punishment. Prior studies have linked plasticity in the γ1 and γ2 compartments to short-term aversive memory and plasticity in the α2 and α3 compartments to long-term aversive memory. The cross-compartmental MBON-to-DAN connections observed in this study suggest an underlying cicuit mechanism for this 'transfer' of short to long term memory. Aversion drives PPL101 (γ1pedc) and depresses the conditioned odor-drive to the GABAergic MBON11 (γ1pedc>α/β). MBON11 is strongly connected with PPL1-γ1pedc and is more weakly connected with PPL105 and PPL106. Depression of MBON11 will therefore also release the PPL105 and PPL106 DANs from MBON11-mediated inhibition, increasing their activity in response to the conditioned odor and making them more responsive during subsequent trials. The net result is that short-term aversive learning by MBON11 (γ1pedc>α/β) promotes long-term learning in MBON18 (α2sc) and MBON14 (α3) by releasing the inhibition on the dopamine neurons that innervate the α2 and α3 compartments. Indeed pairing inactivation of MBON11 with odor presentation can form an aversive memory that requires output from PPL1-DANs and optogenetic stimulation of MBON11 during later trials of odor-shock conditioning impairs long-term memory formation (Li, 2020).

Cross-compartment MBON-DAN feedback may also enable context-dependent valence associations, such as the temporary association of positive valence with neutral stimuli when a fly is repeatedly exposed to aversive conditions. Multiple, spaced aversive conditioning trials were recently shown to form, in addition to an aversive memory for the shock-paired odor, a slowly emerging attraction, a 'safety' memory, for a second odor that was presented over the same training period without shock. The GABAergic MBON09 (γ3β'1) appears to play a critical role in the formation of this safety memory. The synapses from KCs conveying the shock-paired odor are depressed in that portion of MBON09's dendrite that arborizes in γ3, while the synapses from KCs conveying the safety odor are depressed in its β'1 arbor. These combined modulations should gradually release downstream neurons from MBON09 feedforward inhibition, consistent with the proposed mechanism for PAM13/14 (β'1) and PAM05/06 (β'2m and β'2p) DANs becoming more responsive to the safe odor. The connectome data revealed that MBON09 is directly connected to the PAM13/14 (β'1ap) and PAM05 (β'2p) DANs, consistent with release from inhibition underlying the delayed encoding of safety (Li, 2020).

DANs also make direct connections with MBONs in each MB compartment, and optogenetic activation of PAM11 DANs can directly excite MBON07 with slow dynamics. This might provide a mechanism to temporarily suppress memory expression without impairing the underlying memory, which is stored as depressed KC-to-MBON synapses. For example, DANs are known to control hunger and thirst-dependent memory expression, and their excitation of MBONs could provide a possible mechanism (Li, 2020).

Feedforward MBON-MBON connections were postulated, based on behavioral and light microscopic anatomical observations, to propagate local plasticity between compartments. The GABAergic MBON11 is poised to play a key role in such propagation as it is connected with 17 other MBONs (at a threshold of 10 synapses). Thus local depression of KC-MBON11 synapses by the shock responsive PPL1-γ1pedc DAN would be expected to result in disinhibition of MBONs in other compartments. Indeed enhanced CS+ responses of MBON01 and MBON03 after odor-shock conditioning have been ascribed to release from MBON11 inhibition. The consequence of releasing other strongly connected MBONs (MBON07, MBON14 and MBON29) from MBON11 inhibition awaits future study. As discussed above, MBON11 is also connected to DANs innervating its cognate compartment and to DANs innervating other compartments. MBON11 may therefore coordinate MBON network activity via both direct and indirect mechanisms. Analogous feedforward inhibitory connections were found from MBON09 (γ3β'1) to MBON01 and MBON03. Aversive learning therefore reduces both MBON11- and MBON09-mediated inhibition of these MBONs, which further skews the MBON network toward directing avoidance of the previously punished odor (Li, 2020).

Disinhibition likely also plays an important role in appetitive memory. The connectivity of MBON06 (β1>α) revealed here indicates that local plasticity in the β1 compartment can propagate to other MBONs. Similar to MBON11, MBON06 is directly connected with nine other MBONs with a threshold of 10 synapses. MBON06 gradually increases its response to repeated odor exposure and odor-evoked responses of β1 DANs vary with metabolic state. More compellingly, artificially triggering PAM10 (β1) DANs can assign appetitive valence to odors. However, the role of MBON06 in appetitive memory and how the PAM10 (β1) DANs modulate KC synapses to MBON06 has not been investigated. The glutamatergic MBON06 (β1>α) makes a large number of reciprocal axoaxonic connections with the GABAergic MBON11, whose activation favors approach. This reciprocal network motif and the positive sign of behavior resulting from β1 DAN-driven memory suggests that MBON06 released glutamate is likely to be inhibitory to MBON11 and its other downstream targets. MBON06 also makes twice as many connections onto the glutamatergic MBON07 and the cholinergic MBON14 as does MBON11. MBON14 (α3) and MBON07 (α1) have established roles in appetitive memory. Assuming that PAM10 (β1) DANs encode appetitive memory by depressing synapses from odor-specific KCs onto MBON06 (β1>α), MBON06 suppression will release the feedforward inhibition of MBON06 onto MBON07 (α1), freeing it to participate in driving the PAM11-aad (α1) DANs. Release of MBON06 inhibition should also simultaneously potentiate the responses of MBON14 (α3) to the conditioned odor. Lastly, releasing the strong inhibition from MBON06 frees MBON11 to provide weaker inhibition that would further favor odor-driven approach (Li, 2020).

Aversive learning will also alter the function of the MBON06:MBON11 network motif. Aversive reinforcement through the PPL101 (γ1pedc) DAN depresses KC-MBON11 connections. This depression releases MBON06 from MBON11-mediated suppression and allows MBON06 to then suppress output through MBON07 and MBON11, further favoring odor-driven avoidance (Li, 2020).

Several studies have described the influence of internal states such as hunger and thirst on the function and physiology of the MB. In essence, states appear to modulate the DAN-MBON network so that the fly preferentially engages in the pursuit of its greatest need. Since current knowledge suggests these deprivation states employ volume release of modulatory peptides or monoamines to control specific DANs and their downstream MBONs, the connectome analyzed in this study does not provide a complete description of this circuit. Nevertheless, direct connectivity does provide some interesting new insight concerning hunger and thirst-dependent control (Li, 2020).

The MBON06 (β1>α):MBON11 (γ1pedc>α/β) cross-inhibitory network motif is likely to be relevant for the dependence of learning and memory expression on hunger state. PPL101 (γ1pedc) DANs and MBON11 (γ1pedc>α/β) are sensitive to nutrient/satiety status, with MBON11 being more responsive in hungry flies. Therefore, the hungry state favors the activity of the MBONs that are normally repressed by MBON06 (Li, 2020).

Thirst-dependent seeking of water vapor requires the activity of DANs innervating the β'2 compartment. The current work shows that these DANs are likely to directly modulate thermo/hygrosensory KCs. In addition, a recent study showed that thirst-dependent expression of water memory required peptidergic suppression of the activity of both the PPL103(γ2α'1) and PAM02 (β'2a) DANs. Interestingly, blocking the PAM02 (β'2a) DANs released memory expression in water-sated flies whereas blocking the PPL103 (γ2α'1) DANs had no effect. However, if the PPL103 (γ2α'1) DANs were blocked together with PAM02 (β'2a) they further facilitated water memory expression. The connectome suggests circuit mechanisms that could reconcile these observations: MBON12 (γ2α'1) provides strong cholinergic input to the PAM02 (β'2a) DANs, suggesting that PPL103 (γ2α'1) DANs might facilitate water memory expression by suppressing MBON12's excitatory input onto PAM02 (β'2a) DANs (Li, 2020).

A role for the MB in guiding locomotion and navigation in ants and other insects has been proposed. The strong and direct connections observed from the majority of MBONs to the CX, the fly's navigation and locomotion control center, provide one circuit path for the MB to exert influence on motor behaviors. Discovering how this input is utilized by the CX will require additional experimental work (Li, 2020).

Optogenetic activation of Drosophila MBONs can promote attraction or avoidance by influencing turning at the border between regions with and without stimulating light. The effect of MBON activation is additive: coactivation of positive-valence MBONs produced stronger attraction, whereas coactivation of positive and negative-valence MBONs cancelled each other out. Because the fly needs to balance the outputs of different compartments, it is expected that those downstream neurons that integrate inputs from multiple MBONs will have a privileged role in motor control (Li, 2020).

The activity of some DANs has been shown to correlate with motor activity, and the optogenetic activation of PAM-β'2 or PPL1-α3 DANs can attract flies, indicating that DAN activity can itself in some circumstances drive motor behavior. The circuit mechanisms generating the correlation between DAN activity and motor behavior remain to be discovered. Downstream targets of MBONs provide extensive input to DANs, and this study found that neurons downstream of multiple MBONs are twice as likely as other MBON targets to provide such direct input to DANs (Li, 2020).

This study also discovered a direct pathway mediated by atypical MBONs that connects to the descending neurons (DNs) that control turning, an observation that provides additional support for the importance of the MB in the control of movement. The connections of these MBONs appear to be structured so as to promote directional movement, often involving a push-pull arrangement of MBONs signaling approach and avoidance. In addition to direct connections to DNs, there is a network of connections mediated by both local LAL interneurons and interneurons that connect the right and left hemisphere LALs. The atypical MBONs that connect directly to the descending steering system, MBON26, MBON27, MBON31 and MBON32, appear to be among the most integrative neurons in the MB system in the sense that they combine direct KC input from the MB compartments with both input from many other MBONs and non-MB input. At the level of the descending neurons, the highly processed signals from these MBONs are combined with inputs from many other sources, including the central complex, to affect a decision to turn. This high degree of integration presumably reflects the complexity and importance of this decision, with many factors involved that might act individually or in combination (Li, 2020).

Visual input to the MB is over-represented in the output to the descending neurons, predominantly through MBON27. Short- and long-term learning based on features in a visual scene has been reported to involve the CX. Plasticity in the CX enables visual feature input from the sky and surrounding scenery to be mapped flexibly onto the fly's internal compass. The visual input conveyed to the MB and, presumably, the learning at the synapses between visual KCs and MBON27, may be of lower resolution, encoding broader features such as color and contrast. An early study demonstrated that the MB is dispensable for flying Drosophila to learn shapes but that it is required for them to generalize their learning if the visual context changes between training and testing. Memory of visual features and the ability to generalize context could allow visual landmarks to help guide navigation either through the CX or by directly influencing descending neurons. The thermo/hygrosensory features conveyed by MBON26 could play a similar role, as could the large amount of odor-related information present in this MBON output pathway (Li, 2020).

The MB has an evolutionarily-conserved circuit architecture and uses evolutionarily-conserved molecular mechanisms of synaptic plasticity. The dense connectome analysed in this report has uncovered many unanticipated circuit motifs and suggested potential circuit mechanisms that now need to be explored experimentally. Drosophila provides access to many of the required tools such as cell-type-specific driver lines, genetically encoded sensors and microscopy methods to observe whole-brain neuronal activity and fine ultrastructure. These features make the fly an excellent system in which to study many general issues in neuroscience, including: the functional diversity of dopaminergic neurons that carry distributed reinforcement signals, the interactions between parallel memory systems, and memory-guided action selection, as well as the mechanisms underlying cell-type-specific plasticity rules, memory consolidation, and the influence of internal state. It is expected that studies of the MB will provide insight into general principles used for cognitive functions across the animal kingdom (Li, 2020).

As mentioned in the introduction, the MB shares many features with the vertebrate cerebellum, and the results should be informative for studies of the cerebellum proper as well as other cerebellum-like structures such as the dorsal cochlear nucleus and the electrosensory lobe of electric fish. A distinctive feature of these systems, and of the MB, is that learning is driven by a particular mechanism; for example DAN modulation in the MB or complex spiking driven by climbing fiber input in the cerebellum. Studies of learning in cortical circuits have traditionally focused on Hebbian forms of learning driven by the ongoing input and output activity of a neuron. However, recent results from both hippocampal circuits have stressed the importance of plasticity that is driven by dendritic plateau potentials or bursts that resemble the distinct learning events seen in cerebellar and MB circuits. Thus, the form of plasticity seen in the MB and its control by output and modulatory circuits may inform studies of learning in the cerebral cortex as well (Li, 2020).

Go to top

Genetic dissection of aversive associative olfactory learning and memory in Drosophila larvae

Widmann, A., Artinger, M., Biesinger, L., Boepple, K., Peters, C., Schlechter, J., Selcho, M. and Thum, A. S. (2016). PLoS Genet 12: e1006378. PubMed ID: 27768692

Memory formation is a highly complex and dynamic process. It consists of different phases, which depend on various neuronal and molecular mechanisms. In adult Drosophila it was shown that memory formation after aversive Pavlovian conditioning includes-besides other forms-a labile short-term component that consolidates within hours to a longer-lasting memory. Accordingly, memory formation requires the timely controlled action of different neuronal circuits, neurotransmitters, neuromodulators and molecules that were initially identified by classical forward genetic approaches. Compared to adult Drosophila, memory formation was only sporadically analyzed at its larval stage. This study deconstructed the larval mnemonic organization after aversive olfactory conditioning. After odor-high salt conditioning (establishing an aversive olfactory memory) larvae form two parallel memory phases; a short lasting component that depends on cyclic adenosine 3'5'-monophosphate (cAMP) signaling and synapsin gene function. In addition, this study shows for the first time for Drosophila larvae an anesthesia resistant component, which relies on radish and bruchpilot gene function, protein kinase C (PKC) activity, requires presynaptic output of mushroom body Kenyon cells and dopamine function. Given the numerical simplicity of the larval nervous system this work offers a unique prospect for studying memory formation of defined specifications, at full-brain scope with single-cell, and single-synapse resolution (Widmann, 2016).

Memory formation and consolidation usually describes a chronological order, parallel existence or completion of distinct short-, intermediate- and/or long-lasting memory phases. For example, in honeybees, in Aplysia, and also in mammals two longer-lasting memory phases can be distinguished based on their dependence on de novo protein synthesis. In adult Drosophila classical odor-electric shock conditioning establishes two co-existing and interacting forms of memory--ARM and LTM--that are encoded by separate molecular pathways (Widmann, 2016).

Seen in this light, memory formation in Drosophila larvae established via classical odor-high salt conditioning seems to follow a similar logic. It consist of LSTM (larval short lasting component) and LARM (anesthesia resistant memory). Aversive olfactory LSTM was already described in two larval studies using different negative reinforcers (electric shock and quinine) and different training protocols (differential and absolute conditioning). The current results introduce for the first time LARM that was also evident directly after conditioning but lasts longer than LSTM. LARM was established following different training protocols that varied in the number of applied training cycles and the type of negative or appetitive reinforcer. Thus, LSTM and LARM likely constitute general aspects of memory formation in Drosophila larvae that are separated on the molecular level (Widmann, 2016).

Memory formation depends on the action of distinct molecular pathways that strengthen or weaken synaptic contacts of defined sets of neurons. The cAMP/PKA pathway is conserved throughout the animal kingdom and plays a key role in regulating synaptic plasticity. Amongst other examples it was shown to be crucial for sensitization and synaptic facilitation in Aplysia, associative olfactory learning in adult Drosophila and honeybees, long-term associative memory and long-term potentiation in mammals (Widmann, 2016).

For Drosophila larvae two studies by Honjo (2005) and Khurana (2009) suggest that aversive LSTM depends on intact cAMP signaling. In detail, they showed an impaired memory for rut and dnc mutants following absolute odor-bitter quinine conditioning and following differential odor-electric shock conditioning. Thus, both studies support the interpretation of the current results. It is argued that odor-high salt training established a cAMP dependent LSTM due to the observed phenotypes of rut, dnc and syn mutant larvae. The current molecular model is summarized in A molecular working hypothesis for LARM formation. Yet, it has to be mentioned that all studies on aversive LSTM in Drosophila larvae did not clearly distinguish between the acquisition, consolidation and retrieval of memory. Thus, future work has to relate the observed genetic functions to these specific processes (Widmann, 2016).

In contrast, LARM formation utilizes a different molecular pathway. Based on different experiments, it was ascertained, that LARM formation, consolidation and retrieval is independent of cAMP signaling itself, PKA function, upstream and downstream targets of PKA, and de-novo protein synthesis. Instead it was found that LARM formation, consolidation and/or retrieval depends on radish (rsh) gene function, brp gene function, dopaminergic signaling and requires presynaptic signaling of MB KCs (Widmann, 2016).

Interestingly, studies on adult Drosophila show that rsh and brp gene function, as well as dopaminergic signaling and presynaptic MB KC output are also necessary for adult ARM formation. Thus, although a direct comparison of larval and adult ARM is somehow limited due to several variables (differences in CS, US, training protocols, test intervals, developmental stages, and coexisting memories), both forms share some genetic aspects. This is remarkable as adult ARM and LARM use different neuronal substrates. The larval MB is completely reconstructed during metamorphosis and the initial formation of adult ARM requires a set of MB α/β KCs that is born after larval life during puparium formation (Widmann, 2016).

In addition, this study has demonstrated the necessity of PKC signaling for LARM formation in MB KCs. The involvement of the PKC pathway for memory formation is also conserved throughout the animal kingdom. For example, it has been shown that PKC signaling is an integral component in memory formation in Aplysia, long-term potentiation and contextual fear conditioning in mammals and associative learning in honeybees. In Drosophila it was shown that PKC induced phosphorylation cascade is involved in LTM as well as in ARM formation. Although the exact signaling cascade involved in ARM formation in Drosophila still remains unclear, this study has established a working hypothesis for the underlying genetic pathway forming LARM based on the current findings and on prior studies in different model organisms. Thereby this study does not take into account findings in adult Drosophila. These studies showed that PKA mutants have increased ARM and that dnc sensitive cAMP signaling supports ARM. Thus both studies directly link PKA signaling with ARM formation. (Widmann, 2016).

KCs have been shown to act on MB output neurons to trigger a conditioned response after training. Work from different insects suggests that the presynaptic output of an odor activated KCs is strengthened if it receives at the same time a dopaminergic, punishment representing signal. The current results support these models as they show that LARM formation requires accurate dopaminergic signaling and presynaptic output of MB KCs. Yet, for LARM formation dopamine receptor function seems to be linked with PKC pathway activation. Indeed, in honeybees, adult Drosophila and vertebrates it was shown that dopamine receptors can be coupled to Gαq proteins and activate the PKC pathway via PLC and IP3/DAG signaling. As potential downstream targets of PKC radish and bruchpilot are suggested. Interference with the function of both genes impairs LARM. The radish gene encodes a functionally unknown protein that has many potential phosphorylation sites for PKA and PKC. Thus considerable intersection between the proteins Rsh and PKC signaling pathway can be forecasted. Whether this is also the case for the bruchpilot gene that encodes for a member of the active zone complex remains unknown. The detailed analysis of the molecular interactions has to be a focus of future approaches. Therefore, the current working hypothesis can be used to define educated guesses. For instance, it is not clear how the coincidence of the odor stimulus and the punishing stimulus are encoded molecularly. The same is true for ARM formation in adult Drosophila. Based on the working hypothesis it can be speculated that PKC may directly serve as a coincidence detector via a US dependent DAG signal and CS dependent Ca2+ activation (Widmann, 2016).

Do the current findings in general apply to learning and memory in Drosophila larvae? To this the most comprehensive set of data can be found on sugar reward learning. Drosophila larva are able to form positive associations between an odor and a number of sugars that differ in their nutritional value. Using high concentrations of fructose as a reinforcer in a three cycle differential training paradigm (comparable to the one used in this study for high salt learning and fructose learning) other studies found that learning and/or memory in syn97 mutant larvae is reduced to ~50% of wild type levels. Thus, half of the memory seen directly after conditioning seems to depend on the cAMP-PKA-synapsin pathway. The current results in turn suggest that the residual memory seen in syn97 mutant larvae is likely LARM. Thus, aversive and appetitive olfactory learning and memory share general molecular aspects. Yet, the precise ratio of the cAMP-dependent and independent components rely on the specificities of the used odor-reinforcer pairings. Two additional findings support this conclusion. First, a recent study has shown that memory scores in syn97 mutant larvae are only lower than in wild type animals when more salient, higher concentrations of odor or fructose reward are used. Usage of low odor or sugar concentrations does not give rise to a cAMP-PKA-synapsin dependent learning and memory phenotype. Second, another study showed that learning and/or memory following absolute one cycle conditioning using sucrose sugar reward is completely impaired in rut1, rut2080 and dnc1 mutants. Thus, for this particular odor-reinforcer pairing only the cAMP pathway seems to be important. Therefore, a basic understanding of the molecular pathways involved in larval memory formation is emerging. Further studies, however, will be necessary in order to understand how Drosophila larvae make use of the different molecular pathways with respect to a specific CS/US pairing (Widmann, 2016).

Go to top

Different Kenyon cell populations drive learned approach and avoidance in Drosophila

Perisse, E., Yin, Y., Lin, A. C., Lin, S., Huetteroth, W. and Waddell, S. (2013). Neuron 79: 945-956. PubMed ID: 24012007

In Drosophila, anatomically discrete dopamine neurons that innervate distinct zones of the mushroom body (MB) assign opposing valence to odors during olfactory learning. Subsets of MB neurons have temporally unique roles in memory processing, but valence-related organization has not been demonstrated. This study functionally subdivided the αβ neurons, revealing a value-specific role for the approximately 160 αβ core (αβc) neurons. Blocking neurotransmission from αβ surface (αβs) neurons revealed a requirement during retrieval of aversive and appetitive memory, whereas blocking αβc only impaired appetitive memory. The αβc were also required to express memory in a differential aversive paradigm demonstrating a role in relative valuation and approach behavior. Strikingly, both reinforcing dopamine neurons and efferent pathways differentially innervate αβc and αβs in the MB lobes. It is proposed that conditioned approach requires pooling synaptic outputs from across the αβ ensemble but only from the αβs for conditioned aversion (Perisse, 2013).

Olfactory memories are believed to be represented within the ~2,000 intrinsic Kenyon cells (KCs) of the Drosophila mushroom body (MB). Individual odors activate relatively sparse populations of KCs within the overall MB ensemble providing cellular specificity to odor memories. Prior research of fly memory suggests that the KCs can be functionally split into at least three major subdivisions: the αβ, α′β′, and γ neurons (see Anatomical organization of the olfactory nervous system in Drosophila from Davis, R. Traces of Drosophila Memory, Neuron 70: 8-19, 2011). The current consensus suggests a role for γ in short-term memory, for α′β′ after training for memory consolidation, and for αβ in later memory retrieval, with the αβ requirement becoming more pronounced as time passes. Importantly, odor-evoked activity is observable in each of these cell types, consistent with a parallel representation of olfactory stimuli across the different KC classes (Perisse, 2013).

Value is assigned to odors during training by anatomically distinct dopaminergic (DA) neurons that innervate unique zones of the MB. Negative value is conveyed to MB γ neurons in the heel and junction and to αβ neurons at the base of the peduncle and the tip of the β lobe. In contrast, a much larger number of rewarding DA neurons project to approximately seven nonoverlapping zones in the horizontal β, β′, and γ lobes. This clear zonal architecture of reinforcing neurons suggests that plastic valence-relevant KC synapses may lie adjacent to these reinforcing neurons. Furthermore, presumed downstream MB efferent neurons also have dendrites restricted to discrete zones on the MB lobes, consistent with memories being formed at KC-output neuron synapses (Perisse, 2013).

Long before the zonal DA neuron innervation of the MB was fully appreciated, experiments suggested that appetitive and aversive memories were independently processed and stored. Subsequently, models were proposed that represented memories of opposite valence at distinct output synapses on the same odor-activated KCs or on separate KCs. Importantly, memory retrieval through these modified KC-output synapses was predicted to guide either odor avoidance or approach behavior. A KC synapse-specific representation of memories of opposing valence would dictate that it is not possible to functionally separate the retrieval of aversive and appetitive memories by disrupting KC-wide processes. This study therefore tested these models by systematically blocking neurotransmission from subsets of the retrieval-relevant αβ neurons. It was found that aversive and appetitive memories can be distinguished in the αβ KC population, showing that opposing odor memories do not exclusively rely on overlapping KCs. Whereas output from the αβs neurons is required for aversive and appetitive memory retrieval, the αβ core (αβc) neurons are only critical for conditioned approach behavior. Higher-resolution anatomical analysis of the innervation of reinforcing DA neurons suggests that valence-specific asymmetry may be established during training. Furthermore, dendrites of KC-output neurons differentially innervate the MB in a similarly stratified manner. It is therefore proposed that aversive memories are retrieved and avoidance behavior triggered only from the αβ surface (αβs) neurons, whereas appetitive memories are retrieved and approach behavior is driven by efferent neurons that integrate across the αβ ensemble (Perisse, 2013).

When faced with a choice, animals must select the appropriate behavioral response. Learning provides animals the predictive benefit of prior experience and allows researchers to influence behavioral outcomes. After olfactory learning, fruit flies are provided with a simple binary choice in the T-maze. Aversively trained flies preferentially avoid the conditioned odor, whereas appetitively conditioned flies approach it. A major goal of the field is to understand the neural mechanisms through which the fly selects the appropriate direction (Perisse, 2013).

In mammals, mitral cells take olfactory information direct from the olfactory bulb to the amygdala and the perirhinal, entorhinal, and piriform cortices. In doing so, odor information is segregated into different streams, allowing it to be associated with other modalities and emotionally salient events. In contrast, most olfactory projection neurons in the fly innervate the MB calyx and lateral horn or only the lateral horn. The lateral horn has mostly been ascribed the role of mediating innate responses to odors, leaving the MB to fulfill the potential roles of the mammalian cortices (Perisse, 2013).

Although morphological and functional subdivision of the αβ, α′β′, and γ classes of MB neuron has been reported, until now a valence-restricted role has been elusive. This study investigated the functional correlates of substructure within the αβ population. An appetitive memory-specific role was identified for the αβc neurons. Whereas blocking output from the αβs neurons impaired aversive and appetitive memory retrieval, blocking αβc neurons produced only an appetitive memory defect. These behavioral results, taken with functional imaging of odor-evoked activity, suggest that beyond the αβ, α′β′, and γ subdivision, odors are represented as separate streams in subsets of MB αβ neurons. These parallel information streams within αβ permit opposing value to be differentially assigned to the same odor. Training therefore tunes the odor-activated αβc and αβs KCs so that distinct populations differentially drive downstream circuits to generate aversive or appetitive behaviors. Such a dynamic interaction between appetitive and aversive circuits that is altered by learning is reminiscent of that described between the primate amygdala and orbitofrontal cortex. It will be important to determine the physiological consequences of appetitive and aversive conditioning on the αβc and αβs neurons. Positively and negatively reinforced olfactory learning in rats produced bidirectional plasticity of neurons in the basolateral amygdala (Perisse, 2013).

The αβp neurons, which do not receive direct olfactory input from projection neurons in the calyx, are dispensable for aversive and appetitive 3 hr memory and for 24 hr appetitive memory. The αβp neurons were reported to be structurally linked to dorsal anterior lateral (DAL) neurons and both DAL and αβp neurons were shown to be required for long-term aversive memory retrieval. This study found that, like αβp neurons, DAL neurons are not required for appetitive long-term memory retrieval. In addition, the αβp neurons were inhibited by odor exposure, which may reflect cross-modal inhibition within the KC population (Perisse, 2013).

Observing a role for the αβc neurons in the relative aversive paradigm argues against the different requirement for αβc neurons in the routine shock-reinforced aversive and sugar-reinforced appetitive assays being due to different timescales of memory processing. In addition, a pronounced role was observed for αβc neurons in retrieval of 24 hr appetitive LTM, whereas others have reported that αβc neurons are not required for the retrieval of 24 hr aversive LTM . Nevertheless, time and the methods of conditioning may be important variables. Although appetitive and aversive memory retrieval requires output from the αβ ensemble at 3 hr and 24 hr after conditioning, αβ neurons were shown to be dispensable for 2 hr appetitive memory retrieval. Instead, appetitive retrieval required γ neuron output at this earlier point. The current experiments were generally supportive of the γ-then-αβ neuron model but revealed a slightly different temporal relationship. The αβ neurons were dispensable for memory retrieved 30 min after training but were essential for 2 hr and 3 hr memory after training. An early role for γ neurons is further supported by the importance of reinforcing DA input to the γ neurons for aversive memory formation. It will be interesting to determine whether there is a stratified representation of valence within the γ neuron population (Perisse, 2013).

Finding an appetitive memory-specific role for αβc neurons suggests that the simplest model in which each odor-activated KC has plastic output synapses driving either approach or avoidance appears incorrect. Such a KC output synapse-specific organization dictates that it would not be possible to functionally segregate aversive and appetitive memory by blocking KC-wide output. This study however found a specific role for the αβc neurons in conditioned approach that supports the alternative model of partially nonoverlapping KC representations of aversive and appetitive memories. The anatomy of the presynaptic terminals of reinforcing DA neurons in the MB lobes suggests that the functional asymmetry in αβ could be established during training in which αβc only receive appetitive reinforcement. Rewarding DA neurons that innervate the β lobe tip ramify throughout the βs and βc, whereas aversive reinforcing DA neurons appear restricted to the αβs. Consistent with this organization of memory formation, aversive MB-V2α output neurons have dendrites biased toward αs, whereas the dendrites of aversive or appetitive MB-V3 output neurons are broadly distributed throughout the α lobe tip. Therefore, a model is proposed that learned odor aversion is driven by αβs neurons, whereas learned approach comes from pooling inputs from the αβs and αβc neurons (Perisse, 2013).

Another property that distinguishes appetitive from aversive memory retrieval is state dependence; flies only efficiently express appetitive memory if they are hungry. Prior work has shown that the dopaminergic MB-MP1 neurons are also critical for this level of control. Since the MB-MP1 neurons more densely innervate the αβs than αβc, it would seem that satiety state differentially tunes the respective drive from parts of the αβ ensemble to promote or inhibit appetitive memory retrieval (Perisse, 2013).

Go to top

Convergence of multimodal sensory pathways to the mushroom body calyx in Drosophila melanogaster

Yagi, R., Mabuchi, Y., Mizunami, M. and Tanaka, N. K. (2016). Sci Rep 6: 29481. PubMed ID: 27404960

Detailed structural analyses of the mushroom body which plays critical roles in olfactory learning and memory revealed that it is directly connected with multiple primary sensory centers in Drosophila, for example the γ lobe neurons innervating the ventral accessory calyces respond to visual stimuli, the antennal lobe tracts neuron terminating in the lateral accessory calyces shows calcium responses to temperature shifts, and taste activity has been observed in the dorsal accessory calyces. Connectivity patterns between the mushroom body and primary sensory centers suggest that each mushroom body lobe processes information on different combinations of multiple sensory modalities. This finding provides a novel focus of research by Drosophila genetics for perception of the external world by integrating multisensory signals (Yagi, 2016).

Go to top

Montagnese, M., McLachlan, A. D., O'Kane, C. J., Zhang, S. W. and Masuda-Nakagawa, L. M. (2021)

Wong, J. Y. H., Wan, B. A., Bland, T., Montagnese, M., McLachlan, A. D., O'Kane, C. J., Zhang, S. W. and Masuda-Nakagawa, L. M. (2021). Octopaminergic neurons have multiple targets in Drosophila larval mushroom body calyx and can modulate behavioral odor discrimination. Learn Mem 28(2): 53-71. PubMed ID: 33452115

Discrimination of sensory signals is essential for an organism to form and retrieve memories of relevance in a given behavioral context. Sensory representations are modified dynamically by changes in behavioral state, facilitating context-dependent selection of behavior, through signals carried by noradrenergic input in mammals, or octopamine (OA) in insects. To understand the circuit mechanisms of this signaling, this study characterized the function of two OA neurons, sVUM1 neurons, that originate in the subesophageal zone (SEZ) and target the input region of the memory center, the mushroom body (MB) calyx, in larval Drosophila. sVUM1 neurons were found to target multiple neurons, including olfactory projection neurons (PNs), the inhibitory neuron APL, and a pair of extrinsic output neurons, but relatively few mushroom body intrinsic neurons, Kenyon cells. PN terminals carried the OA receptor Oamb, a Drosophila α1-adrenergic receptor ortholog. Using an odor discrimination learning paradigm, this study showed that optogenetic activation of OA neurons compromised discrimination of similar odors but not learning ability. These results suggest that sVUM1 neurons modify odor representations via multiple extrinsic inputs at the sensory input area to the MB olfactory learning circuit (Wong, 2021).

Go to top

Octopamine and Dopamine differentially modulate the nicotine-induced calcium response in Drosophila Mushroom Body Kenyon Cells

Leyton, V., Goles, N. I., Fuenzalida-Uribe, N. and Campusano, J. M. (2013). Neurosci Lett 560: 16-20. PubMed ID: 24334164

In Drosophila associative olfactory learning, an odor, the conditioned stimulus (CS), is paired to an unconditioned stimulus (US). The CS and US information arrive at the Mushroom Bodies (MB), a Drosophila brain region that processes the information to generate new memories. It has been shown that olfactory information is conveyed through cholinergic inputs that activate nicotinic acetylcholine receptors (nAChRs) in the MB, while the US is coded by biogenic amine (BA) systems that innervate the MB. In this regard, the MB acts as a coincidence detector. A better understanding of the properties of the responses gated by nicotinic and BA receptors are required to get insights on the cellular and molecular mechanisms responsible for memory formation. In recent years, information has become available on the properties of the responses induced by nAChR activation in Kenyon Cells (KCs), the main neuronal MB population. However, very little information exists on the responses induced by aminergic systems in fly MB. This study evaluated some of the properties of the calcium responses gated by Dopamine (DA) and Octopamine (Oct) in identified KCs in culture. Exposure to BAs induces a fast but rather modest increase in intracellular calcium levels in cultured KCs. The responses to Oct and DA are fully blocked by a Voltage-gated Calcium Channel (VGCC) blocker, while they are differentially modulated by cAMP. Moreover, co-application of BAs and nicotine has different effects on intracellular calcium levels: while DA and nicotine effects are additive, Oct and nicotine induce a synergistic increase in calcium levels. These results suggest that a differential modulation of nicotine-induced calcium increase by DA and Oct could contribute to the events leading to learning and memory in flies (Leyton, 2013).

Go to top

Aversive learning and appetitive motivation toggle feed-forward inhibition in the Drosophila mushroom body

Perisse, E., Owald, D., Barnstedt, O., Talbot, C.B., Huetteroth, W. and Waddell, S. (2016). Neuron [Epub ahead of print]. PubMed ID: 27210550

In Drosophila, negatively reinforcing dopaminergic neurons also provide the inhibitory control of satiety over appetitive memory expression. This study shows that aversive learning causes a persistent depression of the conditioned odor drive to two downstream feed-forward inhibitory GABAergic interneurons of the mushroom body, called MVP2, or mushroom body output neuron (MBON)-γ1pedc>α/β. However, MVP2 neuron output is only essential for expression of short-term aversive memory. Stimulating MVP2 neurons preferentially inhibits the odor-evoked activity of avoidance-directing MBONs and odor-driven avoidance behavior, whereas their inhibition enhances odor avoidance. In contrast, odor-evoked activity of MVP2 neurons is elevated in hungry flies, and their feed-forward inhibition is required for expression of appetitive memory at all times. Moreover, imposing MVP2 activity promotes inappropriate appetitive memory expression in food-satiated flies. Aversive learning and appetitive motivation therefore toggle alternate modes of a common feed-forward inhibitory MVP2 pathway to promote conditioned odor avoidance or approach (Perisse, 2016).

Prior work in Drosophila indicated that negative reinforcement and hunger-state-dependent motivational control of appetitive memory performance might be controlled by the same dopaminergic neurons (DANs). The presynaptic field of the MP1/PPL1-γ1pedc DANs occupies a defined region of the MB that also contains the MVP2/MBON-γ1pedc >αβ dendrites, suggesting that these DANs modulate the efficacy of this specific KC-MBON connection. The results of the current study demonstrate that the MVP2 MBONs also play a critical role in the expression of short-term aversive memory and the state-dependence of appetitive memory expression. Since these findings directly mirror the described roles for the MP1 DANs, it is concluded that DAN modulation of the KC-MVP2 junction is critical for both negative reinforcement during olfactory learning and the motivational salience of appetitive odor cues (Perisse, 2016).

The GABA-ergic MVP2 neurons have postsynaptic and presynaptic processes in the MB, suggesting that they are interneurons of the MB and feed-forward inhibit other MBON compartments. Dendrites of MVP2 neurons (and the presynaptic terminals of the MP1 DANs) innervate the γ1 region and more densely innervate the αβs than the αβ core (αβc) region of the αβ ensemble. MVP2 are therefore likely to be primarily driven by αβs KCs. Since αβs neurons contribute to conditioned approach and avoidance, whereas αβc are particularly important for conditioned approach (Perisse, 2013), there is an imbalance in the drive to approach and avoidance behaviors at this level of the MBON network (Perisse, 2016).

Artificial activation of MVP2 neurons in naive flies drives approach behavior, consistent with them preferentially inhibiting MBON compartments that direct avoidance -- as opposed to those that drive approach. Anatomical and functional connectivity and odor-directed behavioral data are consistent with such a model. MVP2 stimulation inhibits odor-evoked activity in M4/6 but not in V2αV2α' MBONs. MVP2 stimulation also promotes expression of approach memory in food-satiated flies, yet it inhibits naive odor avoidance behavior. It is concluded that MVP2 directly inhibit the M4/6 class of horizontal lobe MBONs through synapses made on the primary axonal segment as it exits the MB lobes. Inhibition exerted in this area might be expected to control the gain of the MBON responses following integration of KC inputs in the MBON dendrite in a manner similar to perisomatic inhibition in mammals. Consistent with this anatomy and idea, no obvious changes were found in the odor drive to the dendritic region of M4/6 neurons between hungry and satiated flies, but a hunger-dependent decrease was apparent when odor-evoked responses were measured in the efferent neurites. In contrast, MVP2 neurons do not functionally inhibit or densely innervate the neurites of V2αV2α' MBONs, nor does hunger reduce odor-evoked responses in V2αV2α' MBONs. It therefore seems likely that MVP2 neurons contact DANs or other neurons that occupy the α2 compartment of the MB lobe (Perisse, 2016).

The data also demonstrate that aversive learning reduces the relative conditioned odor drive to MVP2 neurons, which would presumably decrease feed-forward inhibition onto the relevant MBON compartments and thereby render them more responsive to odors. Output from the glutamatergic M4/6 neurons, which are postsynaptic to the KCs in the horizontal tip regions, is required for expression of aversive and appetitive memory. Furthermore, the relative odor-drive to M4/6 neurons was shown to be depressed by reward learning and potentiated by aversive learning. Since aversive learning reduces the conditioned odor drive of the MVP2 neuron, it is proposed that the observed increase in odor-drive to M4/6 after aversive learning results from reduced feed-forward inhibition from MVP2. This would mean that bi-directional output plasticity could emerge via a direct junctional plasticity following reward conditioning, but a network property of reduced MVP2 feed-forward inhibition after aversive conditioning. Such a layered feed-forward network architecture linking one site of DAN-driven KC-MBON plasticity to another KC-MBON connection would provide a means to achieve odor-specific bi-directional plasticity at a particular synaptic junction using dopamine-driven synaptic depression in two different places. It is proposed that this circuit design principle in which plasticity at one site of a neuron can, via feed-forward inhibition, indirectly alter the efficacy of output elsewhere in the same neuron, could be a general feature in the brain of the fly and other animals. It is possible that the KC-MVP2 junction also exhibits bi-directional plasticity, notably with inverted polarity relative to M4/6 plasticity traces (Perisse, 2016).

The layered network architecture places the aversive memory relevant MVP2 plasticity on top of the M4/6 plasticity that is relevant for appetitive memory. This organization could accommodate the co-existence of aversive and appetitive olfactory memories following conditioning reinforced by sugar laced with bitter taste. Immediately after such training flies avoid the conditioned odor because the aversive taste memory relieves feed-forward inhibition onto the sites that are depressed by appetitive sugar plasticity and therefore over-rides the expression of approach memory. However, as the aversive memory decays, feed-forward inhibition returns and appetitive memory is then expressed. A similar mechanism might account for the time-dependent switch from conditioned aversion to approach following odor conditioning reinforced by alcohol. It is notable that learning-induced plasticity of relative odor-drive to MVP2 persists for at least 3 hr after training whereas output from MVP2 is dispensable for the expression of aversive memory at that time. Since expression of different phases of aversive memory requires distinct combinations of MBON pathways, it is proposed that more persistent MVP2 plasticity might provide a permissive gate for both the formation of aversive memory in, and the expression from, other parts of the MBON network. This would be reminiscent of fear conditioning in the neural circuitry of the mouse amygdala, where dopamine suppresses feed-forward GABA-ergic inhibition from local interneurons to facilitate the induction of long-term potentiation (Perisse, 2016).

MVP2 neuron output is required for the expression of sugar-reinforced approach memory at all times. Moreover, odors evoked larger MVP2 responses in hungry than in food-satiated flies, and elevating MVP2 activity in satiated flies promoted inappropriate expression of appetitive memory. These results are consistent with the model that hunger generally increases feed-forward inhibition through MVP2 to support appetitive memory expression). This result is also the mirror-image of that with MP1 DANs whose activity increases when the flies are satiated and whose inhibition leads to the expression of appetitive memory in satiated flies. Taken with prior work, it is therefore proposed that hunger increases dNPF, which releases MP1 inhibition over the KC-MVP2 connection. This results in an increase of odor-evoked MVP2 feed-forward inhibition onto the MBON compartments such as M4/6 that contain the KC-MBON synapses that are directly modified by appetitive conditioning. The increase of MVP2 inhibition into these, and other, compartments allows more efficient expression of the appetitive memory-directed approach behavior by effectively raising the motivational salience of learned food-related odors. Appetitive conditioning may also increase odor-specific recruitment of MVP2 feed-forward inhibition (Perisse, 2016).

Two DA receptors that share homology to vertebrate DA type 1 receptors are expressed in Drosophila MB, DAMB (Dopamine 1-like receptor 2 or Dop1R2) and dDA1/DmDOP1/DopR. These receptors have been cloned and expressed in heterologous systems, where they are positively coupled to AC. Moreover, these receptors participate in the generation or modification of new olfactory memories in MB. The third cloned DA receptor, Dop2R/D2R, is not expressed in Drosophila MB. The current results are consistent with the general idea that DA receptors modulate intracellular cAMP levels, which could lead to the modification of the activity of VGCCs in KCs. The EC50 describe in this study is in agreement with previous studies in heterologous systems, which report EC50 in the range of 300-500 nM for both DA type 1 receptors. Thus, it is very likely that DAMB and/or dDA1/DmDOP1 are contributing to the DA-induced calcium response in Drosophila KCs. It would be expected that DA receptors increase cAMP levels to activate calcium currents in fly MB KCs. However, data using SQ22536 suggest that cAMP inhibits calcium fluxes in KCs. Although the cellular mechanisms responsible for this effect are not evident, it has been previously shown that increased cAMP signaling negatively modulates the activity of VGCCs through the activation of specific phosphatases in the vertebrate Nucleus Accumbens. Remarkably, the Nucleus Accumbens and MB are brain regions highly associated to the plastic behavioral effects induced by addictive drugs. Thus, it would not be particularly surprising to find similarities in the mechanisms responsible for the modulation of neuronal communication and excitability byDA in these two brain structures, as previously suggested.On the other hand, several Oct receptors have been previously cloned in Drosophila: one Oct receptor with high sequence homology to vertebrate -type receptors (Oct1R/OAMB) is expressed in dendrites and axons of the MB, and is the main candidate for the calcium responses induced by Oct in KCs, since the other cloned Oct receptors are not expressed in MB. In agreement with this, the calculated EC50, and the description that the response depends on VGCC activation and is independent on cAMP, further support this suggestion (Leyton, 2013).

An interaction of the neural systems responsible for CS and US stimuli in the MB region could mediate the generation of new memories. This interaction could occur at the presynaptic level, for instance, through the nAChR modulation of aminergic innervation to the MB region. However, the most accepted idea is that cholinergic and aminergic receptors expressed in MB KCs gate intracellular cascades that could cross-talk to modify the activity of KCs, a cellular event that could underlie long-lasting changes responsible for the generation of new olfactory memories in the fly (Leyton, 2013).

Go to top

An octopamine-mushroom body circuit modulates the formation of anesthesia-resistant memory in Drosophila

Wu, C. L., Shih, M. F., Lee, P. T. and Chiang, A. S. (2013). Curr Biol 23(23): 2346-54. PubMed ID: 24239122

Drosophila olfactory aversive conditioning produces two components of intermediate-term memory: anesthesia-sensitive memory (ASM) and anesthesia-resistant memory (ARM). Recently, the anterior paired lateral (APL) neuron innervating the whole mushroom body (MB) has been shown to modulate ASM via gap-junctional communication in olfactory conditioning. Octopamine (OA), an invertebrate analog of norepinephrine, is involved in appetitive conditioning, but its role in aversive memory remains uncertain. This study shows that chemical neurotransmission from the anterior paired lateral (APL) neuron, after conditioning but before testing, is necessary for aversive ARM formation. The APL neurons are tyramine, Tβh, and OA immunopositive. An adult-stage-specific RNAi knockdown of Tβh in the APL neurons or Octβ2R OA receptors in the MB α'β' Kenyon cells (KCs) impaired ARM. Importantly, an additive ARM deficit occurred when Tβh knockdown in the APL neurons was in the radish mutant flies or in the wild-type flies with inhibited serotonin synthesis. It is concluded that OA released from the APL neurons acts on α'β' KCs via Octβ2R receptor to modulate Drosophila ARM formation. Additive effects suggest that two parallel ARM pathways, serotoninergic DPM-αβ KCs and octopaminergic APL-α'β' KCs, exist in the MB (Wu, 2013).

The key finding of this study is that OA from the single APL neuron innervating the entire MB is required specifically for ARM formation in aversive olfactory conditioning in Drosophila. This conclusion is supported by five independent lines of evidence. First, blocking neurotransmission from APL neurons after training, but before testing, impaired ARM. Second, the APL neurons are tyramine, Tβh, and OA antibody immunopositive. Third, adult-stage-specific reduction of Tβh levels in the APL neurons, but not in dTdc2-GAL4 neurons that do not include the APL neurons, specifically abolishes ARM without affecting learning or ASM. Fourth, Octβ2R is expressed preferentially in the α'β' lobes, and adult-stage-specific reduction of Octβ2R expression in the α'β' KCs impaired ARM. Fifth, the additive memory impairments demonstrated in flies subjected to Tβh plus inx7 knockdowns and Tβh knockdown plus cold shock, but not inx7 knockdown plus cold shock, confirm that a single APL neuron modulates both ASM and ARM through gap-junctional communication and OA neurotransmission, respectively. Although it has been shown that the APL neurons are also GABAergic, the current results showed that OA is the primary neurotransmitter from the APL neurons involved in ARM formation because reduced GABA levels induced by Gad1RNAi inhibition in the APL neurons did not affect 3 hr memory (Wu, 2013).

In Drosophila olfactory memories, OA and dopamine have been shown to act as appetitive and aversive US reinforcements, respectively. It is important to point out that the original claim that Tβh plays no role in aversive learning only examined 3 min memory, not 3 hr memory or ARM. It is not surprising to find that OA modulates ARM in aversive memory because dopamine has also been attributed to diverse memory roles, including a motivation switch for appetitive ITM and appetitive reinforcement. Intriguingly, dopamine negatively inhibits ITM formation, but OA positively modulates ARM formation (Wu, 2013).

Food deprivation in Drosophila larvae induces behavioral plasticity and the growth of octopaminergic arbors via Octβ2R-mediated cyclic AMP (cAMP) elevation in an autocrine fashion. This study showed that the APL neurons release OA acting on the Octβ2R-expressing α' β' KCs for ARM, instead of inducing autocrine regulation. Applying OA directly onto the adult brain results in an elevation of cAMP levels in the whole MB, and OA has been shown to upregulate protein kinase A (PKA) activity in the MBs. Intriguingly, ARM is enhanced by a decreased PKA activity and requires DUNCE-sensitive cAMP signals. It is speculated that APL-mediated activation of Octβ2R may lead to an intricate regulation of cAMP in the α' β' KCs for ARM formation. (Wu, 2013).

Although it has generally been assumed that, in a particular neuron, the same neurotransmitter is used at all synapses, exceptions continue to accumulate in both vertebrates and invertebrates. Scattered evidence suggests that co-release may be regulated at presynaptic vesicle filling and postsynaptic activation of receptors, but the physiologic significance remains poorly understood. This study reports that the APL neurons co-release GABA and OA. In the APL neurons, a reduced GABA level affects learning, but not ITM, whereas a reduced OA level has no effect on learning, but impairs ITM, suggesting that the two neurotransmitters are regulated in different ways in the same cell (Wu, 2013).

It has been proposed that the APL neurons might be the Drosophila equivalent of the honeybee GABAergic feedback neurons, receiving odor information from the MB lobes and releasing GABA inhibition to the MB calyx. This negative feedback loop for olfactory sparse coding has been supported by electrophysiological recording of the giant GABAergic neuron in locusts. However, the function of Drosophila APL neurons is complicated by the existence of functioning presynaptic processes in the MB lobes, mixed axon-dendrite distribution throughout the whole MB, and GABA/OA cotransmission (Wu, 2013).

Normal performance of ARM behavior requires serotonin from the DPM neurons acting on ab KCs via d5HT1A serotonin receptors and function of RADISH and BRUCHPILOT in the ab KCs. Surprisingly,the current results show that ARM formation also requires OA from the APL neurons acting on the α' β' KCs via Octβ2R OA receptors, suggesting the existence of two distinct anatomical circuits involved in ARM formation. However, it remains uncertain whether two branches of ARM occur in parallel because combination of various molecular disruptions (i.e., TβhRNAi and pCPA feeding/rsh1 mutant) did not completely abolish ARM and partial disruption of one anatomical circuit will allow additive effects of another treatment even if they act on the same ARM. The hypothesis of the existence of two distinct forms of ARM is favored based on the following observations. First, neither d5HT1ARNAi knockdown in α' β' KCs nor Octβ2RRNAi knockdown in ab KCs affects ARM, suggesting that the two signaling pathways act separately in different KCs and do not affect each other in the same KCs. Second, each of the three ways of molecular disruption (i.e., TβhRNAi, pCPA feeding, and rsh1 mutant) results in a similar degree of ARM impairment, but additive effect did not occur in rsh1 mutant flies fed with pCPA and was evident when TβhRNAi treatment combines with either pCPA feeding or rsh1 mutant. It's noteworthy that ARM is also affected by dopamine modulation because calcium oscillation within dopaminergic MB-MP1 and MB-MV1 neurons controls ARM and gates long-term memory, albeit a different view has been brought up. The target KCs of these dopaminergic neurons on ARM remain to be addressed (Wu, 2013).

Both the APL and DPM neurons are responsive to electric shock and multiple odorants, suggesting that they likely acquire olfactory associative information during learning for subsequent ARM formation. However, the DPM neurons may receive ARM information independently because their fibers are limited within MB lobes and gap-junctional communications between the APL and DPM neurons are specifically required for the formation of ASM, but not ARM. Given that all dopamine reinforcement comes in via the γ KCs, it is possible that the DPM neurons obtain ARM information from γ KCs. Together, these data suggest that two parallel neural pathways, serotoninergic DPM-αβ KCs and octopaminergic APL-α'β' KCs, modulate 3 hr ARM formation in the MB (Wu, 2013).

Go to top

Dopaminergic Modulation of cAMP Drives Nonlinear Plasticity across the Drosophila Mushroom Body Lobes

Boto, T., Louis, T., Jindachomthong, K., Jalink, K. and Tomchik, S. M. (2014). Curr Biol 24: 822-831. PubMed ID: 24684937

Activity of dopaminergic neurons is necessary and sufficient to evoke learning-related plasticity in neuronal networks that modulate learning. During olfactory classical conditioning, large subsets of dopaminergic neurons are activated, releasing dopamine across broad sets of postsynaptic neurons. It is unclear how such diffuse dopamine release generates the highly localized patterns of plasticity required for memory formation. This study has mapped spatial patterns of dopaminergic modulation of intracellular signaling and plasticity in Drosophila mushroom body (MB) neurons, combining presynaptic thermogenetic stimulation of dopaminergic neurons with postsynaptic functional imaging in vivo. Stimulation of dopaminergic neurons generated increases in cyclic AMP (cAMP) across multiple spatial regions in the MB. However, odor presentation paired with stimulation of dopaminergic neurons evoked plasticity in Ca2+ responses in discrete spatial patterns. These patterns of plasticity correlated with behavioral requirements for each set of MB neurons in aversive and appetitive conditioning. Finally, broad elevation of cAMP differentially facilitated responses in the gamma lobe, suggesting that it is more sensitive to elevations of cAMP and that it is recruited first into dopamine-dependent memory traces. These data suggest that the spatial pattern of learning-related plasticity is dependent on the postsynaptic neurons' sensitivity to cAMP signaling. This may represent a mechanism through which single-cycle conditioning allocates short-term memory to a specific subset of eligible neurons (gamma neurons) (Boto, 2014).

Dopaminergic neurons are involved in modulating diverse behaviors, including learning, motor control, motivation, arousal, addiction and obesity, and salience-based decision making. In Drosophila, dopaminergic neurons innervate multiple brain regions, including the mushroom body (MB), where they modulate aversive learning, forgetting, state-dependent modulation of appetitive memory retrieval, expression of ethanol-induced reward memory, and temperature-preference behavior (Boto, 2014).

Dopaminergic circuits play a particularly critical role in memory acquisition. During olfactory classical conditioning, where an odor (conditioned stimulus [CS]) is paired with an aversive event (e.g., electric shock; the unconditioned stimulus [US]), dopaminergic neurons respond strongly to the aversive US (Mao, 2009). Dopamine functions in concert with activity-dependent Ca2+ influx to synergistically elevate cyclic AMP (cAMP) (Tomchik, 2009) and PKA (Gervasi, 2010), suggesting that dopamine is one component of a molecular coincidence detector underlying learning. Proper dopamine signaling is necessary for aversive and appetitive memory. Moreover, driving activity of a subset of TH-GAL4+ dopaminergic neurons that differentially innervates the vertical α/α' MB lobes (with less dense innervation of the horizontal β/β'/γ lobes, peduncle, and calyx), is sufficient to induce behavioral aversion to a paired odorant in larvae and adult flies. Conversely, stimulation of a different set of Ddc-GAL4+ dopaminergic neurons, the PAM cluster that innervates mainly the horizontal β/β'/γ lobes, is sufficient to induce behavioral attraction to a paired odorant. Thus, dopaminergic neurons comprise multiple circuits with distinct roles in memory acquisition (Boto, 2014).

Multiple subsets of MB neurons receive CS and US information and express molecules associated with the coincidence detection, making them theoretically eligible to generate dopamine/cAMP-dependent plasticity. Yet only some subsets are required to support memory at any given time following conditioning, leaving open the question of how spatial patterns of plasticity are generated during conditioning. This question has been approached, by using a technique to probe the postsynaptic effects of neuronal pathway activation. Odor presentation was paired with stimulation of presynaptic dopaminergic neurons via ectopic expression of the heat-sensitive channel TRPA1, while monitoring postsynaptic effects with genetically encoded optical reporters for Ca2+, cAMP, and PKA in vivo (Boto, 2014).

The present data demonstrate four major points about how dopaminergic circuits function in neuronal plasticity underlying olfactory classical conditioning. (1) Stimulation of small subsets of dopaminergic neurons evokes consistent, compartmentalized elevations of cAMP across the MB lobes. (2) Broad stimulation of dopaminergic neurons generates broad postsynaptic elevation of cAMP, but Ca2+ response plasticity occurs in discrete spatial regions. (3) Stimulation of TH-GAL4+ neurons and Ddc/R58E02-GAL4+ neurons, which mediate opposing behavioral responses to conditioned stimuli, generates an overlapping pattern of Ca2+ response plasticity in the γ lobe, with additional regions recruited by Ddc/R58E02-GAL4+ stimulation. Finally, (4) the spatial pattern of plasticity coincides with differential sensitivity to cAMP in the γ lobe. Collectively, these data suggest that different subsets of neurons exhibit heterogeneous sensitivity to activation of second messenger signaling cascades, which might shape their responses to neuromodulatory network activity and modulate their propensity for recruitment into memory traces (Boto, 2014).

The data suggest that dopaminergic neurons mediate Ca2+ response plasticity largely in the γ lobe and suggest a potential mechanism for localization of short-term, learning-related plasticity. These data coincide with multiple previous studies that have demonstrated a critical role of γ neurons in short-term memory. Rescue of Rutabaga (Rut) in the γ lobe of rut mutants is sufficient to restore performance in short-term memory, whereas rescue in α/β lobes supports long-term memory. Rescue of the D1-like DopR receptor in the γ lobe is sufficient to rescue both short- and long-term memory in a mutant background, suggesting that the γ neurons mediate the dopaminergic input during conditioning. In addition, stimulating MP1 dopaminergic neurons innervating the heel of the γ lobe is sufficient as an aversive reinforcer. Finally, learning induces plasticity in synaptic vesicle release from MB γ lobes, which depends in part on G(o) signaling (Zhang, 2013). The data support a critical role for the γ lobe in short-term memory. Furthermore, the observation of differential sensitivity of the γ lobe to cAMP might provide an elegant explanation for why it is specifically recruited into short-term memory traces (Boto, 2014).

Direct elevation of cAMP was sufficient to generate localized, concentration-dependent Ca2+ response plasticity in the MB γ lobe in these experiments. Because applying forskolin in the bath is expected to elevate cAMP across the brain, the spatial specificity of the effect is remarkable. This was not an acute effect, because the forskolin was washed out before imaging the first postconditioning odor response. At the concentrations that were tested, only the γ lobe was facilitated. Therefore, it is concluded that the γ lobe is most sensitive to elevation of cAMP, which has the effect of differentially recruiting γ neurons into the representation of short-term memory via dopamine-mediated neuronal plasticity. It is possible that additional signaling cascades are involved in generating learning-related plasticity in α/β and α'/β' neurons, given that no Ca2+ response plasticity was observed in those neurons following forskolin application (Boto, 2014).

The dominant model for cellular mechanisms of olfactory associative learning is that integration of information about the conditioned and unconditioned stimuli are integrated by Rut, which functions as a molecular coincidence detector. This would suggest that MB neurons, which receive CS and US information, would exhibit at least somewhat uniform Ca2+ response plasticity. From this molecular and cellular perspective, the finding that the α/β and α'/β' neurons did not exhibit Ca2+ response plasticity when an odor was paired with stimulation of dopaminergic neurons is surprising. These neurons are theoretically eligible to encode memory, because they receive information about the CS and US. However, the finding that γ neurons differentially exhibit dopamine-dependent plasticity following single-cycle conditioning is consistent with the data from the behavioral experiments. In summary, the present results suggest that differential cAMP sensitivity provides a potential mechanism allowing specific subsets of eligible neurons in an array (γ neurons) to differentially encode CS-US coincidence relative to other subsets (α/β neurons) that also receive CS/US information (Boto, 2014).

Go to top

Dopaminergic neurons write and update memories with cell-type-specific rules

Aso, Y. and Rubin, G.M. (2016). Elife [Epub ahead of print]. PubMed ID: 27441388

Associative learning is thought to involve parallel and distributed mechanisms of memory formation and storage. In Drosophila, the mushroom body (MB) is the major site of associative odor memory formation. The anatomy of the adult MB and 20 types of dopaminergic neurons (DANs) that each innervate distinct MB compartments have been previously defined and described. This study compared the properties of memories formed by optogenetic activation of individual DAN cell types. Extensive differences were found in training requirements for memory formation, decay dynamics, storage capacity and flexibility to learn new associations. Even a single DAN cell type can either write or reduce an aversive memory, or write an appetitive memory, depending on when it is activated relative to odor delivery. These results show that different learning rules are executed in seemingly parallel memory systems, providing multiple distinct circuit-based strategies to predict future events from past experiences. The mechanisms that generate these distinct learning rules are unknown. They could arise from differences in the dopamine release properties of different DAN cell types or from local differences in the biochemical response to dopamine signaling in each MB compartment. For example, KCs express four distinct dopamine receptors, which might be deployed differently in each compartment. Or they could originate from circuit properties (Aso, 2016).

Go to top

Circuit analysis of a Drosophila Dopamine type 2 receptor that supports anesthesia-resistant memory

Scholz-Kornehl, S. and Schwarzel, M. (2016). J Neurosci 36: 7936-7945. PubMed ID: 27466338

Dopamine is central to reinforcement processing and exerts this function in species ranging from humans to fruit flies. It can do so via two different types of receptors (i.e., D1 or D2) that mediate either augmentation or abatement of cellular cAMP levels. Whereas D1 receptors are known to contribute to Drosophila aversive odor learning per se, this study shows that D2 receptors are specific for support of a consolidated form of odor memory known as anesthesia-resistant memory. By means of genetic mosaicism, this function was localized to Kenyon cells, the mushroom body intrinsic neurons, as well as GABAergic APL neurons and local interneurons of the antennal lobes, suggesting that consolidated anesthesia-resistant memory requires widespread dopaminergic modulation within the olfactory circuit. Additionally, dopaminergic neurons themselves require D2R, suggesting a critical role in dopamine release via its recognized autoreceptor function. Considering the dual role of dopamine in balancing memory acquisition (proactive function of dopamine) and its 'forgetting' (retroactive function of dopamine), this analysis suggests D2R as central player of either process (Scholz-Kornehl, 2016).

Go to top

Pre- and postsynaptic role of Dopamine D2 Receptor DD2R in Drosophila olfactory associative learning

Qi, C. and Lee, D. (2014). Biology (Basel) 3: 831-845. PubMed ID: 25422852

Dopaminergic neurons in Drosophila play critical roles in diverse brain functions such as motor control, arousal, learning, and memory. Using genetic and behavioral approaches, it has been firmly established that proper dopamine signaling is required for olfactory classical conditioning (e.g., aversive and appetitive learning). Dopamine mediates its functions through interaction with its receptors. There are two different types of dopamine receptors in Drosophila: D1-like (dDA1, DAMB) and D2-like receptors (DD2R). Currently, no study has attempted to characterize the role of DD2R in Drosophila learning and memory. Using a DD2R-RNAi transgenic line, this study has examined the role of DD2R, expressed in dopamine neurons (i.e., the presynaptic DD2R autoreceptor), in larval olfactory learning. The function of postsynaptic DD2R expressed in mushroom body (MB) was also studied as MB is the center for Drosophila learning, with a function analogous to that of the mammalian hippocampus. These results showed that suppression of presynaptic DD2R autoreceptors impairs both appetitive and aversive learning. Similarly, postsynaptic DD2R in MB neurons appears to be involved in both appetitive and aversive learning. The data confirm, for the first time, that DD2R plays an important role in Drosophila olfactory learning (Qi, 2014).

Go to top

Reward signal in a recurrent circuit drives appetitive long-term memory formation

Ichinose, T., Aso, Y., Yamagata, N., Abe, A., Rubin, G. M. and Tanimoto, H. (2015). Elife 4 [Epub ahead of print]. PubMed ID: 26573957

Dopamine signals reward in animal brains. A single presentation of a sugar reward to Drosophila activates distinct subsets of dopamine neurons that independently induce short- and long-term olfactory memories (STM and LTM, respectively). This study shows that a recurrent reward circuit underlies the formation and consolidation of LTM. This feedback circuit is composed of a single class of reward-signaling dopamine neurons (PAM-alpha1) projecting to a restricted region of the mushroom body (MB), and a specific MB output cell type, MBON-α1, whose dendrites arborize that same MB compartment. Both MBON-α1 and PAM-α1 neurons are required during the acquisition and consolidation of appetitive LTM. MBON-α1 additionally mediates the retrieval of LTM, which is dependent on the dopamine receptor signaling in the MB αβ neurons. These results suggest that a reward signal transforms a nascent memory trace into a stable LTM using a feedback circuit at the cost of memory specificity (Ichinose, 2015).

Go to top

Distinct dopamine neurons mediate reward signals for short- and long-term memories

Yamagata, N., Ichinose, T., Aso, Y., Placais, P., Friedrich, A. B., Sima, R. J., Preat, T., Rubin, G. M. and Tanimoto, H. (2014). Proc Natl Acad Sci U S A 112(2):578-83. PubMed ID: 25548178

Drosophila melanogaster can acquire a stable appetitive olfactory memory when the presentation of a sugar reward and an odor are paired. However, the neuronal mechanisms by which a single training induces long-term memory are poorly understood. This study shows that two distinct subsets of dopamine neurons in the fly brain signal reward for short-term (STM) and long-term memories (LTM). One subset induces memory that decays within several hours, whereas the other induces memory that gradually develops after training. They convey reward signals to spatially segregated synaptic domains of the mushroom body (MB), a potential site for convergence. Furthermore, a single type of dopamine neuron was identified that conveys the reward signal to restricted subdomains of the mushroom body lobes and induces long-term memory. Constant appetitive memory retention after a single training session thus comprises two memory components triggered by distinct dopamine neurons (Yamagata, 2014).

In Drosophila, sugar ingestion in a single training session induces stable appetitive odor memory. The results showed that observed memory represents the composite of STM and LTM that are induced by the distinct and complementary reward signals of dopamine. These separate reward signals very well corroborate parallel processing of STM and LTM in the MB. The stm-PAM neurons (a specific cluster of dopamine neurons) induce appetitive memory that decays within hours, whereas memory by ltm-PAM gradually develops after training. These dopamine neurons convey reward signals to spatially segregated synaptic domains of the MB, whereas the other PAM cluster neurons may also contribute to LTM reward. Furthermore, this study identified a single type of dopamine neurons (PAM-α1) encoding a reward signal for LTM. The PAM-α1 targets a spatially restricted subdomain in the α-lobe of the MB, suggesting local associative modulation in the α-lobe. These results indicate that sugar ingestion activates multiple reward signals of different qualities to form complementary memories, rather than a single reward system forming STM that later transforms into LTM (Yamagata, 2014).

In the defensive siphon and tail withdrawal reflex of Aplysia, different modes of 5-HT application to the sensory neurons in the pedal and pleural ganglion differentially induce short-term and long-term sensitization memories. As both ganglia are innervated by a single identified serotonergic neuron, it is likely that the tail shock activates the same cell to induce both forms of sensitization independently. This cellular configuration is in strong contrast to the reward system in Drosophila appetitive memory, despite parallel formation of STM and LTM in both systems. The sugar reward may be more intricately encoded in the fly, given the importance of long-lasting food-related memory in survival (Yamagata, 2014).

Representations of different reinforcing stimuli by the same transmitter seems to be variable. In Drosophila aversive memory, reinforcing signals of electric shock and heat punishment converge to the same dopamine neurons, whereas distinct dopamine neurons are recruited for different aversive stimuli in mammals. Appetitive memory of Drosophila is closer to the latter case. However, a critical difference is that stm- and ltm-PAM neurons seem to signal different properties of the same reward, again pointing to more complex representation of the sugar reward. It would be interesting to compare neuronal representations of different rewarding stimuli, such as ethanol and water. An important future question would be to understand physiological mechanisms by which the MB computes the distinct dopamine inputs to control approach behavior (Yamagata, 2014).

Memory induced by ltm-PAM develops gradually after training. This gradual increase may reflect the time to implement learning-dependent molecular changes. Many molecules that are specifically required for LTM have been identified, and some of these molecules are involved in learning-dependent gene transcription/translation. For instance, fasting-dependent LTM that is formed without training repetition requires CREB-regulated transcription coactivator (CRTC)-mediated cAMP response element binding protein (CREB) transcription in the MB. Such a transcription-dependent mechanism takes time and may thus underlie gradual formation of memory induced by the ltm-PAM (Yamagata, 2014).

The complementary memory dynamics as a consequence of the distinct reward signals suggest that the 'intent' of appetitive memory undergoes a transition from palatability to caloric content. Similar temporal transitions have been found in feeding choice, where flies initially choose sugars according to sweet taste, but later prioritize caloric contents. Similarly, ethanol exposure initially acts as an aversive reinforcer, but eventually turns into reward and induces LTM. The sequential regulation of appetitive behavior by the same stimulus may be conserved across relevant appetitive stimuli. As palatability is not always a faithful predictor of its nutritional value, it may be a general design of reward systems to balance short-term benefit and long-term fitness (Yamagata, 2014).

Go to top

Sweet taste and nutrient value subdivide rewarding dopaminergic neurons in Drosophila

Huetteroth, W., Perisse, E., Lin, S., Klappenbach, M., Burke, C. and Waddell, S. (2015). Curr Biol 25(6):751-8. PubMed ID: 25728694

Dopaminergic neurons provide reward learning signals in mammals and insects. Recent work in Drosophila has demonstrated that water-reinforcing dopaminergic neurons are different to those for nutritious sugars. This study tested whether the sweet taste and nutrient properties of sugar reinforcement further subdivide the fly reward system. They found that dopaminergic neurons expressing the OAMB octopamine receptor specifically conveyed the short-term reinforcing effects of sweet taste. These dopaminergic neurons projected to the β'2 and γ4 regions of the mushroom body lobes. In contrast, nutrient-dependent long-term memory required different dopaminergic neurons that project to the γ5b regions, and it could be artificially reinforced by those projecting to the β lobe and adjacent α1 region. Surprisingly, whereas artificial implantation and expression of short-term memory occurred in satiated flies, formation and expression of artificial long-term memory required flies to be hungry. These studies suggest that short-term and long-term sugar memories have different physiological constraints. They also demonstrate further functional heterogeneity within the rewarding dopaminergic neuron population (Huetteroth, 2015).

These results demonstrate that the sweet taste and nutrient properties of sugars are independently processed and reinforce memories of different duration. Sweet taste is transduced through octopaminergic neurons whose released octopamine, via the OAMB receptor, activates dopaminergic neurons that project to the β'2am and γ4 regions of the mushroom body. Octopaminergic reinforcement also modulates the state dependence of STM via the OCTβ2R receptor that is required in the dopaminergic MB-MP1 neurons (Huetteroth, 2015).

Nutrient-dependent LTM does not involve octopamine or sweet-taste-reinforcing dopaminergic neurons. Nutrient reinforcement instead requires dopaminergic neurons innervating γ5b of the mushroom body, whereas those going to β1, β2, and the adjacent α1 region are sufficient. More work will be required to understand this distributed process, which apparently has an immediate and delayed dynamic (Huetteroth, 2015).

Whereas formation and expression of sweet-taste-reinforced STM is insensitive to satiety state, artificial formation and expression of nutrient-relevant memory require flies to be hungry. Even direct stimulation of the relevant rewarding dopaminergic neurons cannot implant appetitive LTM in food-satiated flies. These experiments suggest that hunger establishes an internal state that permits the nutrient-reinforcing signals to be effective. It will be interesting to understand what the permissive state involves and where it is required. Others have previously described a role for CREB-regulated transcription co-activator 1 (CRTC) in enabling hunger-dependent LTM in the fly and promoting persistent memory in the mouse. It therefore seems plausible that such a mechanism might be required in the mushroom body neurons to permit nutrient-dependent reinforcement (Huetteroth, 2015).

Go to top

Differential coding of absolute and relative aversive value in the Drosophila brain

Villar, M. E., Pavao-Delgado, M., Amigo, M., Jacob, P. F., Merabet, N., Pinot, A., Perry, S. A., Waddell, S. and Perisse, E. (2022). Curr Biol. PubMed ID: 36103878

Animals use prior experience to assign absolute (good or bad) and relative (better or worse) value to new experience. These learned values guide appropriate later decision making. Even though understanding of how the valuation system computes absolute value is relatively advanced, the mechanistic underpinnings of relative valuation are unclear. This study uncovered mechanisms of absolute and relative aversive valuation in Drosophila. Three types of punishment-sensitive dopaminergic neurons (DANs) respond differently to electric shock intensity. During learning, these punishment-sensitive DANs drive intensity-scaled plasticity at their respective mushroom body output neuron (MBON) connections to code absolute aversive value. In contrast, by comparing the absolute value of current and previous aversive experiences, the MBON-DAN network can code relative aversive value by using specific punishment-sensitive DANs and recruiting a specific subtype of reward-coding DANs. Behavioral and physiological experiments revealed that a specific subtype of reward-coding DAN assigns a "better than" value to the lesser of the two aversive experiences. This study therefore highlights how appetitive-aversive system interactions within the MB network can code and compare sequential aversive experiences to learn relative aversive value (Villar, 2022)

Value-based decisions require animals to make choices between several options based on a prediction of their relative subjective value learned through prior experience. Associative learning provides a means to assign absolute (good or bad) values to experience that can be used to guide future approach or avoidance behaviors. During learning, animals can also compare the value of their current experience with that of prior knowledge and assign a relative value (better or worse) between these experiences to promote more accurate economic-based choices. Notwithstanding that a substantial body of research has investigated mechanisms for relative reward-value coding, less is known about how relative aversive value is computed during learning to guide appropriate value-based decisions. Reinforcement learning models propose that learning occurs when actual outcome value differs from predicted value. This process and the error computed between actual and predicted value are driven by a valuation circuit that includes dopaminergic and GABAergic neurons (Villar, 2022)

However, it is unclear whether similar circuitry compares current and previous experience to assign relative aversive value to sensory stimuli during learning. Anatomically discrete dopaminergic neurons (DANs) in Drosophila and mice provide either positive or negative teaching signals]. In flies, these different DANs project to unique compartments of the mushroom body (MB), a central brain structure essential for olfactory learning and memory as well as several goal-directed behaviors. DANs from the protocerebral posterior lateral 1 (PPL1) cluster projecting to the vertical and proximal horizontal lobes of the MB relay punishment and signal negative value during learning. Many DANs from the protocerebral anterior medial (PAM) cluster projecting to the horizontal lobes of the MB assign positive reward value during learning (Villar, 2022)

Sparse activation within the ~4,000 MB intrinsic Kenyon cells (KCs), which indirectly receive odorant information from sensory neurons in the periphery, provides the specificity of olfactory memories. KCs synapse onto mushroom body output neurons (MBONs), which project into downstream structures to drive (for most of them) approach or avoidance behavior. Some MBONs are synaptically interconnected, providing cross-excitation or -inhibition between MB compartments. Lastly, many MBONs make feedback or feedforward synapses outside the MB onto DAN dendrites (Villar, 2022)

During olfactory associative learning, specific DANs releasing dopamine in individual MB compartments depress synaptic strengths between sparse odor-activated KCs and the MBONs whose dendrites reside within the relevant compartments. As a result, learning-induced plasticity within different MB compartments reconfigures the MBON ensemble output signal to promote either learned approach or avoidance behavior. Olfactory aversive learning reduces odor drive of approach-directing MBONs, hence tilting the MBON network toward promoting odor avoidance. By contrast, appetitive learning reduces odor drive of avoidance-directing MBONs, leaving the network in a configuration promoting odor approach (Villar, 2022)

Flies can perceive, learn, and compare differences in the intensity of punishment and adapt their behavior accordingly. This study combined genetic interventions with behavioral analyses, anatomical characterization, and in vivo two-photon calcium imaging to investigate the detailed circuit requirements that allow flies to write and compare olfactory aversive memories of different intensities during learning to promote appropriate value-based choices. This study found that aversive PPL1 DANs show differential responses to electric shock punishment of varying intensity. As a result, the intensity of shock reinforcement correlates with the magnitude of learning-driven plasticity at the corresponding KC to MBON junctions. Using a specific behavioral paradigm in which flies associate three odors with 0, 60, and 30 V punishment, respectively, the circuits involved in coding relative aversive value were identified. Loss-of-function screening revealed a role for specific aversive DANs, in addition to the rewarding PAM-β'2aγ5n DANs, to learn relative aversive value. Recording from PAM-β'2aγ5n DANs during learning revealed these neurons to signal relative aversive value by increasing their responsiveness when the odor-low shock association is better than a previous odor-high shock association. Recording from three MBONs presynaptic to PAM-β'2aγ5n DANs revealed a positive difference in the odor responses of MBON-γ2α'1 between current and previous aversive experiences. This increased responsiveness of cholinergic MBON-γ2α'1 likely provides excitatory input necessary to drive the PAM-β'2aγ5n DANs 'better than' value signal for the less aversive experience, and they thereby learn relative aversive value. In support of this model, optogenetic activation of PAM-β'2aγ5n reward DANs, during learning, assigns a relative 'better than' value to one of two identical odor-punishment associations (Villar, 2022)

This study addresses how animals assign absolute aversive value during learning and how they compare and ascribe relative aversive value information to consecutive negative experiences for them to make appropriate value-based decisions afterwards. Using the fruit fly Drosophila permitted a cellular resolution view of how the interaction between the appetitive and aversive DAN systems, within the MB network, is at the heart of the mechanistic underpinnings that compute a relative aversive value teaching signal. This work also indicates that coding of a relative 'worse than' aversive value likely involves different circuit mechanisms to those for 'better than' but that there may be some overlap (Villar, 2022)

PPL1 DANs reinforce a range of aversive memories with differing strength and persistence. The data provide new insight into the functional diversity of these anatomically discrete DANs. It was found that individual aversively reinforcing PPL1- γ1pedc, PPL1- γ2 α'1, and PPL1- α2 α'2 DANs exhibit different intensity response profiles when flies were exposed to a series of shock voltages. Importantly, the strength of their responses to electric shocks strongly correlated with the magnitude of plasticity of the odor-evoked responsiveness of their corresponding MBONs after differential conditioning. These results indicate that absolute aversive value is assigned to odors in different ways in the γ1pedc, γ2 α'1, and α2 α'2 MB compartments, consistent with the conclusion of a prior study that artificially activated individual DANs (Villar, 2022)

Of note, no significant shock responses was observed in PPL1- α2 α'2 DANs. These results are also in accordance with the absence of odor-evoked changes in the corresponding MBON- α2sc immediately after training and a lack of reinforcing properties when pairing artificial activation of PPL1- α2 α'2 DANs with an odor. In addition, learning-dependent depression of odor responses in MBON- α2sc has been reported to be most relevant for expression of later forms of memory (Villar, 2022)

This study observed that the stronger the aversive experience, the greater the PPL1- γ1pedc DAN-driven depression of the CS+-evoked response of MBON- γ1ped> αβ. Feedforward GABAergic inhibition from MBON- γ1ped> αβ to the primary axon of MBON- γ5β'2a is therefore reduced in a graded manner by aversive conditioning. MBON- γ5β'2a should therefore display a proportional increase in its CS+-evoked response to drive learned avoidance behavior (Deng, 2022).

These experiments uncovered a very different effect of absolute aversive conditioning at the MBON- γ2 α'1 junction. Although the PPL1- γ2 α'1 DANs were significantly triggered by shocks ≥30 V, their responses were comparable at all voltages between 30 and 90 V. Moreover, aversive conditioning did not significantly depress the CS+ responses of MBON- γ2 α'1. Instead, this study observed that the responses to the CS- odor were specifically increased, and the CS- CS+ differential responses were correlated with the intensity of the shocks applied. The data therefore suggest that any odor that follows the CS+ with ≥45 V presentation during training gains the capacity to drive more activity in the cholinergic MBON- γ2 α'1. In addition, recordings indicate that the more aversive the first experience is, the stronger the cholinergic MBON- γ2 α'1 activity will be to the subsequent 'better than' experience. These data reveal a key role for MBON- γ2 α'1 in coding relative aversive value (Deng, 2022).

MBON- γ2 α'1 input to PAM-β'2a γ5n DANs provides a 'better than' reward signal during relative aversive training. Output from PAM-β'2a γ5n DANs was critical during the odor Z + 30 V presentation for relative aversive learning. These DANs receive direct excitatory cholinergic input from MBON- γ2 α'1, and it is proposes that the strength of this excitation is key for the flies to assign a 'better than' reward value to the lesser of the two aversive experiences. As mentioned above, when odor Y is paired with 60 V shock in a differential conditioning assay, the CS- responses of MBON- γ2 α'1 become elevated. This means that when Y + 60 V is followed by Z + 30 V, the Z odor will more strongly drive MBON- γ2 α'1 and as a result will activate the PAM-β'2a γ5n DANs. In effect, it is hypothesized that any odor that follows a Y + 60 V experience is predisposed to be judged as 'better than,' unless it is itself accompanied by 60 V or a greater voltage. These analyses that subtracted MBON- γ2 α'1 odor-evoked responses are entirely consistent with this model. Odor-driven activity of MBON- γ2 α'1 is greater during the first period of the following Z + 30 V experience than during the same period (just after the first shock) of the prior Y + 60 V experience is observed. Critically, this is also the time period during which an elevation of PAM-β'2a γ5n DAN activity. It is speculated that the first of the 30 V shocks somehow further releases the PAM-β'2a γ5n DAN activity to be fully driven by MBON- γ2 α'1, perhaps as a release of feedforward inhibition in the MBON- γ1ped> αβ to MBON- γ5β'2a to PAM-β'2a γ5n DAN pathway. The results and proposed models of PAM-β'2a γ5n DANs providing a 'better than' reward signal are in accordance with previous reports that PAM-β'2a γ5n activation provides appetitive reinforcement (Villar, 2022).

Are there limits to comparable aversive memories? Individual PPL1-DAN subtypes have different thresholds for activation, and intensity-dependent plasticity in their corresponding MBON junctions have similar thresholds. It was noted that these thresholds seem reflected in the range of comparisons that flies can make in a relative choice between different aversive memories, which point toward a threshold and a difference between voltages of 30 V as being optimal to efficiently estimate a relative difference. In the recordings of this study, 30 V was the threshold for observing shock-evoked responses in PPL1- γ1pedc and PPL1- γ2 α'1, but it did not trigger PPL1- α2 α'2. In addition, 30 V produced significant plasticity of MBON- γ1ped> αβ odor responses, but plasticity was not evident in MBON- γ2 α'1 responses until 45 V. Thus, perhaps every odor paired with a voltage of ≥30 V is considered to be 'not so bad,' because it only depresses the GABAergic MBON- γ1ped> αβ responses and not the cholinergic MBON- γ2 α'1 responses, thereby leaving CS+ odor-driven excitation of PAM-β'2a γ5n DANs from these MBONs. Although flies can differentiate between stronger aversive memories such as 90 versus 60 V, their relative choice performances are less good than 60 versus 30 V.8 While significant shock responses were not observed for PPL1- α2 α'2 DANs, this study found a role for these neurons during learning of relative aversive value. MBON- γ1ped> αβ is GABAergic and is connected to PPL1- α2 α'2 DANs (Villar, 2022)

It is therefore possible that repeated pairing of odor Y + 60 V electric shocks (or anything above their threshold) during relative training induces enough CS+-evoked depression at MBON- γ1ped> αβ to release inhibition in PPL1- α2 α'2 DANs while pairing the odor Z with 30 V shocks. The resulting plasticity in MBON- α2sc could explain the requirement of αβ surface and αβ core KCs during a relative choice between Y60 versus Z (Villar, 2022)

The results of this study show that learning a relative 'better than' aversive value requires an interplay between aversively reinforcing PPL1 DANs modulating KC-MBON connections, which provide feedforward and recurrent feedback input that determines the activity of specific subtypes of rewarding PAM DANs. These results support long-held and recent models in both vertebrates and invertebrates, suggesting that learning requires critical interactions between appetitive and aversive reinforcement systems. In the fly, and likely also in mammals, this process relies on opposing populations of DANs providing predictive signals needed to compare current and previous experience to assign (and update) both absolute and relative value to stimuli during learning. For instance, aversive memory extinction and reversal learning require the reward system in both vertebrates and invertebrates (Villar, 2022)

In all these cases, stimuli that represent the absence of a punishment are rewarded. In humans, the ventral striatum, targeted by numerous DA inputs from the ventral tegmental area (VTA) providing rewarding information, is essential to compare aversive experiences of different intensities. In the orbitofrontal cortex, relative coding of aversive (but also appetitive) experiences seem to require overlapping neuronal ensembles to select a preferred option and promote appropriate economical decisions in a specific spatial and temporal context. In the dopaminergic system, reward is also computed in a relative manner to broadcast value signals in different brain regions. These DANs from the VTA and substantia nigra compute a prediction error to signal positive, but also negative, value. A similar value prediction error calculation has not yet been demonstrated experimentally in the fly. Instead, results from several studies in the fly suggest that errors are registered in the MB network by the action of DANs that signal the opposing value. The current experiments suggest that a similar interplay between opposing populations of DANs, and plasticity at different MBON junctions in the MB network, permits computation of relative aversive value (or difference) between a prior and a new aversive experience. Combined with previous work and current computational models, these data provide key features of how the appetitive-aversive system interactions in the MB network using heterogeneous DANs can compare previous and current experience to 'pre-compute' a relative value during learning that facilitates future value-based decisions (Villar, 2022)

Go to top

Representations of novelty and familiarity in a mushroom body compartment

Hattori, D., Aso, Y., Swartz, K. J., Rubin, G. M., Abbott, L. F. and Axel, R. (2017). Cell 169(5): 956-969. PubMed ID: 28502772

Animals exhibit a behavioral response to novel sensory stimuli about which they have no prior knowledge. This study has examined the neural and behavioral correlates of novelty and familiarity in the olfactory system of Drosophila. Novel odors elicit strong activity in output neurons (MBONs) of the α'3 compartment of the mushroom body that is rapidly suppressed upon repeated exposure to the same odor. This transition in neural activity upon familiarization requires odor-evoked activity in the dopaminergic neuron innervating this compartment. Moreover, exposure of a fly to novel odors evokes an alerting response that can also be elicited by optogenetic activation of α'3 MBONs. Silencing these MBONs eliminates the alerting behavior. These data suggest that the α'3 compartment plays a causal role in the behavioral response to novel and familiar stimuli as a consequence of dopamine-mediated plasticity at the Kenyon cell-MBONα'3 synapse (Hattori, 2017).

Novel stimuli can elicit a behavioral response that alerts an organism to unexpected, potentially salient events. An alerting response to sensory stimuli not previously encountered by an animal, such as the orienting response described by Pavlov, provides an organism the opportunity to explore the potential significance of the novel stimulus. A behavioral response to novelty is elicited by sensory cues about which an animal has no prior knowledge. Most behaviors, in contrast, are based upon past experience acquired either over long periods of evolutionary time (innate behaviors) or by learning over the life of an animal. The observation that sensory cues can be identified as novel and can evoke a behavioral response in the absence of prior knowledge poses an interesting problem (Hattori, 2017).

A neural circuit encoding novelty should respond to all novel stimuli, but this response should be suppressed upon familiarization. The memory of a familiar sensory cue should be stimulus specific and long-lasting, distinguishing it from sensory adaptation. Neural responses that correlate with novelty and familiarity are seen in a number of mammalian brain regions. The transition from novelty to familiarity is associated with suppression of neural responses in higher brain centers that appears distinct from intrinsic or sensory adaptation. Electrophysiologic recordings along the visual and auditory pathways reveal that neurons exhibit activity in response to novel or unexpected cues that diminish upon repeated exposure. In the auditory pathway, neurons in the inferior colliculus and the auditory cortex exhibit responses to novel or unexpected tones that attenuate upon repetition. Similarly, neurons in the perirhinal and inferior temporal cortices respond to novel visual stimuli, and this response attenuates rapidly upon repetition, a phenomenon known as repetition suppression (Hattori, 2017).

Dopaminergic neurons in the substantia nigra (SN) pars compacta and ventral tegmental area (VTA) also exhibit phasic bursting activity in response to novel or unexpected sensory events. Unexpected flashes of light or auditory tones evoke burst firing in 60%-70% of the dopaminergic neurons that attenuates as the novel stimulus becomes familiar. Related neural events may underlie attenuation in the BOLD signal observed in extra striate cortex as well as SN and VTA in fMRI studies of humans upon repeated exposure to sensory stimuli. Thus, mammals have evolved neural systems that distinguish novel from familiar sensory stimuli that may facilitate the determination of the potential salience of unfamiliar environmental events (Hattori, 2017).

This study has analyzed the behavioral and neural correlates of novelty and familiarity in the olfactory system of Drosophila. Olfactory perception in the fly is initiated by the binding of an odor to an ensemble of olfactory sensory neurons in the antennae that results in the activation of a unique and topographically fixed combination of glomeruli in the antennal lobe. Each antennal lobe projection neuron extends dendrites into one of the 54 glomeruli and extends axons that bifurcate to innervate two distinct brain regions, the lateral horn, and the mushroom body (MB). The invariant circuitry of the lateral horn is thought to mediate innate behaviors, whereas the unstructured projections to the MB translate olfactory sensory information into learned behavioral responses. In the MB, each odor activates a sparse representation (5%-10%) of principal neurons, the Kenyon cells (KCs). KCs extend axons that form en passant synapses in the compartments of the MB lobes. The KCs synapse on the MB output neurons (MBONs), which have distinct spatially stereotyped dendritic arbors within compartments that collectively tile the lobes. MBONs provide the only output of the MB and the activity of different MBON combinations biases behavior. Each of the 15 compartments is also innervated by the axons of one to three of 20 dopaminergic cell types (dopaminergic neurons, or DANs). Distinct DANs respond to different unconditioned stimuli and dopamine release elicits plasticity in the synapses between the KCs and MBONs. The alignment of DAN arbors with compartmentalized KC-MBON synapses creates a unit for learning that transforms the disordered KC representation into ordered MBON output to collectively bias behavioral responses to sensory stimuli (Hattori, 2017).

The transition from novelty to familiarity involves memory formation, and learning and memory in the fly are accomplished by the circuitry of the MB. This study has identified a neural circuit in the MB that appears to encode a representation of novelty and familiarity. MBONs were observed innervating the α'3 compartment respond to novel odors and that their activity is rapidly suppressed upon repeated exposure to the same stimulus. This suppression upon familiarization is observed for all novel odors tested regardless of innate valence, is stimulus specific, lasts for more than 20 min, and recovers in 1 hr. Repetition suppression of MBON-α'3 is distinct from sensory adaptation and requires odor-evoked activity in the DAN innervating the α'3 compartment. These data suggest that repeated exposure to an odor mediates dopamine-dependent plasticity at the KC synapses onto MBON-α'3 that suppresses MBON output. Moreover, behavioral experiments demonstrate that the α'3 MBONs mediate an alerting response to novel odors. Exposure of a fly to novel olfactory stimuli evokes an alerting behavior. This behavioral response can be elicited by optogenetic activation of α'3 MBONs and eliminated by α'3 MBON silencing. These observations suggest that the behavioral response to novelty and the transition to familiarity is mediated by the circuitry of KCs, DANs, and MBONs within the α'3 compartment (Hattori, 2017).

This study has examined the neural and behavioral correlates of novelty and familiarity in the olfactory system of Drosophila. Exposure of flies to a novel odor interrupts grooming. This alerting response is dependent upon the activity of output neurons of the α'3 compartment of the MB. Optogenetic activation of the α'3 MBONs elicits an alerting response, whereas silencing these neurons eliminates the behavioral response to novel odors. A neural correlate of this behavioral response is observed in the activity of MBON-α'3. Novel odors elicit strong activity in these neurons that is rapidly suppressed upon repeated exposure to the same odor. This transition in MBON-α'3 response upon familiarization requires the activity of PPL1-α'3, the DAN innervating the α'3 compartment, and dopamine receptors in the MBONs. These data suggest that the α'3 compartment may play a causal role in the behavioral response to novel and familiar stimuli as a consequence of dopamine-mediated plasticity at the KC-MBONα'3 synapse. Although the circuitry of the α'3 compartment is central to the behavioral response to novel and familiar odors, the data do not exclude a contribution from other compartments (Hattori, 2017).

Plasticity at the KC-MBON synapses has been invoked to explain olfactory learning and memory (Heisenberg, 2003). In associative learning, exposure to a conditioned stimulus (CS), when paired with an unconditioned stimulus (US), imposes an associative memory upon the CS. The identity of the CS is represented by activity in a specific ensemble of KCs, whereas USs of different valence activate distinct DANs. Dopamine input depresses the KC-MBON synapse in specific compartments to transform the unstructured KC representation of an odor into an ordered MBON representation encoding behavioral bias. The neural mechanism governing the novelty response differs from this classical model of associative learning. In the α'3 compartment, a novel odor elicits strong MBON output and also activates the DAN. Dopamine release from PPL1-α'3 depresses the KC-MBONα'3 synapse, suppressing MBON output on further exposure to this odor. In this manner, a novel odor effectively serves as both a CS and a US to drive the transition from novelty to familiarity (Hattori, 2017).

A second distinction between the novelty response and associative learning emerges from the observation that the response to novelty is suppressed by learning whereas the conditioned response depends on learning. The stereotyped alerting behavior in response to novel odor does not require learning and is therefore innate. In associative learning models, exposure to odor prior to learning activates all MBONs examined, but this combinatorial of MBONs does not elicit a behavior. Rather, behavioral bias is imposed by the suppression of specific MBON output after learning. In the novelty response, innate alerting behavior is elicited by strong output from MBON-α'3 in response to novel odors prior to learning. Learning that accompanies the transition to familiarity suppresses MBON-α'3 activity and the behavioral response to novelty. If, however, the novel odor is accompanied by salient events in the environment, this will result in activation of additional DANs, leading to the formation of associations in other compartments. The alerting response evoked by MBON-α'3 may enhance the awareness of these environmental events. Thus, α'3 output may elicit an immediate and stereotyped response to an odor independent of its salience, which is then assessed by the remaining MB compartments to mediate more measured associative responses (Hattori, 2017).

Odor-evoked dopamine release by PPL1-α'3 appears to be essential to modulate MBON output in the transition from novelty to familiarity. Dopamine release also contributes to decay in the memory of familiar odors. MBON-α'3 activity in response to novel odors is suppressed upon repeated exposure, but activity is restored after 1 hr. Dopamine release in the absence of odor, following repetition suppression, accelerates this recovery process. These observations are consistent with recent experiments demonstrating that dopamine release within a compartment in the absence of odor can lead to synaptic facilitation and the restoration of MBON output. In this manner, familiarity in the fly is a transient phenomenon and the restoration of the perception of novelty may be accelerated by dopamine (Hattori, 2017).

It is suggested that the neural events responsible for the transition from novelty to familiarity involve depression of only those KC-MBON synapses activated by the novel odor. In this manner, novelty and familiarity can be both universal and odor specific. A given odor activates about 5%-10% of the KCs in the MB and by inference 5%-10% of the KC-MBON synapses. An organism is likely to encounter multiple odors that may be construed as novel in the course of hours. If a single novel odor suppresses 5%-10% of the KC-MBON synapses and this synaptic depression is long term, all KC-MBON synapses within the α'3 compartment would be depressed after exposure to roughly 20 novel odors. Depression of all the synapses would prevent subsequent response to novel odors. The data suggest this problem may be obviated in two ways. First, the depression is relatively short lived. Second, although novel odors result in the depression of active KC-MBON synapses, they also enhance the recovery of previously suppressed but currently inactive synapse. This implies that the rate of synaptic recovery is proportional to the rate of exposure to novel odors. Thus, the α'3 compartment has evolved a mechanism to assure that novelty responses can be generated in both dense and sparse odor environments without saturation (Hattori, 2017).

The MB is an associative center in invertebrate brains thought to impose valence on sensory representations. The current data suggest that the MB not only functions in classical learning paradigms, but also supports novelty detection and the transition to familiarity. An organism can have no knowledge of a novel stimulus, and hence it exhibits an indiscriminate alerting response. The MB also integrates information about the organism's internal state (hunger, satiety, sleep, wakefulness, roaming, dwelling) allowing the fly to more comprehensively contextualize the diverse sensory experiences it may encounter throughout its life. Thus, the MB may afford the fly an 'individuality' allowing different flies to respond differently to the same stimuli in accord with its unique history and current state (Hattori, 2017).

Go to top

Coordinated and compartmentalized neuromodulation shapes sensory processing in Drosophila

Cohn, R., Morantte, I. and Ruta, V. (2015). Cell 163(7): 1742-1755. PubMed ID: 26687359

Learned and adaptive behaviors rely on neural circuits that flexibly couple the same sensory input to alternative output pathways. This study shows that the Drosophila mushroom body functions like a switchboard in which neuromodulation reroutes the same odor signal to different behavioral circuits, depending on the state and experience of the fly (see Compartmentalized Architecture of the Mushroom Body). Using functional synaptic imaging and electrophysiology, it was shown that dopaminergic inputs to the mushroom body modulate synaptic transmission with exquisite spatial specificity, allowing individual neurons to differentially convey olfactory signals to each of their postsynaptic targets. Moreover, the dopaminergic neurons function as an interconnected network, encoding information about both an animal's external context and internal state to coordinate synaptic plasticity throughout the mushroom body. These data suggest a general circuit mechanism for behavioral flexibility in which neuromodulatory networks act with synaptic precision to transform a single sensory input into different patterns of output activity (Cohn, 2015).

This study took advantage of the mushroom body's orderly architecture to gain insight into the circuit mechanisms through which neuromodulation mediates flexible sensory processing. Compartmentalized dopaminergic signaling permits independent tuning of synaptic transmission between an individual KC and its repertoire of postsynaptic MBON targets. As a consequence, the same KC odor representation can evoke different patterns of output activity, depending on the state of the animal and the dopaminergic network. Recent data indicate that the ensemble of MBONs acts in concert to bias an animal's behavioral response to an odor such that altering the balance of their activity can modify the olfactory preferences of both naive and trained animals. In accord with such a model, this study revealed how a distributed neuromodulatory network is poised to orchestrate plasticity across all 15 compartments of the mushroom body and reweight the net output of the MBONs, allowing for adaptive behavioral responses based on the immediate needs or past experience of the animal (Cohn, 2015).

Distinct subsets of DANs are sufficient to drive learned olfactory associations, leading to the suggestion they may act autonomously to encode the rewarding or punishing contextual stimuli that assign meaning to an odor. The current data, however, suggest a more complex circuit architecture, in which rich functional interconnectivity between compartments contributes to coordinated and bidirectional patterns of activity across the DAN population. This raises the possibility that reinforcement experiences may be represented by combinatorial patterns of DAN excitation and inhibition in different compartments, endowing the dopaminergic population with a greater capacity to instruct behavior via the limited repertoire of mushroom body outputs. Intriguingly, midbrain dopaminergic neurons responsive to punishment and reward also project to distinct targets in the mammalian brain and display a similar functional opponency as a consequence of reciprocal network interactions. Thus, the concerted and partially antagonistic action of neuromodulatory pathways may represent a general and conserved circuit principle for generating adaptive behavioral responses (Cohn, 2015).

Distinct DAN network activity states are evoked by electric shock and sugar ingestion, reinforcers classically used in associative olfactory conditioning paradigms because of their strong inherent valence. However, similarly distributed patterns of DAN activity are correlated with the fly's motor activity, implying that an animal's behavioral state might serve as a reinforcement stimulus that itself drives synaptic plasticity to shape odor processing. Metabolic states, such as thirst and hunger, have been shown to gate appetitive reinforcement by water and sugar rewards, permitting state-dependent formation of olfactory associations only in motivated animals. The current data highlight an additional facet of how an animal's internal state can regulate dopamine release to adjust the salience of contextual cues. Together, these observations indicate that the distributed DAN network integrates information about external context and internal state with MBON feedback to represent the moment-by-moment experience of an animal and dynamically regulate the flow of olfactory signals through the mushroom body (Cohn, 2015).

The independent regulation of synapses along an axon is thought to permit a single neuron to convey specialized information to different downstream targets, providing additional flexibility and computational power to neural circuits. In the mushroom body, synapse-specific plasticity is achieved through spatially restricted patterns of dopaminergic modulation that divide a KC axon into functionally distinct segments. Thus, the ensemble of synapses within a compartment, as the site of convergence for sensory and contextual signals, represents the elementary functional unit that underlies experience-dependent mushroom body output (Cohn, 2015).

Within a compartment, multiple neuromodulatory mechanisms appear to shape synaptic signaling. Broad potentiation of KC-MBON synapses is seen after DAN activation, but odor-specific depression is seen if DANs were coincidently activated with KCs, consistent with the synaptic changes previously proposed to occur after learning. Taken together, these findings indicate that neuromodulation in the mushroom body instructs opposing forms of synaptic plasticity, analogous to the bidirectional tuning of synaptic strength by dopamine in mammalian brain centers. The molecular mechanisms through which dopamine can direct diverse synaptic changes within a compartment remain to be elucidated, but they may depend on signaling through different dopamine receptors or downstream signaling cascades that function as coincidence detectors. Indeed, while DopR1 in KCs is essential to the formation of learned olfactory associations, this receptor was found to play only a subtle role in the context-dependent patterning of Ca2+ along their axons. Conversely, DopR2 strongly influences the topography of presynaptic Ca2+ along KC axons, in accord with evidence that tonic release of dopamine during ongoing behavior acts through this receptor to interfere with the maintenance of specific learned olfactory associations. Thus, distinct molecular pathways may transform the same dopaminergic reinforcement signals into synaptic changes of opposite polarity to shape olfactory processing based on both the present context and prior experiences of an individual (Cohn, 2015).

The mushroom body has been most extensively studied as a site for associative learning in which the temporal pairing of an odor with a reinforcement experience selectively alters subsequent behavioral responses to that odor. The current data suggest that the convergence of DAN network activity and KC olfactory representations within the mushroom body lobes may drive associative plasticity in each compartment, allowing the odor tuning of the MBON repertoire to reflect the unique experiences of an individual. However, these observations also provide insight into the mushroom body's broader role in the context-dependent regulation of innate behaviors. The ongoing activity of the distributed DAN network, encoding information about an animal's current environmental context and behavioral state, is poised to continuously reconfigure the activity patterns of the MBON population to allow for adaptive responses based on the acute needs of the animal. This context-dependent synaptic modulation could potentially erode odor-specific learned associations within the mushroom body, permitting the immediate circumstances of an animal to dominate over previously learned olfactory associations that may no longer be predictive or relevant. The axons of MBONs ultimately converge with output pathways from the lateral horn, a Drosophila brain center thought to mediate stereotyped responses to odors, providing a potential substrate for learned and context-dependent output from the mushroom body to influence inherent olfactory preferences (Cohn, 2015).

Thus, the dual role of neuromodulation in the mushroom body-to select among alternative circuit states that regulate both innate and learned behaviors-is reminiscent of its function in other higher integrative brain centers. In the basal ganglia, for example, different temporal patterns of dopamine release are thought to select the relevant circuit configurations that control inherently motivated behaviors as well as reinforcement learning. The generation of flexible behavioral responses based on experience, whether past or present, may therefore rely on common integrative brain structures in which neuromodulatory networks act with exquisite spatial precision to shape sensory processing (Cohn, 2015).

Go to top

Trace conditioning in Drosophila induces associative plasticity in mushroom body kenyon cells and dopaminergic neurons

Dylla, K. V., Raiser, G., Galizia, C. G. and Szyszka, P. (2017). Front Neural Circuits 11: 42. PubMed ID: 28676744

Dopaminergic neurons (DANs) signal punishment and reward during associative learning. In mammals, DANs show associative plasticity that correlates with the discrepancy between predicted and actual reinforcement (prediction error) during classical conditioning. Also in insects, such as Drosophila, DANs show associative plasticity that is, however, less understood. This study examined ssociative plasticity in DANs and their synaptic partners, the Kenyon cells (KCs) in the mushroom bodies (MBs), while training Drosophila to associate an odorant with a temporally separated electric shock (trace conditioning). In most MB compartments DANs strengthened their responses to the conditioned odorant relative to untrained animals. This response plasticity preserved the initial degree of similarity between the odorant- and the shock-induced spatial response patterns, which decreased in untrained animals. Contrary to DANs, KCs (α'/β'-type) decreased their responses to the conditioned odorant relative to untrained animals. No evidence was found for prediction error coding by DANs during conditioning. Rather, the data supports the hypothesis that DAN plasticity encodes conditioning-induced changes in the odorant's predictive power (Dylla, 2017).

Associative learning enables animals to anticipate negative or positive events. The neural mechanisms of associative learning are commonly studied in classical conditioning paradigms, in which animals are trained to associate a cue (conditioned stimulus; CS) with a punishment or reward (unconditioned stimulus; US). In the standard conditioning paradigm CS and US overlap in time, while in the trace conditioning paradigm there is a temporal gap between the CS and US. During both standard conditioning and trace conditioning, the US is mediated by dopaminergic neurons (DANs), in animals as diverse as monkeys and fruit flies (Dylla, 2017).

Genetic tools for monitoring and manipulating neuronal activity in the fruit fly Drosophila melanogaster promoted the understanding of the neural mechanisms of dopamine-mediated learning. Those mechanisms are well-described for standard 'odor-shock conditioning' in Drosophila, in which an olfactory CS is paired with a temporally overlapping electric shock US. During conditioning, an odor-shock association is formed in the mushroom body (MB) neuropil. The intrinsic neurons of the MB, the Kenyon cells (KCs), receive olfactory input in the MB-calyx and project to the vertical (α and α'), and the medial (β, β', and γ) MB-lobes. During odor-shock conditioning, the olfactory CS activates an odorant-specific KC population, and the electric shock US activates DANs that innervate the MB-lobes. In KCs, the CS-induced increase in intracellular calcium and the US-(dopamine)-induced second messengers synergistically activate an adenylyl cyclase, which alters the synaptic strength between KCs and MB output neurons (MBONs). This change in KC-to-MBON synapses is thought to encode the associative odor memory (Dylla, 2017).

The MB-lobes are divided into 15 compartments (α1-3, β1-2, α'1-3, β'1-2, and γ1-5), each of which is innervated by a distinct population of DANs and MBONs. These compartments constitute functional units, which are involved in different forms of associative learning. In compartments such as γ1, γ2, and β2, DANs mediate electric shock reinforcement. Besides mediating reinforcement during classical conditioning, Drosophila DANs are involved in long-term memory formation, forgetting, extinction learning and memory reconsolidation, and in integrating internal states with memory and sensory processing. A single DAN can even serve different functions, for example, PPL1-γ1pedc (also referred to as MB-MP1) signals reinforcement, and controls state-dependent memory retrieval (Dylla, 2017).

The functional complexity of Drosophila DANs is further increased by the fact that DANs show learning-induced associative plasticity: they increase their response to the CS during classical conditioning. Mammalian DANs also increase their CS-induced responses during classical conditioning. In addition, they decrease their response to the US, and when a predicted US does not occur, they decrease their activity below baseline level. This pattern of response plasticity in mammalian DANs is compatible with the hypothesis that animals only learn to associate a CS with a US, when the US occurs unpredictably. Thus, mammalian DANs appear to encode this prediction error. In Drosophila, however, DANs do not change their response to the US. Therefore, Drosophila DANs appear to encode the US prediction by the CS rather than encoding the US prediction error during classical conditioning. It is not clear, whether classical conditioning in insects is driven by US prediction error. There is evidence for prediction error-driven conditioning in crickets, but there is also a controversy about whether or not blocking (a failure to learn, when the US is already predicted by another CS) occurs (Dylla, 2017).

This study reassessed the hypothesis that Drosophila DANs do not encode the prediction error during classical conditioning (Riemensperger, 2005). Different from Riemensperger (2005) who pooled DAN activity across the mushroom body lobes, this study differentiated between DAN types that innervate different compartments of the MB lobes. Moreover, instead of using standard conditioning, trace conditioning with a 5 s gap between the CS and the US was used, allowing for precise distinguishing between responses to either the CS or the US (Dylla, 2017).

This study investigated associative plasticity in the responses of DANs and their synaptic partners, the KCs, across the compartments of the Drosophila MB. Using calcium imaging, CS- and US-induced responses of a subpopulation of DANs (labeled by TH-GAL4) and of KCs (labeled by OK107-GAL4) were recorded during odor-shock trace conditioning. Note, that most compartments are innervated by multiple TH-GAL4-labeled DANs. Therefore, the average activity that was recorded in most of the compartments might mask possible differences in the response properties and plasticity between individual DANs and KCs. Only DAN responses in the compartments γ2 and α'1 reflect the responses of a single neuron (Dylla, 2017).

Across MB compartments, DANs and KCs differed in their response strength to odorants and electric shock, and they differed in CS-US pairing-induced plasticity. Compared to the unpaired control groups, KCs decreased their responses to the CS in all compartments of the β'-lobe and in the junction, while DANs increased their responses to the CS in all compartments of the γ- and β'-lobe, and in the junction. The occurrence of associative plasticity in DANs in the compartments γ3-5 and β'1 is surprising, given that these DANs are not known to be involved in odor-shock conditioning, after training there was neither an associative change in US-induced DAN responses nor a change of activity during US-omission after CS presentation. It is therefore concluded that Drosophila DANs do not encode the US-prediction error during classical conditioning (Dylla, 2017).

Previous studies suggested that DANs in the MB lobes respond strongly to electric shock and weakly to odorants. The compartment-resolved analysis of the calcium imaging data refines this picture: It is confirmed that DANs of all imaged compartments respond to both electric shock and odorants, and it was shown that their relative response strength to odorants and electric shock differs across compartments. For example, DANs innervating γ1 responded stronger to electric shock than to odorants, while DANs innervating β'2 responded equally strong to odorants and electric shock. The strongest DAN responses to electric shock were shown to be in the compartments γ1 and γ2. These compartments receive input from PPL1-γ1pedc and PPL1-γ2α'1 DANs that mediate electric shock reinforcement. In all compartments, except in α1/α'1, the DAN response strength correlated positively with the current strength encountered by individual flies. Thus, DANs are capable of encoding the strength of the electric shock US, and this property may account for the positive dependence between electric shock strength and learning performance in flies (Dylla, 2017).

Calcium responses in KCs differ between MB lobes, and they differ between the compartments of a given lobe, possibly due to compartment-specific modulation by DANs and MBONs. KCs in γ2 and γ3 responded strongest to odorants, confirming the results of Cohn (2015). KCs generally responded only weakly to electric shocks. Previously published strong KC responses to electric shock may be because electric shocks were applied to the flies' abdomen rather than to their legs, which might have resulted in a stronger stimulation (Dylla, 2017).

The associative strengthening of DAN responses to the olfactory CS (as compared to the unpaired control group), confirms the previous report by Riemensperger (2005). Associative plasticity occurred in those DANs that innervate the MBs (PPL1 and PAM cluster DANs; note that the used TH-GAL4 driver line covers only a small subpopulation of PAM neurons but not in DANs that innervate the central complex (PPL1 and PPM3 cluster DANs). This is in line with the established role of MB innervating-DANs in associative memory formation, while central complex-innervating DANs are involved in behaviors such as locomotion, wakefulness, arousal, and aggression, and are therefore not expected to show odor-shock conditioning-induced plasticity (Dylla, 2017).

In contrast to previous studies, this study did not find an associative increase in KC calcium responses to the CS in the MB-lobes after odor-shock conditioning. This may indicate either a difference between trace conditioning and standard conditioning, or a difference in other experimental parameters that may also account for inconsistencies in the published effects of odor-shock conditioning (Dylla, 2017).

The associative decrease in KC responses in the β'-lobe compartments is in line with previous studies that showed conditioning-induced depression of KC-to-MBON synapses (Cohn, 2015; Hige, 2015). Therefore, it propose that the associative decrease in KC responses to the CS reflects a presynaptic depression at KC-to-MBON synapses in β'-lobe compartments (Dylla, 2017).

What is the site of neuronal plasticity that underlies the relative increase in DANs' responses to the olfactory CS? Riemensperger (2005) proposed that DANs get odorant-driven excitatory input via a MBON feedback loop that is strengthened during odor-shock conditioning. However, the DAN population is composed of different neuron types that do not share a common input either from MBONs or from other neurons that could explain the global associative plasticity across MB compartments. Because KCs presumably provide the only common odor-driven input to all MB-innervating DANs, it is suggested that the site of associative plasticity is located in a KC-to-DAN synapse. Indeed, KC-to-DAN synapses have recently been reported in Drosophila (Cervantes-Sandoval, 2017). Associative increase in CS-induced DAN responses occurred despite unaltered or decreased KC responses in the same compartment. This suggests that the associative plasticity occurs post-synaptic in DANs and is not inherited from KCs (Dylla, 2017).

What is the neuronal substrate of CS-US coincidence detection in DANs and KCs? Drosophila trace conditioning depends on dopamine receptor-triggered signaling in KCs, as is the case for standard conditioning. However, the CS-US coincidence detection mechanism in trace conditioning is unknown. In standard conditioning the CS-induced increase in KCs' calcium concentration coincides with the US-(dopamine)-induced second messengers, which is thought to synergistically activate the rutabaga adenylyl cyclase, and ultimately alters the strength of KC-to-MBON synapses. This mechanism would not work for trace conditioning, because (1) at the time the US occurs, CS-induced increase in KCs' calcium concentration is back to baseline levels, and (2) trace conditioning does not involve the rutabaga adenylyl cyclase. It is therefore hypothesized that a non-rutabaga adenylyl cyclase or a protein kinase C could serve as a molecular coincidence detector for the CS trace and the US. For example, the CS-induced calcium and dopamine signaling could lead to a sustained activation of an adenylyl cyclase or protein kinase C in KCs, which then would increase synergistically and drive synaptic plasticity during the US-induced dopamine signaling (Dylla, 2017).

DAN responses to odorants and associative strengthening of DAN responses to the CS-odorant are not included in current models of associative learning in the MB. However, associative plasticity is a common feature of US-mediating neurons, which occurs in mammalian and Drosophila DANs, and in an octopaminergic neuron in honey bees (Dylla, 2017).

What could be the function of odorant-induced responses and odor-shock conditioning-induced plasticity in DANs? MB-innervating DANs strengthened their response to the CS (as compared to the unpaired group) during odor-shock conditioning, in line with Riemensperger (2005). However, other than in monkey DANs, this study did not observe associative plasticity in DANs' response to the US. The data therefore support the idea that Drosophila DANs encode predictive power of the CS, e.g., US-prediction, but not the US-prediction error during classical conditioning (Dylla, 2017).

This study found shock-induced responses and associative plasticity in DANs that are not involved in odor-shock conditioning, for example in DANs innervating β'1, γ3, γ4, and γ5. This suggests that those DANs serve a function in aversive odor learning which is not captured by the commonly applied conditioning paradigms. For example, the relative strengthening of CS-induced responses could mediate reinforcement during second-order conditioning, in which a previously reinforced CS1 can act as US in subsequent conditioning of a second CS2. As Drosophila is capable of second-order learning, this theory can be tested in behavioral experiments: if associative strengthening of DAN responses to the CS underlies CS1-induced reinforcement in second-order conditioning, then preventing associative plasticity in DANs, or blocking their output during CS2-CS1 pairing should abolish second-order conditioning (Dylla, 2017).

The occurrence of CS-induced responses and associative plasticity in most of the MB-innervating DANs suggests that the separation between the CS- and US-pathway and between different US-pathways is less strict than suggested in current models of associative learning in the MB. Associative plasticity in the spatial pattern of CS-induced DAN responses makes them a potential neuronal substrate for encoding the US identity in CS-US memories and the predictive power of a CS (Dylla, 2017).

These data revealed similar response properties and plasticity rules across Drosophila DANs in the γ- and β'-lobe. This contrasts with their anatomical and functional heterogeneity, which indicates yet undiscovered mechanisms and functions of DAN plasticity. Note, that this study could not test whether the flies learned in the imaging setup, as currently no behavioral readout exists for odor-shock conditioning during physiological experiments. Nevertheless, since a conditioning protocol and stimulus application comparable to an established behavioral paradigm was used, it is believed that the associative plasticity in neuronal responses that was found underlies behavioral associative plasticity. Therewith the data lay the foundations for causal studies on the function of associative plasticity in DANs (Dylla, 2017).

Go to top

Suppression of Dopamine Neurons Mediates Reward

Yamagata, N., Hiroi, M., Kondo, S., Abe, A. and Tanimoto, H. (2016). PLoS Biol 14(12): e1002586. PubMed ID: 27997541

Massive activation of dopamine neurons is critical for natural reward and drug abuse. In contrast, the significance of their spontaneous activity remains elusive. In Drosophila melanogaster, depolarization of the protocerebral anterior medial (PAM) cluster dopamine neurons en masse signals reward to the mushroom body (MB) and drives appetitive memory. Focusing on the functional heterogeneity of PAM cluster neurons, a single class of PAM neurons, PAM-γ3, mediates sugar reward by suppressing their own activity. PAM-γ3 is selectively required for appetitive olfactory learning, while activation of these neurons in turn induces aversive memory. Ongoing activity of PAM-γ3 gets suppressed upon sugar ingestion. Strikingly, transient inactivation of basal PAM-γ3 activity can substitute for reward and induces appetitive memory. Furthermore, the satiety-signaling neuropeptide Allatostatin A (AstA) was identified as a key mediator that conveys inhibitory input onto PAM-γ3. These results suggest the significance of basal dopamine release in reward signaling and reveal a circuit mechanism for negative regulation (Yamagata, 2016).

Sugar ingestion triggers multiple reward signals in the fly brain. This study has provided lines of evidence that part of the reward is signaled by inactivating dopamine neurons. The role of PAM-γ3 highlights the striking functional heterogeneity of PAM cluster dopamine neurons. The decrease and increase of dopamine can convey reward to the adjacent compartments of the same MB lobe-γ3 and γ4-. The reward signal by the transient decrease of dopamine is in stark contrast to the widely acknowledged role of dopamine. Midbrain dopamine neurons in mammals were shown to be suppressed upon the presentation of aversive stimuli or the omission of an expected reward, implying valence coding by the bidirectional activity. As depolarization of PAM-γ3 can signal aversive reinforcement, these neurons convey the opposite modulatory signals to the specific MB domain by the sign of their activity. Intriguingly, the presentation and cessation of electric shock act as punishment and reward, respectively. Such bidirectional activity of PAM-γ3 may represent the presentation and omission of reward. (Yamagata, 2016).

While thermoactivation of PAM-γ3 induced robust aversive memory, blocking their synaptic transmission did not affect shock learning, leaving a question regarding their role in endogenous aversive memory process. PAM-γ3 may only be involved in processing aversive reinforcement different from electric shock-like heat. However, two studies show that dopamine neurons mediating aversive reinforcement of high temperature and bitter N,N-Diethyl-3-methylbenzamide (DEET) are part of those for electric shock. Identification of such aversive stimuli that are signaled by PAM-γ3 activation is certainly interesting, as it is perceived as the opposite of sugar reward and thus provides the whole picture of the valence spectrum. Another scenario where sufficiency and necessity do not match is the compensation of the reinforcing effect by other dopamine cell types (e.g. MB-M3). The lack of PAM-γ3 requirements for electric shock memory may be explained by a similar mechanism. (Yamagata, 2016).

How can the suppression of PAM-γ3 modulate the downstream cell and drive appetitive memory? Optogenetic activation of the MB output neurons from the γ3 compartment induces approach behavior. This suggests that the suppression of the PAM-γ3 neurons upon reward leads to local potentiation of Kenyon cell output. This model is supported by recent studies showing the depression of MB output synapses during associative learning. A likely molecular mechanism is the de-repression of inhibitory D2-like dopamine receptors, DD2R. As D2R signaling is a widely conserved mechanism, it may be one of the most ancestral modes of neuromodulation. (Yamagata, 2016).

Furthermore, recent anatomical and physiological studies demonstrated that different MB-projecting dopamine neurons are connected to each other and act in coordination to respond to sugar or shock. Therefore, memories induced by activation or inhibition of PAM-γ3 may well involve the activity of other dopamine cell types (Yamagata, 2016).

The finding that appetitive reinforcement is encoded by both activation and suppression of dopamine neurons raises the question as to the complexity of reward processing circuits (see Reward signals by excitation and inhibition of dopamine neurons). It is, however, reasonable to implement a component like PAM-γ3 as a target of the satiety-signaling inhibitory neuropeptide AstA. Intriguingly, the visualization of AstA receptor distribution by DAR-1-GAL4 revealed expression in two types of MB-projecting dopamine neurons: PAM-γ3 and MB-MV1 (also named as PPL1- γ2α'1). Given the roles of MB-MV1 in aversive reinforcement and locomotion arrest, AstA/DAR-1 signaling may also inhibit a punishment pathway upon feeding. It is thus speculated that this complex dopamine reward circuit may be configured to make use of bidirectional appetitive signals in the brain (Yamagata, 2016).

Go to top

Mushroom body signaling is required for locomotor activity rhythms in Drosophila

Mabuchi, I., Shimada, N., Sato, S., Ienaga, K., Inami, S. and Sakai, T. (2016). Neurosci Res [Epub ahead of print]. PubMed ID: 27106579

In the fruitfly Drosophila melanogaster, circadian rhythms of locomotor activity under constant darkness are controlled by pacemaker neurons. To understand how behavioral rhythmicity is generated by the nervous system, it is essential to identify the output circuits from the pacemaker neurons. The importance of mushroom bodies (MBs) in generating behavioral rhythmicity remains controversial because contradicting results have been reported as follows: (1) locomotor activity in MB-ablated flies is substantially rhythmic, but (2) activation of restricted neuronal populations including MB neurons induces arrhythmic locomotor activity. This study reports that neurotransmission in MBs is required for behavioral rhythmicity. For adult-specific disruption of neurotransmission in MBs, the GAL80/GAL4/UAS ternary gene expression system was used in combination with the temperature-sensitive dynamin mutation shibirets1. Blocking of neurotransmission in GAL4-positive neurons including MB neurons induced arrhythmic locomotor activity, whereas this arrhythmicity was rescued by the MB-specific expression of GAL80. These results indicate that MB signaling plays a key role in locomotor activity rhythms in Drosophila (Mabuchi, 2016).

Go to top

The full-length form of the Drosophila amyloid precursor protein is involved in memory formation

Bourdet, I., Preat, T. and Goguel, V. (2015). J Neurosci 35: 1043-1051. PubMed ID: 25609621

The APP plays a central role in AD, a pathology that first manifests as a memory decline. Understanding the role of APP in normal cognition is fundamental in understanding the progression of AD, and mammalian studies have pointed to a role of secreted APPα in memory. In Drosophila, APPL, the fly APP ortholog, is required for associative memory. This study aimed to characterize which form of APPL is involved in this process. Expression of a secreted-APPL form in the mushroom bodies, the center for olfactory memory, was able to rescue the memory deficit caused by APPL partial loss of function. The study next assessed the impact on memory of the Drosophila α-secretase kuzbanian (KUZ), the enzyme initiating the nonamyloidogenic pathway that produces secreted APPLα. Strikingly, KUZ overexpression not only failed to rescue the memory deficit caused by APPL loss of function, it exacerbated this deficit. Further, in addition to an increase in secreted-APPL forms, KUZ overexpression caused a decrease of membrane-bound full-length species that could explain the memory deficit. Indeed, transient expression of a constitutive membrane-bound mutant APPL form was sufficient to rescue the memory deficit caused by APPL reduction, revealing for the first time a role of full-length APPL in memory formation. This data demonstrates that, in addition to secreted APPL, the noncleaved form is involved in memory, raising the possibility that secreted and full-length APPL act together in memory processes (Bourdet, 2015).

The majority of studies into APP biology have focused on pathogenic mechanisms. However, it remains crucial to understand the normal physiological function of APP, especially as it is possible that APP loss of function elicits early cognitive impairment in AD patients. This study shows that overexpression of secreted APPL rescues the short-term memory deficit caused by a reduction of APPL level. In sharp contrast, overexpression of the α-secretase, KUZ, which produces sAPPL, exacerbates the memory impairment, a phenotype that is likely due to a deficit in full-length APPL protein level. Supporting this hypothesis, it was further demonstrated that expression of a nonprocessed APPL mutant form is able to restore wild-type memory in an APPL partial loss of function background (Bourdet, 2015).

In the past, two main strategies have been considered as therapeutic approaches for AD. First, inhibition of the β- or γ-secretase has been used to achieve an inhibition of Aβ toxic production. However, reduction of Aβ production is not only an ineffective approach for AD, it also can actually promote further pathology, as these enzymes have numerous substrates. A second proposed approach has been to inhibit the amyloidogenic pathway by activating the α-processing of APP. In addition to the potential beneficial inhibition of the amyloidogenic pathway, the advantage of this type of approach is to also increase the production of sAPPα. Indeed, decreased CSF sAPPα levels were found in familial and sporadic AD patients, and correlated with poor memory performance in patients with AD. Thus, in vitro and in vivo studies indicate that sAPPα is downregulated during AD. Numerous analyses have shown that sAPPα ectodomain has neurotrophic and neuroprotective effects in different models of neuronal stress. In addition, sAPPα exhibits memory-enhancing properties. Intracerebroventricular infusion of anti-sAPPα serum was deleterious for memory, while that of sAPPα was beneficial. However, these studies relied on an exogenous excess of sAPPα and mechanisms of action and potential targets remained to be elucidated. With knock-in mice experiments, showed that sAPPα was sufficient to correct the impairments in spatial learning and long-term potentiation that are present in APP KO mice. This study shows in Drosophila that sAPPL is able to fully rescue the STM deficit caused by a reduction in endogenous APPL level, thus establishing that an APPL soluble form plays a role in memory, and giving further support for a role of secreted forms in memory in mammal systems (Bourdet, 2015).

When the fly α-secretase, KUZ, was overexpressed in the adult MB, no STM-enhancing effect was seen and, unexpectedly, KUZ overexpression in the MB of flies with an APPL partial loss of function exacerbated their memory impairment. Thus, KUZ overexpression was actually deleterious for memory, rather than beneficial. These results contrast with a previous study showing that overexpression of the mammalian α-secretase ADAM10 in an AD mice model led to an increase in sAPPα, and was able to overcome APP-related learning deficits. However, these studies showed that α-secretase activation has a positive impact on memory exclusively under conditions where human APP is overexpressed. In wild-type mice, results were not clear because overexpression of either the wild-type or an inactive form of the bovine ADAM10 altered learning and memory. Furthermore, ADAM10 has many substrates, and no evidence was brought to link the memory deficit to APP (Bourdet, 2015).

Interestingly, this study observed that KUZ overexpression decreases membrane nonproteolyzed APPL level, suggesting that its negative impact on memory in APPL LOF flies is linked to a reduction of nonproteolyzed APPL level. Therefore, strategies aimed at increasing APP α-cleavage may not be appropriate as this could provoke a decrease of fl-APP levels that might be deleterious to APP function (Bourdet, 2015).

Transient expression of a constitutive membrane-bound mutant APPL has the capacity to fully rescue the STM deficit caused by APPL partial loss of function. Thus, both sAPPL and fl-APPL appear to be involved in memory processes. This is in apparent contradiction with the observation that mammalian sAPPα was sufficient to correct spatial learning deficit of APP KO mice. However, in this study APP-like proteins APLP1 and ALPL2 were preserved, and as it is known from double KO analyses that the three APP homologs exert functional redundancy, they may have compensated for the loss of essential fl-APP functions. In consequence, one cannot attribute the memory function exclusively to sAPPα (Bourdet, 2015).

If both fl-APPL and sAPPL carry the capacity to restore wild-type STM in APPL partial LOF flies, it is puzzling to observe that KUZ overexpression in this genetic context is deleterious for memory. Indeed, in addition to causing a decrease in fl-APPL, KUZ overexpression leads to a concomitant increase in sAPPL that should be able to complement fl-APPL deficiency. It is suggested that in this context, fl-APPL level is below threshold so that even high levels of sAPPL cannot restore a wild-type memory. This hypothesis is supported by protein quantification experiments showing a 30% decrease in fl-APPL level. Because APPL was extracted from the whole brain, whereas KUZ overexpression was only driven in a subset of neurons, the effective fl-APPL decrease in the MB must be much higher than 30%. In mammalian cells under steady-state levels, ~10% of APP is located at the plasma membrane. APP has long been suggested to act as a cell-surface receptor; however, such a function has not been unequivocally established. Several reports have shown that APP exists as homodimers. Cis-dimerization of APP would represent a potential mechanism for a negative regulation of APP functions and a concomitant impact on Aβ generation via an increase in β-processing. Interestingly, it has been suggested that APP is a receptor for sAPPα as its binding could disrupt APP dimers (Bourdet, 2015).

In Drosophila, it has been reported that the secreted N-terminal ectodomain of APPL acts as a soluble ligand for neuroprotective functions. Furthermore, coimmunoprecipitation experiments from transfected Drosophila MB intrinsic cells revealed a physical interaction between fl-APPL and sAPPL, suggesting that sAPPL could be a ligand for fl-APPL. The current data showing the involvement of both membrane fl-APPL and sAPPL in memory are consistent with the hypothesis that sAPPL could be a ligand for its own fl-APPL precursor (Bourdet, 2015).

In conclusion, these data reveal for the first time a role for membrane fl-APPL in memory, opening new questions about APP nonpathological functions and relations between secreted and full-length forms in memory processes (Bourdet, 2015).

Go to top

Developmental inhibition of miR-iab8-3p disrupts mushroom body neuron structure and adult learning ability

Busto, G. U., Guven-Ozkan, T., Chakraborty, M. and Davis, R. L. (2016). Dev Biol [Epub ahead of print]. PubMed ID: 27634569

MicroRNAs are small non-coding RNAs that inhibit protein expression post-transcriptionally. They have been implicated in many different physiological processes, but little is known about their individual involvement in learning and memory. Several miRNAs have been identified that either increased or decreased intermediate-term memory when inhibited in the central nervous system, including miR-iab8-3p. This paper reports a new developmental role for this miRNA. Blocking the expression of miR-iab8-3p during the development of the organism leads to hypertrophy of individual mushroom body neuron soma, a reduction in the field size occupied by axonal projections, and adult intellectual disability. Four potential mRNA targets of miR-iab8-3p were identified whose inhibition modulates intermediate-term memory including ceramide phosphoethanolamine synthase, which may account for the behavioral effects produced by miR-iab8-3p inhibition. These results offer important new information on a microRNA required for normal neurodevelopment and the capacity to learn and remember normally (Busto, 2016).

Go to top

Aging impairs protein-synthesis-dependent long-term memory in Drosophila

Tonoki, A. and Davis, R.L. (2015). J Neurosci 35: 1173-1180. PubMed ID: 25609631

Although aging is known to impair intermediate-term memory in Drosophila, its effect on protein-synthesis-dependent long-term memory (LTM) is unknown. This study shows that LTM is impaired with age, not due to functional defects in synaptic output of mushroom body (MB) neurons, but due to connectivity defects of dorsal paired medial (DPM) neurons with their postsynaptic MB neurons. GFP reconstitution across synaptic partners (GRASP) experiments revealed structural connectivity defects in aged animals of DPM neurons with MB axons in the α lobe neuropil. As a consequence, a protein-synthesis-dependent LTM trace in the α/β MB neurons fails to form. Aging thus impairs protein-synthesis-dependent LTM along with the α/β MB neuron LTM trace by lessening the connectivity of DPM and α/β MB neurons (Tonoki, 2015).

The data presented in this study offer several important findings about the neural circuitry and the forms of memory disrupted by aging. First, it shows that aging impairs only one of the two mechanistically distinct forms of LTM generated by spaced, aversive classical conditioning in Drosophila. LTM that is independent of protein synthesis remains unaffected by age, whereas that form of LTM requiring protein synthesis becomes impaired. Therefore, there is mechanistic specificity in the effects of aging on LTM. Although aging, in principal, could disrupt processes like protein synthesis at the molecular level leading to a LTM deficit, these results indicate that the problem is traceable to the circuitry involved in generating protein-synthesis-dependent LTM (Tonoki, 2015).

The normal synaptic transmission from DPM neurons onto follower neurons during spaced training that is required for generating LTM is lost with age. This is attributable to the reduction of synaptic contacts between DPM neuron processes and MB axons specifically in the tip of the α lobe neuropil as revealed by GRASP signals. The loss of synaptic contacts between DPM and MB neurons in this region also may explain why synaptic blockade of DPM neurons during acquisition disrupts protein-synthesis-dependent LTM in young but not old flies. Therefore, a second major finding is that neural contacts and subsequent synaptic activity between DPM and α/β MB neurons are required for generating protein-synthesis-dependent LTM, and aging impairs this process. Consistent with this model, it is found that aging blocks the formation of a calcium-based, protein-synthesis-dependent memory trace in the α/β MB neurons (Tonoki, 2015).

It was found previously that ITM is impaired in flies of 30 d of age along with the capacity to form an ITM trace in the DPM neurons. Nevertheless, aging does not compromise the capacity to form an STM trace in the α'/β' MB neurons. Therefore, aging disrupts specific temporal forms of memory, including ITM and protein synthesis LTM, but not STM and protein-synthesis-independent LTM. It is possible that the loss of connectivity of DPM neurons with the α tip neuropil is responsible for the loss of both ITM and protein-synthesis-dependent LTM, along with their respective memory traces. Previous and this study's data indicate that STM appears to bypass the DPM neurons, whereas the reciprocal activity between DPM and MB neurons is required for ITM and LTM. Aging puts a kink in this neural system by impairing connectivity (Tonoki, 2015).

The study offers a model to explain the neural circuitry involved in protein-synthesis-dependent LTM formation and how aging impairs this form of memory. Although DPM neurons make contacts widely throughout the MB lobe neuropil with processes of many cell types, the critical interaction for LTM formation occurs in the vertical lobes of the MB through contacts onto the axons of α/β MB neurons. DPM neuron synaptic activity during spaced training, which occurs due to their stimulation by MB neurons, promotes synaptic changes in the postsynaptic α/β MB neurons and leads to the formation of memory trace in the α/β MB neurons. Aging impairs protein-synthesis-dependent LTM along with a LTM trace that normally forms in the α/β MB neurons by lessening the connectivity of DPM and α/β MB neurons. Identifying the mechanisms by which the DPM neurons lose their connectivity with only the tips of α/β MB neurons might reveal how aging impairs protein-synthesis-dependent LTM (Tonoki, 2015).

Go to top

Parallel circuits control temperature preference in Drosophila during ageing

Shih, H.W., Wu, C.L., Chang, S.W., Liu, T.H., Sih-Yu Lai, J., Fu, T.F., Fu, C.C. and Chiang, A.S. (2015). Nat Commun 6: 7775. PubMed ID: 26178754

The detection of environmental temperature and regulation of body temperature are integral determinants of behaviour for all animals. These functions become less efficient in aged animals, particularly during exposure to cold environments, yet the cellular and molecular mechanisms are not well understood. This study identifies an age-related change in the temperature preference of adult fruit flies that results from a shift in the relative contributions of two parallel mushroom body (MB) circuits-the β'- and β-systems. The β'-circuit primarily controls cold avoidance through dopamine signalling in young flies, whereas the β-circuit increasingly contributes to cold avoidance as adult flies age. Elevating dopamine levels in β'-afferent neurons of aged flies restores cold sensitivity, suggesting that the alteration of cold avoidance behaviour with ageing is functionally reversible. These results provide a framework for investigating how molecules and individual neural circuits modulate homeostatic alterations during the course of senescence (Shih, 2015).

Go to top

Reciprocal synapses between mushroom body and dopamine neurons form a positive feedback loop required for learning

Cervantes-Sandoval, I., Phan, A., Chakraborty, M. and Davis, R. L. (2017). Elife 6. PubMed ID: 28489528

Current thought envisions dopamine neurons conveying the reinforcing effect of the unconditioned stimulus during associative learning to the axons of Drosophila mushroom body Kenyon cells for normal olfactory learning. This study shows, using functional GFP reconstitution experiments, that Kenyon cells and dopamine neurons form axoaxonic reciprocal synapses. The dopamine neurons receive cholinergic input via nicotinic acetylcholine receptors from the Kenyon cells; knocking down these receptors impairs olfactory learning revealing the importance of these receptors at the synapse. Blocking the synaptic output of Kenyon cells during olfactory conditioning reduces presynaptic calcium transients in dopamine neurons (DAn), a finding consistent with reciprocal communication. Moreover, silencing Kenyon cells decreases the normal chronic activity of the dopamine neurons. These results reveal a new and critical role for positive feedback onto dopamine neurons through reciprocal connections with Kenyon cells for normal olfactory learning (Cervantes-Sandoval, 2017).

The results demonstrate that DAn are both pre- and post-synaptic to KC through axoaxonic reciprocal connections, in contrast to current models which envision them as providing only pre-synaptic input. The DAn>KC half of the reciprocal synapse employs DA as neurotransmitter, although the possibility cannot be ruled out that other neurotransmitters are co-released with DA. The KC>DAn half of the reciprocal synapse is cholinergic. However, the fact that it was not possible to observe both cAMP and calcium responses in DAn with KC stimulation suggests that there may be other mediators of this reciprocal connection. Blocking the cholinergic input to DAn attenuates aversive olfactory learning, providing evidence that its function is, at least in part, to provide an amplification signal for the initial DA release due to activating the US pathway. Consistent with this role it was found that silencing KC impairs DAn presynaptic calcium responses to conditioning, odor and shock stimuli, presumably influencing dopamine release and explaining the learning phenotype. Overall, results support the existence of a positive feedback loop required for optimal learning. It is envisioned that DAn receive direct input from the US during conditioning which is conveyed to KC. The KCs also receive coincident olfactory input and this coincidence provides positive feedback onto the DAn through cholinergic synapses to further increase DAn activity (Cervantes-Sandoval, 2017).

The results also show that KC input to DAn shapes their ongoing or chronic activity. It is plausible that ongoing activity in the DAn provides a moment-by-moment update of the external environment and internal states and the behavioral status of the fly that appropriately reconfigures the KC>MBOn flow of information. Thus, the DAn/KC/MBOn circuit may form a recurrent network that serves as the insect's brain center for the rapid integration of sensory information and decision-making. Local feedback loops, achieved by reciprocal connectivity like that described in this study, may provide computational benefits to fine tune and optimize the output. Behavioral flexibility may be achieved by passing information through local reentrant loops with constant updating from the external or internal state of the organism (Cervantes-Sandoval, 2017).

A caveat in this and other studies is that it is not possible to exclude that other, non-KC, cholinergic input to DAn contributes to memory acquisition. Massive efforts to generate 'connectomes' in multiple species may offer resolution to this issue at some point in the future. Alternatively, they may continue to reveal additional connections and complexities that defy an immediate understanding. Studies like the present one, that reveal unexpected relationships between synaptic partners in difficult-to-untangle circuits, expose the need to advance beyond 'connectomics' and develop new tools that allow silencing or activation of specific channels and specific synaptic connections between neurons of interest without affecting other functions in the same cells (Cervantes-Sandoval, 2017).

Go to top

Upregulated energy metabolism in the Drosophila mushroom body is the trigger for long-term memory

Placais, P. Y., de Tredern, E., Scheunemann, L., Trannoy, S., Goguel, V., Han, K. A., Isabel, G. and Preat, T. (2017). Nat Commun 8: 15510. PubMed ID: 28580949

Efficient energy use has constrained the evolution of nervous systems. However, it is unresolved whether energy metabolism may resultantly regulate major brain functions. The observation that Drosophila flies double their sucrose intake at an early stage of long-term memory formation initiated the investigation of how energy metabolism intervenes in this process. Cellular-resolution imaging of energy metabolism reveals a concurrent elevation of energy consumption in neurons of the mushroom body, the fly's major memory centre. Strikingly, upregulation of mushroom body energy flux is both necessary and sufficient to drive long-term memory formation. This effect is triggered by a specific pair of dopaminergic neurons afferent to the mushroom bodies, via the D5-like DAMB dopamine receptor. Hence, dopamine signalling mediates an energy switch in the mushroom body that controls long-term memory encoding. These data thus point to an instructional role for energy flux in the execution of demanding higher brain functions (Placais, 2017).

Go to top

Identification and characterization of mushroom body neurons that regulate fat storage in Drosophila

Al-Anzi, B. and Zinn, K. (2018). Neural Dev 13(1): 18. PubMed ID: 30103787

Two neuronal populations, c673a and Fru-GAL4, regulate fat storage in fruit flies. Both populations partially overlap with a structure in the insect brain known as the mushroom body (MB), which plays a critical role in memory formation. This overlap prompted an examination of whether the MB is also involved in fat storage homeostasis. Using a variety of transgenic agents, the neural activity of different portions of the MB and associated neurons were selectively manipulated to decipher their roles in fat storage regulation. The data show that silencing of MB neurons that project into the &alpha:'β' lobes decreases de novo fatty acid synthesis and causes leanness, while sustained hyperactivation of the same neurons causes overfeeding and produces obesity. The &alpha:'β' neurons oppose and dominate the fat regulating functions of the c673a and Fru-GAL4 neurons. It was also shown that MB neurons that project into the γ lobe also regulate fat storage, probably because they are a subset of the Fru neurons. It was possible to identify input and output neurons whose activity affects fat storage, feeding, and metabolism. The activity of cholinergic output neurons that innervating the β'2 compartment (MBON-β'2mp and MBON-γ5β'2a) regulates food consumption, while glutamatergic output neurons innervating α' compartments (MBON-γ2α'1 and MBON-α'2) control fat metabolism. This study has identified a new fat storage regulating center, the α'β' lobes of the MB. The study also delineated the neuronal circuits involved in the actions of the α'β' lobes, and showed that food intake and fat metabolism are controlled by separate sets of postsynaptic neurons that are segregated into different output pathways (Al-Anzi, 2018).

Go to top

Convergence of monosynaptic and polysynaptic sensory paths onto common motor outputs in a Drosophila feeding connectome

Miroschnikow, A., Schlegel, P., Schoofs, A., Hueckesfeld, S., Li, F., Schneider-Mizell, C. M., Fetter, R. D., Truman, J. W., Cardona, A. and Pankratz, M. J. (2018). Elife 7. PubMed ID: 30526854

This study reconstructed, from a whole CNS EM volume, the synaptic map of input and output neurons that underlie food intake behavior of Drosophila larvae. Input neurons originate from enteric, pharyngeal and external sensory organs and converge onto seven distinct sensory synaptic compartments within the CNS. Output neurons consist of feeding motor, serotonergic modulatory and neuroendocrine neurons. Monosynaptic connections from a set of sensory synaptic compartments cover the motor, modulatory and neuroendocrine targets in overlapping domains. Polysynaptic routes are superimposed on top of monosynaptic connections, resulting in divergent sensory paths that converge on common outputs. A completely different set of sensory compartments is connected to the mushroom body calyx. The mushroom body output neurons are connected to interneurons that directly target the feeding output neurons. These results illustrate a circuit architecture in which monosynaptic and multisynaptic connections from sensory inputs traverse onto output neurons via a series of converging paths (Miroschnikow, 2018).

Motor outputs of a nervous system can be broadly defined into those carried out by the muscles to produce movements and by the glands for secretion. Both of these behavioral and physiological events are regulated by a network of output neurons, interneurons and sensory neurons, and a major open question is how one neural path is selected from multiple possible paths to produce a desired output. Nervous system complexity and tool availability have strongly dictated the type of experimental system and analysis that can be used to address this issue, such as a focus on a particular organism, behavior or type of neuron. In this context, the detailed illustrations of different parts of nervous systems at neuronal level as pioneered by Cajal, to the first complete description of a nervous system wiring diagram at synaptic level for C. elegans, demonstrate the power of systematic neuroanatomical analysis in providing a foundation and guide for studying nervous system function. However, the technical challenges posed by such analysis have limited the type of organisms for which synaptic resolution mapping can be performed at the scale of an entire nervous system (Miroschnikow, 2018).

Analysis of the neural circuits that mediate food intake in the Drosophila larvae offers numerous advantages in meeting the challenge of neuroanatomical mapping at a whole brain level, and combining it with the ability to perform behavioral and physiological experiments. The muscle system that generates the different movements necessary for transporting food from the pharynx to the esophagus, as well as the endocrine system responsible for secreting various hormones for metabolism and growth, have both been well described. These are also complemented by the analysis of feeding behavior in adult flies. Although there is broad knowledge at the morphological level on the organs underlying larval feeding behavior and physiology, as well as on the nerves innervating them in the periphery, the central connectivity of the afferent and efferent neurons within these nerves are largely unknown. At the same time, advances in the EM reconstruction of an entire CNS of a first instar larva offers an opportunity to elucidate an animals' feeding system on a brain-wide scale and at synaptic resolution. As part of this community effort, we recently performed an integrated analysis of fast synaptic and neuropeptide receptor connections for an identified cluster of 20 interneurons that express the neuropeptide hugin, a homolog of the mammalian neuropeptide neuromedin U, and which regulates food intake behavior. This analysis showed that the class of hugin neurons modulating food intake receives direct synaptic inputs from a specific group of sensory neurons, and in turn, makes mono-synaptic contacts to output neuroendocrine cells. The study not only provided a starting point for a combined approach to studying synaptic and neuropeptidergic circuits, but a basis for a comprehensive mapping of the sensory and output neurons that innervate the major feeding and endocrine organs. (Miroschnikow, 2018).

Feeding is one of the most universal and important activities that animals engage in. Despite large differences in the morphology of the external feeding organs, the internal gut structures are quite similar across different animals; indeed, even within closely related species, there can be large differences in the external organs that detect and gather food, whereas the internal organs that transport food through the alimentary canal are much more similar. Recent studies have also pointed out the functional similarities between the subesophageal zone in insects and the brainstem in vertebrates for regulating feeding behavior. In mammals, the different cranial nerves from the medulla innervate distinct muscles and glands of the foregut. For example, the VIIth cranial nerve (facial nerve) carries taste sensory information from anterior 2/3 of the tongue, and innervates the salivary glands, and lip and facial muscles. The IXth cranial nerve (glossopharyngeal nerve) receives taste inputs from the posterior 1/3 of the tongue, and innervates the salivary glands and pharynx muscles. The Xth cranial nerve (vagus nerve) receives majority of the sensory inputs from the enteric nervous system of the gut, and innervates pharynx and esophagus muscles. The XIth cranial nerve (spinal accessory nerve) and the XIIth cranial nerve (hypoglossal nerve) are thought to carry strictly motor information which innervate the pharynx and neck muscles, and the tongue muscles. The distinct cranial nerves project onto topographically distinct areas in the medulla of the brainstem. It is also noted that olfactory information is carried by cranial nerve I, a strictly sensory nerve that projects to the olfactory bulb (OB), an area topographically distinct from the brainstem. In addition, there are direct neuronal connections between the brainstem and the hypothalamus, the key neuroendocrine center of vertebrates (Miroschnikow, 2018).

Analogously, distinct pharyngeal nerves of the Drosophila larva are connected to the subesophageal zone (SEZ), and also carry sensory and motor information that regulate different parts of the body. The AN (antennal nerve) carries sensory information from the olfactory, pharyngeal and internal organs, and innervates the pharyngeal muscles for pumping in food. The serotonergic neurons that innervate the major endocrine center and the enteric nervous system also project through the AN. Note also that the olfactory sensory organs project to the antennal lobe (AL), which abuts the SEZ yet is topographically separate. The MxN (maxillary nerve) carries external and pharyngeal sensory information, and innervates the mouth hooks, whose movements are involved in both feeding and locomotion. The PaN (prothoracic accessory nerve) carries external sensory information from the upper head region, and innervates the muscles involved in head tilting. Furthermore, the SEZ has direct connections to median neurosecretory cells (mNSCs) and the ring gland. In sum, although a large body of knowledge exists on the gross anatomy of the nerves that target the feeding organs in vertebrates and invertebrates, the synaptic pathways within the brain that interconnect the sensory inputs and output neurons of the individual nerves remain to be elucidated (Miroschnikow, 2018).

This paper has reconstructed all sensory, serotonergic modulatory (Se0) and motor neurons of the three pharyngeal nerves that underlie the feeding motor program of Drosophila larvae. The activity of these nerves has previously been shown to be sufficient for generating the feeding motor pattern in isolated nervous system preparations, and that the central pattern generators (CPGs) for food intake lie in the SEZ. This study then identified all monosynaptic connections between the sensory inputs and the motor, Se0 and previously described median neurosecretory ouput neurons, thus providing a full monosynaptic reflex circuit for food intake. Polysynaptic pathways were also mapped that are integrated onto the monosynaptic reflex circuits. In addition, the multisynaptic non-olfactory neuron connections from the sensory neurons to the mushroom body memory circuit were mapped, and these were shown to be different from those involved in monosynaptic reflex circuits. Finally, a set of mushroom body output neurons were traced onto the neurosecretory and other feeding output neurons. Reflex circuits can be seen to represent the simplest synaptic architecture in the nervous system, as formulated by Charles Sherrington. Anatomical reconstructions of monosynaptic and polysynaptic reflex circuits can also be seen in the works of Cajal. A model is proposed of how different mono- and polysynaptic pathways can be traversed from a set of sensory neurons to specific output neurons, which has relevance for understanding the mechanisms of action selection (Miroschnikow, 2018).

This study provides a comprehensive synaptic map of the sensory and output neurons that underlie food intake and metabolic homeostasis in Drosophila larva. Seven topographically distinct sensory compartments, based on modality and peripheral origin, subdivide the SEZ, a region with functional similarities to the vertebrate brainstem. Sensory neurons that form monosynaptic connections are mostly of enteric origin, and are distinct from those that form multisynaptic connections to the mushroom body (MB) memory circuit. Different polysynaptic connections are superimposed on the monosynaptic input-ouput pairs that comprise the reflex arc. Such circuit architecture may be used for controlling feeding reflexes and other instinctive behaviors (Miroschnikow, 2018).

Reflex circuits represent a basic circuit architecture of the nervous system, whose anatomical and physiological foundations were laid down by Cajal and Sherrington. The Drosophila larval feeding reflex circuit comprises the motor neurons that innervate the muscles involved in pharyngeal pumping, as well as the neurosecretory neurons that target the endocrine organs. They also include a cluster of serotonergic neurons that innervate the entire enteric nervous system, and which may have neuromodulatory effects on the feeding system in a global manner. The vast majority of output neurons are targeted monosynaptically from a set of topographically distinct sensory synaptic compartments in the CNS. These compartments target the output neurons in overlapping domains: the first, ACa, targets all neuroendocrine cells as well as the serotonergic neurons; the second, AVa, targets a subset of neuroendocrine cells, the serotonergic neurons and most of the pharyngeal motor neurons, while the third, AVp, targets the serotonergic neurons and a different set of pharyngeal motor neurons. With these outputs, one can in principle fulfill the most basic physiological and behavioral needs for feeding: neurosecretory cells for metabolic regulation and pharyngeal motor neurons for food intake. This set of monosynaptic connections can thus be seen to represent an elemental circuit for feeding, since the connections between the input and output neurons cannot be broken down any further (Miroschnikow, 2018).

Vast majority of the sensory inputs comprising this 'elemental feeding circuit'derive from the enteric nervous system to target the pharyngeal muscles involved in food intake and neuroendocrine output organs. However, there is a small number of monosynaptic reflex connections that originate from the somatosensory compartment. The output neurons targeted by these somatosensory neurons are motor neurons that control mouth hook movements and head tilting, movements which are involved in both feeding and locomotion. In this context, it is noteworthy that monosynaptic reflex connections are found to a much lesser degree in the larval ventral nerve cord, which generates locomotion. An analogous situation exists in C. elegans, where majority of the monosynaptic reflex circuits are found in the head motor neurons and not in the body. One reason could be due to the relative complexity in the response necessary for food intake as compared to locomotion. For example, a decision to finally not to swallow a harmful substance, once in the mouth, may require a more local response, for example muscles limited to a very specific region of the pharynx and esophagus, where monosynaptic arc might suffice. By contrast, initiating escape behaviors requires a more global response with respect to the range and coordination of body movements involved, although it also employs multimodal sensory integration via a multilayered circuit (Miroschnikow, 2018).

The inter-sensory connections show a combination of hierarchical and reciprocal connections, which may increase the regulatory capability and could be especially important for monosynaptic circuits. By contrast, very few monosynaptic connections exist between the larval olfactory, chordotonal or nociceptive class IV sensory neurons in the body. Interestingly, there is also a much higher percentage of intersensory connections between olfactory receptor neurons in the adult as compared to the larva, which could function in gain modulation at low signal intensities. This might be attributable to adults requiring faster processing of olfactory information during flight navigation (or mating), and/or to minimize metabolic cost. Whether such explanation also applies to the differences in intersensory connection between the different types of sensory neurons in the larvae remains to be determined (Miroschnikow, 2018).

Very few cases were found where a monosynaptic path between any sensory-output pair is not additionally connected via a polysynaptic path. An interesting question in the context of action selection mechanism is which path a sensory signal uses to reach a specific target neuron. For example, a very strong sensory signal may result in a monosynaptic reflex path being used. However, a weaker sensory signal may result in using a different path, such as one with less threshold for activation. This would also enable the integration of different types of sensory signals through the usage of multiple interneurons, since the interneurons may receive sensory inputs that are not present in monosynaptic connections. For example, sensory neurons can target the neuroendocrine cells directly (monosynaptically), or through a hugin interneuron (di-synaptically). The sensory compartments that directly target the neuroendocrine cells are of enteric origin; however, when hugin neurons are utilized as interneurons, not only is the number of sensory neurons from the same sensory compartment increased, but sensory neurons are added from a completely new peripheral origin. Thus, the hugin interneurons enable sensory inputs from different peripheral origins, for example to integrate enteric inputs with pharyngeal gustatory inputs, to influence an output response, which, in this case, is to stop feeding (Miroschnikow, 2018).

The coexistence of polysynaptic and monosynaptic paths could also be relevant for circuit variability and compensation: destruction of any given path would still enable the circuit to function, but with more restrictions on the precise types of sensory information it can respond to. In certain cases, this may even lead to strengthening of alternate paths as a form of synaptic plasticity (Miroschnikow, 2018).

An open issue is how the sensory synaptic compartments might be connected to the feeding central pattern generators (CPGs) which have been demonstrated to exist in the SEZ, especially since CPGs are defined as neural circuits that can generate rhythmic motor patterns in the absence of sensory input. However, the modulation of CPG rhythmic activity can be brought about by sensory and neuromodulatory inputs. A complete circuit reconstruction of the larval SEZ circuit may shed some light on the circuit structure of feeding CPGs (Miroschnikow, 2018).

A more complex circuit architecture is represented by the MB, the site of associative learning and memory in insects: a completely different set of sensory synaptic compartments is used to connect the various projection neurons to the MB calyx. Thus, the MB module is not superimposed onto the monosynaptic reflex circuits but rather forms a separate unit. The classical studies by Pavlov demonstrated conditioned reflex based on an external signal and an autonomic secretory response in response to food. Although a comparable autonomic response has not been analyzed in the larvae, analogous associative behavior based on odor choice response has been well studied. It is also noteworthy that in the Aplysia, classical conditioning of the gill withdrawal reflex involves monosynaptic connections between a sensory neuron (mechanosensory) and a motor neuron, and neuromodulation by serotonin. This constellation has similarities with the elemental feeding circuit consisting of sensory, motor and serotonergic modulatory neurons. For more complex circuits of feeding behavior in the mouse, a memory device for physiological state, such as hunger, has been reported involving synaptic and neuropeptide hormone circuits. Functional studies on MB output neurons such as the MBON-f1, which may be part of a 'psychomotor' pathway and which targets a number of interneurons that connect to the neurosecretory, serotonergic and pharyngeal motor neurons, may help address how memory circuits interact with feeding circuits (Miroschnikow, 2018).

Feeding behavior manifests itself from the most primitive instincts of lower animals, to deep psychological and social aspects in humans. It encompasses cogitating on the finest aspects of food taste and the memories evoked by the experience, to sudden reflex reactions upon unexpectedly biting down on a hard seed or shell. Both of these extremes are mediated, to a large degree, by a common set of feeding organs, but the way these organs become utilized can vary greatly. The architecture of the feeding circuit described in this study allows the various types of sensory inputs to converge on a limited number of output responses. The monosynaptic pathways would be used when fastest response is needed. The presence of polysynaptic paths would enable slower and finer control of these reflex events by allowing different sensory inputs, strengths or modalities to act on the monosynaptic circuit. This can be placed in the context in the control of emotions and survival circuits, or by cortex regulation of basic physiological or autonomic processes. In a striking example, pupil dilation, a reflex response, has been used as an indicator of cognitive activity. Here, a major function of having more complex circuit modules on top of monosynaptic circuits may be to allow a finer regulation of feeding reflexes, and perhaps of other reflexes or instinctive behaviors (Miroschnikow, 2018).

As an outlook, this analysis provides an architectural framework of how a feeding circuit is organized in the CNS. The circuit is divided into two main axes that connect the input to the output systems: the sensory-neurosecretory cell axis and the sensory-motor neuron axis. The sensory system targets overlapping domains of the output neurons; for example, a set of sensory neurons targets exclusively the neuroendocrine cells, other targets both neuroendocrine and pharyngeal motor neurons, and another just the pharyngeal motor neurons. The inputs derive mostly from the internal organs. These connections form the monosynaptic reflex circuits. With these circuits, one can perform the major requirements of feeding regulation, from food intake and ingestion to metabolic homeostasis. Additional multisynaptic circuits, such as the CPGs, those involving sensory signaling from the somatosensory system (external inputs), or those comprising the memory circuits, are integrated or added to expand the behavioral repertoire of the animal (see Input-output synaptic organization of the larval feeding system and its connectivity architecture in the brain). Although circuit construction may proceed from internal to the external, the sequence is reversed in a feeding animal: the first sensory cues are external (olfactory), resulting in locomotion (somatic muscles) that can be influenced by memory of previous experience; this is followed by external taste cues, resulting in food intake into the mouth; the final action is the swallowing of food, involving pharyngeal and enteric signals and reflex circuits. However, regardless of the types of sensory inputs, and whether these are transmitted through a reflex arc, a memory circuit or some other multisynaptic circuits in the brain, they will likely converge onto a certain set of output neurons, what Sherrington referred to as the 'final common path'. The current work is a first step towards finding the common paths (Miroschnikow, 2018)

Go to top

Immune Receptor Signaling and the Mushroom Body Mediate Post-ingestion Pathogen Avoidance

Kobler, J. M., Rodriguez Jimenez, F. J., Petcu, I. and Grunwald Kadow, I. C. (2020). Curr Biol. PubMed ID: 33007248

In spite of the positive effects of bacteria on health, certain species are harmful, and therefore, animals must weigh nutritional benefits against negative post-ingestion consequences and adapt their behavior accordingly. This study used Drosophila to unravel how the immune system communicates with the brain, enabling avoidance of harmful foods. Using two different known fly pathogens, mildly pathogenic Erwinia carotovora (Ecc15) and highly virulent Pseudomonas entomophila (Pe), preference behavior was analyzed in naive flies and after ingestion of either of these pathogens. Although survival assays confirmed the harmful effect of pathogen ingestion, naive flies preferred the odor of either pathogen to air and also to harmless mutant bacteria, suggesting that flies are not innately repelled by these microbes. By contrast, feeding assays showed that, when given a choice between pathogenic and harmless bacteria, flies-after an initial period of indifference-shifted to a preference for the harmless strain, a behavior that lasted for several hours. Flies lacking synaptic output of the mushroom body (MB), the fly's brain center for associative memory formation, lost the ability to distinguish between pathogenic and harmless bacteria, suggesting this to be an adaptive behavior. Interestingly, this behavior relied on the immune receptors PGRP-LC and -LE and their presence in octopaminergic neurons. A model is postulated wherein pathogen ingestion triggers PGRP signaling in octopaminergic neurons, which in turn relay the information about the harmful food source directly or indirectly to the MB, where an appropriate behavioral output is generated (Kobler, 2020).

Go to top

A neural circuit arbitrates between persistence and withdrawal in hungry Drosophila

Sayin, S., De Backer, J. F., Siju, K. P., Wosniack, M. E., Lewis, L. P., Frisch, L. M., Gansen, B., Schlegel, P., Edmondson-Stait, A., Sharifi, N., Fisher, C. B., Calle-Schuler, S. A., Lauritzen, J. S., Bock, D. D., Costa, M., Jefferis, G., Gjorgjieva, J. and Grunwald Kadow, I. C. (2019). Neuron 104(3):544-558. PubMed ID: 31471123

In pursuit of food, hungry animals mobilize significant energy resources and overcome exhaustion and fear. How need and motivation control the decision to continue or change behavior is not understood. Using a single fly treadmill, this study shows that hungry flies persistently track a food odor and increase their effort over repeated trials in the absence of reward suggesting that need dominates negative experience. It was further shown that odor tracking is regulated by two mushroom body output neurons (MBONs) connecting the MB to the lateral horn. These MBONs, together with dopaminergic neurons and Dop1R2 signaling, control behavioral persistence. Conversely, an octopaminergic neuron, VPM4, which directly innervates one of the MBONs, acts as a brake on odor tracking by connecting feeding and olfaction. Together, these data suggest a function for the MB in internal state-dependent expression of behavior that can be suppressed by external inputs conveying a competing behavioral drive (Sayin, 2019).

Flexibility is an important factor in an ever in-flux environment, where scarcity and competition are the norm. Without persistence to achieve its goals, however, an animal's strive to secure food, protect its offspring, or maintain its social status is in jeopardy. Therefore, sensory cues related to food or danger often elicit strong impulses. However, these impulses must be strictly controlled to allow for coherent goal-directed behavior and to permit behavioral transitions when sensible. Inhibition of antagonistic behavioral drives at the cognitive and physiological level has been proposed as a major task of a nervous system. Which sensory cues and ultimately which behaviors are prioritized and win depends on the animal's metabolic state, internal motivation, and current behavioral context. How this is implemented at the level of individual neurons, circuit motifs, and mechanisms remains an important open question (Sayin, 2019).

Like most animals, energy-deprived flies prioritize food seeking and feeding behavior. To find food, flies can follow olfactory or visual cues over long distances. External gustatory cues provide information about the type and quality of the eventually encountered food. However, only internal nutrient levels will provide reliable feedback about the quality and quantity of a food source and ultimately suppress food-seeking behaviors. Therefore, food odor, the taste of food, and post-ingestive internal feedback signals induce sequential and partly antagonistic behaviors. Interestingly, chemosensory and internal feedback systems typically mediated by distinct neuromodulators appear to converge in the mushroom body (MB). How neurons and neural circuits signal and combine external and internal cues to maintain or suppress competing behavioral drives is not well understood (Sayin, 2019).

In mammals, norepinephrine (NE) released by a brain stem nucleus, the locus coeruleus, has been implicated in controlling the balance between persistence and action selection. The potential functional counterpart of NE in insects could be octopamine (OA). Flies lacking OA indeed show reduced arousal, for instance upon starvation. Additionally, OA neurons (OANs) gate appetitive memory formation of odors and also modulate taste neurons and feeding behavior. OANs are organized in distinct clusters and project axons to diverse higher brain regions in a cell type-specific manner. The precise roles and important types of OA and NE neurons in state-dependent action selection remain to be elucidated (Sayin, 2019).

Similar to NE and OA, dopamine (DA) is being studied in many aspects of behavioral adaptation and flexibility. Different classes of DA neurons (DANs) innervating primarily the MB signal negative or positive context, or even wrong predictions (Sayin, 2019 and references therein).

This study took advantage of the small number and discrete organization of neuromodulatory neurons in the fly brain to analyze the mechanistic relationship between motivation-dependent persistence in one behavior and the decision to disengage and change to another behavior. Using a single fly spherical treadmill assay, this study found that hungry flies increase their effort to track a food odor with every unrewarded trial. MB output through two identified MBONs (MBON-γ1pedc>α/β and MBON-α2sc) is required for persistent odor tracking. MBON-α2sc provides a MB connection to the lateral horn (LH), where it can modify innate food odor attraction. Furthermore, this study pinpoints a specific type of OAN, VPM4 (ventral paired medial), which connects feeding centers directly to MBON-γ1pedc>α/β and disrupts food odor tracking. Finally, the experimental data suggest that persistent tracking depends on DANs, including PPL1-γ1pedc, and signaling through dopamine receptor Dop1R2 in αβ-type KCs. Based on these results, it is proposed that MB output and a direct external input, depending on internal state and motivation, gradually promote or interrupt ongoing behavior (Sayin, 2019).

What drives gradually increasing persistence in behavior? For the fly, a model is proposed by which a circuit module of KCs, MBONs, and DANs drive gradually increasing odor tracking, which can be efficiently suppressed by extrinsic MBON-innervating feeding-related OANs. Behavioral persistence has been previously analyzed in flies in a different context. For instance, courtship of fly males and copulation with a female are maintained by dopaminergic neurons in the ventral nerve cord, where they counteract GABAergic neurons. In that scenario, DANs in the ventral nerve cord maintain an ongoing behavior and prevent male premature disengagement before successful insemination (Sayin, 2019).

The experimental data also implicate DANs, primarily from within the PPL1 (e.g., PPL1-γ1pedc) and PPL2ab clusters, and Dop1R2 signaling. In particular, inactivation of synaptic output of DANs positive for TH-Gal4 as well as loss of Dop1R2 in αβ-type KCs reduced the increase in odor tracking from trial to trial, while not affecting the speed at first odor stimulation. These data suggest that TH+ DANs promote goal-directed movement, i.e., odor tracking, through a Dop1R2-dependent mechanism in KCs (Sayin, 2019).

MBON-γ1pedc>αβ, which receives dopaminergic input by PPL1-γ1pedc, is required for odor tracking. Moreover, this study also observed a trial-to-trial decrease in odor response of this MBON, matching the dopamine-induced synaptic depression previously observed in MBONs upon learning. Notably, PPL1-γ1pedc activates Dop1R2 in MBON-γ1pedc>αβ, a signal recently found to be critical for appetitive long-term memory. Nevertheless, it appears that, in addition to PPL1-γ1pedc, other DANs regulate behavioral persistence by modulating in particular αβ-KCs. It is intriguing to speculate about a common function of Dop1R2 in the formation of long-lasting aversive memory induced by repeatedly pairing odor with an aversive experience and the behavior examined in this study: increased and persistent expression of a behavior induced by the experience of repeated failure to reach a goal (Sayin, 2019).

The experimental data further implicated MBON-α2sc, which is connected to MBON-γ1pedc>αβ. Calcium imaging data are consistent with an inhibitory interaction between the two MBONs. However, some of the behavioral data and prior imaging data do not support an inhibitory connection. Furthermore, MBON-γ1pedc>αβ projects to other brain regions and downstream targets, and similarly MBON-α2sc receives additional inputs—all of which could be equally or more important for persistent behavior than a direct connection between these two MBONs. Finally, some DANs respond to movement, including PPL1-γ2α'1/MV1. Although no essential role of this particular neuron was found in odor tracking persistence, movement might contribute to the activity of MBONs responding the odorant (Sayin, 2019).

Remarkably, MBON-α2sc connects the MB to neurons within the LH. Thus, it is speculated that the LH might assign an odor to its corresponding behavioral category, such as 'food-related' for vinegar, while the MB acts as a top-down control to gauge the expression of an innate behavior (i.e., tracking an appetitive odor) according to state and experience (Sayin, 2019).

The behavioral data led to the proposal of a circuit model. Using computational modeling, this study tested whether the MB network including DANs and MBONs could, in theory, produce the observed behavior. Indeed, it was found that a simplified recurrent circuit of KCs, DANs, and MBONs can account for the observed behavioral persistence and also the measured MBON-γ1pedc>αβ odor responses. While this model cannot replace experimental evidence, it forms a useful theoretical framework for future studies on the role of the MB in behavioral persistence (Sayin, 2019).

Based on the present data and computational predictions, a model is proposed by which the recurrent circuit architecture of the MB, in addition to storing information for future behavior, is ideally suited to maintain and gradually change ongoing behavior, for instance by modulating output of the LH, according to the animal's internal state and needs (Sayin, 2019).

The use of an olfactory treadmill has allowed dissection of the different aspects of a food search. In particular, how does food and feeding suppress food search if the sensory cue, the odor, is still present? OA-VPM4 connects feeding centers (i.e., SEZ) directly with odor tracking-promoting MBON-γ1pedc>αβ and inhibits its activity suggesting an inhibitory connection between VPM4 and the MBON. Nevertheless, it cannot be excluded that OA-VPM4 signals through multiple mechanisms including OA and possibly other neurotransmitters. In addition, a recent study showed that activation of VPM4 promotes proboscis extension to sugar. Although a direct role in taste detection through pharynx or labellum appears unlikely, it is possible that feeding behavior itself (e.g., lymphatic sugar, food texture, activity of feeding muscles) are detected and/or promoted by these neurons and then brought to the MB. It is proposef that VPM4 is a direct mediator between olfactory-guided food search and the rewarding experience of feeding and related behavior (Sayin, 2019).

The data provide a neural circuit mechanism empowering flies to express and prioritize behavior in a need- and state-dependent manner. It is exciting to speculate that fundamentally similar circuit motifs might exist in NE and DA neuron-containing circuits in the mammalian brain, governing the organization of behavior in a flexible and context-dependent manner by integrating internal and external context. For instance, noradrenergic neurons of the brainstem nucleus of the solitary tract (NST) receive taste information, and input from the gastrointestinal tracts, lungs, and heart. Neurons in the NST project to multiple brain regions including the amygdala, hypothalamus, and insular cortex, all of which receive internal state as well as other sensory information (Sayin, 2019).

The data in the fly provide an experimental and theoretical framework for a better understanding of the fundamental circuit mechanisms underpinning neuromodulation of context-dependent behavioral persistence and withdrawal (Sayin, 2019).

Go to top

A neural algorithm for Drosophila linear and nonlinear decision-making

Zhao, F., Zeng, Y., Guo, A., Su, H. and Xu, B. (2020). Sci Rep 10(1): 18660. PubMed ID: 33122701

It has been evidenced that vision-based decision-making in Drosophila consists of both simple perceptual (linear) decision and value-based (non-linear) decision. This paper proposes a general computational spiking neural network (SNN) model to explore how different brain areas are connected contributing to Drosophila linear and nonlinear decision-making behavior. First, the SNN model could successfully describe all the experimental findings in fly visual reinforcement learning and action selection among multiple conflicting choices as well. Second, the computational modeling shows that dopaminergic neuron-GABAergic neuron-mushroom body (DA-GABA-MB) works in a recurrent loop providing a key circuit for gain and gating mechanism of nonlinear decision making. Compared with existing models, this model shows more biologically plausible on the network design and working mechanism, and could amplify the small differences between two conflicting cues more clearly. Finally, based on the proposed model, the unmanned aerial vehicle (UAV) could quickly learn to make clear-cut decisions among multiple visual choices and flexible reversal learning resembling to real fly. Compared with linear and uniform decision-making methods, the DA-GABA-MB mechanism helps UAV complete the decision-making task with fewer steps (Zhao, 2020).

Go to top

Activation of specific mushroom body output neurons inhibits proboscis extension and sucrose consumption

Chia, J. and Scott, K. (2020). PLoS One 15(1): e0223034. PubMed ID: 31990947

In Drosophila, the mushroom bodies (MBs) are critical for olfactory associative learning and conditioned taste aversion, but how the output of the MBs affects specific behavioral responses is unresolved. In conditioned taste aversion, Drosophila shows a specific behavioral change upon learning: proboscis extension to sugar is reduced after a sugar stimulus is paired with an aversive stimulus. While studies have identified MB output neurons (MBONs) that drive approach or avoidance behavior, whether the same MBONs impact innate proboscis extension behavior is unknown. This study tested the role of MB pathways in altering proboscis extension and identified MBONs that synapse onto multiple MB compartments that upon activation significantly decreased proboscis extension to sugar. Activating several of these lines also decreased sugar consumption, revealing that these MBONs have a general role in modifying feeding behavior beyond proboscis extension. The MBONs that decreased proboscis extension and ingestion are different from those that drive avoidance behavior in another context. These studies provide insight into how activation of MB output neurons decreases proboscis extension to taste compounds (Chia, 2020)

Go to top

Social attraction in Drosophila is regulated by the mushroom body and serotonergic system

Sun, Y., Qiu, R., Li, X., Cheng, Y., Gao, S., Kong, F., Liu, L. and Zhu, Y. (2020). Social attraction in Drosophila is regulated by the mushroom body and serotonergic system. Nat Commun 11(1): 5350. PubMed ID: 33093442

Sociality is among the most important motivators of human behaviour. However, the neural mechanisms determining levels of sociality are largely unknown, primarily due to a lack of suitable animal models. This study reports the presence of a surprising degree of general sociality in Drosophila. A newly-developed paradigm to study social approach behaviour in flies reveal that social cues perceive through both vision and olfaction converged in a central brain region, the γ lobe of the mushroom body, which exhibit activation in response to social experience. The activity of these γ neurons control the motivational drive for social interaction. At the molecular level, the serotonergic system is critical for social affinity. These results demonstrate that Drosophila are highly sociable, providing a suitable model system for elucidating the mechanisms underlying the motivation for sociality (Sun, 2020).

Go to top

Predictive olfactory learning in Drosophila

Zhao, C., Widmer, Y. F., Diegelmann, S., Petrovici, M. A., Sprecher, S. G. and Senn, W. (2021). Sci Rep 11(1): 6795. PubMed ID: 33762640

Olfactory learning and conditioning in the fruit fly is typically modelled by correlation-based associative synaptic plasticity. It was shown that the conditioning of an odor-evoked response by a shock depends on the connections from Kenyon cells (KC) to mushroom body output neurons (MBONs). Although on the behavioral level conditioning is recognized to be predictive, it remains unclear how MBONs form predictions of aversive or appetitive values (valences) of odors on the circuit level. This study presents behavioral experiments that are not well explained by associative plasticity between conditioned and unconditioned stimuli, and two alternative models are suggested for how predictions can be formed. In error-driven predictive plasticity, dopaminergic neurons (DANs) represent the error between the predictive odor value and the shock strength. In target-driven predictive plasticity, the DANs represent the target for the predictive MBON activity. Predictive plasticity in KC-to-MBON synapses can also explain trace-conditioning, the valence-dependent sign switch in plasticity, and the observed novelty-familiarity representation. The model offers a framework to dissect MBON circuits and interpret DAN activity during olfactory learning (Zhao, 2021).

Predicting the future from sensory input is fundamental for survival. Co-appearing stimuli can be used for improving a prediction, or for predicting important events themselves, as observed in classical conditioning. In fruit fly odor conditioning, an odor that will become the conditioned stimulus (CS), is paired with the unconditioned stimulus (US), in this study an electroshock, that triggers an avoidance behavior and in internal representation of a negative value (valence). After conditioning, and the negative value representation (although not the full unconditioned response) previously elicited by the electroshock will be reproduced by the odor itself. Classical conditioning theories posit that throughout learning the odor becomes predictive for the electroshock. During learning, the prediction error decreases, and learning stops when the predictive odor value matches the strength of the electroshock (Zhao, 2021).

Predictive olfactory learning in fruit flies is a widely recognised concept in the experimental literature, and dopaminergic neurons (DANs) in the mushroom body (MB) have been suggested to predict punishment or reward. Yet, despite the acknowledgment of its predictive nature, computational models on fruit fly conditioning are mostly guided by the formation of associations, a notion that relates more to memories rather than predictions. Similarly, the concept of predictive learning is well recognized for olfactory conditioning in insects in general, but synaptic plasticity models are not formulated in terms of explicit predictions, but rather in terms associations and correlations, with plasticity being driven by two or three factors, each representing a temporal nonlinear function of the pre- or postsynaptic activities or of a modulatory signal, sometimes combined with homeostatic plasticity. This type of associative models exist for fruit flies, locusts or honey bees. They differ from target learning, where the unconditioned stimulus sets a target that is learned to be reproduced by the conditioned stimulus. Target learning becomes predictive learning when including a temporal component. It involves a difference operation, and learning stops when the target is reached. The stop-learning feature is difficult to be reproduced by purely correlation-based associative learning, while a purely predictive model also intrinsically captures associative properties (Zhao, 2021).

Associative learning was suggested to be implemented through spike- or stimulus-timing dependent plasticity (STDP) that would underlie conditioning. STDP strengthens or weakens a synapse based on the temporal correlation between the US (electroshock) and the CS (odor), both on the neuronal time scale of 10s of milliseconds and on the behavioral time scale of 10s of seconds. Whether an association is strengthened by just repeating the pairing until the behavioral saturation is reached, or the association saturates due to a faithful prediction, however, has not been investigated in the fruit fly so far. This study shows that olfactory conditioning in Drosophila is better captured as predictive plasticity that stops when a US-imposed target is reached, rather than by correlation-based plasticity, such as STDP, that does not operate with an explicit error or a target. According to this scheme, it is only the aversive/appetitive value of the US that is predicted by the CS after faithful learning, not the US itself. Based on the common value representation in the mushroom body output neurons (MBONs), the corresponding avoidance/approach reaction as one aspect of the unconditioned response is elicited by the CS alone (Zhao, 2021).

The Drosophila olfactory system represents a unique case for studying associative /predictive learning, and the MB is known to be essential in olfactory learning. The Kenyon cells (KCs) receive olfactory input from olfactory projection neurons and form a sparse representation of an odor. The parallel axons of the Kenyon cells (KCs) project to the MB lobes, along which the compartmentalized dendritic arbors of the MB output neurons (MBONs) collect the input from a large number of KCs. Reward or punishment activates specific clusters of DANs PAM and PPL1, respectively which project to corresponding compartments of the MB lobes, modulating the activity of the MBONs and the behavioral response. Recently, a detailed mapping of the MB connectome has been accomplished for larvae and of the vertical lobe for the adult Drosophila. Several studies show that not only the feedforward modulation from DANs to MBONs, but also the feedback from MBONs to DANs play an important role in olfactory learning (Zhao, 2021).

Previous studies have given insights into the possible cellular and subcellular mechanisms of olfactory conditioning. Yet, the suggested learning rules remain correlation-based and miss the explicit predictive element postulated by the classical conditioning theories. This study presents distinctive conditioning experiments showing that olfactory learning is best explained by predictive plasticity. These experiments, in contrast, could not be reproduced by various types of correlation-based associative learning rules. A mathematical model captures the new and previous data on olfactory conditioning, including trace conditioning. The model encompasses the odor/shock encoding and the learning of the aversive odor value with the stochastic response. It is further suggested how the predictive plasticity could be implemented in the MB circuit, with MBONs encoding the value ('valence') of the odor stimulus, and DANs calculating either the error or the target that drives the KC-to-MBON plasticity. The predictive plasticity rule for the KC-to-MBON synapses is shown to be consistent with the experimental results showing the involvement of these synapses in the novelty-familiarity representation (Zhao, 2021).

This study has reconsidered classical odor conditioning in the fruit fly and presents experimental and modeling evidence showing that olfactory learning, also on the synaptic level, is better described as predictive rather than associative. The key observation is that repetitive and time-continuous odor-shock pairing stops strengthening the conditioned response after roughly 1 minute of pairing, even if the shock intensity is below the behavioral saturation level. During conditioning, the odor is learned to predict the co-applied shock stimulus. As a consequence, the odor-evoked avoidance reaction stops strengthening at a level that depends on the shock strength, irrespective of the pairing time beyond 1 min. Associative synaptic plasticity, defined by a possibly nonlinear function of the CS-US correlation strength, as suggested by STDP models, fails to reproduce the early saturation of learning (Zhao, 2021).

A simple phenomenological model for predictive plasticity is suggested according to which synapses change their strength proportionally to the prediction error. This error is expressed as a difference between the internal shock representation and the value representation of the odor. The model encompasses a description of the shock and value representation, the stochastic response behavior of individual flies, and the synaptic dynamics (using a total of 5 parameters). It faithfully reproduces the conditioning experiments (with a total of 28 data points from 3 different types of experiments) as well as previously studied trace conditioning experiments (without need for further fitting). As compared to the associative rules (Hebbian, linear and nonlinear STDP, covariance rule), the predictive plasticity rule obtained the best fits with the least number of parameters. The model was further compared by the Akaike information criterion that considers the number of parameters beside the fitting quality. This criterion yields a likelihood for the predictive plasticity rule to be the best one that is at least 7 orders of magnitude larger as compared to the other four associative rules considered (Zhao, 2021).

The same phenomenological model of predictive learning may be implemented in two versions by the recurrent MB circuitry. In both versions the MBONs code for the odor value ('valence') that drives the conditioned response. For the error-driven predictive plasticity, the DANs directly represent the shock-prediction error by comparing the shock strength with its MBON estimate, and this prediction error modulates the KC-to-MBON plasticity (see Suggested implementation of error- and target-driven predictive plasticity). For the target-driven predictive plasticity, the DANs represent the shock stimulus itself that is then provided as a target for the KC-to-MBON plasticity. In this target-driven predictive learning, the DANs may also learn to predict the shock stimulus based on the MBON feedback, preventing a fast extinction of the KC-to-MBON memory (Zhao, 2021).

Predictive plasticity for both types of implementation has its experimental support. In general, MBON activity is well recognized to encode the aversive or appetitive value of odors and to evoke the corresponding avoidance or approach behavior, while KC-to-MBON synapses were mostly shown to undergo long-term depression, but also potentiation. DAN responses are shown to be involved in both the representation of punishment and reward that drive the aversive or appetitive olfactory conditioning. This conditioning further involves the recurrent feedback from MBONs to DANs that may be negative or positive. Moreover, the connectome from the larvae and adult fruit fly MBON circuit reveals feedback projections from DANs to the presynaptic side on the KC and the postsynaptic side on the MBONs at the KC-to-MBON synaptic connection giving different handles to modulate synaptic plasticity (Zhao, 2021).

With regard to the specific implementations, the error-driven predictive plasticity is consistent with the observation that DAN activity decreases during the conditioning. The two models have opposite predictions for learning while blocking MBON activity. The error-driven predictive plasticity would yield a higher learning index (LI) while the target-driven predictive plasticity would yield a lower LI. It was also shown that some DANs increased their activity with learning while other DANs, in the same PPL1 cluster that is supposed to represent aversive valences, decreased their activity. In fact, error- and target-driven predictive plasticity may both act in concert to enrich and stabilize the representations. DAN activity would decrease in those DANs involved in error-driven predictive plasticityand increase in those involved in target-driven predictive plasticity (Zhao, 2021).

While error-driven predictive plasticity offers access to an explicit error representation in DANs, target-driven predictive plasticity has its own merits. If DANs and MBONs code for similar information, they can support a positive feedback-loop to represent a short-term memory beyond the presence of an odor or a shock, as it was observed for aversive valences in PPL1 DANs and for appetitive PAM DANs. A positive feedback-loop between MBONs and DANs is further supported by the persistent firing between these cells after a rejected courtship that may consolidate memory of the rejection, linked to a specific pheromone (Zhao, 2021).

Target-driven plasticity has further functional advantages in terms of memory retention time. Any odor-related input to the DANs, arising either through a forward hierarchy from KC48 or a recurrence via MBONs to the DANs, will extend the memory life-time in a 2-stage prediction process: the unconditioned stimulus (s) that drives the DAN activity (d) to serve as a target for the value learning in the MBONs via KC-to-MBON synapses, will itself be predicted in the DANs. Extending the memory life-time through circuit plasticity might be attractive under the light of energy efficiency, showing that long-term memory in a synapse involving de novo protein synthesis can be costly, while cheaper forms of individual synaptic memories likely have limited retention times. Moreover, distributed memory that includes the learning of an external target representation offers more flexibility, including the regulation of the speed of forgetting (Zhao, 2021).

Target-driven predictive plasticity may also explain the novelty-familiarity representation observed in the recurrent triple of KCs, DANs and MBONs. The distributed representation of valences allows for expressing temporal components of the memories. Spontaneous activity in the KCs and their downstream cells injures a minimal strength of the KC-to-MBON synapses through predictive plasticity. A novel odor that drives KCs will then also drive MBONs and, to a smaller extent (as is assumed in this study), also DANs. If the DANs that represent the target for the KC-to-MBON plasticity are only weakly activated by the odor, the KC-to-MBON synapses learn to predict this weaker activity and depress. The depression results in a repetition suppression of MBONs and the corresponding familiarization of the fly to the ongoing odor. However, when the odor is cleared away, the MBON activity induced by spontaneously active KCs via depressed synapses now becomes lower than the spontaneous DAN activity, and predictive plasticity recovers the original synaptic strength. Eventually the spontaneous MBON and DAN activites match again and the response to the originally novel odor is also recovered, as seen in the experiment (Zhao, 2021).

Olfactory learning is likely distributed across several classes of synapses in the MB. The acquisition of olfactory memories was shown to be independent of transmitter release in KC-to-MBON synapses, although the behavioral recall of these memories required the intact transmission. In fact, learning may also be supported by plasticity upstream of the MBONs such that the effect of blocking KC-to-MBON transmission during learning is behaviorally compensated. Predictive plasticity at the KC-to-MBON synapses requires the summed synaptic transmissions across all synapses to be compared with the target d, also during the memory acquisition. This type of plasticity would therefore be impaired by blocking the release (Zhao, 2021).

Distributed learning also offers flexibility in acquiring predictions from new cues. While the original Rescorla-Wagner rule would predict blocking, this has not been observed in the fruit fly. Blocking refers to the phenomenon that, if the first odor of a compound-CS is pre-conditioned, the second odor of the compound will not be learned to become predictive for the shock. Because the predictive plasticity rules are expressed at the neuronal but not at the phenomenological level, predictions about blocking will depend on the neuronal odor representation. If the two odors activate the same MBONs, blocking would be observed since the MBONs are already driven to the correct value representation by the first odor. If they activate different MBONs, however, blocking would not be observed since the MBONs of the second odor did not yet have the chance to learn the correct value during the first conditioning. Hence, since blocking has not been observed in the fruit fly, it is postulated that the odors of the compound-CS in these experiments were represented by different groups of MBONs (Zhao, 2021).

How does this model relate to the concentration-specificity and the timing-specifity of odor conditioning? First, olfactory learning was found to be specific to the odor concentration, with different concentrations changing the subjective odor identity. The response behavior was described to be non-monotonic in the odor intensity, with the strongest response for the specific concentration the flies were conditioned with. It was suggested that this may arise from a non-monotonic odor representation in the KC population as a function of odor intensity. Given such a presynaptic encoding of odor concentrations, the predictive olfactory learning in the KC-to-MBON connectivity would also inherit the concentration specificity from the odor representation in the KCs. This predictive plasticity, and also the Rescorla-Wagner model, further predicts that learning with a higher odor concentration (but the same electroshock strength) only speeds up learning, but would not change the asymptotic performance (Zhao, 2021).

Second, olfactory conditioning was also shown to depend on the timing of the shock application before or after the conditioning odor. While a shock application 30s after an odor assigns this odor an aversive valence, an appetitive valence is assigned if the shock application arises 30s before the odor presentation. Modeling the approaching behavior in the context of predictive plasticity would require duplicating this model to also represent appetitive valences, and the action selection would depend on the difference between aversive and appetitive valences. Inverting the timing of CS and US may explain 'relief learning' if a stopping electroshock would cause a decrease of the target for aversive MBONs and an increase of the target for appetitive MBONs. An odor presented after the shock would then predict the increased appetitive target and explain the relief from pain behavior, similar to the model of relief learning in humans (Zhao, 2021).

Overall, these behavioral experiments and the plasticity model for the KC-to-MBON synapses support the notion of predictive learning in olfactory conditioning, with the DANs representing either the CS-US prediction error or the prediction itself. While predictive coding is recognized as a hierarchical organization principle in the mammalian cortex that explains animal and human behavior it may also offer a framework to investigate the logic of the MB and the multi-layer MBON readout network as studied by various experimental work (Zhao, 2021).

Go to top

A pair of dopamine neurons mediate chronic stress signals to induce learning deficit in Drosophila melanogaster

Jia, J., He, L., Yang, J., Shuai, Y., Yang, J., Wu, Y., Liu, X., Chen, T., Wang, G., Wang, X., Song, X., Ding, Z., Zhu, Y., Zhang, L., Chen, P. and Qin, H. (2021). Proc Natl Acad Sci U S A 118(42). PubMed ID: 34654742

Chronic stress could induce severe cognitive impairments. Despite extensive investigations in mammalian models, the underlying mechanisms remain obscure. This study shows that chronic stress could induce dramatic learning and memory deficits in Drosophila melanogaster The chronic stress-induced learning deficit (CSLD) is long lasting and associated with other depression-like behaviors. This study demonstrates that excessive dopaminergic activity provokes susceptibility to CSLD. Remarkably, a pair of PPL1-γ1pedc dopaminergic neurons that project to the mushroom body (MB) γ1pedc compartment play a key role in regulating susceptibility to CSLD so that stress-induced PPL1-γ1pedc hyperactivity facilitates the development of CSLD. Consistently, the mushroom body output neurons (MBON) of the γ1pedc compartment, MBON-γ1pedc>&alpha/β neurons, are important for modulating susceptibility to CSLD. Imaging studies showed that dopaminergic activity is necessary to provoke the development of chronic stress-induced maladaptations in the MB network. The data supports PPL1-γ1pedc mediates chronic stress signals to drive allostatic maladaptations in the MB network that lead to CSLD (Jia, 2021).

Go to top

Innate and learned odor-guided behaviors utilize distinct molecular signaling pathways in a shared dopaminergic circuit

Noyes, N. C. and Davis, R. L. (2023). Cell Rep 42(2): 112026. PubMed ID: 36701232

Odor-based learning and innate odor-driven behavior have been hypothesized to require separate neuronal circuitry. Contrary to this notion, innate behavior and olfactory learning were recently shown to share circuitry that includes the Drosophila mushroom body (MB). But how a single circuit drives two discrete behaviors remains unknown. This study defines an MB circuit responsible for both olfactory learning and innate odor avoidance and the distinct dDA1 dopamine receptor-dependent signaling pathways that mediate these behaviors. Associative learning and learning-induced MB plasticity require rutabaga-encoded adenylyl cyclase activity in the MB. In contrast, innate odor preferences driven by naive MB neurotransmission are rutabaga independent, requiring the adenylyl cyclase ACXD. Both learning and innate odor preferences converge on PKA and the downstream MBON-γ2α'1. Importantly, the utilization of this shared circuitry for innate behavior only becomes apparent with hunger, indicating that hardwired innate behavior becomes more flexible during states of stress (Noyes, 2023).

These data reveal the shared use of a discrete circuit for both state-dependent odor-driven behavior and experience-dependent odor learning. The shared components include the upstream DA neurons, the MBn-expressed DA receptor dDA1, and the downstream MBON. Odor response processing for state-dependent behavior and odor learning diverge at the level of the dDA1 receptor-activated adenylyl cyclase, with ACXD employed for innate state-dependent odor driven behavior and rut employed for olfactory learning. The unique activation of rut for olfactory learning is explained by the fact that this adenylyl cyclase functions as a coincidence detector, synergistically responding to both DA receptor activation from the unconditioned stimulus and Ca2+ increases due to the conditioned stimulus (Noyes, 2023).

ACXD is a transmembrane AC that is expressed in a number of tissues including the brain (flyatlas.org) and is orthologous to the mammalian AC2. Mammalian AC2 activity is Ca2+ independent. If ACXD is also Ca2+ independent, it would provide a mechanism for the engagement of distinct cAMP pathways by dDA1 for state-dependent versus experience-dependent olfactory behavior. Thus, common neural circuitry is employed for both state-dependent and conditioned behaviors with the unique changes of MBn output influenced by the intracellular signaling pathways that are mobilized. Dopaminergic input to the dDA1 DA receptor expressed in the γ2 compartment of MBn activates an intracellular signaling pathway that includes the ACXD adenylyl cyclase, PKA activity, and the release of ACh. The downstream MBON-γ2α'1 responds to the MB ACh release through the α2 nACh receptor, with the activity of the MBON-γ2α'1 ultimately dictating the balance in state-dependent odor approach/avoidance. The simplest model to account for the state-dependent MBON activity would have the internal state modulating DA input into to the MBn to increase or decrease ACh release onto the MBON. However, the current data failed to detect a significant change in ACh release between the fed and starved conditions. Nevertheless, the activity of the PPL1-γ2α'1 that influences the MB γ2 compartment is required for state-dependent behavioral responses to odor. A proposed model for reconciling these observations envisions that the basal activity of this circuit is required for behavioral-state odor choice but that starvation mobilizes a qualitative or quantitative signal independent of the magnitude ACh release by the MBn to increase MBON activity. An unidentified signal representing hunger could directly enhance MBON excitability. For example, octopamine has been proposed as a feeding signal that acts directly on MBONs (Noyes, 2023).

Similarly, a hunger signal could act on neurons elsewhere in the brain that ultimately connect to MBONs through intermediary neurons. The hunger-responsive neuropeptide leucokinin acts on DAns that connect to MBONs (Noyes, 2023).

Alternatively, there may be a co-neurotransmitter released by the MBn due to starvation that works to increase MBON activity. Finally, the possibility that starvation does modulate MBn ACh output is left open, since the reporters employed lack the sensitivity to detect this change. Future investigations into state-dependent changes in MBON-γ2α'1 physiology will need to grapple with numerous competing hypotheses (Noyes, 2023).

Changes in odor responses in MBONs have suggested that learning induces a change in connectivity in the MBn-MBON synapse. In addition, compartment specific plasticity in MB ACh release was discovered that fits with the idea that plasticity observed in MBONs occurs from the input of MB compartments (Noyes, 2023).

However, there has been a lack of data connecting the MBn ACh release plasticity with MBON plasticity and particularly to the central role for the rut adenylyl cyclase. The current data offer this important connection. It shows that the MB γ2 compartment undergoes a rapid depression in response to odor/shock pairings during aversive learning and that rut is required for the acquisition of this depression. Downstream of the MB γ2 compartment, MBON-γ2α'1 drives approach and also undergoes a learning-induced depression (Noyes, 2023).

These results put prior speculation about how the genetic regulation of cAMP signaling through the rut adenylyl cyclase drives Drosophila memory on concrete ground. The work does not conclusively delineate a role for dDA1 in the MB plasticity. Loss of MB dDA1 dramatically reduced naive odor responses in MB γ2. This precluded attempts to measure dDA1 effects on MB γ2 depression because the naive responses were already low. Interestingly, both learning- and starvation-dependent odor avoidance require PKA. A likely explanation is that rut and ACXD are spatially segregated, creating distinct cAMP microdomains or signaling platforms. Thus, PKA activity would result in the phosphorylation of unique substrates within those microdomains (Noyes, 2023).

The characterization of the MB as a brain region for learned, but not innate, olfactory behavior was motivated by experiments eliminating MBn or blocking MB output. Disrupting the MB eliminates odor-associated memory but has no effect on innate avoidance of those same odors. Recent work has overturned this simple categorization demonstrating some DANs, MBONs, and MBns do contribute to innate olfactory behavior in certain circumstances. Interestingly, the majority of reports define a role for MBns in innate behavioral responses to food-related odors (Noyes, 2023).

The results, using more general, non-food odors, puts the hypothesis that MBn regulates innate behavior on more solid ground. Importantly, it was found that MB dDA1 is required for state-dependent behavior to general odors. This is in contrast to a report finding that dDA1 is not involved but that DAMB is required for state-dependent behavior to food odors (Noyes, 2023).

This difference will be a key element to understand state-dependent behavior moving forward. The Drosophila and mammalian olfactory systems are remarkably similar in terms of anatomical organization and function. In both, odorant molecules activate olfactory sensory neurons (OSNs), with each OSN only expressing one type of odorant receptor (OR). Each OSN expressing the same type of OR project to the same glomeruli (Noyes, 2023).

Within the glomeruli, the OSNs synapse onto projection neuron (PN) dendrites, and PN activity is modified in the glomeruli by local inhibitory interneurons before being sent on to multiple higher-order brain regions. PN neurons connect to downstream neurons in the mammalian piriform cortex and in the Drosophila MB in a seemingly random manner. Like the Drosophila MB, the piriform cortex is critical for olfactory memory. It is not clear how the piriform cortex is involved in state-dependent olfactory behavior. However, in humans, odor coding changes in the piriform cortex with hunger and sleep deprivation, and piriform cortex neuron activity levels are inversely correlated with sexual satiety in rats (Noyes, 2023).

We conclude from these results demonstrating dDA1-dependent MBn Ach release and a dDA1-dependent MBON-γ2α'1 Ca2+ in response to odor that dDA1 directly modulates the MBn/MBON connection. However, due to a limitation in the sensitivity of the ACh sensor employed, it was not possible to directly record ACh input to MBON-γ2α'1. Based on the established direct cholinergic connection between these MBn and MBONs and the lack of any known non-MB cholinergic innervation to this brain region, it is believed that the conclusions are merited. However, a formal possibility that other intermediary neurons mediate this relationship must be left open (Noyes, 2023).

Go to top

Context-dependent representations of movement in Drosophila dopaminergic reinforcement pathways

Zolin, A., Cohn, R., Pang, R., Siliciano, A. F., Fairhall, A. L. and Ruta, V. (2021). Nat Neurosci 24(11): 1555-1566. PubMed ID: 34697455

Dopamine plays a central role in motivating and modifying behavior, serving to invigorate current behavioral performance and guide future actions through learning. This study examined how this single neuromodulator can contribute to such diverse forms of behavioral modulation. By recording from the dopaminergic reinforcement pathways of the Drosophila mushroom body during active odor navigation, this study reveals how their ongoing motor-associated activity relates to goal-directed behavior. Dopaminergic neurons were found to correlate with different behavioral variables depending on the specific navigational strategy of an animal, such that the activity of these neurons preferentially reflects the actions most relevant to odor pursuit. Furthermore, this study shows that these motor correlates are translated to ongoing dopamine release, and acutely perturbing dopaminergic signaling alters the strength of odor tracking. Context-dependent representations of movement and reinforcement cues are thus multiplexed within the mushroom body dopaminergic pathways, enabling them to coordinately influence both ongoing and future behavior (Zolin, 2021).

Go to top

Sleep deprivation results in diverse patterns of synaptic scaling across the Drosophila mushroom bodies

Weiss, J. T. and Donlea, J. M. (2021). Curr Biol. PubMed ID: 34107302

The mushroom bodies of Drosophila contain circuitry compatible with race models of perceptual choice. When flies discriminate odor intensity differences, opponent pools of αβ core Kenyon cells (on and off αβ(c) KCs) accumulate evidence for increases or decreases in odor concentration. These sensory neurons and "antineurons" connect to a layer of mushroom body output neurons (MBONs) which bias behavioral intent in opposite ways. All-to-all connectivity between the competing integrators and their MBON partners allows for correct and erroneous decisions; dopaminergic reinforcement sets choice probabilities via reciprocal changes to the efficacies of on and off KC synapses; and pooled inhibition between αβc KCs can establish equivalence with the drift-diffusion formalism known to describe behavioral performance. The response competition network gives tangible form to many features envisioned in theoretical models of mammalian decision making, but it differs from these models in one respect: the principal variables-the fill levels of the integrators and the strength of inhibition between them-are represented by graded potentials rather than spikes. In pursuit of similar computational goals, a small brain may thus prioritize the large information capacity of analog signals over the robustness and temporal processing span of pulsatile codes (Vrontou, 2021).

Two-alternative forced-choice tasks, in which a subject must commit to one of two alternatives, sometimes under time pressure and nearly always with uncertain information, are a commonly studied laboratory simplification of real-world decision making. The neural processes that culminate in a binary choice have been compared to the deliberations of a jury before a verdict: neurons, like jurors, gather evidence from witnesses over the course of a trial and then reconcile their divergent views in a majority vote (Vrontou, 2021).

The problem of how neural circuits implement this form of trial by jury has been approached in a range of species, from primates and rodents to fish and flies. A pioneering and influential body of work is built on a two-alternative forced-choice task in which monkeys distinguish directions of motion in a noisy random dot display. Recordings of correlated neuronal activity suggest that motion-sensitive neurons in the middle temporal visual area (MT or V5) provide momentary evidence that is temporally integrated in lateral intraparietal cortex (LIP) before passing an unspecified thresholding mechanism. Although the precise role attributed to LIP is a matter of debate, the principle that ephemeral sensory signals flow into integrators whose fill levels rise to a response threshold appears general; similar arrangements have been inferred to support visual motion discrimination in zebrafish and odor intensity discrimination in the fly (Vrontou, 2021).

In Drosophila, a rate-limiting integration step takes place in a particular group of third-order olfactory neurons. When flies decide on the direction of an odor concentration change, the membrane potentials of Kenyon cells (KCs) in the αβ core (αβc) division of the mushroom bodies drift noisily toward action potential threshold, just as accumulating evidence would drift toward a response bound. Consistent with the proposed correspondence of membrane voltage and integrated sensory information, and of action potential and decision thresholds, neurometric functions based on the average timing of the first odor-evoked spikes in the αβc KC population can account for the speed and accuracy of the decision-making animal; psychophysical estimates of noise in the decision process match the measured membrane potential noise of αβc KCs; and genetically targeted manipulations that alter the latencies of αβc KC spikes have the expected impact on reaction times (Vrontou, 2021).

Two functionally separate groups of αβc KCs, termed up and down or on and off cells, respond to increases or decreases in odor concentration and can therefore represent the strength of evidence for either of the two alternatives in the choice. This explicit representation of support for each of the competing hypotheses (as opposed to an aggregate representation of the extent to which one hypothesis is favored over the other) suggests that a decision involves a race between two integrators-one built from neurons that accumulate evidence for an increase in odor concentration and another composed of 'antineurons' that do the opposite. Changes in odorant receptor occupancy at the periphery alter the baseline activity of olfactory receptor neurons and the second-order projection neurons (PNs) with which they form receptor-specific glomerular channels. Large odor concentration changes in a channel's preferred direction drive high-frequency transmission from PNs to αβc KCs that promotes steep depolarizations to spike threshold and fast, accurate decisions, whereas small concentration changes in the preferred direction, or any change in the null direction, cause only a trickle of synaptic release; shallow, undulating membrane potential rises; and long spike delays that lead to slow, error-prone choices (Vrontou, 2021).

This study examined whether the circuitry downstream of αβc KCs is compatible with a model of two competing integrators. Three predictions of such a model were tested. First, to adjudicate the rival hypotheses advocated by on and off αβc KCs, mushroom body output neurons (MBONs) sampling the cores of the αβ lobes must listen to both. It is therefore expected that each core-innervating MBON is excited by increases as well as decreases in odor concentration. Second, as an animal learns the rules of the two-alternative forced-choice task-that an increase in odor intensity predicts imminent electric shock, whereas a decrease signals protection-the influence of αβc KCs championing the correct choice should be enhanced while that of proponents of the incorrect choice should be diminished. In other words, antagonistic changes are expected in the strengths of connections of on and off αβc KCs with the same action selection neurons if evidence for the competing alternatives is accumulated separately. Third, race models become equivalent to a drift-diffusion process-the formalism shown accurately to describe the psychophysics of the decision-making animal-only if they include an element of mutual or pooled inhibition to establish response competition between the integrators. Inhibition is needed to ensure that the integrators are anti-correlated so that evidence for one choice simultaneously counts as evidence against the other. This study therefore predicts the existence of inhibitory interactions between αβc KCs (Vrontou, 2021).

The idea that decisions are based on the accumulated spikes of oppositely tuned sensors was born in early attempts to unite psychophysical and neurophysiological measurements under the umbrella of signal detection theory. The recorded spike count distributions of direction-selective units in the monkey's area MT to motion in the preferred or null directions were taken to represent the responses of two neurons-the recorded neuron and its imagined antineuron conjugate-to movement in the neuron's preferred direction. The likely direction of motion can then be inferred as the probability that a draw from the neuron's response distribution yields a larger spike count than a draw from the antineuron's. At minimal motion strengths, when the two distributions are congruent, these odds are even and choices are random, but as the neuron responds ever more vigorously to increasingly coherent motion while the antineuron's response stays flat, the distributions unmix and the probability of a correct choice rises toward one. Comparing the spike counts of two sensors rather than thresholding the output of one removes shared sources of variation and with them the need of adjusting the discrimination threshold to achieve the best separation of the changing response distributions: a neuron-antineuron pair always returns a quantity proportional to likelihood ratio, the optimal hypothesis test. Although opponent sensory channels in one or another guise feature prominently in many decision-making models, their involvement in the brain is unproven: neurons and antineurons owe their status to each other, as inputs to comparator circuits, but these circuits remain uncharacterized (Vrontou, 2021).

This study draws back the curtain on one such circuit in the fly. Changes in odor intensity are registered by pools of on and off αβc KCs, which represent the strengths of the accumulated evidence for an increase or decrease in odor concentration. These pools of sensory neurons and antineurons couple to a second layer of neurons and antineurons, the core-innervating MBONs, which bias behavioral intent in opposite ways. Members of both neuronal pools in the sensory layer connect to both types of MBON in the action selection layer via plastic synapses. With two sets of neurons and antineurons and all-to-all feedforward connectivity between them, the comparator circuit allows for approach or avoidance following judgments of upward or downward changes in odor intensity-that is, it comprises neural pathways representing the possible contingencies seen behaviorally. The perceptual decision is won-correctly or incorrectly-by the αβc KC pool that reaches spike threshold first, and it is expressed in the behavior instructed by that pool's favored MBON partners (Vrontou, 2021).

Unlike neurons comprising the ON and OFF pathways of motion vision, on and off αβc KCs cannot be distinguished and manipulated genetically. This study has therefore exploited the sensitivity of KC-to-MBON synapses to the timing of reinforcement to reveal the convergence of separate on and off channels onto the same MBONs. KC-to-MBON synapses in their ground state exert finely balanced drive on the MBON ensemble, so that votes cast by its members cancel one another as in a hung jury, but experience can shift the synaptic weight distribution and the resulting pattern of MBON activation away from net zero. This study has documented such shifts for the approach-advocating MBON-γ1pedc>αβ: pairing odor on- or offset with electric shock weakens transmission from the on αβc KC pool and strengthens transmission from the off αβc KC pool (or vice versa), synergistically changing odor preference. The underlying mechanism is a switch from synaptic depression to synaptic potentiation when the order of odor-evoked KC activation and dopaminergic reinforcement is reversed. This mechanism operates at KC connections with all core-innervating MBONs but is likely engaged at different timescales that may reflect sequential memory phases; to demonstrate the mechanism's ubiquity, this temporal sequence was artificially collapsed by photostimulating DANs directly. Within the short time frame of these behavioral experiments, only PPL1-γ1pedc, but not PPL1-α'2α2, shows significant pain responses that modulate its sensitivity to a predictive odor, consistent with the view that PPL1-γ1pedc and its cognate MBON-γ1pedc>αβ represent the core circuit for the storage and expression of short-term aversive memories (Vrontou, 2021).

A crucial element of many neural network models of decision making is inhibitory feedback from a common interneuron pool driven by the competing integrators, which helps to amplify small differences in conflicting sensory evidence until, eventually, one integrator prevails. The response competition circuit this study has delineated contains such an inhibitory element but with the intriguing twist that the key variables are represented by membrane voltages rather than spikes. Analog processing may be a consequence of numerical constraints: if the mushroom bodies lack the neuron numbers needed to approximate continuous quantities with discrete-time action potentials, there may be little choice but to swap the advantages regenerative spikes could provide (such as long time windows for adding and retaining sensory evidence) for the greater information capacity of graded potentials. Perhaps more is different (Vrontou, 2021).

Go to top

A non-canonical on-demand dopaminergic transmission underlying olfactory aversive learning

Saitoe, M., Naganos, S., Miyashita, T., Matsuno, M. and Ueno, K. (2022). Neurosci Res 178: 1-9. PubMed ID: 34973292

Dopamine (DA) is involved in various brain functions including associative learning. However, it is unclear how a small number of DA neurons appropriately regulates various brain functions. DA neurons have a large number of release sites and release DA non-specifically to a large number of target neurons in the projection area in response to the activity of DA neurons. In contrast to this "broad transmission", recent studies in Drosophila ex vivo functional imaging studies have identified "on-demand transmission" that occurs independent on activity of DA neurons and releases DA specifically onto the target neurons that have produced carbon monoxide (CO) as a retrograde signal for DA release. Whereas broad transmission modulates the global function of the target area, on-demand transmission is suitable for modulating the function of specific circuits, neurons, or synapses. In Drosophila olfactory aversive conditioning, odor and shock information are associated in the brain region called mushroom body (MB) to form olfactory aversive memory. It has been suggested that DA neurons projecting to the MB mediate the transmission of shock information and reinforcement simultaneously. However, the circuit model based on on-demand transmission proposes that transmission of shock information and reinforcement are mediated by distinct neural mechanisms; while shock transmission is glutamatergic, DA neurons mediates reinforcement. On-demand transmission provides mechanical insights into how DA neurons regulate various brain functions (Saitoe, 2022).

Go to top

Serotonin Signals Modulate Mushroom Body Output Neurons for Sustaining Water-Reward Long-Term Memory in Drosophila

Lee, W. P., Chiang, M. H., Chang, L. Y., Shyu, W. H., Chiu, T. H., Fu, T. F., Wu, T. and Wu, C. L. (2021). Front Cell Dev Biol 9: 755574. PubMed ID: 34858982

Memory consolidation is a time-dependent process through which an unstable learned experience is transformed into a stable long-term memory; however, the circuit and molecular mechanisms underlying this process are poorly understood. The Drosophila mushroom body (MB) is a huge brain neuropil that plays a crucial role in olfactory memory. The MB neurons can be generally classified into three subsets: γ, αβ, and α'β'. This study reports that water-reward long-term memory (wLTM) consolidation requires activity from α'β'-related mushroom body output neurons (MBONs) in a specific time window. wLTM consolidation requires neurotransmission in MBON-γ3β'1 during the 0-2 h period after training, and neurotransmission in MBON-α'2 is required during the 2-4 h period after training. Moreover, neurotransmission in MBON-α'1α'3 is required during the 0-4 h period after training. Intriguingly, blocking neurotransmission during consolidation or inhibiting serotonin biosynthesis in serotoninergic dorsal paired medial (DPM) neurons also disrupted the wLTM, suggesting that wLTM consolidation requires serotonin signals from DPM neurons. The GFP Reconstitution Across Synaptic Partners (GRASP) data showed the connectivity between DPM neurons and MBON-γ3β'1, MBON-α'2, and MBON-α'1α'3, and RNAi-mediated silencing of serotonin receptors in MBON-γ3β'1, MBON-α'2, or MBON-α'1α'3 disrupted wLTM. Taken together, these results suggest that serotonin released from DPM neurons modulates neuronal activity in MBON-γ3β'1, MBON-α'2, and MBON-&alpha'1&alpha'3 at specific time windows, which is critical for the consolidation of wLTM in Drosophila (Li, 2021).

Go to top

Fast-Scan Cyclic Voltammetry (FSCV) Reveals Behaviorally Evoked Dopamine Release by Sugar Feeding in the Adult Drosophila Mushroom Body

Shin, M. and Venton, B. J. (2022). Fast-Scan Cyclic Voltammetry (FSCV) Reveals Behaviorally Evoked Dopamine Release by Sugar Feeding in the Adult Drosophila Mushroom Body. Angew Chem Int Ed Engl 61(44): e202207399. PubMed ID: 35989453

Drosophila melanogaster, the fruit fly, is an excellent model organism for studying dopaminergic mechanisms and simple behaviors, but methods to measure dopamine during behavior are needed. This study developed fast-scan cyclic voltammetry (FSCV) to track in vivo dopamine during sugar feeding. First, acetylcholine stimulation was used to evaluate the feasibility of in vivo measurements in an awake fly. Next, sugar feeding was tested by placing sucrose solution near the fly proboscis. In the mushroom body medial tip, 1 pmol acetylcholine and sugar feeding released 0.49±0.04 μM and 0.31#177;0.06 μM dopamine, respectively but sugar-evoked release lasted longer than with acetylcholine. Administering the dopamine transporter inhibitor nisoxetine or D2 receptor antagonist flupentixol significantly increased sugar-evoked dopamine. This study develops FSCV to measure behaviorally evoked release in fly, enabling Drosophila studies of neurochemical control of reward, learning, and memory behaviors (Shin, 2022).

Go to top

Sleep deprivation results in diverse patterns of synaptic scaling across the Drosophila mushroom bodies

Weiss, J. T. and Donlea, J. M. (2021). Curr Biol. PubMed ID: 34107302

Sleep is essential for a variety of plastic processes, including learning and memory. However, the consequences of insufficient sleep on circuit connectivity remain poorly understood. To better appreciate the effects of sleep loss on synaptic connectivity across a memory-encoding circuit, changes were examined in the distribution of synaptic markers in the Drosophila mushroom body (MB). Protein-trap tags for active zone components indicate that recent sleep time is inversely correlated with Bruchpilot (BRP) abundance in the MB lobes; sleep loss elevates BRP while sleep induction reduces BRP across the MB. Overnight sleep deprivation also elevated levels of dSyd-1 and Cacophony, but not other pre-synaptic proteins. Cell-type-specific genetic reporters show that MB-intrinsic Kenyon cells (KCs) exhibit increased pre-synaptic BRP throughout the axonal lobes after sleep deprivation; similar increases were not detected in projections from large interneurons or dopaminergic neurons that innervate the MB. These results indicate that pre-synaptic plasticity in KCs is responsible for elevated levels of BRP in the MB lobes of sleep-deprived flies. Because KCs provide synaptic inputs to several classes of post-synaptic partners, a fluorescent reporter for synaptic contacts was used to test whether each class of KC output connections is scaled uniformly by sleep loss. The KC output synapses that were observed in this study can be divided into three classes: KCs to MB interneurons; KCs to dopaminergic neurons, and KCs to MB output neurons. No single class showed uniform scaling across each constituent member, indicating that different rules may govern plasticity during sleep loss across cell types (Weiss, 2021).

This study used genetic reporters to quantify the effects of sleep loss on pre-synaptic active zone markers and putative synaptic contacts in the Drosophila MB lobes. Abundance of Brp, dSyd-1, and Cacophony broadly increase across all MB lobes after overnight sleep deprivation and that acutely increasing sleep for 6 h is sufficient to reduce Brp levels across the α, 'β, γ, and 'β' lobes. KCs strongly contribute to the increase in Brp across each MB lobe following sleep loss, while pre-synapses of other MB cell types are less sensitive to sleep disruption. Because release of Drosophila neuromodulators likely occurs through a combination of classical neurotransmission and extrasynaptic release, these studies do not rule out the possibility that BRP-independent secretion of dopaminergic dense core vesicles might be altered in the MB lobes by sleep loss. The elevated levels of Brp present in KCs of sleep-deprived flies return to control levels within 48 h of ab libitum recovery sleep. While associative learning can recover within only a few hours after sleep deprivation, these studies indicate that some synaptic consequences of prolonged waking may persist for at least 24 h of recovery. These findings parallel those from humans and rodents, suggesting that some measures of cognition and neurophysiology recover rapidly after acute sleep loss while others last much longer, even for several days in some cases. The tractability of Drosophila may provide opportunities for future studies to investigate the processes that mediate recovery from sleep loss and to test whether similar trends in plasticity occur in other neuropil regions across the brain (Weiss, 2021).

Interestingly, sleep deprivation does not seem to increase other active zone components: Rim and Syt1 only show localized changes in some MB lobes, and the primarily post-synaptic marker Dlg shows no significant changes across the MB after sleep loss. Additionally, it was found that the abundance of vesicular proteins Rab3 and nSyb decreases across all MB lobes following overnight sleep deprivation. The varying responses between pre-synaptic components may indicate that sleep deprivation may alter the abundance of some active zone constituents along differing time courses or that active zone release machinery may be regulated differently from synaptic vesicle pools. The varied responses of each synaptic reporter that was observed suggests that Brp, dSyd-1, and Cac levels may underlie the consequences of sleep loss on MB functioning, but the precise physiological consequences of these changes on KC neurotransmitter release are unclear. Previous work finds that increasing BRP gene copy number drives changes in other active zone proteins that recapitulate protein levels observed in short sleeping mutants and also increases sleep in a dose-dependent manner (Weiss, 2021).

It is tempting to speculate that increases in Brp with sleep loss may drive concomitant increases in some core active zone scaffolding components and compensatory decreases in some proteins regulating synaptic vesicle release. Experiments at the Drosophila larval NMJ indicate that elevated Brp levels increase the rate of spontaneous release and enhance facilitation with pairs of stimuli, while other markers of synapse strength, including the amplitudes of evoked and spontaneous junction potentials, remained unchanged (Weiss, 2021).

It is unclear whether acute changes in Brp with sleep loss induce the same physiological changes at MB-output synapses, and additional studies will be required to understand how plastic mechanisms that contribute to memory formation might be altered by the pre-synaptic changes described above. Recent work finds that pan-neuronal knockdown of dSyd-1 can reduce sleep and dampen homeostatic rebound, even in flies with elevated BRP (Weiss, 2021).

Consistent with the idea that dSyd-1 levels may influence sleep pressure, decreased dSyd-1MI05387-GFSTF abundance was found in previously sleep-deprived flies after 48 h of recovery. While the MB contains several different cell types, pre-synapses in the axons of KCs appear to be uniquely plastic during sleep loss. Use of an activity-dependent fluorescent GRASP reporter of synaptic contacts observed that sleep loss altered synaptic contacts between KCs and distinct post-synaptic partners in different ways (Weiss, 2021).

Among these changes, it was found that GRASP fluorescence reporting contacts from KCs to PPL1 DANs is strongly decreased after sleep loss, indicating a weakening of the KC > PPL1 DAN contacts. Interestingly, these connections may be vital for recurrent activation within MB compartments during learning and could contribute to prediction error signals (Weiss, 2021).

While further studies will be required to examine the contribution of these particular connections to learning deficits after sleep loss, human subjects have been reported to exhibit impaired error prediction and affective evaluation in learning tasks following sleep loss (Weiss, 2021).

Because reduced GRASP signal was observed in KC > PPL1 DAN connections, which mediate aversive reinforcement, and not in KC > PAM DAN connections, which influence appetitive reinforcement, it is also possible that sleep loss may not equally degrade the encoding of reinforcement signals across all valences or modalities. Recent findings also suggest that not all forms of memory require sleep for consolidation; appetitive olfactory memories can be consolidated without sleep when flies are deprived of food, and sleep-dependent and independent memory traces in these conditions are stored in separate MB zones (Weiss, 2021).

The KC > MBON connections that contribute to sleep-dependent memory (KC > γ2α'1) also show an overall decrease in GRASP signal with sleep loss, while those that are vital for sleep-independent memory (MBON-γ1pedc) show no GRASP change after sleep deprivation. These compartment-specific variations in the effects of sleep on both memory and synaptic distribution further indicate that local MB zones may follow distinct plasticity rules under physiological stressors, including sleep loss (Weiss, 2021).

Additionally, GRASP signal from KCs to APL is significantly elevated following sleep loss, suggesting a strengthening of KC > APL connections. KCs and APL form a negative feedback circuit, where KCs activate APL and APL inhibits KCs: this feedback inhibition maintains sparseness of odor coding and odor specificity of memories (Weiss, 2021).

It is possible that KCs compensate for increased synaptic abundance accumulated during sleep loss by recruiting inhibition from APL. While further experimentation is needed to examine the role of these connections in the regulation of net synaptic strength during sleep loss, sleep deprivation results in increased cortical excitability in humans and rodents, and hyperexcitability is often counteracted by increased synaptic inhibition (Weiss, 2021).

Conversely, sleep loss reduces connectivity between KCs and DPM, a second large interneuron that may facilitate recurrent activity in the MB lobes. The current results also indicate that KC > MBON synaptic contacts exhibit a variety of changes in response to sleep deprivation. The specific KC > MBON connections that show significantly elevated or reduced GRASP signal in this study are not clearly assorted based on valence encoding, contribution to specific associative memory assays, or influence on sleep/wake regulation (Weiss, 2021).

Activity in several MB cell types, including α'/'β' KCs, MBON-γ5'β'2, MBON-γ2 α'1, DPM, and PAM DANs regulate sleep. The observation that KC > MBON-γ5'β'2a labeling is reduced with sleep loss complements previous observations of reduced electrical activity in MBON-γ5'β'2 following sleep deprivation (Weiss, 2021).

Other sleep-promoting MB neurons, however, such as DPM, do not show an overall increase in BRP abundance, suggesting either that other changes in excitability, synaptic drive, or post-synaptic adaptations might drive homeostatic sleep regulation in these cells or that distinct subsets of connections within the populations that were labelled in this study might be sleep regulatory. The compartment-to-compartment variance in KC > MBON responses to sleep loss also parallels previous findings that plasticity rules can vary between MBONs during heterosynaptic plasticity (Weiss, 2021).

While GRASP results suggest diverse changes in putative synaptic contacts with sleep loss, the functional effects of these changes require further study. It is important to note that a significant portion of MB synapses are composed of connections between either pairs or groups of KCs. The genetic strategies that were used in this study have prevented reliable visualization and quantification of these connections. As a result, the effect of sleep loss on KC > KC synapses has not been examined in this study but may comprise a portion of the increase in KC pre-synaptic abundance that was observed in this study. While these studies identify synaptic classes that exhibit altered GRASP labeling across sleep loss, future studies using super resolution imaging and/or physiology could examine the structural and molecular changes that underlie this plasticity. Connections between neurons in the MB may be also influenced by non-neuronal cell types, including astrocytes. Astrocytic contact with KCs can be reduced by sleep loss and astrocytic calcium levels correlate with sleep pressure, which suggests that astrocytic processes could be positioned to mediate sleep-dependent plasticity in the MB (Weiss, 2021).

The broad conservation of release machinery across active zones within and between cell types has simplified examination of pre-synaptic plasticity during sleep loss. Assays of both Hebbian and homeostatic plasticity have also identified a variety of post-synaptic adaptations. Interestingly, post-synaptic densities isolated from rodent cortex show significant reorganization of post-synaptic GluR5 receptors. This depends upon the activity of Homer and sleep-dependent phosphorylation of CaMKII and GluR1, that contribute to consolidation of visual cortex plasticity (Weiss, 2021).

Because MBONs exhibit post-synaptic plasticity during other contexts, including the formation of associative memories, sleep deprivation may also alter post-synaptic organization of MBONs or other cell types in the MB. Although the distribution of Dlg is not significantly changed by sleep loss, the rich variety of post-synaptic receptors for acetylcholine, dopamine, GABA, and other signals in the MB requires development of additional reporters to examine these post-synaptic consequences of insufficient sleep in MB neurons. Additionally, while the data outline changes in pre-synaptic protein abundance and pre-synaptic KC contacts that result from sleep loss, the possibility that these synaptic changes may be accompanied by homeostatic compensation in neuronal excitability or firing patterns remains to be tested. Because sleep-deprived flies can recover the capacity to learn after only a brief nap, homeostatic adjustments in post-synaptic strength and/or excitability may permit MBs to compensate for pre-synaptic changes that appear to persist for at least 24 h after sleep deprivation. Further, recovery sleep or pharmacological sleep enhancement may not simply reverse the effects of sleep loss, and it is unclear how particular subsets of synaptic proteins or connections may be selected for removal during times of elevated sleep (Weiss, 2021).

The consequences of sleep loss on synaptic connectivity are not clearly understood, but previous work has found net changes in synaptic abundance or size across brain regions. This study characterized a diverse array of synaptic responses to sleep loss among different cell types within the same circuit. These findings may suggest that distinct cell types and connections within the MB are governed by heterogeneous plasticity rules during sleep disruption. While previous studies have characterized the synaptic effects of sleep history on individual cell types within plastic circuits, the data provide a more comprehensive understanding of the consequences of sleep loss on MB circuits. While this project outlines the local effects of sleep loss on MB connectivity, it is unclear whether specific neural subsets also drive BRP increases within other neuropil compartments of sleep-deprived brains (Weiss, 2021).

This study found an overall increase in the abundance of reporters for some, but not all, pre-synaptic proteins. These pre-synaptic changes are not distributed equally across all cell types: they are most pronounced in MB-intrinsic KCs. Further, output connections from KCs to different classes of synaptic partners show varying patterns of plasticity in MB sub-circuits that contribute to encoding odor valence, comprise recurrent feedback loops, or relay reinforcement signals. The results indicate that sleep loss may degrade MB-dependent memory by altering several different classes of synapses, but future studies will be required to test the specific roles of changes at individual synapse types and the mechanisms by which prolonged waking reorganizes MB connectivity (Weiss, 2021).

Go to top

Neurofibromin 1 in mushroom body neurons mediates circadian wake drive through activating cAMP-PKA signaling

Almeida, M. P., Lago Solis, B., Stickley, L., Feidler, A. and Nagoshi, E. (2021). Nat Commun 12(1): 5758. PubMed ID: 34599173

Neurofibromin 1 in mushroom body neurons mediates circadian wake drive through activating cAMP-PKA signaling

Various behavioral and cognitive states exhibit circadian variations in animals across phyla including Drosophila melanogaster, in which only ~0.1% of the brain's neurons contain circadian clocks. Clock neurons transmit the timing information to a plethora of non-clock neurons via poorly understood mechanisms. This study addresses the molecular underpinning of this phenomenon by profiling circadian gene expression in non-clock neurons that constitute the mushroom body, the center of associative learning and sleep regulation. This study shows that circadian clocks drive rhythmic expression of hundreds of genes in mushroom body neurons, including the Neurofibromin 1 (Nf1) tumor suppressor gene and Pka-C1. Circadian clocks also drive calcium rhythms in mushroom body neurons via NF1-cAMP/PKA-C1 signaling, eliciting higher mushroom body activity during the day than at night, thereby promoting daytime wakefulness. These findings reveal the pervasive, non-cell-autonomous circadian regulation of gene expression in the brain and its role in sleep (Almeida, 2021).

Go to top

The lateral posterior clock neurons of Drosophila melanogaster express three neuropeptides and have multiple connections within the circadian clock network and beyond

Reinhard, N., Bertolini, E., Saito, A., Sekiguchi, M., Yoshii, T., Rieger, D. and Helfrich-Forster, C. (2021). J Comp Neurol. PubMed ID: 34961936

Drosophila's lateral posterior neurons (LPNs) belong to a small group of circadian clock neurons that is so far not characterized in detail. Thanks to a new highly specific split-Gal4 line, this study describes LPNs' morphology in fine detail, their synaptic connections, daily bimodal expression of neuropeptides, and a putative role of this cluster in controlling daily activity and sleep patterns is proposed. The three LPNs were found to be heterogeneous. Two of the neurons with similar morphology arborize in the superior medial and lateral protocerebrum and most likely promote sleep. One unique, possibly wakefulness-promoting, neuron with wider arborizations extends from the superior lateral protocerebrum toward the anterior optic tubercle. Both LPN types exhibit manifold connections with the other circadian clock neurons, especially with those that control the flies' morning and evening activity (M- and E-neurons, respectively). In addition, they form synaptic connections with neurons of the mushroom bodies, the fan-shaped body, and with many additional still unidentified neurons. Both LPN types rhythmically express three neuropeptides, Allostatin A, Allostatin C, and Diuretic Hormone 31 with maxima in the morning and the evening. The three LPN neuropeptides may, furthermore, signal to the insect hormonal center in the pars intercerebralis and contribute to rhythmic modulation of metabolism, feeding, and reproduction. These findings are discussed in the light of anatomical details gained by the recently published hemibrain of a single female fly on the electron microscopic level and of previous functional studies concerning the LPN.

Go to top

Compartment specific regulation of sleep by mushroom body requires GABA and dopaminergic signaling

Driscoll, M., Buchert, S. N., Coleman, V., McLaughlin, M., Nguyen, A. and Sitaraman, D. (2021). Sci Rep 11(1): 20067. PubMed ID: 34625611

Sleep is a fundamental behavioral state important for survival and is universal in animals with sufficiently complex nervous systems. Biogenic amines like dopamine, serotonin and norepinephrine have been shown to be critical for sleep regulation across species but the precise circuit mechanisms underlying how amines control persistence of sleep, arousal and wakefulness remain unclear. The fruit fly, Drosophila melanogaster, provides a powerful model system for the study of sleep and circuit mechanisms underlying state transitions and persistence of states to meet the organisms motivational and cognitive needs. In Drosophila, two neuropils in the central brain, the mushroom body (MB) and the central complex (CX) have been shown to influence sleep homeostasis and receive aminergic neuromodulator input critical to sleep-wake switch. Dopamine neurons (DANs) are prevalent neuromodulator inputs to the MB but the mechanisms by which they interact with and regulate sleep- and wake-promoting neurons within MB are unknown. This study investigated the role of subsets of PAM-DANs that signal wakefulness and project to wake-promoting compartments of the MB. This study found that PAM-DANs are GABA responsive and require GABA(A)-Rdl receptor in regulating sleep. In mapping the pathways downstream of PAM neurons innervating γ5 and β'2 MB compartments it was found that wakefulness is regulated by both DopR1 and DopR2 receptors in downstream Kenyon cells (KCs) and mushroom body output neurons (MBONs). Taken together, a dopamine modulated sleep microcircuit has been identified within the mushroom body that has previously been shown to convey information about positive and negative valence critical for memory formation. These studies will pave the way for understanding how flies balance sleep, wakefulness and arousal (Driscoll, 2021).

Go to top

Mushroom body evolution demonstrates homology and divergence across Pancrustacea

Strausfeld, N. J., Wolff, G. H. and Sayre, M. E. (2020). Elife 9. PubMed ID: 32124731

Descriptions of crustacean brains have focused mainly on three highly derived lineages of malacostracans: the reptantian infraorders represented by spiny lobsters, lobsters, and crayfish. Those descriptions advocate the view that domeor cap-like neuropils, referred to as 'hemiellipsoid bodies,' are the ground pattern organization of centers that are comparable to insect mushroom bodies in processing olfactory information. This study challenges the doctrine that hemiellipsoid bodies are a derived trait of crustaceans, whereas mushroom bodies are a derived trait of hexapods. It was demonstrated that mushroom bodies typify lineages that arose before Reptantia and exist in Reptantia thereby indicating that the mushroom body, not the hemiellipsoid body, provides the ground pattern for both crustaceans and hexapods. Evolved variations of the mushroom body ground pattern are, in some lineages, defined by extreme diminution or loss and, in others, by the incorporation of mushroom body circuits into lobeless centers. Such transformations are ascribed to modifications of the columnar organization of mushroom body lobes that, as shown in Drosophila and other hexapods, contain networks essential for learning and memory (Strausfeld, 2020).

As demonstrated in Drosophila melanogaster, paired mushroom bodies in the insect brain play manifold roles in learning and memory. One diagnostic tool for identifying putative mushroom body homologues in other mandibulates is an antibody raised against the catalytic subunit of protein kinase A, encoded by the Drosophila gene DC0, and required for effective learning and memory. The antibody, known as 'anti-DC0', selectively identifies the columnar neuropils of mushroom bodies of insects and of other arthropods with, until very recently, the notable exception of crustaceans. Indeed, numerous studies have disputed or expressed ambivalence about centers in comparable locations in the crustacean brain being mushroom body homologues) (Strausfeld, 2020).

The discovery of mushroom bodies in mantis shrimps has led to a review mushroom bodies in crustaceans and a search for these centers in other lineages of this taxon. This study provides evidence that mushroom bodies generally occur in Crustacea, thereby further supporting Hexapoda + Crustacea as the established subphylum Pancrustacea (Strausfeld, 2020).

Historically, identification of the insect mushroom body relied on just four neuroanatomical traits: an overall fungiform shape; a location exclusively in the brain's rostrolateral protocerebrum; an internal composition comprising hundreds to tens of thousands of intrinsic neurons originating from minute cell bodies clustered at the brain's rostral surface; and the possession of columnar lobes formed by the extended axon-like processes of intrinsic neurons. An often-included fifth character typical of almost all insects is a cap of neuropil called the calyx comprising the dendrites of intrinsic neurons, which receive inputs from the antennal lobes and other sensory neuropils. Even flightless Zygentoma (silverfish and firebrats) possess calyces, as do Diplura and Collembola, two sister groups of Insecta. However, Ephemeroptera (mayflies) and Odonata (dragonflies, darters) are calyxless; inputs to their mushroom bodies supply their columnar lobes directly. Such distinctions demonstrate a general property of all well-defined protocerebral brain centers: retention of ancestral ground patterns despite evolved modifications such as losses of some components and elaborations of others. For the insect mushroom body these modifications include: variations of columnar lobe organization; expanded representations of sensory modalities; and loss, or hypertrophy, of the calyces (Strausfeld, 2020).

Despite such variations, the insect mushroom body has additional features that are always present. Foremost is that the columnar lobes are serially partitioned down their length such that inputs to the lobes are constrained to discrete synaptic domains. These domains receive functionally defined inputs and supply functionally defined mushroom body output neurons (MBONs) that extend to circumscribed regions of the medial and rostral protocerebrum. There, they contribute to further higher processing and have arrangements comparable to those relating the mammalian hippocampus to cortical areas. Certain outputs from the columns also project recurrently to more distal levels of the mushroom body itself. These arrangements are defined by neurons usually expressing GABA. The same organization typifies the columnar lobes of the stomatopod mushroom bodies where input and output neurons are arranged along the length of the lobes (Strausfeld, 2020).

Also apparent across insect species is that subsets of intrinsic neurons parse the mushroom body lobes into longitudinal subunits that are differentiated from each other with regard to their clonal lineages, as well as by their aminergic and peptidergic identities. Comparable longitudinal divisions of the stomatopod mushroom body are represented by four columnar lobes that extend together approximately in parallel. In Drosophila and other insects these characteristic attributes have been shown to be critical in supporting functions relating to the modality, valence, stored memories and behavioral relevance of information computed by the mushroom body (Strausfeld, 2020).

In total, thirteen characters have been identified that define both the insect (Drosophila) and mantis shrimp (Stomatopod) mushroom bodies. This study shows that crown eumalacostracan species belonging to lineages originating early in evolutionary history share the same expanded set of diagnostic characters that define insect mushroom bodies, including the partitioning of columns into discrete circuit-defined domains. These can occur not only as segment-like partitions of the columnar lobes but also as tuberous outswellings, as they do in basal insects belonging to Zygentoma and Pterygota (Strausfeld, 2020).

Because stomatopods are unique amongst crustaceans in possessing elaborated optic lobes that serve a multispectral color and polarization photoreceptor system, it could be argued that mushroom bodies in the mantis shrimp are unique apomorphies that have evolved specifically to serve those modalities. However, Stomatopoda are an outgroup of Eucarida (Euphausiacea + Decapoda) and the status of mushroom bodies as the ancestral ground pattern is supported by corresponding centers in the lateral protocerebrum of later evolving eumalacostracan lineages. For example, Lebbeus groenlandicus, a member of the caridid family Thoridae, has been shown to possess paired mushroom bodies each comprising a layered calyx supplying intrinsic neuron processes to columnar lobes (Sayre, 2019). This study describe neuroanatomical characters defining mushroom bodies also in cleaner shrimps (Stenopodidae) and several groups of carideans (Strausfeld, 2020).

The recognition that mushroom bodies occur in crustaceans is not new. In his 1882 description of stomatopod brains, the Italian neuroanatomist Giuseppe Bellonci identified domed neuropils (corpo emielissoidale), from which extend columnar neuropils (corpo allungato). Bellonci explicitly homologized these with, respectively, the insect mushroom body calyx and its columnar lobes. This terminology was adopted for studies of Reptantia, a relatively recent malacostracan lineage that includes crayfish and lobsters. The terms 'hemiellipsoid body' or corpora pedunculata were used to denote centers lacking columnar lobes that were considered to be homologues of the mushroom body. However, over the last four decades numerous studies on Reptantia have insisted that hemiellipsoid bodies are apomorphic - genealogically distinct from mushroom bodies - and that they represent the ancestral ground pattern of crustacean learning and memory centers. Apart from a fundamental misunderstanding of the original meaning of the term 'hemiellipsoid body,' that viewpoint is incorrect and is in conflict with the demonstration in this study, and in two recent studies, that mushroom bodies hallmark eumalacostracan lineages that diverged earlier than Reptantia . Studies of Reptantia also show that at least one anomuran group has mushroom bodies equipped with columnar lobes. Comparisons across Eumalacostraca described in this study indicate that mushroom bodies are indeed ubiquitous across Crustacea. However, in contrast to mushroom bodies in insects, evolved modifications of the mushroom body ground pattern in crustaceans have resulted in highly divergent morphologies including, both within Caridea and in certain lineages of Reptantia, centers lacking defined lobe (Strausfeld, 2020).

The results demonstrate substantial variation of the mushroom body ground pattern in pancrustaceans where it has undergone major transformations in Reptantia (see Retention and divergence of the mushroom body ground pattern in Mandibulata), a monophyletic group for which fossil-calibrated molecular phylogenies estimate a time of origin at 400-385 mya. Reptantians are distinguished by their apomorphic accessory lobes, spherical or crescent-shaped neuropil comprising diminutive synaptic islets that receive inputs from the adjacent olfactory lobe. In reptantian infraorders Achelata, Astacidea, and Axiidea the accessory lobe supplies inputs to the 'hemiellipsoid bodies', which are distinct from the 'standard' mushroom body morphology. In Achelata and Astacidea these consist of two well-defined neuropils lacking columnar lobes, arranged side by side (as in spiny lobsters) or one over the other (as in crayfish). In Axiidea they are single highly condensed neuropil (Strausfeld, 2020).

The great majority of descriptions have focused on hemiellipsoid body neuropils some promoting them as apomorphic equivalents. However, two lines of neuroanatomical evidence support hemiellipsoid bodies as mushroom body homologues. First, their organization reflects an evolved transformation of the ancestral ground pattern, in which the columnar lobe neuropil has been subsumed into the calyx, which in Reptantia comprises elaborate networks of short-axon and anaxonal intrinsic neurons, and parallel fibers. These are associated with anti-DC0-positive synaptic microglomeruli, comparable to calycal microglomeruli reported for the stomatopod and hexapod calyces (Strausfeld, 2020).

Second, outputs (parasol cells) relaying multisensory associations from the hemiellipsoid bodies correspond to mushroom body output neurons (MBONs). Like MBONs, parasol cells, which target other regions of the lateral protocerebrum, are resolved by antisera against TH and 5HT (Strausfeld, 2020).

These astacid arrangements contrast with centers, also referred to as hemiellipsoid bodies, in the reptantian lineage Anomura. In the marine hermit crab, Pagurus hirsutiusculus, two adjoining calyces are composed of a hybrid arrangement of intrinsic neurons contributing to stratified arrangements of orthogonal networks corresponding to arrangements observed in the columnar mushroom body lobes of insects. Studies of these circuits in the more recent Coenobitidae have demonstrated that their synaptic organization corresponds to that of an insect mushroom body's column. In land hermit crabs (Birgus latro and Coenobita clypeatus) although the hemiellipsoid bodies are greatly inflated, matching the huge olfactory lobes that define these species' deutocerebra, they nevertheless possess the same morphological attributes as Pagurus, except that their stratified arrangements are reiterated many times over to provide stacked synaptic strata. In both marine and land hermit crabs, large efferent neurons, identified by their immunoreactivity to anti-5HT and anti-TH, insinuate dendrites that extend across and between these stratifications to provide yet another example of neurons that correspond to MBONs leading to protocerebral centers. Hemiellipsoid bodies in the squat lobster Munida quadrispina, belonging to the superfamily Galatheoidea, which shares a common ancestor with Paguroidea, have a comparable arrangement of strata, indicating this as a defining character of Anomura (Strausfeld, 2020).

In true crabs (Brachyura) identification of a possible mushroom body homologue is problematic. The few recent studies on the brains of crabs include descriptions of their optic lobes and reniform body, in addition to the brains of the shore crab Carcinus maenason, and species of fully terrestrial crabs. However, no clearly defined center, comparable to that of an astacid or anomuran 'hemiellipsoid body,' has been convincingly identified (Strausfeld, 2020).

Anti-DC0 immunohistology of the shore crab Hemigrapsus nudus reveals a prominent rostral immunoreactive domain, denoted by fissures and lobes, occupying almost the entire rostral half of the lateral protocerebrum immediately adjacent to the reniform body. This corresponds to the adjacent arrangement of the mushroom body and reniform body seen in Stomatopoda. In the fiddler crab Uca minax, two immunoreactive territories similarly occupy a rostral location in the lateral protocerebrum (Strausfeld, 2020).

These observations do not align with studies based on synapsin immunohistology approximatiing where hemiellipsoid bodies with astacid-like substructures might be. And their designation as astacid-type hemiellipsoid bodies implies derivation from a dome-like ancestral morphology. There is as yet no evidence for this, and the anti-DC0-immunoreactive volumes in Brachyura suggest a folded neuropil reminiscent of anti-DC0-immunoreactive centers of 'whip spiders' belonging to the arachnid order Amblypygi (Strausfeld, 2020).

Arachnids offer a paradigm for an evolutionary scenario where basal and later evolving taxa demonstrate ancestral morphologies, whereas intermediate lineages may show highly derived morphologies. In Arachnida, the more basal Solifugae and Scorpiones possess 'standard' anti-DC0-immunoreactive mushroom bodies with a columnar lobe comprising parallel intrinsic neurons intersected by afferent and efferent arborizations. The most recent arachnids, 'whip scorpions' (Thelyphonida), also possess 'standard' mushroom bodies, whereas the intermediate Amblypygi have greatly enlarged and highly folded anti-DC0-immunoreactive centers. Arachnida have evolved modifications departing from, as well as retaining, even in the most recent lineages, the ancestral mushroom body ground pattern (Strausfeld, 2020).

A similar evolutionary scenario appears to have occurred across Caridea and Reptantia where the ground pattern may reappear in younger lineages. In Crangonidae, Crangon franciscorum has a mushroom body-like calyx and columnar lobes, whereas its sister species Paracrangon echinata has a bulbous hemiellipsoid body-like center and a diminutive lobe, indiscernible unless labelled with anti-DC0. Crangon and Paracrangon, which occupy different habitats, are estimated to have diverged in the early Eocene, about 56 mya . In reptantian Anomura, there has been both evolved loss (Munididae, Coenobitidae) as well as retention (Paguridae) of the mushroom body's columnar lobe during divergence times spanning 200 million years (Strausfeld, 2020).

Second-order olfactory integration centers may not be required in species that live in conditions where associative memories or place memories are of negligible relevance. The minute anti-DC0-immunoreactive centers of Penaeus suggest evolved reduction in a species that spends part of its adult life in the featureless ecology of the off-shore water column. In certain isopods, possibly olfactory specialists, the mushroom body is reduced to a mere vestige. In terrestrial isopods, absence of a hemiellipsoid body has been suggested to relate to diminution of the olfactory pathway. Complete loss is also reported for species constrained to unusual ecologies, such as fresh water caverns. The marine isopod Saduria entomon also seems to have a greatly reduced hemiellipsoid body. But that the olfactory globular tract (OGT) terminates in a much reduced lateral protocerebrum does not imply the presence of a hemiellipsoid body (Strausfeld, 2020).

Nevetheless, miniaturizing of the mushroom body is exemplified in Leptostraca, sister to all Eumalacostraca. Until now, evidence for such a center has been tenuous and its absence would be problematic: if mushroom bodies and their modifications are a defining feature of Pancrustacea, then mushroom bodies would be expected to occur in this basal malacostracan group (Strausfeld, 2020).

Application of anti-RII to the brain of the leptostracan Nebalia pugettensis identifies an immunoreactive neuropil that comprises a relatively large shallow cap over the rostro-dorsal lateral protocerebrum. The cap provides two very short columnar lobes, in which anti-&α;-tubulin and anti-RII immunoreactivity suggest parallel arrangements of axon-like processes typical of the mushroom body ground pattern. Small perikarya, suggestive of globuli cells, are contiguous with a larger population also supplying optic lobe neuropils. The smallness of the mushroom body in Nebalia suggests a possible evolved miniaturization, comparable to that described from hexapod Collembola. The mid-Silurian Cascolus ravitis is the oldest known fossil stem representative of this group (approximately 430 mya), and leptostracans have since undergone morphological and probably behavioral simplification. Extant species are mostly epibenthic scavengers or suspension feeders, many limited to burrowing habits in simple ecologies (Strausfeld, 2020).

Nonetheless, life in the water column though devoid of visual landmarks is rich in other sensory cues: tactile, vibrational, thermal and chemosensory. In certain predatory Copepoda, small intertidal members of Multicrustacea, glomerular antennular lobes are connected by heterolaterally projecting OGTs to prominent hemiellipsoid bodies located in their lateral protocerebra (Strausfeld, 2020).

Likewise, the allotriocarid Remipedia, which are also blind predatory crustaceans, have prominent anti-DC0-immunoreactive centers in both lateral protocerebra supplied heterolaterally by the OGT. These centers give rise to a small columnar extension, also visible in thin sections as a short stubby column extending from the larger neuropil. This is reminiscent of the arrangement in Nebalia. Yet molecular phylogenies identify Remipedia as the closest crustacean relative of Hexapoda - a conundrum, because overall brain organization in Remipedia is typical of Malacostraca (Strausfeld, 2020).

The primary sensory input to mandibulate mushroom bodies is olfactory, and the olfactory pathways of crustaceans and insects, as far as the lateral protocerebrum, have been promoted as homologous. However, there are profound differences between the crustacean and insect olfactory systems. This paper proposes that these differences may underlie the finding that mushroom body morphology varies considerably across Malacostraca, whereas it is highly conserved across dicondylic insects (Strausfeld, 2020).

Differences between the crustacean and (dicondylic) insect olfactory system begin at the level of the uniramous appendages of the pancrustacean deutocerebrum: the crustacean paired antennules and the insect paired antennae. Although these are segmental homologues, the crucial difference between the two applies to their olfactory sensilla and olfactory receptor neurons (ORNs). Aesthetascs, the flexible odorant sensilla of crustaceans, are distributed in malacostracans on the antennule's lateral flagellum. Each sensillum may contain some hundreds of ORNs, the dendrites of which may further branch to provide extensive membrane surfaces. In insects, the four principal morphologies of odorant sensilla can contain 1-6 poorly branched or unbranched ORNs. An exception is the hymenopteran placode sensillum serving 30+ ORNs (Strausfeld, 2020).

These receptor-level distinctions are amplified by differences of sensory transduction mechanisms. Crustacean ORNs are exclusively equipped with ionotropic channels, whereas in dicondylic insects, in addition to ionotropic ORNs, the majority of ORNs employ ligand-gated channels, there being as many variants of these in a species as there are genes that encode them (Strausfeld, 2020).

Differences between insects and crustaceans are profound in the olfactory lobes, which in crustaceans comprise identical wedge- or lozenge-shaped synaptic glomeruli. ORN axons terminate in one or several of these, but there is no hard evidence that any specific type of ORN targets a specific subset of glomeruli. This contrasts with insects. As shown for Drosophila, axons from different genetically determined ionotropic ORNs target discrete volumes of the antennal lobes. In Drosophila, and insects generally, each glomerulus has a specific location (and size and shape) in its antennal lobe and is functionally unique in receiving converging terminals from, mainly, a single genetically identical set of ligand-gated ORNs. There are as many glomeruli as there are genes encoding ligand-gated receptors. These arrangements comprise an odotypic map, a feature apparently not present in crustaceans (Strausfeld, 2020).

Mushroom bodies are supplied by relay neurons (projection neurons: PNs) originating from the olfactory glomeruli. This feature is entirely different in crustaceans and insects. Hundreds - in some species thousands - of axons belonging to PNs extend rostrally from the crustacean olfactory lobes (or in certain reptantians also from their adjacent accessory lobes) to reach the lateral protocerebrum, many axons dividing into two tributaries extending to both sides. In crustaceans, PNs have wide-field dendrites extending to numerous glomeruli; no PN confines its dendrites to a single unit of the olfactory lobe. This sharply contrasts with insects where each glomerulus contains the dendritic trees of 3-5 PNs, and only a small population of additional PNs is associated with groups of glomeruli. Thus, relatively few PN axons extend rostrally from the insect antennal lobe and their axons exclusively target the ipsilateral protocerebrum to end in the mushroom body and, or, distally adjacent neuropils of the lateral horn. The only crustacean with comparable homolateral PN projections is the cephalocarid Hutchinsoniella macracantha, an allotriocarid basal to Branchiopoda, Hexapoda and Remipedia (Strausfeld, 2020).

As far as the lateral protocerebrum, each level of the olfactory pathway is distinctly crustacean or insect. No crustacean possesses ligand-gated ORNs, as far as is known. In crustaceans, the arrangement amongst ORN terminals and widely branching PNs implies all-to-all connectivity in the olfactory lobe. There is, as yet, no evidence for functional differentiation of olfactory glomeruli, and it is currently assumed that reconstruction of chemical and temporal patterning of the odorant milieu is distinct from its encoding by the insect system. Molecular phylogenetics implies that the insect odorant receptor family (ORs) appeared after terrestrialization and that the 'labelled-line' organization of ORN terminals in specific insect olfactory ('antennal') glomeruli is an evolved innovation (Strausfeld, 2020).

The contrast between the general uniformity of mushroom bodies in insects and their diversity of form in crustaceans may be considered in light of these distinctions between olfactory organization in insects and crustaceans. The suggestion here is that the 'labelled line' odotopic organization of the insect olfactory pathway from ORNs to antennal lobe glomeruli may have constrained divergent evolution of the insect mushroom body, whereas the all-to-all connections amongst ORNs and PNs within the crustacean olfactory lobe may have permitted relaxed mushroom body evolution (Strausfeld, 2020).

Along with differences in mushroom body morphology, ranging from columnar lobes to the hemiellipsoid body morphotype, pancrustacean species examined in this study also vary in their behavioral ecologies. The columnar lobes, and the calyces from which they protrude, are variously accentuated in certain genera, as exemplified by Lebbeus groenlandicus (Thoridae) and Alpheus bellulus (Alpheidae), whereas in others, circuits characteristic of the lobes are incorporated into the calyx to provide the dome-like structures Hanström referred to as hemiellipsoid bodies (Strausfeld, 2020).

Considering what is known about the habits of species investigated in this studuy, the observations suggest that highly mobile species occupying dynamic three-dimensional ecologies such as reefs have retained the mushroom body's columnar lobes. In insects, neurons associated with the columnar lobes provide continuous updating of spatial associations and their valences . Essentially, the aerial-terrestrial ecology exploited by dicondylic insects, all possessing mushroom bodies with columnar lobes, is comparable to elaborate three-dimensional ecologies of reefs and escarpments exploited by stomatopods and many caridean species likewise equipped with mushroom bodies having columnar lobes. The proposition is that crustaceans and insects possessing columnar mushroom bodies share the attribute of negotiating structurally elaborate three-dimensional habitats, often within defined territories. Ancestral mandibulates likely evolved in comparable ecologies. Micro-CT studies of their appendicular morphologies suggest that the oldest pancrustaceans from the lower Cambrian were locomotory adepts rather than simple animals that crawled on the seabed. That mushroom bodies are crucial for spatial awareness is supported by studies demonstrating that the lobes increase in size during the acquisition of information about three-dimensional space. Anomurans such as land hermit crabs, in which circuits characterizing mushroom body lobes have been subsumed into the calyces, have refined spatial memory for exploration and homing (Strausfeld, 2020).

Large columnar mushroom bodies typify crustacean species that show place memory. Examples are cleaner shrimps such as Stenopus hispidus, motile reef dwellers such as Lebbeus groenlandicus or Spirontocaris lamellicornis, and active hunters such as Stomatopoda. Life on the ocean floor, intertidal flats, or the bed of a stream or lake, such as that adopted by reptantian Achelata, Astacidea, and Axiidea, may require altogether different computational networks for coping with what is predominantly a planar world. Such distinctions are also suggested by recent Crangonidae, as discussed above (Strausfeld, 2020).

The entrenched opinion is that central to its optic lobe neuropils, the crustacean lateral protocerebrum comprises an integrative center called the 'medulla terminalis,' with which the hemiellipsoid body is invariably associated. The original term 'masse médullaire terminale,' was to indicate numerous neuropils proximal to the optic lobes. However, another study used the term for a large (nonexistent) fourth optic neuropil and even another used it to include all neuropils proximal to the optic lobes, usually without distinction except for the hemiellipsoid bodies (Strausfeld, 2020).

The designation 'medulla terminalis' should be abandoned. First, it suggests a single synaptic neuropil rather than an elaborate organization of discrete centers. Second, it misleads by assuming the presence of a hemiellipsoid body-like center even when not resolvable (e.g., Parhyale hawaiensis). Proposing equivalence of the 'medulla terminalis' with a specific center in the insect brain (e.g. the lateral horn), could distract from deeper scrutiny. The lateral horn is a neuropil complex situated between the mushroom body calyces and the lobula where it is supplied by projection neurons from olfactory and optic glomeruli. A possible, yet uncertain, crustacean homologue might be the reniform body, an integrative neuropil in a corresponding location that connects to the lobula and mushroom body in Stomatopoda and other Malacostraca. That some eyestalk neuropils in Malacostraca likely correspond to those of the insect lateral protocerebrum is supported by a previous study, which, to date, is the only comprehensive study to name neuropil volumes that illustrate a level of elaboration comparable to the insect lateral protocerebrum; it is also the first study to appreciate Hanströumm's deployment of the term hemiellipsoid body (Strausfeld, 2020).

The difficulty of relinquishing the hemiellipsoid body as a derived trait of crustaceans is exemplified by excluding the mushroom body's columnar lobes, which is problematic for reaching consensus about pancrustacean brain evolution. That mushroom bodies occur in Stomatopoda and caridean lineages necessitates reconsideration of the relationship between the mushroom body and structures referred to as hemiellipsoid bodies. One goal of this paper has been to show how transformations from a columnar lobed morphotype to a calycal (hemiellipsoid) morphotype can be followed across eumalacostracan evolution while unifying this center in Eumalacostraca and Allotriocarida. Observations of Myriapoda (Diplopoda + Chilopoda), show this mandibulate group likewise defined by paired anti-DC0-immunoreactive mushroom bodies. Neuroanatomical traits identifying insect mushroom bodies further resolve these centers in Chelicerata and Onychophora and, across phyla, in spiralian Annelida and Platyhelminthes suggesting an ancient origin of this center (Strausfeld, 2020).

It is indisputable that the diversity of crustacean mushroom bodies contrasts with the uniformity of mushroom bodies across Hexapoda other than Archaeognathawhere evolved loss of the center is implied by Archaeognatha's phylogenetic position in Hexapoda. No member of Hexapoda has been identified with an evolutionarily modified mushroom body comparable to a hemiellipsoid body. The one fascinating exception is an experimentally induced point mutation that transforms the wild-type insect mushroom body to a hemiellipsoid body-like center. This is in the Drosophila brain mutant ks63 (mbd 'deranged'), the paired centers of which lack columnar lobes because their Kenyon cell axons are constrained to within the voluminous hybrid (calyx + lobe) neuropil, supplied by normal projection neurons from the olfactory (antennal) lobes. The mutant flies are able to discriminate odors and show odor-induced behavior, but were not shown to be capable of learning odors (Strausfeld, 2020).

That the mushroom body ground pattern has persisted with far less variation across Hexapoda than it has across Crustacea is no reason to dispute that it unites Hexapoda and Crustacea. Evidence provided here demonstrates that (to paraphrase Richard Owen's famous statement), in Crustacea mushroom bodies are the same center under every variety of form and function. The notion that hemiellipsoid bodies are something phyletically separate and distinct from mushroom bodies is based on a historical misapprehension that diverts attention from why such a plentitude of variations evolved and what drove their evolution (Strausfeld, 2020).

Go to top

Aversive training induces both pre- and postsynaptic suppression in Drosophila

Zhang, X., Noyes, N. C., Zeng, J., Li, Y. and Davis, R. L. (2019). J Neurosci. 39(46):9164-9172. PubMed ID: 31558620

The alpha'beta' subtype of Drosophila mushroom body neurons (MBn) is required for memory acquisition, consolidation and early memory retrieval after aversive olfactory conditioning. However, in vivo functional imaging studies have failed to detect an early forming memory trace in these neurons as reflected by an enhanced G-CaMP signal in response to presentation of the learned odor. This study shows that aversive olfactory conditioning suppresses the calcium responses to the learned odor in both alpha'3 and alpha'2 axon segments of alpha'beta' MBn and in the dendrites of alpha'3 MBOn immediately after conditioning using female flies. Notably, the cellular memory traces in both alpha'3 MBn and alpha'3 MBOn are short-lived and persist for less than 30 min. The suppressed response in alpha'3 MBn is accompanied by a reduction of acetylcholine (ACh) release, suggesting that the memory trace in postsynaptic alpha'3 MBOn may simply reflect the suppression in presynaptic alpha'3 MBn. Furthermore, this study shows that the alpha'3 MBn memory trace does not occur from the inhibition of GABAergic neurons via GABAA receptor activation. Since activation of the alpha'3 MBOn drives approach behavior of adult flies, the results demonstrate that aversive conditioning promotes avoidance behavior through suppression of the alpha'3 MBn-MBOn circuit (Zhang, 2019).

Animals learn to avoid a neutral stimulus that is repeatedly coupled with an unpleasant one. This type of learning, aversive associative learning, induces cellular memory traces in engram cells in the brain and changes the representation of the neutral stimulus. In Drosophila, several memory traces detected with the calcium indicator G-CaMP have been observed in the mushroom body (MB), a brain region critical for olfactory learning and memory. These memory traces are detectable across discrete time periods extending from 30 min to several days after training. However, memory traces that form immediately in the MB after conditioning have not been detected with in vivo Ca2+ imaging (Zhang, 2019).

The MB is composed of ~2000 intrinsic neurons in each hemisphere that integrates olfactory cues received from antennal lobe projection neurons with aversive or rewarding stimuli from two clusters (PPL1, PAM) of dopamine neurons. MBn are classified into three major subtypes: α'β', αβ, and γ neurons, based on their birth order and projection patterns of their axons in the brain. The axons of α'β' and αβ MBn bifurcate and project within the vertical α'/α lobe and horizontal β'/β lobe neuropil, whereas the axons of γ neurons project only within the horizontal γ lobe neuropil. Although each of these MBn subtypes contributes to aversive olfactory memory, they do so at different times after conditioning, with synaptic transmission from the α'β' and γ MBn required for robust expression of early and intermediate-term memory (immediate to 3 h) and synaptic transmission from the αβ MBn having a more pronounced role for memory expression after 3 h. Importantly, although the α'β' MBn are required for memory acquisition, consolidation and early memory retrieval, no immediate memory trace in α'β' MBns has been detected using in vivo Ca2+ imaging (Zhang, 2019).

Five different types of MB output neurons (MBOns) tile the α'β' lobe with their dendritic trees into five discrete compartments, matching the tiling by axon terminals from presynaptic DAns. Several of these MBOns are required for aversive memory or appetitive memory expression, and intermediate-term memory traces (~1-2 h after conditioning) have been detected in some of these neurons. However, early memory traces have not been documented in these MBOns, and the relationship between such putative traces and those in the presynaptic MBn is unexplored. Connectome studies revealed that DAns make direct connection with MBOns, opening the possibility that MBOns form traces independently of the MBn (Zhang, 2019).

This study shows that a cellular memory trace forms immediately after conditioning in the MBn axons occupying the α'3 compartment and in the downstream α'3 MBOn. Functional Ca2+ imaging reveals that aversive conditioning suppresses subsequent responses to the learned odor in both the presynaptic α'3 compartment and the postsynaptic α'3 MBOn across a similar time period, suggestive of a causal relationship. In vivo ACh imaging revealed that the suppressed Ca2+ responses are accompanied by reduced ACh release in the α'3 compartment, supporting the model that the α'3 MBOn memory trace occurs from suppressed presynaptic activity. This study also shows that the conditioning-induced suppression in the α'3 compartment does not occur from increased inhibition through the Resistance to dieldrin (Rdl) GABAA receptor, indicating that mechanisms other than Rdl receptor activation are responsible for the suppression of activity (Zhang, 2019).

This study provides evidence for the existence of immediate cellular memory traces that form in at least two adjacent segments of the axons in the vertical lobe neuropil of the α'β' MBn and at least one (α'3 MBOn) of the corresponding output neurons. These memory traces, detected as decreased Ca2+ responses to the CS+ odor immediately after conditioning when compared with preconditioning responses, and persisting for >30 min before the response properties return to the naive state, are consistent with the fact that α'β' MBn are required for memory acquisition, consolidation and early memory retrieval. Several other previously characterized early memory traces due to odor conditioning provide an interesting background to these newly discovered traces. The neurites of the DPM neurons innervating the vertical MB lobe neuropil exhibit an increased Ca2+ response to the learned odor from ~30-70 min after conditioning. A memory trace forms in the antennal lobe, registered as the recruitment of new projection neuron activity in response to the learned odor, that lasts <10 min after conditioning. The activity of GABAergic APL neurons that synapse in the vertical lobe neuropil of the MBn is suppressed for a period of a few minutes after conditioning. Further, in vivo functional imaging of the α'β' MBn axons revealed an early memory trace displayed as increased Ca2+ influx by 30 min after conditioning that persists for at least 1 h. The action of these five memory traces, together along with other unknown traces, may provide the cellular modifications required for behavioral performance gains to be made across the first hour after conditioning. Memory traces in compartments other than α'3 and/or their MBOns may underlie the requirement of α'β' MBn for memory retrieval beyond the first hour (Zhang, 2019).

However, the developmental trajectory of the memory traces forming in the α'β' MBn lobe is of additional interest. As indicated above, a cellular memory trace forms in these neurons by 30 min after conditioning that is manifested as an increased Ca2+ response to the conditioned odor. The data presented in this study show that the α'β' MBn axons become suppressed across the first ~15 min after conditioning. The combined studies thus indicate that the CS+ odor response properties in the α'β' MBn axons are initially suppressed after conditioning but then become enhanced at later times. The time courses for the two cellular memory traces do not match exactly (0-15 min for the suppression and ~30-60 min for the increase) given the current data showing no detectable increase at 30 or 45 min, but this is easily explained by variation in the strength of conditioning or minor technical differences between the two studies. Thus, the most parsimonious conclusion is that the vertical axon compartments of the α'β' MBn initially exhibit a suppressed response to the CS+ followed by an increased response with the transition from suppression to enhancement occurring somewhere between ~30-45 min after conditioning. How this evolution in response properties from negative to positive with time translates into behavioral memory expression remains unclear (Zhang, 2019).

The suppressed responses to the CS+ odor were found in both the axon segments of the α'β' MBn and the dendrites of α'3 MBOn. Given that activation of α'3 MBOn drives approach behavior, the results are consistent with the model that aversive conditioning promotes avoidance through suppressing the MBn-MBOn circuits that signal positive valence, at least across the time that the MBOn responses are suppressed. Notably, the memory traces in α'3 MBn and α'3 MBOn persisted for the similar time, raising the question of whether the suppressed responses form independently or whether the α'3 MBOn memory trace simply reflects the presynaptic one. The data support the model in which the suppressed α'β' MBn responses are simply transmitted to the MBOn from reduced synaptic activity: the suppressed Ca2+ response in α'3 MBn axon compartment is correlated with reduced ACh release and the suppressed response in the α'3 MBOn dendrites. Behavioral data suggest that if the early cellular memory traces that form in the α'3 MBn-MBOn circuit cannot be readout precisely, the expression of behavioral memory early after conditioning becomes impaired. However, the possibility that memory traces can be formed independently in α'3 MBOn cannot be excluded (Zhang, 2019).

This study formulated the hypothesis that the immediate suppression in α'3 MBn axons after aversive conditioning might be due to enhanced GABAergic input to the α'3 compartment in an effort to delineate the underlying mechanism. However, attempts to detect any impairment of the immediate suppression in α'3 axonal compartment failed when Rdl GABAA receptor was knocked down in α'β' neurons. Thus, the data argue against attributing the suppression in the α'3 compartment to GABAergic inhibition through GABAA receptor (Zhang, 2019).

Go to top

Nitric oxide acts as a cotransmitter in a subset of dopaminergic neurons to diversify memory dynamics

Aso, Y., Ray, R. P., Long, X., Bushey, D., Cichewicz, K., Ngo, T. T., Sharp, B., Christoforou, C., Hu, A., Lemire, A. L., Tillberg, P., Hirsh, J., Litwin-Kumar, A. and Rubin, G. M. (2019). Elife 8. PubMed ID: 31724947

Animals employ diverse learning rules and synaptic plasticity dynamics to record temporal and statistical information about the world. However, the molecular mechanisms underlying this diversity are poorly understood. The anatomically defined compartments of the insect mushroom body function as parallel units of associative learning, with different learning rates, memory decay dynamics and flexibility. This study shows that nitric oxide (NO) acts as a neurotransmitter in a subset of dopaminergic neurons in Drosophila. NO's effects develop more slowly than those of dopamine and depend on soluble guanylate cyclase in postsynaptic Kenyon cells. NO acts antagonistically to dopamine; it shortens memory retention and facilitates the rapid updating of memories. The interplay of NO and dopamine enables memories stored in local domains along Kenyon cell axons to be specialized for predicting the value of odors based only on recent events. These results provide key mechanistic insights into how diverse memory dynamics are established in parallel memory systems (Aso, 2019).

An animal's survival in a dynamically changing world depends on storing distinct sensory information about their environment as well as the temporal and probabilistic relationship between those cues and punishment or reward. Thus it is not surprising that multiple distributed neuronal circuits in the mammalian brain have been shown to process and store distinct facets of information acquired during learning. Even a simple form of associative learning such as fear conditioning induces enduring changes, referred to as memory engrams, in circuits distributed across different brain areas. Do these multiple engrams serve different mnemonic functions, what molecular and circuit mechanisms underlie these differences, and how are they integrated to control behavior? Localizing these distributed engrams, understanding what information is stored in each individual memory unit and how units interact to function as one network are important but highly challenging problems. The Drosophila mushroom body (MB) provides a well-characterized and experimentally tractable system to study parallel memory circuits. Olfactory memory formation and retrieval in insects requires the MB. In associative olfactory learning, exposure to an odor paired with a reward or punishment results in formation of a positive- or negative-valence memory, respectively. In the MB, sensory stimuli are represented by the sparse activity of ~2,000 Kenyon cells (KCs). Each of 20 types of dopaminergic neurons (DANs) innervates compartmental regions along the parallel axonal fibers of the KCs. Similarly, types of mushroom body output neurons (MBONs) arborize their dendrites in specific axonal segments of the KCs; together, the arbors of the DANs and MBONs define the compartmental units of the MB. Activation of individual MBONs can cause behavioral attraction or repulsion, depending on the compartment in which their dendrites arborize, and MBONs appear to use a population code to govern behavior (Aso, 2019).

A large body of evidence indicates that these anatomically defined compartments of the MB are also the units of associative learning. Despite the long history of behavioral genetics in fly learning and memory, many aspects of the signaling pathways governing plasticity -- especially whether they differ between compartments -- remain poorly understood. Nevertheless, dopaminergic neurons and signaling play a key role in all MB compartments, and flies can be trained to form associative memories by pairing the presentation of an odor with stimulation of a single dopaminergic neuron. Punishment or reward activates distinct sets of DANs that innervate specific compartments of the MB. Activation of the DAN innervating a MB compartment induces enduring depression of KC-MBONs synapses in those specific KCs that were active in that compartment at the time of dopamine release. Thus, which compartment receives dopamine during training appears to determine the valence of the memory, while which KCs were active during training determines the sensory specificity of the memory (Aso, 2019).

Compartments operate with distinct learning rules. Selective activation of DANs innervating specific compartments has revealed that they can differ extensively in their rates of memory formation, decay dynamics, storage capacity, and flexibility to learn new associations. For instance, the dopaminergic neuron PAM-α1 can induce a 24h memory with a single 1-minute training session, whereas PPL1-α3 requires ten repetitions of the same training to induce a 24h memory. PPL1-γ1pedc (aka MB-MP1) can induce a robust short-lasting memory with a single 10-second training, but cannot induce long-term memories even after 10 repetitions of a 1-minute training. PAM-α1 can write a new memory without compromising an existing memory, whereas PPL1-γ1pedc extinguishes the existing memory when writing a new memory. What molecularand cellular differences are responsible for the functional diversity of these compartments? Some differences might arise from differences among KC cell types, but memory dynamics are different even between compartments that lie along the axon bundles of the same Kenyon cells (for example, α1 and α3). This paper shows that differences in memory dynamics between MB compartments can arise from the deployment of distinct cotransmitters by the DAN cell types that innervate them (Aso, 2019).

Evidence from a wide range of organisms establishes that dopaminergic neurons often release a second neurotransmitter, but the role of such cotransmitters in diversifying neuronal signaling is much less clear. In rodents, subsets of dopaminergic neurons co-release glutamate or GABA. In mice and Drosophila, single-cell expression profiling reveals expression of diverse neuropeptides in dopaminergic neurons. EM connectome studies of the mushroom body in adult and larval Drosophila reveal the co-existence of small-clear-core and large- dense-core synaptic vesicles in individual terminals of dopaminergic neurons; moreover, the size of the observed large-dense-core 02 vesicles differs between DAN cell types (Aso, 2019).

This study found that NOS, the enzyme that synthesizes NO, was located in the terminals of a subset of DAN cell types. NOS catalyzes the production of nitric oxide (NO) from L-arginine. Drosophila NOS is regulated by Ca2+/calmodulin, raising the possibility that NO synthesis might be activity dependent. Furthermore, the localization of the NOS1 protein in the axonal terminals of DANs is consistent with NO serving as a cotransmitter. The conclusion that NO acts as a neurotransmitter is supported by the observation that NO signaling requires the presence of a putative receptor, soluble guanylate cyclase, in the postsynaptic Kenyon cells. This role contrasts with the proposed cell-autonomous action of NOS in the ellipsoid body, in which NO appears to target proteins within the NOS-expressing ring neurons themselves, rather than conveying a signal to neighboring cells. The valence-inversion phenotype observed when PPL1-γ1pedc was optogenetically activated in a dopamine-deficient background can be most easily explained if NO induces synaptic potentiation between odor-activated KCs and their target MBONs. Modeling work is consistent with this idea, but testing this idea and 19 other possible mechanisms for NO action will require physiological experiments (Aso, 2019).

During olfactory learning, the concentration of Ca2+ in KC axons represents olfactory information. The coincidence of a Ca2+ rise in spiking KCs and activation of the G-protein-coupled Dop1R1 dopamine receptor increases adenylyl cyclase activity. The resultant cAMP in turn activates protein kinase A, a signaling cascade that is important for synaptic plasticity and memory formation throughout the animal phyla. In contrast, when DANs are activated without KC activity, and thus during low intracellular Ca2+ in the KCs, molecular pathways involving the Dop1R2 receptor, Rac1 and Scribble facilitate decay of memory (Aso, 2019).

This study found that NOS in PPL1-γ1pedc shortens memory retention, while facilitating fast updating of memories in response to new experiences. These observations could be interpreted as indicating that NO regulates forgetting. Indeed, NO-dependent effect requires scribble in KCs, a gene previously reported as a component of active forgetting. However, it is an open question whether the signaling pathways for forgetting, which presumably induce recovery from synaptic depression, are related to signaling cascades downstream of NO and guanylate cyclase, which appear to be able to induce memory without prior induction ofsynaptic depression by dopamine. Lack of detectable 1-day memory formation after spaced training with PPL1-γ1pedc can be viewed as a balance between two distinct, parallel biochemical signals, one induced by dopamine and the other by NO, rather than the loss of information (that is, forgetting). Confirming this interpretation will require better understanding of the signaling pathways downstream of dopamine and NO. The search for such pathways will be informed by the prediction from modeling that dopamine and NO may alter two independent parameters that define synaptic weights with a multiplicative interaction (Aso, 2019).

In the vertebrate cerebellum, which has many architectural similarities to the MB, long-term-depression at parallel fiber-Purkinje cell synapses (equivalent to KC-MBON synapses) induced by climbing fibers (equivalent to DANs) can coexist with long-term-potentiation by NO. In this case, the unaltered net synaptic weight results from a balance between coexisting LTD and LTP rather than recovery from LTD. This balance was suggested to play an important role in preventing memory saturation in the cerebellum 59 and allowing reversal of motor learning. In the Drosophila MB, a similar facilitation was observed of reversal learning by NO. The antagonistic roles of NO and synaptic depression may be a yet another common feature of the mushroom body and the cerebellum (Aso, 2019).

Opposing cotransmitters have been observed widely in both invertebrate and vertebrate neurons. A common feature in these cases is that the transmitters have distinct time courses of action. For instance, hypothalamic hypocretin-dynorphin neurons that are critical for sleep and arousal synthesize excitatory hypocretin and inhibitory dynorphin. When they are released together repeatedly, the distinct kinetics of their receptors result in an initial outward current, then little current, and then an inward current in the postsynaptic cells. In line with these observations, this study found that dopamine and NO show distinct temporal dynamics: NO-dependent memory requires repetitive training and takes longer to develop than dopamine-dependent memory. What molecular mechanisms underlie these differences? Activation of NOS may require stronger or more prolonged DAN activation than does dopamine release. Alternatively, efficient induction of the signaling cascade in the postsynaptic KCs might require repetitive waves of NO input. Direct measurements of release of dopamine and NO, and downstream signaling events by novel sensors will be needed to address these open questions (Aso, 2019).

Decades of behavioral genetic studies have identified more than one hundred genes underlying olfactory conditioning in Drosophila. Mutant andtargeted rescue studies have been used to map the function of many memory-related genes encoding synaptic or intracellular signaling proteins (for example, rutabaga, DopR1/dumb, DopR2/DAMB, PKA-C1/DC0, Synapsin, Bruchpilot, Orb2 and Rac1) to specific subsets of Kenyon cells. However, it is largely unknown if these proteins physically colocalize at the same KC synapses to form intracellular signaling cascades. Some of 97 these proteins might preferentially localize to specific MB compartments. Alternatively, they may distribute uniformly along the axon of Kenyon cells, but be activated in only specific compartments. Identification of cell-type-specific cotransmitters in DANs enabled a beginning to the exploration of this question (Aso, 2019).

Optogenetic activation of specific DANs was used to induce memory in specific MB compartments, while manipulating genes in specific types of KCs. This approach allowed mapping and characterization of the function of memory-related genes at a subcellular level. For example, the Gycbeta100B gene, which encodes a subunit of sGC, has been identified as 'memory suppressor gene' that enhances memory retention when pan-neuronally knocked down, but the site of its action was unknown. Gycbeta100B appears to be broadly dispersed throughout KC axons, based on the observed distribution of a Gycbeta100B-EGFP fusion protein. The experiments ectopically expressing NOS in PPL1-α3 DANs that do not normally signal with NO is most easily explained if sGC is available for activation in all MB compartments. What are the molecular pathways downstream to cGMP? How do dopamine and NO signaling pathways interact in regulation of KC synapses? Previous studies and RNA-Seq data suggest several points of possible crosstalk. In cultured KCs from cricket brains, cGMP-dependent protein kinase (PKG) mediates NO-induced augmentation of a Ca2+ channel current. However, no expression of either of the genes encoding Drosophila PKGs (foraging and Pkg21D) was detected in KCs in RNA-Seq studies. On the other hand, cyclic nucleotide-gated channels and the cGMP-specific phosphodiesterase Pde9 are expressed in KCs. Biochemical studies have shown that the activity of sGC is calcium dependent and that PKA can enhance the NO-induced activity of sGC by phosphorylating sGC; sGC isolated from flies mutant for adenylate cyclase, rutabaga, show lower activity than sGC from wild-type brains, suggesting crosstalk between the cAMP and cGMP pathways (Aso, 2019).

All memory systems must contend with a tension between the strength and longevity of the memories they form. The formation of a strong immediate memory interferes with and shortens the lifetimes of previously formed memories, and reducing this interference requires a reduction in initial memory strength that can only be overcome through repeated exposure. Theoretical studies have argued that this tension can be resolved by memory systems that exhibit a heterogeneity of timescales, balancing the need for both fast, labile memory and slow consolidation of reliable memories. The mechanisms responsible for this heterogeneity, and whether they arise from complex signaling within synapses themselves), heterogeneity across brain areas, or both, have not been identified (Aso, 2019).

NO acts antagonistically to dopamine and reduces memory retention while facilitating the rapid updating of memory following a new experience. Viewed in isolation, the NO-dependent reduction in memory retention within a single compartment may seem disadvantageous, but in the presence of parallel learning pathways, this shortened retention may represent a key mechanism for the generation of multiple memory timescales that are crucial for effective learning. During shock conditioning, for example, multiple DANs respond to the aversive stimulus, including PPL1-γ1pedc, PPL1-γ2α'1, PPL1-α3. This study has shown that optogenetic activation of these DAN cell types individually induces negative-valence olfactory memory with distinct learning rates. The NOS-expressing PPL1-γ1pedc induces memory with the fastest learning rate in a wild-type background, and this study shows that it induces an NO-dependent memory trace when dopamine synthesis is blocked, with a much slower learning rate and opposite valence (Aso, 2019).

Robust and stable NO-dependent effects were only observed when training was repeated 10 times. Under such repeated training, compartments with slower learning rates, such as α3, form memory traces in parallel to those formed in γ1pedc. Thus, flies may benefit from the fast and labile memory formed in γ1pedc without suffering the potential disadvantages of 58 shortened memory retention, as long-term memories are formed in parallel in other compartments. The Drosophila MB provides a tractable experimental system to study the mechanisms and benefits of diversifying learning rate, retention, and flexibility in parallel memory units, as well as exploring how the outputs from such units are integrated to drive behavior (Aso, 2019).

Go to top

Distinct dopamine receptor pathways underlie the temporal sensitivity of associative learning

Handler, A., Graham, T. G. W., Cohn, R., Morantte, I., Siliciano, A. F., Zeng, J., Li, Y. and Ruta, V. (2019). Cell 178(1): 60-75. PubMed ID: 31230716

Animals rely on the relative timing of events in their environment to form and update predictive associations, but the molecular and circuit mechanisms for this temporal sensitivity remain incompletely understood. This study shows that olfactory associations in Drosophila can be written and reversed on a trial-by-trial basis depending on the temporal relationship between an odor cue and dopaminergic reinforcement. Through the synchronous recording of neural activity and behavior, this study shows that reversals in learned odor attraction correlate with bidirectional neural plasticity in the mushroom body, the associative olfactory center of the fly. Two dopamine receptors, DopR1 and DopR2, contribute to this temporal sensitivity by coupling to distinct second messengers and directing either synaptic depression or potentiation. These results reveal how dopamine-receptor signaling pathways can detect the order of events to instruct opposing forms of synaptic and behavioral plasticity, allowing animals to flexibly update their associations in a dynamic environment (Handler, 2019).

While memories are often thought of as windows into the past, their adaptive value lies in the ability to predict the future. This study, took advantage of the concise circuitry of the Drosophila mushroom body to investigate how the precise timing of dopaminergic reinforcement allows animals to form and maintain predictive associations between cues and outcomes. While studies of associative learning have often focused on sensory cues that anticipate punishment or reward, equally informative are cues associated with their termination. This study demonstrates that shifting the relative timing of an odor and reinforcement by <1 s can switch the valence of an olfactory memory, underscoring the exquisite temporal sensitivity of this circuit. As a consequence, flies can form equivalent appetitive associations with odors that anticipate rewards or follow punishments, or aversive associations with odors that predict punishments or follow rewards. The symmetry of this behavioral modulation permits Drosophila to take advantage of all the temporally correlated features of their environment that can be used to infer causal relationships. Together, this work suggests a model in which the steep temporal sensitivity of associative learning arises from the concerted action of two dopamine receptor-signaling pathways that work in opposition to bidirectionally regulate the strength of KC-MBON signaling (see DopR1 and DopR2 Are Required for Behavioral Flexibility), allowing animals to maintain an accurate model of a complex and changing world (Handler, 2019).

In a dynamic environment, memories must be continually retouched and rewritten to maintain their relevance and predictive value. By monitoring how individual flies adapt their odor preferences over 50 conditioning trials, this study has revealed that Drosophila can form and reverse learned associations on a trial-by-trial basis, pointing to the fundamental flexibility of memory updating mechanisms (Handler, 2019).

Prior work in both Drosophila and mammals has suggested memory retention is regulated by multiple mechanisms at different timescales. If not reinforced, memories may passively fade over time, reflecting the natural turnover of molecular and neural hardware. Alternatively, memories can be actively eroded either by re-exposure to the learned odor in the absence of the anticipated dopaminergic reinforcement or the reinforcement in the absence of the odor, violating the expected contingency between these two events. In contrast, the brief episodes of odor and dopaminergic reinforcement (1-2 s) used in the current study are insufficient to overwrite an olfactory association when presented independently but can immediately reverse a prior association when paired together in time. The convergence of olfactory and DAN input to the mushroom body thus conveys information about their causal relationship, offering a mechanism to rapidly update memories to reflect the changing temporal structure of the environment (Handler, 2019).

While memory updating could rely on plasticity at various sites within this circuit, this study demonstrated that the bidirectional modulation of behavior is highly correlated with bidirectional changes in the strength of the same KC-MBON synapses within the mushroom body. Such bidirectional synaptic plasticity has been proposed to confer reversibility to learning circuits. For example, spike-timing dependent plasticity (STDP) can bidirectionally tune the strength of synaptic connections between neurons depending on the relative timing of spikes in pre- and post-synaptic neurons, mirroring the sensitivity to temporal order observed in associative learning. However, STDP requires nearly coincident firing patterns on a millisecond timescale, far more rapid than the temporal relationships between stimuli typically required for associative learning. In this study, by examining neural and behavioral modulation over the same timescales and even concurrently within the same individuals, the modulation of synaptic signaling within the mushroom body is linked to reversible changes in behavior (Handler, 2019).

Within the mushroom body, each compartment serves as a site of convergence between odor signaling conveyed by KCs and dopaminergic reinforcement, allowing dopamine-receptor pathways within KC axons to detect the temporal order of these inputs. The spatial patterns of dopamine release and dopamine receptor second-messenger cascades are found to adhere to the compartmentalized architecture of the lobes, permitting different synapses along the same KC axon to be independently regulated. These observations suggest that within a compartment, multiple neuromodulatory mechanisms tune neurotransmission depending on the temporal structure of conditioning. As a consequence, the distinct complement of KC-MBON synapses activated by odors that precede or follow a reinforcement are differentially regulated, allowing a single dopaminergic reinforcement to drive the synchronous formation of multiple odor associations, effectively enhancing the coding capacity of a compartment (Handler, 2019).

Dopamine shapes circuit function in diverse ways by engaging distinct classes of receptors that couple to different signaling cascades. In Drosophila, DopR1 and DopR2 have been proposed to play opposing roles in olfactory memory regulation at the behavioral level, with DopR1 essential to memory formation and DopR2 necessary for memory erosion. Yet the contribution of these receptors to synaptic modulation within the mushroom body has remained unclear. The current work reveals that the opposing behavioral roles of DopR1 and DopR2 are mirrored by their antagonistic regulation of KC-MBON signaling, with DopR1 required for the depression ensuing from forward pairing, while DopR2 is essential for the potentiation that follows backward pairing. Although DANs selectively innervating the mushroom body are sufficient to instruct bidirectional behavioral modulation, the broader expression of DopR1 and DopR2 leaves open the possibility that these receptors may also act at other sites within the nervous system to shape the temporal sensitivity of associative learning (Handler, 2019).

Using fluorescent sensors of DopR1 and DopR2 second messengers allowed the gaining insight into the spatial and temporal patterning of these intracellular signaling pathways during conditioning. While a potential limitation of optical reporters is their restricted sensitivity and dynamic range, these sensors nevertheless reveal that the selective recruitment of dopamine-receptor signaling cascades is sufficient to account for the temporal dependence of neural and behavioral modulation. Monitoring cAMP production during conditioning reveals that, while the DopR1 pathway serves as a coincidence detector, in accord with the coordinate regulation of adenylate cyclases by Gαs and calcium, it cannot autonomously encode the temporal order of events. In contrast, DopR2 signaling through Gαq strictly depends on the temporal sequence of KC and DAN activation (Handler, 2019).

Which component of the DopR2-signaling cascade is sensitive to the temporal ordering of odor and reinforcement? IP3 receptors that gate calcium release from the ER lumen represent an intriguing candidate, as their complex regulation by both IP3 and cytosolic calcium renders them inherently sensitive to the sequence of agonist binding: IP3 binding unmasks a calcium regulatory site required for channel opening, while high calcium in the absence of IP3 inhibits channel activity. Indeed, this study observed that ER calcium release is time locked to KC stimulation suggesting that the precise order dependence of this pathway relies on calcium entry subsequent to IP3 production. In the cerebellum, bidirectional plasticity at parallel fiber-Purkinje neuron synapses has been proposed to similarly rely on calcium release from the ER lumen via IP3 receptors. While the analogous circuit organization of the mushroom body and cerebellum has been well described, the current observations suggest they may share conserved molecular mechanisms for temporally precise synaptic modulation (Handler, 2019).

Together, this work points to dopamine receptor signaling pathways in KC axons as a key site of temporal coincidence and order detection during associative learning. While this study focused on the role of dopaminergic signaling within the γ4 compartment, timing-dependent bidirectional plasticity was observed to be shared characteristic of KC-MBON synapses in multiple compartments of the γ lobe. Therefore, the reversible modulation of behavior instructed by both the PAM or PPL DANs likely reflects bidirectional plasticity driven synchronously in the multiple compartments innervated by these DAN drivers. Aversive electric shock and sugar rewards evoke distributed patterns of activity across the DAN population, implying that these naturalistic reinforcers likewise instruct coordinated bidirectional plasticity across different compartments to rapidly shape the net output of the mushroom body. Similar patterns of DAN activity are also elicited by a fly's locomotion, raising the possibility that, in the context of an odor plume, an animal's behavior may serve as a reinforcement stimulus that itself drives bidirectional synaptic plasticity to regulate odor processing (Handler, 2019).

The ability to form or overwrite associations on a trial-by-trial basis allows for adaptive behavior in a noisy and uncertain environment where the temporal relationships between events may quickly change. However, animals must also have the capacity to store relevant memories persistently, even for a lifetime. Therefore, the reversible plasticity observed must co-exist with additional molecular and circuit mechanisms that underlie the formation and retention of longer-term associations. Indeed, recent work has described intrinsic differences between mushroom body compartments in their susceptibility to memory erosion as well as differences in second-messenger signaling in distinct KC subpopulations. Together, these results suggest that the differential expression or coupling of dopamine-receptor signaling pathways in different KC classes may tune synaptic plasticity rules to regulate the persistence of information storage. While this work connects molecular pathways within a sub-population of KCs to the emergence of short-term associations, functional dissection of these signaling cascades across the different lobes of the mushroom body will provide insight into the distinct timescales of memory formation and erosion (Handler, 2019).

Go to top

Identification of Dopaminergic Neurons That Can Both Establish Associative Memory and Acutely Terminate Its Behavioral Expression

Schleyer, M., Weiglein, A., Thoener, J., Strauch, M., Hartenstein, V., Kantar Weigelt, M., Schuller, S., Saumweber, T., Eichler, K., Rohwedder, A., Merhof, D., Zlatic, M., Thum, A. S. and Gerber, B. (2020). J Neurosci 40(31): 5990-6006. PubMed ID: 32586949

An adaptive transition from exploring the environment in search of vital resources to exploiting these resources once the search was successful is important to all animals. The neuronal circuitry that allows larval Drosophila melanogaster of either sex to negotiate this exploration-exploitation transition was examined. This was done by combining Pavlovian conditioning with high-resolution behavioral tracking, optogenetic manipulation of individually identified neurons, and EM data-based analyses of synaptic organization. Optogenetic activation of the dopaminergic neuron DAN-i1 was found to both establish memory during training and acutely terminate learned search behavior in a subsequent recall test. Its activation leaves innate behavior unaffected, however. Specifically, DAN-i1 activation can establish associative memories of opposite valence after paired and unpaired training with odor, and its activation during the recall test can terminate the search behavior resulting from either of these memories. These results further suggest that in its behavioral significance DAN-i1 activation resembles, but does not equal, sugar reward. Dendrogram analyses of all the synaptic connections between DAN-i1 and its two main targets, the Kenyon cells and the mushroom body output neuron MBON-i1, further suggest that the DAN-i1 signals during training and during the recall test could be delivered to the Kenyon cells and to MBON-i1, respectively, within previously unrecognized, locally confined branching structures. This would provide an elegant circuit motif to terminate search on its successful completion (Schleyer, 2020).

Go to top

Input Connectivity Reveals Additional Heterogeneity of Dopaminergic Reinforcement in Drosophila

Otto, N., Pleijzier, M. W., Morgan, I. C., Edmondson-Stait, A. J., Heinz, K. J., Stark, I., Dempsey, G., Ito, M., Kapoor, I., Hsu, J., Schlegel, P. M., Bates, A. S., Feng, L., Costa, M., Ito, K., Bock, D. D., Rubin, G. M., Jefferis, G. and Waddell, S. (2020). Curr Biol. PubMed ID: 32619479

Different types of Drosophila dopaminergic neurons (DANs) reinforce memories of unique valence and provide state-dependent motivational control. Prior studies suggest that the compartment architecture of the mushroom body (MB) is the relevant resolution for distinct DAN functions. This study used a recent electron microscope volume of the fly brain to reconstruct the fine anatomy of individual DANs within three MB compartments. The 20 DANs of the γ5 compartment, at least some of which provide reward teaching signals, can be clustered into 5 anatomical subtypes that innervate different regions within γ5. Reconstructing 821 upstream neurons reveals input selectivity, supporting the functional relevance of DAN sub-classification. Only one PAM-γ5 DAN subtype γ5(fb) receives direct recurrent feedback from γ5β'2a mushroom body output neurons (MBONs) and behavioral experiments distinguish a role for these DANs in memory revaluation from those reinforcing sugar memory. Other DAN subtypes receive major, and potentially reinforcing, inputs from putative gustatory interneurons or lateral horn neurons, which can also relay indirect feedback from MBONs. The single aversively reinforcing PPL1-γ1pedc DAN was similarly reconstructed. The γ1pedc DAN inputs mostly differ from those of γ5 DANs and they cluster onto distinct dendritic branches, presumably separating its established roles in aversive reinforcement and appetitive motivation. Tracing also identified neurons that provide broad input to γ5, β'2a, and γ1pedc DANs, suggesting that distributed DAN populations can be coordinately regulated. These connectomic and behavioral analyses therefore reveal further complexity of dopaminergic reinforcement circuits between and within MB compartments (Otto, 2020).

Go to top

Recurrent architecture for adaptive regulation of learning in the insect brain

Eschbach, C., Fushiki, A., Winding, M., Schneider-Mizell, C. M., Shao, M., Arruda, R., Eichler, K., Valdes-Aleman, J., Ohyama, T., Thum, A. S., Gerber, B., Fetter, R. D., Truman, J. W., Litwin-Kumar, A., Cardona, A. and Zlatic, M. (2020). Nat Neurosci 23(4): 544-555. PubMed ID: 32203499

Dopaminergic neurons (DANs) drive learning across the animal kingdom, but the upstream circuits that regulate their activity and thereby learning remain poorly understood. This study provides a synaptic-resolution connectome of the circuitry upstream of all DANs in a learning center, the mushroom body of Drosophila larva. Afferent sensory pathways and a large population of neurons were discovered that provide feedback from mushroom body output neurons and link distinct memory systems (aversive and appetitive). This was combined with functional studies of DANs and their presynaptic partners and with comprehensive circuit modeling. It was found that DANs compare convergent feedback from aversive and appetitive systems, which enables the computation of integrated predictions that may improve future learning. Computational modeling reveals that the discovered feedback motifs increase model flexibility and performance on learning tasks. This study provides the most detailed view to date of biological circuit motifs that support associative learning (Eschbach, 2020).

To behave adaptively in an ever-changing environment, animals must be able to learn new associations between sensory cues (conditioned stimuli, CS) and rewards or punishments (aversive and appetitive unconditioned stimuli, US), and continuously update previous memories, depending on their relevance and reliability (Eschbach, 2020).

Modulatory neurons (for example, DANs) convey information about rewards and punishments, and provide the teaching signals for updating the valence associated with CS in learning circuits across the animal kingdom (for example, the vertebrate basal ganglia or the insect mushroom body, MB). The co-occurrence of CS and modulatory neuron activity tuned only to the received US can support simple associative memory formation. To account for more complex behavioral phenomena, theories have been developed in which learning can be regulated by previously formed associations. According to reinforcement learning theories, learning is driven by errors between predicted and actual US (prediction errors) which are thought to be represented by the activity of DANs. Indeed, in many model organisms, the responses of modulatory neurons have been shown to be adaptive, including monkeys, rodents and insects. Despite recent progress the basic principles by which DAN activity is adaptively regulated and teaching signals are computed are not well understood (Eschbach, 2020).

A prerequisite for the adaptive regulation of modulatory neuron activity is convergence of afferent pathways that convey information about received US with feedback pathways that convey information about previous experiences. A comprehensive synaptic-resolution connectivity map of the feedback circuits that regulate modulatory neurons would provide a basis for understanding how learning is adaptively regulated by prior learning, but it has previously been out of reach (Eschbach, 2020).

Insects, especially in their larval stages, have small and compact brains that have recently become amenable to large-scale electron microscopy (EM) circuit mapping. Both adult and larval insect stages possess a brain center that is essential for associative learning, the MB. The MB contains neurons called Kenyon cells (KCs) that sparsely encode CS, MB modulatory neurons (collectively called MB input neurons, MBINs) that provide the teaching signals and MB output neurons (MBONs) whose activity represents learnt valences of stimuli. In the Drosophila larva, most modulatory neurons are DANs, some are octopaminergic neurons (OANs) and some have unidentified neurotransmitters (simply called MBINs). Modulatory neurons and MBONs project axon terminals and dendrites, respectively, onto the KC axons in a tiled manner, defining MB compartments, in both adult and larval Drosophila. In adult Drosophila, it has been shown that coactivation of KCs and DANs reduces the strength of the KC-MBON synapse in that compartment. Different compartments have been implicated in the formation of distinct types of memories, for example, aversive and appetitive, or short and long term. However, despite a good understanding of the structure and function of the core components of the MB in both adult and larval Drosophila, the circuits presynaptic to modulatory neurons that regulate their activity have remained relatively uncharacterized (Eschbach, 2020).

This study therefore reconstructed all neurons presynaptic to all modulatory neurons in an EM volume that spans the entire nervous system of a first instar Drosophila larva, in which all the core components of the MB were previously reconstructed. This study also determined which individual modulatory neurons are activated by punishments and reconstructed their afferent US pathways from nociceptive and mechanosensory neurons. The neurotransmitter profiles of some of the neurons in the network were characterized and some of the identified structural connections were functionally confirmed. Finally, a model was developed of the circuit constrained by the connectome, the neurotransmitter data and the functional data, and it was used to explore the computational advantages offered by the recently discovered architectural motifs for performing distinct learning tasks (Eschbach, 2020).

Modulatory neurons (for example, DANs) are key components of higher-order circuits for adaptive behavioral control, and they provide teaching signals that drive memory formation and updating. This study provides a synaptic-resolution connectivity map of a recurrent neural network that regulates the activity of modulatory neurons in a higher-order learning center, the Drosophila larval MB. Some of the recently identified structural pathways were functionally tested, and a model of the circuit was developed to explore the roles of these motifs in different learning tasks (Eschbach, 2020).

A large population was discovered of 61 feedback neuron pairs that provide one- and/or two-step feedback from the MBONs to modulatory neurons. Strikingly, it was found that many modulatory neurons receive more than 50% of their total dendritic input from feedback pathways. These results suggest that prior memories as represented by the pattern of MBON activity can strongly influence modulatory neuron activity (Eschbach, 2020).

Learning and memory systems in vertebrates and insects are often organized into distinct compartments implicated in forming distinct types of memories (for example, aversive and appetitive or short and long term). Interestingly, it was found that the majority of the discovered feedback pathways link distinct memory systems, suggesting that the MB functions as an interconnected ensemble during learning. Thus, prior memories formed about an odor in one compartment can influence the formation and updating of distinct types of memories about that odor in other compartments (Eschbach, 2020).

In adult Drosophila, functional connections between some MBONs and DANs have been reported, and some have been shown to play a role in short-term memory formation, long-term memory consolidation, extinction and reconsolidation, or in synchronizing DAN ensemble activity in a context-dependent manner. In some cases, direct MBON-to-DAN connections have been demonstrated. Although direct connections from several MBONs onto DANs exist in the larva, this study found that indirect connections via the feedback neurons account for a much larger fraction of a modulatory neuron's dendritic input than direct MBON synapses. This suggests that adaptive DAN responses may be largely driven by such indirect feedback (Eschbach, 2020).

Some of the one-step within-compartment feedback motifs that were found are analogous to the feedback motifs so far described for the DANs in the vertebrate midbrain. Although the diversity and the inputs of striatal feedback neurons have not yet been fully explored, in the future it will be interesting to determine whether many of the striatal feedback neurons also link distinct memory systems (Eschbach, 2020).

The use of internal predictions can dramatically increase the flexibility of a learning system. This study reveals candidate circuit motifs that could compute integrated predicted value signals across appetitive and aversive memory systems. A prominent motif that was identified is convergence of excitatory and inhibitory connections from MBONs from compartments of opposite valence onto DANs. In naive animals, odor-evoked MBON excitation in all compartments is thought to be similar. However, associative learning selectively depresses conditioned odor drive to MBONs in compartments where modulatory neuron activation has been paired with the odor. It is proposed that by comparing the conditioned odor-evoked MBON excitation in compartments of opposite valence via cross-compartment feedback connections, modulatory neurons compute an integrated predicted value signal across appetitive and aversive domains (Eschbach, 2020).

Convergence of feedback and US pathways could allow the computation of prediction errors An important aspect of reinforcement learning theories is the idea that modulatory neurons compare predicted and actual US (to compute so-called prediction errors) and drive memory formation or extinction depending on the sign of the prediction error. Although Drosophila modulatory neurons have not yet been directly shown to represent prediction errors, adult and larval Drosophila are capable of extinction, and the current study reveals candidate motifs that could support the comparison of expected and actual US. Modulatory neurons were found to receive convergent input from feedback pathways from MBONs and from US pathways. Modulatory neurons could therefore potentially compute prediction errors by comparing inhibitory drive from the feedback pathways with the excitatory drive from the US pathways, or vice versa. Consistent with this idea, some DANs were observed in this model that are inhibited by US alone and activated by CS+ alone, or vice versa (Eschbach, 2020).

This study also reveals that US pathways and feedback pathways converge at two levels: not only at the modulatory neurons themselves, but also at the FB2Ns. Actual and expected outcomes could therefore also be compared by FB2Ns. A recent study in the mouse ventral tegmental area has found that some pre-DAN neurons encoded only actual or only expected reward, whereas others encoded both variables. Thus, both in vertebrates and in insects, comparing predicted and actual outcomes may be a complex computation involving multiple levels of integration that eventually converge onto an ensemble of modulatory neurons (Eschbach, 2020).

An assumption in many reinforcement learning models is that all modulatory neurons receive a global scalar reward prediction error signal. The current study was able to analyze the comprehensive set of inputs of every individual uniquely identifiable modulatory neuron in a learning center. This revealed that each modulatory neuron receives a unique set of feedback inputs that could enable each neuron to compute a unique set of features. Consistent with this, a diversity of adaptive response types in the modulatory neurons was observed in this model This suggests that instead of computing a single global reward prediction error that is distributed to all modulatory neurons, the network uses a range of distinct compartmentalized and distributed teaching signals (Eschbach, 2020).

The connectivity and modeling studies revealed two architectural features of the circuit that provide input to the modulatory neurons that increase its performance and flexibility on learning tasks. The first is the multilevel feedback architecture that includes not only the previously known direct MBON feedback, but also multiple levels of indirect feedback. The second is the extensive set of cross-compartment connections. Modeling suggests that these motifs support improved performance on complex tasks that require the computation of variables such as predictions, prediction errors and context (Eschbach, 2020).

In summary, this study presents a complete circuit diagram of a recurrent network that computes teaching signals in a biological system, providing insights into the architectural motifs that increase its computational power and flexibility. The connectome-constrained model provides numerous predictions that can be tested in the future in a tractable model organism, for which genetic tools can be generated to monitor and manipulate individual neurons. The connectome, together with the functional and modeling studies, therefore provides exciting opportunities for elucidating the biological implementation of reinforcement learning algorithms (Eschbach, 2020).

Go to top

Circuits for integrating learned and innate valences in the insect brain

Eschbach, C., Fushiki, A., Winding, M., Afonso, B., Andrade, I. V., Cocanougher, B. T., Eichler, K., Gepner, R., Si, G., Valdes-Aleman, J., Fetter, R. D., Gershow, M., Jefferis, G. S., Samuel, A. D., Truman, J. W., Cardona, A. and Zlatic, M. (2021). Elife 10. PubMed ID: 34755599

Animal behavior is shaped both by evolution and by individual experience. Parallel brain pathways encode innate and learned valences of cues, but the way in which they are integrated during action-selection is not well understood. This study used electron microscopy to comprehensively map with synaptic resolution all neurons downstream of all Mushroom Body output neurons (encoding learned valences) and characterized their patterns of interaction with Lateral Horn neurons (encoding innate valences) in Drosophila larva. The connectome revealed multiple convergence neuron types that receive convergent Mushroom Body and Lateral Horn inputs. A subset of these receives excitatory input from positive-valence MB and LH pathways and inhibitory input from negative-valence MB pathways. Functional connectivity from LH and MB pathways and behavioral roles of two of these neurons was confirmed. These neurons encode integrated odor value and bidirectionally regulate turning. Based on this it is speculated that learning could potentially skew the balance of excitation and inhibition onto these neurons and thereby modulate turning. Together, this study provides insights into the circuits that integrate learned and innate valences to modify behavior (Eschbach, 2021).

Selecting whether to approach or avoid specific cues in the environment is essential for survival across the animal kingdom. Many cues have both innate valences acquired through evolution and learned valences acquired through experience that can guide action selection. Innate and learned valences are thought to be computed in distinct brain areas, but the circuit mechanisms by which they are integrated and by which learned valences can override innate ones are poorly understood. Using the tractable Drosophila larva as a model system, this study describes with synaptic resolution the patterns of convergence between the output neurons of a learning center (the MB) and an innately attractive pathway in the lateral horn (the LH). 62 neurons per brain hemisphere were identified that represent direct points of convergence between the MB and the LH, that fall into a number of different subtypes based on their patterns of MB and LH inputs and potentially encode a number of distinct features. One subtype of 10 convergence neurons (CNs) receives excitatory input from positive-valence MB and LH pathways and inhibitory input from negative-valence MB pathways. Functional connectivity from LH and MB pathways and behavioral roles of two of these neurons were confirmed. These neurons encode integrated odor value (coding for positive value with an increase in their activity) and regulate turning. They are activated by an attractive odor, and when activated they repress turning. Conversely, when inactivated, they increase turning. Based on this, it is speculated that learning could potentially skew the balance of excitation and inhibition onto these neurons and thereby modulate turning. For one of these neurons, this study indeed verified that aversive learning skews inputs towards inhibition. Together, this study provides insights into the circuit mechanisms by which learned valences could interact with innate ones to modify behavior (Eschbach, 2021).

The brain areas that compute innate and learned valences of stimuli interact with each other, but despite recent progress, their patterns of interaction are not fully understood. In principle, MBONs could synapse directly onto LHNs thereby directly modifying innate valences. Alternatively, LHNs could directly synapse onto MBONs. Finally, learned and innate valences could initially be kept separate, and MBONs and LHNs could converge on downstream neurons. This study found that in Drosophila larva (1) some MBONs synapse directly onto some LHNs; (2) some MBONs received direct synaptic input from LHNs and (3) many MBONs and LHNs converge onto downstream CNs, similar to findings in the adult. One MBON (m1) was also a CN, receiving significant direct input from other MBONs and from LHNs. Overall, the architecture suggests some early mixing of representations of innate and learned valences, but to some extent these representations are also kept separate in the LH and MB, and then integrated by the downstream CNs. Maintaining some initial separation of representations of innate and learned valences prior to integration could offer more flexibility and independent regulation, for example by context or internal state. Convergence neurons could compute learned valence by comparing odor-drive to positive- and negative-valence MBONs (Eschbach, 2021).

The prevailing model of MB function in adult Drosophila proposes that in naive animals the odor drive to positive- and negative-valence MBONs is equal such that their outputs cancel each other out and the LH circuits guide olfactory behavior. Learning alters the overall output towards positive- or negative-valence MBONs by modifying specific KC-to-MBON connections. This model raises several important questions. First, how is the output from MBONs of opposite valence integrated to compute a learned valence? Second, how does it interact with the output of the LH. The current findings provide further support for this model and shed insight into these questions (Eschbach, 2021).

EM reconstruction combined with neurotransmitter information has revealed a class of 10 CNs that receive excitatory input from positive-valence MBONs and inhibitory input from negative-valence MBONs. These CNs are well poised to compute the learned odor valence by comparing the odor drive to positive- and negative-valence MBONs (Eschbach, 2021).

For two members of this class (MBON-m1 and CN-33), their MB drive was tested in untrained larvae, in vivo or in explants. On average, across the population of untrained individuals, KC activation did not induce a significant change in ∂F/F0 in either CN. However, it was found that in some individuals the MB drive onto the CNs was excitatory, and in others inhibitory, indicating that the MB can provide both excitatory and inhibitory drive to the CNs. Consistently, when inhibitory neurotransmission was blocked using PTX, MBON-m1 and CN-33 showed strong excitatory responses to KC activation (Eschbach, 2021).

High variability was observed in the responses of both CN-33 and MBON-m1 both to activation of the whole MB pathway, as well as to odor presentation. This variability may in part be due to technical reasons. Indeed, the variable responses to pan-KC optogenetic activation could be due to differences in CsChrimson expression in KCs across individuals (the expression of CsChrimson was verified for each individual recording but it was not quantified). However, such a difference is unlikely to be a reason for a range of responses that spans inhibition to excitation following the stimulation of the same neuronal population. Likewise, the exposure of the sensory organ to an odor may vary from one animal to the other depending on the location of the larva's head in the chip's channel. Together with other technical aspects, this may account for some of the variance observed in in vivo. Indeed, a decomposition of the variance within the datasets by ANOVA revealed a significant effect of the identity of the individual on the variance. Interestingly though, the inter-individual variability contributed to a higher fraction of the overall variance in odor response in larvae with functional MBs than in larvae with silenced MBs. This analysis suggests that the MB pathway is a significant source of variability in odor-evoked responses in MBON-m1 and CN-33 and is consistent with the highly variable responses of these CNs to direct KC activation. The variability in response of CNs to MB inputs across untrained individuals could be due to different experiences prior to these experiments, as suggested also by the fact that MBON-m1 responses to trained odors are modified with training. High variability in MBON responses to odor across individuals has previously also been observed in the adult flies and has been related to different individual experiences. The MBs can also compute other kinds of information, such as internal state which may modulate an individual's disposition toward sensory stimuli. Therefore, the average response across untrained individuals might be similar to the response of a single individual in a naive neutral state (with any interindividual differences averaged out across a population), and the variability may represent the degree of freedom for the MB to tune this response to a stimulus depending on previous experience or state (Eschbach, 2021).

How do the learned valences modify the innate ones? How is conflict between opposite innate and learned valences resolved during action selection? One possibility could involve the integration of conflicting signals into a unified representation of value, a notion similar to common currency valuation of options, which could then be used to promote or suppress specific actions. The findings suggest that the 10 CNs that can read out the learned valence by comparing MBON inputs of opposite valence also integrate the learned valence with the innate one. Thus, the 10 CNs that receive inhibitory input from negative-valence MBONs and excitatory input from positive-valence MBONs, also receive inputs from a positive-valence LHN pathway. These neurons are therefore well poised to compute an integrated odor value and code for a positive value with an increase in their activity. For two members of this class (MBON-m1 and CN-33) it was possible to confirm these predictions. This study has shown they are activated by an innately attractive odor via the LH pathway in untrained animals. Furthermore, when the innate attractiveness of an odor was reduced through aversive training, the activity of MBON-m1 was also reduced. Finally, it was shown that activating MBON-m1 and CN-33 represses turning, further supporting the idea that their activation encodes positive value. Interestingly, it was also found that a decrease in their activity promotes turning, raising the possibility that they could bi-directionally encode value, coding for negative value with a decrease in their activity (Eschbach, 2021).

In principle, a single CN of this type could potentially be sufficient to compute integrated odor value and regulate turning, but a population of 10 CNs with similar patterns of input from positive- and negative-valence pathways was found that likely operate partially redundantly with each other. Each CN also had unique aspects of connectivity, raising the possibility that they may also encode partially complementary features and that the integrated value could be distributed across the CN population (Eschbach, 2021).

Based on these findings the following model is proposed that could explain the way in which learning could modulate innate responses to odors through a population of integrative CNs: (1) In naive animals, the CNs are activated by innately attractive odors, mainly via the LH pathway (3) when activated these neurons repress turning, which enables crawling towards an attractive odor source; (3) in naive animals the net MB drive onto these CNs is close to 0; (4) aversive learning can skew the net MB output onto CNs towards inhibition, so that aversively-conditioned odor fails to activate these neurons; (5) if CNs are not activated, turning rate remains high when crawling towards the odor; (6) without sufficient suppression of turning by an odor, the animals' ability to approach the odor source is impaired. This proposed model could explain how aversive learning suppresses approach of innately attractive odors. To fully suppress approach, multiple CNs of this class would likely need to be silenced (Eschbach, 2021).

The model presented above could readily be extended to explain how appetitive learning could enhance odor approach and how strong aversive learning could switch innate odor approach to learned avoidance. Thus, appetitive learning could skew the net MB drive towards excitation thereby repressing turning even more. In contrast, strong aversive learning could skew the conditioned odor drive towards inhibitory negative-valence MBONs so much that the inhibition would become stronger than the excitatory LH drive. The CN would then be inhibited by the aversively conditioned odor. Since inhibition of CNs promotes turning, the aversively conditioned odor could promote odor avoidance by inducing turning. Consistent with this idea, inhibitory responses to the innately attractive odor were observed in MBON-m1 and in CN-33, in some untrained individuals, and strong excitatory responses in others, proving that odor drive onto CNs can range all the way from strong excitation to inhibition. Inhibition of MBON-m1 following aversive learning in the microfluidic device was not observed, but this could be because the animals did not form a very strong aversive memory under these conditions. Testing these proposed extensions of the model will require the future development of automated single-animal training assays and calcium imaging tracking microscopes to correlate the strength of the learned behavior with the conditioned-odor response of the CNs (Eschbach, 2021).

Finally, EM reconstruction did reveal a potentially opposite population of CNs that are inhibited by MBONs of positive value. It will be interesting to test in the future whether these neurons potentially encode negative value with an increase in their activity and positive value with a decrease in their activity and regulate turning in the opposite way: promoting turning when they are activated and repressing it when they are inhibited. Having two different populations of neurons that encode value in opposite ways could further increase the dynamic range of a distributed value code (Eschbach, 2021).

This study found that many of the CNs that receive input from both MBONs and LHNs also provide direct (including CN-33) or indirect (including MBON-m1) feedback to MB modulatory neurons that provide teaching signals for learning (n = 18). In a former study, it was shown that at least some of these feedback connections are functional and can influence memory formation. For example, CN-33/FAN-7 is capable of generating an olfactory memory when it is paired with an odor. This type of connectivity is consistent with learning theories that propose that future learning is influenced by predicted value computed based on prior learning. A major role of the CNs discovered in this study may therefore be not only to organize current actions, but also to regulate learning (Eschbach, 2021).

In summary, the comprehensive synaptic-resolution architecture of the circuits downstream of the learning center output neurons presented in this study is a valuable resource for constraining future modelling and function studies of value computation, action selection, and learning (Eschbach, 2021)

Go to top

Interactions between amyloid precursor protein-like (APPL) and MAGUK scaffolding proteins contribute to appetitive long-term memory in Drosophila melanogaster

Silva, B., Niehage, C., Maglione, M., Hoflack, B., Sigrist, S. J., Wassmer, T., Pavlowsky, A. and Preat, T. (2020). J Neurogenet 34(1):92-105. PubMed ID: 31965876

Amyloid precursor protein (APP), the precursor of amyloid beta peptide, plays a central role in Alzheimer's disease (AD), a pathology characterized by memory decline and synaptic loss upon aging. Understanding the physiological role of APP is fundamental in deciphering the progression of AD, and several studies suggest a synaptic function via protein-protein interactions. Nevertheless, it remains unclear whether and how these interactions contribute to memory. In Drosophila, previous work has shown that APP-like (APPL), the fly APP homolog, is required for aversive associative memory in the olfactory memory center, the mushroom body (MB). The present study shows that APPL is required for appetitive long-term memory (LTM), another form of associative memory, in a specific neuronal subpopulation of the MB, the alpha'/beta' Kenyon cells. Using a biochemical approach, this study identified the synaptic MAGUK (membrane-associated guanylate kinase) proteins X11, CASK, Dlgh2 and Dlgh4 as interactants of the APP intracellular domain (AICD). Next, this study shows that the Drosophila homologs CASK and Dlg are also required for appetitive LTM in the alpha'/beta' neurons. Finally, using a double RNAi approach, it was demonstrated that genetic interactions between APPL and CASK, as well as between APPL and Dlg, are critical for appetitive LTM. In summary, these results suggest that APPL contributes to associative long-term memory through its interactions with the main synaptic scaffolding proteins CASK and Dlg. This function should be conserved across species (Silva, 2020).

AD is the principal neurodegenerative disorder affecting the elderly, and it is characterized by amyloid β (Aβ) deposition derived from proteolytic processing of amyloid precursor protein (APP). A pathological hallmark of AD is a progressive memory decline that correlates intimately with synaptic loss. One of the main hypotheses for the cognitive deficits observed in AD is thus a dysfunction of synapses leading ultimately to synaptic loss and alteration of neural network activity. Therefore, it is essential to understand the physiological role of APP at the synapse. APP is a transmembrane protein expressed on both sides of the synapse. The APP extracellular domain can mediate dimerization across the synapse or interact with extracellular matrix components, growth factors and receptor-like proteins. These interactions are involved in synapse stabilization during development and also in regulating synapse plasticity in mature neuronal networks. APP can undergo two types of proteolytic processing, including the non-amyloidogenic pathway, which is initiated by α-secretase and produces a secreted form of APP (sAPPα), and the amyloidogenic pathway, which successively involves β- and then γ-secretase to release Aβ peptide and an APP intracellular C-terminal domain (AICD). Although the manner in which proteolytic processing of APP and its derivatives interferes with neuronal physiology has been extensively studied, little is known about the function of APP intracellular domain at the synapse or its synaptic partners (Silva, 2020).

Studies in mammals suggest that APP can interact via its intracellular domain with synaptic MAGUK proteins such as X11, CASK, or PSD-95. MAGUK proteins are involved in the assembly, maintenance and remodeling of the scaffolding in synaptic compartments mainly via regulation of the targeting of receptors and ion channels to the synapse. Therefore, understanding the interactions between APP and MAGUKs should help decipher the synaptic function of APP (Silva, 2020).

The three mammalian orthologs APP, APLP1 and APLP2 are partially functionally redundant, whereas Drosophila expresses a single APP homolog named APP-like (APPL) that has been implicated in olfactory memory and visual memory. APPL is strongly expressed in the adult mushroom body (MB), the main olfactory memory center in insects. Previous work has investigated the function of APPL in Drosophila aversive olfactory memory. However, whether the APP synaptic partners and their interactions might contribute to memory is still unexplored. Several MAGUK homologs in Drosophila have been identified such as dX11, CASK/Caki, and Discs-large (Dlg). Similar to its mammalian counterpart, dX11 binds APPL, and both are necessary for synaptic remodeling at the Drosophila neuromuscular junction. Drosophila CASK regulates CaMKII activity, interacts with dX11. Mammalian Dlg1/SAP97, Dlg2/PSD-93, Dlg3/SAP102 and Dlg4/PSD-95 share similarities with the fly Dlg proteins DlgA and DlgS97), which are encoded by a single dlg gene. Both CASK and Dlg play key roles in neurotransmission, synaptogenesis and plasticity (Silva, 2020).

This study has aimed to decipher the role of APPL and its synaptic partners in appetitive olfactory memory. Additionally, to investigate the AICD interactome, this study used a proteo-liposome recruitment method and found that AICD interacts with the MAGUK synaptic proteins X11, CASK and Dlg. It was then found in flies that APPL, CASK and Dlg are required specifically in the same neuronal subpopulation (the α'/β' KCs) for appetitive LTM. Finally, a double RNAi strategy was used to demonstrate that genetic interaction between APPL and MAGUKs is critical for appetitive LTM. To determine whether this memory deficit could be due to a major disorganization of the synaptic structure, both the pre-synaptic and the post-synaptic sites of MB α'/β' neurons were investigated using confocal immuno-labeling of synaptic proteins (Silva, 2020).

Previous work has showed that APPL is required in the α/β and γ KCs for aversive LTM. Additional work demonstrated that APPL is required for aversive LTM and MTM in the DPM neurons, a pair of serotonergic and GABAergic neurons that project to each of the MB lobes, where they connect both pre- and post-synaptically to the KCs. This study has found that APPL is also required in another form of long-lasting protein synthesis-dependent memory, appetitive LTM, albeit surprisingly in a different subpopulation of KCs, the α'/β' neurons. The requirement of APPL in α'/β' but not in the other KCs for appetitive LTM suggests that it has a specific role in appetitive LTM consolidation, as consolidation of appetitive LTM requires synaptic neurotransmission from α'/β'. Interestingly, recurrent activity of the α'/β'-DPM loop has been described as necessary to consolidate appetitive memories, with LTM eventually being stored in the α/β neurons. An involvement of APPL in memory consolidation may rely on transsynaptic APPL interactions and may also contribute to the molecular support of the α'/β'-DPM loop. Eventually, such a role of APPL in memory consolidation through transsynaptic interaction would be consistent with published research in mammals. Indeed, at the cellular level APP is expressed in pre- and postsynaptic compartments and can form trans-dimers that have been suggested as necessary for synaptic function. In addition, perturbation of APP function by intraventricular infusion of an antibody against APP induced memory impairments only when it was performed during the memory consolidation phase of a passive avoidance task. As shown for other synaptic cell adhesion molecules, the regulation of APP expression at the neuronal membrane is critical for hippocampal-dependent memory consolidation in the dentate gyrus, suggesting a potential involvement of APP in synaptic remodeling. Altogether, the present findings combined with research on mammalian models suggest that APP might have a conserved function across species in memory consolidation processes via transsynaptic interactions (Silva, 2020).

This study has shown that in addition to the previously known X11 proteins, the intracellular domain of APP interacts with other scaffolding proteins (CASK, Dlg2/PSD-93 and Dlg4/PSD-95). However, it is not clear whether these proteins interact directly with APP or if their interactions are mediated by the X11 adaptor proteins, as it has been described for CASK. This study demonstrates that the Drosophila homologs of these proteins, i.e. CASK and Dlg, are required in the same neuronal subpopulation as APPL for proper long-term memory. Several arguments in favor of an APPL/X11/CASK/Dlg macromolecular complex can be found in previous studies on mammals or Drosophila. In the Drosophila visual system, the MAGUK complex Lin-7/Dlg/CASK is involved in synaptic stabilization and the interaction between CASK and Dlg proteins has been described as direct. The role of CASK in the recruitment of Dlg1 to the membrane in various cell types has been confirmed in a recent study showing that the N-terminal domain of Dlg1 is critical for its interaction with CASK . In mammals, the existence of the APP/X11/CASK ternary complex has also been documented, and the regulation of neuronal excitability through potassium Kir2 channels involves a macromolecular complex consisting of Lin-7/SAP97/CASK/X11. Altogether these studies demonstrate that the interaction between the Dlg, CASK and X11 proteins is both possible and functionally relevant for neuronal physiology. Finally, CASK binds to the Dlg protein SAP97 in mammalian hippocampal neurons to regulate its conformation state and thus its role in glutamate receptor trafficking and insertion at the synapse. Thus, the existence of the APPL/X11/CASK/Dlg macromolecular complex is consistent with previously published reports, and even if this study only demonstrate the genetic interaction between APPL, CASK and Dlg for appetitive LTM, it is likely that such a complex exists in Drosophila α'/β' MB neurons. Complementary studies using genetic tools to impair the interactions between these proteins such as overexpression of an interfering peptide corresponding to the N-terminus of Dlg as in a previous study could bring the confirmation that protein-protein interactions between APPL/X11/CASK/Dlg are required for appetitive LTM via synaptic stabilization of the α'/β' neurons. The present study looked at synaptic organization using confocal microscopy and immuno-labeling of either pre-synaptic proteins of the active zone or post-synaptic scaffold proteins. The levels of the analyzed pre- or post-synaptic proteins were not affected by the concomitant knockdown of APPL and MAGUKs in the adult α'/β' neurons, indicating no obvious alterations in MB synaptic structure. It is noted that as expected this study observed a significant decrease in Dlg levels in the MB calyx upon knockdown of APPL;Dlg in α'/β' neurons, demonstrating the efficacy of the knockdown. To determine the requirement of these proteins for pre- or post-synaptic subtle organization, higher-resolution imaging studies would be required. However, such modification might be difficult to observe if, as suggested by the role of CASK and Dlg proteins, the APPL/X11/CASK/Dlg macromolecular complex is involved in synaptic stabilization specifically when synapses are modified during a plastic event, and not as a basal mechanism for synaptic organization and formation as described for APPL and FasII interactions at the NJM (Silva, 2020).

Interestingly, the Dlg protein SAP97 is involved in trafficking the α-secretase ADAM10 to the synapse through direct interaction, consequently regulating APP processing and the production of the neurotrophic and neuroprotective secreted-APPα fragment. Previous work has shown in Drosophila that a secreted fragment of APPL is involved in aversive memory, as well as KUZBANIAN, the Drosophila homolog of ADAM10. Therefore, an APPL/X11/CASK/Dlg supramolecular complex could also be involved in recruiting α-secretase at the synaptic site as well as generating sAPPα (Silva, 2020).

The subcellular localization of APP and MAGUK interactions in the KCs is still an open question. In Drosophila, Dlg and CASK are known to be present at both the pre- and postsynaptic compartments. However, APPL has been described mainly in the neuropil of KCs. Furthermore, the APPL binding protein X11 targets APPL to the axonal compartments and excludes it from MB dendrites via endocytosis. These data suggest that the APPL-MAGUKs complexes would be localized in the α'/β' KC axonal compartment (i.e. the α'/β' lobes), which is importantly also the site of the DPM/KC dialog for appetitive LTM consolidation (Silva, 2020).

In conclusion, this work highlights a novel role of APPL and its synaptic partners in appetitive long-term memory in Drosophila. Genetic interactions between APPL and the MAGUKs are critical for appetitive LTM in the α'/β' KCs, a neuronal sub-population known to be involved in the consolidation of appetitive LTM. A model is proposed in which the role of the interactions between APPL, CASK and Dlg might be the synaptic stabilization of the α'/β'-DPM loop through transsynaptic interactions (Silva, 2020).

Go to top


date revised: 27 December 2021

Home page: The Interactive Fly © 2020 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.