Transcripts of the bowel gene initially accumulate at the cellular blastoderm stage in three domains that are more terminally located than the odd and sob stripes; these include a strong cap at the posterior pole (from 0 to 11% EL), a relatively weak and nonuniform domain at the anterior pole (84%-100% EL), and a broad transverse stripe (approximately 6 cells wide) that lies just anterior to the presumptive cephalic furrow (extending from 76% EL dorsally to 67% EL ventrally). While some cells in these regions may also express odd and sob (e.g., the anterior cap of odd partially overlaps the anterior bowl domain, as do the anteriormost secondary stripes of odd/sob and the bowl cephalic stripe), the overall patterns have little in common. Therefore, the initial regulation of bowl is clearly distinct from that of sob and odd. However, at early gastrulation, the terminal expression of bowl is supplemented by transverse stripes in the trunk primordium. As with odd and sob, these initially appear in a pair-rule pattern of seven stripes, but rapidly evolve to a segment polarity-like pattern as the number of stripes doubles. The intensity of these stripes is very low compared to the terminal domains and the single cephalic stripe, suggesting that bowl may retain only vestigial regulatory elements in common with odd and/or sob (Hart, 1996).
The embryonic expression patterns of the Drosophila odd-family genes can be placed into two classes, one exemplified by drm, and the other by bowl. The three family members that map closest to one another, drm, sob, and odd, display striking similarities in patterns (but not levels) of mRNA expression. Expression of these genes begins in seven pair-rule-like stripes at the cellular blastoderm stage; these then resolve into 14 stripes (Green, 2002 and Hart, 1996). drm, sob, and odd are also expressed in the invaginating stomodeum and proctodeum at their respective junctions with the midgut primordia. By stage 13, drm, sob, and odd are expressed in the keyhole region that will form the proventriculus, and in a wide ring at the most posterior portion of the midgut (Green, 2002 and Hart, 1996). Their expression continues in these regions through the remainder of embryogenesis (Johansen, 2003b and references therein).
bowl expression is overall quite different from that of drm, sob, and odd; in contrast to the spatially localized expression of drm, sob, and odd, the expression of bowl is relatively uniform throughout the anlagen, primordia, and epithelia of the foregut and hindgut (Wang, 1996). By stage 13, expression of bowl in the hindgut is greatly reduced, but is maintained strongly throughout the foregut, where it remains high until stage 17 (Johansen, 2003b).
odd expression during leg development has previously been defined by an embryonic lethal enhancer trap insertion in the odd locus, oddrK111 (Rauskolb, 1999). To confirm that this expression faithfully represents the endogenous gene expression pattern, in situ hybridization to leg discs using an odd probe was performed. Moreover, since the odd-skipped family genes are segmentally expressed during embryogenesis, attempts were made to determine their expression during leg development and how this expression compares with that of odd. During development, the leg disc ultimately gives rise to a cylindrical adult leg, and thus the expression of genes in a segmental pattern is observed as a reiterated pattern of concentric rings in the leg imaginal disc. All four odd-skipped family genes are expressed during leg development in segmentally repeated patterns; the expression profiles of odd, sob, and drm are identical, while that of bowl is distinct (Hao, 2003).
odd, sob, and drm are expressed in a narrow band of cells at the distal edge of each presumptive leg segment, except in tarsal segments 1-4. Robust expression of drm is consistently observed, while both odd and sob are expressed at lower levels. Endogenous expression of odd is identical to that previously observed for oddrK111. Comparison of drm and sob expression to that of oddrK111-driven ß-galactosidase expression has confirmed that all three genes are expressed in the same cells. Expression of drm was further characterized at different stages of leg development. In early third instar discs, expression is first observed in a proximal ring, corresponding to the presumptive coxa. As development proceeds, progressively more distal rings of drm expression are added, until the full complement of five segmental rings is observed by the end of larval development. This progressive increase in the number of rings of drm expression is similar to the previously observed (Rauskolb, 2001) progressive increase in Notch ligand expression in the leg (Hao, 2003).
Notch activation occurs in a narrow stripe of cells within each presumptive leg segment, just distal to cells expressing the ligands Serrate and Delta. The expression of oddrK111 (and thus presumably also odd, sob, and drm) was compared with that of Serrate and Delta to determine whether odd-driven ß-galactosidase expression is observed in cells in which Notch is activated. Indeed, oddrK111 expression was observed in cells directly adjacent, and distal, to ligand-expressing cells. Thus, odd (and drm and sob, as well) might be direct targets of Notch activation. In support of this notion, ectopic activation of Notch within the leg induces the expression of oddrK111, drm, and sob (Rauskolb, 1999). Further, expression was reduced in cells along the anterior-posterior (A-P) axis of the leg when Notch activity was inhibited in a stripe of anterior cells at the A-P compartment boundary by expression of the extracellular domain of Notch, which acts as a dominant-negative receptor (Hao, 2003).
bowl is also segmentally expressed in the developing leg, although its expression is not as restricted as that observed for the other three genes. bowl is expressed in a broad stripe within each presumptive leg segment, including tarsal segments 1-4, in a pattern that appears overlapping with, but also broader than, that of odd, sob, and drm. The bowlk08617 enhancer trap is also expressed in developing leg discs, in both larval third instar and pupal discs, in a pattern similar to that observed for the endogenous bowl transcripts (Hao, 2003).
Taken together, these results indicate that the odd-skipped family genes are candidate effectors of the influence of Notch on Drosophila leg segmentation. However, given that odd, sob, and drm are expressed identically, they may act redundantly in this process. bowl, in contrast, may act independently, especially within the distal leg, as it is expressed more broadly and is the only gene family member expressed within tarsal segments 1-4 (Hao, 2003).
In the developing leg, Notch mutant cells result in fused leg segments, owing to the absence of joints, and in severely reduced growth. Accumulation of pigmented tissue also occurs at joints in proximal regions. Mutations in bowl result in similar phenotypes; the mutant cells are associated with fusions and truncations of tarsal segments as well as with melanotic patches at the proximal joints. Because the gene is essential at earlier stages in development (bowl mutant embryos do not hatch; Wang, 1996), it is difficult to determine the consequence of completely eliminating bowl in the leg. However, when the mutant clones cover most of the distal part of the leg, the limb is severely truncated, with little or no segmentation/joint tissue evident. Conversely, clones that are restricted to the central part of the tarsus often fail to result in a detectable phenotype (de Celis Ibeas, 2003).
The bowl gene encodes a zinc-finger transcription factor and is closely related to odd-skipped, a gene that has already been implicated in leg segmentation (Wang, 1996; Rauskolb, 1999). Phenotypes produced by odd mutant cells were examined and defects similar to bowl were found -- segmental fusions and truncations in the tarsal region, and melanotic patches in proximal joints. The two genes therefore have similar but essential roles in leg segmentation, although their functions elsewhere appear to be distinct (Wang, 1996; de Celis Ibeas, 2003).
Neither Bowl nor Odd appear to be present within the tarsus (Rauskolb, 1999), yet mutations in either gene produce defects in this part of the leg (fusion of tarsomeres and growth defects). There are three models to explain this. (1) Bowl and Odd influence tarsomere segmentation indirectly, by regulating production of a long-range signal. (2) The two genes are expressed at intertarsomere joints but at a level too low to be detected. (3) They are involved in an earlier patterning event that influences subsequent tarsal segmentation. The first hypothesis can be ruled out, because fusions only occur within tarsal bowl clones and the effects on downstream genes are autonomous. Although it is difficult to rule out the low levels of tarsal expression implied by the second model, the data argue that this is an unlikely explanation for several reasons. (1) bowl clones that spanned only intertarsomere joints (e.g. t2/t3/t4) appeared normal. Almost all clones that resulted in observable phenotypes spanned the tibia/t1 boundary and, even within these clones, there was considerable variation in the number of tarsomere joints affected. This suggests that the effect on tarsomere joints is a secondary consequence of the mutations. (2) In the case of Odd, the pattern of expression detected in the larva persists throughout pupal leg morphogenesis (Mirth, 2002), ruling out later expression in the tarsomeres. (3) When high levels of bowl mRNA were produce using GAL4/UAS system, very little protein is detected in the tarsal domain of late-third-instar/early-pupal legs, suggesting either that the protein is unstable or that the mRNA is poorly translated in this region (de Celis Ibeas, 2003).
To investigate the third model (that bowl has an early patterning function in the leg), two questions were asked. (1) Whether E(spl)mß-CD2 expression in t2-t4 is affected by bowl mutations was tested, as predicted if Bowl acts prior to tarsomere boundary formation. As in the adult legs, there is considerable variation between clones, but clear disruptions to E(spl)mß-CD2 were observed in three out of nine clones. These defects are confined to the clone but do not strictly follow clone boundaries because some of the mutant cells retain wild-type levels of E(spl)mß-CD2 even in the most severe cases. This demonstrates that bowl function precedes Notch activation at the tarsomere boundaries and supports the hypothesis that its effects on tarsomere boundaries are secondary (de Celis Ibeas, 2003).
(2) It was asked whether mutations in bowl alter the expression of genes involved in the initial regionalization of the leg. Several genes have been identified that confer distinct regional identities and are expressed in broad domains within the leg disc. These include dachshund (dac), which is expressed in more proximal regions including t1, the two genes of the bric-a-brac (bab) complex (bab1 and bab2), which are expressed in the presumptive tarsal region, and two Bar genes that are expressed in distal tarsal segments t4 and t5 (BarH1 and BarH2). Expression of all three types of regional genes are affected by mutations in bowl. In late-third-instar/early-pupal leg discs, Bab2 expression normally extends to the proximal edge of t1. In bowl clones, ectopic Bab2 is detected in proximal parts of t1 and the levels in t2 are also altered. Conversely, when these clones also encompass the distal part of Dac domain, there is a reduction in Dac. Mutant clones that lie at the distal side of the tarsal domain again show derepression of Bab2 (in distal t5, where Bab2 expression is low or absent), and this is coupled with a reduction in the levels of BarH1. In all cases, the effects are autonomous to the clone and precisely follow clone boundaries (de Celis Ibeas, 2003).
It is therefore concluded that Bowl regulates the expression of patterning genes, promoting development of the proximal (t1/t2) and distal (t5) extremities of the tarsus. Thus, bowl mutations lead to an expansion of a 'central tarsal fate' that is characterized by uniform Bab2 and decreased BarH1 and Dac. This disruption in tarsal patterning would in turn affect the expression of Notch ligands in the tarsus and hence lead to defects in tarsomere joints, as seen in the disruption of E(spl)mß-CD2. However, there is still a discrepancy between the apparent function of Bowl and its site of expression: Bowl is necessary to inhibit/lower Bab2 expression in t1/t2 and t5, but is not present in these regions in late-third-instar/early-pupal discs. Nevertheless, the effects of bowl mutations on Bab2 and Dac are strictly cell autonomous. These observations can be reconciled if Bowl (and likewise Odd) is expressed within the cells that give rise to t1/t2 and t5 at early stages when the domains of the regional genes (bab, dac and Bar) are first established. This expression must subsequently disappear from these regions and become restricted to the boundaries of the tarsus (de Celis Ibeas, 2003).
If Bowl and Odd regulate tarsomere segmentation via effects on regional genes like bab2, they should be expressed at the boundaries of the Bab2 domain during early stages. At ~76-80 hours, both Bowl and Odd are first detected in a 2-3-cell-wide ring that surrounds the Bab2 expressing cells and corresponds to the proximal edge of the Dll domain and the distal edge of the Dac domain. Most of these Bab2-expressing cells also express BarH1; on the proximal side only a 1-2-cell-wide ring contains Bab2 and not BarH1. At this stage, therefore, the tarsus consists primarily of one identity, which has Bab2 and BarH1 expression and appears to approximate to t4. This early domain is surrounded by cells expressing Bowl and Odd (de Celis Ibeas, 2003).
Subsequently, a further ring of Bowl and Odd-expressing cells appears in the centre of the Bab2 domain, at the boundary with the pretarsus, and Bab2 is rapidly lost from within this ring. Bab2 is now flanked both proximally and distally by Bowl/Odd. At later stages, gaps appear between Bab2 and the flanking rings of Bowl/Odd. These gaps expand and, at the same time, Bab2 expression becomes more graded, with decreasing levels at the edges of its expression domain. This is most marked in the proximal (t2) direction, and is even more evident with Bab1 than with Bab2. Dac expression also extends distally beyond the proximal ring of Bowl/Odd so that it occupies t1 and a small part of t2 by the time the leg disc everts. As a consequence, a series of distinct territories is established within the tarsus by late third instar. Bowl/Odd mark the extreme (tibia/t1 and t5/pretarsal) boundaries of the tarsus and Bab2 expression spans t2-t5 with the peak of its expression in t3/t4 (de Celis Ibeas, 2003).
Both the expression patterns and the phenotypes suggest that cells within the t1/t2/t3 and t5 regions of the tarsal domain contain Bowl/Odd at 76-86 hours. It is proposed that Bowl/Odd expression is gradually lost from the tarsal cells as they proliferate, giving rise to a temporal gradient of Bowl/Odd (prolonged expression in t1, shorter period of expression in t3/t2). If this is the case, expression from the odd-lacZ line might be visible within t1/t2/t3 because of the endurance of ß-galactosidase. Indeed, in 96-hour-old odd-lacZ discs, ß-galactosidase is detected at lower levels within many cells of the tarsus. A temporal gradient cannot definitively be shown by this method, but the expression of odd-lacZ is most persistent close to the final domain of Bowl and Odd, consistent with this model (de Celis Ibeas, 2003).
Because bowl mutations result in expansion of 'central tarsal' (t3/t4) fates, it was anticipated that persistent expression of Bowl within the tarsus would have the converse effect, expanding proximal (t1/t2) and distal (t5) fates. To test this, GAL4 driver lines were used to direct expression of UAS-bowl within the tarsus, scoring phenotypes in the adult male pro-thoracic legs, using the sex comb as a marker of t1. Expression of UAS-bowl throughout the tarsal region (Dll-Gal4) gave rise to legs with expanded t1 fates manifest in the ectopic sex-combs on distal tarsal segments. In the more strongly expressing lines, the tarsus became severely distorted and carried sex-comb bristles throughout its length. Even with more restricted production of Bowl (e.g. klumpfuss-GAL4), similar transformations occurred, with ectopic sex-comb bristles present in t2 and t3 (de Celis Ibeas, 2003).
To determine whether expansion of t1 fates occur at the expense of 'central tarsal' fates, the effects of ectopic Bowl on Bab, Dac and BarH1, were assayed. In leg discs from Dll-GAL4/UAS-bowl, levels of Bab2 were strongly reduced and more patchy than in the wild type, consistent with central tarsal identity being compromised. Conversely, the domains of Dac and BarH1 were extended so that they were almost contiguous in the middle of the tarsus, demonstrating the expansion of t1 and t5 fates. Ectopic expression of Bowl in a more restricted domain (e.g. with Ptc-Gal4) also reduced Bab specifically within the domain of ectopic expression. The inhibition of bab2 by Bowl fits with the phenotypes of bab/bab2 loss-of-function alleles, which are similar to that of ectopic Bowl (ectopic sex combs on distal segments). In analysing the levels of Bowl produced by the directed misexpression, it was noted that high levels of protein only accumulate close to the normal sites of expression. Elsewhere, such as within the tarsus, protein levels remain low and patchy, even though mRNA levels are fairly uniform throughout the domain of misexpression. Despite the low levels of protein, inhibition of Bab is still seen within the tarsus (de Celis Ibeas, 2003).
Further support for the role of bowl in patterning the proximal tarsus comes from analysing spineless mutant legs. This gene is essential for antenna development but is also expressed transiently in the tarsus of the early third-instar leg. The phenotype observed in weak spineless mutants (ssa/ss114) resembles that of ectopic Bowl (with ectopic sex combs in t2 and an ectopic joint within t1), and the domain of Bowl expression remains broader in spineless larval discs and ectopic patches of Bowl persist in t1 and t2 of early pupal discs. Sometimes, the ectopic Bowl forms a discrete ring within t1 that corresponds to the site of the ectopic joint. Persistent Bowl therefore alters P/D patterning, promoting t1-like fates and, in some cases, resulting in an ectopic tibia/t1-like joint. These data suggest that spineless is involved in keeping Bowl absent from in the tarsus. In agreement with this, ectopic Spineless results in loss of Bowl, although these conditions also result in transformation to antenna fates, complicating the interpretations. It is concluded that the translation or the stability of Bowl protein are regulated. Candidates to participate in Bowl regulation include Spineless and Lines, a protein that acts antagonistically to Bowl and Drm in hindgut morphogenesis (de Celis Ibeas, 2003).
To assess the normal role of odd and its cognate genes sob and drm, during leg development, mutant clones were analyzed for those genes for which mutants are available. It was found that neither odd nor drm mutant clones had a detectable effect on leg segmentation or leg growth. Thus, odd and drm may not have a role during leg development, they may act redundantly, or their function may be masked by the presence of sob. Since neither a single mutation in sob nor a deficiency removing only odd, sob, and drm is available, the contribution of sob to leg development or potential redundancy amongst these genes could not easily be addressed. The Df(2L)drmP2 (Green, 2002), a large deficiency removing drm, sob, and odd, as well as about 30 other genes, appears to be cell lethal, as mutant clones could not be recovered in the adult leg (Hao, 2003).
Although determination of the requirement for odd, sob, and drm in leg segmentation is complicated by their potential redundancy, ectopic expression has the possibility to reveal whether any of these three genes is sufficient to promote aspects of segmentation in the developing leg. Moreover, if odd, sob, and drm are important downstream effectors of Notch activity, then ectopic expression of these genes should give phenotypes similar to those observed with ectopic Notch activation, namely induction of segmentation and/or tissue growth. To test the functional potential of odd, sob, or drm, UAS expression constructs were made, allowing for either patterned misexpression or generation of FLP-out clones during leg development, and the resulting phenotypes were examined in adult legs (Hao, 2003).
Importantly, ectopic expression of odd, sob, or drm can induce deep creases in the leg cuticle that resemble ectopic joint-like structures. The phenotype obtained upon ectopic expression of these genes along the A-P axis of the leg is similar to that observed with ectopic expression of a constitutively activated form of Notch. The ectopic expression of these genes was not associated with altered Notch ligand expression, and thus these genes appear to function solely downstream of Notch activation. In addition, broad misexpression of odd, sob, and drm resulted in tarsal segment fusions, as has also been observed with broad misexpression of activated-Notch. This result implies that, for normal leg development to occur, odd, sob, and drm must be expressed in a segmentally repeated pattern (Hao, 2003).
To address whether odd, sob, and drm behave cell autonomously, small patches of cells expressing the gene of interest were made using the FLP-out GAL4 technique. Significantly, FLP-out clones of odd, sob, or drm were each capable of inducing an indentation in the cuticle, in a largely cell-autonomous manner, and such alterations can occur anywhere along the proximodistal axis of the leg. In some instances, a crease in the cuticle was induced around the FLP-out clone, while in other instances, all the cells of the clone appeared to invaginate and form an indentation in the cuticle. In these latter instances, a lack of bristles was often found within the clone. One interpretation of these observations is that these cells are now fated, by virtue of their expression of odd, sob, or drm to become "joint-like" cells, which indent and do not make bristles. FLP-out clones of odd, sob, or drm never induce leg tissue outgrowths. By contrast, FLP-out activated-Notch clones or FLP-out four-jointed clones (another Notch target gene) do induce leg tissue outgrowths (Hao, 2003).
Together, these results indicate that the role of odd, sob, and drm during leg development may be to initiate cellular changes associated with joint formation. Since ectopic expression of these genes does not induce growth of the leg tissue, they appear to regulate only a subset of the functions downstream of Notch activation during leg development (Hao, 2003).
The formation of a joint is a complex process involving cell movements and changes in cell shape. To determine whether odd, sob, or drm might contribute to joint formation by altering cellular morphology, these genes were ectopically expressed along the A-P axis of the leg and cellular morphology was examined by staining with anti-E-cadherin or with phalloidin. Both of these reagents preferentially label the apical surface of cells, at the level of the adherens junctions. ptcGAL4 drives the expression of the gene of interest in a gradient within the anterior compartment, with the highest levels of expression at the A-P compartment boundary. The expression of each gene was monitored by coexpression of GFP (Hao, 2003).
Normally, in late third instar discs, cells at the tip of the leg, within the pretarsus, are arranged as a flat epithelial sheet, with their apical surfaces densely packed and in register. This is revealed by a straight line of either E-cadherin or phalloidin staining seen when the discs are examined in cross-section. By contrast, when constitutively activated-Notch is expressed along the A-P axis (ptcGAL4 UASGFP UASN*), cells are induced to fold into the pretarsus, as an indentation in the epithelial sheet is observed. When observed in cross-section, the cells at the A-P border expressing activated-Notch appear to have invaginated, as they are now located below the plane of the pretarsal epithelium. However, they do not lose contact with their neighbors and are still a continuous part of the epithelial sheet. This invagination occurs at the border between cells expressing activated-Notch and nonexpressing cells; in fact, cells in which Notch is activated appear to nonautonomously induce neighboring cells to invaginate. Although similar results were obtained with E-cadherin and phalloidin staining when examining the behavior of cells expressing activated Notch, it is noted that, while E-cadherin expression is not altered, the invaginating cells have higher levels of phalloidin staining than their neighbors. Interestingly, increased phalloidin staining is observed at the endogenous joints during leg segmentation (Hao, 2003).
Since odd, sob, and drm are targets of Notch activation, it was asked whether their ectopic expression might induce changes in epithelial morphology like those observed with constitutive Notch activation. Strikingly, ectopic expression of odd along the A-P axis of the leg gave identical cellular changes to those observed upon activated-Notch expression. An invagination with increased phalloidin-staining levels at its apical surface was induced and nonautonomous effects on neighboring nonexpressing cells were observed. Similar results were obtained upon ectopic expression of sob, while ptcGAL4-driven drm expression was lethal. Taken together, these data showing similar morphological changes associated with FLP-out clones of odd, sob, and drm, and similar cellular changes associated with ptcGAL4-driven odd and sob, strongly suggest that each gene can act as an effector of Notch activation to promote the epithelial cellular changes driving joint formation (Hao, 2003).
The developing leg joint is composed of different cell populations that ultimately contribute to a particular structure of the adult leg joint. These different cell populations can be identified in the developing leg disc by the genes that they express. Although these studies show that the odd-skipped family genes can promote one aspect of joint morphogenesis, epithelial invagination, it remained possible that this phenotype was the result of a more general role for these genes in regulating joint fate. However, this latter possibility does not seem to be the case, since ectopic odd expression does not activate the expression of other genes that serve as markers of the developing joint. One of these joint markers is nubbin, which is expressed in proximal joint tissue, just proximal to those cells expressing odd, sob, and drm. Ectopic odd expression does not alter the expression of nubbin in any region of larval and pupal leg discs. In particular, closer examination of the pretarsus, in which the aforementioned epithelial invaginations were characterized, shows no change in nubbin expression. Examination of other markers of the presumptive joint, including E(spl)mß, big brain, odd, and drm, indicates that the expression of these genes is not induced upon ectopic odd expression, but rather expression also remains unaffected within the pretarsus. It was noted that, during late third instar stages and pupal development, when ptcGAL4-driven odd expression results in extensive tarsal segment fusions, the expression of E(spl)mß and big brain is repressed within the developing tarsus. It is believed that this disruption of their expression is a secondary consequence of the fusion of these segments, since earlier in development their expression is unaffected by ectopic odd expression. Together, the data are consistent with the idea that the odd-skipped family genes promote the epithelial cellular changes characteristic of invaginating joint cells, without inducing the expression of other joint markers, and hence these genes appear to have a specific role in joint morphogenesis (Hao, 2003).
These studies show that the odd-skipped family is a key group of genes induced upon Notch activation that promotes morphological changes associated with joint formation during leg development. Expression of odd, sob, and drm is induced in cells responding to Notch activation; these cells lie distal to Notch ligand-expressing cells. bowl expression is also regulated by Notch. Ectopic expression of odd, as well as loss of bowl function, does not alter the expression of Notch ligands. Hence, the morphological changes induced by expression of these genes appear to be mediated downstream of Notch activity. Ectopic expression of odd, sob, and drm, like ectopic Notch activation, can cause alterations in the leg cuticle that resemble those that occur at joints, including deep creases within the cuticle and an absence of bristles. Importantly, their ectopic expression, like ectopic Notch activation, induces cells to form an invaginating furrow, while still remaining part of the disc epithelium. Interestingly, during normal leg development, mid-distal joint cells express odd; these are the same cells that will invaginate and ultimately fold under proximal joint cells. Moreover, the cells that invaginate because of their expression of odd, both in ectopic expression studies and in wild-type legs, accumulate high levels of apical filamentous actin. Further support for the idea that odd, sob, and drm control a specific aspect of cell morphogenesis, an invagination, as opposed to being more generally required for specifying joint fate, comes from the observation that ectopic odd expression does not induce the expression of other markers of joint fate, including nubbin, E(spl)mß, big brain, odd, and drm. This contrasts with the effect of Notch, which induces nubbin, E(spl)mß, big brain, odd, and drm expression. Moreover, FLP-out clones of Notch induce outgrowths of leg tissue, whereas FLP-out clones of odd, drm, and sob are not associated with leg outgrowths. Thus, while both ectopic Notch activation and ectopic odd or sob expression are capable of inducing an invagination into the disc epithelium, Notch activation must further organize additional aspects of joint formation and also leg growth. Taken together, the ectopic expression studies indicate that odd, sob, and drm are Notch target genes that mediate a subset of the activities of Notch during leg development, namely, they promote a cell morphological change, an epithelial invagination, which normally occurs during joint formation (Hao, 2003).
Interestingly, the involvement of the odd-skipped family of segmentation genes in promoting epithelial cellular changes may not be unique to the leg joint. odd-LacZ is expressed in the apodemes of the developing leg, which are tubes of invaginating cells that serve as muscle attachment sites. Thus, these genes may have a role in promoting the apodeme invagination as well. odd is also required for embryonic segmentation during which segmental borders are defined by intersegmental furrows; cells at the prospective segment boundary elongate and fold into the epidermis. While the relationship between odd expression and cell morphology during embryonic segmentation has not been elucidated, it is possible that one function of odd in segmentation is to orchestrate epithelial invaginations. Thus, the odd-skipped gene family may be required in multiple developmental contexts to induce epithelial cellular changes, such as promoting an invagination, as has been have described here. Since the odd-skipped family genes encode transcriptional regulators, it is hypothesized that they regulate the expression of genes involved in cytoskeletal architecture or cell adhesion (Hao, 2003).
These studies on the function of odd, sob, and drm suggest that these genes may have a similar function during leg development. They share a common expression pattern at all stages of leg development, consistent also with their overlapping expression in the embryo. Importantly, ectopic expression of each gene is capable of inducing the same morphological changes in the adult cuticle and, for odd and sob, the same cellular changes in the leg disc epithelium. It is thus suggested that odd, sob, and drm act redundantly during leg segmentation. Hence, it is not surprising that, when only one of the genes is removed, no effect on leg development is observed. It will ultimately be of interest to determine the phenotype of leg tissue triply mutant for odd, sob, and drm (Hao, 2003).
bowl, in contrast, may have adopted functions that are independent of and/or not obscured by the other three members of the family. bowl expression appears largely distinct from the other three genes; its expression encompasses a broader domain that overlaps that of the other genes in proximal segments and tarsal segment 5, while bowl is the only odd-skipped family gene expressed in tarsal segments 1–4. The identical expression profile of odd, sob, and drm, yet distinct pattern of bowl, is also observed in other tissues. The observation that the odd-skipped family genes are expressed in overlapping domains in a number of different developmental contexts, yet are not always genetically redundant, suggests that their contribution to a particular morphogenetic process may depend on their relative expression levels or their interaction with other proteins in that particular tissue (Hao, 2003).
In fact, recently, it has been shown that a physical interaction between one of the odd-skipped family members, Drm, and another transcriptional regulator, Lines, is important during hindgut morphogenesis [Green, 2002]. By interacting with Lines, Drm inhibits Lines activity in the embryonic hindgut, thereby allowing specification of the small intestine. As the functionally significant Drm-Lines interaction was mapped to the first zinc-finger of Drm, it is conceivable that Odd, Sob, and Bowl may also interact with Lines in other developing tissues [Green, 2002]. Indeed, this does seem to be the case, since regulatory interactions among Drm, Bowl, and Lines operate during the patterning of the embryonic dorsal epidermis and the foregut [Hatini, 2000 and Johansen, 2003b]. In these contexts, Lines inhibits Bowl, resulting in a particular cell type. The remaining cell types are controlled by Drm, which activates Bowl by causing inhibition of Lines (Hao, 2003).
These results are consistent with this molecular genetic circuit also functioning during Drosophila leg development. Clones of cells mutant for bowl are unable to participate in joint formation, resulting in melanotic protrusions from the leg cuticle in proximal segments and in a fusion of tarsal segments. The difference in the phenotype of bowl clones in proximal versus tarsal segments may be because proximal joints do not form in the same way as tarsal joints, although some of the changes in cell behavior are presumably conserved. Notably, the difference does not appear to be due to redundancy, because loss-of-function bowl mutations result in the fusion of the tibia and tarsal segment ta1, despite the fact that all four genes are expressed in the developing tibia. Thus, odd, sob, and drm are insufficient to induce tibia-tarsal 1 joint formation in the absence of bowl; the ability of these genes to induce morphogenesis might be dependent in some way on the expression of bowl. It has also recently been reported that lines has a role during leg development [Green, 2002]. These results are all consistent with a molecular model in which Bowl and Lines interact to regulate joint formation during leg development, although it remains to be determined whether Lines inhibits Bowl function or whether a Bowl–Lines complex regulates the expression of genes effecting joint formation. It is further proposed that formation of proximal leg joints requires the additional contribution of Odd, Sob, and Drm, which act redundantly to relieve the repression of Bowl by Lines. In such a model, ptcGAL4-driven ectopic bowl expression would be insufficient to induce ectopic segmentation in the leg, as was observed, since Lines would repress Bowl and hence render Bowl inactive. Also, drm bowl mutant clones would behave similar to bowl mutant clones, as was observed, since odd and/or sob would compensate for the loss of drm. While odd, sob, drm, and bowl may act together to regulate proximal leg segmentation, it appears that only bowl is essential to tarsal segmentation, as both these data and that of [Mirth, 2002] indicates that odd, sob, and drm are not expressed in tarsal segments 1–4. This would suggest that within tarsal segments 1–4 an alternative mechanism regulates Bowl activity (Hao, 2003).
The Drosophila embryonic hindgut is a robust system for the study of patterning and morphogenesis of epithelial organs. In a period of about 10 h, and in the absence of significant cell division or apoptosis, the hindgut epithelium undergoes morphogenesis by changes in cell shape and size and by cell rearrangement. The epithelium concomitantly becomes surrounded by visceral mesoderm and is characterized by distinct gene expression patterns that forecast the development of three morphological subdomains: small intestine, large intestine, and rectum. At least three genes encoding putative transcriptional regulators, drumstick (drm), bowl, and lines (lin), are required to establish normal hindgut morphology. The defect in hindgut elongation in drm, bowl, and lin mutants is due, in large part, to the requirement of these genes in the process of cell rearrangement. Further, drm, bowl, and lin are required for patterning of the hindgut, i.e., for correct expression in the prospective small intestine, large intestine, and rectum of genes encoding cell signals (wingless, hedgehog, unpaired, Serrate, dpp) and transcription factors (engrailed, dead ringer). The close association of both cell rearrangement and patterning defects in all three mutants suggest that proper patterning of the hindgut into small intestine and large intestine is likely required for its correct morphogenesis (Iwaki, 2001).
Focusing on hindgut elongation that occurs after stage 10, neither apoptosis nor cell proliferation contribute significantly to the process. Thus, hindgut morphogenesis occurs normally in the apoptosis-deficient DfH99 mutant, and the only cell proliferation occurring in the hindgut after stage 10 is in a small domain at the anterior of the small intestine. The morphogenesis of the hindgut after stage 10, in particular its elongation and narrowing, must therefore be driven by changes in cell size, shape, and rearrangement (Iwaki, 2001).
After the cessation of the postblastoderm mitoses, an endoreplication cycle increases the size of the cells of the large intestine (but not small intestine or rectum). Inhibition of this endoreplication by different genetic manipulations results in a shorter large intestine with a smaller cell size, but roughly normal diameter. Endoreplication thus appears to be required to bring the large intestine to its full length, but not to play a critical role in reducing hindgut diameter. There is a change in cell shape, from columnar to cuboidal, as the hindgut elongates; such a change increases epithelial surface area and thus could contribute to hindgut elongation, but not to a reduction in its diameter (Iwaki, 2001).
The threefold elongation of the hindgut is accompanied by a three- to four-fold reduction of circumferential cell number, but not by appreciable cell proliferation or apoptosis. The major process driving this stereotypic elongation and narrowing must therefore be cell rearrangement. Elongation by cell rearrangement is a morphogenetic process of broad significance: it has been shown to drive gastrulation and embryonic axis elongation, and elongation of various tissues, throughout the bilateria. To date, few molecules required for this process have been identified. Elongation by cell rearrangement of the Drosophila germband, ovarian terminal filaments, and stigmatophore requires the Evenskipped homeodomain, Bric a brac BTB, and Grain GATA proteins, respectively, while that of the C. elegans dorsal epidermis requires the DIE-1 zinc finger protein. The genetic pathways in which these presumed transcriptional regulators function have not yet been determined. Only the Xenopus Brachyury transcription factor has been shown to affect cell rearrangement by controlling expression of a specific target, Wnt11, which acts via the planar cell polarity pathway to orient cell intercalation. A fuller understanding of the molecular basis of oriented cell rearrangement clearly depends on the identification of additional required genes and genetic pathways (Iwaki, 2001).
Since the hindguts of their mutant embryos are shorter and wider than normal, drm, bowl, and lin have been identified as possible regulators of the cell rearrangement that drives hindgut elongation. Analysis of hindgut morphology and gene expression patterns in mutants indicates that drm, bowl, and lin function in hindgut development after the primordium has already been established and internalized by gastrulation. No massive apoptosis in the hindgut (as seen in fkh, cad, and byn) is observed in drm, bowl, or lin hindguts. The number of cells in the hindgut epithelium of drm, bowl, or lin mutants is within 20% of wild type, demonstrating that cell proliferation is roughly normal in these mutants. The byn and fkh genes are expressed normally throughout drm, bowl, and lin hindguts, and otp is expressed throughout drm and bowl hindguts. The hindgut visceral mesoderm, on the basis of its expression of Connectin, appears to be established normally in drm, bowl, and lin mutants. Taken together, these results indicate that early events in hindgut development, namely the establishment and maintenance of the primordium (including initiation of gene activity, and cell proliferation throughout the primordium), its internalization during gastrulation, and its investment with visceral mesoderm, all occur more or less normally in drm, bowl, and lin mutants. The shorter overall length, and the two- to three-fold greater circumferential cell number seen in drm, bowl, and lin hindguts, must therefore be a result of a failure to complete the cell rearrangement that elongates and narrows the wild-type hindgut (Iwaki, 2001).
Patterning of the Drosophila hindgut serves as a microcosm of the complex anteroposterior and dorsoventral patterning that takes place during vertebrate gut development. In the Drosophila hindgut, patterning along the anteroposterior axis gives rise to the small intestine, large intestine, and rectum; patterning along the dorsoventral axis gives rise to the large intestine ventral and large intestine dorsal domains, and the boundary cells. Previous studies described gene expression patterns in the different domains of the developing Drosophila hindgut (as well as the requirement of fkh for these expression patterns), but did not identify any genetic activity that distinguished among or specified the different domains (Iwaki, 2001).
This study shows that drm, bowl, and lin are required for the gene expression patterns that distinguish these three domains: lin is required for expression characteristic of large intestine (dpp, dri, and en) and rectum (Ser, hh, and wg); drm and bowl are required for expression characteristic of small intestine (hh and upd). By both morphological criteria (cell shape, presence or absence of boundary cells) and gene expression patterns (expanded expression of genes expressed in the small intestine), lin hindguts appear to consist of a greatly expanded small intestine and to lack the large intestine and rectum. In contrast, both morphological and gene expression characteristics of drm and bowl hindguts indicate that they lack most or all of the small intestine, and consist only of large intestine (which remains unelongated) and rectum. A model consistent with these data is that lin functions in the hindgut to repress small intestine fate and to promote large intestine and rectum fate, while establishment of the small intestine requires the activity of drm and bowl. The requirement for drm (but not bowl) for wg expression at the most anterior of the hindgut could be explained if the domain of bowl function in the small intestine does not extend to the most anterior of the hindgut (consistent with the expression of bowl). Since they have opposite effects on Ser expression, bowl and drm may function in different ways, possibly in different pathways, to promote small intestine fate (Iwaki, 2001).
The function of lin as both an activator and repressor of gene activity in the developing hindgut is consistent with molecular and genetic characterization of its function in other embryonic tissues. In the developing dorsal epidermis, lin is required for transcriptional regulation (both activation and repression) of targets downstream of wg signaling. In the developing posterior spiracles, lin is required for the activation by Abd-B of its transcriptional targets. lin encodes a novel protein that is expressed globally throughout the embryo, including the developing hindgut. When ectopically expressed, Lin protein is detected in nuclei of cells signaled by Wg. The early expression of wg throughout the hindgut primordium, starting at the blastoderm stage and continuing to stage 10, might, analogous to its effect in the dorsal epidermis, activate Lin. This might be required for Lin to carry out its function, demonstrated here, of promoting expression of genes characteristic of large intestine identity (otp, dpp, en, and dri), and repressing expression of genes characteristic of small intestine identity (hh, upd, and Ser) (Iwaki, 2001).
It has been shown by genetic analysis that bowl and drm function to establish the small intestine. bowl encodes a zinc finger protein related to Odd-skipped and is expressed strongly in the hindgut primordium starting at the blastoderm stage and continuing through stage 11. Although the Bowl protein has not been shown to be nuclear or to bind DNA, the fact that it encodes five tandem zinc fingers suggests that it is a transcription factor. Thus, Bowl might function in the hindgut as an activator or coactivator of transcription of genes characteristic of small intestine fate. Finally, drm encodes a zinc finger protein related to Bowl and Odd-skipped and is expressed during stage 10 in the anterior of the developing hindgut, consistent with its required role in establishing the small intestine. drm, like bowl, is required for gene expression characteristic of small intestine fate. The drm protein may, like Bowl, function as a transcriptional regulator in the small intestine primordium (Iwaki, 2001).
Thus drm, bowl, and lin are required for both patterning and cell rearrangement of the hindgut. At least one other putative transcriptional regulator expressed in the hindgut has similar properties: Dichaete encodes a Sox protein required for en, hh, and dpp expression in and elongation of the hindgut. The question therefore arises whether any of the genes expressed in different hindgut domains are mediators of the required role of drm, bowl, lin, or Dichaete in hindgut morphogenesis (Iwaki, 2001).
The phenotypes described for wg, hh, dpp, dri, Ser, and en do not suggest a role for these genes in hindgut elongation by cell rearrangement. Mutations in Ser, dri, and en do not appear to affect overall hindgut morphology. The hindgut in wg mutants is extremely small, suggesting that the critical function of wg in hindgut development is in establishing and maintaining the primordium, but not in elongation. dpp mutant hindguts are shorter, consistent with the role of dpp in endoreplication in the large intestine; nevertheless, dpp hindguts have a roughly normal diameter. In hh mutant embryos, the rectum degenerates and hindgut length is reduced, but the overall morphology, particularly the narrowing of the large intestine, appears normal (Iwaki, 2001).
Only in upd embryos is a defect in both elongation and narrowing of the hindgut observed; significantly, upd is expressed only in the small intestine, the same domain that is largely missing from drm and bowl mutants. Shorter and wider hindguts are seen in younger upd embryos, but the majority of hindguts in mature upd embryos appear normal. Thus, while upd may at least partially mediate drm and bowl function in the hindgut, there must be other targets of these genes that are required for cell rearrangement in the hindgut (Iwaki, 2001).
It is concluded that, if correct patterning of the hindgut is a prerequisite for its elongation by cell rearrangement, either all the targets of drm, bowl, and lin that are the essential components of the necessary patterning have not been identified, or the genes presently identified have overlapping or redundant function. Consistent with the idea that cell rearrangement in the Drosophila hindgut requires its correct patterning, convergent extension during vertebrate gastrulation has been shown to depend on patterning of cell fates along the dorsoventral axis of the embryo. It is, of course, possible that hindgut patterning and cell rearrangement, although closely associated both temporally and in the drm, bowl, and lin mutant phenotypes, do not have a necessary relationship to each other. A number of genes are known that, without affecting patterning, control cell rearrangement by directly affecting morphogenetic movements. This is a property of the Drosophila GATA transcription factor-encoding grain in stigmatophore elongation and of the zebrafish trilobite locus in body axis elongation. Thus, drm, bowl, lin, and Dichaete, in addition to patterning the hindgut, might be regulating other genes that independently control cell rearrangement. Nonetheless, the relationship between patterning of both small intestine and large intestine, on the one hand, and cell rearrangement, on the other hand, is striking. drm and bowl hindguts have a substantial cohort of large intestine cells, yet fail to complete cell rearrangement, presumably due to absence of the small intestine. lin hindguts have an excess of small intestine cells and also fail to complete cell rearrangement, presumably due to absence of the large intestine. The connection between hindgut patterning and cell rearrangement observed in drm, bowl, and lin mutants supports the idea that interaction between two correctly patterned anteroposterior subdomains, the small and large intestine, is a requirement for cell rearrangement in the hindgut tubule (Iwaki, 2001).
In addition to their roles in patterning and morphogenesis of the hindgut, the Drosophila genes drumstick (drm) and bowl are required in the foregut for spatially localized gene expression and the morphogenetic processes that form the proventriculus. drm and bowl belong to a family of genes encoding C2H2 zinc finger proteins; the other two members of this family are odd-skipped (odd) and sob. In both the fore- and hindgut, drm acts upstream of lines (lin), which encodes a putative transcriptional regulator, and relieves the lin repressive function. In spite of its phenotypic similarities with drm, bowl was found in both foregut and hindgut to act downstream, rather than upstream, of lin. These results support a hierarchy in which Drm relieves the repressive effect of Lin on Bowl, and Bowl then acts to promote spatially localized expression of genes (particularly the JAK/STAT pathway ligand encoded by upd) that control fore- and hindgut morphogenesis. Since the odd-family and lin are conserved in mosquito, mouse, and humans, it is proposed that the odd-family genes and lin may also interact to control patterning and morphogenesis in other insects and in vertebrates (Johansen, 2003b).
By following the spatially localized expression domains of Wg, Dead ringer (Dri), and Connectin (Con), it is possible to identify at least four processes that contribute to morphogenesis of the three-layered valve of the proventriculus. During stages 13-15, a bulge referred to as the 'keyhole' forms in the foregut ectoderm, at its junction with the endoderm of the anterior midgut; this appears to be a process of evagination, similar to that described for formation of the vertebrate optic cup. The keyhole can be distinguished from the anterior ectoderm of the esophagus and from the anterior endoderm of the anterior midgut, since it is not surrounded by visceral mesoderm and is flanked by two domains of Wg expression. During stage 16, the anterior hemisphere of the keyhole reverses its curvature, and is approached by the posterior hemisphere; this has been described as a folding process. During stage 17, the posterior portion of the keyhole moves interiorly over the anterior lip of the developing proventriculus, a movement that appears similar to involution over the dorsal lip of the blastopore in Xenopus. Also during stage 17, the most posterior portion of the esophagus, i.e. that portion just anterior to the keyhole, inserts posteriorly into the pocket consisting of the most anterior endoderm of the anterior midgut. During this step of insertion, there is significant elongation of the most interior cells, which are derived from the anterior of the keyhole. At the conclusion of proventricular morphogenesis, the domain of Wg expression initially at the anterior of the keyhole is now inserted most distally into the proventriculus, while the domain of Wg initially posterior to the keyhole is now just interior, at the most anterior of the proventriculus (Johansen, 2003b and references therein).
drm and bowl mutants, which exhibit similar defects in hindgut elongation (Iwaki, 2001), are also similar in that they have related effects on proventriculus folding; in both mutants, rather than undergoing the first step of evagination to form the keyhole, the entire foregut remains as a narrow tube (Johansen, 2003b).
In both foregut and hindgut, localized expression of signaling molecules is required for gut morphogenesis; in the hindgut, this patterned expression has been shown to depend on drm and bowl (Iwaki, 2001). In the foregut, drm and bowl are also required for patterned gene expression, specifically for the two stripes of Wg that bracket the keyhole primordium, as well as for expression of upd and dri in the posterior hemisphere of the keyhole primordium. Another indicator of foregut patterning is its investment with visceral musculature, which surrounds the foregut epithelium with the exception of the keyhole region. In both drm and bowl mutants, the entire foregut (up to the anterior midgut) is surrounded by visceral mesoderm, indicating that the keyhole region is not established in these mutants. The odd mutant has no detectable foregut or hindgut defect. In a screen of over 10,000 chromosomes, no lethal mutation was isolated in sob; it therefore seems unlikely that sob plays a significant role in gut morphogenesis (Johansen, 2003b).
The similar sequence, phenotype, and partially overlapping expression of drm and bowl suggest that these genes might play redundant roles in foregut and hindgut morphogenesis. To address this possibility, double, triple, and quadruple mutants were constructed. Proventriculus and hindgut morphologies, patterned gene expression, and investment with foregut visceral musculature indicate that the drm bowl double mutant has hindgut and foregut phenotypes similar to those of drm and bowl single mutants. Further, drm sob odd (drmP2) and drm sob odd bowl (drmP2 bowl) mutants also have foregut and hindgut phenotypes that resemble those of drm and bowl single mutants. Since no additional phenotypes were revealed when drm, sob, odd, and bowl mutants were combined, it is concluded that, in the gut, members of the odd family, in particular drm and bowl, do not have redundant or overlapping function (Johansen, 2003b).
One odd-family member, drm, has previously been shown to interact with lin; like drm, lin is required for both hindgut patterning and the cell rearrangement that elongates the hindgut (Green, 2002 and Iwaki, 2001). In the foregut, the processes of folding and involution that form the proventriculus fail to occur in both drm and lin mutant embryos; the resulting phenotypes, however, are distinct: the drm foregut is long and narrow, while that of lin is short and bloated. Consistent with what has been observed for the hindgut, the foregut phenotype of the drm lin double mutant is very similar to that of the lin single mutant. In addition, the region that will become the keyhole, i.e., that portion of the foregut bracketed by Wg expression and lacking Con-expressing visceral musculature, is expanded in both lin and drm lin embryos. Since the foregut in the drm lin double mutant (both in terms of gene expression and morphology) is similar to that seen in the lin single mutant, it is concluded that in the foregut, as has been shown in the hindgut (Green, 2002), lin is epistatic to (acts downstream of) drm (Johansen, 2003b).
Since the bowl phenotype is similar to that of drm in both foregut and hindgut (Iwaki, 2001; Wang, 1996), the epistatic relationship between bowl and lin in both fore- and hindgut was investigated. Strikingly, while the drm lin phenotype is similar to that of lin, the bowl lin mutant phenotype, in both fore- and hindgut, appears the same as that of bowl. Further, the hindgut and foregut of drm bowl lin and drm sob odd bowl lin (drmP2 bowl lin) embryos are indistinguishable from those of bowl embryos. These results are consistent with observations that odd and sob are not required for gut morphogenesis (Wang, 1996). Most importantly, they show that in both hindgut and foregut, bowl is epistatic to (acts downstream of) lin (Johansen, 2003b).
More detailed analysis of gene expression confirms the epistasis of bowl to lin in both the hindgut and foregut. upd expression, which is observed throughout the small intestine (anterior portion of the hindgut) in wild-type (Iwaki, 2001), is barely detectable in bowl and bowl lin, but greatly expanded in lin hindguts. Similarly, hh expression, seen in both small intestine and rectum, is greatly reduced in the anterior of bowl and bowl lin, but expanded in lin hindguts. En expression, seen on the dorsal side of the large intestine (Iwaki, 2002), is expanded to both dorsal and ventral sides in bowl and bowl lin, but missing from lin hindguts. dri, expressed strongly in the two boundary cell rows and at a lower level in all cells of the small intestine (Iwaki, 2001), is expressed in duplicated boundary cell rows in bowl and bowl lin hindguts, but at a low level throughout the lin hindgut. All of these results confirm the epistasis of bowl to lin in the hindgut (Johansen, 2003b).
Characterization of gene expression also supports the epistasis of bowl to lin in the foregut. The expression of upd in the foregut epithelium, in a region that will become the posterior keyhole, is not seen in bowl or bowl lin mutant foreguts, but is greatly expanded in lin mutants. hh expression, which extends throughout the keyhole and part of the foregut anterior to it, is significantly reduced in bowl and bowl lin, while it appears to be expanded in lin embryos. dri, expressed in a narrow ring in the posterior keyhole region, is missing in bowl and bowl lin, but expanded in lin embryos. It is concluded that in both foregut and hindgut, lin is epistatic to drm, and bowl is epistatic to lin; in other words, lin acts downstream of drm, and bowl acts downstream of lin (Johansen, 2003b).
The byn-GAL4 construct drives posterior gut-specific expression, specifically, a uniformly high level of expression in the hindgut starting at stage 8 and continuing through embryogenesis (Iwaki, 2002). When this construct is used to drive ectopic expression of drm throughout the hindgut, a lin-like phenotype, i.e., expansion of small intestine, is observed, while ectopic expression of lin throughout the hindgut results in a drm-like defect, i.e., loss of small intestine (Green, 2002). This is demonstrated by morphology, by expanded expression of upd and hh, and by absence of expression of En (Johansen, 2003b).
In contrast to the dramatic effect of ectopic drm, ectopic expression of bowl throughout the hindgut has little effect on morphology and patterning. The morphology of the hindgut is altered only modestly: the small intestine is somewhat wider, the large intestine shorter, and the rectum a little longer. Similarly, patterning of the hindgut does not appear different: expression of upd, hh, and En is normal. Therefore, although bowl is required to specify the small intestine and for normal hindgut elongation, it functions differently from drm (Johansen, 2003b).
A genetic hierarchy of drm, lin, and bowl activity controls the localized expression of upd in both foregut and hindgut. Previous work has shown that, most likely by establishing a gradient of JAK/STAT activity, localized expression of upd is required for hindgut morphogenesis, specifically the cell rearrangement that drives hindgut elongation (Green, 2002 and Johansen, 2003a). upd is expressed in the region of the foregut that will become the posterior hemisphere of the keyhole, and remains restricted to this region through the remainder of embryogenesis. Stat92E, a transcriptional target of JAK/STAT signaling in the embryo, is expressed in the foregut in a domain that overlaps with, but extends beyond (both anteriorly and posteriorly), the ring of upd expression in the keyhole. Analysis of anti-Crb stained embryos shows that, in upd mutants, the evagination and folding steps of proventriculus morphogenesis occur, but the involution and insertion steps do not take place. It is concluded that upd is required for late steps in proventricular morphogenesis, and that Upd is likely to affect cells beyond the domain in which it is expressed (Johansen, 2003b).
For each of the Drosophila odd-family genes drm, sob, odd, and bowl, a specific ortholog could be identified in the Anopheles gambiae genome; this was possible because of the high similarity in amino acid sequence between specific zinc fingers. Three of the four Anopheles odd-family genes (the fourth was not mapped at the time of this submission) are, as in Drosophila, clustered on one chromosome arm (Johansen, 2003b).
There is a high degree of identity not only between the sequence of each of the five zinc fingers in the Drosophila and Anopheles Sob and Bowl proteins, but also between these and the five zinc fingers encoded by a splice variant of mouse Odd-skipped related 2. Remarkably, these proteins share 88% identity over the 135 amino acids that comprise five zinc fingers, implying that this group of fingers may have conserved molecular function (Johansen, 2003b).
Like the odd-family genes, lin is highly conserved between Drosophila and Anopheles, showing 43% identity over 858 amino acids, and a striking 76% identity in the 165 amino acid 'Lines homology domain' at the carboxy terminus (Johansen, 2003b and references therein).
From the data presented here and previously, it is concluded that drm, lin, and bowl have the same relationship to each other in both the foregut and the hindgut. drm and bowl in one case, and lin in the other, affect specification of the small intestine in different ways: bowl, expressed throughout the hindgut, and drm, expressed at the anterior of the hindgut, are both required to establish the small intestine, while lin, expressed throughout the hindgut, represses the small intestine (Green, 2002; Iwaki, 2001). Similarly, in the foregut, drm and bowl are both required to establish the keyhole, while lin, expressed throughout the foregut, represses formation of the keyhole region. For mutants in all three genes, failure to establish the keyhole results in an early failure in morphogenesis of the proventriculus (Johansen, 2003b).
These observations can be integrated with the results of epistasis and ectopic expression experiments presented in this study to yield a hierarchical model for the mechanism by which drm, lin, and bowl gene activities interact to specify discrete domains in the gut. The epistasis of bowl to lin as demonstrated in this study means that the repressive activity of lin acts through bowl. Bowl promotes specification of the small intestine and keyhole, while Lin represses specification of these regions. Drm, expressed in the small intestine and in the keyhole, relieves repression of the small intestine and keyhole by repressing Lin, thereby allowing Bowl to function (Johansen, 2003b).
Although bowl is necessary to specify small intestine fate, overexpression studies suggest that it is not sufficien to specify this fate in the context of gene activity in the hindgut. At least two possible reasons can be imagined for this: either the level of bowl activity generated by the byn-GAL4 driver is not sufficient to overcome the repressive effect of lin, or another activity in addition to bowl (and present only in the anterior hindgut) is required to promote small intestine fate (Johansen, 2003b).
The drm-lin-bowl hierarchy that patterns the epithelial foregut and hindgut tubes functions by a distinctly different mechanism from the segmentation hierarchy that patterns the blastoderm embryo. During embryonic segmentation, a cascade of transcription factor-encoding genes (gap, pair-rule, and segment polarity) is sequentially expressed in more and more restricted domains. During gut development, in contrast, the activator Bowl and the repressor Lin are expressed throughout both foregut and hindgut; it is the spatially localized expression of Drm that is required for patterning of the distal foregut and hindgut (at the junction with the midgut). Drm, presumably by its direct binding to Lin (Green, 2002), relieves the repression of Bowl by Lin. The Drm-Lin-Bowl genetic hierarchy defined in this study is thus based, not on transcriptional regulatory interactions, but on protein-protein interactions (Johansen, 2003b).
Foregut and hindgut comprise the most distal portions of the gut tube, connecting the endodermal midgut to the exterior; both are ectodermal in origin, arising from the invagination of the stomodeum and proctodeum, respectively. Although the proventriculus undergoes a different type of morphogenesis than the hindgut (evagination, folding, involution, and insertion, versus elongation by cell rearrangement), it is significant that the drm-lin-bowl hierarchy in the hindgut obtains in the foregut as well. The drm-lin-bowl hierarchy acts in both ectodermal gut tissues to specify a small domain at the most interior position of the tube (i.e., the most posterior of the foregut and the most anterior of the hindgut) (Johansen, 2003b).
Specification of these domains results in the localized expression of upd, encoding the Drosophila JAK/STAT pathway ligand. In the hindgut, upd is required for morphogenetic behavior (rearrangement) of cells distant from the site of its expression, suggesting that a gradient of Upd may orient cell rearrangement (Johansen, 2003a). Since upd is required for proventriculus formation, and a domain of Stat92E expression is observed in the foregut extending beyond the domain of localized upd expression, there may also be a gradient of Upd in the foregut that plays a role in the morphogenetic processes that form the proventriculus (Johansen, 2003b).
These results support the proposal that foregut and hindgut morphogenesis are controlled in parallel by the some of the same genes. In addition to the drm-lin-bowl-upd pathway delineated here, the transcription factor encoded by forkhead (fkh) is required for expression of wg, hh, and decapentaplegic (dpp), which each contribute to aspects of both foregut and hindgut morphogenesis. Thus, in Drosophila, a number of molecular pathways are similarly deployed to promote both fore- and hindgut morphogenesis; some of this pathway conservation may extend to other organisms (Johansen, 2003b).
The one-to-one correspondence among each of the four members of the Drosophila and Anopheles (both members of the order Diptera) Odd family proteins, indicates that the four members were present prior to the divergence of the suborder Cyclorrhapha (which include Drosophila) from the suborder Nematocera (which include Anopheles), approximately 250 MYA. Drosophila chromosome 2L and Anopheles chromosome 3R, on which the odd family of each species is located, respectively, are the most conserved pair of chromosome arms between the two species. Based on their high sequence similarity to Drosophila drm and bowl, the Anopheles gambiae orthologs are likely to control similar developmental processes, in particular, gut morphogenesis (Johansen, 2003b and references therein).
odd-like genes are present in mammals and in Ciona intestinalis, a hemichordate. The mouse Osr2 and human OSR1 proteins display 65% and 70% respective identity to Drosophila Odd, Sob, and Bowl over their first three zinc fingers. The Ciona Odd-family protein, with two zinc fingers, has approximately 85% identity to the first and second zinc fingers of Drm, Sob, Odd, and Bowl. Intriguingly, human OSR1 is expressed in the adult colon, the mammalian equivalent to the Drosophila hindgut, and the Ciona homolog shows expression in the esophagus of the young adult. An important question to be addressed, therefore, is whether mammalian and Ciona odd-like genes also play roles in gut development (Johansen, 2003b).
The presence of lin-like genes in other genomes suggests that the interaction between odd family members and lin demonstrated for Drosophila may also obtain in other organisms. It is striking that the Lin ortholog in Anopheles is 45% identical overall, and 76% identical over 165 amino acids, to Drosophila Lin. Since each of the four members of the Anopheles odd family has an ortholog in the Drosophila odd family, the epistatic relationships observed between Drosophila drm, lin, and bowl are also likely to obtain in Anopheles. It is proposed that drm, lin, and bowl may play similar roles in Drosophila and mosquito gut development; analysis of expression of the Anopheles orthologs will be an important step toward testing this hypothesis. The human WINS1 and mouse Wins2 proteins show a much lower, although significant, 29 and 27% respective identity to Drosophila Lin. Since these genomes do not contain 1:1 orthologs of each of the odd family members, it is unlikely that there is a drm-lin-bowl pathway in these species. Nevertheless, interactions between Osr1/2 and Lin proteins may very well play important roles in embryonic patterning and morphogenesis (Johansen, 2003b).
Reference names in red indicate recommended papers.
Buckley, M. S., Chau, J., Hoppe, P. E. and Coulter, D. E. (2004). odd-skipped homologs function during gut development in C. elegans. Dev. Genes Evol. 214(1): 10-18. 14648222
de Celis Ibeas, J. M. and Bray. S. J. (2003). Bowl is required downstream of Notch for elaboration of distal limb patterning. Development 130: 5943-5952. 14573519
Goldstein, R. E., et al. (2005). An eh1-like motif in Odd-skipped mediates recruitment of Groucho and repression in vivo. Mol. Cell. Biol. 25(24): 10711-20. 16314497
Green, R. B., et al. (2002). Drumstick is a zinc finger protein that antagonizes Lines to control patterning and morphogenesis of the Drosophila hindgut. Development 129: 3645-3656. 12117814
Hao, I, Green, R. B., Dunaevsky, O. Lengyel, J. A. and Rauskolb, C. (2003). The odd-skipped family of zinc finger genes promotes Drosophila leg segmentation. Dev. Biol. 263: 282-295 14597202
Hart, Wang. L. and Coulter, D. E. (1996). Comparison of the structure and expression of odd-skipped and two related genes that encode a new family of zinc finger proteins in Drosophila. Genetics 144: 171-182. 8878683
Hatini, V., et al. (2000). Tissue- and stage-specific modulation of Wingless signaling by the segment polarity gene lines. Genes Dev. 14: 1364-1376.
Hatini, V., Green, R. B., Lengyel, J. A., Bray, S. J. and Dinardo, S. (2005). The Drumstick/Lines/Bowl regulatory pathway links antagonistic Hedgehog and Wingless signaling inputs to epidermal cell differentiation. Genes Dev. 19(6): 709-18. 15769943
Iwaki, D. D., et al. (2001). drumstick, bowl, and lines are required for patterning and cell rearrangement in the Drosophila embryonic hindgut. Dev. Biol. 240(2): 611-26. 11784087
Iwaki, D. D. and Lengyel, J. A. (2002). A Delta-Notch signaling border regulated by engrailed/invected repression specifies boundary cells in the Drosophila hindgut. Mech. Dev. 114: 71-84. 12175491
Johansen, K. A., Iwaki, D. D. and Lengyel, J. A. (2003a). Localized JAK/STAT signaling is required for oriented cell rearrangement in a tubular epithelium. Development 130: 135-145. 12441298
Johansen, K. A., Green, R. B., Iwaki, D. D., Hernandez, J. B. and Lengyel, J. A. (2003b). The Drm-Bowl-Lin relief-of-repression hierarchy controls fore- and hindgut patterning and morphogenesis. Mech. Dev. 120(10): 1139-51. 14568103
Mirth, C. and Akam, M. (2002). Joint development in the Drosophila leg: cell movements and cell populations. Dev. Biol. 246: 391-406. 12051824
Rauskolb, C. and Irvine, K. D. (1999). Notch-mediated segmentation and growth control of the Drosophila leg. Dev. Biol. 210: 339-350. 10357895
Rauskolb, C. (2001). The establishment of segmentation in the Drosophila leg. Development 128: 4511-4521. 11714676
Wang, L. and Coulter, D. E. (1996). bowel, an odd-skipped homolog, functions in the terminal pathway during Drosophila embryogenesis. EMBO J. 15: 3182-3196. 8670819
date revised: 25 July 2005
Home page: The Interactive Fly
© 2003 Thomas B. Brody, Ph.D.
The Interactive Fly resides on the
Society for Developmental Biology's Web server.