Drosophila has robust behavioral plasticity to avoid or prefer the odor that predicts punishment or food reward, respectively. Both types of plasticity are mediated by the mushroom body (MB) neurons in the brain, in which various signaling molecules play crucial roles. However, important yet unresolved molecules are the receptors that initiate aversive or appetitive learning cascades in the MB. D₁ dopamine receptor dDA1 (FlyBase name: Dopamine receptor) has been shown to be highly enriched in the MB neuropil. This study demonstrates that dDA1 is a key receptor that mediates both aversive and appetitive learning in pavlovian olfactory conditioning. Two mutants, dumb¹ and dumb², have abnormal dDA1 expression. When trained with the same conditioned stimuli, both dumb alleles showed negligible learning in electric shock-mediated conditioning while they exhibited moderately impaired learning in sugar-mediated conditioning. These phenotypes are not attributable to anomalous sensory modalities of dumb mutants because their olfactory acuity, shock reactivity, and sugar preference are comparable to those of control lines. Remarkably, the dumb mutant's impaired performance in both paradigms is fully rescued by reinstating dDA1 expression in the same subset of MB neurons, indicating the critical roles of the MB dDA1 in aversive as well as appetitive learning. Previous studies using dopamine receptor antagonists implicate the involvement of D₁/D₅ receptors in various pavlovian conditioning tasks in mammals; however, these have not been supported by the studies of D₁- or D₅-deficient animals. The findings described in this study here unambiguously clarify the critical roles of D₁ dopamine receptor in aversive and appetitive pavlovian conditioning (Kim, 2007b).

Pavlovian (classical) olfactory conditioning in Drosophila tests the animal's ability to learn and remember the odor [conditioned stimulus (CS)] associated with diverse unconditioned stimuli (US) and is instrumental in investigating the neural and cellular mechanisms underlying distinct learning and memory processes. When subjected to concurrent odor (CS+) and electric shock (aversive US) presentation, flies learn to avoid the CS+ odor in the absence of shock. Conversely, flies learn to prefer the CS+ odor after concurrent odor (CS+) and sugar (appetitive US) exposure. Thus, the same CS+ triggers either avoidance or preference behavior depending on previous experience of the flies. Several key questions arise regarding the underlying mechanisms. Specifically, are common or separate neural systems required for aversive versus appetitive learning and memory? What are the critical molecular and cellular events that distinguish reward versus punishment information (Kim, 2007b)?

Two key components are essential for both aversive and appetitive conditioning. One component is the cAMP signaling pathway. Flies defective in cAMP metabolism, such as dunce (cAMP-specific phosphodiesterase) and rutabaga [rut; calcium/calmodulin (CaM)-dependent adenylyl cyclase (AC)], or flies with altered activities of cAMP effectors Protein kinase A and dCREB2 are impaired in learning and/or memory in aversive conditioning. Likewise, appetitive conditioning requires cAMP because rut mutants display poor learning. The other component is the mushroom body (MB) brain structure. Flies with ablated MB structures or functions are completely defective in aversive learning. Moreover, synaptic output of different MB lobes is involved in memory formation or retrieval in aversive and appetitive conditioning. These indicate the MB as a central neural substrate for olfactory learning and memory. This poses a fundamental question regarding the neuromodulators and their receptors that initiate the cAMP cascade in the MB for aversive and appetitive learning and their memories (Kim, 2007b).

The neuromodulators that are crucial for olfactory conditioning and activate cAMP increases are dopamine and octopamine (see Tyramine β hydroxylase). Previous studies of Drosophila larvae and adults show that dopaminergic neuronal activities are essential for aversive, but not for appetitive, learning, whereas octopamine or octopaminergic neuronal activities are necessary only for appetitive learning (Schwaerzel, 2003; Schroll, 2006). Consistently, the activities of dopaminergic neurons projecting to the MB are mildly increased by odor stimuli and strongly by electric shock (Riemensperger, 2005). Moreover, duration of their activities is prolonged when the CS+ odor is presented, suggesting the role of dopamine neurons in US prediction. However, it is yet unknown whether dopamine directly activates the MB for aversive learning. To uncover the signal(s) activating the learning and memory cascade, three receptors have been identifed that are highly enriched in the MB and increase cAMP levels, and they are two dopamine receptors, dDA1 and DAMB, and an octopamine receptor, OAMB (Han, 1996; Han, 1998; Kim, 2003). This study shows that dDA1 is required in the MB for aversive and appetitive learning (Kim, 2007b).

Previous research on olfactory conditioning in Drosophila has primarily focused on the intracellular components, many of which are involved in learning and/or memory processes in the MB. Although those studies have revealed important insights, the receptors that initiate signaling cascades into motion in the MB are unknown. The findings presented here provide the first demonstration of a MB receptor essential for aversive learning. The role of dDA1 in this behavioral plasticity is physiological, rather than developmental. This is consistent with the observations that synaptic output of dopaminergic neurons is necessary during training (Schwaerzel, 2003), and the learning phenotype of rut mutants is rescued by the restricted expression of rut-AC, a potential dDA1 effector, in the adult MB (McGuire, 2003; Mao, 2004). Moreover, the dopaminergic processes projecting to the MB gamma lobe strongly respond to electric shock (US) and show altered activities after CS+ exposure (Riemensperger, 2005). Together, these strongly implicate dDA1 as a receptor conveying aversive US information in the MB lobes for memory formation (Kim, 2007b).

Two additional neuromodulator systems are previously implicated in aversive olfactory conditioning in Drosophila. One neuromodulator system is the glutamate NMDA receptor composed of dNR1 and dNR2 subunits. Flies with decreased dNR1 expression show diminished performance in aversive conditioning. Although dNR1 in the MB is crucial for anesthesia-resistant and midterm memories, dNR1-dependent learning occurs outside of the MB. Another putative modulator involved in olfactory learning is Amn, which has sequence homology with mammalian neuropeptide PACAP. Although amn mutants are mostly defective in midterm memory, they are mildly impaired in learning when BA is used as CS+ and learning of BA depends on synaptic output of Amn-expressing DPM neurons projecting to the MB lobes. Thus, it has been suggested that putative amn-encoded neuropeptides, by binding to their receptor(s) in the MB neuropil, may mediate memory formation; however, the predicted Amn neuropeptides or their receptors remain unidentified. Therefore, dDA1 represents the only MB receptor identified to date that is essential for aversive learning. Notably, dumb mutants, similar to MB-less flies, show negligible learning. This indicates that the MB neurons absolutely require dDA1 for aversive memory formation (Kim, 2007b).

The data presented in this study demonstrate the crucial role of dDA1 in sugar-mediated olfactory learning as well. Interestingly, dumb mutants have diminished, yet significant, performance scores, implicating an additional receptor(s) for this type of learning. Indeed, tßh mutants lacking octopamine show severe impairment in appetitive conditioning, which is rescued by feeding octopamine before, but not after, training (Schwaerzel, 2003). Thus, octopamine represents another neuromodulator crucial for appetitive learning. Because the MB is a primary neural substrate for appetitive conditioning, reward memory formation is likely mediated by dDA1 and an octopamine receptor(s) in the MB (Kim, 2007b).

The previous study of TH-GAL4/UAS-Shi^ts flies, in which endocytosis of the dopamine neurons expressing TH-GAL4 can be temporally controlled by dominant-negative dynamin Shi^ts, suggests that dopamine is not involved in appetitive conditioning (Schwaerzel, 2003; Kim, 2007a). This is contrary to the learning phenotype of D₁ dopamine receptor mutants dumb. Nonetheless, the discrepancy may be reconciled by several reasonable possibilities: (1) TH-GAL4 used in the previous study to drive Shi^ts may not be expressed, or expressed at low levels, in a subset of the dopamine neurons critical for appetitive learning; (2) dopamine neuronal output conveying sugar information may not be completely inhibited by Shi^ts; (3) dopamine crucial for appetitive learning may be secreted by a dynamin-independent pathway. These possibilities may be tested by investigating pale mutants that are unable to synthesize dopamine; however, such flies die during development because of an essential role of dopamine in cuticle formation (Budnik, 1987). Future studies on conditional pale mutants should help resolve this issue (Kim, 2007b).

dDA1 expression driven by MB247-GAL4 fully rescues the learning phenotypes of dumb mutants in electric shock- as well as sugar-mediated conditioning. This indicates that appetitive and aversive memory formations are mediated by dDA1 in the same subset of the MB neurons (~30% of all MB neurons). This poses an intriguing question as to how those MB neurons distinguish punishment versus reward information delivered by dDA1 to generate avoidance versus preference behavioral output. The key to answering this question may be intracellular effectors in the MB neuropil. rut-AC is crucial in the MB247-GAL4-expressing MB neurons for both aversive and appetitive learning (Schwaerzel, 2003). Notably, rut mutants retain some learning capacities in electric shock- and sugar-mediated conditioning, whereas MB-less flies or the flies with inhibited MB synaptic output exhibit no trace of learning in both assays. This implicates additional cellular components crucial for aversive and appetitive memory formation. Because dumb mutants are rather completely impaired in aversive learning, dDA1 may activate rut-AC and other cellular components in the MB to process punishment information (Kim, 2007b).

G-protein-coupled receptors including dopamine receptors can recruit multiple effector systems through heteromeric G-proteins or through cross-interactions of diverse signaling components. Thus, for aversive learning, dDA1 activated by electric shock US input may recruit the mitogen-activated protein (MAP) kinase cascade in addition to the cAMP pathway. The activated protein kinase A and MAP kinases may act on ion channels or cell adhesion molecules such as integrin and FasII to modify MB synaptic output, leading to avoidance behavior. Consistently, the flies defective in 14-3-3 and S6KII, which are involved in the MAP kinase cascade, and α-integrin and fasII mutants are poor learners in electric shock-mediated conditioning (Kim, 2007b).

For reward-mediated learning, reward US input may impinge on at least two receptors, dDA1 and an octopamine receptor, in the MB. Their simultaneous activities may recruit multiple effectors that possibly include rut-AC, MAP kinases, protein kinase C, and CaM kinase II. The biochemical changes collectively activated by these effectors may alter MB synaptic output to generate preference behavior. Interestingly, OAMB (octopamine receptor) activates the increases in intracellular calcium as well as cAMP (Han, 1998) and is a good candidate that can turn on the aforementioned effectors for processing reward information in the MB. The punishment and reward effectors may be at work in separate areas of the same MB neuropil or in different MB neurons or neuropils, which are differentially innervated by dopaminergic axons conveying electric shock input or by dopaminergic and octopaminergic axons conveying sugar input. At present, there is limited information on intracellular components involved in appetitive learning. Future studies in this venue will help attest this model. Together, concurrent CS+ and US received during training may activate dDA1 (for punishment US) or dDA1 and an octopamine receptor (for reward US) to induce distinctive biochemical changes, leading to avoidance or preference behavior, respectively (Kim, 2007b).

Multiple lines of evidence indicate that dopamine in the amygdala, the nucleus accumbens, and the medial prefrontal cortex in mammals is crucial for acquisition, expression, and/or extinction in aversive pavlovian conditioning (for review, see Pezze, 2004). However, the receptors mediating the functions of dopamine are unclear. Studies using D₁/D₅ dopamine receptor antagonist R(+)-7-chloro-8-hydroxy-3-methyl-1-phenyl-2,3,4,5-tetrahydro-1H-3-benzazepine hydrochloride (SCH23390) in rats suggest the significant roles of the D₁-type receptor in the amygdala and the nucleus accumbens during acquisition in fear conditioning and conditioned taste aversion (CTA), respectively (Guarraci, 1999; Fenu, 2001). However, the mice lacking D₁ or D₅ receptor show normal acquisition in fear conditioning (El-Ghundi, 2001; Holmes, 2001). Likewise, D₁-deficient mice show normal learning of CTA to salt (CS) paired with LiCl (US), although they do not develop CTA to sucrose (Cannon, 2005). The discrepant findings of the pharmacological and genetic studies may be attributable to either other receptor types affected by SCH23390 or compensatory adaptations in D₁ or D₅ knock-out mice. The studies reported in this study support the latter and clarify the indispensable role of D₁ receptor in aversive pavlovian conditioning. Additionally, pharmacological studies reveal the significant role of D₁-type receptors in appetitive pavlovian conditioning in mammals (Schroeder, 2000; Baker, 2003; Eyny, 2003; Dalley, 2005) and possibly in Aplysia (Reyes, 2005), although no information is available on D₁ or D₅ knock-out mice in this type of behavioral plasticity. Therefore, the studies described here elucidate, for the first time, the critical role of D₁ receptor in appetitive pavlovian conditioning (Kim, 2007b).

GENE STRUCTURE

cDNA clone length - 2538

Bases in 5' UTR - 365

Exons - 6

Bases in 3' UTR - 637

PROTEIN STRUCTURE

Amino Acids - 511

Structural Domains

Members of the superfamily of G-protein coupled receptors share significant similarities in sequence and transmembrane architecture. A Drosophila homologue of the mammalian dopamine receptor family has been isolated using a low stringency hybridization approach. The deduced amino acid sequence is approximately 70% homologous to the human D1/D5 receptors. When expressed in HEK 293 cells, the Drosophila receptor stimulates cAMP production in response to dopamine application. This effect was mimicked by SKF 38393, a specific D1 receptor agonist, but inhibited by dopaminergic antagonists such as butaclamol and flupentixol. In situ hybridization revealed that the Drosophila dopamine receptor is highly expressed in the somata of the optic lobes. This suggests that the receptor might be involved in the processing of visual information and/or visual learning in invertebrates (Gotzes, 1994).

Molecular cloning revealed the existence of at least five pharmacologically different dopamine receptors in vertebrates. Functionally, dopamine receptors either activate (D1-type), inhibit or do not interact with adenylate cyclase (D2-type). A recently cloned dopamine receptor from Drosophila melanogaster shares many structural and functional properties with vertebrate D1-type receptors but the pharmacological properties are very different. In contrast to most aminergic receptors, DmDop1 contains a long N-terminal extension. This study describes a deletion-mutagenesis approach to study whether the N-terminus of DmDop1 participates in receptor-ligand interactions. All mutants gave rise to functional receptors after heterologous expression in HEK 293 cells. The pharmacological properties, however, remained unchanged. A comparison of DNA and deduced amino acid sequences revealed that some Drosophila strains express a truncated version of the DmDop1 receptor (Gotzes, 1996).

A cDNA clone has been isolated from Drosophila displaying, within putative transmembrane domains, highest amino acid sequence homology (49%-53%) to members of the vertebrate dopamine D1-like receptor family. When expressed in either Sf9 or COS-7 cells, dDA1 did not bind the specific D1-like receptor antagonist [3H]SCH-23390 or numerous other dopaminergic, adrenergic or serotoninergic ligands with high affinity. However, like vertebrate dopamine D1-like receptors, dDA1 stimulated the accumulation of cAMP in response to DA. The dopaminergic rank order of potency (DA --> NE -->> 5-HT) and the lack of stimulation by other possible neurotransmitters (octopamine, tyramine, tryptamine) or DA metabolites (e.g. N-acetyl dopamine) found in Drosophila suggests that this receptor functionally belongs to the dopamine D1-like subfamily. Benzazepines, which characteristically bind to vertebrate dopamine D1-like receptors with high affinity, were relatively poor in stimulating dDA1-mediated accumulation of cAMP. Of the numerous compounds tested, a few dopaminergic antagonists inhibited DA-stimulated production of cAMP in a concentration-dependent manner, albeit with considerably reduced affinity. In situ hybridization revealed that dDA1 receptor mRNA is expressed as a maternal transcript, and at later blastoderm stages is restricted to apical regions of the cortical peripheral cytoplasm. The generation of inter-species D1 receptor chimeras may help to identify those particular sequence-specific motifs or amino acid residues conferring high affinity benzaepine receptor interactions (Sugamori, 1995; full text of article).

date revised: 22 September 2007

Home page: The Interactive Fly © 2006 Thomas Brody, Ph.D.