apterous

cis-Regulatory Sequences and Functions

Hox genes control segment identity in the mesoderm as well as in other tissues. Most evidence indicates that Hox genes act cell-autonomously in muscle development, although this remains a controversial issue. apterous expression in the somatic mesoderm is under direct Hox control. A small enhancer element of apterous (apME680) has been identified that regulates reporter gene expression in the LT1-4 muscle progenitors. The product of the Hox gene Antennapedia is present in the somatic mesoderm of the second and third thoracic segments. Through complementary alterations in the Antennapedia protein and in its binding sites on apME680, it has been shown that Antennapedia positively regulates apterous in a direct manner, demonstrating unambiguously its cell-autonomous role in muscle development. LT1-4 muscles contain more nuclei in the thorax than in the abdomen and it is proposed that one of the segmental differences under Hox control is the number of myoblasts allocated to the formation of specific muscles in different segments (Capovilla, 2001).

A fragment of 680 bp, located in the second largest ap intron, is capable of directing lacZ reporter expression starting from stage 10 in clusters of cells very similar to those expressing ap at this stage. This fragment is called apME680 (for ap-muscle-enhancer-680) because it directs muscle-specific reporter gene expression. At stage 13, beta-galactosidase is detected in one continuous cluster in T2 and T3, while two smaller clusters, located at the dorsal and ventral limits of the thoracic clusters, are detected in segments A1-A7. In segment A8, a unique smaller cluster is detected. These beta-galactosidase-positive cells contribute to the formation of muscles LT1-4 in segments T2-A7 and to muscle LT1 in A8. These are a subset of the muscles originating from ap-expressing cells, since ap is expressed also in the progenitors of muscles VA2 and VA3. Thoracic muscles LT1-4 are differ slightly from the same abdominal muscles. In particular, muscle LT4 extends more dorsally and ventrally in the thorax than in the abdomen (Capovilla, 2001).

The question of the significance of the homeotic regulation of ap by Antp was addressed. The perdurance of beta-galactosidase allows the labeling of thoracic and abdominal LT1-4 mature muscles originating from the cells expressing ap starting from the early germ band extended stage. LT1-4 muscles present different characteristics in the thorax and in the abdomen. In the thorax, they contain more beta-galactosidase, they are more tightly packed and, at least in the case of muscle LT4, extend more dorsally and ventrally. These differences may be a consequence of more myoblasts contributing to the thoracic muscles than to the corresponding abdominal muscles. To investigate this hypothesis, double labeling experiments were performed using anti-beta-galactosidase to label muscles LT1- 4 and anti-MEF2 antibodies, which label all muscle nuclei. In wild-type embryos, LT1-4 thoracic muscles do contain more MEF2-positive nuclei than the same abdominal muscles. The number of nuclei was compared in the T2, T3 and A1 hemisegments of ten independent embryos. This quantitative analysis shows that, on average, T3 muscles contain a total of 28 nuclei, while A1 muscles contain 19 nuclei. This difference is statistically significant. No significant differences were observed between the number of nuclei in T2 and T3. Consistently, highly packed nuclei are present in the medial portion of T2 and T3 muscles, but are absent in the same region of abdominal muscles (Capovilla, 2001).

Enhancer blocking and transvection at the Drosophila apterous locus

Intra- and interchromosomal interactions have been implicated in a number of genetic phenomena in diverse organisms, suggesting that the higher-order structural organization of chromosomes in the nucleus can have a profound impact on gene regulation. In Drosophila, homologous chromosomes remain paired in somatic tissues, allowing for trans interactions between genes and regulatory elements on the two homologs. One consequence of homolog pairing is the phenomenon of transvection, in which regulatory elements on one homolog can affect the expression of a gene in trans. This paper reports a new instance of transvection at the Drosophila apterous (ap) locus. Two different insertions of boundary elements in the ap regulatory region were identified. The boundaries are inserted between the ap wing enhancer and the ap promoter and have highly penetrant wing defects typical of mutants in ap. When crossed to an ap promoter deletion, both boundary inserts exhibit the interallelic complementation characteristic of transvection. To confirm that transvection occurs at ap, a deletion of the ap wing enhancer was generated by FRT-mediated recombination. When the wing-enhancer deletion is crossed to the ap promoter deletion, strong transvection is observed. Interestingly, the two boundary elements, which are inserted ~10 kb apart, fail to block enhancer action when they are present in trans to one another. This study demonstrates that this is unlikely to be due to insulator bypass. The transvection effects described here may provide insight into the role that boundary element pairing plays in enhancer blocking both in cis and in trans (Gohl, 2008).

This study presents evidence for transvection at the Drosophila apterous locus. While interallelic complementation at ap has been previously reported, the ap alleles were not molecularly characterized. Consequently, it was not clear whether the complementation between these alleles involved trans-regulatory interactions or occurred at the level of the mutant ap gene products. This study has observed trans-regulatory interactions with several different classes of ap mutations (Gohl, 2008).

The first type is the transvection seen in trans combinations between mutations that disrupt enhancers and mutations that disrupt the promoter. At the ap locus, this is illustrated by the ap^DG/ap^UGO35 combination. Interestingly, the transvection observed between ap^DG and ap^UGO35 is sufficient to express ap at or near wild-type levels, as >90% of the wings are completely wild type. ap mutants are recessive, so there is likely a range of ap activity that is sufficient to produce wild-type wings (Gohl, 2008).

It is unknown to what extent Dipterans have learned to exploit this interesting feature of their genomes for normal gene regulation. For example, it is unlikely that trans regulation occurs at the endogenous y locus in wild-type flies, since the enhancers appear to be strongly tethered in cis by the promoter. Instead, trans regulation is observed only at y when the enhancers are freed by deletion of the cis promoter. ap is clearly different from y in this respect, since relatively strong trans regulation is also observed when the enhancer deletion, ap^DG, is combined with presumed ap-coding region mutations that are likely to retain an intact promoter. Since the suppression of these coding region mutants by ap^DG is not as strong as that observed with the promoter deletion ap^UGO35, cis interactions between the upstream wing enhancer and the promoter of the mutant gene must compete with the ap^DG promoter in trans (Gohl, 2008).

The second type of trans-regulatory interaction observed at ap is the transvection effects observed with boundary elements. Two different boundary insertions were observed in the ap regulatory region. ap^MM-Mcp is an experimentally generated insertion of the Mcp-containing Flipper 2 transposon 403 bp upstream of the ap transcriptional start site between the wing enhancer and the ap promoter. MCP is a polycomb response element from the Drosophila bithorax complex. Although the Mcp element in this transgene contains both a boundary element and a PRE, the results indicate that the wing defects seen in homozygous or hemizygous ap^MM-Mcp flies are due to the enhancer-blocking activity of the boundary and not due to silencing by the Mcp PRE. In the absence of an Mcp boundary insertion that lacks the PRE, the possibility remains that the Mcp PRE contributes to the ap wing phenotype. However, if this is the case, it is likely that the role of the PRE is a modulatory one, as the bxd PRE alone is not sufficient to cause wing defects. ap^f00451 is a su(Hw)-containing piggyBac element and was also experimentally inserted between ap enhancer elements and the ap promoter (Gohl, 2008).

One version of this boundary-element-induced transvection is that seen in the interallelic complementation between the boundary insertions and the ap promoter deletion, ap^UGO35. This trans-regulatory interaction is observed with both the Mcp and su(Hw) elements. The Mcp insert, ap^MM-Mcp, has a strong ap wing phenotype, but when it is combined with the promoter deletion, ap^UGO35, the wing defects are partially suppressed. The fact that full suppression is not observed in this combination, while it is observed when the enhancer deletion is combined with the promoter deletion, indicates that the Mcp element must be capable of partially blocking trans interactions between the ap^UGO35 wing enhancers and the ap^MM-Mcp promoter. This suggestion is substantiated by a comparison of the wing phenotypes in combinations between ap^UG035 and the enhancer deletion with (ap^DG-Mcp) and without (ap^DG) the Mcp element. While nearly full suppression is observed in the latter case, the suppression of the wing defects in ap^DG-Mcp/ap^UGO35 flies is comparatively modest. This difference can be attributed to the ability of the Mcp element to block the ap enhancers in trans from activating the ap promoter in cis to the boundary. In contrast, a comparison of the wing phenotype of the ap^DG-Mcp/ap^UGO35 trans combination with flies that are either hemizygous or homozygous for the Mcp insertion, ap^MM-Mcp, reveals that the enhancer-blocking activity of this boundary element is stronger when the enhancer and promoter are in cis than when they are in a trans configuration (Gohl, 2008).

The other version of boundary-element-induced transvection that was observed is the trans combination between the boundary insertions and the ap wing-enhancer deletion, ap^DG. This combination was tested for the Mcp and su(Hw) inserts and in both cases the wing phenotype of the enhancer deletion was suppressed. Since the extent of suppression in both cases is considerably less than seen when the enhancer deletion ap^DG is combined with the promoter deletion ap^UGO35, it would appear that the boundary in cis to the enhancer is able to partially block its interactions with the ap promoter in trans. As noted above, the converse is also true: boundary elements in trans to the enhancer are able to partially block interactions with the ap promoter in cis (Gohl, 2008).

Since these results demonstrate that the Mcp and su(Hw) boundaries can act not only in cis but also in trans, one might predict either that no interallelic complementation would be observed when two different boundary inserts are combined or that the phenotype would actually become even stronger because of the ability of boundaries to inhibit regulatory interactions in trans. Surprisingly, however, neither of these expectations holds. Instead, flies trans-heterozygous for the Mcp insert ap^MM-Mcp, and the su(Hw) insert ap^f00451 have completely wild-type wings. One mechanism that could account for this unexpected result is insulator bypass. Studies on the su(Hw) insulator have shown that enhancer-blocking activity is neutralized when there are two copies of this element in tandem between the enhancer and the promoter. While bypass is thought to involve su(Hw)-pairing interactions, other insulators, including Mcp, can be substituted for one of the two su(Hw) elements. A strong prediction of the insulator bypass model is that interallelic complementation should also be observed when the su(Hw) element in ap^f00451 is in trans to the enhancer deletion that retains an intact Mcp element, ap^DG-Mcp. However, this is not the case as the wing phenotype of ap^DG-Mcp/ap^f00451 trans-heterozygotes is the same as that of ap^f00451 alone. This result indicates that the Mcp element is able to prevent trans activation of the ap promoter in cis by the wing enhancers on the ap^f00451 chromosome. The ability to block enhancers on the trans chromosome from contacting the promoter in cis to a boundary element was also observed when ap^MM-Mcp is combined with the Mcp-containing enhancer deletion ap^DG-Mcp (Gohl, 2008).

Thus, the interallelic complementation observed in ap^MM-Mcp/ap^f00451 flies is not likely to be an instance of insulator bypass. Instead, it seems that the additive effects of the unblocked, ap proximal portion of the ap^f00451 enhancer and trans activation by the enhancer on the ap^MM-Mcp chromosome can account for the wild-type wings of ap^MM-Mcp/ap^f00451 flies (Gohl, 2008).

Including the studies reported in this study on boundary insertions in the ap locus, there are now several examples in which the blocking activity of a boundary element can be partially bypassed by interactions between enhancers on one chromosome and the target gene/promoter on the other chromosome. These findings raise the question of why boundary elements are more permissive for regulatory interactions in trans than they are for interactions in cis (Gohl, 2008).

Answering this question depends upon how enhancers communicate with promoters and how boundaries block this communication. Two general models have been proposed to explain how enhancers interact with their target promoters. In the first model, the enhancer (or an activator molecule recruited by the enhancer) processively tracks along the chromosome (perhaps modifying the intervening chromatin) until it encounters the promoter. In this model, boundary elements function as roadblocks (or 'promoter decoys'), stopping the tracking activator and/or the spread of active chromatin. As this model requires the enhancer to act in cis, it is difficult to reconcile it with the phenomenon of transvection, which depends upon regulatory interactions occurring in trans. In addition, if transvection is explained in this model by postulating that the tracking activator skips from one paired chromosome to the other, then it is hard to understand how a boundary element would ever be able to prevent an enhancer from activating a promoter since an activator molecule that can skip freely in trans should also be able to skip over a boundary in cis (Gohl, 2008).

The second model, which is strongly supported by recent studies, hypothesizes that the sliding of the chromatin fiber against itself within a higher-order chromatin domain brings the enhancer and promoter into contact while looping out the intervening DNA. This is more easily reconciled with transvection since the enhancer could interact with a promoter in trans by a similar sliding-looping mechanism as long as the chromatin fibers of the two chromosomes are paired. Indeed, chromosomal rearrangements that disrupt pairing also tend to disrupt transvection. In this model, boundary elements prevent enhancer-promoter contact by isolating the enhancer and the promoter from each other in topologically independent looped domains. It is thought that boundaries generate topologically independent looped domains through pairing interactions with the neighboring boundaries (or by interacting with some fixed structure such as the nuclear matrix). This mechanism is supported by studies on su(Hw), scs/scs', and several boundaries from the Drosophila BX-C. For example, pairing between tandem su(Hw) insulators neutralizes their boundary function, enabling an upstream enhancer to activate a downstream promoter. According to this model for enhancer blocking, the Mcp [or su(Hw)] boundary would isolate the ap wing enhancer from the ap promoter in cis through interactions with the hypothetical upstream and downstream boundaries that define the ap domain (Gohl, 2008).

This mechanism for boundary function in cis still leaves open the question of why boundaries can be partially bypassed in trans. One possibility is that pairing interactions between boundaries occur not only in cis but also in trans. In this model, the arrangement of loop domains would be the same on each chromosome when they both contain the Mcp or su(Hw) boundary insert-there would be two loops, one containing the ap enhancer and the other containing the ap promoter. These loops would be generated by interactions between Mcp and the neighboring proximal and/or distal boundaries. The situation would be more complicated when one chromosome has the boundary element insertion and the other does not. In this case, the wild-type chromosome should have a single ap loop containing both the enhancer and the promoter, while the chromosome containing Mcp should have two loops, one containing the enhancer and the other the promoter. However, this arrangement of loops on the two chromosomes might be dynamically unstable if trans-boundary interactions also tend to stabilize cis contacts between the boundary elements that flank the ap locus. This dynamic instability could disrupt or weaken cis interactions between Mcp and the boundaries flanking the ap locus. In this case, the arrangement of loops on the Mcp-containing chromosome might switch back and forth from two to one, permitting a partial bypass of Mcp through trans-regulatory interactions (Gohl, 2008).

While both the Mcp and su(Hw) boundary elements can be partially bypassed by interactions between the ap enhancer and promoter in trans, trans interactions do not occur when the same boundary insertion is present on both homologs. In contrast, when the Mcp and su(Hw) boundary insertions are present in trans on the two chromosomes (ap^MM-Mcp/ap^f00451), this seems to abrogate their blocking activity. One explanation for this effect is that Mcp and su(Hw) are unable to interact with each other; however, it was previously demonstrated that su(Hw) and Mcp can pair with one another, possibly through the interaction of GAGA factor and Mod(mdg4). Since the Mcp and su(Hw) boundary insertions are located at distant sites within the ap locus, another possibility is that the pairing of the two structurally dissimilar alleles in this arrangement results in conformational stress that precludes the formation of stable Mcp/su(Hw) interactions either with each other or with the hypothetical flanking ap boundaries. In this model for enhancer bypass of the Mcp and su(Hw) boundaries in trans, homologous pairing between sequences in the ap locus would loop out the transposons containing the Mcp and su(Hw) boundary elements, preventing them from blocking enhancer-promoter contacts. An alternative possibility is that boundary interactions occur only in pairwise combinations. Thus, instead of interacting simultaneously with the boundaries that flank the ap locus, Mcp and su(Hw) might be paired only with either the upstream or the downstream ap boundary at a given time. If the pairing of Mcp and su(Hw) with the flanking boundaries occurs independently [or if Mcp and su(Hw) differ in their pairing preferences], either of these distinct domains might be predicted to confer enhancer blocking to both homozygous or hemizygous flies. However, when these two alleles are crossed together, the domains in effect would be complementary, with one unblocked enhancer and one unblocked ap gene. It may be possible to distinguish between these different models by generating new insertions into the ap locus in which the Mcp and su(Hw) boundaries are brought closer together and by substituting other boundary elements for Mcp or su(Hw) (Gohl, 2008).

Enhancer loops appear stable during development and are associated with paused polymerase

Developmental enhancers initiate transcription and are fundamental to our understanding of developmental networks, evolution and disease. Despite their importance, the properties governing enhancer-promoter interactions and their dynamics during embryogenesis remain unclear. At the β-globin locus, enhancer-promoter interactions appear dynamic and cell-type specific, whereas at the HoxD locus they are stable and ubiquitous, being present in tissues where the target genes are not expressed. The extent to which preformed enhancer-promoter conformations exist at other, more typical, loci and how transcription is eventually triggered is unclear. This study generated a high-resolution map of enhancer three-dimensional contacts during Drosophila embryogenesis, covering two developmental stages and tissue contexts, at unprecedented resolution. Although local regulatory interactions are common, long-range interactions are highly prevalent within the compact Drosophila genome. Each enhancer contacts multiple enhancers, and promoters with similar expression, suggesting a role in their co-regulation. Notably, most interactions appear unchanged between tissue context and across development, arising before gene activation, and are frequently associated with paused RNA polymerase. These results indicate that the general topology governing enhancer contacts is conserved from flies to humans and suggest that transcription initiates from preformed enhancer-promoter loops through release of paused polymerase (Ghavi-Helm, 2014).

Drosophila embryogenesis proceeds very rapidly, taking 18 h from egg lay to completion. Underlying this dynamic developmental program are marked changes in transcription, which are in turn regulated by characterized changes in enhancer activity. However, the role and extent of dynamic enhancer looping during this process remains unknown. To address this, 4C-seq (chromosome conformation capture sequencing) experiments were performed, anchored on 103 distal or promoter-proximal developmental enhancers (referred to as 'viewpoints'), and absolute and differential interaction maps were constructed for each, varying two important parameters: (1) developmental time, using embryos at two different stages, early in development when cells are multipotent (3-4 h after egg lay; stages 6-7), and mid-embryogenesis during cell-fate specification (6-8 h; stages 10-11); and (2) tissue context, comparing enhancer interactions in mesodermal cells versus whole embryo. To perform cell-type-specific 4C-seq in embryos, a modified version of BiTS-ChIP (batch isolation of tissue-specific chromatin for immunoprecipitation) was established. Nuclei from covalently crosslinked transgenic embryos, expressing a nuclear-tagged protein only in mesodermal cells, were isolated by fluorescence-activated cell sorting (FACS; (>98% purity) and used for 4C-seq on 92 enhancers at 6-8 h and a subset of 14 enhancers at 3-4 h. The same 92 enhancers, and 11 additional regions, were also used as viewpoints in whole embryos at both time points. The enhancers were selected based on dynamic changes in mesodermal transcription factor occupancy between these developmental stages and the expression of the closest gene. This study was thereby primed to detect dynamic three-dimensional (3D) interactions, focusing on developmental stages during which the embryo undergoes marked morphological and transcriptional changes (Ghavi-Helm, 2014).

All 4C-seq experiments had the expected signal distribution, with high concordance between replicates. To assess data quality further, ten known enhancer-promoter pairs (of the ap, Abd-b, E2f, pdm2, Con, eya, stumps, Mef2, sli and slp1 genes) were compared, and in all cases the expected interactions were recovered. For example, using an enhancer of the apterous (ap) gene, the expected interaction was detected with the ap promoter, 17 kilobases (kb) away, illustrating the high quality and resolution of the data (Ghavi-Helm, 2014).

In chromosome conformation capture assays, interaction frequencies decrease with genomic distance between regions. To adjust for this, the 4C signal decay was modelled as a function of distance using a monotonously decreasing smooth function. Subtracting this trend, the residual interaction signal was converted to z-scores and interacting regions defined by merging neighbouring high-scoring fragments within 1 kb. Using this stringent approach, 4,247 high-confidence interactions were identified across all viewpoints and conditions, representing 1,036 unique interacting regions (Ghavi-Helm, 2014).

Each enhancer (viewpoint) interacted with, on average, ten distinct genomic regions, less than half (41%) of which were annotated enhancers or promoters. Distal enhancers had a higher than expected interaction frequency with other enhancers. Similarly, promoter-proximal elements had extensive interactions with distal active promoters, 98% of which are >10 kb away. Enhancer-promoter interactions, although not significantly enriched, involve active promoters, with high enrichment for H3K27ac and H3K4me3, and active enhancers, defined by H3K27ac, RNA Pol II and H3K79me3. These results are similar to recent findings in human cells and the mouse β-globin locus, indicating similarities in 3D regulatory principles from flies to human (Ghavi-Helm, 2014).

The extent of 3D connectivity is surprising given the relative simplicity of the Drosophila genome. On average, each promoter-proximal element interacted with four distal promoters and two annotated enhancers, whereas each distal enhancer interacted with two promoters and three other enhancers. These numbers are probably underestimates, as 60% of interactions involved intragenic or intergenic fragments containing no annotated cis-regulatory elements. Despite this, the level of connectivity is similar to that recently observed in humans, where active promoters contacted on average 4.75 enhancers and 25% of enhancers interacted with two or more promoters. The multi-component contacts that were observed for Drosophila enhancers indicate topologically complex structures and suggest that, despite its non-coding genome being an order of magnitude smaller than humans, Drosophila may require a similar 3D spatial organization to ensure functionality (Ghavi-Helm, 2014).

Insulators, and associated proteins, are thought to have a major role in shaping nuclear architecture by anchoring enhancer-promoter interactions or by acting as boundary elements between topologically associated domains (TADs). Occupancy data from 0 to 12 h Drosophila embryos revealed a 50% overlap of interacting regions with occupancy of one or more insulator protein. Insulator-bound interactions are enriched in enhancer elements, suggesting that insulators may have a role in promoting enhancer-enhancer interactions. In contrast to mammalian cells, this study observed no association between insulator occupancy and the genomic distance spanned by chromatin loops, although there was a modest increase in average interaction strength. Conversely, 50% of interacting regions are not bound by any of the six Drosophila insulator proteins, suggesting that these 3D contacts are formed in an insulator-independent manner, or are being facilitated by neighbouring interacting regions (Ghavi-Helm, 2014).

If enhancer 3D contacts are involved in transcriptional regulation, then genes linked by interactions with a common enhancer should share spatio-temporal expression. For the four loci examined-pdm2, engrailed, unc-5 and charybde-this is indeed the case. For example, the pdm2 CE8012 enhancer interacts with both the pdm2 and nubbin (nub, also known as pdm1) promoters, located 2.5 and 47 kb away, respectively. Both genes, producing structurally related proteins, are co-expressed in the ectoderm, overlapping the activity of the pdm2 enhancer. Although there are examples of long-range interactions in Drosophila, often involving Polycomb response elements (PREs) and insulator elements, the vast majority of characterized enhancers are within 10 kb of their target gene, with few known to act over 50 kb. However, as investigators historically tested regions close to the gene of interest, characterized Drosophila enhancers are generally close to the gene they regulate. In contrast, although 4C cannot assess the full extent of short-range interactions, it provides an unbiased systematic measurement of the distance of enhancer interactions, far beyond 10 kb (Ghavi-Helm, 2014).

The distance distribution of all significant interactions reveals extensive long-range interactions within the ~180 megabase (Mb) Drosophila genome; 73% span >50 kb, with the median interaction-viewpoint distance being 110 kb. Two striking examples of long-range interactions are the unc-5 and charybde loci. The unc-5 promoter interacts with multiple regions, including a weak but significant interaction with the promoter of slit (sli), at a distance of >500 kb. These genes produce structurally unrelated proteins that are co-expressed in the heart, and are essential for heart formation (Ghavi-Helm, 2014).

A promoter-proximal element near the charybde (chrb) promoter has a strong interaction with the promoter of the scylla (scyl) gene, almost 250 kb away. Both genes are closely related in sequence and co-expressed throughout embryogenesis. These long-range interactions were confirmed by reciprocal 4C, using either the promoter of chrb or scyl, or an interacting putative enhancer as viewpoint. This interaction was further verified using DNA fluorescence in situ hybridization (FISH) in embryos. As a control, the distance was assessed between the chrb promoter (probe A) and an overlapping probe A' or a region on another chromosome (probe D), to determine the distances between regions very close or far away, respectively. Comparing the distance between the chrb and scyl promoters (probes A and B) showed a high, statistically significant co-localization, in contrast to the distance between the chrb promoter and a non-interacting region with equal genomic distance (probes A and C) (Ghavi-Helm, 2014).

The reciprocal 4C revealed several intervening interactions that are consistently associated with loops to both the scyl and chrb promoter. The activity was examined of two of these in transgenic embryos. Both interacting regions can function as enhancers in vivo, recapitulating chrb expression in the visceral mesoderm and nervous system (Ghavi-Helm, 2014).

When considering a 1-Mb scale around this region, the 4C interaction signal drops to almost zero just after the promoters of both genes. This 'contained block' of interactions is reminiscent of TADs, although the boundaries don't exactly match TADs defined at late stages of embryogenesis, which may reflect differences in the developmental stages used. However, the boundaries do overlap a block of conserved microsynteny between drosophilids spanning ~50 million years of evolution, suggesting a functional explanation underlying the maintained synteny. Expanding this analysis across all viewpoints, ~60% of interactions are located within the same TAD and the same microsyntenic domain as the viewpoint. In the case of the chrb and scyl genes, this constraint may act to maintain a regulatory association between a large array of enhancers, facilitating their interaction with both genes' promoters (Ghavi-Helm, 2014).

These examples, and the other 555 unique interactions >100 kb, provide strong evidence that long-range interactions are widely used within the Drosophila genome, potentially markedly increasing the regulatory repertoire of each gene. As enhancer-promoter looping can trigger gene expression, it follows that enhancer contacts should reflect the dynamics of transcriptional changes during development and therefore be temporally associated with gene expression. To assess this, looping interactions were directly compared between the two different time points and tissue contexts. Given the non-discrete nature of chromatin contacts, the quantitative 4C-seq signal was used to identify differential interactions based on a Gamma-Poisson model, and they were defined as having >2-fold change and false discovery rate <10% (Ghavi-Helm, 2014).

Despite the marked differences in development and enhancer activity between these conditions, surprisingly few changes were found in chromatin interaction frequencies, with ~6% of interacting fragments showing significant changes between conditions. Of these, 87 interactions were significantly reduced during mid-embryogenesis (6-8 h) compared to the early time point (3-4 h), and 90 interactions significantly increased. Similarly, 105 interactions had a higher frequency in mesodermal cells, compared to the whole embryo, and For example, a promoter-proximal viewpoint in the vicinity of the Antp promoter identified many interactions, two of which are significantly decreased at 6-8 h, although the expression of the Antp gene itself increases. For one region, the reduction in 4C interaction at 6-8 h corresponds to a loss in a H3K4me3 peak from 3-4 h to 6-8 h, suggesting that this 3D contact is associated with the transient expression of an unannotated transcript. The activity of the other interacting peak was examined in transgenic embryos, and it was shown to act as an enhancer, driving specific expression in the nervous system overlapping the Antp gene at 6-8 h. Along with the two enhancers discovered at the chrb locus, this demonstrates the value of 3D interactions to identify new enhancer elements, even for well-characterized loci like Antp (Ghavi-Helm, 2014).

A viewpoint in the vicinity of the Abd-B promoter interacted with a number of regions spanning the bithorax locus, three of which correspond to previously characterized Abd-B enhancers; iab-5, iab-7 and iab-8. The iab-7 and iab-8 enhancers are active in early embryogenesis, and have much reduced or no activity at the later time point. Notably, although the loop to those two enhancers is strong at the early time point, it becomes significantly reduced later in development, when both enhancers' activities are reduced. Conversely, the iab-5 enhancer contacts the promoter at a much higher frequency later in development, at the stage when the enhancer is most active. This locus therefore exhibits dynamic 3D promoter-enhancer contacts that reflect the transient activity of three developmental enhancers. It is interesting to note that in all loci examined, the dynamic contacts of specific elements are neighboured by stable contacts, as seen in the Antp and Abd-B loci. Dynamic changes, therefore, appear to operate in the context of larger, more-stable 3D landscapes (Ghavi-Helm, 2014).

Ninety-four per cent of enhancer interactions showed no evidence of dynamic changes across time and tissue context, which is remarkable given the marked developmental transitions during these stages. To investigate this further, enhancer-promoter interactions were examined of genes switching their expression state between time points or tissue contexts. The ap gene, for example, is not expressed at 2-4 h but is highly expressed during mid-embryogenesis (6-8 h). Despite the absence of expression, the interaction between the apME680 enhancer and the ap promoter is already present at 3-4 h, several hours before the gene's activation. To examine this more globally, differentially expressed genes, going either from on-to-off or off-to-on, were selected. Even for these dynamically expressed genes, there was no correlation with changes in their promoter-enhancer contacts. Similar 'stable' interactions were observed between tissue contexts. Genes predominantly expressed in the neuroectoderm at 6-8 h, for example, have interactions at the same locations in whole embryos and purified mesodermal nuclei at 6-8 h, despite the fact that they are not expressed in the mesoderm at this stage (Ghavi-Helm, 2014).

Pre-existing loops were recently observed in human and mouse cells, and suggested to prime a locus for transcriptional activation. However, why they are formed and how transcription is eventually triggered remains unclear. To investigate this, this study focused on the subset of genes that have both off-to-on expression and no evidence for differential interactions (20 genes; differentially expressed with stable loops (DS) genes). Despite changes in their overall expression, DS genes have similar levels of RNA polymerase II (Pol II) promoter occupancy at both time points. The presence of promoter-bound Pol II in the absence of full-length transcription is indicative of Pol II pausing. Using global run-on sequencing (GRO-seq) data to define a stringent set of paused genes, it was observed that most (75%) DS genes are paused (15 of 20 DS genes), and have a significantly higher pausing index. This percentage is significantly higher than expected by chance when sampling over all off-to-on genes, and is robust to using a strict or more relaxed) definition of Pol II pausing. This association is very evident when examining specific loci, showing Pol II occupancy, short abortive transcripts, and loop formation before the gene's expression. Taken together, these results indicate that 'stable' chromatin loops are associated with the presence of paused Pol II at the promoter (Ghavi-Helm, 2014).

To understand how transcription is ultimately activated, changes were examined in DNase I hypersensitivity at the promoter of DS genes. DNase I hypersensitivity is significantly increased at interacting promoters at the stages when the gene is expressed, suggesting that the recruitment of additional transcription factor(s) later in development might act as the trigger for transcriptional activation (Ghavi-Helm, 2014).

In summary, these data reveal extensive long-range interactions in an organism with a relatively compact genome, including pairs of co-regulated genes contacting common enhancers often at distances greater than 200 kb. Comparing enhancer contacts in different contexts revealed that chromatin interactions are very similar across developmental time points and tissue contexts. Enhancers therefore do not appear to undergo long-range looping de novo at the time of gene expression, but are rather already in close proximity to the promoter they will regulate. Within this 3D topology, highly dynamic and transient contacts would not be visible when averaging over millions of nuclei. As transcription factor binding is sufficient to force loop formation, these results suggest a model where through transcription factor-enhancer occupancy, an enhancer loops towards the promoter and polymerase is recruited, but paused in the majority of cases. The subsequent recruitment of transcription factor(s) or additional enhancers at preformed 3D hubs most likely triggers activation by releasing Pol II pausing. Such preformed topologies could thereby promote rapid activation of transcription. At the same time, as paused promoters can exert enhancer-blocking activity, the presence of paused polymerase within these 3D landscapes could safeguard against premature transcriptional activation, but yet keep the system poised for activation (Ghavi-Helm, 2014).

Transcriptional Regulation

The Wingless protein, in a role surprisingly distinct from its embryonic segment polarity function, appears to be the earliest-acting member of the hierarchy of regulatory genes that subdivide the wing disc into discrete subregions. Wingless is crucial for distinguishing the notum/wing subfields, and for the compartmentalization of the dorsal and ventral wing surfaces. wingless signaling is required to restrict the expression of the apterous gene to dorsal cells and to promote the expression of the vestigial and scalloped genes that demarcate the wing primordia and act in concert to promote morphogenesis (Williams, 1993).

Dual role for Drosophila epidermal growth factor receptor signaling in early wing disc development

Cell fate decisions in the early Drosophila wing disc assign cells to compartments (anterior or posterior and dorsal or ventral) and distinguish the future wing from the body wall (notum). Egf receptor signaling stimulated by its ligand, Vein, has a fundamental role in regulating two of these cell fate choices: (1) Vn/EGFR signaling directs cells to become notum by antagonizing wing development and by activating notum-specifying genes; (2) Vn/EGFR signaling directs cells to become part of the dorsal compartment by induction of apterous, the dorsal selector gene, and consequently also controls wing development, which depends on an interaction between dorsal and ventral cells (Wang, 2000).

To determine when Vn/EGFR signaling is required for notum development, the temperature-sensitive alleles, Egfr^tsla and vn^tsWB240 were used. Inactivating Vn/Egfr activity during the second instar (a 24 hr period) causes loss of the notum. The wing develops but shows pattern abnormalities characteristic of vn hypomorphs. Later shifts during the third instar does not cause loss of the notum. This demonstrates that Vn/Egfr activity is required for notum development in the second instar when wg is required to specify the wing. Thus, Vn and Wg appear to have complementary roles and this relationship has been examined by following their expression in mutants (Wang, 2000).

In second instar wild-type wing discs, wg is expressed distally in a wedge of anterior ventral cells and vn is expressed proximally. In vn null mutants, the initiation of wg expression is normal as is expression of its target gene optomotor-blind (omb). In wg mutants, however, there is a dramatic and early expansion of vn expression to include distal cells, presaging the development of these cells as an extra notum. Together these results suggest that Vn has an early role in establishing the notum and that Wg signaling is required to define a distal domain that is reduced in Egfr activity to allow wing development (Wang, 2000).

To test the role of Vn/Egfr signaling in specifying notum an examination was carried out to see whether the Iroquois complex (Iro-C) genes, ara and cap are targets of the pathway. The Iro-C genes have been implicated in specifying notum cell fate because loss of function causes a transformation of notum to hinge. Furthermore, misexpression of ara causes loss of the wing and a duplication of notum. Ectopic expression of an activated form of the receptor, Egfr^lambdatop4.2 greatly reduces the size of the wing and a small ectopic notum forms. vn is expressed in the presumptive notum in early second instar discs and Caup/Ara are expressed in the presumptive notum at the end of the second instar. In early third instar wing discs, Caup/Ara are expressed in a domain that overlaps with vn. In vn mutants, this expression of Caup/Ara is lost and loss of Egfr signaling, in Egfr^ts clones, in the medial notum results in a loss of Caup/Ara expression. However, clones in the lateral notum continued to express Caup/Ara, suggesting other factors regulate Iro-C gene expression in these cells at this stage (Wang, 2000).

Activation of Iro-C genes could account for the requirement for Egfr activity to specify the notum at the end of the second instar as this correlates with when these genes are first expressed. However, loss of Egfr signaling at a slightly earlier time (mid-first instar to mid-second instar, see below), prior to activation of the Iro-C genes, also results in loss of the notum. A possible explanation for this comes from the finding that vn expression is lost in vn mutants. This suggests Egfr activity must be sustained, via a positive feedback loop involving transcriptional activation of vn, during the second instar, to activate the Iro-C genes and hence specify notum at the end of this period. Interestingly, the vn gene is also a target of Egfr signaling in the embryo (Wang, 2000 and references therein).

It is suggested that the mechanisms by which wg and vn specify alternate cell fates in the early wing disc, wing, or notum are antagonistic. This is based on the observation that loss of Wg results in the spread of vn expression and the supposition that the resulting ectopic Egfr activity causes loss of the wing and a double notum phenotype. Further evidence that Vn/Egfr signaling represses wing development comes from the results of misexpressing a constitutive receptor, Egfr^lambdatop4.2, in the presumptive wing. In these flies, the wing is reduced to a stump covered with sensilla characteristic of the proximal wing (hinge) region and expression of the wing specific gene vestigial (vg) is repressed. Ectopic notal structures also form from the ventral pleura. The ability of ectopic Egfr signaling to suppress wing development is cell autonomous because clones of cells expressing Egfr^lambdatop4.2 lack vg expression. In adult wings these clones produced outgrowths lacking wing characteristics but are otherwise difficult to characterize (Wang, 2000).

Although vn expression expands in wg mutants, no reciprocal spread of wg expression was observed in vn mutants that would have been indicative of a double wing phenotype. However, when Vn/Egfr signaling is inhibited in the notum by expressing a ligand antagonist (Vn::Aos-EGF) under the control of ptc-Gal4, ectopic wings are induced in ~10% of the flies. This result demonstrates that presumptive notal tissue can be transformed to wing by reducing Egfr signaling. However, the transformation occurs only when Egfr signaling is reduced in a subset of cells, rather than all cells in the notum (as in a vn mutant). This may reflect the indirect requirement for Egfr activity to also promote wing development (Wang, 2000).

The loss of notum phenotype is characteristic of vn hypomorphs but in null vn alleles and some Egfr alleles both the wing and notum primordia fail to develop and the wing discs remain tiny. Thus, although ectopic activity of Egfr in the distal disc represses wing development, the pathway is nevertheless normally required for wing development. Using the temperature-sensitive Egfr^tsla allele it was found that this requirement is restricted to the period from mid-first to mid-second instar. Key genes involved in wing development that are active at this time include wg and apterous (ap). ap is expressed in dorsal cells and acts as a selector gene to divide the disc into dorsal and ventral compartments. Regulation of Notch ligands by Ap leads to Notch signaling at the DV boundary and the formation of an organizer for wing outgrowth and expression of the wing-specific transcription factor vg (Wang, 2000 and references therein).

Of these two candidates, wg and ap, it seemed unlikely that wg was the key gene affected by Egfr signaling from mid-first to mid-second instar because wg expression is normal in vn mutants at mid-second instar. However, later in the second instar, wg expression normally expands to fill the growing wing pouch and it was noted that in vn mutants, wg expression fails to undergo this expansion. A similar defect in wg expression is seen in ap mutants consistent with Ap function being impaired in vn mutants. Remarkably, ap expression is completely absent in second instar vn mutant discs. Thus, loss of Ap can explain why there is no wing in vn mutants. This is supported by the demonstration that ectopic ap is capable of rescuing wing development in vn mutants (Wang, 2000).

Several additional lines of evidence demonstrate that ap is a cell autonomous target of Vn/Egfr signaling and that this relationship exists only transiently in early wing development: (1) ap expression partially overlaps that of vn in the second instar; (2) ap can be induced ectopically in ventral clones misexpressing an activated form of the receptor, Egfr^lambdatop4.2; (3) Egfr^tsla mutant clones generated in the first instar show autonomous loss of ap expression, whereas clones generated in the second instar express ap normally. Finally, loss of Egfr activity in whole discs from mid-first to mid-second instar results in complete loss of ap expression, whereas ap is still expressed in discs from larvae given a temperature shift slightly later during the second instar (Wang, 2000).

The results described here suggest that division of the early wing disc into presumptive wing and body wall regions is defined by the action of two secreted signaling molecules, Wg and Vn. wg, a pro-wing gene, is required to repress vn expression, which at high levels antagonizes wing development. Antagonism between Wg and Egfr signaling has also been demonstrated in segmental patterning of the embryo and in development of the head and third instar wing pouch, suggesting such a relationship between these pathways may be a common theme in a number of cell fate choices. Finding that one of the main functions of Wg in early wing specification is to repress Vn/Egfr signaling in the distal region of the early disc raises the question as to whether this is the only role of Wg in wing specification and hence if wing-cell fate can be specified in the absence of both signals. This seems unlikely, because nubbin, an early wing cell marker, is not misexpressed proximally in a vn mutant, where cells would lack both signals (Wang, 2000).

Vn/Egfr signaling promotes development of the notum by maintaining its own activity through transcriptional activation of vn itself, and also promotes expression of ap. Thus, both vn and ap appear to be targets of Egfr signaling, but the domain of ap is clearly wider than that of vn, indicating that ap can be activated at a lower signaling threshold than vn. Vn is a secreted molecule and thus could generate a gradient of Egfr activity. This provides an explanation for how Egfr signaling can regulate both wing and notum development: vn autoregulation and notum development requires high Egfr signaling activity while ap expression and subsequent wing development requires lower signaling activity (Wang, 2000).

Interestingly, vertebrate Egfr and its ligands are expressed in the chick limb bud in a pattern that appears to overlap with the vertebrate ap homolog Lhx2, and these factors are required for limb outgrowth in the chick. In light of the present results it will be important to determine whether Egfr signaling controls Lhx2 expression and thus plays a role in regulating outgrowth of the vertebrate limb. These results may also have implications for the evolution of insect wings. If the control of body wall development by Egfr signaling is ancestral, and comparative analysis of other arthropods will be required to assert this, then one of the first steps towards evolution of wings could have occurred when Egfr signaling assumed control of ap (Wang, 2000).

Control of growth and patterning of the Drosophila wing imaginal disc by EGFR-mediated signaling

Growth and patterning of the Drosophila wing imaginal disc depends on its subdivision into dorsoventral (DV) compartments and limb (wing) and body wall (notum) primordia. Evidence is presented that both the DV and wing-notum subdivisions are specified by activation of the Drosophila Epidermal growth factor receptor (Egfr). Egfr signaling is necessary and sufficient to activate apterous (ap) expression, thereby segregating the wing disc into D (ap-ON) and V (ap-OFF) compartments. Similarly, Egfr signaling directs the expression of Iroquois Complex (Iro-C) genes in prospective notum cells, rendering them distinct from, and immiscible with, neighboring wing cells. However, Egfr signaling acts only early in development to heritably activate ap, whereas it is required persistently during subsequent development to maintain Iro-C gene expression. Hence, as the disc grows, the DV compartment boundary can shift ventrally, beyond the range of the instructive Egfr signal(s), in contrast to the notum-wing boundary, which continues to be defined by Egfr input (Zecca, 2002b).

The subdivision of the wing imaginal disc into AP and DV compartments, as well as prospective body wall (notum) and limb (wing) territories is marked by the expression of particular regulatory genes, such as the selector gene engrailed (en) in the P compartment, the selector gene apterous (ap) in the D compartment, and the genes of the Iroquois Complex (Iro-C) [mirror (mirr), auracan (ara) and caupolican (caup)] in the lateral notum. In mature third instar wing discs, the Iro-C genes are expressed not only within the prospective lateral notum, but in additional locations, including a thin stripe of cells that extends ventrally along the edge of the disc, as well as in specific subpopulations of cells in the prospective wing blade. This study addresses the role of Egfr signaling in controlling notum development and Iro-C gene expression therein, and then focuses on the role of Egfr signaling in inducing ap expression and establishing the DV compartments (Zecca, 2002b).

Egfr/Ras signaling is both necessary and sufficient to activate ap expression in early wing disc cells. Furthermore, evidence is provided that each wing disc cell chooses to express, or not to express, ap at this time, depending on its level of Egfr/Ras activation. However, in contrast to the Iro-C genes, the descendents of each cell then inherit this initial choice without further reference to Egfr/Ras signaling. The results of eliminating Egfr/Ras activity before the establishment of the DV compartments are particularly striking. Early loss of Egfr activity causes dorsally positioned cells within the disc to choose, incorrectly, to become V compartment founders. These cells and their descendents generally sort into the existing V compartment or out of the disc epithelium. In rare cases, they can form an ectopic V compartment within the D compartment. By contrast, later loss of Egfr activity has no effect on the DV compartmental segregation. These findings establish that Egfr signaling is responsible for establishing the D and V compartments through the heritable activation of ap (Zecca, 2002b).

Although the Iro-C and ap genes are activated in overlapping dorsoproximal sectors of the early wing disc, the domain of ap expression expands relative to that of Iro-C gene expression during subsequent development, causing the DV boundary to be positioned up to 30 cell diameters ventral to the notum-wing boundary. It is suggested that this shift occurs because ap-expressing cells no longer depend on Egfr/Ras input to continue to express ap. Hence, as ap-expressing cells within the notum primordium proliferate, some will move out of range of the instructive Egfr ligand, cease to express Iro-C genes and enter the wing primordium. In the accompanying paper (Zecca and Struhl, 2002b), evidence is provided that this shift must occur in order for D and V compartment cells to interact to induce Wg and stimulate wing growth and differentiation (Zecca, 2002b).

These results raise intriguing questions about the mechanism of ap activation. For example, Egfr signaling induces ap expression only during a discrete window of opportunity during the second larval instar, even though Egfr signaling both precedes the initial activation of ap and continues thereafter. What makes the ap gene responsive to Egfr signaling only during this early window of opportunity? In addition, the state of ap gene expression during this period, whether 'on' or 'off', is inherited for the remainder of development. How are both states of expression rendered heritable? It is possible that a temporal signal, such as a flux of a unique combination of hormones (for example, ecdysone and juvenile hormone) or the unique prior history of signaling events in the early wing disc, might prime the ap locus for activation by Egfr signaling during this period. The state of expression chosen during this period might then be maintained subsequently by mechanisms involving positive autoregulation (for the 'on' state) or heritable silencing mediated by the Polycomb Group proteins (for the 'off' state). However, there is little evidence at present to support these speculations and the actual mechanisms remain unknown (Zecca, 2002b).

The subdivision of the Drosophila wing imaginal disc into dorsoventral (DV) compartments and limb-body wall (wing-notum) primordia depends on Epidermal growth factor receptor (Egfr) signaling, which heritably activates apterous (ap) in D compartment cells and maintains Iroquois Complex (Iro-C) gene expression in prospective notum cells. The source, identity and mode of action of the Egfr ligand(s) that specify these subdivisions has been examined. Of the three known ligands for the Drosophila Egfr, only Vein (Vn), but not Spitz or Gurken, is required for wing disc development. Vn activity is required specifically in the dorsoproximal region of the wing disc for ap and Iro-C gene expression. However, ectopic expression of Vn in other locations does not reorganize ap or Iro-C gene expression. Hence, Vn appears to play a permissive rather than an instructive role in organizing the DV and wing-notum segregations, implying the existance of other localized factors that control where Vn-Egfr signaling is effective. After ap is heritably activated, the level of Egfr activity declines in D compartment cells as they proliferate and move ventrally, away from the source of the instructive ligand. Evidence is presented that this reduction is necessary for D and V compartment cells to interact along the compartment boundary to induce signals, like Wingless (Wg), which organize the subsequent growth and differentiation of the wing primordium (Zecca, 2002b).

All cells within the wing imaginal disc require a minimum level of Egfr/Ras activity to sustain a normal rate of proliferation. It is not known whether this activity reflects the basal activity of the Egfr/Ras transduction pathway, or the response of the receptor to a specific ligand. However, it is clear that this low level of Egfr/Ras activity does not require Vn dependent Egfr signaling, since it has been shown that ectopic expression of Ap in vn mutant discs can rescue growth and differentiation of the wing primordium. This result demonstrates that the absence of wing development in vn mutant discs is an indirect consequence of the failure to establish an ap^ON-ap^OFF interface (Zecca, 2002b).

During normal development, the ap and Iro-C genes are initially activated in overlapping dorsoproximal domains in response to Egfr signaling, and hence, at this early stage, it appears that most or all D compartment cells are exposed to relatively high levels of Egfr/Ras signaling. Thereafter, as the wing disc grows, ventrally situated D compartment cells inherit the 'on' state of ap expression, even as they populate areas of the disc progressively farther from the domain of high Egfr/Ras signaling and sustained Iro-C expression. It is suggested that the progressive reduction of Egfr/Ras activity in these ventrally situated D cells enables them to interact with neighboring V compartment cells to induce Wg and Vg expression and stimulate growth of the wing primordium. By contrast, early induced clones of Ras^V12-expressing cells autonomously express ap and experience persistent high levels of Ras activation, as indicated by sustained expression of the Iro-C genes. As a consequence, the ectopic DV boundary cannot shift outside of the domain of high Egfr/Ras signaling. Cells flanking this ectopic DV boundary fail to engage in the reciprocal induction of Wg and Vg expression or to stimulate growth. Hence, the ap^ON-ap^OFF interface may normally have to shift to a region of relatively low Egfr activity for the DV boundary to acquire wing organizer activity (Zecca, 2002b).

The ap^ON-ap^OFF interface may only be able to function as an organizer when cells on both sides are of prospective wing type. Prior to the initial activation of ap and the Iro-C genes, the nascent wing disc appears to be subdivided into mutually antagonistic domains of Egfr and Wg signaling that at least transiently define the incipient notum and wing primordia. Because ap and the Iro-C genes are initially activated in response to a common source of Egfr signaling, most or all D cells at this stage may be notum type. It is only later, when ventrally situated D cells move out of range of Vn-dependent Egfr signaling and switch to being wing type, that inductive interactions occur across the DV boundary to create a new and stable source of Wg signaling. It is suggested that cells on both sides of the DV boundary may have to be of wing type for the boundary to have organizer activity. One possible reason for why this might be the case is that vg, the selector-like gene that defines the wing state, is itself an integral component of the reciprocal signaling mechanism that allows D and V cells to induce the expression of DV boundary genes. High levels of Egfr/Ras signaling actively maintain Iro-C gene expression (and hence the notum state) and block vg expression. Hence, the DV boundary may normally have to shift ventrally, into a domain of low Egfr/Ras signaling and high Wg signaling that defines the incipient wing state, to allow the positive feedback loop of inductive signaling to initiate across the DV compartment boundary. Once this loop is established, it would provide a stable source of Wg and other signals generated along the DV boundary that govern the subsequent growth and differentiation of the wing blade (Zecca, 2002b).

Control of apterous by vestigial drives indirect flight muscle development in Drosophila

Drosophila thoracic muscles are comprised of both direct flight muscles (DFMs) and indirect flight muscles (IFMs). The IFMs can be further subdivided into dorsolongitudinal muscles (DLMs) and dorsoventral muscles (DVMs). The correct patterning of each category of muscles requires the coordination of specific executive regulatory programs. DFM development requires key regulatory genes such as cut (ct) and apterous (ap), whereas IFM development requires vestigial (vg). Using a new vg^null mutant, a total absence of vg is shown to lead to DLM degeneration through an apoptotic process and to a total absence of DVMs in the adult. vg and scalloped (sd), the only known Vg transcriptional coactivator, are coexpressed during IFM development. Moreover, an ectopic expression of ct and ap, two markers of DFM development, is observed in developing IFMs of vg^null pupae. In addition, in vg^null adult flies, degenerating DLMs express twist (twi) ectopically. Evidence is provided that ap ectopic expression can induce per se ectopic twi expression and muscle degeneration. All these data seem to indicate that, in the absence of vg, the IFM developmental program switches into the DFM developmental program. Moreover, the muscle phenotype of vg^null flies can be rescued by using the activity of ap promoter to drive Vg expression. Thus, vg appears to be a key regulatory gene of IFM development (Bernard, 2003).

Thus the absence of Vg leads to IFM degeneration. Some IFM phenotypes have been reported for the vg^83b27R allele, a strong allele of vg. In these flies, the DVMs are absent and some DLMs are missing. It has been shown that this phenotype is fully penetrant in vg^null flies and that apoptosis is involved in loss of IFMs. Since muscle attachment sites are normal in vg^null flies, the process of degeneration is different from that described in ap mutants. Phenotypic analysis shows that degeneration occurs during late metamorphosis (after 48 h APF) (Bernard, 2003).

In vg^null mutants all adepithelial cells express high levels of Ct, while this is normally only the case of DFM-forming myoblasts. Is DLM degeneration in vg^null mutants the result of engagement of DLMs toward a DFM-like differentiation process? To answer this question, ap expression was examined in vg^null developing and adult DLMs. In wild-type flight muscles, ap expression is specific to DFMs and begins at 17-19 h APF. In vg^null flies, ap expression is found in developing DLMs at 21 h APF, in myoblasts surrounding DLMs and in adult muscles. Moreover, an absence of actin 88F expression was found in vg^null developing IFMs, suggesting that IFM differentiation is disrupted. Interestingly, as in wild-type flies, no expression was found in adepithelial cells. These data show that ap starts to be expressed at the same stage in DLMs of vg^null flies and in DFMs of the wild type strain (Bernard, 2003).

In summary, the following has been demonstrated in vg^null flies: (1) DLM-forming myoblasts express high levels of Ct, an early marker for DFM-forming myoblasts and (2) myoblasts and developing and adult degenerating DLMs express ap, a specific late DFM marker, whereas actin 88F expression, an IFM-specific differentiation marker, is lost. According to these data, it is supposed that in the vg^null mutants, adepithelial cells and developing DLMs enter into a DFM-like development. The suggestion that ap ectopic expression may impose a DFM identity on the IFMs has already been proposed. However, an IFM-to-DFM transformation was not observed; rather, IFMs degenerated through an apoptotic process. Similarly, DFMs were not transformed into IFMs upon overexpression of Vg in DFM-forming myoblasts. Instead, DFM degeneration was obtained. This suggests that Vg and AP are major actors but are not sufficient for IFM and DFM development, respectively. Other signals and factors must be required to specify these muscles. Nerve-muscle interaction is associated with IFM development. Wnt oncogene analog 2 (Dwnt-2) expression is required in the vicinity of the developing DFMs for patterning of DFMs. Thus, it appears that adult muscle development requires complex interactions between several kinds of signals delivered in specific localizations. In vg^null homozygous flies, adepithelial cells and swarming myoblasts express DFM markers, but their position on the wing imaginal disc and in the pupa remains unchanged with respect to wild type. Thus, developing IFMs receive IFM signaling (at least nerve-muscle interactions), but myoblasts express apterous, a DFM maker. Moreover, they lack information necessary for formation of either DFMs or IFMs (absence of vg expression). It is suggested that IFM degeneration in vg^null homozygous flies is the result of this complex interaction between two contradictory signals (IFM and DFM) associated with incomplete signaling for formation of either type of muscle (Bernard, 2003).

Attempts were made to rescue the vg^null muscle phenotype by targeted Vg overexpression using the UAS-GAL4 system. Significant rescue was obtained with the ap-GAL4 driver. It is therefore likely that ectopic activation of the ap-GAL4 transgene in vg^null DLMs and myoblasts occurs when Vg is required for DLM formation. Since ap activation in vg^null myoblasts and developing DLMs occurs after puparium formation, it is concluded that a late Vg expression is sufficient to restore the DLM developmental process. This implies that adepithelial cell determination by the level of Ct at the wing disc is reversible. Thus, even though earlier Ct levels distinguish two adepithelial cell populations that will differentiate into DFMs or IFMs, definitive DFM versus IFM determination is a later event that takes place during metamorphosis. vg and ap could be key genes during specification of IFMs and DFMs, respectively. To support this hypothesis, ubiquitous overexpression of ap was shown to be sufficient to induce specific DLM degeneration. The way in which AP and Vg direct muscle development toward a DFM or IFM fate remains unclear. However, it is well known that muscle fibers express specific structural genes or isoforms. Since ap and vg encode transcription factors, they are probably involved in specific genes activation. For example, misexpression of ap in developing IFMs represses the expression of actin 88F, an IFM-specific actin gene. Moreover, no actin 88F expression is found in a vg^null context. However, it is not currently known whether AP or Vg can directly activate or repress structural genes. Interestingly, the Sd mammalian homolog (Transcription Enhancer Factor-1, TEF-1) has been shown to bind muscle-specific promoters, like the cardiac alpha-Myosin Heavy Chain and the cardiac Troponin T promoters. It is therefore possible that the Sd-Vg dimer plays a similar role in Drosophila, directly activating structural genes. Further studies are necessary to address this question (Bernard, 2003).

Thus DLMs degenerate by apoptosis in homozygous vg^null flies. This degeneration could be due to a misprogramming of myoblasts surrounding DLMs during development. The process that leads to apoptosis in these muscles remains to be determined. DLM degeneration is associated with an ectopic expression of Twi transcription factor. During flight muscle development, Twi expression is restricted to myoblasts and that persistent expression in developing muscles leads to muscle degeneration. Thus, Twi expression in vg^null mutants could be responsible for DLM degeneration. Finally, it has been shown that ectopic ap expression induces Twi expression in DLMs. Since AP and twi are known to be, respectively, activator and target of the N pathway, it can be hypothesized that AP activates Twi ectopically in vg^null DLMs through the N pathway. If this hypothesis is confirmed, it can be asked why AP does not activate Twi during normal DFM development. It is likely that numerous genes, other than vg and ap, are differentially activated during DFM and IFM development. Twi activation by AP could be repressed by one of these genes during DFM development (Bernard, 2003).

In this study, evidence is provided that vg is required to change DFM-forming myoblasts into IFM-forming myoblasts. As in wing development where Vg is considered as a selector gene, Vg could be a key gene in IFM specification. Its function would be equivalent to that of Ap for DFM development. DFM fate inhibition through repression of ct and ap by Vg seems therefore to be a key regulation feature of IFM development. Thus, correct programming and regulation of these three genes are necessary for correct patterning of Drosophila flight muscles (Bernard, 2003).

A concerted action of a paired-type homeobox gene, aristaless, and a homolog of Hox11/tlx homeobox gene, clawless, is essential for the distal tip development of the Drosophila leg

The distal region of the Drosophila leg, the tarsus, is divided into five segments (ta I-V) and terminates in the pretarsus, which is characterized by a pair of claws. Several homeobox genes are expressed in distinct regions of the tarsus, including aristaless (al) and lim1 in the pretarsus, Bar (B) in ta IV and V, and apterous (ap) in ta IV. This pattern is governed by regulatory interactions between these genes; for example, Al and Bar are mutually antagonistic, resulting in exclusion of Bar expression from the pretarsus. Although Al is necessary, it is not sufficient to repress Bar, indicating another factor is required. This factor has been identified as the product of the C15 gene, also termed clawless, a homeodomain protein that is a homolog of the human Hox11 oncogene. C15 is expressed in the same cells as al -- together, C15 and Al appear to directly repress Bar and possibly to activate Lim1. C15/Al also act indirectly to repress ap in ta V, i.e., in surrounding cells. To do this, C15/Al autonomously repress expression of the gene encoding the Notch ligand Delta (Dl) in the pretarsus, restricting Dl to ta V and creating a Dl+/Dl− border at the interface between ta V and the pretarsus. This results in upregulation of Notch signaling, which induces expression of the bowl gene, the product of which represses ap. Similar to aristaless, the maximal expression of C15 requires Lim1 and its co-factor, Chip. Bar attenuates aristaless and C15 expression through Lim1 repression. Aristaless and C15 proteins form a complex capable of binding to specific DNA targets, which cannot be well recognized solely by Aristaless or Clawless (Campbell, 2005; Kojima, 2005).

Bar expression is absent from the center of the leg, specifically from the cells expressing Al and C15. However, other genes, including ap and bab, are absent from a more extensive region in the center, and there is a gap between the C15 expression domain and Ap and Bab. Consequently, Ap expression is restricted to presumptive tarsal segment IV, where it overlaps with Bar. It has been suggested that, as well as activating genes such as al and Bar, EGFR signaling may directly repress genes in the center of the disc, possibly accounting for the absence of ap and bab in this location. Surprisingly, ap and bab expression, as well as Bar, is regulated by C15/Al. In both C15 and al mutant discs, Ap and Bab expression expands into the center of the disc. Consequently, in regard to Ap expression, the distal region of the leg adopts a tarsal segment IV-like fate. However, Nub, which is normally only expressed in ta V, is now co-expressed with Ap in the very center, indicating that the distal-most segment in C15 legs has characteristics of both ta IV and V (Campbell, 2005).

In wild-type discs, Ap expression is first detected slightly later than Bar, Al, or C15, but even at this time there is a clear gap between Ap expression and C15, indicating that C15/Al acts non-autonomously to repress ap. This is supported by two further studies: (1) unless there is a complete loss of C15 in homozygous mutant discs, Ap expression is not derepressed in C15 mutant clones in the center if the clones are not too large, indicating surrounding wild-type C15-expressing cells can rescue the mutant tissue; (2) ectopic expression of C15 results in non-autonomous repression of Ap (Campbell, 2005).

These results suggest that EGFR signaling represses gene expression in the center of the disc only indirectly through activation of C15/Al. This is also supported by two other observations. (1) Al is still expressed in C15 mutant discs, indicating that EGFR signaling levels are still very high in the center of these discs, but ap is not repressed (if ap is repressed directly by EGFR, its threshold for this would be lower than the threshold for activation of al because ap is repressed further from the source in the center than al is activated). (2) Ectopic expression of C15 results in non-autonomous repression of ap, but, if this is due to increased EGFR signaling in surrounding cells, then it should result in activation of EGFR targets such as Bar immediately adjacent to the cells expressing C15 (outside of the normal Bar domain), but does not. Consequently, it seems very likely that C15 uses an alternative mechanism to repress ap, most likely by upregulation of a signaling pathway in surrounding cells (i.e., ta V) (Campbell, 2005).

Examination of Bowl and Ap expression in leg discs reveals that there is a gap between their expression domains, even at a time when Ap expression is first detected in mid-third instars. This could indicate that Bowl acts non-autonomously to repress ap. However, the clonal analysis clearly shows that Bowl acts autonomously: any wild-type cells expressing Bowl has no influence on Ap expression in surrounding mutant tissue. It is possible that there is low-level Bowl expression in the 'gap' that cannot be detected with antibody staining. Another possible explanation is one of timing, and that Bowl is expressed in the cells in the 'gap' slightly earlier and that this is sufficient to silence the ap gene even before its expression can be detected more proximally. The possibility that bowl is expressed transiently in cells has been proposed to explain the observation that bowl mutant clones have effects in central regions of tarsus, i.e., in regions where its expression cannot be detected later (Campbell, 2005).

Thus, Bowl is required to repress ap expression in tarsal segment V and this predicts that C15 regulates bowl expression. This was confirmed by analysis of C15 mutant discs, in which Bowl expression in the center is lost, although other, more proximal, domains of expression are normal. The ring of Bowl in the distal tarsus is usually just two cells in width with the inner cell overlapping with C15, but the outer cell being outside the C15 domain, suggesting C15 can induce bowl non-autonomously. This is supported by the ability of cells ectopically expressing C15 to activate Bowl expression in surrounding cells. This ability is fairly limited, but would be expected because the endogenous C15-expressing cells only appear able to induce bowl in their immediate neighbor (resulting in a ring of bowl expression in a single row of cells surrounding the C15 domain (Campbell, 2005).

Regulation of expression of Vg and establishment of the dorsoventral compartment boundary in the wing imaginal disc by Suppressor of Hairless

The transcription factor Suppressor of Hairless [Su(H)] belongs to the CSL transcription factor family, the main transcriptional effector of the Notch-signaling pathway. Su(H) is the only family member in the Drosophila genome and should therefore be the main transcriptional effector of the Notch pathway in this species. Despite this fact, in many developmental situations, the phenotype caused by loss of function of Su(H) is too weak for a factor that is supposed to mediate most or all aspects of Notch signaling. One example is the Su(H) mutant phenotype during the development of the wing, that is weaker in comparison to other genes required for Notch signaling. Another example is the complete absence of a phenotype upon loss of Su(H) function during the formation of the dorsoventral (D/V) compartment boundary, although the Notch pathway is required for this process. Recent work has shown that Su(H)/CBF1 has a second function as a transcriptional repressor, in the absence of the activity of the Notch pathway. As a repressor, Su(H) acts in a complex together with Hairless (H), which acts as a bridge to recruit the co-repressors Groucho and CtBP, and acts in a Notch-independent manner to prevent the transcription of target genes. This raises the possibility that a de-repression of target genes can occur in the case of loss if function of Su(H). This study shows that the weak phenotype of Su(H) mutants during wing development and the absence of a phenotype during formation of the D/V compartment boundary are caused by the concomitant loss of the Notch-independent repressor function. This loss of the repressor function of Su(H) results in a de-repression of expression of target genes to a different degree in each process. Loss of Su(H) function during wing development results in a transient de-repression of expression of the selector gene vestigial (vg). This residual expression of vg is responsible for the weaker mutant phenotype of Su(H) in the wing. During the formation of the D/V compartment boundary, de-repression of target genes seems to be sufficiently strong, to compensate for the loss of Su(H) activity. Thus, de-repression of its target genes obscures the involvement of Su(H) in this process. Furthermore, evidence that is provided Dx does not signal in a Su(H)-independent manner as has been suggested previously (Koelzer, 2006).

This work provides an answer to the observation that the patterning defects of Su(H) mutant wing imaginal discs is weaker than anticipated for a gene that encodes a factor that mediates most of the transcriptional activity of the Notch-signaling pathway. Su(H) is required for the formation of the D/V compartment boundary despite any obvious defect in this process in the absence of its function. In both processes, the explanation for the phenotype of Su(H) mutants is the loss of its function as repressor of transcription along with its function as an activator (Koelzer, 2006).

Loss of function of Su(H) leads to an arrest in the development of the sensory organ precursor cell of the bristle sense organ. Although it was possible to demonstrate genetically that de-repression of expression of some genes of the Enhancer of split-complex are responsible for the arrest, it was not possible to detect the expression of any of these genes directly. This work shows that de-repression of vg is a consequence of loss of Su(H) function during wing development. Although this de-repression is weak and transient, it is sufficient to establish more distal elements than in mutants of other genes necessary for Notch signaling. The results are in agreement with two reports that show de-repression of target genes in Su(H) mutants in other developmental processes such as mesectoderm specification and bristle development. Thus, de-repression of target genes appears to be a common phenomenon during Drosophila development, if Su(H) function is lost. Importantly, this de-repression can even become strong enough to obscure an involvement of Su(H) in a developmental process, the formation of the D/V compartment boundary. De-repression of target genes upon loss of the repressor function of Su(H) is an attractive explanation for the paradox that loss of Notch function during the first larval instar stage is cell lethal, but loss of Su(H) function is not. Presumably, the de-repression of expression of target genes that are required for cell survival guarantees the survival of Su(H) mutant cells. In contrast, a similar de-repression cannot occur in Notch mutant cells, and the cells undergo apoptosis. Although the repressor function has been initially found in cell culture experiments with the vertebrate ortholog CBF1, reports analyzing the consequences of loss of its repressor function during vertebrate development are missing. The presented results should encourage researchers to search for such an effect in their vertebrate model systems (Koelzer, 2006).

The results have important implications on the use of various mutants in order to analyze the function of the Notch pathway in a particular developmental process. They show that the phenotype of loss of function of Su(H), or its vertebrate ortholog CBF1, is not necessarily identical to that of loss of the Notch-signaling activity. It is possible that de-repression of Notch target genes occurs upon loss of function of Su(H) but not upon inactivation of the pathway by other means. Previous work indicates that only a subset of genes might be de-repressed in a developmental process if Su(H) is absent. For example, de-repression of expression of wg along the D/V compartment boundary has never been observed upon loss of Su(H) function. The de-repression of only a subset of target genes could cause a phenotype that is difficult to interpret. Thus, it is better to use alleles of genes such as Psn, kuz or nic, which do not affect the repressor function of Su(H), to determine the function of the Notch pathway within a process of interest (Koelzer, 2006).

The weaker phenotype of Su(H) mutants during wing development was considered an argument for the existence of a Su(H)-independent mechanism of Notch signal transduction. The current findings strongly argue against the existence of such a mechanism in the analyzed processes. Evidence has been provided for the existence of a Su(H)-independent Notch-signaling pathway that is mediated by Dx. Since the existence of such a pathway has been excluded in the two other situations, it was of interest to discover whether an alternative explanation might exist for observations on the role of Dx. Indeed no evidence was found that Dx participates in a Su(H)-independent Notch signal during wing development. The results suggest that in this case, the confusion came from analyzing a domain of the vgBE (domain 2) that appears not to be completely dependent on the function of Su(H). Using the MARCM technique to generate Dx expressing Su(H) mutant cell clones, it was clearly show that Dx depends on the function of Su(H) to induce target gene expression in ectopic places as well as along the D/V boundary. Thus, the results abolish three arguments for the existence of a Su(H)-independent signal transduction mechanism during wing development. However, this does not imply that such a pathway does not exist. Indeed, evidence exists that during dorsal closure of the embryo, Notch acts independently of Su(H), through the JNK pathway (Koelzer, 2006).

Recent work indicates that cell-cell interactions are required for the establishment of both the A/P as well as the D/V compartment boundaries. While it is clear that a transcriptional response mediated by the transcription factor Cubitus interruptus (Ci) is necessary to establish the A/P boundary, the situation at the D/V boundary was unclear. The possibility has been raised of a Su(H)-independent mechanism that is used to establish the D/V boundary. This mechanism might not even require a transcriptional response to the Notch signal. The results demonstrate that this is not the case: similar to the formation of the A/P boundary compartment boundary, a transcriptional response to the Notch signal is required for the segregation of dorsal and ventral cells, and this response is mediated by Su(H). Similar to Ci, Su(H) acts as a transcriptional activator at the D/V boundary, where Notch is active and as a transcriptional repressor in a complex with H, and probably Groucho and dCtBP away from the boundary. The results suggest that the loss of this repressor function results in the de-repression of the relevant target genes in a manner sufficient to allow the formation of the D/V compartment boundary even in absence of Su(H). Overall the scenario at the D/V boundary seems to be very similar to that proposed for the formation of the A/P compartment boundary. In this situation, En endows the posterior fate and regulates the expression of Hedgehog that signals to anterior cells. As a response to Hh, the transcription factor Ci is transformed from a repressor to an activator of transcription and activates the expression of target genes in a stripe along the anterior side of the A/P boundary. The results suggest a similar scenario for the formation of the D/V compartment boundary: similar to En, Ap imposes the dorsal fates on cells and activates the expression of Ser. Ser signals to the ventral cells at the D/V boundary. Similar to Hh transforming Ci from a repressor into an activator of transcription, Ser induced activation of the Notch pathway transforms Su(H) from a repressor into an activator. In analogy to En, it was found that Ap has a second, Notch-independent function during D/V boundary formation. As in the case for En, an attractive possibility is that Ap acts to repress activation of the relevant target genes of Su(H) in dorsal cells. This repression creates a strong difference in expression of these genes at the D/V boundary and eventually leads to a strong difference in adhesion between the dorsal and ventral cells. This repressor function of Ap would also explain why the compartment boundary can form in the absence of Su(H) function, since the de-repression of target genes of Su(H) would be still restricted to ventral cells leading to a similar, albeit weaker difference in expression of these genes and in adhesion at the D/V boundary. Furthermore, it explains why the formation of the boundary fails in the absence of the function of ap and Su(H), since in this case no strong difference in expression of target genes will be created (Koelzer, 2006).

It appears that very similar strategies are exploited at both compartment boundaries to achieve segregation of the cell lineages. However, in each situation, a set of different but mechanistically similar acting signaling molecules are used to achieve the segregation of cell populations and formation of a compartment boundary (Koelzer, 2006).

Targets of Activity

alpha PS1 integrin is expressed at high levels in the dorsal domain of the third instar wing disc, while PS2 integrin is restricted to the ventral epithelium. Ectopic ventral expression of ap induces ectopic ventral alpha PS1 integrin (Blair, 1994).

apterous also regulates two other genes: fringe and Serrate. Each have distinct roles in a novel cell recognition and signal induction process. FNG serves as a boundary-determining molecule such that Ser is induced wherever cells expressing fng and cells not expressing fng are juxtaposed. SER in turn triggers the expression of genes involved in wing growth and patterning on both sides of the DV boundary. Fringe induces Serrate by a apterous independent mechanism. Serrate, through interaction with Notch induces vestigial and wingless. The expression of wingless is induced through Notch is independent of the earlier expression of wingless involved in inducing dorsal apterous expression (Williams, 1993 and Kim, 1995).

The product of the Drosophila gene Serrate acts as a short-range signal during wing development to induce the organizing center at the dorsal/ventral compartment boundary, from which growth and patterning of the wing is controlled. Regulatory elements reflecting the early Serrate expression in the dorsal compartment of the wing disc have recently been confined to a genomic fragment in the 5'-upstream region of the gene (from -8 to -18 kb). This fragment, termed the dorsal wing regulator or DWR, responds to various positive and negative inputs required for the early Serrate expression. Activation and maintenance of expression in the dorsal compartment of the wing discs of second and early third instar larvae depend on apterous, as revealed by reporter gene expression in discs either lacking or ectopically expressing apterous. The DWR is not activated by ectopic fringe expression in the ventral compartment, suggesting that the observed induction of Serrate protein by ectopic Fringe is mediated by a different enhancer, which is active at later stages during wing development. The lack of Suppressor of Hairless results in a precocious repression of reporter gene expression along the margin, suggesting that the DWR of Ser responds to the postulated feedback loop mediated by the Notch signaling cascade to maintain expression in cells adjacent to the dorsal wing margin (Bachmann, 1998b).

Transcriptional downregulation during third larval instar is mediated by hiiragi. hiiragi, which has not yet been cloned, develops a notched wing phenotype when homozygous and enhances the notched wing phenotype of Ser^D/+. Strikingly, in hir^P1 homozygous third instar larvae the expression domain of the DWR not only persists on the dorsal wing pouch, but expands into the ventral compartment from mid-third instar onwards. hiiragi is a good candidate to be involved in the downregulation of the DWR of Ser. The lack of nubbin (nub) leads to the loss of wing structures. In discs mutant for nub expression, the DWR along the D/V boundary is upregulated and persists longer than in wild-type discs. This is in agreement with the observation that Serrate protein expression appears to be more pronounced along the dorsal wing margin in nubbin mutant discs. This regulatory element also responds to Delta signaling in a nonautonomous way to maintain Serrate expression along the dorsal margin. The results clearly show that some of the previously described transactivators of Serrate protein expression, e.g. fringe, act on elements required for later aspects of Serrate expression (Bachmann, 1998b).

In the ventral ganglion, apterous is expressed in up to ten of the ~350 neurons in each hemi-segment. In each of six thoracic hemisegments and in the third subesophageal hemisegments, the ap neurons include a ventrolateral cluster of four or five cells. Double-labeling with antiserum to the FMRFalpha propeptide shows that one of the neurons in each cluster is the Tv neuron, a neuron committed to neuropeptide production. Double-labeled Tv neurons shows cytoplasmic FMRFalpha immunoreactivity and nuclear beta-galactosidase immunoreactivity, marking cells expressing Gal4 under the direction of an apterous promoter. The ap gene is also expressed in several brain cells, one of which is the FMRFalpha-positive SP2 neuron. Other larval brain FMRFalpha neurons, such as the neighboring SP1 neuron, do not express ap. The restriction of ap and FMRFalpha co-expression to the Tv and SP2 neurons is constant throughout mature larval stages; in the adult, additional neurons begin to express dFMRFalpha, and some of these, including the Tva and several subesophageal neurons, also express ap (Benveniste, 1998).

Tv neurons first stain with antibodies to the tetrapeptide FMRFalpha during stage 17. FMRFalpha gene expression was measured using reporter expression driven be a 446 bp Tv neuron-specific enhancer sequence located within the first kB of FMRFalpha 5' flanking region. Ap is required for normal initiation of neuropeptide expression by the Tv neurons. Apterous is shown not to be required for the survival or morphological differentiation of the Tv neuron cluster. Apterous contributes to the initiation of FMRFalpha expression in Tv neurons, but not in those FMRFalpha neurons that do not express Apterous. Apterous is not required for Tv neuron survival or morphological differentiation. Apterous contributes to the maintenance of FMRFalpha expression by postembryonic Tv neurons, although the strength of its regulation is diminished. Apterous regulation of FMRFalpha expression includes direct mechanisms, although ectopic Apterous does not induce ectopic FMRFalpha. These findings show that, for a subset of neurons that share a common neurotransmitter phenotype, the Apterous LIM homeoprotein helps define neurotransmitter expression with very limited effects on other aspects of differentiation (Benveniste, 1998).

The hypothesis that Ap regulates FMRFalpha in the Tv neurons directly was tested by seeking potential Ap-binding sites within the dFMRFa gene regulatory sequences. The search was confined to the 446 bp Tv neuron-specific enhancer, which is highly responsive to Ap levels and located in the 5' flanking region. The 446 bp enhancer contains three sequences corresponding to the six-nucleotide consensus binding site for homeodomain proteins. All three of these sequences are shared between the homologous regions of the FMRFalpha genes of two Drosophila species: D. melanogaster and D. virilis. Electrophoretic mobility shift assays (EMSA) were used to test the ability of Ap protein to bind in vitro to these three sequences, as represented by three different 25 bp oligonucleotide probes. Recombinant Ap homeodomain binds all three oligonucleotide probes with different affinities, and at stoichiometries comparable to those observed for other LIM homeoproteins binding in vitro. Ap binding to these probes in vitro is sequence-specific: mutant oligonucleotide probes, with clusters of 6-point mutations replacing the TAATNN sequences do not bind Ap in these assays. It was then asked whether these Ap-binding sequences are functionally important in vivo. Two mutant Tv-lacZ constructs were used incorporating the same clustered point mutations in the Ap-binding sequences used in the EMSA. In first instar larvae, a construct containing mutations in Ap-binding site C [(mC)Tv-lacZ ] shows slightly decreased activity in Tv neurons and in ectopic cells relative to the wild-type enhancer. Construct (mABC)Tv-lacZ, which includes mutations in all three Ap-binding sequences, shows no detectable activity in Tv neurons or ectopic cells. These results show that at least two of the three elements within the Tv neuron-specific enhancer that bind Ap in vitro are critical for proper enhancer activity in vivo (Benveniste, 1998).

It is found that ap is expressed in more than 100 neurons in the larval CNS, but that FMRFalpha is expressed in only eight of these. Therefore, co-factors must be required to activate FMRFalpha transcription in the Tv neurons or to repress FMRFalpha transcription in other neurons that express Ap. Two lines of evidence suggest that positively acting co-factors are required for FMRFalpha gene activation by Ap. (1) Widespread ectopic expression of Ap (ubiquitously or throughout the CNS) does not induce ectopic FMRFalpha expression. (2) Ap expression in embryonic Tv neurons begins soon after the birth of the cell and precedes dFMRFa expression by at least 3-6 hours (Benveniste, 1998 and references).

Dorsoventral axis formation in the Drosophila wing depends on the activity of the selector gene apterous. Although selector genes are usually thought of as binary developmental switches, Apterous activity has been found to be negatively regulated during wing development by its target gene dLMO. Apterous-dependent expression of Serrate and fringe in dorsal cells leads to the restricted activation of Notch along the dorsoventral compartment boundary. Evidence is presented that the ability of cells to participate in this Apterous-dependent cell-interaction is under spatial and temporal control. Apterous-dependent expression of dLMO causes downregulation of Serrate and fringe and allows expression of Delta in dorsal cells. This limits the time window during which dorsoventral cell interactions can lead to localized activation of Notch and induction of the dorsoventral organizer. Overactivation of Apterous in the absence of dLMO leads to overexpression of Serrate, reduced expression of Delta and concomitant defects in differentiation and cell survival in the wing primordium. Thus, downregulation of Apterous activity is needed to allow normal wing development (Milan, 2000).

Removing Apterous activity at different stages of wing development shows that Ap is needed throughout larval stages to confer dorsal cell identity, but its role in Notch activation along the DV boundary is temporally and spatially modulated. This can be explained in terms of changes in Serrate and fringe expression. Some of the changes in Serrate and fringe expression are caused by reducing Ap activity, whereas others are Ap independent. In early second instar wing discs, Ap activity is required in the entire dorsal compartment. Removing Ap activity in mitotic recombination clones at this stage induces Notch activation at the interface between wild-type and mutant cells. This response is independent of the position of the clone within the wing pouch. In early third instar wing discs, Ap-dependent expression of Serrate and fringe is reduced by dLMO. Serrate expression gradually becomes restricted to the region near the DV boundary and, subsequently, by mid-third instar is induced only in cells adjacent to the boundary. The effects of removing Ap activity in clones reflects the gradual retraction of Serrate expression toward the DV boundary. Clones of cells lacking Ap activity induced in early third instar activate the Notch pathway and induce Wg if they are located close to the DV boundary. Clones located more proximally do not show this response. This spatial difference can be overcome by providing Serrate in proximal cells (Milan, 2000).

By mid-third instar, new Ap-independent patterns of Serrate and fringe expression are observed. Serrate is expressed on both sides of the DV boundary by the activity of Wg, and fringe is expressed in four quadrants flanking the DV and AP compartment boundaries. Maintenance of Notch activation along the DV boundary is now under control of a feedback loop between Wg and Serrate and Delta. Ap is no longer required for Notch activation at the DV boundary and removing Ap activity no longer leads to activation of the Notch pathway. In the absence of dLMO, Ap activity remains at high early levels as development proceeds. Serrate and fringe expression remain high throughout the dorsal compartment and fail to undergo normal modulation. In addition, Delta is not expressed in dorsal cells. Ap-dependent repression of Delta at early stages is needed to prevent ectopic activation of Notch in dorsal cells, which are inherently Delta-sensitive due to the activity of Fringe. Some of the defects observed in dLMO mutant wings are correlated with excess Serrate activity and insufficient Delta activity. In addition, abnormally high levels of cell death in the dorsal compartment of the dLMO mutant wing disc are due to excess Ap activity and this leads to overall reduction in the size of the wing. These findings indicate the need to downregulate Ap activity to allow normal wing development. However, Ap activity continues to be required for dorsal cell fate specification and for proper adhesion of D and V wing surfaces. Thus it is proposed that different target genes may be controlled at different levels of Ap activity. Serrate, fringe and Delta may be regulated by a higher level of Ap activity than the target genes involved in surface apposition or fate specification. Temporal changes in the level of Ap activity may be required to modulate activity of different genes at different times to allow normal wing development (Milan, 2000).

Capricious and Tartan, two transmembrane proteins with leucine-rich repeats, contribute to formation of the affinity boundary between dorsal and ventral compartments during Drosophila wing development. Engrailed/Invected expression confers posterior (P) identity and Apterous confers dorsal (D) identity in the wing disc. P compartment cells lacking engrailed/invected activity do not respect the anterior-posterior boundary. Likewise, dorsal cells lacking ap activity fail to respect the dorsal-ventral (DV) boundary in the wing disc. Modulation of Notch signaling has been implicated in DV boundary formation. Fringe acts as a glycosyltransferase to modify the receptor protein Notch in the dorsal compartment. Fringe activity makes D cells more sensitive to Delta, a ligand expressed by V cells and less sensitive to Serrate, the ligand expressed by D cells. Consequently, signaling by each ligand is limited to nearby cells on the opposite side of the boundary, with the result that high levels of Notch activity are limited to a narrow band of cells along the DV boundary. Although altering the signaling properties of cells by modulation of Fringe activity has been shown to allow cells to cross the boundary, Fringe activity has been shown to be insufficient to support boundary formation. This observation, together with the fact that Notch signaling is activated symmetrically has suggested that other Apterous-dependent cell interactions might be needed for formation of the DV affinity boundary. Evidence suggests that capricious and tartan are targets of Apterous that contribute to DV boundary formation in the wing disc. caps and tartan are expressed in the D compartment during boundary formation. Caps and Tartan confer affinity for D cells, assessed by sorting-out behavior. Caps supports boundary formation without conferring D signaling properties. Fringe, in contrast, confers dorsal signaling properties without affecting DV affinity. Thus, Caps, Tartan, and Fringe have complementary roles in boundary formation (Milán, 2001a).

Drosophila limbs develop from imaginal discs that are subdivided into compartments. Dorsal-ventral subdivision of the wing imaginal disc depends on apterous activity in dorsal cells. Apterous protein is expressed in dorsal cells and is responsible for (1) induction of a signaling center along the dorsal-ventral compartment boundary; (2) establishment of a lineage restriction boundary between compartments, and (3) specification of dorsal cell fate. The homeobox gene msh (muscle segment homeobox) acts downstream of apterous to confer dorsal identity in wing development (Milán, 2001b).

Four structural features distinguish the dorsal and ventral surfaces of the adult wing: (1) bristle morphology in the anterior wing margin; (2) the presence or absence of bristles in the alula; (3) the surface on which the veins are corrugated, and (4) the location of certain sensory organs. The anterior wing margin (AWM) is composed of three rows of bristles, two located in the dorsal surface and one in the ventral. The dorsal wing margin differentiates a row of thick, densely aligned, mechanosensory bristles and a second row of thinner, curved, chemosensory bristles. The dorsal AWM produces one chemosensory bristle per five mechanosensory bristles. The ventral row is composed of thin bristles interspersed with chemosensory bristles in every fifth position. The alula is located in the posterior compartment. It produces a single row of long thin bristles along the margin on the ventral surface. The dorsal surface of the alula lacks bristles. The adult wing differentiates five longitudinal veins. L1 is located on both dorsal and ventral sides of the wing margin and L2-L5 veins are located in the wing blade. Veins L2-L5 are asymmetrical on the dorsal and ventral surfaces of the wing. One side contains more rows of tightly packed cells ('corrugated vein'). The opposite side is thinner ('ghost vein'). Corrugated veins consist of three rows of strongly pigmented and densely packed cells. Ghost veins consist of a single row of cells. Longitudinal veins L3, L5 and the distal tip of L4 are dorsally corrugated. Veins L2 and proximal L4 are ventrally corrugated (Milán, 2001b).

The msh gene belongs to the msh/Msx family of homeobox genes involved in dorsal cell fate specification in the Drosophila neuroectoderm. Since msh is expressed in the dorsal compartment of the wing disc, an investigation was carried out to see whether msh is also involved in dorsal identity specification in the Drosophila wing. For this purpose, msh mutant clones were generated in the wing and the DV identity of the bristles located along the AWM, in the alula and the DV corrugation of longitudinal veins in mutant cells, was assessed. Clones mutant for msh have no aberrant phenotype in the ventral surface of the wing. When mutant for msh, the dorsal anterior wing margin differentiates ventral bristles. A single row of thin bristles interspersed with chemosensory bristles in every fifth position is observed. Thus, the anterior wing margin differentiates a ventral pattern of bristles symmetrically on both surfaces (Milán, 2001b).

When covered with mutant cells, the dorsal surface of the alula differentiates bristles. This reflects transformation to a ventralized cell fate. Absence of msh activity also induces a change in the pattern of corrugation of the longitudinal veins. In wild-type wings, veins L2 and L4 differentiate as 'ghost veins' on the dorsal surface. When mutant for msh, these veins are corrugated and differentiate three rows of strongly pigmented cells, thus mimicking a ventral-like pattern. Veins L3 and L5 are normally corrugated on the dorsal surface. When mutant for msh, they lose pigmentation and consist of a single row of aligned cells. Thus veins differentiate ventral characteristics in the dorsal surface when mutant for msh. It is concluded that msh is required in the dorsal compartment of the Drosophila wing to confer dorsal cell identity. In the absence of msh, symmetric wings are observed that differentiate ventral characteristics on both surfaces (Milán, 2001b).

Apterous is expressed in dorsal cells and is required to confer dorsal cell identity. It was therefore necessary to determine whether msh expression in the dorsal compartment is regulated by Apterous activity. MSH mRNA and msh-lacZ reporter genes are expressed in the dorsal compartment of the wing disc. MSH mRNA is expressed at a low level throughout the dorsal compartment, except in the region of the anterior margin where it is expressed at higher level. Ectopic expression of Apterous in the ventral compartment under control of dppGal4 induces ectopic expression of MSH mRNA at a level comparable to the overall low dorsal level. In apterous mutant discs, msh expression is lost from dorsal cells of the reduced wing pouch, but expression in the anterior mesopleura and hinge region remains. Finally, overexpression of dLMO, a repressor of Apterous activity in the Drosophila wing, represses expression of the msh-lacZ reporter gene. These results indicate that msh is indeed a target of Apterous. Additional studies show that ectopic expression of msh in the ventral surface is sufficient to confer dorsal identity on ventrally located cells (Milán, 2001b).

The results presented thus far indicate that mshis necessary and sufficient to specify dorsal identity in the Drosophila wing. A dominant mutation Dlw¹ has been identified that shows partial dorsalization of the AWM. Both surfaces of Dlw¹/+ AWMs have dorsal bristles, similar to what is observed when msh is ectopically expressed in the ventral compartment. Interestingly, Dlw alleles are associated with breakpoints located 30-90 kb upstream of the msh gene, raising the possibility that Dlw alleles may be regulatory mutants of msh. Indeed, a lethal allele of msh, mshDelta68, has proved to be lethal when heterozygous with Dlw¹ and the recessive lethal alleles Dlw³ and lw⁴. Dorsal clones mutant for Dlw³differentiate ventral structures. Genetic evidence is provided that supports the proposal that the msh gene is expressed in an Apterous-independent manner in Dlw¹ wings (Milán, 2001b).

msh mRNA levels are reduced throughout the wing pouch in discs heterozygous for Dlw¹. The low level of msh expression in the Dlw¹ background may explain the loss of function characteristics exhibited by the Dlw¹ allele in homozygous mutant clones. Dlw¹/Dlw¹ mutant clones located in the dorsal surface of the wing differentiate ventral structures. Thus, Dlw¹ causes a dominant transformation of ventral cells to dorsal identity when heterozygous and an opposite transformation of dorsal cell to ventral identity when homozygous mutant in clones. Interestingly, the dominant mutation Drop, which affects eye development, has been recently shown to be a gain-of-function allele of msh (Mozer, 2001). Drop mutants contain lesions in the same region as Dlw mutants (i.e. upstream of the msh transcription start site) and ectopic expression of msh in the eye phenocopies the Drop phenotype. However, Mozer was not able to find detectable misexpression of msh in Drop mutants. Thus, undetectably low levels of msh misexpression in eye and wing seem to be associated with the dominant adult phenotypes associated with the Dlw and Drop alleles of msh (Milán, 2001).

Apterous activity is required to confer dorsal identity and dorsal-type signaling properties. Fringe and Serrate expression in dorsal cells induce a cascade of short-range interactions between dorsal and ventral compartments that lead to the expression of the organizing molecule Wingless along the DV compartment boundary. The results reported in this study suggest that msh confers dorsal identity without affecting DV signaling. In order to verify that this is the case, the ability of msh to restore dorsal identity and dorsal signaling properties in the absence of Apterous activity was examined. In ap^Gal4/ap^UGO35 flies, the wing margin is reduced and the wing is considerably smaller than normal owing to reduced Apterous activity. When present, margin bristles have ventral identity in this genotype. Expression of msh in ap^Gal4/ap^UGO35;uas-msh flies does not restore outgrowth of the wing. The few margin bristles observed in the dorsal surface of these wings have dorsal identity. Growth and wing margin formation can be restored in the ap^Gal4/ap^UGO35 mutant background by expression of Fringe under ap^Gal4 control. In these wings, both surfaces differentiate ventral structures: the AWM and the alula differentiate ventral bristles on both surfaces and the pattern of vein corrugation is ventral. Co-expression of msh with EP-fringe confers dorsal differentiation in the bristles of the dorsal AWM in these rescued wings. It was also noted that overexpression of msh in dorsal cells reduces the size of the dorsal wing pouch, induces differentiation of ectopic bristles in the wing blade and affects vein differentiation. This was also observed in ap^Gal4/+; uas-msh/+ flies and presumably reflects defects caused by higher than normal Msh levels in dorsal cells. Note that the endogenous levels of msh expression in the wing pouch are quite low. These results suggest that developmental regulation of Msh protein levels may be crucial for proper wing development and differentiation of patterning elements. All these results indicate that msh confers dorsal identity without affecting dorsal signaling properties (Milán, 2001).

Two apterous homologs, Lmx1 and Lhx2, have been implicated in vertebrate limb development. Interestingly, these two genes appear to have separable functions in conferring dorsal identity and limb outgrowth. Lmx1 is expressed in the dorsal compartment of vertebrate limbs and is necessary and sufficient to confer dorsal identity. Lhx2 induces Radical-fringe expression in the apical ectodermal ridge, which is required for limb outgrowth. This contrasts with the situation in Drosophila where Apterous is responsible for both dorsal fate specification and for establishing the Fringe-dependent signaling center at the DV boundary. The findings reported here implicate msh as the principle target gene through which Apterous confers dorsal cell fate. msh is necessary and sufficient to induce dorsal cell fate, but has no role in DV boundary signaling. Intriguingly, the msh/Msx family of homeobox genes is also differentially expressed along the DV axis of the embryo and msh is required in the Drosophila neurectoderm to specify dorsal fate (Milán, 2001).

Apterous regulates neuropeptide cell identity by the integration of retrograde BMP signaling and a combinatorial transcription factor code

Individual neurons express only one or a few of the many identified neurotransmitters and neuropeptides, but the molecular mechanisms controlling their selection are poorly understood. In the Drosophila ventral nerve cord (VNC), the six Tv neurons express the neuropeptide gene FMRFamide (FMRFa). Each Tv neuron resides within a neuronal cell group specified by the LIM-homeodomain (LIM-HD) gene apterous (ap). The zinc-finger gene squeeze acts in Tv cells to promote their unique axon pathfinding to a peripheral target. There, the BMP ligand Glass bottom boat activates the Wishful thinking receptor, initiating a retrograde BMP signal in the Tv neuron. This signal acts together with apterous and squeeze to activate FMRFamide expression. Reconstituting this 'code,' by combined BMP activation and apterous/squeeze misexpression, triggers ectopic FMRFamide expression in peptidergic neurons. Thus, an intrinsic transcription factor code integrates with an extrinsic retrograde signal to select a specific neuropeptide identity within peptidergic cells (Allan, 2003).

FMRFa is specifically expressed in the six Tv neuroendocrine neurons located bilaterally in the three thoracic (T1-3) segments of the embryonic and larval VNC. apterous is expressed in three interneurons per VNC hemisegment, as well as in a lateral cluster of four neurons (the ap-cluster) in each of the T1-3 hemisegments. One of the four ap-cluster cells is the FMRFa-expressing Tv neuron. All ap interneurons in the VNC, except for the Tv, join a common ipsilateral axon tract termed the ap-fascicle. The Tv axon instead projects to the midline and exits the VNC dorsally to innervate the dorsal neurohemal organ (DNH). The DNH is a club-like neuroendocrine structure formed by two glial cells protruding from the midline of each thoracic segment. Anteriorly, two additional FMRFa-expressing cells are found, denoted SE2 cells. The SE2 cells do not express, nor depend upon, any regulators described in this study for their FMRFa expression. ap is important for the expression of FMRFa in the Tv neurons, but since most ap neurons do not express FMRFa, other regulators are likely needed for FMRFa regulation (Allan, 2003).

Rotund, a zinc finger protein of the C₂H₂ Krüppel-type belongs to a conserved subfamily of zinc finger proteins together with Drosophila CG5557, C. elegans Lin-29, and rat CIZ. Squeeze is most closely related to Rotund, with identity greater than 90% throughout the zinc finger region; Squeeze is 78% identical to LIN-29 in the conserved zinc finger region. Both rotund and CG5557 are expressed in subsets of cells in the developing CNS. CG5557 has a larval lethal phase. Mutants eclosed at a low frequency as immotile adults that died within 24 hr. Mutant larvae display a motility defect whereby the body wall musculature over-contract radially during the peristaltic wave typical of insect larval motility, apparent as a 'squeezing' of the intestine. Since this motility phenotype is fully penetrant and scored with 100% accuracy (sqz^lacZ/sqz^Df), CG5557 was renamed squeeze (sqz) (Allan, 2003).

The expression of sqz is largely restricted to subsets of cells in the CNS throughout embryonic and first instar larval (L1) development. Using sqz^GAL4 to drive expression of the axonal reporter, UAS-τ-myc, sqz was found to be expressed in a population of lateral interneurons, primarily projecting axons in the anterior and posterior commissures. In sqz mutants, expressing neurons are generated and appear to extend axons along the appropriate tracts. Using both sqz^lacZ and sqz^GAL4, tests were performed for overlap with ap; sqz and ap were found to be co-expressed specifically within the thoracic ap cluster. Co-expression of sqz and ap is evident from the onset of ap expression at stage 14, with one neuron typically expressing higher levels of sqz. By stage 17, sqz expression is restricted to two neurons within the ap-cluster, with one neuron typically continuing to display higher levels of expression. Expression overlap between sqz and FMRFa was tested in late stage 17 embryos, when FMRFa expression commences; sqz is indeed selectively expressed at higher levels within the FMRFa Tv neuron. Thus, the six neurons within the VNC that co-express ap and higher levels of sqz selectively express the neuropeptide FMRFa and innervate the three specialized neuroendocrine glands -- the dorsal neurohemal organs (Allan, 2003).

To determine whether sqz regulates FMRFa expression, immunoreactivity for the FMRFa peptide was compared in wild-type and sqz mutant L1 larvae. In wild-type, FMRFa immunoreactivity is robust (98%) in all six Tv neurons. In sqz mutants (sqz^lacZ/sqz^Df), FMRFa staining was found to be reduced in all Tv neurons and was detected in 75% of cells. The T1 segment was most affected, with FMRFa expressed in 40% of T1 Tv neurons. To verify that the observed effects reflected regulation of the FMRFa gene, antibodies recognizing the C-terminal of the FMRFa precursor peptide (proFMRF) were used, as well as an FMRFa-lacZ reporter that faithfully reports FMRFa expression in Tv neurons. An equivalent effect on proFMRF (75%) and FMRFa-lacZ (77%) was found in sqz mutants (sqz^lacZ/sqz^Df and sqz^GAL4/sqz^Df, respectively) when compared to wild-type. Again, segment T1 is most affected with FMRFa-lacZ expressed in only 50% of T1 Tv neurons. These results show that sqz in part regulates the expression of the FMRFa gene in Tv cells (Allan, 2003).

To determine whether sqz regulates axon pathfinding of the Tv neuron, ap^GAL4 was used to drive the expression of a membrane-targeted reporter (UAS-EGFP^F). In sqz mutants, a frequent failure of the Tv axon to innervate the DNH was observed, instead, it apparently joins the ap-fascicle. This phenotype is most pronounced within the most anterior thoracic segment (T1). In wild-type embryos, the DNH was innervated in 100% of thoracic segments, whereas sqz mutants (ap^GAL4/+;sqz^ie/sqz^Df,UAS-EGFP^F) show axonal innervation in 69% of T1 segment DNHs. Failure of innervation did not result from the absence of the DNH itself, since its profile was evident in affected segments. These results show that sqz is important for proper pathfinding of Tv axons and that the Tv axon often fails to diverge from the ap-fascicle in sqz mutants, apparently reverting to an 'ap-only' phenotype (Allan, 2003).

Several determinants critical for proper FMRFa expression have been identified. These include a general peptidergic cell identity, co-expression of sqz and ap, axon projection out of the VNC, and competence to respond to a retrograde signal by activating the BMP pathway. When these criteria are met, either in the endogenous or ectopic case, FMRFa expression is triggered. Importantly, none of these events are individually exclusive to the Tv cell, but they are uniquely combined in only these 6 out of the 10,000 cells in the VNC. Reconstituting this scenario in other peptidergic neurons can trigger FMRFa expression. These results are in line with the emerging theme of a critical interplay between combinatorial transcription factor codes and signal transduction pathways in regulating gene expression and provide a clear example of how these general mechanisms also apply to the specific regulation of a terminal differentiation gene in the nervous system (Allan, 2003).

Why is ectopic FMRFa expression restricted to peptidergic neurons? Conceivably, cells responding to BMP activation and sqz/ap co-misexpression may arise from precursor cells utilizing a common genetic program, resulting in a chromatin state where the FMRFa gene is accessible to activation. Currently, the lineage from which most neuropeptidergic neurons arise is unknown, and any common theme behind their generation is uncertain. FMRFa expression may also be constrained by the presence of activators common to peptidergic neurons and/or by repressors present in non-peptidergic neurons. Common properties of peptidergic neurons, such as the dense core vesicle secretory machinery and the processing of precursor peptides, may indicate the existence of common regulatory programs for all peptidergic neurons. In support of this notion, recent studies of a novel basic helix-loop-helix transcription factor, dimmed, show that this gene is specifically expressed in most if not all peptidergic neurons. In dimmed mutants, peptidergic and secretory properties of the majority of peptidergic neurons are affected, including the expression of processing enzymes and several neuropeptides, such as FMRFa. This shows that dimmed plays a key role in specifying the peptidergic fate and supports the notion of a common regulatory program for this cell type (Allan, 2003).

Previous studies found that ap is essential for axon pathfinding of the majority of ap-neurons. However, ap does not affect Tv axon pathfinding, suggesting that the role of ap in Tv cells may exclusively be to regulate FMRFa expression. In line with these results, ap mutants do not show any apparent loss of pMad accumulation in the Tv neurons. In contrast, sqz mutants have Tv axon pathfinding phenotypes, and, consequently, a partial loss of pMad staining specifically in Tv neurons. Observations of Tv axons at the midline suggest that in the absence of sqz, the Tv axon likely reverts to an 'ap-only' axonal phenotype and turns to grow along the common ap-fascicle. Given the importance of DNH innervation for FMRFa expression, axon pathfinding defects in sqz mutants likely contribute to the loss of FMRFa in some hemisegments. However, the great difference in the loss of FMRFa expression between sqz (75%) and wit (0%) argues that sqz is not critical for BMP signaling, but rather affects it indirectly by affecting Tv axon pathfinding. Moreover, the sqz axon pathfinding phenotype is only partially penetrant and fails to explain either the reduction of FMRFa expression observed in all hemisegments, or the potency of sqz (acting together with ap) to trigger ectopic expression in Va and Vap peptidergic neurons (cells whose axons already exit the VNC and are pMad-positive). Misexpression of sqz in all ap cells occasionally leads to an additional pMad/FMRFa positive cell in the ap-cluster. In these cases, no ectopic FMRFa expression is detected in any axons extending in the common ap-fascicle, only in axons projecting into the DNH. Therefore, sqz misexpression likely alters the identity of another ap-cluster cell, imposing a Tv-like axonal pathfinding behavior and causing it to ectopically innervate the DNH. Thus, it appears that sqz regulates two critical features of Tv cell identity: differential pathfinding, and FMRFa expression (both directly and indirectly) (Allan, 2003).

Why do sqz and ap function to activate FMRFa expression within only three neuropeptidergic cell types (the Tv, Va, and Vap cells) which together comprise only 18 out of ~200 peptidergic neurons in the developing Drosophila VNC? Using the specific GAL4 lines, ap^GAL4, Va^GAL4, and Vap^GAL4 to drive the expression of UAS-EGFP^F, it was found that all three neuronal subsets exit the VNC: Tv axons via the DNHs, Va axons via the transverse nerves, and Vap axons via the posterior A8 nerves. This observation is important in light of previous studies of tinman (tin) mutants. In tin mutants, a number of mesodermally derived tissues, including the DNHs, fail to develop. As a result, Tv axons stall at the presumptive midline exit point and, intriguingly, FMRFa expression is strongly reduced. This suggests that the DNHs may be necessary for proper FMRFa expression in Tv cells. These findings have been confirmed; in tin mutants, the DNHs are absent, and proFMRF staining is weak and only detected in 10% of Tv neurons. To address the putative target requirement for FMRFa expression in an alternative way, ap^GAL4 was used to express molecules that either alter Tv axon pathfinding or interfere with Tv axonal transport. roundabout (robo), a receptor that mediates repulsion from the VNC midline, was tested. In ap^GAL4/UAS-robo L1 larvae, Tv axons avoid the midline and fail to innervate the DNH. As predicted, this results in a loss (2%) of FMRFa-lacZ expression. Next, dominant-activated rac (UAS-rac^V12) was tested; it causes Tv axons to stall before reaching the midline and they fail to innervate the DNHs. This results in a complete loss (0%) of FMRFa-lacZ expression. To interfere with axonal transport, ap^GAL4 was used to express a dominant-negative version of the P150/Glued dynactin motor complex component (UAS-Glued^DN), a molecule shown to specifically interfere with retrograde axonal transport. In ap^GAL4/UAS-Glued^DN L1 larvae, a complete loss (0%) of FMRFa-lacZ expression was detected. Similarly, expression of the microtubule binding Tau protein, shown to interfere with axonal transport in Drosophila led to a near complete loss (4%) of FMRFa-lacZ expression (ap^GAL4/UAS-τ-myc). In both UAS-Glued^DN and UAS-τ-myc, normal Tv axon innervation of the DNH was observed in all segments (Allan, 2003).

By co-expressing UAS-EGFP^F in all scenarios outlined above, it was found that loss of FMRFa expression was not due to loss of the Tv cell, since the number of cells within the ap-cluster was unaltered in tin, UAS-robo, UAS-rac^V12, UAS-Glued^DN, and UAS-τ-myc. Using α-Glutactin, it was found that the DNH itself is only affected in tin mutants, not in the other genotypes. Together, these results show that innervation of the DNH and retrograde signaling is essential for the expression of FMRFa (Allan, 2003).

What is the identity of the retrograde FMRFa-inducing signal? Recently, a Drosophila BMP type-II receptor, wishful thinking (wit), was implicated in mediating a retrograde signal from muscles to motor neurons, responsible for presynaptic maturation. Signaling by the TGF-β/BMP superfamily occurs via activation of a receptor complex, consisting of two type I and two type II receptors, leading to phosphorylation and nuclear translocation of a receptor Smad protein. In Drosophila, BMP signaling leads to the phosphorylation and nuclear translocation of the Smad protein Mothers against dpp (Mad), which can be monitored using antibodies specific to phosphorylated Mad (pMad) (Allan, 2003).

Using antibodies to pMad, BMP activation in peptidergic neurons was assayed. Nuclear pMad was detected not only in motor neurons, but also in the Tv, Va, and Vap neurons, demonstrating that peptidergic neurons projecting out of the VNC also show evidence of BMP activation. Accumulation of pMad in the Tv neurons commences during stage 17, immediately following DNH innervation. These results led to a test of whether Tv innervation of the DNH would be critical for pMad accumulation and consequently for FMRFa expression. Indeed, it was found that the absence of the DNH (in tin mutants), Tv axon pathfinding alterations (in ap^GAL4/UAS-robo and ap^GAL4/UAS-rac^V12) and interference with Tv axonal transport (in ap^GAL4/UAS-Glued^DN and ap^GAL4/UAS-τ-myc) are all accompanied by loss of pMad staining specifically in Tv neurons. The ectopic ap-cluster FMRFa-expressing cell induced by sqz misexpression is also pMad positive. Given the role of sqz in Tv axon pathfinding, this is interpreted as resulting from sqz dominantly altering the projection of one other ap-cluster cell, forcing it to innervate the DNH. Thus, in all genotypes examined, Tv axonal projection to the DNH is critical for pMad accumulation (Allan, 2003).

Since Wit is expressed in a restricted pattern in the developing VNC, attempts were made to address whether the Tv neurons express Wit. However, single-cell resolution could not be obtained with the Wit antibody and Wit could not be definitely localized in Tv cells. However, the wit-dependent pMad accumulation in Tv neurons, the ap^GAL4/UAS-tkv^A, UAS-sax^A-mediated rescue of wit mutants, and the UAS-gbb-mediated 'rescue' of UAS-robo misexpression, provide genetic evidence supporting the expression of wit in Tv cells. Previous studies have shown that gbb is expressed in developing endoderm and visceral mesoderm, but it has not been detected in the VNC. By in situ hybridization, no apparent expression was detected in the DNH. Given that the DNH only contains two cell bodies, low-level gbb expression may be beyond detection. Moreover, since the anterior midgut is positioned in very close proximity to the DNHs, it is possible that Gbb diffuses from the visceral mesoderm to the DNH (Allan, 2003).

Why is BMP activation necessary for FMRFa expression? Neither forced axonal exit from the VNC (ap^GAL4/UAS-Unc5) nor autocrine presentation of the Gbb ligand (ap^GAL4/UAS-gbb) leads to activated pMad and FMRFa expression in ap cells other than the Tv cell. This indicates that the Tv cell is uniquely predetermined to respond to the Gbb ligand. In fact, even direct activation of the BMP pathway (UAS-sax^A, -tkv^A;ap^GAL4/+) in all ap neurons does not trigger ectopic FMRFa expression, showing that the Tv cell is further uniquely capable of responding to BMP activation. The misexpression results show that both of these properties of the Tv cell are specified by sqz/ap co-expression. Given this level of Tv cell predetermination, it begs the question as to why Tv cell FMRFa expression evolved to be dependent upon a retrograde BMP signal. Perhaps dependence upon a retrograde signal provides precise control over the onset of FMRFa expression during embryogenesis. In fact, Tv neurons are born by stage 14 (as evident by ap expression) but do not activate FMRFa expression until late stage 17, upon DNH innervation. Alternatively, the presence of a small number of sqz/ap co-expressing cells in the developing brain that do not express FMRFa may necessitate additional regulatory control over FMRFa expression. Dependence upon a signal transduction pathway also provides several unique means of control and amplification of target gene expression. Finally, the fact that sqz, ap, and BMP activation only act to trigger FMRFa expression within a neuropeptidergic cellular context reveals additional complexity underlying the control of specific neuropeptide expression. Given the large number of diverse cell types in the CNS, what may appear to be an almost excessive complexity of combinatorial coding may in fact be essential for high fidelity of gene expression (Allan, 2003).

Ap-let neurons--a peptidergic circuit potentially controlling ecdysial behavior in Drosophila

A set of peptidergic neurons is conserved throughout all developmental stages in the Drosophila central nervous system. A small complement of 28 apterous-expressing cells (Ap-let neurons) in the ventral nerve cord (VNC) of Drosophila larvae co-express numerous gene products. The products include the neuroendocrine-specific bHLH regulator called Dimmed (Dimm), four neuropeptide biosynthetic enzymes (PC2, Fur1, PAL2, and PHM), and a specific dopamine receptor subtype (dDA1). For the PC2, Fur1, and PAL2 enzymes, and for the dDA1 receptor, this neuronal pattern represents the vast majority of their total expression in the VNC. In addition, while Dimm and PHM are present in the peritracheal Inka cells in larvae, pupae, and adults, Ap, PC2, Fur1, PAL2, and dDA1 are not. PC2, PAL2, and DA1 receptor expression are all controlled by both dimm and ap. Previous genetic analysis of animals deficient in PC2 revealed an abnormal larval ecdysis phenotype. Together, these data support the hypothesis that the small cohort of Ap-let interneurons regulates larval ecdysis behavior by secretion of an unidentified amidated peptide(s). This hypothesis further predicts that the production of the Ap-let neuropeptide(s) is dependent on each of four specific enzymes, and that a certain aspect(s) of its production and/or release is regulated by dopamine input (Park, 2004).

Regulators acting in combinatorial codes also act independently in single differentiating neurons

In the Drosophila ventral nerve cord, a small number of neurons express the LIM-homeodomain gene apterous (ap). These ap neurons can be subdivided based upon axon pathfinding and their expression of neuropeptidergic markers. ap, the zinc finger gene squeeze, the bHLH gene dimmed, and the BMP pathway are all required for proper specification of these cells. Here, using several ap neuron terminal differentiation markers, how each of these factors contributes to ap neuron diversity has been resolved. These factors interact genetically and biochemically in subtype-specific combinatorial codes to determine certain defining aspects of ap neuron subtype identity. However, it was also found that ap, dimmed, and squeeze additionally act independently of one another to specify certain other defining aspects of ap neuron subtype identity. Therefore, within single neurons, single regulators acting in numerous molecular contexts differentially specify multiple subtype-specific traits (Allan, 2005).

Within every VNC hemisegment, ap is expressed by one dorsal neuron (dAp) and two ventral neurons (vAp). Additionally, in thoracic VNC hemisegments, ap is expressed by a lateral cluster of four neurons (the ap cluster), termed the Tv, Tvb, Tva, and Tvc neurons. These ap neurons are phenotypically diverse. The axons of most ap neurons project within an ipsilateral fascicle (ap fascicle) that projects to the brain, whereas the axons of the Tv cell exit the VNC at the midline to innervate the dorsal neurohemal organs (DNH). A subset of ap neurons is peptidergic (the Tv, Tvb, and dAp neurons). As is characteristic for the vast majority of Drosophila peptidergic neurons, these cells express high levels of the peptide biosynthetic enzyme peptidylglycine alpha-hydroxylating monooxygenase (PHM). However, this peptidergic subset is also diverse: Tv cells selectively express the dFMRFa neuropeptide, whereas Tvb and dAp cells selectively coexpress three peptide biosynthetic enzymes -- PC2, Furin1, and PAL2 -- although the identity of their secreted neuropeptide(s) remains unknown. This coexpression in Tvb and dAp cells suggested a functional grouping and a common name, 'Ap-let' cells. For clarity, the ap neurons will be considered as three classes: (1) Tv cells express dFMRFa and PHM and innervate the DNH; (2) Ap-let (Tvb and dAp) cells express PHM, PC2, Furin1, and PAL2; (3) the vAp, Tva, and Tvc cells are nonpeptidergic (Allan, 2005).

ap, sqz, dimm, and the BMP pathway act in a combinatorial code to regulate dFMRFa in the Tv cell (ap, sqz, dimm, and the BMP pathway) and furin1 (ap, dimm) in Ap-let cells. Importantly, however, each regulator also plays critical roles within these ap neurons independent of the other regulators. Ap independently acts to regulate axon pathfinding by all ap cells except the Tv. Dimm independently controls PHM in the Tv and Ap-let cells. Moreover, Sqz independently acts via the N pathway to regulate cell identity within the ap cluster, upstream of both Ap and Dimm, apparently by suppressing the Tvb cell fate to favor the Tv fate. The Ap-let cells do not express Sqz, nor do they have an activated BMP pathway. In these neurons, Ap activates the expression of Dimm, and both act together to activate the expression of the peptide-processing enzyme Fur1. The Tva and Tvc cells of the ap cluster do not express Dimm and do not have an activated BMP pathway. Remarkably, the differences inferred between regulatory circuits for the two classes of peptidergic cells are highly reminiscent of differences in regulatory circuits that operate during the differentiation of distinct noradrenergic neurons. Together, these sets of studies support the proposition that epistatic relations between regulators underlying the production of a common phenotype may differ according to cell type (Allan, 2005).

The loss-of-function and gain-of-function phenotypes presented for ap, sqz, dimm, and the BMP pathway, suggest that they act in a combinatorial fashion to regulate dFMRFa expression in the Tv neuron. Likewise, the results indicate that ap and dimm, in the absence of sqz and the BMP pathway, combine to activate Fur1 in the Ap-let neurons, Tvb and dAp. In order to determine whether these regulators act simultaneously on dFMRFa and Fur1, rather than in a genetic hierarchy, the epistatic and biochemical relationship between these regulators were studied. Only one clear epistatic relationship was found; Ap activates the expression of Dimm in the majority of ap neurons. Therefore, it was important to determine whether Dimm acted downstream of Ap to independently and more directly regulate dFMRFa and Fur1 expression. This hypothesis was tested in two complementary tests. (1) Rescuing Dimm function in ap neurons that were absent for Ap function, yielded a nearly complete rescue of dFMRFa in Tv neurons, but only relatively weak rescue of Fur1 in Ap-let neurons. (2) Panneuronal co-misexpression of both ap and dimm triggers ectopic dFMRFa expression in a much greater number of neurons than does either regulator alone. These data indicate that Dimm functions together with Ap to achieve wild-type levels of dFMRFa and, more notably, Fur1. Thus, ap and dimm appear to display both hierarchical and combinatorial interactions. This hypothesis has precedent in studies of the developing pancreas, in which Foxa2 is required for pdx-1 transcription in β cells and later interacts directly with PDX-1 protein to regulate target gene expression, including maintained pdx-1 expression. Biochemical data are also consistent with the possibility that a combinatorial Ap, Dimm, and Sqz code that activates dFMRFa and dFur1 involves direct protein interactions. These may exist within larger complexes bridged by Chip, since Dimm can interact directly with both Ap and Chip, and this in turn may explain why Dimm partially rescues both the ap mutant dFMRFa and Fur1 phenotypes. These multiple interactions are reminiscent of synergistic interactions suggested between mammalian bHLH proteins, LIM-HD proteins, and the Chip homolog, LDB1/NLI. The simplest explanation for restricted patterns of neuropeptides and certain neuropeptide biosynthetic enzymes features a combinatorial hypothesis. More specifically, it is proposed that different combinatorial codes of transcription factors act cell specifically to effect differing patterns of neuropeptides and associated processing enzymes (Allan, 2005).

Ap expression is an early marker of ap cell differentiation, and it is required for proper axonal pathfinding by most ap neurons, although not by the Tv cell. In contrast, neither Sqz nor Dimm appear to control ap cell morphogenesis. An independent role for Sqz occurs early in ap cell differentiation, at a time when postmitotic cell fates are being determined. It is surprising that such cell fate changes can be rescued by UAS-Dl. Why would the frequently used N pathway signaling system depend upon a much more restricted regulator like sqz for proper activity? Increasing evidence points to major mechanistic differences between N signaling during neuroblast specification and during asymmetric division, where asymmetric divisions specifically require neuralized, numb, and sanpodo. No expression of sqz is found in neuroblasts, but expression is evident in many VNC cells. Therefore, it is proposed that factors like Sqz coordinate late N signaling with cell specification and/or cell cycle genes (Allan, 2005).

Dimm acts independently of Ap, Sqz, and the BMP pathway to activate expression of the neuropeptide-processing enzyme PHM. The evidence regarding the independent role of Dimm suggests that it is a master regulator of neuroendocrine cell fate. dimm expression is highly correlated with a neuroendocrine/peptidergic cellular identity, where it regulates the expression of almost all neuropeptides and their processing enzymes examined to date, especially within those neurons that express peptides that are processed to include an α-amidated C terminus. This is a significant cellular pattern, because more than 90% of Drosophila neuropeptides are amidated. Furthermore, high-level expression of the PHM enzyme is absolutely required for amidation and serves as an excellent marker for most peptidergic neurons in Drosophila. Finally, PHM expression appears to be dedicated to neuroendocrine peptide biosynthesis; it is exclusively found within the luminal domain of secretory vesicles. Thus, PHM expression provides a faithful marker for the peptidergic/neuroendocrine cell fate. This study has shown that PHM is dominantly induced by dimm overexpression throughout most or all of the CNS. This evidence, together with the loss-of-function data argues strongly that dimm is a neuroendocrine master regulator, with properties akin to those of other bHLH proteins in regulating cell fate (Allan, 2005).

As anticipated, more restricted peptidergic traits such as dFMRFa and Fur1 expression are dependent upon combinatorial codes. Importantly, however, the selection of cell-specific peptidergic markers arises from a deterministic interaction between a peptidergic master regulator and a cell-specific combinatorial code. There exists a clear analogy between the action of dimm in developing neurons and results regarding the glial cells missing (gcm) gene. Studies have shown that gcm is both necessary and sufficient for glial cell specification within the DrosophilaVNC. gcm is able to ectopically activate generic glial genes, such as reversed polarity, and also activates subclass-specific glial genes, but only in certain prescribed subsets of cells. Thus, similar to gcm, it is predicted that dimm is a master regulator of core neuroendocrine genes in most peptidergic/neuroendocrine cells. It will be of interest to determine which genes beyond PHM are under dimm control. In parallel, dimm combines with local-acting factors to help activate subclass-specific genes (e.g., neuropeptide-encoding genes) within peptidergic cell subsets (Allan, 2005).

The genes studied here combine to regulate dFMRFa and Fur1 but also have independent roles within the same cells. This raises the issue of how Dimm, for instance, can complex with Ap/Sqz on dFMRFa and also act independently on PHM within the same nucleus. Surprisingly, no clear evidence of an antagonistic relationship between the individual roles of Ap, Sqz, and Dimm was found. For example, co-misexpression of ap with dimm does not obviously suppress the ectopic PHM expression observed when dimm alone is misexpressed. Likewise, misexpression of sqz in the Fur1-expressing dAp/Tvb cells does not suppress Fur1. Thus, it appears that the independent mechanisms of regulator action are robust and can coexist with combinatorial functions. Therefore, it is proposed that these regulators operate within a bistable organizational mechanism. With respect to independent roles, it is proposed that Dimm operates independently of Ap and Sqz to dominantly induce specific target genes (e.g., PHM) within all neuronal lineages by forming heterodimers with a class A bHLH like Da, or by forming homodimers. The Drosophila bHLH Twist protein has distinct regulatory roles in vivo, acting either as a heterodimer with Da, or as a homodimer. Notably, the mammalian ortholog of Dimm, Mist1, forms functional homodimers to promote the differentiation of pancreatic secretory cells (Allan, 2005).

The TGFβ/BMP signal transduction pathway plays critical roles during a number of developmental events, and mutants affecting the Drosophila BMP pathway show dramatic defects in embryonic development. In contrast, in the Tv neuron, BMP signaling plays a much more subtle role, and although it is critical for dFMRFa expression, no effects were found upon the expression of sqz, ap, or dimm or on the general peptidergic marker PHM in wit mutants. Although these studies cannot rule out other roles for the BMP pathway in Tv neurons, it is tempting to speculate that target-derived BMP signaling in neurons may have quite a limited set of nuclear readouts in each specific neuronal subclass (Allan, 2005).

Apterous directly regulates a Serrate promoter element

Drosophila wing development is a useful model to study organogenesis, which requires the input of selector genes that specify the identity of various morphogenetic fields and cell signaling molecules. In order to understand how the integration of multiple signaling pathways and selector proteins can be achieved during wing development, the regulatory network that controls the expression of Serrate (Ser), a ligand for the Notch (N) signaling pathway, which is essential for the development of the Drosophila wing, as well as vertebrate limbs, was examined. A 794 bp cis-regulatory element located in the 3' region of the Ser gene can recapitulate the dynamic patterns of endogenous Ser expression during wing development. Using this enhancer element, Apterous (Ap, a selector protein), and the Notch and Wingless (Wg) signaling pathways, are shown to sequentially control wing development through direct regulation of Ser expression in early, mid and late third instar stages, respectively. In addition, later Ser expression in the presumptive vein cells is controlled by the Egfr pathway. Thus, a cis-regulatory element is sequentially regulated by multiple signaling pathways and a selector protein during Drosophila wing development. Such a mechanism is possibly conserved in the appendage outgrowth of other arthropods and vertebrates (Yan, 2004).

Ser is expressed in the dorsal compartment during the early stages of wing disc development. This expression pattern is identical to that of the selector gene of the dorsal compartment, ap, which encodes a homeodomain transcription factor. It has been hypothesized that early Ser expression in the dorsal compartment is under the direct control of Ap. However, no direct evidence has been shown to support this hypothesis. To determine whether Ser is a direct target gene of Ap, whether the 794 bp Ser minimal wing enhancer is regulated by Ap was tested. Construct 10, Ser-lacZ containing the 794 bp Ser minimal enhancer, is expressed in a stripe in the dorsal compartment flanking the DV boundary at 24 hours after the L2/L3 molt in early third instar. A constitutively active form of Ap (ChAp) was expressed using the Gal4/UAS system and Ser-lacZ expression was examined. When Dpp-Gal4 was used to drive ChAp expression at the anteroposterior (AP) boundary, ectopic Ser-lacZ expression was found in the ventral wing regions along the AP boundary, overlapping dpp-Gal4 expression in early and late third instar. This indicated that Ap is sufficient to activate Ser expression, probably cell-autonomously. To determine whether Ap function is necessary for Ser expression, an Ap antagonist, dLMO, was expressed in cells along the AP boundary, using a patched (ptc) promoter. This led to the loss of Ser-lacZ expression in the early third instar and partial reduction of Ser-lacZ in the late third instar, suggesting that Ap is required in vivo for Ser expression in the dorsal compartment (Yan, 2004).

To test whether early Ser expression can be directly regulated by Ap, DNaseI footprinting analysis was used to determine the interaction sites between the 794 bp DNA sequence and Ap. A total of 14 protected Ap binding sites were detected spanning the 794 bp element. The binding of Ap to this Ser minimal wing enhancer is sequence specific with two major binding sequences, TAATNN and CAATNN. The TAATNN consensus sequence matches the six-nucleotide consensus binding sequence for homeodomain proteins. There is also the non-canonical CAATNN consensus sequence derived from the aligned sequences, which matches the consensus binding sites for some homeodomain proteins, such as murine S8. The existence of four CAATNN sites suggests that Ap may bind the CAATNN sequences specifically, in addition to the canonical TAATNN sites (Yan, 2004).

To test whether these Ap binding sites are functionally important in vivo, nucleotides in the Ap-binding sequences of Ser-lacZ construct 10 were mutagenized from TAATNN and CAATNN to AAAANN or TTTTNN, in most cases. The (mAp)Ser-lacZ construct, which included mutations in all 14 Ap-binding sites, showed no enhancer activity in the wing and haltere discs in early third instar, as compared with Ser-lacZ expression, which was first detected in much of the dorsal compartment and then as a dorsal stripe. In mid and late third instar, (mAp)Ser-lacZ expression was reduced or eliminated. These results show that the Ap-binding sites identified in vitro are crucial for the activity of the 794 bp Ser minimal wing enhancer in vivo. In summary, Ser expression is mediated by direct Ap interaction with the 794 bp wing enhancer during the early third instar stage were mutagenized (Yan, 2004).

Given that the Ser-Fng-N pathway is evolutionarily conserved in appendage development between insects and vertebrates, the mechanism by which Ser is sequentially regulated by Ap, N, Wg and Egfr may also be conserved in appendage outgrowth of other arthropods and vertebrates. Consistent with this hypothesis, the Ap, Wg/Wnt and Egfr/Fgf pathways are also involved in appendage development in vertebrates, as well as D. melanogaster. Indeed, a BLAST search of the Drosophila pseudoobscura genome identified a putative homolog of the Ser minimal wing enhancer. Interestingly, this enhancer region is also located less than 1 kb downstream of the putative D. pseudoobscura Ser 3'UTR. Sequence comparisons between the Ser minimal wing enhancer from D. melanogaster and the putative D. pseudoobscura enhancer show a significant degree of similarity, whereas the similarities in the 5' and 3' flanking regions are lower. Importantly, sequences of putative Ap, Su(H) and dTCF binding sites are highly conserved in D. pseudoobscura and D. melanogaster. Although the strong conservation of sequence and location suggests that the putative D. pseudoobscura Ser enhancer may be a functional homolog of the D. melanogaster Ser minimal wing enhancer, it remains to be tested whether this enhancer drives reporter gene expression at the identical time and location in the D. melanogaster wing discs (Yan, 2004).

Squeeze and Apterous involvement in the specification of Drosophila leucokinergic neurons

One of the most widely studied phenomena in the establishment of neuronal identity is the determination of neurosecretory phenotype, in which cell-type-specific combinatorial codes direct distinct neurotransmitter or neuropeptide selection. However, neuronal types from divergent lineages may adopt the same neurosecretory phenotype, and it is unclear whether different classes of neurons use different or similar components to regulate shared features of neuronal identity. This question was addressed by analyzing how differentiation of the Drosophila larval leucokinergic system, which is comprised of only four types of neurons, is regulated by factors known to affect expression of the FMRFamide neuropeptide. All leucokinergic cells express the transcription factor Squeeze (Sqz). However, based on the effect on LK expression of loss- and gain-of-function mutations, three types of Lk regulation are described. In the brain LHLK (lateral horn leukokinin) cells, both Sqz and Apterous (Ap) are required for LK expression, but surprisingly, high levels of either Sqz or Ap alone are sufficient to restore LK expression in these neurons. In the s SELK cells, Sqz, but not Ap, is required for LK expression. In the abdominal ABLK neurons, inhibition of retrograde axonal transport reduces LK expression, and although sqz is dispensable for LK expression in these cells, it can induce ectopic leucokinergic ABLK-like cells when over-expressed. Thus, Sqz appears to be a regulatory factor for neuropeptidergic identity common to all leucokinergic cells, whose function in different cell types is regulated by cell-specific factors (Herrero, 2007).

It has been shown that Ap is required for LK expression only in LHLK cells. Ap is also necessary for proper transcription of the Fmrf gene in the thoracic Tv neurons. In attempts to understand the mechanisms underlying leucokinergic differentiation, it was asked whether other factors known to control expression of the FMRFamide neuropeptide, i.e., Sqz and the BMP signalling pathway, affected LK expression. Indeed, the number of LK-immunopositive cells is strongly reduced in sqz^lacZ mutant larvae. It has been proposed that Apterous is not necessary for the emergence and maintenance of LHLK cells. Earlier reports have established that the Tv cells are present in sqz mutants, although they do not express Fmrf. Are the LK cells present in sqz mutants? This question could not be directly addressed due to the lack of independent markers for following the fate of the LK cells, but the results presented in this study indicate that in sqz mutants, leucokinergic cells are born, but fail to express LK. First, LK expression is restored by postmitotic expression of Sqz. Second, LK-immunoreactive cells are detected in sqz mutant brains during early larval stages, and later disappear. Finally, the few LK-expressing cells detected in sqz mutants often show very faint immunostaining, which suggests that reduction of Sqz protein decreases Lk transcription but does not affect cell survival. The reduction in the number of leucokinergic cells in sqz mutants even in first instar larvae indicates that sqz is necessary for induction of LK expression, and the weaker phenotype observed in early larval stages may reflect perdurance of the wild type product supplied by the mother, or a requirement for sqz also for maintenance of LK expression (Herrero, 2007).

Although all leucokinergic neurons express sqz, they differ in how this transcription factor affects LK expression. In this study, at least three neuronal types have been identified based on the components that regulate LK expression: (1) in the LHLK neurons, LK expression is controlled by the transcription factors Ap and Sqz; (2) in the SELK neurons, Sqz, but not Ap, is necessary for wild type Lk transcription; (3) in the ABLK neurons, Sqz is dispensable for LK expression even though it can induce leucokinergic ABLK-like cells, and LK expression is affected by inhibition of retrograde axonal transport only in the ABLK cells. Regarding the ALK neurons, the number of this cell type is highly variable, and thus was not further analyzed in this study (Herrero, 2007).

By analyzing two different sqz alleles, it was shown that, besides differences in the components that control LK expression, the amount of Sqz protein required to achieve wild type LK expression also varies. RealTime PCR analysis indicates that the sqz^lacZ allele has undetectable levels of sqz transcripts, while the sqz^GAL4 allele has reduced, but measurable sqz transcription. Consistent with these results, sqz^lacZ mutant larvae show a large reduction in the number of LHLK and SELK cells detected by anti-LK immunostaining, while sqz^GAL4 mutants do not affect LK expression in the LHLK neuron, but do show a significant decrease in the number of LK-immunopositive SELK cells. Moreover, restoring neuronal Sqz protein with the elav^GAL4 driver can completely rescue LK expression in LHLK neurons in sqz^lacZ mutants, but only partially rescue LK in SELK neurons. The differential effect of these two sqz alleles suggests that there is a threshold of Sqz protein below which Lk transcription is prevented, and this threshold is higher in SELK cells than in LHLK cells (Herrero, 2007).

This study has demonstrated that the reduced LK expression observed in the LHLK cells of sqz mutants is fully rescued when over-expressing the lost protein, indicating that this protein is indeed responsible for the observed phenotype. In addition, the data show that LK expression in these cells does not depend on the Gbb/Wit or Babo/Activin signalling cascades, or any extrinsic retrograde signal. Thus, neuropeptidergic identity of the LHLK neurons is controlled by Ap and Sqz, but not BMP. These cells also express the bHLH transcription factor Dimm, which has been shown to act post-tanscriptionally on the regulation of LK expression in these cells. This combination of factors (i.e., Ap, Sqz, Dimm, but not BMP signalling) is different from any of the known codes that regulate the neuropeptidergic differentiation of the peptidergic neurons in the thoracic ap cluster, which includes the Furin 1-expressing Ap-let cells that do not express Sqz, and the Fmrf-expressing Tv cell, which requires extrinsic signalling through Gbb/Wit (Herrero, 2007).

These results demonstrate that high levels of either Ap or Sqz protein alone are sufficient to induce LK expression in the LHLK cells. Thus, based on the cross-rescue experiments, it can be inferred that Sqz can form transcriptionally active complexes with proteins other than Ap to promote Lk expression. Similarly, Ap is able to promote Lk expression when it complexes with proteins other than Sqz. However, wild type levels of LK expression are only achieved when both proteins are present. It has been shown that Sqz and Ap can physically interact, and that both proteins can form separate complexes with Chip. Chip is involved in Lk regulation, because LK expression in LHLK cells is affected when dLMO, a cofactor that binds Chip and prevents Ap action, is expressed in ap-expressing cells. Thus, Ap and Sqz could weakly activate Lk transcription independently by interacting with Chip and/or other transcription factors, but strong activation would require a synergistic interaction mediated by complexes containing both proteins. A similar mechanism is exerted by Brn2 and Otp to stimulate transcription of the corticotrophin-releasing hormone gene in the neuroendocrine hypothalamic neurons, and by Cdx-2 and Brn-4 to activate expression of the proglucagon gene in pancreatic B-cell lines (Herrero, 2007).

It has been reported that Lk transcription in dimm mutants is downregulated in ABLK cells, while LHLK and SELK cells show slightly upregulated Lk transcription but reduced levels of LK peptide. It was found that the ABLK neurons also differ from LHLK and SELK neurons in that Sqz appears to be dispensable for LK expression in these cells. However, neuronal over-expression of Sqz can frequently induce ectopic LK expression in ABLK-like cells in an ap-independent fashion. Likewise, misexpression of the mammalian homeodomain transcription factor Phox2a in the sympathetic ganglion induces ectopic noradrenergic neurons, even though in the absence of Phox2a, sympathetic development is largely normal (Herrero, 2007).

Another peculiarity of the ABLK cells is that LK expression in these cells seems to depend on a retrograde signal, different from BMP, because the number of ABLK LK-expressing cells is reduced when axonal transport is inhibited. Although the degree of reduction is variable, it is indeed highly significant, because as few as two cells can be found when Glued^DN is pan-neurally expressed. This variability could be due to partial inhibition of retrograde transport, as suggested by the late pupal lethality of Glued^DN-expressing flies, instead of the earlier embryonic lethality of Glued amorphic mutations. In the thoracic ap-cluster, projection of the Tv cell outside the CNS is essential for its response to Gbb and subsequent Fmrf expression. Likewise, ABLK cells are known to project outside the CNS, and this may be essential for them to react to an extrinsic signal important for induction and/or maintenance of LK expression. Alternatively, extrinsic signalling may control guidance and/or targeting of ABLK axons, as suggested by the ectopic leucokinergic varicosities found in the abdominal neuropile of Glued^DN-expressing ganglia, and proper axonal targeting may be an important factor in the regulation of LK expression. Further experiments will be necessary to test these hypotheses, and rule out a possible deleterious effect of Glued^DN on ABLK viability (Herrero, 2007).

Ectopic leucokinergic cells can be induced by over-expressing Sqz, but not by Ap. Moreover, ectopic LK neurons can also be classified into three different groups: first, the ectopic LHLK-like cells, which appear to depend on the relative amount of Sqz and Ap; second, the ABLK-like cells, which require high levels of Sqz expression, but not ap expression; and finally, the ectopic brain cells, which are the only leucokinergic ectopic cells generated by ectopic expression of Sqz. These latter cells must have unidentified components present in other leucokinergic cells that enable them to activate Lk transcription when Sqz is present (Herrero, 2007).

An ectopic LHLK-like cell is present in a small percentage of the sqz^GAL4 mutants. Appearance of this ectopic cell requires Ap, because ap is expressed in this cell and its appearance is prevented in the ap^UGO35 null allele. In the thoracic ganglion, sqz determines the number and identities of cells of the ap cluster, so that in szq null mutants an extra Dimm and Furin-expressing cell was generated in every ap cluster. Attempts to analyze changes in the number of ap and dimm expressing cells in sqz mutants was precluded by the large number of ap cells surrounding the LHLK cell, and the inconsistent expression of the dimm-GAL4 driver c929 in the LHLK cells, in both wild type and mutant brains. However, the results obtained when Sqz and/or Ap were over-expressed in a wild type background suggest a more complex scenario. First, a phenotype was obtained equivalent to that observed in sqz^GAL4 mutants, i.e., the appearance of ectopic LHLK-like cells, when Sqz was over-expressed with sqz^GAL4 and with the postmitotic ap- and elav- GAL4 drivers. Moreover, these ectopic LK cells disappeared if Ap was coexpressed with Sqz, even though over-expression of Ap alone had no effect whatsoever on LK expression. Based on these data, and on the different levels of Ap protein in cells surrounding the LHLK neuron, it is hypothesized that the relative dose of Ap and Sqz, rather than the absolute amount of Sqz, is essential for specifying the correct number of LHLK cells with detectable LK, and that this process is controlled postmitotically. According to this model, when Sqz is reduced, those cells with low Ap levels would reach an optimum stoichiometric Sqz/Ap ratio leading to ectopic LK expression, while when Sqz is increased, the correct ratio will be present in cells with high Ap levels. In this last case, further increasing Ap would drive the Sqz/Ap ratio away from the optimum for LK expression, and ectopic cells would not be produced (Herrero, 2007).

Two other situations lead to ectopic leucokinergic LHLK-like cells. (1) A small percentage of ectopic LHLK cells was detected in dac mutants. Because Ap has been shown to repress dac expression in the thoracic Tvb cell, it is possible that repression of dac is required to restrict LK expression to the LHLK neuron. (2) Over-expression of the transcription factor Dimm produces ectopic LHLK-like cells, albeit at much higher frequency. Ap also regulates dimm transcription in most cells. Thus, Dimm over-expression might overcome the requirement for a Sqz/Ap optimum ratio to induce LK expression, provided additional necessary factors are also present in the cell. More experiments will be needed to understand the role of Dac and Dimm, and their interaction with Ap and Sqz, in regulating LK expression in the LHLK cell (Herrero, 2007).

Deciphering how cells of different origin acquire the same neurosecretory identity is one of the major challenges in neuronal development. Much of the progress in this field has been achieved by studying the vertebrate catecholaminergic system, in which central and peripheral noradrenergic neurons use a similar combination of factors to specify their neurotransmitter phenotype, but they differ in the hierarchical relations between these factors, and recruit components specific to each neuronal type. It is tempting to speculate that analogous mechanisms may control Lk expression in leucokinergic cells in Drosophila. Consistent with this hypothesis, Sqz, and probably Dimm, appear to be regulatory factors for neuropeptidergic identity common to all leucokinergic cells. However, sqz affects LK expression in each cell subclass in very different ways, and cell-type-specific regulatory modes on Lk have also been described for Dimm. Moreover, additional cell-type-specific factors regulate LK expression in different leucokinergic cells, such as Ap in the LHLK cells, and an unidentified retrograde signal in the ABLK cells. Thus, the leucokinergic neurons comprise a simple, genetically amenable system for understanding how complex regulatory networks confer a similar neurosecretory identity on cells of different origins (Herrero, 2007).

Transcriptional regulation by CHIP/LDB complexes

It is increasingly clear that transcription factors play versatile roles in turning genes 'on' or 'off' depending on cellular context via the various transcription complexes they form. This poses a major challenge in unraveling combinatorial transcription complex codes. This study used the powerful genetics of Drosophila combined with microarray and bioinformatics analyses to tackle this challenge. The nuclear adaptor CHIP/LDB is a major developmental regulator capable of forming tissue-specific transcription complexes with various types of transcription factors and cofactors, making it a valuable model to study the intricacies of gene regulation. To date only few CHIP/LDB complexes target genes have been identified, and possible tissue-dependent crosstalk between these complexes has not been rigorously explored. SSDP proteins protect CHIP/LDB complexes from proteasome dependent degradation and are rate-limiting cofactors for these complexes. By using mutations in SSDP, 189 down-stream targets of CHIP/LDB were identified; these genes are enriched for the binding sites of Apterous (AP) and Pannier (PNR), two well studied transcription factors associated with CHIP/LDB complexes. Extensive genetic screens were performed and target genes were identified that genetically interact with components of CHIP/LDB complexes in directing the development of the wings (28 genes) and thoracic bristles (23 genes). Moreover, by in vivo RNAi silencing, novel roles were uncovered for two of the target genes, xbp1 and Gs-alpha, in early development of these structures. Taken together, these results suggest that loss of SSDP disrupts the normal balance between the CHIP-AP and the CHIP-PNR transcription complexes, resulting in down-regulation of CHIP-AP target genes and the concomitant up-regulation of CHIP-PNR target genes. Understanding the combinatorial nature of transcription complexes as presented here is crucial to the study of transcription regulation of gene batteries required for development (Bronstein, 2011).

Drosophila SSDP was identified on the basis of its ability to bind the nuclear adaptor protein CHIP/LDB (van Meyel, 2003; Chen, 2002). Both nuclear localization of SSDP and its ability to modulate the transcription activity of the CHIP-AP complex during wing development depend on its interaction with CHIP/LDB. This study implemented a combination of molecular, bioinformatic and genetic approaches that allowed has led to insight into the effect of SSDP on the transcriptional activity of CHIP/LDB complexes and their role in development. A genome wide screen was conducted for SSDP target genes in Drosophila using expression microarrays with mRNA isolated from larvae bearing hypomorphic alleles of ssdp. Analysis of transcription factor binding site enrichment served as an orthogonal assay that validates and extends the microarray results and thus contributes to understanding of the relation between the CHIP-AP and CHIP-PNR transcription complexes in specific tissues (e.g. wing and thorax) (Bronstein, 2011).

SSDP proteins directly bind DNA and mouse SSDP1 activates the expression of a reporter gene in both yeast and mammalian cells indicating that it is capable of regulating transcription activity. Enrichment was found for SSDP binding sites upstream of the genes identified in the microarray experiments on flies lacking SSDP. Moreover, in agreement with the positive transcriptional role of SSDP, enrichment for SSDP binding sites was restricted to the genes showing decreased expression in mutants. This strongly suggests that a significant number of these genes are bona fide SSDP target genes (Bronstein, 2011).

Consistent with the involvement of SSDP with the CHIP-AP complex, it was found that upstream regulatory regions of the SSDP putative target genes are also enriched for the AP binding site and the SSDP binding site. These sites are likely to be functionally significant, since loss of ssdp enhances the wing notching phenotype of a dominant allele of ap. Additionally, over-expression of Dlmo, whose product negatively regulates the CHIP-AP complex, also interacts with mutants of SSDP target genes, demonstrating that SSDP target genes are involved in the CHIP-AP pathway. The efficiency of finding genetic interactions among the genes differentially expressed in the microarray experiments, demonstrated the power of this approach. Specifically, 72% of the loci tested with Dlmo^Bx2 is more than an order of magnitude higher than an EP insertion screen (1.3% interacting) in a Dlmo^Bx1 sensitized background. Combined microarray and genetic loss of function screen allowed the identification of a similar number of Dlmo-interacting genes by screening a much smaller group of putative target genes (Bejarano, 2008). Of the 35 genes identified by Bejarano only CG1943 was found in the 189 genes identified in the current microarray screen. This study specifically identified down-stream targets of SSDP, while Bejarano searched for any modifiers of the Dlmo wing notching phenotype and thus uncovered genes that function in other regulatory pathways or genes that are upstream of the CHIP-AP complexes. This may explain the limited overlap between the current results and those of Bejarano (Bronstein, 2011).

In contrast to the enrichment of SSDP binding sites in the genes down-regulated in ssdp mutants, the PNR binding site was enriched specifically in the genes up-regulated in the ssdp mutants. A model is therefore presented in which loss of SSDP disrupts the balance between the CHIP-AP and CHIP-PNR complexes. Mammalian SSDP proteins protect LDB, LHX and LMO proteins from ubiquitination and subsequent proteasome-mediated degradation by interfering with the interaction between LDB and the E3 ubiquitin ligase, RLIM. It is therefore possible that in the absence of SSDP proteins, CHIP/LDB and LMO can escape degradation by interacting with GATA and beta-HLH proteins that are not subjected to proteasome-mediated regulation. The N-terminus of CHIP/LDB proteins is responsible for interaction with both PNR and RLIM. Thus, PNR/GATA proteins may partially interfere with the interaction between CHIP/LDB and RLIM making the CHIP/LDB-PNR/GATA complex more resistant to proteasome regulation and less dependant on the levels of SSDP proteins then the CHIP/LDB-LHX/AP complex (Bronstein, 2011).

According to the current model, in cells where both the CHIP-AP and CHIP-PNR complexes are active, loss of SSDP should result in the same phenotype as over-expression of PNR. Indeed, it was found that ssdp^L7/+ flies display duplications of scutellar sensory bristles, similar to gain of function mutations in pnr. In addition, lowered levels of pnr in ssdp^L7/+; pnr^VX6/+ flies suppresses scutellar bristle duplication. This indicates that the duplicated scutellar bristle phenotype of ssdp^L7/+ flies depends on the presence of PNR. As predicted by the model, since both AP and PNR regulate bristle formation, the functional interactions between SSDP target genes and ssdp^L7 and/or Chip^e5.5 resulted in either suppression or enhancement of the duplicated scutellar bristle phenotype (Bronstein, 2011).

These results in flies indicate that SSDP contributes differentially to CHIP/LDB complexes containing AP versus PNR. By contrast, mouse SSDP proteins positively contribute to the transcription activity and assembly of both LDB-GATA and LDB-LHX complexes, but the relative contribution of mammalian SSDP proteins to LDB complexes containing LHX proteins versus GATA proteins has not been specifically examined. It is possible that SSDP alters the balance of LIM-based CHIP/LDB complexes and GATA-containing CHIP/LDB complexes in the development of mice, as occurs in flies (Bronstein, 2011).

The search for enrichment of transcription factor binding sites upstream of the putative SSDP target genes identified additional transcription factors that may warrant future study. Some of these factors are associated with SSDP and CHIP/LDB complexes. For example, the binding sites for PNR and ZESTE (Z) were both enriched in the up-regulated putative SSDP target genes. This is in agreement with previous studies showing that Z can recruit the BRAHMA (BRM, the Drosophila homolog of the yeast SWI2/SNF2 gene) complex via its member OSA, which together negatively regulate the CHIP-PNR complex during sensory bristle formation through direct and simultaneous binding of OSA to both CHIP and PNR (Bronstein, 2011).

Some of the additional regulatory inputs at SSDP target genes may be evolutionarily conserved. For example, enrichment of STAT92E and SSDP binding sites was found in the down-regulated SSDP target genes. This may be significant, as a known role of ssdp is regulation of the JAK/STAT pathway during Drosophila eye development. Interestingly, mammalian STAT1 confers an anti-proliferative response to IFN-γ signaling by inhibition of c-myc expression. Similarly, expression of mammalian SSDP2 in human acute myelogenous leukemia cells and prostate cancer cells leads to cell cycle arrest and inhibits proliferation accompanied by down-regulation of C-MYC. These findings indicate that both in Drosophila and in mammals SSDP and STAT proteins have similar functions and may share common target genes (Bronstein, 2011).

While the transcription factor binding site analysis utilized all of the 189 putative SSDP target genes, genetic screens were conducted on a subset of them due to the availability of mutants. This suggests that more genetic interactions will be found among the untested genes. Even among this more limited subset, there are interesting new stories that suggest future experimental directions. For example, an insertion mutation in the Xbp1 gene suppressed the duplicated scutellar bristle phenotype characteristic of ssdp^L7/+ and Chip^e5.5/+ flies, indicating that XBP1 contributes positively to bristle formation. In contrast, when Xbp1 was silenced in ap-expressing cells both the wings and the scutum displayed a marked excess of sensory bristles while the scutellum was not affected. These results suggest that in the wing and scutum XBP1 acts as a negative regulator of bristle formation. Silencing of Xbp1 in pnr-expressing cells caused a similar excess of bristle on the scutum, accompanied by a reduced number of scutellar bristles, further emphasizing the opposing effects of XBP1 in these two distinct parts of the thorax. Such contrasting phenotypes have been previously documented for several pnr mutants as well. In flies and mammals XBP1 regulates the ER stress response, also termed the unfolded protein response (UPR). Since one of the functions of the ER is the production of secreted proteins, UPR-related pathways are widely utilized during the normal differentiation of many specialized secretory cells. In this respect it would be interesting to examine whether SSDP and CHIP/LDB complexes affect the production of secreted morphogens, such as Wingless (WG), the secreted ligands of the EGFR receptor, Spitz (SPI) and Argos (AOS), or the secreted Notch binding protein Scabrous (SCA) via XBP1 during wing and sensory bristle formation. Alternatively, the transcription factor XBP1 may directly regulate the expression of genes required for differentiation of the wing and sensory bristles. Indeed, carbohydrate ingestion induces XBP1 in the liver of mice, which in turn directly regulates the expression of genes involved in fatty acid synthesis. This role of XBP1 is independent of UPR activation and is not due to altered protein secretory function. Curiously, the two GO function categories 'cellular carbohydrate metabolism' and 'cellular lipid metabolism' which are enriched among Xbp1 target genes in mouse skeletal muscle and secretory cells were also enriched in the list of putative SSDP target genes. Whether this reflects a secondary effect due to the down-regulation of Xbp1 in ssdp mutants or a direct regulation of these processes by SSDP is yet to be determined (Bronstein, 2011).

Additional novel functions for CHIP/LDB complexes are implied by the results regarding the Gs-alpha60A (a.k.a. CG2835) gene. G protein coupled receptors are important regulators of development by for example, signaling via the protein kinase A (PKA) pathway. Activation or inhibition of PKA signaling during pupal wing maturation perturb proper adhesion of dorso-ventral wing surfaces resulting in wing blistering. This phenotype may be due to miss-regulation of wing epithelial cell death in ap-expressing cells. Interestingly, similar wing blisters occur in the wing of Dlmo^Bx2 flies. Moreover, it was found that mutant alleles of Gs-alpha60A enhanced the wing blistering phenotype of Dlmo^Bx2. Silencing of G-salpha60A in ap-expressing cells caused a curled wing phenotype. Such a phenotype can result from differences in the size of the dorsal and ventral wing blade surfaces. In addition, silencing of this gene in pnr-expressing cells caused the posterior pair of scutellar bristles to form in reversed orientation. Bristle orientation have been proposed to be regulated by planar cell polarity genes. Taken together these results point to novel aspects of regulation of wing and sensory bristle development by SSDP and CHIP/LDB complexes mediated by G-alpha proteins (Bronstein, 2011).

This genome-wide expression profiling and bioinformatics analysis of ssdp mutant larvae, combined with genetic screens resulted in gained insight into the intricate context-dependent transcriptional regulation by CHIP/LDB complexes. It was possible to identify 28 putative SSDP target genes that are involved in wing development and 23 putative SSDP target genes that play a role in scutellar bristle formation. Examination of two of these, xbp1 and Gs-alpha60A, suggests novel aspects of developmental regulation such as the involvement of SSDP and CHIP/LDB complexes in ER function and PKA signaling. Furthermore, it was shown that SSDP proteins contribute differentially to transcription activity, and probably to the balance in formation of CHIP-AP and CHIP-PNR complexes. Furthermore potential novel partners of SSDP in regulating transcription of downstream genes during fly development were. It stands to reason that an extension of the genetic analysis to mammals and other vertebrates will reveal a host of additional functions of SSDP and CHIP/LDB during the multifaceted process of transcriptional regulation that underlies the development of multicellular organisms (Bronstein, 2011).

Neuronal cell fate specification by the molecular convergence of different spatio-temporal cues on a common initiator terminal selector gene

The extensive genetic regulatory flows underlying specification of different neuronal subtypes are not well understood at the molecular level. The Nplp1 neuropeptide neurons in the developing Drosophila nerve cord belong to two sub-classes; Tv1 and dAp neurons, generated by two distinct progenitors. Nplp1 neurons are specified by spatial cues; the Hox homeotic network and GATA factor grn, and temporal cues; the hb -> Kr -> Pdm -> cas -> grh temporal cascade. These spatio-temporal cues combine into two distinct codes; one for Tv1 and one for dAp neurons that activate a common terminal selector feedforward cascade of col -> ap/eya -> dimm -> Nplp1. This study molecularly decodes the specification of Nplp1 neurons, and finds that the cis-regulatory organization of col functions as an integratory node for the different spatio-temporal combinatorial codes. These findings may provide a logical framework for addressing spatio-temporal control of neuronal sub-type specification in other systems (Stratmann, 2017).

The Drosophila ventral nerve cord (VNC; defined here as thoracic segments T1-T3 and abdominal A1-A10) contains ~10,000 cells at the end of embryogenesis, which are generated by a defined set of ~800 neuroblasts (NBs). The Apterous neurons constitute a small sub-group of interneurons, identifiable by the selective expression of the Apterous (Ap) LIM-homeodomain factor, as well as the Eyes absent (Eya) transcriptional co-factor and nuclear phosphatase. A subset of Ap neurons express the Nplp1 neuropeptide, but can be sub-divided into the lateral thoracic Tv1 neurons, part of the thoracic Ap cluster of four cells, and the dorsal medial row of dAp neurons. In line with the distinct location of the Tv1 and dAp neurons, studies have revealed that they are generated by distinct NBs; NB5-6T and NB4-3, respectively. A number of studies have addressed the genetic mechanisms underlying the specification of the Tv1 and dAp neurons, and the regulation of the Nplp1 neuropeptide. These have revealed that two distinct spatio-temporal combinatorial transcription factor codes, one acting in NB5-6T and the other in NB4-3, converge on a common initiator terminal selector gene; collier, encoding a COE/EBF transcription factor. Col in turn is necessary and sufficient to trigger a feed forward loop (FFL) consisting of Ap, Eya and the Dimmed (Dimm) bHLH transcription factor, which ultimately activates the Nplp1 gene. Strikingly, the combinatorial coding selectivity of the spatio-temporal cues combined with the information-coding capacity of the FFL results in the selective activation of Nplp1 in only 28 out of the ~10,000 cells within the VNC. While these genetic studies have helped resolve the regulatory logic of this cell specification event, they have not addressed the molecular mechanisms by which the two different spatio-temporal combinatorial codes intersect upon the col initiator terminal selector, to trigger a common terminal FFL, or the molecular nature of the FFL (Stratmann, 2017).

To address this issue, this study has identified enhancers for Tv and dAp neuron expression for the genes in the common Tv1/dAp FFL: col, ap, eya, dimm and Nplp1. Transgenic reporters were generated for these enhancers, both wildtype and mutant for specific transcription factor binding sites, to test their regulation in mutant and misexpression backgrounds. CRISPR/Cas9 technology was used to delete these enhancers in their normal genomic location to test their necessity for gene regulation. Strikingly, this study found that the distinct upstream spatio-temporal combinatorial codes, which trigger col expression in Tv1 versus dAp neurons, converge onto different enhancer elements in the col gene. Hence, the col Tv1 neuron enhancer is triggered by Antp, hth, exd, lbe and cas, while the dAp enhancer is triggered by Kr, pdm and grn. In contrast to this subset-specific enhancer set-up for col activation, the subsequent, col-driven Nplp1 FFL feeds onto common enhancers in each downstream gene. These findings reveal that distinct spatio-temporal cues, acting in different neural progenitors, can trigger the same FFL by converging on discrete enhancer elements in an initiator terminal selector, to thereby dictate the same ultimate neuronal subtype cell fate (Stratmann, 2017).

This study has been able to molecularly decode the Tv1/dAp genetic FFL cascades, bolstering evidence for a complex molecular FFL, based upon sequential transcription factor binding to the downstream genes. The NB4-3 and NB5-6T neuroblasts are born in different regions of the VNC, and express different spatial determinants i.e., Antp, Lbe, Hth, Exd and Gr. As lineage progression commences, they undergo a programmed cascade of transcription factor expression; the temporal cascade. Early temporal factors Kr and Pdm integrate with Grn in NB4-3, while the late temporal factor Cas integrates with Antp, Lbe, Hth and Exd in NB5-6T, to create two distinct combinatorial spatio-temporal codes. These two codes converge on two different enhancers in the col gene, triggering Col expression, and hence the Nplp1 FFL. The FFL, in this case a so-called coherent FFL, where regulators act positively at one or several steps of a cascade, was first identified in E.coli and yeast regulatory networks, but have also been identified in C.elegans and Drosophila. Coherent FFLs can act as regulatory timing devices, exemplified by the action of col in NB5-6T: The initial expression of col in Ap cluster cells triggers a generic Ap/Eya interneuron fate in all four cells, while its downregulation in Tv2-4 and maintenance in Tv1 helps propagate the FFL leading to Nplp1 expression (Stratmann, 2017).

This study has found that the two different spatio-temporal programs converge on col, but on different enhancer elements. However, neither enhancer element gave complete null effects when deleted. Specifically, the 6.3kb col-Tv-CRM shows robust reporter expression, overlaps with endogenous col expression, responds to the upstream mutants, and is affected by TFBS mutations. However, when deleted (generating the col^ΔTv-CRM mutant), it had weak effects upon endogenous col expression in NB5-6T, and no effect upon Eya and Nplp1 expression. Deletion of the col-dAp-CRM (generating the col^ΔdAp-CRM mutant), gave more robust effects with reduction of Col, Eya and Nplp1 in dAp cells, although the expression was not lost completely (Stratmann, 2017).

Early developmental genes, which often are dynamically expressed, may be controlled by multiple enhancer modules, to thereby ensure robust onset of gene expression. This has been reported previously in studies of early mesodermal and neuro-ectodermal development, in which several genes i.e., twist, sog, snail are controlled by multiple distal enhancer fragments, so called 'shadow enhancers', in order to ensure reliable onset of gene expression. The shadow enhancer principle is also supported by recent findings on the Kr gene. Moreover, extensive CRM transgenic analysis, scoring thousands of fragments in transgenic flies, has also supported the shadow enhancer idea, revealing that a number of early regulators, several of which encode for transcription factors, indeed have shadow enhancers. The dichotomy between the col transgenic reporter results and the partial impact on col expression upon deletion of its Tv1 and dAp enhancers, gives reason to speculate that col may be under control of additional enhancers, some of which may be referred to as shadow enhancers (Stratmann, 2017).

The results on the eya, ap, dimm and Nplp1 enhancer mutants stand in stark contrast to the col CRMs findings. For these four genes, the enhancer deletion resulted in robust, near null effects, on expression. It is tempting to speculate that the current findings, combined with previous studies, points to a different logic for early regulators, with highly dynamic patterns, requiring several functionally overlapping enhancers for fidelity, and late regulators and terminal differentiation genes, which may operate with one enhancer that is inactive until the pertinent combinatorial TF codes have been established (Stratmann, 2017).

Analysis of the ap and eya enhancers indicates that Col directly interacts with these enhancers. Both of these enhancer-reporter transgenes are affected in col mutants, and can be activated by ectopic col. Moreover, mutation of one Col binding site in the ap enhancer and two sites in the eya enhancer, was enough to dramatically reduce enhancer activity. Direct action of Col on ap and eya is furthermore supported by recent data on Col genome-wide binding, using ChIP, which demonstrated direct binding of Col to these regions of ap and eya in the embryo. The regulation of ap is an excellent example of the complexity of gene regulation, and studies have identified additional enhancers controlling ap expression in the wing, muscle and brain (Stratmann, 2017).

In contrast to regulation of ap and eya, a direct action of Col on dimm and Nplp1 is less clear. Analysis of the dimm and Nplp1 enhancers did not reveal perfectly conserved Col binding sites. Mutation of multiple non-perfect Col binding sites in the dimm enhancer did not affect reporter expression in the Ap cluster, but did however reduce levels in the dorsal Ap cells. Mutation of non-perfect Col binding sites in the Nplp1 enhancer had no impact on enhancer activity, neither in Tv1 nor dAp. These findings support a model where Col is crucial for directly activating ap and eya, which in turn directly activate dimm and Nplp1, with some involvement of Col on dimm. However, support for a direct role for Col on Nplp1 comes from RNAi studies in larvae or adult flies, showing that knockdown of col resulted in loss of Nplp1, while Ap, Eya and Dimm expression was unaffected (Stratmann, 2017).

It is tempting to speculate that Col regulates Nplp1 not via direct interaction with its enhancer, but rather as a chromatin state modulator, keeping the chromatin around the Nplp1 locus in an accessible state, in order for Dimm, Ap and Eya to be able to access the Nplp1 gene. Support for this notion comes from studies on the mammalian Col orthologue EBF, which is connected to the chromatin remodeling complex SWI/SNF during EBF-mediated gene regulation in lymphocytes (Gao, 2009). Moreover, the central SWI/SNF component Brahma was recently identified in a genetic screen for Ap cluster neurons, and found to affect FMRFa neuropeptide expression in Tv4 without affecting Eya expression, indicating a late role in Ap cluster differentiation. Alternatively, Col may activate Nplp1 via unidentified, low affinity sites, similar to the mechanism by which Ubx regulates some of its embryonic target genes (Stratmann, 2017).

ap encodes a LIM-HD protein, a family of transcription factors well known to control multiple aspects of terminal neuronal subtype fate, including neurotransmitter identity, axon pathfinding and ion channel expression. The current results indicate that Ap in turn acts upon dimm, and subsequently with Dimm on Nplp1. eya encodes an evolutionary well-conserved phosphatase and does not bind DNA directly, instead acting as a transcriptional co-factor. Eya (and its orthologues) have been found to interact with several transcription factors in different systems, but whether it forms complexes with Col and Ap is not known (Stratmann, 2017).

The final transcription factor in the FFL is Dimm, a bHLH protein. Dimm is selectively expressed by the majority of neuropeptide neurons in Drosophila, and is important for expression of many neuropeptides. Intriguingly, Dimm is also both necessary and sufficient to establish the dense-core secretory machinery, found in neuropeptide neurons. Based upon these findings Dimm has been viewed as a cell type selector gene, acting to up-regulate the secretory machinery. This study found evidence for that Dimm acts directly on the Nplp1 enhancer, and this raises the possibility that Dimm is both a selector gene for the dense-core secretory machinery, and can act in some neuropeptide neurons to directly regulate specific neuropeptide gene expression (Stratmann, 2017).

Gene regulatory mechanisms underlying the spatial and temporal regulation of target-dependent gene expression in Drosophila neurons

Neuronal differentiation often requires target-derived signals from the cells they innervate. These signals typically activate neural subtype-specific genes, but the gene regulatory mechanisms remain largely unknown. Highly restricted expression of the FMRFa neuropeptide in Drosophila Tv4 neurons requires target-derived BMP signaling and a transcription factor code that includes Apterous. Using integrase transgenesis of enhancer reporters, this study functionally dissected the Tv4-enhancer of FMRFa within its native cellular context. Two essential but discrete cis-elements were identified, a BMP-response element (BMP-RE) that binds BMP-activated pMad, and a homeodomain-response element (HD-RE) that binds Apterous. These cis-elements have low activity and must be combined for Tv4-enhancer activity. Concatemers of the HD-RE and BMP-RE cis-elements were found to independently generate the same spatiotemporal expression as the Tv4-enhancer. The activation of target-dependent genes is assumed to 'wait' for target contact. This was tested directly, and it was unexpectedly found that premature BMP activity could not induce early FMRFa expression; also, that the BMP-insensitive HD-RE cis-element is activated at the time of target contact. seven up (svp) was uncovered as a repressor of FMRFa induction prior to target contact. Svp is normally downregulated immediately prior to target contact, and maintaining Svp expression prevents cis-element activation, whereas reducing svp gene dosage prematurely activates cis-element activity. It is concluded that the target-dependent FMRFa gene is repressed prior to target contact, and that target-derived BMP signaling directly activates FMRFa gene expression through an atypical gene regulatory mechanism (Berndt, 2015).

Home page: The Interactive Fly © 1995, 1996 Thomas B. Brody, Ph.D.

The Interactive Fly resides on the
Society for Developmental Biology's Web server.