Stem cells have the remarkable ability to give rise to both self-renewing and differentiating daughter cells. Drosophila neural stem cells segregate cell-fate determinants from the self-renewing cell to the differentiating daughter at each division. This study shows that one such determinant, the homeodomain transcription factor Prospero, regulates the choice between stem cell self-renewal and differentiation. The in vivo targets of Prospero have been identified throughout the entire genome. Prospero represses genes required for self-renewal, such as stem cell fate genes and cell-cycle genes. Surprisingly, Prospero is also required to activate genes for terminal differentiation. In the absence of Prospero, differentiating daughters revert to a stem cell-like fate: they express markers of self-renewal, exhibit increased proliferation, and fail to differentiate. These results define a blueprint for the transition from stem cell self-renewal to terminal differentiation (Choksi, 2006).
To identify sites within the Drosophila genome to which Prospero binds, use was made of an in vivo binding-site profiling technique, DamID. DamID is an established method of determining the binding sites of DNA- or chromatin-associated proteins. Target sites identified by DamID have been shown to match targets identified by chromatin immunoprecipitation (ChIP). DamID enables binding sites to be tagged in vivo and later identified on DNA microarrays. In brief, the DNA or chromatin-binding protein of interest is fused to an Escherichia coli adenine methyltransferase (Dam), and the fusion protein is expressed in vivo. The DNA-binding protein targets the fusion protein to its native binding sites, and the Dam methylates local adenine residues in the sequence GATC. The sequences near the protein-DNA interaction site are thereby marked with a unique methylation tag, over approximately 2–5 kilobase pairs (kb) from the binding site. The tagged sequences can be isolated after digestion with a methylation-sensitive restriction enzyme, such as DpnI (Choksi, 2006).
Dam was fused to the N terminus of Prospero, and transgenic flies were generated. The fusion protein is expressed from the uninduced minimal Hsp70 promoter of the UAS vector, pUAST, as high levels of expression of Dam can result in extensive nonspecific methylation and cell death. As a control for nonspecific Dam activity, animals expressing Dam alone were generated. To assess the sites to which Prospero binds during neurogenesis, genomic DNA was extracted from stage 10–11 embryos, approximately 4–7 hr after egg laying (AEL), expressing either the Dam-Prospero fusion protein or the Dam protein alone. The DNA was digested with DpnI and amplified by PCR. DNA from Dam-Prospero embryos was labeled with Cy3, and control DNA with Cy5. The samples were then cohybridized to genomic microarrays. Microarrays were designed that tile the entire euchromatic Drosophila genome. A 60 base oligonucleotide was printed for approximately every 300 bp of genomic DNA, resulting in roughly 375,000 probes on a single array (Choksi, 2006).
Log-transformed ratios from four biological replicates (two standard dye configurations plus two swapped dye configurations) were normalized and averaged. Regions of the genome with a greater than 1.4-fold log ratio (corresponding to approximately a 2.6-fold enrichment) of Dam-Prospero to the control over a minimum of four adjacent genomic probes were identified as in vivo Prospero binding sites. Using these parameters, a total of 1,602 in vivo Prospero binding sites were identified in the Drosophila genome. This work demonstrates that it is possible to map in vivo binding sites across the whole genome of a multicellular organism (Choksi, 2006).
Prospero is known to regulate the differentiation of photoreceptors in the adult eye, and recently sites have been characterized to which Prospero can bind upstream of two Rhodopsin genes, Rh5 and Rh6. A variant of the Prospero consensus sequence is found four times upstream of Rh5 and four times upstream of Rh6. Prospero was shown to bind this sequence in vitro, by band shift assay, and also by a 1-hybrid interaction assay in yeast. In addition, deletion analysis demonstrated that the consensus sequence is required for the Pros-DNA interaction both in vivo and in vitro. It was found that 67% of in vivo binding sites contain at least one Prospero binding motif. Combining in vivo binding-site data with searches for the Prospero consensus sequence reveals 1,066 distinct sites within the Drosophila genome to which Prospero binds during embryogenesis (Choksi, 2006).
A total of 730 genes have one or more of the 1,066 Prospero binding sites located within 1 kb of their transcription unit. Statistical analyses to determine GO annotation enrichment on the members of the gene list that had some associated annotation (519) was performed by using a web-based set of tools, GOToolbox. Using Biological Process (GO: 0008150) as the broadest classification, a list was generated of overrepresented classes of genes (Choksi, 2006).
The three most significant classes of genes enriched in the list of putative Prospero targets are Cell Fate Commitment, Nervous System Development, and Regulation of Transcription. Utilizing GO annotation, it was found that nearly 41% of all annotated neuroblast fate genes (11 of 27) are located near Prospero binding sites and that approximately 9% of known cell-cycle genes are near Prospero binding sites. These include the neuroblast genes achaete (ac), scute (sc), asense (ase), aPKC, and mira and the cell-cycle regulators stg and CycE. In addition, it was found that the Drosophila homolog of the mammalian B lymphoma Mo-MLV insertion region 1 (Bmi-1) gene, Posterior sex combs, is located near a Prospero binding site. Bmi-1 is a transcription factor that has been shown to regulate the self-renewal of vertebrate hematopoetic stem cells. It is concluded that Prospero is likely to regulate neuroblast identity and self-renewal genes as well as cell-cycle genes directly, repressing their expression in the GMC (Choksi, 2006).
Prospero enters the nucleus of GMCs, and its expression is maintained in glial cells but not in neurons . Therefore the list of targets was searched for genes annotated as glial development genes. Prospero binds near 45% of genes involved in gliogenesis. Among the glial genes, it was found that the master regulator of glial development, glial cells missing (gcm), and gilgamesh (gish), a gene involved in glial cell migration, are both near Prospero binding sites and are likely directly activated by Prospero in glia (Choksi, 2006).
In summary, Prospero binds near, and is likely to regulate directly, genes required for the self-renewing neural stem cell fate such as cell-cycle genes. It was also found that Prospero binds near most of the temporal cascade genes: hb, Kruppel (Kr), nubbin (nub/pdm1), and grainyhead (grh) and to genes required for glial cell fate. The in vivo binding-site mapping experiments are supportive of a role for Prospero in regulating the fate of Drosophila neural precursors by directly controlling their mitotic potential and capacity to self-renew (Choksi, 2006).
The Drosophila ventral nerve cord develops in layers, in a manner analogous to the mammalian cortex. The deepest (most dorsal) layer of the VNC comprises the mature neurons, while the superficial layer (most ventral) is made up of the mitotically active, self-renewing neuroblasts. Neuroblast cell-fate genes and cell-cycle genes are normally expressed only in the most ventral cells, while Prospero is found in the nucleus of the more dorsally lying GMCs. If in GMCs, Prospero normally acts to repress neuroblast cell-fate genes and cell-cycle genes, then in a prospero mutant, expression of those genes should expand dorsally. Conversely, ectopically expressed Prospero should repress gene expression in the neuroblast layer.
The neuroblast genes mira, ase, and insc and the cell cycle genes CycE and stg show little or no expression in differentiated cells of wild-type stage 14 nerve cords. Expression of these neuroblast-specific genes was examined in the differentiated cells layer of prospero embryos and it was found that they are derepressed throughout the nerve cord of mutant embryos. mira, ase, insc, CycE, and stg are all ectopically expressed deep into the normally differentiated cell layer of the VNC. To check whether Prospero is sufficient to repress these genes, Prospero was expressed with the sca-GAL4 driver, forcing Prospero into the nucleus of neuroblasts. Prospero expression is sufficient to repress mira, ase, insc, CycE, and stg in the undifferentiated cell layer of the VNC. These data, combined with the Prospero binding-site data, demonstrate that Prospero is both necessary and sufficient to directly repress neuroblast genes and cell-cycle genes in differentiated cells. This direct repression of gene expression is one mechanism by which Prospero initiates the differentiation of neural stem cells (Choksi, 2006).
Having shown that Prospero directly represses genes required for neural stem cell fate, it was asked whether Prospero also directly activates GMC-specific genes. Alternatively, Prospero might regulate a second tier of transcription factors, which are themselves responsible for the GMC fate. Of the few previously characterized GMC genes, it was found that Prospero binds to eve and fushi-tarazu (ftz). In the list of targets several more GMC genes were expected to be found, but not genes involved in neuronal differentiation, since Prospero is not expressed in neurons. Surprisingly, however, it was foudn 18.8% of neuronal differentiation genes are located near Prospero binding sites (Choksi, 2006).
To determine Prospero's role in regulating these neuronal differentiation genes, in situ hybridization was carried out on prospero mutant embryos. Prospero was found to be necessary for the expression of a subset of differentiation genes, such as the adhesion molecules FasciclinI (FasI) and FasciclinII (FasII), which have roles in axon guidance and/or fasciculation. Netrin-B, a secreted protein that guides axon outgrowth, and Encore, a negative regulator of mitosis, also both require Prospero for proper expression. Therefore, in addition to directly repressing genes required for neural stem cell self-renewal, Prospero binds and activates genes that direct differentiation. These data suggest that Prospero is a binary switch between the neural stem cell fate and the terminally differentiated neuronal fate (Choksi, 2006).
To test to what extent Prospero regulates the genes to which it binds, genome-wide expression profiling was carried out on wild-type and prospero mutant embryos. While the DamID approach identifies Prospero targets in all tissues of the embryo, in this instance genes regulated by Prospero were assayed in the developing central nervous system. Small groups of neural stem cells and their progeny (on the order of 100 cells) were isolated from the ventral nerve cords of living late stage 12 embryos with a glass capillary. The cells were expelled into lysis buffer, and cDNA libraries generated by reverse transcription and PCR amplification. cDNA libraries prepared from neural cells from six wild-type and six prospero null mutant embryos were hybridized to full genome oligonucleotide microarrays, together with a common reference sample. Wild-type and prospero mutant cells were compared indirectly through the common reference (Choksi, 2006).
In the group of Prospero target genes that contain a Prospero consensus sequence within 1 kb of the transcription unit, 91 show reproducible differences in gene expression in prospero mutants. Seventy-nine percent of these genes (72) exhibit at least a 2-fold change in levels of expression. Many of the known genes involved in neuroblast fate determination and cell-cycle regulation (e.g., asense, deadpan, miranda, inscuteable, CyclinE, and string) show increased levels in a prospero mutant background, consistent with their being repressed by Prospero. Genes to which Prospero binds, but which do not contain an obvious consensus sequence, are also regulated by Prospero: CyclinA and Bazooka show elevated mRNA levels in the absence of Prospero, as does Staufen, which encodes a dsRNA binding protein that binds to both Miranda and to prospero mRNA (Choksi, 2006).
Expression of genes required for neuronal differentiation is decreased in the prospero mutant cells, consistent with Prospero being required for their transcription. These include zfh1 and Lim1, which specify neuronal subtypes, and FasI and FasII, which regulate axon fasciculation and path finding (Choksi, 2006).
The stem cell-like division of neuroblasts generates two daughters: a self-renewing neuroblast and a differentiating GMC. Prospero represses stem cell self-renewal genes and activates differentiation genes in the newly born GMC. In the absence of prospero, therefore, neuroblasts should give rise to two self-renewing neuroblast-like cells (Choksi, 2006).
The division pattern of individual neuroblasts was studied in prospero mutant embryos by labeling with the lipophilic dye, DiI. Individual cells were labeled at stage 6, and the embryos allowed to develop until stage 17. S1 or S2 neuroblasts were examined, as determined by their time of delamination. Wild-type neuroblasts generate between 2 and 32 cells, producing an average of 16.2 cells. Most of the clones exhibit extensive axonal outgrowth. In contrast, prospero mutant neuroblasts generate between 8 and 51 cells, producing an average of 31.8 cells. Moreover, prospero mutant neural clones exhibit few if any projections, and the cells are smaller in size. Thus, prospero mutant neuroblasts produce much larger clones of cells with no axonal projections, suggesting that neural cells in prospero mutants undergo extra divisions and fail to differentiate (Choksi, 2006).
Recently it was shown, in the larval brain, that clones of cells lacking Prospero or Brat undergo extensive cell division to generate undifferentiated tumors. Given that Prospero is nuclear in the GMC but not in neuroblasts, the expanded neuroblast clones in prospero mutant embryos might arise from the overproliferation of GMCs: the GMCs lacking Prospero may divide like neuroblasts in a self-renewing manner. It is also possible, however, that neuroblasts divide more frequently in prospero mutant embryos, giving rise to supernumerary GMCs that each divide only once. To distinguish between these two possibilities, the division pattern of individual GMCs was followed in prospero mutant embryos (Choksi, 2006).
S1 or S2 neuroblasts were labeled with DiI as before. After the first cell division of each neuroblast, the neuroblast was mechanically ablated, leaving its first-born GMC. All further labeled progeny derive, therefore, from the GMC. Embryos were allowed to develop until stage 17, at which time the number of cells generated by a single GMC was determined (Choksi, 2006).
To determine whether mutant GMCs are transformed to a stem cell-like state, stage 14 embryos were stained for the three neuroblast markers: Miranda (Mira), Worniu (Wor), and Deadpan (Dpn). In wild-type embryos at stage 14, the most dorsal layer of cells in the VNC consists mostly of differentiated neurons. As a result, few or none of the cells in this layer express markers of self-renewal. Mira-, Wor-, and Dpn- expressing cells are found on the midline only or in lateral neuroblasts of the differentiated cell layer of wild-type nerve cords. In contrast, a majority of cells in the differentiated cell layer of stage 14 prospero mutant embryos express all three markers: Mira is found cortically localized in most cells of the dorsal layer of prospero nerve cords; Wor is nuclear in most cells of mutant VNCs; Dpn is ectopically expressed throughout the nerve cord of prospero mutants (Choksi, 2006).
Expression of neuroblast markers in the ventral-most layer of the nerve cord (the neuroblast layer), to exclude the possibility that a general disorganization of cells within the VNC contributes to the increased number of Mira-, Wor-, and Dpn-positive cells in the dorsal layer. The number of neuroblasts in a prospero mutant embryo is normal in stage 14 embryos, as assayed by Wor, Dpn, and Mira expression. Thus, the increased expression of neuroblast markers in prospero mutants is the result of an increase in the total number of cells expressing these markers in the differentiated cell layer. It is concluded that prospero mutant neuroblasts divide to give two stem cell-like daughters. GMCs, which would normally terminate cell division and differentiate, are transformed into self-renewing neural stem cells that generate undifferentiated clones or tumors (Choksi, 2006).
Therefore, Prospero directly represses the transcription of many neuroblast genes and binds near most of the temporal cascade genes: hb, Kruppel (Kr), nubbin (nub/pdm1), and grainyhead (grh), which regulate the timing of cell-fate specification in neuroblast progeny. Prospero maintains hb expression in the GMC, and it has been suggested that this is through regulation of another gene, seven-up (svp). Prospero not only regulates svp expression directly but also maintains hb expression directly. In addition, Prospero maintains Kr expression and is likely to act in a similar fashion on other genes of the temporal cascade. Intriguingly, Prospero regulates several of the genes that direct asymmetric neuroblast division (baz, mira, insc, aPKC). aPKC has recently been shown to be involved in maintaining the self-renewing state of neuroblasts (Choksi, 2006).
Prospero initiates the expression of genes necessary for differentiation. This is particularly surprising since prospero is transcribed only in neuroblasts, not in GMCs or neurons. Prospero mRNA and protein are then segregated to the GMC. Prospero binds near eve and ftz, which have been shown to be downstream of Prospero, as well as to genes required for terminal neuronal differentiation, including the neural-cell-adhesion molecules FasI and FasII. Prospero protein is present in GMCs and not neurons, suggesting that Prospero initiates activation of neuronal genes in the GMC. The GMC may be a transition state between the neural stem cell and the differentiated neuron, providing a window during which Prospero functions to repress stem cell-specific genes and activate genes required for differentiation. There may be few, or no, genes exclusively expressed in GMCs (Choksi, 2006).
Prospero acts in a context-dependent manner, functioning alternately to repress or activate transcription. This implies that there are cofactors and/or chromatin remodeling factors that modulate Prospero's activity. In support of this, although Prospero is necessary and sufficient to repress neuroblast genes, forcing Prospero into the nuclei of neuroblasts is not sufficient to activate all of the differentiation genes to which it binds (Choksi, 2006).
Neuroblasts decrease in size with each division throughout embryogenesis. By the end of embryogenesis, they are similar in size to neurons. A subset of these embryonic neuroblasts becomes quiescent and is reactivated during larval life: they enlarge and resume stem cell divisions to generate the adult nervous system. Neuroblasts in prospero mutant embryos divide to produce two self-renewing daughters but still divide asymmetrically with respect to size, producing a large apical neuroblast and a smaller basal neuroblast-like cell. The daughter may be too small to undergo more than three additional rounds of division during embryogenesis. prospero mutant cells eventually stop dividing, and a small number occasionally differentiate. This suggests that there is an inherent size limitation on cell division. The segregation of Brat, or an additional cell fate determinant, to the daughter cell may also limit the potential of the prospero mutant cells to keep dividing (Choksi, 2006).
The Prox family of atypical homeodomain transcription factors has been implicated in initiating the differentiation of progenitor cells in contexts as varied as the vertebrate retina, forebrain, and lymphatic system. Prospero/Prox generally regulates the transition from a multipotent, mitotically active precursor to a differentiated, postmitotic cell. In most contexts, Prox1 acts in a similar fashion to Drosophila Prospero: to stop division and initiate differentiation (Choksi, 2006).
It is proposed that Prospero/Prox is a master regulator of the differentiation of progenitor cells. Many of the vertebrate homologs of the Drosophila Prospero targets identified in this study may also be targets of Prox1 in other developmental contexts. Prospero directly regulates several genes required for cell-cycle progression, and it is possible that Prox1 will regulate a similar set of cell-cycle genes during, for example, vertebrate retinal development. In addition, numerous Prospero target genes have been identified whose orthologs may be involved in the Prox-dependent differentiation of retina, lens, and forebrain precursors (Choksi, 2006).
Synaptic stimulation activates signal transduction pathways, producing persistently active protein kinases. PKMzeta is a truncated, persistently active isoform of atypical protein kinase C-zeta (aPKCzeta), which lacks the N-terminal pseudosubstrate regulatory domain. Using a Pavlovian olfactory learning task in Drosophila, it was found that induction of the mouse aPKMzeta (MaPKMzeta) transgene enhances memory. The enhancement requires persistent kinase activity and is temporally specific, with optimal induction at 30 minutes after training. Induction also enhances memory after massed training and corrects the memory defect of radish mutants, but does not improve memory produced by spaced training. The 'M' isoform of the Drosophila homolog of MaPKCzeta (DaPKM) is present and active in fly heads. Chelerythrine, an inhibitor of PKMzeta, and the induction of a dominant-negative MaPKMzeta transgene inhibits memory without affecting learning. Finally, induction of DaPKM after training also enhances memory. These results show that atypical PKM is sufficient to enhance memory in Drosophila and suggest that it is necessary for normal memory maintenance (Drier, 2002).
The study of PKC in memory formation has a long history. However, most previous work was done before the current appreciation of the complexity of the PKC gene family. The PKC family can be divided into three classes based on their cofactor requirements. Whereas all PKC proteins require phosphatidylserine for activation, the 'conventional' (cPKC) isotypes require diacylglycerol (DAG) and Ca2+ for full activity; 'novel' (nPKC) isotypes are Ca2+ independent but still require DAG, and the 'atypical' (aPKC) isotypes are both DAG and Ca2+ independent. Structurally, these kinases can be divided into an N-terminal regulatory domain, which contains a pseudosubstrate region as well as the binding sites for the required cofactors, and the C-terminal catalytic domain. Removal of the N-terminal regulatory domain produces a persistently active kinase, referred to as PKM. Persistently active kinases have received attention as components of memory mechanisms (Drier, 2002).
The roles of PKC in hippocampal models of synaptic plasticity, long-term potentiation (LTP) and long-term depression (LTD) have been studied extensively. PKC/M activities may have several roles in the mechanisms that initiate and sustain LTP. However, Western blot analyses with antibodies specific for each of the rat PKC isoforms demonstrate that the only one whose levels specifically increase and remain elevated during the maintenance phase of LTP is PKMzeta, which is the truncated form of the atypical isozyme PKCzeta. Expression analyses also show that the maintenance of LTD is associated with decreasing levels of PKMzeta. Most interestingly, LTP maintenance is abolished by sustained application of low concentrations of the PKC inhibitor chelerythrine, whereas perfusion of PKMzeta into CA1 pyramidal cells produces an increase in AMPA receptor-mediated synaptic transmission (Drier, 2002 and references therein).
Experiments in honeybees also indicate a role for PKC in memory formation. Biochemical analyses of extracts made from the antennal lobes of associatively trained bees show a sustained increase in cytosolic, Ca2+-independent PKC activity. This persistent increase correlates with long-lasting bee memory in four ways: it requires multiple training trials; it persists for up to three days; it is insensitive to a drug that blocks cPKC activity, and it is blocked by protein-synthesis inhibitors. Together with the LTP data, these studies point to an important role for a nonconventional PKC activity in the maintenance of memory (Drier, 2002).
In Drosophila, the best characterized assay for associative learning and memory is an odor-avoidance behavioral task. This classical (Pavlovian) conditioning involves exposing the flies to two odors (the conditioned stimuli, or CS), one at a time, in succession. During one of these odor exposures (the CS+), the flies are simultaneously subjected to electric shock (the unconditioned stimulus, or US), whereas exposure to the other odor (the CS-) lacks this negative reinforcement. After training, the flies are placed at a 'choice point', where the odors come from opposite directions, and they decide which odor to avoid. By convention, learning is defined as the fly's performance when testing occurs immediately after training. A single training trial produces strong learning: a typical response is that >90% of the flies avoid the CS+. Performance of wild-type flies from this single-cycle training decays over a roughly 24-hour period until flies once again distribute evenly between the two odors. Flies can also form long-lasting associative olfactory memories, but normally this requires repetitive training regimens (Drier, 2002).
This task was used in Drosophila to examine the role of atypical PKM in memory formation. Induction of the mouse aPKMzeta (MaPKMzeta) transgene enhances memory, and corrects the memory defect of radish mutants. There is a single atypical PKC in Drosophila, and the truncated 'M' isoform, DaPKM, is preferentially expressed and active in fly heads. Both pharmacological and dominant-negative genetic intervention of DaPKC/M activity disrupts normal memory. Finally, induction of the predicted DaPKM also enhances memory, further suggesting a general role of aPKM in memory processes (Drier, 2002).
To investigate the role of PKC in learning and memory in Drosophila, transgenic lines of flies were made bearing heat shock-inducible, murine atypical PKC (MaPKC) isoforms. Considering that LTP experiments indicate that MaPKMzeta levels increase after the presentation of the stimuli required for long-lasting potentiation, whether inducing MaPKMzeta after training affected olfactory memory was tesed. Induction by mild heat shock (32°C) after training strongly enhances 24-hour memory. This enhancement is not due to transgene-independent heat-shock effects, because the wild-type flies show no enhanced memory when exposed to heat shock. The transgenic flies were made in this wild-type strain, so the enhancement is not due to differences in genetic background. Finally, the memory enhancement does not result from an insertional mutation caused by the transgene, because two independent lines (MaPKMzeta-14 and MaPKMzeta-43) have similar effects (Drier, 2002).
Whether 24-hour memory could be enhanced after single-cycle training was tested by inducing MaPKMzeta with a strong heat shock (37°C) 3 hours before training, but this regimen has no effect. Because transgene induction after behavioral training enhances memory, whereas induction before training does not, the temporal specificity of this MaPKMzeta-dependent effect was examined. Optimal enhancement occurs when heat-shock induction begins 30 minutes after training ends, and the effect is absent if heat shock occurs before, or is delayed until 2 hours after training (Drier, 2002).
The memory enhancement is not observed when a kinase-inactive (KI) mutant of MaPKMzeta is induced either before or after training. The enhancement is also not observed when full-length (FL)-MaPKCzeta is induced before or after training. The failure of either the KI-MaPKMzeta or the FL-MaPKCzeta transgene to enhance memory is not due to lack of expression, because both are expressed at levels comparable to the MaPKMzeta protein. Together, these results indicate that the memory enhancement requires a persistently active aPKM isoform (Drier, 2002).
Inducible increases in MaPKMzeta protein levels and kinase activity have been detected in extracts made from Drosophila heads. Western blot analyses shows that both the mild and strong heat-shock regimens induce the MaPKMzeta and MaPKCzeta isoforms, and that these proteins persisted for ~18 hours after heat shock. The induced MaPKMzeta protein is active, since enhancement of Ca2+/DAG-independent PKC activity is observed in fly head extracts from induced but not from uninduced transgenic flies (Drier, 2002).
Drosophila can form associative olfactory memories lasting 24 hours and longer, but this normally requires repetitive training. Multiple-trial training regimens have been established that produce both anesthesia-resistant memory (ARM) and long-term memory (LTM). ARM can be produced by 10 cycles of 'massed' training with no rest intervals between the individual training trials, and lasts 2-3 days. LTM results from repetitive training that contains rest intervals (15 min each), and 10 cycles of this 'spaced' training generates LTM that lasts at least 7 days. To test whether MaPKMzeta can enhance ARM or LTM, flies were subjected to massed or spaced training regimens; the transgene was induced for 30 minutes after training, and then 4-day memory was measured (Drier, 2002).
MaPKMzeta induction substantially increases 4-day memory after massed training but does not improve 4-day memory after spaced training. These data indicate that MaPKMzeta induction enhances massed training-induced, but not spaced training-induced memory (Drier, 2002).
Previous work indicates that consolidated memory in Drosophila consists of two biochemically separable components: ARM and LTM. ARM is produced by either massed or spaced training, and it is insensitive to cycloheximide treatment. LTM is produced by spaced training and is blocked by cycloheximide treatment; thus it is considered to require acute protein synthesis. A previously identified Drosophila memory mutant, radish, is deficient in ARM, since this mutation blocks memory produced by massed training. Spaced training of radish mutants does produce memory, but this memory can be completely blocked by treating the mutants with cycloheximide. These results led to a two-pathway model of consolidated memory, one dependent on the Radish gene product (ARM) and the other dependent on activity-induced, acute protein synthesis (LTM) (Drier, 2002).
Because MaPKMzeta induction enhances memory after massed but not after spaced training, the dependence of this effect on radish was tested. The radish gene is on the X chromosome in Drosophila, and homozygous radish mutant females were crossed to males homozygous for an autosomal copy of the heat shock-inducible MaPKMzeta transgene. The radish mutant is recessive, thus the heterozygous female progeny of this mating will have normal memory after massed training, whereas the hemizygous males will display the radish memory deficit in the absence of induction. The progeny were subjected to massed training, followed by the standard MaPKMzeta induction after training, and then tested at 24 hours. Males and females were trained and tested en masse, and then separated and counted. The radish mutation did not block the memory effect of MaPKMzeta induction. The memory defect of radish males was apparent in the absence of heat-shock induction (HS-), but memory was clearly present in induced males (HS+). A lesser, but significant induction-dependent memory enhancement of the heterozygous radish females by MaPKMzeta was also observed (Drier, 2002).
There is a single atypical PKC (DaPKC) gene in the Drosophila genome, and it is highly homologous to the MaPKCzeta gene that was used. (The kinase domain shows 76% identity and 87% similarity). A Western blot of extracts made from wild-type fly heads and bodies shows that the antiserum used to detect MaPKMzeta and MaPKCzeta recognizes two bands in fly extracts, the smaller of which is enriched in head extracts. This antiserum is directed against the C-terminal 16 amino acids of MaPKC/Mzeta, which shares substantial homology with DaPKC. Antiserum from mice immunized with peptides derived from DaPKC recognizes these same bands. The molecular weights of these two bands indicate that they are probably the DaPKC (~73 kDa) and DaPKM (~55 kDa) isoforms (Drier, 2002).
The N-terminal sequence of the lower molecular weight band has not been established; however, it likely represents an endogenous DaPKM isoform. The immunoreactivity is competitively reduced by a peptide from the corresponding region of DaPKC, but not one outside of this epitope. In agreement with the Western blot data, fly heads contains more Ca2+ and DAG-independent PKC activity than does bodies. The presence of the putative DaPKM correlates strongly with this enriched activity, suggesting that most, if not all, of the endogenous atypical kinase activity measured in head extracts is due to this DaPKM isoform. These data indicate that flies possess both 'C' and 'M' forms of an atypical PKC that is highly homologous to MaPKC/M, and that the DaPKM is enriched in heads (Drier, 2002).
A P-element insertional mutant in DaPKC has been described; however, it is an embryonic lethal and thus is not suitable for examining a possible role in adult learning and memory formation. To assess whether this gene's product is necessary for memory formation, two approaches were taken. First, the effects on memory of feeding flies the PKC inhibitor chelerythrine were monitored. This drug is reported to selectively inhibit PKMzeta at low concentrations; however, its specificity is controversial, and it inhibits other PKC isotypes at higher concentrations. Memory effects produced by inducing the kinase-inactive KI-MaPKMzeta protein were tested: this form of the protein displays 'dominant-negative' activity that is likely to be specific to the atypical PKCs, leaving cPKC and nPKC responses intact (Drier, 2002).
Feeding flies chelerythrine inhibits 24-hour memory formation in a dose-dependent manner, and induction of the KI-MaPKMzeta inhibits 24-hour memory after massed training. The inhibitory effects of both chelerythrine and the KI-MaPKMzeta are not likely due to effects on olfactory acuity or shock reactivity because learning is unaffected by either treatment (Drier, 2002).
The memory enhancement produced by MaPKMzeta could have been due to properties unique to this mammalian protein. The expression data showing that DaPKM is expressed and active in Drosophila heads, when combined with the chelerythrine and dominant-negative data, suggests that DaPKM is involved in normal memory processes in Drosophila. The extensive structural homology between MaPKMzeta and DaPKM also argues against functional uniqueness. The hypothesis of functional homology makes a strong prediction: induction of DaPKM after training should also enhance memory (Drier, 2002).
Based on the approximate molecular weight of the DaPKM, the DaPKC gene was truncated within the hinge region separating the regulatory from the catalytic domains such that the putative DaPKM gene begins at methionine 223. Induction of the DaPKM transgene after training enhances 24-hour memory after single-cycle training. One of these lines was then used to show that 4-day memory after massed training is also enhanced. As with the MaPKMzeta transgenes, the DaPKM lines shows rapid heat-shock induction. These results confirm those obtained with MaPKMzeta, and thus indicate that aPKM is fundamental in the mechanisms underlying memory across species (Drier, 2002).
These results provide strong evidence that atypical PKM activity is sufficient to enhance memory in Drosophila. Ideally, necessity should have been tested by assessing potential memory deficits of flies bearing null mutations in the DaPKC/M gene. However, the lethality of such mutants precluded these analyses, and no special alleles exist that might have preserved the gene's vital function while disrupting its role in memory. In an attempt to circumvent these problems, both pharmacological and dominant-negative interventions were used. Chelerythrine inhibits normal memory in a dose-dependent manner, and induction of a predicted dominant-negative atypical PKM produces the same memory deficit (Drier, 2002).
It was found that heat-shock induction of MaPKMzeta does not enhance long-term memory, because it does not improve memory after spaced training. One explanation for this is that spaced training induces endogenous maintenance mechanisms, and thus occludes the effect of inducing the MaPKMzeta transgene. Thus, memory after single-cycle or massed training may be prolonged by transgene induction because these training regimens do not normally induce prolonged atypical PKM activity. Work in honeybees shows that single-cycle training produces neither persistent PKC activity nor long-lasting memory, but multiple-cycle training produces both. The memory enhancement observed when inducing MaPKMzeta may simply bypass the endogenous requirements (normally provided by spaced training) for prolonged activation of aPKM (Drier, 2002).
The MaPKMzeta-induced enhancement of massed, but not spaced training prompted an examination of the involvement of the radish gene product in this process. If radish were required for the enhancement, the radish mutation would have blocked the MaPKMzeta-induced effect, and this was clearly not the case. Although MaPKMzeta induction phenotypically rescues the memory defect of radish, it does not do so because radish encodes for the Drosophila aPKM. DaPKM is on the second chromosome and radish is on the X, and no Drosophila PKC gene maps to the genetically defined radish locus. There are two principal possibilities explaining how MaPKMzeta-induced memory enhancement bypasses the defect of radish mutants: (1) MaPKMzeta is downstream of radish or (2) MaPKMzeta activates a pathway that is parallel to and independent of radish. The first interpretation is favored because memory after massed training can be either enhanced or disrupted and the radish phenotype can be partially rescued (Drier, 2002).
The temporal specificity of the MaPKMzeta-dependent memory enhancement implies restrictions on its biochemical mechanism(s); enhancement requires that prior activity-dependent mechanisms be in place, and MaPKMzeta has a narrow post-training interval in which to act. If these kinetic restrictions do exist, the rapid induction achievable with the heat-shock promoter is essential for the detection of memory enhancement in these experiments (Drier, 2002).
There are two general interpretations of these data: PKMzeta acts to increase either (1) the magnitude or (2) the duration of the synaptic potentiation that underlies the behavior. In the first model, PKMzeta enhances the synaptic machinery induced by training, making a 'stronger' synaptic connection that decays more slowly. In the second model, PKMzeta acts solely to maintain the synapses previously modified by experience, with no effect on the induction of the potentiation. If one considers the behavioral measurements of learning (testing done immediately after training) and memory (testing done after a longer time) with induction and maintenance, respectively, the chelerythrine and dominant-negative data argue for a role in maintenance. Neither of these treatments affect learning, but each inhibits memory. No enhancement of learning was detected by prior induction of PKMzeta, nor was there an improvement of 3-hour memory if PKMzeta was induced 30 minutes after training. Although the magnitude and duration models may be artificially exclusive, taken together these data are most consistent with a role of PKMzeta in the maintenance of experience-dependent synaptic plasticity (Drier, 2002).
The stability of a synapse varies in response to different regimens of stimuli. Long-lasting changes normally require multiple stimuli and depend on new protein synthesis. Recent experiments support the existence of a synaptic marking system that enables neurons to tag recently active synapses, thus maintaining synaptic specificity during the cell-wide process of protein synthesis-dependent long-term memory formation. A synapse that would normally be stable for only a short period of time can be potentiated for a much longer period of time. However, to do so it must be activated within 2-4 hours of stimulation that produces long-term changes at a second and separate synapse within the same neuron. Although there is no direct evidence for a role of PKMzeta in this process, the similarity between the temporal windows for the proposed synaptic tag and the memory enhancement observed suggests a mechanistic relationship between them (Drier, 2002).
DaPKC is part of a multiprotein complex important for both cell polarity and the asymmetrical cell divisions of early Drosophila neurogenesis. These processes show strong structural and functional parallels with the first asymmetrical cell division of Caenorhabditis elegans embryogenesis. The Drosophila homologs of C. elegans proteins important for this process, Par-3 (Bazooka) and Par-6 (DmPar-6), interact with each other and with DaPKC to direct a specific and interdependent subcellular localization of the complex. During early Drosophila embryogenesis, Bazooka, DmPar-6, and DaPKC are localized to the zonula adherens, a cell junction structure. Mutation in any one of these genes disrupts the ability of the remaining two proteins to localize to this structure properly, and this disrupts cell polarity. This mutual dependence for localization is also apparent during neurogenesis, and causes the inappropriate segregation of cell determinants. This multiprotein complex is critical in mammalian cell polarity and in organizing junctions between epithelial cells. The mouse homologs of Bazooka and Par-6 are expressed in various regions of the CNS, and their subcellular localization within CA1 hippocampal neurons is consistent with a role in synaptic plasticity. Bazooka and DmPar-6 are expressed in Drosophila heads, as are DaPKC and DaPKM. It remains unclear how DaPKM activity is regulated during memory mechanisms; however, the subcellular localization affected by the Bazooka-DmPar-6-DaPKC complex provides hypotheses with attractive physical properties (Drier, 2002).
Atypical PKM is sufficient to enhance memory in Drosophila, and the chelerythrine and dominant-negative data suggest that it is also necessary for normal memory. Strikingly corroborative results have also been obtained for the role of PKMzeta in the maintenance phase of LTP. Injection of MaPKMzeta into CA1 pyramidal cells is sufficient to potentiate evoked excitatory postsynaptic currents. The potentiation occludes LTP and is reversed by chelerythrine. The introduction of the KI-MaPKMzeta into a CA1 cell abolishes its ability to support LTP. The non-NMDA receptor antagonist CNQX blocks this potentiation, indicating that it occurs via AMPA receptors. When these physiological results, obtained in rat hippocampal slice preparations, are combined with the Drosophila behavioral data, they point to a central role of atypical PKM in the mechanism of memory maintenance. Understanding the regulation of atypical PKMzeta, as well as what it in turn regulates, may be critical to unraveling this process (Drier, 2002 and references therein).
In dividing Drosophila sensory organ precursor (SOP) cells, the fate determinant Numb and its associated adaptor protein Partner of numb (Pon) localize asymmetrically and segregate into the anterior daughter cell, where Numb influences cell fate by repressing Notch signaling. Asymmetric localization of both proteins requires the protein kinase aPKC and its substrate Lethal (2) giant larvae (Lgl). Because both Numb and Pon localization require actin and myosin, lateral transport along the cell cortex has been proposed as a possible mechanism for their asymmetric distribution. This study used quantitative live analysis of GFP-Pon and Numb-GFP fluorescence and fluorescence recovery after photobleaching (FRAP) to characterize the dynamics of Numb and Pon localization during SOP division. It was demonstrated that Numb and Pon rapidly exchange between a cytoplasmic pool and the cell cortex and that preferential recruitment from the cytoplasm is responsible for their asymmetric distribution during mitosis. Expression of a constitutively active form of aPKC impairs membrane recruitment of GFP-Pon. This defect can be rescued by coexpression of nonphosphorylatable Lgl, indicating that Lgl is the main target of aPKC. It is proposed that a high-affinity binding site is asymmetrically distributed by aPKC and Lgl and is responsible for asymmetric localization of cell-fate determinants during mitosis (Mayer, 2005).
In order to study the dynamics of asymmetric protein localization, a time series of the division of an SOP cell expressing GFP-Pon and Histone2B-RFP was recorded under the control of a specific promoter. Histone2B-RFP was used to visualize DNA, thus allowing correlation of distinct steps of GFP-Pon localization with other mitotic events. In interphase, some GFP-Pon is cortical, but a large part localizes to the cytoplasm. As the cell enters mitosis, it rounds up and undergoes strong membrane blebbings, indicative of local rearrangements of the cortical cytoskeleton. Interestingly, similar blebbing events have also been observed in the first division of the C. elegans zygote. Unlike in SOP cells, however, they only occur on the anterior side of the C. elegans zygote, where Par-3/6 localize. Shortly after blebbing has started, chromosomes condense and GFP-Pon accumulates on random sites of the cell cortex. The accumulations are transient and do not necessarily predict the position of the final Pon crescent. This suggests that the process leading to Pon accumulation can take place all around the cell but is reinforced specifically in the crescent region. Some GFP-Pon was also observed at the nucleus. This signal might be due to GFP-Pon binding to the nuclear envelope or to the endoplasmic reticulum, and it disappears slowly after nuclear-envelope breakdown. At nuclear-envelope breakdown, cortical blebbing ceases, the cell cortex smoothes, and first signs of asymmetric localization of GFP-Pon into an anterior cortical crescent are observed. As the cell progressed into metaphase, the GFP-Pon signal in the crescent area becomes stronger. Surprisingly, the intensity of the cortical area opposite of the crescent is almost not changed during this process. Thus, GFP-Pon might actually be recruited to the crescent directly from the cytoplasm rather than being transported along the cell cortex. Indeed, quantification of fluorescence intensity showed that GFP-Pon recruitment at the cell cortex is accompanied by a comparable loss of cytoplasmic GFP-Pon. Note that local degradation of GFP-Pon in the cytoplasm is not responsible for this reduction because total GFP-Pon remains unchanged (Mayer, 2005).
Subsequently, the metaphase plate was oriented with respect to the crescent, and during cytokinesis, GFP-Pon segregated largely into the anterior daughter cell. It is proposed that GFP-Pon localization is a two-step process involving the establishment of a cortical area where the crescent will form and the progressive recruitment of protein to the predefined site until metaphase (Mayer, 2005).
Asymmetry of Numb and Pon could be created by lateral movement along the cell cortex or by direct recruitment from the cytoplasm to one side of the cell cortex. To quantify the exchange of Numb and Pon between the cell cortex and the cytoplasm, fluorescence recovery after photobleaching (FRAP) was used of GFP fusions to Numb and Pon. Numb-GFP can partially rescue the numb mutant phenotype, indicating that it is functional. GFP-Pon contains just the asymmetric-localization domain. Its rescue behavior is unknown, but it colocalizes with endogenous Pon throughout mitosis. When cytoplasmic GFP-Pon is photobleached, fluorescence recovers with a half-time of 0.48 s, indicating that diffusion is not limiting. Recovery of cortical GFP-Pon fluorescence occurred with single exponential kinetics and a half-time of 35 s, whereas the half-time for Numb-GFP was 27 s. Therefore, Numb and Pon showed a surprisingly dynamic association with the cell cortex (Mayer, 2005).
Either cortical recruitment of cytoplasmic GFP-Pon or lateral diffusion/transport of cortical GFP-Pon could be responsible for fluorescence recovery. To measure the exchange between cortical and cytoplasmic Pon, an area covering approximately 40% of the cytoplasm was repeatedly photobleached in an SOP cell expressing GFP-Pon. Fluorescence intensity was simultaneously recorded at the cortex. Cortical fluorescence intensity dropped to less than 5% with a half-time of 52 s. Thus, the cortical and cytoplasmic pools of GFP-Pon rapidly interchange with a mobile fraction of more than 95% (Mayer, 2005).
When the dynamic association with the cell cortex is taken into account, Pon asymmetry could be explained either by fast and continuous lateral transport or by directed recruitment to an asymmetric cortical binding site. To determine the contribution of lateral transport, FRAP rates were compared on the edge and in the center of a photobleached region within the GFP-Pon crescent. The bleached region was defined such that a region of nonbleached molecules was left behind at the edges of the crescent after photobleaching. To avoid recovery from above and below the image plane, a protocol was used in which the region of interest was bleached in several planes. The efficiency of this procedure was confirmed by 3D reconstruction after photobleaching in fixed tissue. FRAP curves from ten experiments were averaged. Their superposition shows that the two regions recover nearly identically with half-times of 32 s for a region close to nonbleached GFP-Pon and of 35 s for a region farther away. Taken together, these observations suggest a model where Pon is preferentially recruited from the cytoplasm to the site of crescent formation. It is proposed that a cortical high-affinity binding site for Pon is established during mitosis and mediates specific recruitment of Pon to one side of the cell cortex (Mayer, 2005).
To test the role of Lgl in asymmetric protein localization in SOP cells, cortical recruitment of GFP-Pon was measured in lgl1 mutant clones. In a similar experiment, Lgl has been shown to be dispensable for Pon localization, although Pon recruitment seemed to be delayed. The ratio between total cortical and total cytoplasmic fluorescence was calculated. Because GFP fluorescence intensity is proportional to GFP-Pon concentration, this ratio should give a good estimate of the fraction of GFP-Pon localized at the cell cortex. Although GFP-Pon was still asymmetric, quantitative analysis revealed that the cortical GFP-Pon fraction was slightly but significantly reduced in lgl1 mutant clones. This might be a hypomorphic phenotype caused by small residual amounts of Lgl protein present in the mutant clones. Therefore expression of deregulated aPKC (aPKC-deltaN) was used as another means to inactivate Lgl. Expression of aPKC-deltaN was shown to phenocopy lgl mutants in embryonic tissues, presumably because it phosphorylates and inactivates Lgl all around the cell. In contrast to lgl1 mutant SOP cells, a much stronger decrease of cortical GFP-Pon recruitment was observed upon aPKC-deltaN expression. Still, a slight cortical asymmetry was observed, which is thought is due to the presence of endogenous aPKC. Even at anaphase, the degree of recruitment hardly reached that of control cells in prophase. To test whether Lgl phosphorylation was responsible for this phenotype, aPKC-deltaN was coexpressed with nonphosphorylatable lgl3A. Expression of lgl3A completely rescued the cortical-recruitment defect. The observed differences are not due to increased protein levels because total cellular GFP-Pon fluorescence remains constant (Mayer, 2005).
Thus, active, nonphosphorylated Lgl is needed for cortical recruitment of GFP-Pon although lgl1 mutant clones did not show a very strong phenotype. The easiest explanation for the discrepancy between the lgl1 mutant and ectopic Lgl phosphorylation is the perdurance of residual Lgl protein in mutant tissue. This is supported by previous observations describing Numb-localization defects in temperature-sensitive alleles of lgl. It is possible that Lgl can mediate its effects even at protein concentrations below the detection limit of the antibody. Thus, Lgl may not be needed at stoichiometric levels for asymmetric protein localization in SOP cells, but it instead plays a catalytic or signaling role (Mayer, 2005).
How could Lgl recruit Pon to the cell cortex? Formally, it is possible that Pon simply binds Lgl in a phosphorylation-dependent manner. However, no direct interaction has been described and such a model would not explain why Pon is cortical even when Lgl levels are strongly reduced. Two other models are more likely: Either cortical binding sites for Numb and Pon are present all around the cell, but their affinity depends on Lgl and its phosphorylation status and therefore varies along the cell cortex (Model 1); or a limiting number of cortical binding sites are present only on one side of the cell, and Lgl is responsible for their asymmetric distribution (Model 2). To distinguish between these models, FRAP rates were measured for cortical GFP-Pon in different genetic backgrounds. The FRAP rate is a function of the rate constants for both association and dissociation of GFP-Pon with its postulated cortical binding site. In Model 1, expression of activated lgl (lgl3A) or deregulated aPKC (aPKC-?N) should alter the affinity of the binding site and therefore change the rate constants, resulting in a variation of the FRAP rate. Because the FRAP rate is independent of receptor concentration, however, it would remain constant under the same conditions in Model 2. Cortical GFP-Pon FRAP rates were measured in wild-type SOP cells, in cells expressing lgl3A, and in cells where Lgl was inactivated by expression of aPKC-deltaN. Although expression of aPKC-deltaN dramatically reduced the amount of GFP-Pon present at the cortex, it did not influence the kinetics of GFP-Pon binding to the cortical binding site. Thus, the number of Pon binding sites at the cell cortex, and not their affinity for Pon, seems to be reduced by aPKC-deltaN expression (Mayer, 2005).
To gain independent evidence for the two models, the fraction of GFP-Pon present at the cell cortex was quantitated. If Lgl regulated GFP-Pon binding site affinity, expression of lgl3A would change the entire SOP cell cortex to high affinity, and therefore it would increase the cortical GFP-Pon fraction. If Lgl regulated only the distribution of binding sites, however, the cortical fraction of GFP-Pon should remain the same. Cortical recruitment was quantified by measuring the ratio of cortical to cytoplasmic fluorescence for GFP-Pon and Numb-GFP at different time points in mitosis. Compared to wild-type cells, expression of lgl3A did not cause a significant increase in cortical recruitment. This is not because cytoplasmic GFP-Pon is limiting; increased GFP-Pon expression predominantly increased the cytoplasmic signal. Taken together, these results favor Model 2, in which Lgl acts by asymmetrically distributing a limiting number of cortical GFP-Pon binding sites. The loss of cortical fluorescence upon aPKC-deltaN expression indicates that lgl is also required for binding site formation, in addition to binding site positioning. However, this second role of lgl does not seem to be rate limiting under normal conditions because lgl3A expression does not increase the cortical GFP-Pon fraction. Although these results are most consistent with Model 2, more-complex models cannot be excluded. For example, lgl could distribute a limiting adaptor protein that links Pon to a receptor but is not the receptor itself (Mayer, 2005).
The direct cortical binding partners for Pon or Numb have not yet been identified. Thus, it is only possible to speculate on the molecular mechanisms of their postulated asymmetric distribution. Although the results are inconsistent with lateral transport of GFP-Pon, they do not exclude lateral transport of its cortical anchor. Similar to what has been proposed for asymmetric cell division in C. elegans, a possible mechanism could be local tearing and contraction of the cortical actin cytoskeleton. Lgl was shown to inhibit the cortical localization of myosin II, and it has been proposed that cortical myosin II might exclude asymmetrically segregating proteins. These data could be integrated with the model if myosin II excludes the cortical binding sites rather than influencing determinant localization directly. Alternatively, transmembrane receptors for Pon or Numb could be delivered to the position of crescent formation by vesicle transport. Such a mechanism in which transmembrane receptors are present on vesicles that dock at the membrane in an Lgl-dependent fashion would be consistent with the quantitative observations. It would also explain why Lgl is essential for crescent formation but not needed in metaphase for maintenance of asymmetric protein localization. It is remarkable that lateral diffusion of transmembrane proteins is slow enough to allow a stable asymmetric distribution, if the delivery of the protein is asymmetric, both in yeast and in SOP cells. The yeast Lgl orthologs Sro7p and Sro77p have been implicated in plasma-membrane fusion of secretory vesicles, and Lgl has been proposed to regulate vesicular targeting to specific membrane domains. Furthermore, asymmetric protein localization in Drosophila requires myosin VI, a motor whose main function is vesicle movement, suggesting that vesicle trafficking plays some role (Mayer, 2005).
These data provide insight into the dynamic protein movements of cell-fate determinants and their associated adaptor proteins during asymmetric cell division. It is proposed that these determinants are preferentially recruited from the cytoplasm to a high-affinity binding site during late prophase. Establishment of this binding site is regulated by the phosphorylation status of Lgl. The role of Lgl is more to concentrate binding sites on one side of the cell than to act as a receptor itself or change the affinity of another Numb or Pon binding site (Mayer, 2005).
The choice of self-renewal versus differentiation is a fundamental issue in stem cell and cancer biology. Neural progenitors of the Drosophila post-embryonic brain, larval neuroblasts (NBs), divide asymmetrically in a stem cell-like fashion to generate a self-renewing NB and a ganglion mother cell (GMC), which divides terminally to produce two differentiating neuronal/glial daughters. Aurora-A (AurA) acts as a tumor suppressor by suppressing NB self-renewal and promoting neuronal differentiation. In aurA loss-of-function mutants, supernumerary NBs are produced at the expense of neurons. AurA suppresses tumor formation by asymmetrically localizing atypical protein kinase C (aPKC), an NB proliferation factor. Numb, which also acts as a tumor suppressor in larval brains, is a major downstream target of AurA and aPKC. Notch activity is up-regulated in aurA and numb larval brains, and Notch signaling is necessary and sufficient to promote NB self-renewal and suppress differentiation in larval brains. These data suggest that AurA, aPKC, Numb, and Notch function in a pathway that involved a series of negative genetic interactions. This study has identified a novel mechanism for controlling the balance between self-renewal and neuronal differentiation during the asymmetric division of Drosophila larval NBs (Wang, 2006).
When aurA function is compromised, mutant NBs acquire some features of cancer stem cells. They divide to generate a large number of daughter cells capable of self-renewal. This excessive self-renewal occurs at the expense of neuronal differentiation, suggesting that the normally asymmetric NB divisions have been altered such that the mutant NBs can divide symmetrically to generate two NB-like daughters. Cell cycle regulator CycE and cell growth factor dMyc are expressed in most of these tumor-like cells. Up-regulation of CycE is required for aurA overgrowth phenotype. AurA also regulates proper orientation of the mitotic spindle probably by controlling asymmetric localization of Mud. Both proteins are localized to centrosomes and are required for centrosome function. Centrosome abnormality and chromosome segregation defects in aurA could lead to aneuploidy, and many cancer cells exhibit centrosome defects and chromosome instability. Mammalian AurA when overexpressed can be oncogenic. However, future studies on its possible role as a tumor suppressor will be particularly interesting (Wang, 2006).
The data suggest that aurA negatively regulates aPKC function to regulate NB self-renewal. aPKC appears to act as a NB proliferation factor since overexpression of a modified membrane-targeted version, aPKC-CAAX, which exhibits ectopic cortical localization throughout the NB cortex, leads to overproliferation and tumor formation, similar to loss of aurA. AurA is required for the asymmetric localization of aPKC and restrict aPKC to the cortical region associated with the future NB daughter and loss of aurA results in delocalization of aPKC to the entire cortex. Consistent with and supporting this notion, loss of aPKC can suppress, albeit partially, the aurA mutant overgrowth phenotype (Wang, 2006).
In contrast to the well-studied role of Numb as a cell fate determinant during asymmetric divisions of embryonic GMCs, SOPs, or muscle progenitors, a role for Numb during NB asymmetric divisions has not been described. This study shows that Numb also acts as a tumor suppressor in Drosophila larval brains, and that Numb is a key downstream target of AurA and aPKC in the regulation of NB self-renewal. In both aurA mutant NBs or NBs overexpressing aPKC-CAAX, the asymmetric localization of Numb is compromised and the resultant overgrowth phenotype is consistent with that of numb loss-of-function. numb and aurA mutant NBs also share several common features including excessive self-renewal at the expense of neuronal differentiation as well as the membrane enrichment of Spdo, a positive regulator of Notch signaling. These data suggest that AurA positively regulates Numb function. Genetic analysis is consistent with the notion that this is achieved through the negative regulation of aPKC that in turn negatively regulates Numb (Wang, 2006).
Numb is known to be a negative regulator of Notch signaling. The current findings indicate that Notch is necessary and sufficient for promoting larval NB proliferation and suppressing neuronal differentiation. Genetic epistasis studies suggest that an AurA-aPKC-Numb-Notch genetic hierarchy acts to regulate self-renewal of Drosophila neural progenitor cells. During a wild-type larval NB asymmetric division, aurA acts to negatively regulate aPKC and restrict its localization to the cortical region associated with the future NB daughter; aPKC negatively regulates Numb and ensures that its localization/activity is restricted to the future GMC where Numb acts to antagonize Notch. The net effect is that Notch is asymmetrically activated in the NB daughter where it acts to promote self-renewal and suppress differentiation. Although these data suggest that aurA acts through the aPKC/Numb/Notch pathway, given the partial suppression seen in the double mutants aPKC;aurA and Notchts-1;aurA, the possibility that additional mechanisms may be involved cannot be excluded (Wang, 2006).
Regulation of stem cell self-renewal versus differentiation is critical for embryonic development and adult tissue homeostasis. Drosophila larval neuroblasts divide asymmetrically to self-renew, and are a model system for studying stem cell self-renewal. This study identified three mutations showing increased brain neuroblast numbers that map to the aurora-A gene, which encodes a conserved kinase implicated in human cancer. Clonal analysis and time-lapse imaging in aurora-A mutants show single neuroblasts generate multiple neuroblasts (ectopic self-renewal). This phenotype is due to two independent neuroblast defects: abnormal atypical protein kinase C (aPKC)/Numb cortical polarity and failure to align the mitotic spindle with the cortical polarity axis. numb mutant clones have ectopic neuroblasts, and Numb overexpression partially suppresses aurora-A neuroblast overgrowth (but not spindle misalignment). Conversely, mutations that disrupt spindle alignment but not cortical polarity have increased neuroblasts. It is concluded that Aurora-A and Numb are novel inhibitors of neuroblast self-renewal and that spindle orientation regulates neuroblast self-renewal (Lee, 2006b).
Mutations in aurA lead to a massive increase in larval brain neuroblasts. The major cause of this phenotype appears to be misregulation of neuroblast cortical polarity. One cortical polarity defect is increased basal localization of aPKC, which is sufficient to induce ectopic neuroblasts. Consistent with this hypothesis, aPKC aurA double mutants show strong suppression of the aurA supernumerary neuroblast phenotype, consistent with aPKC functioning downstream from AurA. While it is possible that loss of aPKC suppresses the phenotype in a nonspecific way (e.g., by arresting neuroblast cell proliferation or inducing neuroblast apoptosis), ni similarly strong suppression of the brat supernumerary neuroblast phenotype was observed in aPKC brat double mutants. This shows that aPKC functions more specifically in the AurA pathway than in the Brat pathway (Lee, 2006b).
The only other detectable cortical polarity defect seen in aurA mutant neuroblasts is a delocalization of Numb from the basal cortex. A similar Numb defect is seen during asymmetric cell division of pupal SOPs in aurA mutants, perhaps reflecting a specific and direct regulation of Numb by AurA, although Numb is not phosphorylated by AurA in vitro. The importance of the Numb delocalization phenotype is revealed by the ability of Numb overexpression in neuroblasts to rescue most of the aurA mutant phenotype (all except the component due to spindle orientation defects). Thus, Numb acts downstream from AurA to inhibit neuroblast self-renewal. Numb joins Mira/Pros/Brat as proteins that are partitioned into the GMC during neuroblast asymmetric cell division, where they function to inhibit neuroblast self-renewal (Lee, 2006b).
Where does AurA function to inhibit neuroblast self-renewal? AurA appears to be required in the neuroblast lineage, and not in surrounding glial cells or nonneuronal tissues of the larva, because neuroblast-specific expression of either AurA or the downstream component Numb can rescue most of the aurA supernumerary neuroblast phenotype. This shows that AurA is not required outside the neuroblast lineage to inhibit neuroblast self-renewal. Within the neuroblast, AurA appears to function in the cytoplasm and not at the centrosome, because cnn mutants lack all detectable AurA centrosomal localization yet do not match the aurA supernumerary neuroblast phenotype. It is concluded that AurA acts in the neuroblast cytoplasm to promote aPKC/Numb cortical polarity and spindle-to-cortex alignment (Lee, 2006b).
How does Numb inhibit neuroblast self-renewal in the GMC? Numb is a well-characterized inhibitor of Notch signaling that is segregated into the GMC, and Notch signaling is active in larval neuroblasts but not in GMCs. Thus the most obvious model is that Numb blocks Notch receptor signaling in the GMC. However, Notch mutant clones generated in larval neuroblasts do not affect neuroblast survival or clone size. Similarly, no change has been seen in neuroblast number in two different Notch-ts mutants (although the expected small wing imaginal disc phenotype was observed. In addition, no supernumerary neuroblasts were observed in larval neuroblast clones overexpressing the constitutively active Notch intracellular domain, although the same Notch intracellular domain generates the expected sibling neuron phenotype when expressed in the embryonic CNS. Thus, Notch is an excellent candidate for promoting neuroblast self-renewal, but additional experiments will be needed to test this model more rigorously. In this context, it is interesting to note that Notch promotes stem cell self-renewal in mammals (Lee, 2006b).
aurA mutant neuroblasts have essentially random orientation of the mitotic spindle relative to the apical/basal cortical polarity axis, resulting in a some neuroblasts dividing symmetrically (in size and cortical polarity markers). This phenotype may arise due to lack of astral microtubule interactions with the neuroblast cortex; aurA mutant neuroblasts have reduced astral microtubule length. Alternatively, AurA may affect spindle orientation by phosphorylating proteins required for spindle orientation, such as Cnn, Pins, or Mud. For example, Mud has a consensus AurA/Ipl1 phosphorylation site within its microtubule-binding domain, and it will be interesting to determine if this site needs to be phosphorylated for Mud to bind microtubules. Spindle orientation defects only generate part of the supernumerary neuroblast phenotype in aurA mutant brains, however, because overexpression of Numb can rescue most of the phenotype without rescuing spindle alignment, and cnn or mud mutants have nearly random spindle alignment but only a modest increase in neuroblast number. Thus, it is proposed that spindle orientation defects and cortical polarity defects combine to generate the dramatic supernumerary neuroblast phenotype seen in aurA mutants (Lee, 2006b).
Mammalian aurA has been termed an oncogene due to its overexpression in several cancers, its ability to promote proliferation in certain cell lines, and the fact that reduced levels lead to multiple centrosomes, mitotic delay, and apoptosis. However, an in vivo aurA mutant phenotype has not yet been reported. In contrast, aurA loss-of-function mutations result in a neuroblast 'brain tumor' phenotype, including prolonged neuroblast proliferation during pupal stages when wild-type neuroblasts have stopped proliferating. aurA mutants do not, however, have the imaginal disc epithelial overgrowth seen in other Drosophila tumor suppressor mutants, and aurA mutant neuroblasts have a delay in cell cycle progression. It is proposed that the aurA supernumerary neuroblast phenotype is not due to loss of growth control or a faster cell cycle time, but rather due to a cell fate transformation from a differentiating cell type (GMC) to a proliferating cell type (neuroblast) (Lee, 2006b).
It is concluded that AurA restrains neuroblast numbers using two pathways: first by promoting Numb localization into the GMC, and second by promoting alignment of the mitotic spindle with the cortical polarity axis. Absence of the first pathway leads to increased neuroblasts at the expense of GMCs, whereas absence of the second pathway leads to increased neuroblasts due to symmetric cell division. It will be interesting to determine whether mammalian AurA uses one or both pathways to regulate stem cell asymmetric division and self-renewal (Lee, 2006b).
To test whether aPKC and Bazooka are associated in a protein complex, coimmunoprecipitation experiments were performed. Baz was immunoprecipitated from extracts of S2 cells overexpressing Baz. S2 cells express endogenous aPKC. The immune complex was subjected to SDS-PAGE and Western analysis with anti-aPKC antibody C20. A significant amount of aPKC coimmunoprecipitates with Baz. To determine whether binding of DaPKC to Baz is direct, interaction studies were performed with the yeast two-hybrid system. A construct containing full-length aPKC fused to the transactivation domain of GAL4 was cotransformed into yeast with bait constructs containing different regions of Baz fused to the GAL4 DNA-binding domain. Interaction of the bait constructs with DaPKC was assayed by X-Gal filter assays. Bait constructs containing the second and third PDZ domain of Baz show interaction with full length DaPKC, whereas all constructs lacking the second or third PDZ domain give negative results. It is concluded that binding of DaPKC to Baz is direct and that the region from amino acid 401 to 737 of Baz is sufficient for binding to DaPKC (Wodarz, 2000).
To generate different cell types, some cells can segregate protein determinants into one of their two daughter cells during mitosis. In Drosophila neuroblasts, the Par protein complex localizes apically and directs localization of the cell fate determinants Prospero and Numb and the adaptor proteins Miranda and Pon to the basal cell cortex, to ensure their segregation into the basal daughter cell. The Par protein complex has a conserved function in establishing cell polarity but how it directs proteins to the opposite side is unknown. A principal function of this complex is to phosphorylate the cytoskeletal protein Lethal (2) giant larvae [Lgl; also known as L(2)gl]. Phosphorylation by Drosophila atypical protein kinase C (aPKC), a member of the Par protein complex, releases Lgl from its association with membranes and the actin cytoskeleton. Genetic and biochemical experiments show that Lgl phosphorylation prevents the localization of cell fate determinants to the apical cell cortex. Lgl promotes cortical localization of Miranda, and it is proposed that phosphorylation of Lgl by aPKC at the apical neuroblast cortex restricts Lgl activity and Miranda localization to the opposite, basal side of the cell (Betschinger, 2003).
Recent results have demonstrated the critical role of the mammalian p62-atypical protein kinase C (aPKC) complex in the activation of NF-kappaB in response to different stimuli. Using the RNA interference technique on Schneider cells it has been shown that Drosophila aPKC (DaPKC) is required for the stimulation of the Toll-signaling pathway, which activates the NF-kappaB homologs Dif and Dorsal. However, DaPKC does not appear to be important for the other Drosophila NF-kappaB signaling cascade, which activates the NF-kappaB homolog Relish in response to lipopolysaccharides. Interestingly, DaPKC functions downstream of the nuclear translocation of Dorsal or Dif, controlling the transcriptional activity of the Drosomycin promoter. The Drosophila Ref(2)P protein is the homolog of mammalian p62, since it binds to DaPKC: its overexpression is sufficient to activate the Drosomycin but not the Attacin promoter, and its depletion severely impairs Toll signaling. Collectively, these results demonstrate the conservation of the p62-aPKC complex for the control of innate immunity signal transduction in Drosophila melanogaster (Avila, 2002).
Drosophila represents an ideal system in which to determine the primary role of the aPKCs in NF-kappaB signal transduction because it encodes only one aPKC isoform. According to the data presented in this study, aPKC is selectively required for the innate immune Toll-signaling pathway, acting downstream of the translocation of Dorsal and Dif and playing a critical role in the induction (a typical NF-kappaB-dependent process) of the antimicrobial peptide gene for Drosomycin. Therefore, it can be argued that the primary role of the aPKCs, particularly that of zetaPKC in higher eukaryotic cells, is to somehow control the transcriptional activity of NF-kappaB through a still not completely understood mechanism that most likely involves the direct phosphorylation of RelA and Dif. Interestingly, in Drosophila it is well documented that the phosphorylation of Dorsal is required not only for its transcriptional activity but also for its nuclear translocation. In Drosophila aPKC-depleted cells, a strong inhibition of Dorsal or Dif nuclear translocation is not observed, suggesting that the role of Drosophila aPKC is independent of the previously characterized role for Dorsal phosphorylation in regulating nuclear translocation. Based on experiments in mammalian systems, which demonstrate that p65 transcriptional activity must be stimulated by phosphorylation, it is possible that the residues that control the transcriptional activities of both Dorsal and Dif are different from those controlling the nuclear import of the protein. It is also possible that Drosophila aPKC-mediated phosphorylation has a subtle, yet important, role in the nuclear translocation of Dif and/or Dorsal. Future studies will address this important issue (Avila, 2002).
These studies also demonstrate that Ref(2)P is most likely the functional homolog of p62 in Drosophila. Like p62, Ref(2)P interacts physically with the aPKCs. Therefore, it appears that the p62-aPKC signaling module, like the Par/aPKC complex, is highly conserved. Importantly, a functional role of Ref(2)P in Toll signaling is demonstrated. Thus, the ectopic expression of Ref(2)P is capable by itself of activating the Drosomycin promoter. More interestingly, its depletion severely impairs the Toll pathway (Drosomycin induction) but not the LPS pathway (Attacin induction). Thus, the Ref(2)P/DaPKC complex is critical for Toll signaling (Avila, 2002).
The results presented here also demonstrate that, similar to the p62-TRAF6 connection in mammals, Ref(2)P and Drosophila TRAF2 physically and functionally interact. Together with the results demonstrating that Drosophila aPKC and Ref(2)P are essential for a downstream event in the Toll-signaling pathway, this suggests that a putative Ref(2)P/aPKC/TRAF2 complex might function in the signal-induced stimulation of Dif or Dorsal transcriptional activity. In this regard, it is noteworthy that recent results suggest that TRAF6, in addition to its role in IKK recruitment and activation, may also be involved in the control of RelA transcriptional activity. However, the role of Drosophila TRAF2 in Toll signaling requires further investigation, since the effect of inhibiting (or mutating) TRAF2 has not yet been reported. Further studies will also address the precise mechanism whereby aPKC controls the Toll pathway. The data presented here clearly establish the conserved role of the homolog of the p62/aPKC cassette in NF-kappaB signaling in Drosophila (Avila, 2002).
Apicobasal cell polarity is crucial for morphogenesis of photoreceptor rhabdomeres and adherens junctions (AJs) in the Drosophila eye. Crumbs (Crb) is specifically localized to the apical membrane of photoreceptors, providing a positional cue for the organization of rhabdomeres and AJs. The Crb complex consisting of Crb, Stardust (Sdt) and Discs-lost (Dlt) colocalizes with another protein complex containing Par-6 and atypical protein kinase C (aPKC) in the rhabdomere stalk of photoreceptors. Loss of each component of the Crb complex causes age-dependent mislocalization of Par-6 complex proteins, and ectopic expression of Crb intracellular domain is sufficient to recruit the Par-6 complex. The absence of Par-6 complex proteins results in severe mislocalization and loss of Crb complex. Dlt directly binds to Par-6, providing a molecular basis for the mutual dependence of the two complexes. These results suggest that the interaction of Crb and Par-6 complexes is required for the organization and maintenance of apical membranes and AJs of photoreceptors (Nam, 2003).
The strong dependence of Crb localization on Sdt and Dlt suggests that Crb may be destabilized or may not be targeted to the membrane in the absence of Sdt or Dlt. It is intriguing that Sdt and Dlt are lost only partially in the absence of Crb. The findings of a direct interaction between Dlt and Par-6 suggest that Sdt-Dlt can still be targeted to the membrane in the absence of Crb through the binding of Dlt to the Par-6 complex. However, it is important to note that Dlt is essentially lost in sdt mutant clones and vice versa. This raises an intriguing possibility that Dlt or Sdt are dependent on each other in vivo to be targeted to the apical membrane via binding to either Crb or Par-6. This mutual dependency between Dlt and Sdt may explain why Dlt and Sdt are lost in the absence of the other, rather than being associated with the Par-6 complex (Nam, 2003).
The interaction between the Crb and Par-6 complexes is mediated by the PDZ3 region of Dlt and the N-terminal domain of Par-6. The N-terminal domain of Par-6 is also used for binding aPKC. Therefore, a potential function of Dlt is to bind Par-6 in competition with aPKC or to facilitate the interaction of Par-6 with aPKC or other Par-6 binding proteins. Mutant analysis indicates that loss of Dlt and Sdt in sdt- clones causes mislocalization of both Crb and Par-6 complex proteins. This suggests that Sdt-Dlt interaction provides a scaffold to recruit Crb complex to the Par-6 complex and enhance the stability of these two complexes rather than functioning as a competitor for aPKC (Nam, 2003).
Proteins in Crb and Par-6 complexes consist of multiple functional domains which may be involved in diverse protein-protein interactions. A recent study has shown that in mammalian cell culture systems the PDZ domain of Par-6 binds not only Par-3 but also the N terminus of Pals1. These results suggest that the crosstalk between the Crb and Par-6 complexes is mediated by multiple domain-specific interactions. Evidence from genetic analysis using mutants suggests that the crosstalk between the two complexes is mutually required for normal organization of apical membranes and AJs in vivo, and also provides a basis for partial redundancy of these complexes in the organization of photoreceptor cell polarity. Interestingly, when either Crb or Sdt is lost, mislocalization or elimination of other associated components including Par-6 complex proteins becomes more severe in the age-dependent manner. This suggests that the Crb complex may be required for the maintenance rather than the formation of the Par-6 complex. The age-dependent degenerative phenotype may be related to the requirement of extensive apical membrane growth to make rhabdomeres and AJs along the growing axis of photoreceptors during pupal stage. Loss of any one component of the Crb complex is likely to be increasingly more detrimental as the process of membrane reorganization proceeds. In crb- or sdt- mutants, significant fractions of Par-6 complex proteins remain in the membrane despite the age-dependent and progressive mislocalization of apical markers. By contrast, loss of Par-6 or aPKC results in mislocalization of Dlt from the apical membrane. This suggests that the Par-6 complex plays essential functions for membrane localization of Crb complex proteins. Furthermore, both Par-6 and aPKC seem to be important for survival and/or proliferation of retinal cells because mutant clones were very small compared with adjacent twin spots and often completely disrupted, probably due to cell death. This is consistent with the findings of frequent apoptosis in aPKC- or par-6- embryos (Nam, 2003).
An important distinction of Par-6 complex in the photoreceptors from other epithelia is the localization of Baz. Baz localizes with Crb complex in the subapical membrane or both the subapical region and AJ in the Drosophila embryonic epithelia. Vertebrate Par-3 also localizes to the apical tight junction in vertebrate epithelial cells. By contrast, Baz in the photoreceptors is specifically positioned in the AJs basal to the all other proteins in the Crb/Par-6 complexes. Baz and Arm are recruited together to ectopic membrane sites by misexpression of CrbJM, suggesting that Baz is an integral component of AJ. However, Baz is not recruited by CrbPBM, whereas Par-6 and aPKC can be ectopically recruited by CrbPBM rather than CrbJM. Therefore, Baz appears to be recruited to AJ independently of Par-6/aPKC (Nam, 2003).
Intriguingly, despite its specific localization to AJs, loss of Baz results in most severe disruption of AJ as well as the more apical Dlt domain. It has been proposed that the Par-6/aPKC cassette is recruited to the site of cell-cell contact and then moves along the most apical zone of the developing cell-cell contact. In this process, an important step for cell polarity formation is to tether the cytoplasmic Par-6/aPKC complex to the site of cell-cell contact at the membrane, which is mediated by the interaction of Par-3 and a membrane protein JAM. Therefore, the results that baz mutation causes loss of Dlt and AJs support the crucial role of Baz in the initial step of cell polarization. However, the distinct localization of Baz from Par-6 and aPKC in the photoreceptors suggests that the mode of Baz localization varies in different systems. In photoreceptors, Baz may be targeted to the membrane with Par-6 but be sorted out from Par-6 in subsequent steps of polarization to remain in the AJs, whereas Par-6-aPKC-Baz cassette remains together in the complex in other epithelia. In contrast to Baz, aPKC localizes to both rhabdomere stalk and AJ, suggesting that Baz and Par-6 are completely separated during polarization while aPKC is not sorted from both Par-6 and Baz. The critical function of Baz in the localization of Crb complex in the rhabdomere stalk is consistent with the requirement of Baz for Crb localization in embryonic epithelia. However, the requirement of Baz in the embryo appears to be dependent on the stage of development since Crb distribution in the absence of Baz becomes normal in late embryos. On the contrary, such stage-dependent recovery of Crb complex localization has not been observed in baz- photoreceptor cells (Nam, 2003).
Recent studies have shown that mutations in human CRB1 cause RP12 and LCA, severe recessive retinal diseases, emphasizing the importance of Crb family proteins in the eyes of mammals including humans. The Drosophila Crb and human CRB1 are localized in analogous subcellular membrane domains of photoreceptors, the rhabdomere stalk and the inner segment in Drosophila and human photoreceptors, respectively. Besides similar subcellular localization, Crb and human CRB1 are functionally conserved. Age-dependent photoreceptor defects in the crb mutant also provide analogy to age-dependent retinal degeneration in RP12/LCA patients. These studies here imply that hCRB1 may function as a protein complex with homologs of Sdt and Dlt and such a complex may interact with a homologous Par-6 complex. Whether such homologous human genes are the targets of inherited retinal diseases such as RP remains to be studied (Nam, 2003).
How epithelial cells subdivide their plasma membrane into an apical and a basolateral domain is largely unclear. In Drosophila embryos, epithelial cells are generated from a syncytium during cellularization. Polarity is established shortly after cellularization when Par-6 and the atypical protein kinase C concentrate on the apical side of the newly formed cells. Apical localization of Par-6 requires its interaction with activated Cdc42 and dominant-active or dominant-negative Cdc42 disrupt epithelial polarity, suggesting that activation of this GTPase is crucial for the establishment of epithelial polarity. Maintenance of Par-6 localization requires the cytoskeletal protein Lgl. Genetic and biochemical experiments suggest that phosphorylation by aPKC inactivates Lgl on the apical side. On the basolateral side, Lgl is active and excludes Par-6 from the cell cortex, suggesting that complementary cortical domains are maintained by mutual inhibition of aPKC and Lgl on opposite sides of an epithelial cell (Hutterer, 2004).
These results describe the first steps of a molecular pathway that leads to the establishment of polarity in epithelial cells of the Drosophila ectoderm. The Par-6 protein localizes to the apical cell cortex by binding to Cdc42. Par-6 recruits Bazooka and aPKC and is essential for establishment of the apical domain. Maintenance of Par-6 localization requires Lgl, a substrate of aPKC. Phosphorylation by aPKC inactivates Lgl at the apical cell cortex and restricts Lgl to the basolateral cortex to establish the basolateral domain (Hutterer, 2004).
Apical localization of Par-6 is a key event in the establishment of epithelial polarity. How is Par-6 recruited to the apical cell cortex? In C. elegans, the proteins Par-3, Par-6, and aPKC are localized to the anterior cell cortex before and during the first cell division. Their asymmetric localization is initiated by interaction of the sperm aster with the overlying cell cortex that excludes Par-6 from the posterior cell cortex. During Drosophila cellularization, centrosomes are located apically and it is therefore unlikely that a similar cortical microtubule interaction is responsible for the apical localization of Par-6 (Hutterer, 2004).
Although a distinct apical domain with sharp boundaries is established in epithelial cells only after cellularization, elegant membrane tracer experiments have revealed a subdivision of the plasma membrane into distinct regions already during cellularization. Are these membrane compartments prefiguring the future apical and basolateral domains and is Par-6 localizing apically by recognizing a preformed membrane domain? The first membrane domain is the furrow canal at the tip of the ingrowing cellularization front that is marked by Patj. This domain disintegrates after cellularization and is therefore unlikely to participate in Par-6 localization. During later stages, new membrane is preferentially inserted apically, then apicolaterally. At these stages, newly inserted membrane displaces the pre-existing membrane toward both the apical and basolateral side, indicating that a distinct apical membrane compartment is not established by the end of cellularization. It is therefore unlikely that Par-6 recognizes a preformed apical membrane compartment although these experiments do not rule out a more general role of the vesicle transport machinery in Par-6 localization (Hutterer, 2004).
The results indicate that Par-6 needs to bind to activated Cdc42 in order to localize apically. Since cdc42 mutants cannot be analyzed at this stage, a conserved proline in the CRIB domain was mutated to generate a Par-6 version that no longer binds Cdc42. The structure of the Par-6 Cdc42 complex shows that this residue comes to lie in a hydrophobic groove of the Cdc42 molecule. This may explain why it can be replaced by alanine without affecting Cdc42 binding. When it is deleted, however, one of the adjacent highly charged amino acids will occupy the position of the proline. This could strongly inhibit interaction with the hydrophobic pocket and eliminate binding to Cdc42 both in vertebrates and in flies. Since both Lgl and aPKC still bind Par-6-DeltaP and the protein is expressed at almost wild-type levels from the endogenous promoter in an otherwise null mutant background, par-6-DeltaP embryos are specifically defective in binding of Cdc42 to the Par-6/aPKC complex (Hutterer, 2004).
How does activated Cdc42 localize Par-6? Cdc42 might be required for association of an unidentified Par-6 binding partner that is essential for apical localization of the protein. The conformation of Par-6 changes upon binding to Cdc42, and this could affect interactions with other proteins. However, aPKC and Lgl are the only proteins identified in the Par-6 complex, and their interaction does not depend upon Cdc42 binding. In vertebrates, Par-6 interacts with the Stardust homolog Pals1, and this interaction is regulated by Cdc42. Stardust acts together with its binding partner Crumbs, but apical protein localization is initiated correctly in crumbs mutants. Therefore, it is unlikely that Stardust binding to Par-6 is critical for the initial apical localization of Par-6. It is more likely that Cdc42 activation provides an instructive cue for Par-6 localization. Cdc42 could be preferentially activated on the apical side, for example by localization of an exchange factor, and this could recruit Par-6 to the apical cell cortex. This hypothesis is supported by the ectopic patches of Par-6, which are observed after overexpression of constitutively active Cdc42. Asymmetric activation of Cdc42 is known to polarize other cell types. In yeast, the exchange factor Cdc24 is localized to the incipient bud site. This locally activates Cdc42 and polarizes the actin cytoskeleton toward the site. In migrating neutrophils, Cdc42 is locally activated in response to a chemoattractant gradient by the exchange factor PIXalpha. A clear Drosophila ortholog of PIXalpha exists, but whether it is involved in epithelial polarity remains to be determined (Hutterer, 2004).
Maintenance of Par-6 localization requires the cytoskeletal protein Lgl. Lgl acts at the basolateral cortex where it inhibits cortical localization of Par-6. How Lgl excludes Par-6 from the cortex is unclear, but it is remarkable that in other tissues, Lgl actually promotes cortical protein localization. In MDCK cells, Lgl was suggested to regulate basolateral exocytosis and it could recruit a Par-6 antagonizing factor to the basolateral plasma membrane. Since Lgl and Bazooka binding to Par-6 seem to be mutually exclusive, Lgl could also inactivate the Par protein complex by displacing Bazooka. To perform its role in epithelial polarity, Lgl needs to be phosphorylated by aPKC. This modification has been shown to inactivate the protein and release it from its association with membranes and the cytoskeleton. These results suggest that in epithelial cells, apically localized aPKC phosphorylates Lgl to displace the protein from the apical cell cortex. A simple model is proposed in which mutual inhibition between Par-6/aPKC on the apical and Lgl on the basolateral cell cortex maintains epithelial polarity. This model is in agreement with previous studies that demonstrate negative genetic interactions between lgl and proteins that localize to the apical domain. Furthermore, it provides a molecular explanation for the recently described suppression of the lgl mutant epithelial polarity phenotype by reduction of aPKC levels. Negative interactions between the apical and basolateral domains of epithelial cells have been described before. In the Drosophila follicular epithelium, Bazooka is phosphorylated and inhibited by Par-1, a protein kinase located on the basolateral domain, thus restricting the Par protein complex to the apical domain (Hutterer, 2004).
The proteins Par-6, Bazooka, and aPKC localize to the apical cell cortex of both neuroblasts and epithelial cells, but the mechanism of apical localization seems to be different in the two cell types. In epithelial cells, Lgl is required for maintaining Par proteins at the apical cell cortex, while Par protein localization in neuroblasts is Lgl independent. Expression of nonphosphorylatable Lgl disrupts asymmetric cell division in neuroblasts but is without effect in epithelial cells. In addition, overexpression of dominant-active or -negative Cdc42 disrupts epithelial polarity but has no effect on neuroblast division. What is the basis for these differences (Hutterer, 2004)?
Epithelial cells rely on adherens junctions for maintaining distinct membrane compartments. Such junctions are absent from neuroblasts, and in fact, distinct membrane compartments do not seem to exist. Instead, Par protein localization in neuroblasts requires a protein called Inscuteable that is recruited apically by binding to Bazooka and aPKC and activates heterotrimeric G proteins through an adaptor molecule called Pins. Both Inscuteable and G proteins are essential for maintaining Par protein localization in neuroblasts but not epithelial cells. It is possible that a feedback loop operates downstream of the G proteins to maintain polarity in the absence of diffusion barriers and cellular junctions. Mechanistic differences in the way Par proteins localize are also observed between species. In C. elegans, neither Lgl nor G proteins are required for Par-3 or Par-6 localization. Instead, a Ring finger protein called Par-2 maintains Par-3 and Par-6 at the anterior pole. Cdc42 plays a role, but only in maintenance and not establishment of polarity. Clearly, key players are missing that might help in an understanding of these mechanistic differences (Hutterer, 2004).
Cdc42 binds vertebrate Par-6. Both proteins are implicated in polarizing vertebrate epithelial cells, and their conserved interaction suggests that they achieve this via a conserved mechanism. Although in vertebrates both proteins primarily act on tight junctions, the role of Cdc42 in localizing the Par proteins seems conserved since overexpression of an activated form inhibits the localization of Par-3 to tight junctions in MDCK cells. However, current experiments do not confirm a previously demonstrated role of Cdc42 in activating Par-6-associated aPKC in vitro. Unlike in vertebrates, aPKC is shown to be equally active - at least toward Lgl - when bound to a form of Par-6 that does not interact with Cdc42. Whether species-specific differences or the different experimental setups are responsible for this apparent discrepancy remains unclear. Besides their function in polarity, the Par proteins are involved in proliferation control of vertebrate epithelial cells. Par-6 cooperates with Cdc42 in transforming cells, suggesting a role in oncogenic transformation. In Drosophila, Cdc42, Lgl, and Bazooka were shown to cooperate with activated ras in the formation of metastatic tumors. It can be anticipated that the powerful tool of Drosophila genetics will help to identify other components of this pathway that might clarify its role in carcinogenesis (Hutterer, 2004).
Planar cell polarity (PCP) is a common feature of many vertebrate and
invertebrate epithelia and is perpendicular to their apical/basal (A/B) polarity
axis. While apical localization of PCP determinants such as Frizzled (Fz1) is
critical for their function, the link between A/B polarity and PCP is poorly
understood. A direct molecular link is described between A/B determinants
and Fz1-mediated PCP establishment in the Drosophila eye.
Patj binds the cytoplasmic tail of Fz1 and is proposed to recruit aPKC, which
in turn phosphorylates and inhibits Fz1. Accordingly, components of the aPKC
complex and dPatj produce PCP defects in the eye. During PCP
signaling, aPKC and dPatj are downregulated, while Bazooka is upregulated,
suggesting an antagonistic effect of Bazooka on dPatj/aPKC. A model is proposed
whereby the dPatj/aPKC complex regulates PCP by inhibiting Fz1 in cells where it
should not be active (Djiane, 2005).
The C tail of Fz receptors regulates their localization and signaling activity. A short Fz Cterm governs apical localization, which is critical for effective Fz-PCP signaling. In contrast, a long Cterm (like that of Fz2) governs baso-lateral localization, promoting β-catenin signaling and preventing PCP activity. Thus, a striking feature of all core PCP proteins, including Fz1, is their apical localization within imaginal disc cells. Fz1 colocalizes partially with several components that regulate A/B polarity such as the Crumbs/Sdt/dPatj and Baz/aPKC/Par-6 complexes within the marginal domain, even though it is also present more basally relative to these complexes (Djiane, 2005).
Detailed sequence analysis of the Fz1 Cterm has revealed the presence of
two clustered conserved PKC phosphorylation sites (Ser554 and Ser560 in Fz1).
Given that aPKC expression in the apical domain overlaps with Fz1, a
test was performed to see if aPKC can phosphorylate the Fz1 Cterm on the
two conserved PKC sites in an in vitro kinase assay. Purified human aPKC protein phosphorylates in vitro a GST::Fz1 Cterm fusion protein.
Furthermore, mutations of the two PKC consensus sites (Ser to Ala) prevent
aPKC-mediated Fz1 phosphorylation, confirming that these sites are targets of aPKC (Djiane, 2005).
To investigate the importance of these phosphorylation sites in vivo, flies were generated carrying UAS-inducible transgenes of Fz1 mutant derivatives with either
both serines mutated to alanine (Fz1-AA), inactivating the two prospective PKC
sites, or both Serines mutated to Glutamic acid (Fz1-EE), mimicking
phosphorylation. These transgenes were analyzed under sevenless
(sev)-Gal4 control, which is expressed specifically in R3/R4
precursor cells just posterior to the MF during PCP establishment.
Overexpression of wild-type Fz1 provides too much activity and interferes with
the balance of Fz1 regulation within the R3/R4 pair, resulting in ommatidia with random R3/R4 cell fate decision and chirality, as well as symmetrical R3/R3 type ommatidia. Similarly, overexpression of Fz1-AA (with both
aPKC sites inactivated; sev>Fz1-AA) induces ommatidia with random
chirality and symmetrical clusters. In contrast,
the phosphomimetic Fz1-EE (sev>Fz1-EE) shows hardly any effect. These data suggest that aPKC-mediated Fz1
phosphorylation inhibits Fz-PCP signaling activity (Djiane, 2005).
Since apical Fz1 localization is critical for its proper PCP signaling activity (Wu, 2004), it was hypothesized that
the Fz1-EE mutation could affect the localization of the receptor. To
investigate this possibility, the expression of the different myc
tagged Fz1 transgenes was examined in imaginal discs (under en-Gal4 or dpp-Gal4 control). No difference between the expression of either Fz1-AA or
Fz1-EE with that of wild-type Fz1 was found. These mutant Fz1 isoforms were expressed at similar levels and colocalized apically with aPKC,
indicating that phosphorylation of Fz1
by aPKC does not affect Fz1's localization (Djiane, 2005).
A second feature critical
to Fz1 signaling activity is its ability to recruit Dsh to the membrane.
Interestingly, the aPKC sites partly overlap with the region of the Fz Cterm
known to bind Dsh, raising the possibility that phosphorylation by aPKC could interfere with Dsh recruitment. Thus tests were performed to see whether Dsh recruitment is affected by phosphorylation of the Fz1 aPKC sites. S2 cells, which have no endogenous Fz, were transfected with Dsh-GFP
and the different Fz1 mutants. In this assay, wild-type and both mutant forms of Fz1 recruit Dsh-GFP efficiently to the membrane. Then, whether
overexpression of Fz1-AA and Fz1-EE can recruit Dsh-GFP to apical membranes like
wild-type Fz1 in vivo (expressed with en-Gal4 in the posterior
compartment of wing discs, where there is a sharp boundary between expressing
and nonexpressing cells) was examined. Both Fz1-AA and Fz1-EE
behave like wild-type Fz1, sequestering Dsh to the apical cell membrane in
imaginal disc cells (Djiane, 2005).
In summary, it was shown that Fz1 can be phosphorylated in
vitro by aPKC. Together with the in vitro results, the in vivo analysis of a
phosphomimetic Fz1 mutant suggests that aPKC phosphorylation regulates Fz1
activity negatively and that this effect is not mediated by affecting Fz1
localization or Fz1-mediated Dsh membrane recruitment (Djiane, 2005).
In light of the importance of the aPKC sites in Fz1, it was of interest to determine how aPKC is recruited to the
receptor. This could be mediated by direct binding or through a bridging factor,
the most likely candidates being the A/B determinants that bind aPKC. Thus a two-hybrid interaction screen was conducted using the Fz1 Cterm as bait and components of different A/B protein
complexes as prey. The closely related Fz2 Cterm was included as well as the Stbm
Cterm as control baits. All components of the aPKC/Par-6/Bazooka
apical complex were tested. For Baz, three different fragments were used: an N-terminal fragment involved in Baz dimerization (BazA), a central fragment with three PDZ domains involved in Par-6 binding (BazB), and a C-terminal fragment that
binds aPKC (BazC). Similarly, all components of the Crb/Sdt/dPatj apical complex were tested except Crb (since Crb
is a transmembrane protein) and the components of the more baso-lateral
Scrib/Dlg/Lgl complex were analyzed. No direct interaction was detected between the Fz1 Cterm and aPKC, but, interestingly, Patj was found to be a specific binding partner of the Fz1 Cterm. No other protein was found to
interact with the Fz1 Cterm, and in turn Patj did not interact with the Fz2 or
Stbm Cterms. The interaction between the Stbm Cterm and Dlg was confirmed as was the interaction of aPKC with Par-6 and BazC (Djiane, 2005)
To confirm the Patj-Fz1 interaction in vivo, coimmunoprecipitation (CoIP) experiments were performed from Drosophila S2 cell extracts transfected with GFP fusion proteins with the Fz1 or Fz2 Cterms (GFP::Fz1 and GFP::Fz2, respectively). Patj could be co-immunoprecipitated
from cells transfected with GFP::Fz1 but importantly not with
GFP::Fz2 or GFP alone , demonstrating that Fz1
and Patj interact in Drosophila cells. A weak interaction was found
between Fz1 and endogenous Baz and aPKC, suggesting the existence of one or
several multiprotein complexes among Fz1, Patj, Baz, and aPKC. In contrast, other components of A/B protein complexes, such as Par-6 or Dlg, did not CoIP with either GFP::Fz1 or GFP::Fz2 (Djiane, 2005).
To map the Patj interaction domain with the
Fz1 Cterm, GST pull-down experiments were performed. Patj is a modular protein
containing a N-terminal L27 domain (previously referred to as MRE), mediating
its interaction with Sdt, and four PDZ domains.
Consistent with the yeast two-hybrid and CoIP results, in
vitro translated full-length Patj bind the GST-Fz1 Cterm protein. The fourth PDZ
domain of Patj is sufficient for direct binding to the Fz1 Cterm (Djiane, 2005).
Using the CoIP approach, the residues in the Fz1 Cterm required for Patj
interaction were also mapped. Whereas the
full-length Fz1 Cterm interacts with Patj, removing the last three residues
(Fz1ΔBS) abolishes this interaction. The removal of an internal Cterm
motif, encompassing the tryptophan critical for Dsh binding,
retains Fz1 ability to bind Patj albeit to a lesser extent (Keyes, 2005).
These results support a direct interaction between the apical determinant Patj and the Fz1 Cterm and suggest that Patj could provide a link between Fz1 and aPKC, since Patj was shown in vitro to bind to aPKC either directly or indirectly through Par-6. The Fz1/Patj interaction is mediated by
the fourth PDZ domain of Patj, requires the last three residues of Fz1, and is
largely independent of the Fz motif that mediates Dsh binding (Djiane, 2005).
Apical localization is critical for PCP protein activity and
particularly for Fz1, but until now no direct link between A/B polarity and PCP
establishment has been described. This study shows that the apical determinants
aPKC and Patj negatively regulate Fz-PCP signaling while Bazooka antagonizes
this regulation. Patj binds directly to the Fz1 cytoplasmic tail, possibly
recruiting aPKC, whose phosphorylation of two serine residues within the Fz1
Cterm inhibits the activity of the receptor in cells where signaling should not
occur. This reveals a direct link
between A/B polarity determinants and PCP establishment (Djiane, 2005).
This work provides the first evidence for a direct molecular link between A/B polarity
determinants and PCP by demonstrating that the apical determinants aPKC, Patj,
and Baz regulate Fz1 activity. This regulation is independent of Fz1 recruitment
to the apical membrane, however, since none of the tested A/B determinants is
actively responsible for it. For instance, deleting the Patj binding site in
Fz1 or replacing the Fz1 Cterm for a shortened Fz2 Cterm, which cannot bind
Patj, has no effect on Fz1 apical localization (Wu, 2004), excluding Patj as a
recruiting or targeting factor in Fz1 apical localization. Furthermore, Fmi
apical localization is unaffected in Patj and Baz mutants.
Thus, although an intact A/B polarity is a prerequisite for PCP signaling, there
is no mutual dependency for localizing the Patj/aPKC and the Fz-PCP complexes
to the apical side of imaginal disc cells, where they can functionally
interact (Djiane, 2005).
Other studies also support the existence of a link between A/B
polarity and PCP. In the mouse, Looptail (Lp), the homolog of the
Drosophila PCP gene stbm/Vang, interacts genetically with
mScribble, a baso-lateral determinant conserved in flies. In particular, transheterozygous Lp/mScribble mice show PCP defects in the inner ear. In Drosophila, it has also been shown that PCP factors interact with A/B determinants. Recent work in the sensory organ precursor (SOP) cells has shown
that the orientation of the two opposing domains of Dlg (anterior) and Baz
(posterior) is dependent on Stbm and Fz activity (Djiane, 2005).
The downregulation of aPKC and Patj
in the R3/R4 cells when Fz1 signals to induce PCP is consistent with a model
whereby inhibitory phosphorylation of Fz1 mediated by aPKC is occurring
throughout all eye disc cells, except in those that are required for PCP
establishment at the time Fz1 signaling occurs. Fz1 activity is therefore always
kept low outside of the PCP signaling window, and a release of this inhibition
is required for PCP signaling to take place. It is noteworthy that
overexpression of Fz1 always gives a robust GOF effect without requiring
additional “input,” arguing that either the receptor is
constitutively active or that a ligand is always present in nonlimiting amounts.
In either scenario, it would be important to control Fz1 activity to prevent
signaling at the wrong time and to allow limiting signaling components, such as
Dsh, to be available for canonical Wnt/Fz-β-cat signaling when PCP
signaling is not needed (Wu, 2004). This is particularly true in the eye disc, where cell fate
determination and PCP occur almost simultaneously within a short time window. It is
thus proposed that the downregulation of aPKC/Patj in the R3/R4 precursors, at
the time of PCP establishment, determines
when and in which cells Fz1 is active. A detailed
analysis of the expression of Fz1 and Fmi in the non- R3/R4 cells reveals that
they extend more basally than aPKC and Patj. Since
the precise localization of the active Fz1 is unknown, it is
possible that either another mechanism inactivates Fz1 more basally or that
inactivation is not needed there (Djiane, 2005).
Furthermore, these results argue that high Baz
levels in R3/R4 cells promote Fz1 signaling, possibly by antagonizing the
inhibitory regulation of Fz1 by aPKC.
Indeed, several lines of evidence suggest an inhibitory role of Baz on the
activity of an aPKC complex. (1) In Drosophila embryonic neuroblasts,
aPKC phosphorylates Lgl on the apical side of the cell to inhibit its function,
restricting the active Lgl to the basal domain of the cell. This is mediated
through direct binding of a Par-6/aPKC complex to Lgl, which can only occur
after Baz is released from the Par-6/aPKC complex, suggesting a negative role of
Baz on aPKC function.
(2) Direct measurements of aPKC kinase activity on an exogenous
substrate reveal that addition of purified Par-3, the vertebrate Baz homolog,
inhibits aPKC kinase activity, whereas Par-6 enhances it. However, whether the aPKC
inhibition by Par-3 is direct or indirect remains unclear.
This antagonizing role
of Bazooka on the aPKC-mediated inhibition of Fz1 activity in R3/R4 cells is
further evidence of the tight regulation required for PCP establishment in the
eye (Djiane, 2005).
In this model, the A/B determinants are acting upstream of PCP.
Consistent with this, there is no effect on either aPKC or Patj expression in
cell clones mutant for PCP genes. Similarly, the initial Baz enrichment in R3/R4
precursors is Fz/PCP independent. The later enrichment of Baz in R4 is, however,
dependent on PCP signaling. This could correspond to a similar situation as
observed in the SOP, in which the posterior relocalization of Baz is dependent
on Fz1 activity (Djiane, 2005).
How does aPKC regulate Fz-PCP
activity? The aPKC-mediated phosphorylation of the Fz1 Cterm inhibits its
activity without affecting its apical localization or ability to recruit Dsh.
The negative regulation must therefore occur
by a different mechanism. One possibility is that Fz1 phosphorylation by aPKC
inhibits a PCP-specific signal transduction to Dsh. Consistent with this
hypothesis, similar point mutations in the conserved PKC sites of the canonical
Wnt/β-cat-dedicated Fz2 (Fz2-AA and Fz2-EE), do not affect Fz2 ability to
trigger a Wnt/β-cat response when overexpressed in the wing.
Another possibility is that aPKC regulates Fz1 activity by
promoting its destabilization or by increasing its turnover through the
recycling pathway at the apical membrane. Further investigation will be required
to answer these questions (Djiane, 2005).
The selective downregulation of Patj and upregulation of Baz in R3/R4 precursors
define when and where Fz1, and therefore Fz-PCP signaling, is active. This
scenario represents a permissive rather than an instructive requirement of aPKC,
Patj, and Baz during PCP. Fz-PCP signaling components are widely expressed but
only required at specific time points and in specific subsets of cells. As no
activating PCP specific ligand is known, it is possible that alternate
mechanisms control their activity. This study provides evidence for a negative regulation
of PCP signaling by A/B polarity determinants,
unveiling new mechanisms for regulating PCP. In addition to their importance
during A/B polarity, a function has been revealed for the apical determinants
Patj, Baz, and aPKC in regulating PCP and evidence is provided for a molecular link
between apical-basal and planar cell polarity (Djiane, 2005).
Cdc42 recruits Par-6-aPKC to establish cell polarity from worms to mammals. Although Cdc42 is reported to have no function in Drosophila neuroblasts, a model for cell polarity and asymmetric cell division, this study shows that Cdc42 colocalizes with Par-6-aPKC at the apical cortex in a Bazooka-dependent manner, and is required for Par-6-aPKC localization. Loss of Cdc42 disrupts neuroblast polarity: cdc42 mutant neuroblasts have cytoplasmic Par-6-aPKC, and this phenotype is mimicked by neuroblast-specific expression of a dominant-negative Cdc42 protein or a Par-6 protein that lacks Cdc42-binding ability. Conversely, expression of constitutively active Cdc42 leads to ectopic Par-6-aPKC localization and corresponding cell polarity defects. Bazooka remains apically enriched in cdc42 mutants. Robust Cdc42 localization requires Par-6, indicating the presence of feedback in this pathway. In addition to regulating Par-6-aPKC localization, Cdc42 increases aPKC activity by relieving Par-6 inhibition. It is concluded that Cdc42 regulates aPKC localization and activity downstream of Bazooka, thereby directing neuroblast cell polarity and asymmetric cell division (Atwood, 2007).
Little is currently known about how the Par complex is localized or regulated in Drosophila neuroblasts, despite the importance of this complex for neuroblast polarity, asymmetric cell division and progenitor self-renewal. This study shows that Cdc42 plays an essential role in regulating neuroblast cell polarity and asymmetric cell division. Baz localizes Cdc42 to the apical cortex where it recruits Par-6-aPKC, leading to polarization of cortical kinase activity that is essential for directing neuroblast cell polarity, asymmetric cell division, and sibling cell fate (Atwood, 2007).
Asymmetric aPKC kinase activity is essential for the restriction of components such as Mira and Numb to the basal cortex. The aPKC substrates Lgl and Numb are thought to establish basal polarity either by antagonizing activity of myosin II or by direct displacement from the cortex. This study found that Cdc42 recruits Par-6-aPKC to the apical cortex and that Cdc42 relieves Par-6 inhibition of aPKC kinase activity. In the absence of Cdc42, aPKC is delocalized and has reduced activity, resulting in uniform cortical Mira. Expression of Cdc42-DN leads to cortical overlap of inactive Par-6-aPKC and Mira indicating the importance of Cdc42-dependent activation of aPKC kinase activity. Expression of Cdc42-CA leads to cortical aPKC that displaces Mira from the cortex, presumably because Lgl is phosphorylated at the entire cell cortex. This is similar to what is seen when a membrane-targeted aPKC is expressed (Atwood, 2007).
Baz, Par-6 and aPKC have been considered to be part of a single complex (the Par complex). This study found that, when Cdc42 function is perturbed, Par-6 and aPKC localization is disrupted but Baz is unaffected. Why is Baz unable to recruit Par-6-aPKC in the absence of Cdc42? One explanation is that Cdc42 modulates the Par-6-Baz interaction, although Cdc42 has no direct effect on Par-6-Baz affinity. Alternatively, Baz might only be transiently associated with the Par-6-aPKC complex (e.g. as an enzyme-substrate complex); this is consistent with the observation that Baz does not colocalize with Par-6-aPKC in Drosophila embryonic epithelia and its localization is not dependent on either protein. How does Baz recruit Cdc42 to the apical cortex? Like other Rho GTPases, Cdc42 is lipid modified (prenylated), which is sufficient for cortical localization. Baz is known to bind GDP-exchange factors (GEFs), which may induce accumulation of activated Cdc42 at the apical cortex (Atwood, 2007).
The requirement of Par-6 for robust Cdc42 apical enrichment suggests that positive feedback exists in this pathway, a signaling pathway property that is also found in polarized neutrophils. More work is required to test the role of feedback in neuroblast polarity but one attractive model is that Baz establishes an initial polarity landmark at the apical cortex in response to external cues, which leads to localized Par-6-aPKC activity through Cdc42. Phosphorylation of Baz by aPKC might further increase asymmetric Cdc42 activation, perhaps by increased GEF association, thereby reinforcing cell polarity. Such a mechanism could generate the robust polarity observed in neuroblasts and might explain why expression of dominant Cdc42 mutants late in embryogenesis does not lead to significant defects in polarity (Atwood, 2007).
This study argues that Cdc42 functions downstream of Baz. Cdc42 is required for Baz-Par-6-aPKC localization in C. elegans embryos and mammalian neural progenitors. In C. elegans embryos, RNA interference of cdc42 disrupts Par-6 localization, whereas PAR-3 localization is slightly perturbed. In this case, Cdc42 is required for the maintenance but not establishment of PAR-3-Par-6 asymmetry; however, other proteins have been shown to localize Par complex members independently of Cdc42. Conditional deletion of cdc42 in the mouse brain causes significant Par-3 localization defects, although this may be caused by the loss of adherens junctions. More work will be required in these systems to determine if the pathway that has been proposed is conserved (Atwood, 2007).
This study has identified at least two functions of Cdc42 in neuroblasts: first, to recruit Par-6-aPKC to the apical cortex by direct interaction with its CRIB domain and, second, to promote aPKC activity by relieving Par-6 repression. aPKC activity is required to partition Mira and associated differentiation factors into the basal GMC; this ensures maintenance of the apical neuroblast fate as well as the generation of differentiated neurons. Polarized Cdc42 activity may also have a third independent function in promoting physically asymmetric cell division, because uniform cortical localization of active Cdc42 leads to same-size sibling cells. Loss of active Cdc42 at the cortex by overexpression of Cdc42-DN still results in asymmetric cell division, suggesting that other factors also regulate cell-size asymmetry, such as Lgl and Pins. In conclusion, these data show that Cdc42 is essential for the establishment of neuroblast cell polarity and asymmetric cell division, and defines its role in recruiting and regulating Par-6-aPKC function. These findings now allow Drosophila neuroblasts to be used as a model system for investigating the regulation and function of Cdc42 in cell polarity, asymmetric cell division and neural stem cell self-renewal (Atwood, 2007).
To see where aPKC is expressed during embryonic development, the mRNA distribution was analyzed by RNA in situ hybridization. aPKC mRNA is already detectable in freshly laid eggs before the onset of zygotic transcription and thus must be deposited maternally during oogenesis. At the cellular blastoderm stage, aPKC mRNA is present in all cells except for the pole cells. During gastrulation, strong expression of aPKC is detectable in tissues that undergo morphogenetic movements; e.g., the invaginating mesoderm, the proctodeum, and the cephalic furrow. In embryos at the extended germ band stage, prominent aPKC expression is detectable in neuroblasts. In several epithelial tissues, in particular in the fore- and hind-gut and in the Malpighian tubules, aPKC mRNA is highly enriched in the apical cytocortex, reminiscent of the polarized localization of baz and crumbs (crb) mRNAs (Wodarz, 2000).
The distribution of aPKC protein was analyzed using anti-PKCzeta antibody C20. During cellularization of the embryo, aPKC becomes localized to the apical cytocortex of all cells except for the pole cells, which do not contain detectable amounts of aPKC. Already at this stage aPKC is enriched at apico-lateral cell borders, giving rise to a honeycomb pattern in en face views of the blastoderm. After completion of cellularization, DaPKC is highly concentrated in the apico-lateral cortex and shows little overlap with the basolateral marker Nrt. Apical localization is maintained throughout embryonic development in most epithelia that are derived from the ectoderm (e. g. epidermis, fore-, and hind-gut), Malpighian tubules, and the tracheal system. The only ectodermal epithelium devoid of DaPKC expression is the amnioserosa (Wodarz, 2000).
Strong expression of DaPKC is also detected in neuroblasts, the stem cells of the embryonic CNS. During delamination of neuroblasts, aPKC is localized in the apical stalk that is wedged between adjacent cells of the neuroectodermal epithelium. In pro- and meta-phase, aPKC forms apical cortical crescents. In anaphase, aPKC staining is strongly diminished and expands over a broader region of the neuroblast cortex, but is clearly excluded from the budding ganglion mother cell. Thus, from delamination through pro- and meta-phase, localization of DaPKC in neuroblasts is very similar to that of Baz and Insc. Simultaneously with aPKC, cell outlines of epithelial cells and neuroblasts were visualized with the Nrt antibody. Similar to epithelial cells, Nrt expression is clearly polarized in neuroblasts. Strong staining for Nrt is detectable in the basal and lateral membrane of neuroblasts, whereas staining is strongly reduced in apical regions where DaPKC is expressed (Wodarz, 2000).
Cell polarity is critical for epithelial structure and function. Adherens
junctions (AJs) often direct this polarity, but it has been found that Bazooka
(Baz) acts upstream of AJs as epithelial polarity is first established in
Drosophila. This prompted an investigation into how Baz is positioned and how downstream polarity is elaborated. Surprisingly, it was found that Baz localizes to an apical domain below (basally to) its typical binding partners atypical protein kinase C (aPKC) and
partitioning defective (PAR)-6 as the Drosophila epithelium first forms. In
fact, Baz positioning is independent of aPKC and PAR-6 relying instead on
cytoskeletal cues, including an apical scaffold and dynein-mediated
basal-to-apical transport. AJ assembly is closely coupled to Baz positioning,
whereas aPKC and PAR-6 are positioned separately. This forms