The 3.9 kb transcript first accumulates after 8 hours of embryogenesis, peaks in first instar larvae and is present at low levels in third instar larvae and pupae. Both col transcripts are detected at very low levels in male and female adults and are absent from dissected ovaries (Crozatier, 1996).

collier expression is detected at the beginning of mitotic cycle 14 in two laterally symmetrical stripes whose posterior limit is one or two rows of cells anterior to the postion where the cephalic furrow forms. This restricted expression persists during early gastrulation, when the cephalic furrow demarcates the head from the trunk and the posterior gnathal region (maxillary and labial primordia). During formation of the ventral furrow, the stripe of col expression widens from one to four cells. Comparison with the blastoderm fate map indicates that col expression extensively overlaps the mandibular segment anlage but is slightly displaced anteriorly. The col expression domain corresponds with the 'Engrailed intercalary spot' as well as in cells directly posterior to it. Thus col appears to be specifically expressed in posterior cells of the intercalary segment (parasegment -1) and in anterior cells of the mandibular segment [the first pregnathal (in front of the mouth) segment], a region that possibly corresponds to parasegment 0 (Crozatier, 1996).

By stage 11, col transcripts are still present in the mandibular bud region but additional expression is seen in a few cells of the procephalon and in segmentally repeated groups of cells in the trunk. In stage 13 and 14 embryos, col is expressed in a segmentally reiterated pattern in the ventral nerve cord and in lateral and dorsal groups of cells, including cells of the peripheral nervous system, as well as in patches of cells in the brain (Crozatier, 1996).

It has been hypothesized that the hypopharyngeal lobe is the ventral aspect of the intercalary segment, in which only the dorsal component expresses lab during stages 10 and 11. The anterior expression boundary of most homeotic genes differs in the ventral ectoderm compared to the dorsolateral ectoderm, so the expression of lab in only the dorsolateral intercalary segment would not be unusual. However, kn ectodermal expression, which appears to mark the hypopharyngeal lobe, is separated by hh-expressing cells from the intercalary anterior compartment in which lab is expressed (the short stripe of hh-expressing cells extends from the ventral margin of the ectoderm where it meets the endoderm in the stomatodeal opening to the dorsal limit of kn ectodermal expression). Furthermore, while the ventral portion of the stripe of kn expression gives rise to the hypopharyngeal lobe, the dorsolateral portion does not appear to give rise to any ectodermal derivatives, but instead to cephalic mesoderm. Thus, the hypopharyngeal lobe is not thought to represent the ventral component of the intercalary segment, but rather is a separate domain posterior to it. Because the anterior compartment cells of the hypopharyngeal lobe primordium are not separated by intervening hh- or en-expressing cells from the anterior mandibular segment, the hypopharyngeal lobe should be considered nominally part of the mandibular segment, although it appears to be defined independently at blastoderm stages, as judged by kn activation. This is in keeping with the mandibular defects found in kn mutants that eliminate the hypopharyngeal lobe (Seecoomar, 2000 and references therein).

Ample evidence indicates, however, that in other insects the hypopharyngeal lobe is part of the intercalary segment. Unlike Drosophila, lab is expressed in the hypopharyngeal lobe in the flea and other insects. In all insects, en-expressing cells marking the posterior compartment of the intercalary segment, form on the posterior margin of the lab-expressing cells. Perhaps this difference in lab expression indicates that what is called the 'hypopharyngeal lobes' in Drosophila are not homologous to the hypopharyngeal lobes of other insects, but rather reflect a novel proturberance in the mandibular segment. Alternatively, it is possible that the Drosophila 'hypopharyngeal lobes' are indeed homologous to the hypopharyngeal lobes of other species, but there has been a concerted shift in Drosophila relative to the head morphology of both lab expression and posterior compartment of the intercalary segment, such that the hypopharyngeal lobe no longer resides in the intercalary segment. Because the hypopharyngeal primordium (via activation of kn) appears to be set aside independently at blastoderm stage, much earlier than the segmentation of the cephalic segments, such a shift in segmentation vis-a-vis the hypopharyngeal lobe seems plausible. Further analysis of the expression of the homeotic genes (such as lab, Dfd and cnc) and the upstream patterning genes (such as kn, btd and other gap/pair-rule genes) in a number of closely and distantly related insects will be necessary to resolve these hypotheses (Seecoomar, 2000 and references therein).

Effects of Mutation or Deletion

Hedgehog (Hh) plays an important role in Drosophila wing patterning by inducing expression of Dpp, which serves to organize the wing globally across the A-P axis. Hh signaling also plays a direct role in patterning the medial wing through the activation of the Hh-target gene, knot (kn). kn is expressed in Hh-responsive cells near the A-P compartment boundary, where its expression is dependent on fu, a component of Hh signaling. kn is required for the proper positioning of veins 3 and 4 and to prevent ectopic venation between them. Furthermore, the anterior expansion of the normal kn expression domain causes an associated anterior shift in the position of vein 3 in the resultant wing. Ectopic expression of kn elsewhere in the wing imaginal disc results in the failure to properly activate the vein initiation genes, rho and Dl. Expression of the gene encoding the EGF-receptor (Egfr), which is required for vein initiation and subsequent differentiation, is normally depressed in the 3-4 intervein region. This downregulation of Egfr in the medial portion of the imaginal disc is dependent on kn activity and ectopic expression of kn inactivates Egfr elsewhere in the wing primordium. It is proposed that kn expression in Hh-responsive cells of the wing blade anlagen during the late third instar creates a zone of cells in the medial wing in which vein primordia cannot be induced. The primordia for veins 3 and 4 are laid down adjacent to the kn-imposed vein-free zone, presumably by a signaling factor (such as Vn) also synthesized in the medial region of the wing (Mohler, 2000).

The original viable kn allele causes veins 3 and 4 to form closer together and produces a corresponding shift in their primordia (as detected by rho expression) in the late third instar disc (Sturtevant, 1995). Mosaic analysis of strong, embryonic kn alleles by Nestoras (1997) indicates that kn is important in suppressing vein formation between veins 3 and 4, but not in other regions of the wing. Because vein 4 runs just posterior to the A-P compartment boundary, the region affected by kn mutant clones corresponds approximately to the Hh-responsive cells along the A-P compartment boundary. kn is also required for suppressing vein formation in ptc minus clones in the anterior compartment; these ptc minus clones mimic Hh-responsive cells in which Hh has bound to the Ptc receptor. Nestoras (1997) proposed that kn functions to separate veins 3 and 4 by imposing a vein-free region in response to Hh signaling. Mosaic analysis of strong, pupal lethal alleles of fu, a serine/thronine kinase required in Hh-responsive cells for a normal response, shows a similar requirement for fu in preventing ectopic vein induction between veins 3 and 4, suggesting a direct role for Hh signaling in controlling the 3-4 intervein space (Mohler, 2000 and references therein).

In situ hybridization of the kn cDNA to wing imaginal discs indicates that strong kn expression is limited primarily to a stripe in the middle of the wing blade region of the imaginal disc. Weaker kn expression can also be detected in more posterior regions of the wing pouch and in portions of the hinge region of the disc. Significantly, kn medial expression is reduced in fu mutants, which encode a ser-thr kinase that functions in many aspects of Hh signaling (Mohler, 2000).

Colabelling of KN mRNA expression with a ptc-lacZ reporter construct, reveals that the medial expression of kn is located within the ptc-expressing cells (the Hh-responsive cells). The posterior margin of kn expression matches the posterior margin of ptc-lacZ expression at the anterior-posterior compartment boundary, although kn expression does not extend quite as far anterior as the ptc-lacZ reporter. In contrast, most of the kn expression is posterior to the expression domain of a dpp-lacZ reporter construct (marking a region of strong dpp expression), overlapping this domain only slightly. Thus, kn, like en and ptc is expressed close to the A-P boundary, whereas strong dpp expression is displaced slightly to the anterior, with respect to kn and the A-P boundary (Mohler, 2000).

Expression of the rho gene has been demonstrated to be a determinant for the vein primordia laid down in the late third instar wing disc. Double-labelled in situ hybridizations reveal that kn is expressed precisely between the vein primordia for veins 3 and 4. Thus, KN mRNA expression matches the region determined by mosaic analysis for kn function: between veins 3 and 4 in the Hh-responsive cells of the wing blade (Mohler, 2000).

No alteration of vein expression is found in knot mutant discs or in discs in which kn has been ectopically expressed, indicating that the expression of vn in this region is not controlled by kn. Egfr expression, however, is regulated by knot. Although transcription of Egfr is initially uniform throughout the wing blade, by the end of the third instar Egfr transcription has been repressed along the wing margin and in a medial stripe across the wing blade region. Double in-situ hybridization reveals that the medial region of Egfr downregulation coincides with the region of kn gene expression. To determine whether downregulation of Egfr is necessary for the formation of the 3-4 intervein region, Egfr was ectopically expressed in the region in which it is usually downregulated. Ecoptic expression of EGFR in the medial wing driven by the ptc-GAL4 driver generates wings with fusion of veins 3 and 4 in the proximal portion of the wing. Ectopic expression of Egfr driven throughout the wing blade dorsally does not cause formation of ectopic veins between veins 3 and 4, although a significant amount of ectopic vein material is induced anterior to vein 3 and posterior to vein 4. This suggests that the creation of a vein-free zone between veins 3 and 4 is not likely to be explained solely by downregulation of Egfr by kn (Mohler, 2000).

These results indicate that a second mechanism, in addition to downregulation of the Egf-receptor, must exist to position veins 3 and 4 and prevent vein formation between them, at least in the distal portion of the wing. This secondary mechansim for blocking vein formation between veins 3 and 4 may reflect the fact that Vn is not the only vein-promoting factor synthesized in the medial wing: local Dpp signaling may also play a significant role in the activation of vein 3. Dpp signaling in concert with Hh signaling is required for the activation in the medial wing of the iroquois-complex genes, araucan and caupolican, which in turn are required for the formation of vein 3. The expression of the iro genes is not sufficient to activate vein 3 initiation, since kn overexpression in this region does not affect iro gene expression, but does eliminate expression of the vein initiation genes. The following hypothesis is favored: whereas vein 4 is activated primarily by Vn signaling onto the adjacent Egfr-active domain, vein 3 is combinatorily activated by Vn signaling and by Dpp signaling mediated through iro complex activation. While the Vn and Dpp-inducers may be expressed over a fairly broad domain in the medial wing, because of the block on vein initiation provided by kn activity, vein 3 must form just outside the kn expression domain. The gaps that form in vein 3 when the kn expression domain is expanded to the anterior may be caused by moving the border of the kn expression domain out of range of Vn signaling and either high-level Dpp signaling or the resultant iro expression domain. This process is most likely to produce gaps in the distal portion of vein 3, since the expression of both vn and iro is not as wide in the distal regions of the disc as in the proximal regions (Mohler, 2000 and references therein).

Whereas the segmental nature of the insect head is well established, relatively little is known about the genetic and molecular mechanisms governing this process. The phenotypic analysis is reported of mutations in collier (col), which encodes the Drosophila member of the COE family of HLH transcription factors and is activated at the blastoderm stage in a region overlapping a parasegment (PS0: posterior intercalary and anterior mandibular segments) and a mitotic domain, MD2. col mutant embryos specifically lack intercalary ectodermal structures. col activity is required for intercalary-segment expression both of the segment polarity genes hedgehog, engrailed, and wingless, and of the segment identity gene cap'n'collar. The parasegmental register of col activation is controlled by the combined activities of the head-gap genes buttonhead and empty spiracles and the pair-rule gene even skipped; col therefore integrates inputs from both the head and trunk segmentation systems, which were previously considered to be essentially independent. After gastrulation, positive autoregulation of col is limited to cells of anterior PS0. Conversely, heat-pulse induced ubiquitous expression of Col leads to disruption of the head skeleton. Together, these results indicate that col is required for establishment of the PS(-1)/PS0 parasegmental border and formation of the intercalary segment. These data support neither a simple combinatorial model for segmental patterning of the head nor a direct activation of segment polarity gene expression by head-gap genes, but rather argue for the existence of parasegment-specific second order regulators acting in the head, at a level similar to that of pair-rule genes in the trunk (Crozatier, 1999b).

Three independent alleles of collier (designated as col1, col2 and col3) display lethality at the late embryonic/first instar larval stage with specific defects in the head skeleton similar to those observed upon heat-shock induced expression of col antisense RNA in early gastrula embryos. These defects are a complete lack of the ventral arms (VA) and a strong reduction of the lateral gräten (LG), two structures thought to be derived from the intercalary/mandibular segment anlagen. The T-ribs in the floor of the pharynx and the antennal, maxillary or hypopharyngeal sensory organs are present, as well as a normal pattern of internal sensory structures, detected by 22C10 antibody staining. The mutant embryos that hatch give rise to larvae that do not grow and tend to crawl out of the medium, suggesting that they are unable to feed. That these defects result from col mutations was further substantiated by the lack of a somatic dorsal muscle in which col is specifically expressed. The same head and muscle phenotypes are observed for the 3 alleles in homozygous or hemizygous combinations, suggesting that they are strong hypomorphic or null alleles. Further support was provided by sequence analysis of the col1 allele, where a G to A transition (amino-acid position 228) eliminates a splice acceptor site. This lesion should result in the non-removal of intron 6 and the production of a truncated Col protein; indeed, no Col protein can be detected in col1 mutant embryos using anti-Col polyclonal antibodies directed against the divergent carboxy-terminal end of the protein (Crozatier, 1999b).

The embryonic head phenotype of col1 hemizygous mutant embryos indicates a loss of skeletal structures derived from the intercalary, and possibly mandibular, segments without transformation toward another segment identity. To investigate this segmentation phenotype in more detail, col expression was compared with that of the segment polarity genes hh and wg. At the blastoderm stage, the posterior limit of col expression is parasegmental (PS0/PS1), since it precisely abuts the mandibular stripe of hh-expressing cells. Whether its anterior limit is also parasegmental cannot be answered at this stage because the expression of segment polarity genes in pre-gnathal segments is not yet established at this stage. Examination of early stage 11 embryos shows that col expression overlaps the intercalary hh stripe and abuts the intercalary Wg spot, indicating a parasegmental anterior border for col expression. At this stage however, col expression has been lost from the posterior part of PS0, since it does not overlap mandibular Wg expression. The cnc gene, which codes for a b-ZIP transcription factor, has been postulated to act as a segment identity gene in the mandibular segment. Consistent with col being expressed in PS0, col and cnc expression only partly overlap, in the region corresponding to the anterior mandibular segment. Together, these data indicate a parasegmental register of col expression at the blastoderm stage, which is subsequently restricted to anterior PS0 (Crozatier, 1999b).

A determination was made of whether col mutations affect the expression of wg and En, which mark the anterior and posterior compartments of each segment, respectively. In col1 hemizygous embryos, both the intercalary stripe of En and the spot of wg expression are missing. Since col expression does not overlap the intercalary Wg spot, the loss of this spot in col mutant embryos suggested that col does not regulate wg expression directly but possibly by an hh-dependent mechanism. It has indeed been found that in col mutant embryos, the intercalary stripe of hh is also absent, or much reduced. Together, these results show that col controls hh, en and wg expression in the intercalary segment and is required for establishing the PS(-1)/PS0 parasegmental border. The head skeleton structures ventral arm (VA) and lateral-gräten (LG), which are, respectively, either missing or reduced in col mutant embryos, are also affected in two other head mutants: crocodile (croc), which codes for a forkhead-domain protein, and cnc. These structures are also affected in embryos mutant for the homeotic genes Dfd and lab, which are expressed, respectively, in the mandibular and maxillary segments, and in the intercalary segment. col expression was examined in embryos mutant for croc, cnc, Dfd or lab. In none of these embryos was there a change in col transcription. Conversely, no changes could be detected for croc, Dfd or Lb expression in col1 hemizygous embryos, indicating that expression of each of these three genes is independent of col. In contrast, col is required for cnc transcription in the posterior intercalary segment at stage 9-10. Because this region is anterior to the region of overlap between col and cnc expression at the blastoderm stage, it is concluded that this region corresponds to a secondary site of cnc expression initiated at stage 9, under control of col activity. In cnc mutant embryos, intercalary hh expression is normal, indicating that hh and cnc are regulated by col, independent of one another (Crozatier, 1999b).

col mutations provoke cell death in the forming hypopharynx. The normal pattern of Lab expression and morphology of the intercalary lobe (sometimes referred to as hypopharyngeal lobe) at stage 11 indicate that lack of hh and cnc expression in col1 hemizygous embryos are probably not linked to cell death. In order to address this question more directly, the distribution of reaper (rpr) mRNA, which marks cells fated to undergo apoptosis, was examined. In wild-type embryos at stage 11, there are two invariant sites of rpr expression: inside the epidermal layer of the gnathal region where the primordium of the salivary gland invaginates, and near the caudal tip of the extended germ band. In col1 embryos, an additional site of rpr expression is observed, which does not correspond to cells of the intercalary lobe but rather to cells of the hypopharyngial epithelium (floor of the forming pharynx). These cells have been previously shown to derive from the ventral side of the intercalary segment primordium and lineage tracing analysis has linked them to MD2. This specific cell death is consistent both with the domain of col expression at the onset of gastrulation and the head skeleton phenotype of col mutant larvae (Crozatier, 1999b).

col expression is first detected during the interphase of mitotic cycle 14, when expression of head-gap genes has already resolved from initial broad domains into defined stripes. The stripe of col expression is included in that of btd, overlaps that of ems, and is restricted both dorsally and ventrally to neuroectodermal cells. Examination of dorsal (dl) mutant embryos shows that Dl is required for col repression in the mesodermal plate. The ectopic expression of col observed in twist (twi) and snail (sna) mutant embryos suggests that Dl target genes, rather than Dl itself, are involved. Embryos lacking ems function also show a ventral derepression of col expression. Further, at stage 10, ems mutant embryos show an abnormal pattern of col mRNA accumulation, with a mandibular stripe in addition to intercalary stripe of col-expressing cells. This suggests a second role for ems in regulating col. In btd mutant embryos, there is a complete loss of col expression, whereas there is no change in embryos lacking both slp (slp1 and slp2) genes, consistent with previous data establishing that btd but not slp is required for intercalary en and wg expression (Crozatier, 1999b).

Together, these results confirm that col acts downstream of head gap genes in the transcriptional cascade patterning the head. The fact that col is expressed in a parasegment immediately anterior to the trunk, raises the possibility of a regulation by components of the trunk segmentation system such as pair rule genes. In paired (prd) mutant embryos, col expression starts normally but, soon after gastrulation, becomes restricted to a few cells located anterior to the abnormal cephalic furrow which forms in these embryos. col mis-expression might thus be a rather indirect effect of the prd mutation. In addition to prd, the only pair rule gene whose mutations appear to affect col expression is eve. At the beginning of cycle 14, the stripe of col expression is located immediately anterior to Even-skipped (Eve) stripe 1, these two domains becoming separated by a single row of cells during the process of cellularization. col is ectopically expressed in eve mutant embryos, in a region roughly corresponding to PS1, indicating that Eve acts as a repressor of col in this parasegment. During the interphase of cycle 14, the broad band of ectopic col activation resolves into a distinctive stripe, separated from the normal PS0 stripe by one to three cells going from ventral to dorsal. col expression precisely overlaps the expression of string (stg), in the region prefiguring mitotic domain 2. Expression of stg, which triggers the G2/M transition, is unchanged in embryos deficient for col and vice versa, arguing that MD2 cells undergo a concerted mitotic and differentiation program, set upstream of both col and stg. This led to an examination of whether eve was also involved in defining the position of MD2, using antibodies against the phosphorylated form of histone H3 as a marker of mitosis. Like col transcription, MD2 expands posteriorly in eve mutant embryos at early cycle 14 to form a second, ectopic, stripe of mitotic cells at the beginning of gastrulation. Together, these results show that col expression and MD2 position integrate inputs from both the head and trunk segmentation systems, which were previously considered as being essentially independent (Crozatier, 1999b).

Late control of col expression involves restriction to the anterior PS0 and positive autoregulation. At the onset of gastrulation (stage 7), col is expressed in the entire PS0. At stage 10, the two separate patches of cells that keep expressing col correspond to lateral cells of the intercalary lobe and ventral cells invaginating within the atriopharyngeal cavity, respectively, indicating a restriction of col expression to anterior PS0. To determine when this restriction occurs, use was made of the greater stability of the beta-gal protein compared to lacZ (or col) mRNA and the two patterns were compared in P[col5-lacZ] embryos. In order to follow col transcription rather than transcript accumulation, a col intron probe was used that labels nascent nuclear transcripts. Before mitosis 14, the lacZ and col mRNA and beta-gal protein patterns completely overlap. After completion of mitosis 14, all the ectodermal cells derived from the MD2 domain can be visualized by beta-galactosidase immunostaining. Only the most anterior cells continue to transcribe col-lacZ (or col). This observation suggests that, at cycle 14, an asymmetric cell division occurs with respect to the maintenance of col transcription. After stage 11, col transcription is only maintained in a subset of the cells of the intercalary lobe. In col mutant embryos, this transcription is lost, indicating a direct or indirect, positive auto-regulation (Crozatier, 1999b).

To investigate further the role of col transcriptional regulation in head morphogenesis, the functional consequences of ubiquitous expression of the Col protein were examined, using hs-col transgenic lines. One 45 minute heat shock was applied at different times during embryonic development. Whereas heat-shock treatment of hs-col embryos later than 5 hours after egg laying (AEL) has no apparent effect on embryonic development and viability, heat-shock treatment between 3 and 5 hours AEL (stages 6 to 9) causes embryonic death. Cuticle preparations of dead embryos display a normal segmental array of thoracic and abdominal denticles, but a disrupted head skeleton. The H piece, which originates in part from the maxillary, and partly from the labial segment anlagen, is consistently missing and the LG are also affected. Because col activity is required for maintaining its own expression in the head ectoderm, it was asked whether ubiquitous expression of the Col protein was altering endogenous col expression. In situ hybridization to hs-col embryos with a col intron probe, which allows hs-col and endogenous col transcripts to be distinguished, revealed that col transcription is ectopically activated in mandibular cells. Whether this ectopic col transcription results from activation and/or maintenance in all MD2-derived cells after mitosis 14 remains to be established. No specific change in the expression pattern of En, cnc, Dfd and Lab could be detected, however, suggesting that ectopic activation of other Col targets is responsible for the induced phenotype (Crozatier, 1999b).

This paper summarizes the regulatory cascade controlling formation of the intercalary segment. Activation of col in PS0 at the blastoderm stage (stage 6) is strictly dependent upon btd. Col is ventrally repressed by ems and the mesodermal genes sna and twi; by the trunk pair rule gene eve in the posterior, and by a yet unknown factor in the anterior. At stage 10, col expression has become restricted to anterior PS0, where it activates intercalary expression of hh and cnc, and possibly en. Loss of hh activity leads to the non cell-autonomous loss of wg expression in anterior adjacent cells [PS(-1) cells]. btd activation of eve in PS1 creates a regulatory loop allowing for differential gene expression between PS0 and PS1. cnc expression in the mandibular segment is dependent upon btd (Crozatier, 1999b).

The knot gene of Drosophila is required for the formation of the hypopharyngeal lobe in the germ-band stage embryo and the ventral arms and lateralgräte of the larval head skeleton. Rearrangement mutations of knot disrupt the previously characterized gene, collier. kn is required for the activation of cnc (cap'n'collar, a bZIP homeotic selector gene) in the progenitors of the hypopharyngeal lobe. kn is also required for the activation of CK01299, an EST marker for the cephalic mesoderm. Based on the relative expression of kn and the posterior compartment-specific gene hedgehog, additional evidence is provided that in Drosophila, unlike many other insects, the hypopharyngeal lobe arises from part of the mandibular segment (Seecoomar, 2000).

knot is expressed in early embryogenesis from stage 7 to 11 in a single domain corresponding to mitotic domain 2 within the mandibular segment. Comparison of the expression of kn in the mandibular segment with cnc, a homeotic gene required for labral and mandibular segment identity, in late stage 10 reveals that kn is expressed in the anterior-most region of the mandibular segment, whereas cnc is expressed throughout the anterior compartment of the mandibular region. Unlike cnc, kn is also significantly expressed in the cephalic mesoderm. Injecting antisense COL RNA into embryos prior to blastoderm formation has revealed a role for kn in mandibular development. This results in mandibular defects, particularly the shortening of the lateralgräte of the head skeleton. Embryos mutant for knSA1, the rearrangement allele broken within the transcription unit, or for any of four kn mutations isolated with EMS show similar disruptions of head skeleton. In zygotically mutant kn embryos the observed defects of mandibular structures are somewhat variable. However, for each of the four EMS mutations, kn mutant embryos derived from homozygous germ-line clone mothers had consistently stronger head skeleton defects. For the two strongest EMS alleles germ-line clone-derived zygotic mutant embryos ('zygotic + maternal') consistently lack the ventral arms and have foreshortened lateralgräte, but retain the adjacent ventral plate, ventral T-ribs and posterior pharyngeal wall. The variability in the zygotic mutant embryos is primarily due to variation in the extent of the lateralgräte. Because the difference between the zygotic alone and zygotic + maternal mutant embryos is simply in the extension and shape of the lateralgräte, it is not clear whether the difference is due to loss of maternal kn expression or to other potential differences in genetic background between these two lines. No significant maternal expression can be detected by in situ hybridization to early embryos, although low levels of RNA may not be easily discerned from background staining and the potential for maternal protein stores exists. Due to the possibility of maternal contribution to kn early embryonic function, subsequent molecular analysis of kn function was conducted in homozygous embryos derived from germ-line clones mutant for a strong allele (Seecoomar, 2000).

The development of multicellular organisms requires the establishment of cell populations with different adhesion properties. In Drosophila, a cell-segregation mechanism underlies the maintenance of the anterior (A) and posterior (P) compartments of the wing imaginal disc. Although engrailed (en) activity contributes to the specification of the differential cell affinity between A and P cells, recent evidence suggests that cell sorting depends largely on the transduction of the Hh signal in A cells. The activator form of Cubitus interruptus (Ci), a transcription factor mediating Hh signaling, defines anterior specificity, indicating that Hh-dependent cell sorting requires Hh target gene expression. However, the identity of the gene(s) contributing to distinct A and P cell affinities is unknown. A genetic screen based on the FRT/FLP system has been to search for genes involved in the correct establishment of the anteroposterior compartment boundary. By using double FRT chromosomes in combination with a wing-specific FLP source, 250,000 mutagenized chromosomes were screened. Several complementation groups affecting wing patterning have been isolated, including new alleles of most known Hh-signaling components. Among these, a class of patched (ptc) alleles was identified exhibiting a novel phenotype. These results demonstrate the value of this setup in the identification of genes involved in distinct wing-patterning processes (Végh, 2003).

A total of 250,000 mutant chromosomes covering the X chromosome and both major autosomes were screened. Four complementation groups were identified that affected wing patterning similar to mutations in smo. The largest of these groups represents alleles in smo itself. Two groups exhibiting a subset of smo phenotypes represent new alleles of fused and collier/knot. Fused is a positive regulator of Hh signaling, and collier/knot is an Hh target gene required for the formation of the L3/L4 intervein region. Surprisingly, the remaining complementation group turned out to consist of novel ptc alleles with striking characteristics. Molecularly, they represent point mutations causing an amino acid substitution in either the first or the second large extracellular loop. In contrast to ptc null alleles, homozygous mutant clones failed to upregulate Hh target genes even in the presence of Hh. Together these findings suggest that the mutant proteins repress Smo constitutively, most likely because they fail to bind Hh. Animals mutant for trans-heterozygous combinations of these new ptc alleles with ptcS2 are fully viable. The ptcS2 product lacks the ability to repress Smo but is able to sequester, and hence bind to, Hh. The intragenic complementation that was observed suggests that both functions of Ptc, binding of Hh and repression of Smo, can be provided by individual proteins that possess only one of each. Recently, it was shown that a combination of two proteins, one consisting of the N- and the other the C-terminal half of Ptc, reconstitutes Ptc function. Although these experiments cannot be directly compared with the findings in this study, together they do suggest that Ptc function can be separated intramolecularly into independent modules of N- vs. C-terminal and extra- vs. intracellular domains. One possible scenario that could explain the intragenic complementation would be if Ptc proteins act in a multimeric complex (Végh, 2003).

Cellular immune response to parasitization in Drosophila requires the EBF orthologue Collier

Drosophila immune response involves three types of hemocytes ('blood cells'). One cell type, the lamellocyte, is induced to differentiate only under particular conditions, such as parasitization by wasps. This study investigated the mechanisms underlying the specification of lamellocytes. collier (col), the Drosophila orthologue of the vertebrate gene encoding early B-cell factor (EBF), is expressed very early during ontogeny of the lymph gland, the larval hematopoietic organ. In this organ, Col expression prefigures a specific posterior region recently proposed to act as a signalling centre, the posterior signalling centre (PSC). The complete lack of lamellocytes in parasitized col mutant larvae revealed the critical requirement for Col activity in specification of this cell type. In wild-type larvae, Col expression remains restricted to the PSC following parasitization, despite the massive production of lamellocytes. It is therefore proposed that Col endows PSC cells with the capacity to relay an instructive signal that orients hematopoietic precursors towards the lamellocyte fate in response to parasitization. Considered together with the role of EBF in lymphopoiesis, these findings suggest new parallels in cellular immunity between Drosophila and vertebrates. Further investigations on Col/EBF expression and function in other phyla should provide fresh insight into the evolutionary origin of lymphoid cells (Crozatier, 2004).

Col is expressed in Drosophila lymph glands at the end of embryogenesis. In the absence of a specific molecular marker, the embryonic anlage of lymph glands has been mapped to the thoracic lateral mesoderm by lineage analysis of transplanted cells. By histochemical staining, it was observed that Col is expressed in two discrete clusters of cells in the dorsal mesoderm of thoracic segments T2 and T3, starting at the germ-band extension, when lymph gland hemocyte precursors become specified (stage 11). These clusters of Col-expressing cells grow closer during germ-band retraction before coalescing to form the paired lobes of the lymph glands (early stage 13). Double staining for Col and Odd-skipped, a lymph gland marker expressed from that stage onward, confirmed that Col-expressing cells are lymph gland precursors. Thereafter, only three to five cells located at the posterior tip of each lobe maintain high levels of Col expression, although low levels are still detected in the other cells of the lymph glands and in some pericardial cells. Col expression thus identifies a few cells of the thoracic dorsal mesoderm as the lymph gland primordium and distinguishes a specific posterior region of this hematopoietic organ. The embryonic hematopoietic primordium has been defined as the cephalic domain of Srp expression at the blastoderm stage. Srp is not detected, however, in lymph gland precursors prior to stage 12. Consistent with this result, larval hematopoietic progenitors expressing Col are observed in srp6G (an amorphic allele) mutant embryos, indicating that the specification of the embryonic and larval lymph gland progenitors may involve different processes (Crozatier, 2004).

B- and T-lymphocytes mediate adaptive immunity, a phylogenetically recent component of the immune system as it is found only in gnathostomes. How adaptive immunity emerged during evolution, and was built on top of the innate immune system by which it is controlled and assisted, remains a fascinating question. The requirement for Col function in the Drosophila cellular immune response, and EBF function in B-cell development in vertebrates, suggests that Col/EBF function was co-opted early during the evolution of cellular immunity. A puzzling question remains, however, of how the cell-autonomous function of EBF in B-cell development, and the non-cell-autonomous function of Col in lamellocyte development, could relate to an ancestral Col/EBF function. It is proposed that the ancestral expression of Col/EBF in a subset of hematopoietic cells conferred on these cells the ability to respond to signals from circulating immune supervisors (generically designated as macrophages) and provide a secondary line of defence against specific intruders. This cell-specific property in turn laid the ground for the emergence of the vertebrate lymphoid cells on one side and the Drosophila PSC on the other. Although admittedly highly speculative, this proposal takes into account the following considerations. B-cell development represents the default fate of lymphoid progenitors. Although specification of B-cells critically depends on EBF (and the basic helix-loop-helix protein E2A), commitment depends on another gene, Pax5. The Pax5-/- pro-B-cells retain the ability to generate a whole range of both 'innate' myeloid and lymphoid cells. Thus, the ontogeny of the B-cell lineage from preexisting myeloid cell types has occurred through several steps, one key event being the co-opting of Pax5, acting downstream of EBF, for which there is no known counterpart in Drosophila hematopoiesis. Second, the co-opting of Col activity for lamellocyte differentiation in larval hematopoiesis most likely came on top of a preexisting hematopoietic system, such as that operating in Drosophila embryos. Further investigation of Col/EBF functions in intermediate phyla should provide more insight into the diversity of myeloid lineages and ontogeny of the lymphoid lineages during evolution (Crozatier, 2004).

Control of multidendritic neuron differentiation in Drosophila: the role of Collier

Proper sampling of sensory inputs critically depends upon neuron morphogenesis and expression of sensory channels. The highly stereotyped organisation of the Drosophila peripheral nervous system (PNS) provides a model to study neuronal determination and morphogenesis. This study reports that Collier/Knot (Col/Kn), the Drosophila member of the COE family of transcription factors, is transiently expressed in the subset of multidendritic arborisation (da) sensory neurons that display an highly branched dendritic arborisation, class IV neurons. When lacking Col activity, class IV da neurons are formed but display a reduced dendrite arborisation. Col control on dendrite branching is distinct from that exerted by Cut, another transcription factor expressed in class IV neurons and necessary for proper dendrite morphogenesis. Col is also required for the class IV da-specific expression of pickpocket (ppk), which encodes a degenerin/epithelial sodium channel subunit required for larval locomotion. Characterisation of the col upstream region identified a 9-kb cis-regulatory region driving col expression in all class IV md neurons, even though these originate from two types of sensory precursor cells. Altogether, these findings indicate that col is required in at least two distinct programs that control the morphological and sensory specificity of Drosophila md neurons (Crozatier, 2008).

Each individual cell of the Drosophila embryonic PNS has been described, providing a unique model system to investigate the mechanisms of coding of neural identity. This study shows that expression of the COE transcription factor Collier/Knot is specific to a subset of md neurons, the class IV neurons. Col is required both for the specific expression of ppk, a gene encoding a subunit of a sodium-gated channel involved in locomotion and aspects of the elaborate dendritic branching pattern of class IV md neurons. It should be noted that an independent study of col function in md neurons came to very similar conclusions (Hattori, 2007: Crozatier, 2008).

Systematic screens for transcription factors required for proper morphogenesis of Drosophila sensory neuron dendrites have revealed that a number of distinct transcription factors are involved in different aspects of dendrite extension, lateral branching and arborisation as well as in restriction of the dendritic coverage. However, to date the expression pattern of very few of these transcription factors is known during embryonic and post embryonic development. Col is the only known transcription factor whose expression is restricted to a subtype of md neurons. One particularly intriguing feature is that the three class IV md neurons found per abdominal hemisegment originate from at least two different types of lineages, the md-es (v'ada) or md-solo lineages (ddaC and vdaB), pointing to the existence of a convergence process towards the same neuronal phenotype. Since col remains transcribed in these three neurons in embryos mutant for col, it suggests that they are specified independently of col activity. The regulation of Col expression and the function in class IV md neurons provide a unique paradigm to study how the transcriptional control of md specification operates in independent lineages (Crozatier, 2008).

col transcription is first observed in the pII cells at the origin of class IV md neurons, indicating that it becomes restricted to these neurons following asymmetric division of the pIIb cell. Two or three Col-positive cells in place of each dorsal md neuron were often detected in embryos mutant for sanpodo, which encodes a four-pass transmembrane protein which interacts with the Notch receptor. Conversely, md-specific col transcription is often lost in embryos mutant for numb (numb1, a null allele), which down regulates Notch signalling in one of the pIIb daughter cells. It is therefore concluded that col transcription is repressed by Notch signalling. Previous studies on col requirement for the formation of the DA3 embryonic muscle showed that col transcription in the sibling DO5 muscle lineage is also repressed by Notch. It thus appears that col repression by Notch is used in several independent cell lineage decisions. Whether the same effectors of the Notch pathway are involved both in the md neuronal and DA3 muscle lineages remains to be determined. One intriguing feature of col transcription in post-mitotic md neurons is that it is only transient. This could be due to a direct control by maternal and/or early zygotic transcription activators that dilute away during embryogenesis. Alternatively, it could involve the progressive accumulation in md neurons of a repressor able to shut down col transcription. Preliminary dissection of the Drosophila col upstream region mapped the cis-regulatory information necessary for col transcription in md neurons to a fragment located between − 9 and − 5 kbp upstream of the transcription start site (Dubois, 2007). Expression of the reporter gene was not detected, however, in the neurons innervating the dh2 and vp5 es organs, indicating that Col expression in md neurons and es neurons is under the control of separate cis-elements. Whether (partly) different cis-elements are also involved in md-solo, versus md-es col activation remains an open question. An identical pattern of Col expression in Class IV md neurons is observed in Drosophila virilis, indicating that the transcriptional regulatory network controlling this expression has been conserved between these two Drosophila species. The − 9- to − 5-kbp fragment of col upstream DNA contains many sequences, between 20 and 80 bp in length, which are identical between D. melanogaster and D. virilis and represent as many potential regulatory sequences. Further dissection of this fragment should allow identifying which combinations of transcription factors are involved in the specific activation and temporal restriction of col transcription in class IV md neurons (Crozatier, 2008).

Activation of ppk transcription in class IV md neurons requires Col activity. Unlike Col, ppk remains expressed in these neurons until the end of larval development, showing that once activated, ppk expression is maintained independently of Col. This suggests the occurrence of a relay mechanism. The several hours delay between Col accumulation and ppk activation, in both normal embryos and upon pan-neuronal expression of Col, also indicates that the ability of Col to activate ppk depends upon another md-specific factor. Whether this factor(s) is itself a target of Col only activated in md neurons or an md-specific factor whose activity is potentiated by Col remains to be investigated. Finally, whether and how the delay and maintenance mechanisms are coupled is also unknown, at present. The transient character of Col expression does not favour a feed-forward mechanism such as that proposed for the specification of two lineage-related neurons in the CNS, where Col and its downstream targets act together to activate lineage-specific neuropeptides. The cis-regulatory region driving ppk expression in class IV md neurons has been identified. Parallel studies on col and ppk cis-regulation should now allow deciphering more precisely the transcriptional control of class IV md neuron differentiation (Crozatier, 2008).

In addition to ppk expression, class IV md neurons differ from the other md neurons by the length and degree of arborisation of their dendrite network. This dendritic arborisation is unchanged in ppk mutant larvae, showing that ppk expression and the dendritic network of class IV md neurons are specified by two independent programs. Formation of the primary branches of the md dendrite network starts to form at the end of embryogenesis and continues to elongate during larval development. At the same time, a more elaborate pattern of secondary arbors develops. The recent identification of no less than 76 transcription factors influencing (class I, in that case) dendrite formation suggests that the formation and maintenance of the dendritic network is regulated at many different, likely successive levels. One of these factors is Cut, which regulates distinct dendrite branching patterns in different md classes in a dose-dependent manner. It is therefore proposed that Col plays a dual function in implementing the class IV md neuron identity. According to this model, at least one unidentified class IV md neuron-specific TF is required for activating Col expression and regulating the level of Cut (Cutmedium) expression. On one side, Col cooperates with other md-specific TFs to activate ppk transcription specifically in class IV md neurons, independent of Cut. On the other, Col and Cutmedium are involved in establishing the secondary complex dendrite network, typical of these neurons that develops in larvae. Since col transcription is not maintained beyond embryogenesis, its role in secondary dendrite branching and maintaining ppk transcription in larvae must be indirect and likely involves (an) intermediate TF(s). Systematic identification of col and cut targets in embryonic class IV md neurons should allow to better understand their respective roles. col involvement in two parallel regulatory networks (or dual function) links two salient features of class IV md neurons, an extended dendritic field and ppk expression. Dendrites act as information-integrating centers as they receive sensory or synaptic inputs. How expression of the Na+ channel subunit Ppk and extended dendrite arborisation are physiologically linked is a next question (Crozatier, 2008).

Combinatorial coding of Drosophila muscle shape by Collier and Nautilus

The diversity of Drosophila muscles correlates with the expression of combinations of identity transcription factors (iTFs) in muscle progenitors (see Restricted gene expression patterns in somatic muscles). This study addresses the question of when and how a combinatorial code is translated into muscle specific properties, by studying the roles of the Collier and Nautilus iTFs that are expressed in partly overlapping subsets of muscle progenitors. The three dorso-lateral (DL) progenitors which express Nautilus and Collier are specified in a fixed temporal sequence, and each expresses additionally other, distinct iTFs. Removal of Collier leads to changes in expression of some of these iTFs and mis-orientation of several DL muscles, including the dorsal acute DA3 muscle which adopts a DA2 morphology. Detailed analysis of this transformation revealed the existence of two steps in the attachment of elongating muscles to specific tendon cells: transient attachment to alternate tendon cells, followed by a resolution step selecting the final sites. The multiple cases of triangular-shaped muscles observed in col mutant embryos indicate that transient binding of elongating muscle to exploratory sites could be a general feature of the developing musculature. In nau mutants, the DA3 muscle randomly adopts the attachment sites of the DA3 or DO5 muscles that derive from the same progenitor, resulting in a DA3, DO5-like or bifid DA3-DO5 orientation. In addition, nau mutant embryos display thinner muscle fibres. Together, these data show that the sequence of expression and combinatorial activities of Col and Nau control the pattern and morphology of DL muscles (Enriquez, 2012).

The larval Drosophila somatic musculature is made of a stereotyped set of about 30 uniquely identifiable muscles per hemisegment. This study shows that the combinatorial activities of Col and Nau are required to establish the pattern of DL muscles and confer upon these muscles their distinctive shapes and epidermal attachment sites (Enriquez, 2012).

Col is expressed in a promuscular cluster and the three derived PCs at the origin of DL muscles. Each of these PCs is specified at a stereotypic position and according to a precise temporal sequence, with the dorsal DA3/DO5 PC being born first and the ventral LL1/DO4 PC being born last. In addition to Col and Nau, each expresses a combination of specific iTFs, including Kr, Poxm and S59. Expression of a specific set of iTFs in each DL PC could thus integrate both positional and temporal cues. The textbook view is that, similar to neuroblast selection in the neuroectoderm, each muscle PC is selected via the process of lateral inhibition from an equivalence group of mesodermal cells. The parallel between neuroblast and PC selection is supported by the co-expression of the proneural gene l(1)sc and iTFs such as Eve or S59 in specific promuscular clusters. However, a deficiency of l(1)sc results in only minor defects of somatic muscle development and does not prevent the selection of the DA3/DO5 and DT1/DO3 progenitors. A possibility is that several PCs could be selected from large competence domains defined by expression of specific iTFs and l(1)sc clusters play only a limited or redundant role. Selection of the three DL PCs from a cluster of Col-expressing cells supports this view. The fact that the DA3/DO5 and LL1/DO4 PCs are born sequentially and positioned adjacent to one another, suggests a reiterative selection process (Enriquez, 2012).

The most obvious muscle pattern defects that are observed in col mutant embryos, are DA3 > DA2 and DA3 > LL1 transformations. Since the DA2, DA3 and LL1 muscles are derived from different PCs, these defects indicate changes in progenitor identity. In one case, Poxm and S59 expression in the DT1/DO3 progenitor requires Col activity. In contrast, Kr expression in the LL1/DO4 PC is independent of Col. Interestingly, Kr and S59 are expressed together in the ventral VA1/VA2 PC but, in this case, Kr regulates S59 expression. Together, these expression data strengthen the concept of combinatorial coding of muscle identity at the PC stage and show that hierarchies of interactions between different iTFs are progenitor-specific (Enriquez, 2012).

In Poxm mutants, the DO3 muscle is often duplicated, likely at the expense of a DT1 muscle that is often missing. In S59 mutant embryos, the DO3 and DT1 sibling muscles share ventral attachment sites and form a single syncitium in a fraction of segments. Since Col acts upstream of Poxm and S59 in the DT1/DO3 progenitor, it was expected the col mutant phenotype to overlap with the poxm and S59 phenotypes. It may not be so simple, however, since the DT1 muscle is absent only in a small fraction of segments in col mutant embryos. Interestingly, while the DA3 muscle is transformed into a DA2-like muscle in absence of Col activity, the LL1 muscle can adopt a DA3 morphology The LL1 muscle is mis-oriented in col as well as in Kr mutant embryos. Together, these muscle re-orientation phenotypes suggest that there is a range of possible attachment sites for each elongating DL muscle and that the final pattern results from a global combinatorial control. The propensity of elongating muscles to explore several attachment sites, could explain why a coordinate, global regulation by combinations of iTFs is essential. The term regulatory state has been used to describe the total set of active transcription factors in a given cell at a given time. In essence, each PC iTF code is an example of a regulatory state. The loss of one iTF reveals an alternative regulatory state and PC identity, suggesting that a given iTF is able to exert its activity only in the presence of other specific iTFs. A global analysis of this mutual dependency now requires the identification of all DL iTFs, including those expressed in the DO3, DO4 or DO5 muscles (Enriquez, 2012).

Nau differs from other well characterised iTFs, in that it is expressed in most, if not all FCs, before being restricted to specific muscle precursors. SEM analysis shows that most muscles are much thinner in nau mutant than wt embryos. Detailed examination of the mutant DA3 muscle showed that, despite being thinner, it contained a number of nuclei close to normal. nau activity is thus required for embryonic muscle fibre size, but not the muscle fusion programme, per se. Whether Nau directly or indirectly regulates the synthesis and/or assembly of myofibril proteins remains to be determined. DL muscles, including the DA3 muscle, are more severely affected. Taken together, these data lead to the conclusion that Nau performs both general myogenic functions and specific functions in selected muscle lineages. A different threshold level of MRF activity might be needed to initiate myogenesis in different trunk and craniofacial muscles. The different Nau functions in establishing the Drosophila muscle pattern suggest that Nau activity is, in part conditioned by interactions with other iTFs such as Col (Enriquez, 2012).

Co-expression of Nau and Col in the DA3/DO5 progenitor provides a good model to challenge the concept of combinatorial control of muscle identity. While transformed towards a DA2 muscle in absence of Col activity, the DA3 muscle adopts the morphology of its sibling, DO5 muscle in absence of Nau. Thus, while co-expressed in the DA3/DO5 PC, Col and Nau act at different steps in the DA3 lineage. The following regulatory cascade is proposed: Col expression in a large cluster of myoblasts and the three derived PCs, under control of an early CRM and Hox activity, defines a domain of competence for DL muscle development. Col activity, either upstream, and/or in parallel to other iTFs, contributes to confer each DL progenitor its particular identity. The restricted ability of Col in maintaining its own expression in the DA3 FC, by direct binding to a late, DA3-specific CRM, reveals a context-dependence provided by the iTF combination specific to the DA3/DO5 PC. This PC-specific handover process may explain why the DA3 muscle is the most frequently affected in col mutant embryos. Asymmetric division of each DL PC generates two FCs with different regulatory states. Whereas two DA3 and two DO5 muscles form in Notch (N) loss- and gain of function conditions, respectively, Nau confers robustness to the DA3 versus DO5 differentiation programme. This Nau function involves positive regulation of col transcription in the DA3 syncytium nuclei and is independent of Nau function in ensuring normal fibre size (Enriquez, 2012).

In conclusion, these data show that the sequence of expression and combinatorial activities of Col and Nau are required to establish the pattern of DL muscles and confer upon the DA3 muscle its distinctive size and epidermal attachment sites. Identification of the gene targets of this combination is now essential to link a sequence of regulatory states to the architecture of a specific Drosophila muscle. Interestingly, a recent report suggests that EBF cooperates with MyoD in driving aspects of differentiation in Xenopus muscle cells, suggesting that there may be an ancient, evolutionarily conserved, transcriptional relationship between the COE/EBF and MyoD gene families (Enriquez, 2012).

Embryonic muscles connect to the chitinous exoskeleton of the developing embryo via tendon cells, which are specialised epidermal cells. Proper attachment of muscles requires the specific targeting of tendon cells at segmental or intra-segmental, stereotypic positions. The general view is that growing myotubes extend filopodia at their two ends, in search of attachment sites, and that muscle extension ceases when muscles have reached their targeted tendon cells. Some muscle guidance components have been described, such as the Derailed receptor tyrosine kinase for the lateral transverse muscles and the Robo and Robo2 receptors, the transmembrane protein Kon-Tiki and its associated intracellular signalling protein dGrip for ventral-longitudinal muscles. How the precise matching of specific muscles to specific tendon cells is achieved, however, is far from being understood. SEM analysis and phalloidin staining of col mutant embryos showed many mis-oriented muscles, suggesting targeting defects. Many fibres showed more than two attachment sites to the epidermis, however, a phenotype difficult to reconcile with a bipolar extension of muscle precursors until they connect to the epidermis. Rather, the observation that the wt DA3 muscle is transiently attached to three sites, before acquiring its fully extended bipolar morphology, indicates the existence of an exploratory step, followed by a resolution step that selects the final attachments sites. The allelic series of col phenotypes, which revealed many triangular shape fibres, indicates a defect in the resolution process, without ruling out that ventral elongation of the DA3 myofibre is also defective. Terminal differentiation of tendon cells is dependent upon their interaction with muscles and tendon cells could play a role in the resolution step. Triangular shape LO1 muscles were previously observed in mutants for dgit, which encodes a GTPase activator protein that is involved in myotube guidance. Based on the dgit phenotype, and the current observations, it is proposed that the migratory path of muscles towards their targeted tendon cells can involve exploratory attachment to tendon cells along this path. Deciphering how the final, stereotyped, pattern is controlled now requires the identification of how various iTF combinations differentially regulate guidance cues (Enriquez, 2012).

Developmental transcriptional networks are required to maintain neuronal subtype identity in the mature nervous system

During neurogenesis, transcription factors combinatorially specify neuronal fates and then differentiate subtype identities by inducing subtype-specific gene expression profiles. But how is neuronal subtype identity maintained in mature neurons? Modeling this question in two Drosophila neuronal subtypes (Tv1 and Tv4), tests were performed to see whether the subtype transcription factor networks that direct differentiation during development are required persistently for long-term maintenance of subtype identity. By conditional transcription factor knockdown in adult Tv neurons after normal development, it was found that most transcription factors within the Tv1/Tv4 subtype transcription networks are indeed required to maintain Tv1/Tv4 subtype-specific gene expression in adults. Thus, gene expression profiles are not simply 'locked-in,' but must be actively maintained by persistent developmental transcription factor networks. The cross-regulatory relationships were examined between all transcription factors that persisted in adult Tv1/Tv4 neurons. Certain critical cross-regulatory relationships that had existed between these transcription factors during development are no longer present in the mature adult neuron. This points to key differences between developmental and maintenance transcriptional regulatory networks in individual neurons. Together, these results provide novel insight showing that the maintenance of subtype identity is an active process underpinned by persistently active, combinatorially-acting, developmental transcription factors. These findings have implications for understanding the maintenance of all long-lived cell types and the functional degeneration of neurons in the aging brain (Eade, 2012).

The data provide novel insight supporting the view of Blau and Baltimore (1991) that cellular differentiation is a persistent process that requires active maintenance, rather than being passively 'locked-in' or unalterable. Two primary findings are made in this study regarding the long-term maintenance of neuronal identity. First, all known developmental transcription factors acting in postmitotic Tv1 and Tv4 neurons to initiate the expression of subtype terminal differentiation genes are then persistently required to maintain their expression. Second, it was found that key developmental cross-regulatory relationships that initiated the expression of certain transcription factors were no longer required for their maintained expression in adults. Notably, this was found to be the case even between transcription factors whose expression persists in adults (Eade, 2012).

In this study, all transcription factors implicated in the initiation of subtype-specific neuropeptide expression in Tv1 and Tv4 neurons were found to maintain subtype terminal differentiation gene expression in adults (see Summary of changes in subtype transcription network configuration between initiation and maintenance of subtype identity). In Tv1, col, eya, ap and dimm are required for Nplp1 initiation during development. In this study, knockdown of each transcription factor in adult Tv1 neurons was shown to dramatically downregulate Nplp1. In Tv4 neurons, FMRFa initiation during development requires eya, ap, sqz, dac, dimm and retrograde BMP signaling. Together with previous work showing that BMP signaling maintains FMRFa expression in adults (Eade, 2009), this study now demonstrates that all six regulatory inputs are required for FMRFa maintenance. Most transcription factors, except for dac, also retained their relative regulatory input for FMRFa and Nplp1 expression. In addition, individual transcription factors also retained their developmental subroutines. For example, as found during development, dimm was required in adults to maintain PHM (independently of other regulators) and FMRFa/Nplp1 expression (combinatorially with other regulators) (Eade, 2012).

The few genetic studies that test a persistent role for developmental transcription factors support their role in initiating and maintaining terminal differentiation gene expression. In C. elegans, where just one or two transcription factors initiate most neuronal subtype-specific terminal differentiation genes, they then also appear to maintain their target terminal differentiation genes. In ASE and dopaminergic neurons respectively, CHE-1 and AST-1 initiate and maintain expression of pertinent subtype-specific terminal differentiation genes. In vertebrate neurons, where there is increased complexity in the combinatorial activity of transcription factors in subtype-specific gene expression, certain transcription factors have been demonstrated to be required for maintenance of subtype identity. These are Hand2 that initiates and maintains tyrosine hydroxylase and dopa ß-hydroxylase expression in mouse sympathetic neurons, Pet-1, Gata3 and Lmx1b for serotonergic marker expression in mouse serotonergic neurons, and Nurr1 for dopaminergic marker expression in murine dopaminergic neurons (Eade, 2012).

However, while these studies confirm a role for certain developmental transcription factors in subtype maintenance, it had remained unclear whether the elaborate developmental subtype transcription networks, that mediate neuronal differentiation in Drosophila and vertebrates, are retained in their entirety for maintenance, or whether they become greatly simplified. This analysis of all known subtype transcription network factors in Tv1 and Tv4 neurons now indicates that the majority of a developmental subtype transcription network is indeed retained and required for maintenance. Why would an entire network of transcription factors be required to maintain subtype-specific gene expression? The combinatorial nature of subtype-specific gene expression entails cooperative transcription factor binding at clustered cognate DNA sequences and/or synergism in their activation of transcription. In such cases, the data would indicate that this is not dispensed with for maintaining terminal differentiation gene expression in mature neurons (Eade, 2012).

How the transcription factors of the subtype transcription networks are maintained is less well understood. An elegant model has emerged from studies in C. elegans, wherein transcription factors stably auto-maintain their own expression and can then maintain the expression of subtype terminal differentiation genes. The transcription factor CHE-1 is a key transcription factor that initiates and maintains subtype identity in ASE neurons. CHE-1 binds to a cognate DNA sequence motif (the ASE motif) in most terminal differentiation genes expressed in ASE neurons, as well as in its own cis-regulatory region. Notably, a promoter fusion of the che-1 transcription factor failed to express in che-1 mutants, indicative of CHE-1 autoregulation, and for the cooperatively-acting TTX-3 and CEH-10 transcription factors in AIY neurons. Thus, subtype maintenance in C. elegans is anchored by auto-maintenance of the transcription factors that initiate and maintain terminal differentiation gene expression (Eade, 2012).

In contrast, all available evidence in Tv1 and Tv4 neurons fails to support such an autoregulatory mechanism. An ap reporter (apC-t-lacZ) is expressed normally in ap mutants, and in this study apdsRNAi was not found to alter apGAL4 reporter activity. Moreover, col transcription was unaffected in col mutants that express a non-functional Col protein. This leaves unresolved the question of how the majority of the transcription factors are stably maintained. For transcription factors that are initiated by transiently expressed inputs, a shift to distinct maintenance mechanisms have been invoked and in certain cases shown. In this study, this was found for the loss of cas expression in Tv1 (required for col initiation) and the loss of cas, col and grh in Tv4 (required for eya, ap, dimm, sqz, dac initiation). However, it was surprising to find that the cross-regulatory relationships between persistently-expressed transcription factors were also significantly altered in adults. Notably, eya initiated but did not maintain dimm in Tv4. In Tv1, col initiated but did not maintain eya, ap or dimm. This was particularly unexpected as eya remained critical for FMRFa maintenance and col remained critical for Nplp1 maintenance. Indeed, although tests were performed for cross-regulatory interactions between all transcription factors in both the Tv1 and Tv4 subtype transcription networks, only Dimm was found to remain dependent upon its developmental input; Eya and Ap in Tv1 as well as Ap in Tv4. However, even in this case, the regulation of Dimm was changed; it no longer required eya in Tv4, and in Tv1 it no longer required col, in spite of the fact that both col and eya are retained in these neurons. It is anticipated that such changes in transcription factor cross-regulatory relationships will be found in other Drosophila and vertebrate neurons, which exhibit high complexity in their subtype transcription networks. Indeed, recent evidence has found that in murine serotonergic neurons, the initiation of Pet-1 requires Lmx-1b, but ablation of Lmx-1b in adults did not perturb the maintenance of Pet-1 expression (Eade, 2012).

The potential role of autoregulation for the other factors in the Tv1/Tv4 subtype transcription networks is being pursued. However, there are three additional, potentially overlapping, models for subtype transcription network maintenance. First, regulators may act increasingly redundantly upon one another. Second, unknown regulators may become increasingly sufficient for transcription factor maintenance. Third, transcription factors may be maintained by dedicated maintenance mechanisms, as has been shown for the role of trithorax group genes in the maintenance of Hox genes and Engrailed. Moreover, chromatin modification is undoubtedly involved and likely required to maintain high-level transcription of Tv transcription factors as well as FMRFa, Nplp1 and PHM. However, the extent to which these are instructive as opposed to permissive has yet to be established. In this light, it is intriguing that MYST-HAT complexes, in addition to the subtype transcription factors Che-1 and Die-1, are required for maintenance of ASE-Left subtype identity in C. elegans (Eade, 2012).

Taken together, these studies have identified two apparent types of maintenance mechanism that are operational in adult neurons. On one hand, there are sets of genes that are maintained by their initiating set of transcription factors. These include the terminal differentiation genes and the transcription factor dimm. On the other, most transcription factors appear to no longer require regulatory input from their initiating transcription factor(s). Further work will be required to better understand whether these differences represent truly distinct modes of gene maintenance or reflect the existence of yet unidentified regulatory inputs onto these transcription factors. One issue to consider here is that the expression of certain terminal differentiation genes in neurons, but perhaps not subtype transcription factors, can be plastic throughout life, with changes commonly occurring in response to a developmental switch or physiological stimulus. Thus, terminal differentiation genes may retain complex transcriptional control in order to remain responsive to change. It is notable, however, that FMRFa, Nplp1 and PHM appear to be stably expressed at high levels in Tv1/4 neurons, and no conditions were found that alter their expression throughout life. Thus, these are considered to be stable terminal differentiation genes akin to serotonergic or dopaminergic markers in their respective neurons that define those cells' functional identity and, where tested, are actively maintained by their developmental inputs. Tv1/4 neurons undoubtedly express a battery of terminal differentiation genes, and sets of unknown transcription factors are likely required for their subtype-specific expression. Subtype transcription networks are considered to encompass all regulators required for differentiating the expression of all subtype-specific terminal differentiation genes. Further, differentiation of subtype identity is viewed as the completion of a multitude of distinct gene regulatory events in which each gene is regulated by a subset of the overall subtype transcription network. As highly restricted terminal differentiation genes expressed in Tv1 and Tv4 neurons, it is believed that Nplp1, FMRFa and PHM provide a suitable model for the maintenance of overall identity, with the understanding that other unknown terminal differentiation genes expressed in Tv1 and Tv4 may not be perturbed by knockdown of the transcription factors tested in this study. In the future, it will be important to incorporate a more comprehensive list of regulators and terminal differentiation genes for each neuronal subtype. However, it is believed that the principles uncovered in this study for FMRFa, Nplp1 and PHM maintenance will hold for other terminal differentiation genes (Eade, 2012).

Finally, it is proposed that the active mechanisms utilized for maintenance of subtype differentiation represent an Achilles heel that renders long-lived neurons susceptible to degenerative disorders. Nurr1 ablation in adult mDA neurons reduced dopaminergic markers and promoted cell death. Notably, Nurr1 mutation is associated with Parkinson's disease, and its downregulation is observed in Parkinson's disease mDA neurons. Adult mDA are also susceptible to degeneration in foxa2 heterozygotes, another regulator of mDA neuron differentiation that is maintained in adult mDA neurons. Studies in other long-lived cell types draw similar conclusions. Adult conditional knockout of Pdx1 reduced insulin and ß-cell mass and, importantly, heterozygosity for Pdx1 leads to a rare monogenic form of non-immune diabetes, MODY4. Similarly, NeuroD1 haploinsufficiency is linked to MODY6 and adult ablation of NeuroD in β-islet cells results in β-cell dysfunction and diabetes. These data, together with current results, underscore the need to further explore the transcriptional networks that actively maintain subtype identity, and hence the function, of adult and aging cells (Eade, 2012).

knot/collier: Biological Overview | Evolutionary Homologs | Regulation | Developmental Biology | References

Home page: The Interactive Fly © 1997 Thomas B. Brody, Ph.D.

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