The secreted proteins Wingless and Hedgehog are essential to the elaboration of the denticle pattern in the epidermis of Drosophila embryos. We show that signaling by Wingless and Hedgehog regulates the expression of veinlet (rhomboid) and Serrate, two genes expressed in prospective denticle belts. Thus, Serrate and veinlet (rhom) partake in the last layer of the segmentation cascade. Ultimately, Wingless, Hedgehog, Veinlet (an indirect activator of the Egfr) and Serrate (an activator of Notch) are expressed in non-overlapping narrow stripes. The interface between any two stripes allows a reliable prediction of individual denticle types and polarity suggesting that contact-dependent signaling modulates individual cell fates. Attributes of a morphogen can be ascribed to Hedgehog in this system. However, no single morphogen organises the whole denticle pattern.

The embryonic epidermis of Drosophila provides a unique system for studying pattern formation at single cell resolution. At the end of embryonic development, the ventral epidermis is decorated by a segmentally repeated denticle pattern (DiNardo et al., 1994). One segmental unit is about 11 cells wide. It comprises 6 rows of cells that secrete denticles while the remaining cells make bald cuticle. Patterning occurs progressively from the deployment of coarsely expressed gap genes to the more refined segment polarity genes (reviewed by Ingham, 1988). Segment polarity genes are known to regulate each other’s expression. Most familiar is the non-cell autonomous feedback loop that links the expression of wingless and engrailed (or hedgehog) (DiNardo et al., 1988; Martinez-Arias et al., 1988). In addition, since they are at the bottom of the segmentation cascade, segment polarity gene products are believed to contribute to the specification of individual pattern elements (Bejsovec and Wieschaus, 1993). So far, only the bald fate is well understood: it arises because shavenbaby is repressed in response to Wingless signaling (Payre et al., 1999). Individual denticles (numbered 1 to 6) are distinguishable by their shape, size and polarity (Fig. 1A). The precision and resolution of this pattern imply sophisticated mechanisms of positional information. What are those mechanisms and how are they regulated by segment polarity genes?

Fig. 1.

Wild-type denticle pattern and fine structure of veinlet (rhom) expression. (A) Denticle belt from the third abdominal segment. Note the stereotypical shape and polarity of denticles in individual rows, numbered 1 to 6. (B) Expression of veinlet (rhom) at stage 13 as revealed by in situ hybridisation. (C) Embryo carrying a veinlet-nuclear lacZ enhancer trap gene triple stained for actin and Fasciclin 3 in red and anti-β-galactosidase in green. Actin bundles form in the middle of the apical surface. At the anterior (left) one row of bundles can be seen in β-galactosidase-negative cells while the second row of bundles corresponds to the most anterior veinlet (rhom)-expressing cells. Fat actin bundles (anticipating row 5 denticles) can be recognised. These are just posterior to (outside of) the veinlet (rhom) expression domain. Note the cells’ elongated aspect along the dorsoventral axis. (D) First instar veinlet-nuclear lacZ larva stained with X-gal. The domain of veinlet (rhom) expression corresponds to denticle types 2 to 4 with weak expressing cells underlying type 3 denticles (arrowheads).

Fig. 1.

Wild-type denticle pattern and fine structure of veinlet (rhom) expression. (A) Denticle belt from the third abdominal segment. Note the stereotypical shape and polarity of denticles in individual rows, numbered 1 to 6. (B) Expression of veinlet (rhom) at stage 13 as revealed by in situ hybridisation. (C) Embryo carrying a veinlet-nuclear lacZ enhancer trap gene triple stained for actin and Fasciclin 3 in red and anti-β-galactosidase in green. Actin bundles form in the middle of the apical surface. At the anterior (left) one row of bundles can be seen in β-galactosidase-negative cells while the second row of bundles corresponds to the most anterior veinlet (rhom)-expressing cells. Fat actin bundles (anticipating row 5 denticles) can be recognised. These are just posterior to (outside of) the veinlet (rhom) expression domain. Note the cells’ elongated aspect along the dorsoventral axis. (D) First instar veinlet-nuclear lacZ larva stained with X-gal. The domain of veinlet (rhom) expression corresponds to denticle types 2 to 4 with weak expressing cells underlying type 3 denticles (arrowheads).

Studies of the wing imaginal disks of Drosophila have led to a general model of pattern formation (reviewed by Lawrence and Struhl, 1996). In essence, at the interface between the anterior and posterior compartments, a morphogen, the TGF-β homolog Dpp, is produced and spreads across the entire disk to specify different cell fates at different thresholds of activity. The boundary separating the anterior and posterior compartments (the compartment boundary) is a clonal boundary established during embryogenesis (Garcia-Bellido et al., 1973; Vincent and O’Farrell, 1992). In the embryo, this boundary is called the parasegment border (Martinez-Arias and Lawrence, 1985) and is marked by the expression of wingless at the anterior and engrailed (along with hedgehog) at the posterior. By analogy with the disk system, this cellular interface could be a source of morphogen (possibly Hedgehog or Wingless itself) that organizes the whole segmental pattern. In support for such a notion, in the absence of both signals, segmental organization and denticle diversity are seemingly lost (Bejsovec and Wieschaus, 1993). However, this loss of segmental patterning could equally be explained if wingless (wg) and hedgehog (hh) were at the top of a genetic hierarchy leading to ever more refined expression of short-range signals (Martinez-Arias et al., 1988). Indeed, additional signaling pathways such as those of Epidermal growth factor receptor (Egfr) and Notch are activated in the epidermis of the late embryo.

Patterned activation of Egfr in the embryonic epidermis is primarily modulated by veinlet, also known as rhomboid. A current hypothesis is that Veinlet (rhom), a 7-transmembrane protein, regulates the cleavage of Spitz (a TGF-α homolog) and turns it into a soluble active ligand of Egfr (Schweitzer et al., 1995). In the ventral ectoderm of late embryos, veinlet (rhom) is expressed in stripes at the posterior of the engrailed domain, a region due to make denticles (Bier et al., 1990; O’Keefe et al., 1997; Szüts et al., 1997). Since Veinlet (rhom) activates Spitz and since Spitz is expressed everywhere (Rutledge et al., 1992), the stripes of veinlet (rhom) expression are believed to be a source of active Spitz. Consistent with a role for Egfr signaling in denticle formation, veinlet (rhom) contributes to the formation of denticle types 1-4 (Wiellette and McGinnis, 1999) and Spitz has even been proposed to specify different denticle types in a concentration-dependent manner (Szüts et al., 1997). The role of Notch signaling in denticle patterning was suggested more recently by the finding that mutants in Serrate, a gene that encodes a Notch ligand, have denticle defects (Wiellette and McGinnis, 1999). Moreover, Serrate is also expressed within prospective denticle belts, in a pattern that does not overlap with that of veinlet (rhom) expression (Wiellette and McGinnis, 1999). Since both veinlet (rhom) and Serrate are involved in denticle patterning, we have investigated the regulation of their expression by Hedgehog and Wingless, the two signals originating from the parasegment boundary.

It is already known that Wingless signaling represses veinlet (rhom) expression (Sanson et al., 1999; Gritzan et al., 1999). We now find that Hedgehog and Serrate both contribute positively to veinlet (rhom) expression while Serrate is negatively regulated by both Hedgehog and Wingless. Thus, at the time when cuticular fates are specified, 4 molecules activating 4 signaling pathways (Wingless, Hedgehog, Egfr and Notch) are expressed in a precise spatial pattern within an 8-cell-wide band. We argue that juxtacrine (contact-dependent) interactions mediated by these pathways can account for all denticle types. Nevertheless, we find that, in this system, Hedgehog has differential quantitative effects across three cell diameters. Wingless too acts over several cell diameters, although not in a concentration-dependent manner within the realm of our assays.

Drosophila stocks

wgcx4 (Baker, 1988), hhAC (Lee et al., 1992), ptc9 (Hooper and Scott, 1989), Df(2R)enE (Tabata et al., 1995) and serRX106 (obtained from Sarah Bray, Department of Anatomy, Cambridge) are presumed null alleles (see Flybase at http://gin.ebi.ac.uk:7081/). The serRX106hhAC recombinant was created by standard genetic methods and the wgcx4Df(2R)enE recombinant was obtained from Peter Lawrence (LMB, Cambridge).

The following transgenic stocks were used: armadillo-Gal411 (Sanson et al., 1996), engrailed-Gal4 (gift from Andrea Brand, Wellcome Institute, Cambridge, UK), tubulin-Gal4 (Pignoni and Zipursky, 1997), UAS-Ci (Alexandre et al., 1996), UAS-CiN/Zn, which we call UAS-Ci[repressor] (Hepker et al., 1997), UAS-Nintra (gift from M. Haenlin, CNRS Toulouse), veinlet-lacZ (line X81; Freeman et al., 1992), UAS-armS10 (Pai et al., 1997).

Genotypes of analysed embryos are as follow: tubulin-Gal4/UAS-Ci (Fig. 2B); wgcx4/wgcx4; tubulin-Gal4/UAS-Ci (Fig. 2C); armadillo-Gal411/UAS-Nintra (Fig. 3B and 3C); hhACserRX106/hhACserRX106 (Fig. 3D); armadillo-Gal411/UAS-Ci (Fig. 4B); armadillo-Gal411/UAS-armS10 (Fig. 4C); hhAC/hhAC (Fig. 4D), hhACftz-lacZ/hhACftz-lacZ (Fig. 4E); wgcx4/wgcx4; ftz-lacZ engrailed-Gal4/UAS-Ci[repressor] (Fig. 5A-C); wgcx4/wgcx4 (Fig. 6A,B); ptc9/ptc9 (Fig. 6C,D); wgcx4Df(2R)enE/wgcx4Df(2R)enE (Fig. 6E,F).

Fig. 2.

Hedgehog signaling activates veinlet (rhom) expression. (A) Wild-type stage 13 embryos stained with anti-Engrailed in ochre and a veinlet (rhom) RNA probe in purple. The weak-staining anterior veinlet (rhom) substripe abuts the domain of Engrailed expression, the source of secreted Hedgehog. The segment boundary (marked with an arrowhead in the cartoon) lies at the interface between these cell rows. Bracket indicates the ventralmost cells, whose behavior is the object of this paper. (B) Pattern of veinlet (rhom) expression in a stage 13 embryo expressing Ci uniformly (driven by the tubulin-Gal4 driver). The domain of veinlet (rhom) expression is widened. Anterior edge of the expanded veinlet (rhom) domain still corresponds to the segmental grooves (arrowhead in diagram) implying that widening is towards the posterior. (C) Overexpression of Ci in a wingless mutant. This results in widespread veinlet (rhom) expression.

Fig. 2.

Hedgehog signaling activates veinlet (rhom) expression. (A) Wild-type stage 13 embryos stained with anti-Engrailed in ochre and a veinlet (rhom) RNA probe in purple. The weak-staining anterior veinlet (rhom) substripe abuts the domain of Engrailed expression, the source of secreted Hedgehog. The segment boundary (marked with an arrowhead in the cartoon) lies at the interface between these cell rows. Bracket indicates the ventralmost cells, whose behavior is the object of this paper. (B) Pattern of veinlet (rhom) expression in a stage 13 embryo expressing Ci uniformly (driven by the tubulin-Gal4 driver). The domain of veinlet (rhom) expression is widened. Anterior edge of the expanded veinlet (rhom) domain still corresponds to the segmental grooves (arrowhead in diagram) implying that widening is towards the posterior. (C) Overexpression of Ci in a wingless mutant. This results in widespread veinlet (rhom) expression.

Fig. 3.

Notch signaling activates veinlet (rhom) expression. (A) Wild-type stage 13 embryos stained with a veinlet (rhom) RNA probe in black and a Serrate probe in red. The Serrate domain is at the posterior of the veinlet (rhom) domain. (B) Ubiquitous activation of the Notch pathway from uniform expression of Nintra leads to an increase of veinlet (rhom) expression. Serrate expression is seemingly unchanged and is still at the posterior of the increased veinlet (rhom) domain. (C) Embryos of the same genotype as in B, stained with anti-Engrailed in ochre and a veinlet (rhom) RNA probe in purple. This shows that veinlet (rhom) expression is still bounded by Engrailed when the Notch pathway is overactivated. In both B and C, the domain of veinlet (rhom) expression contain more cells despite being bounded by Engrailed and Serrate as in the wild-type. Maybe an extra cell division has occurred in this domain. (D) Lack of veinlet (rhom) expression in a hhser double mutant. No expression is detectable in the ectoderm although there may be weak expression in underlying tissues.

Fig. 3.

Notch signaling activates veinlet (rhom) expression. (A) Wild-type stage 13 embryos stained with a veinlet (rhom) RNA probe in black and a Serrate probe in red. The Serrate domain is at the posterior of the veinlet (rhom) domain. (B) Ubiquitous activation of the Notch pathway from uniform expression of Nintra leads to an increase of veinlet (rhom) expression. Serrate expression is seemingly unchanged and is still at the posterior of the increased veinlet (rhom) domain. (C) Embryos of the same genotype as in B, stained with anti-Engrailed in ochre and a veinlet (rhom) RNA probe in purple. This shows that veinlet (rhom) expression is still bounded by Engrailed when the Notch pathway is overactivated. In both B and C, the domain of veinlet (rhom) expression contain more cells despite being bounded by Engrailed and Serrate as in the wild-type. Maybe an extra cell division has occurred in this domain. (D) Lack of veinlet (rhom) expression in a hhser double mutant. No expression is detectable in the ectoderm although there may be weak expression in underlying tissues.

Fig. 4.

Expression of Serrate is negatively regulated by Wingless, Hedgehog and possibly Engrailed. (A) Expression of Serrate in the wild-type at stage 11 (detected with an RNA probe). In the ventral region, Serrate is expressed in segmental stripes. Lateral expression (at the edge of the picture) corresponds to the tracheal pits. (B) Overactivation of the Hedgehog pathway by uniform expression of Ci (with armadillo-Gal4) eradicates ventral expression. Expression in the tracheal pits is unaffected. (C) Repression of Serrate is also seen when the Wingless pathway is uniformly activated as in embryos carrying armadillo-Gal4 and UAS-ArmadilloS10 (an activated form of Armadillo; Pai et al., 1997). (D) Stage 13 hh embryos stained with a Serrate RNA probe in red and a veinlet (rhom) probe in purple. Serrate expression is clearly expanded compared to the wild-type and veinlet (rhom) expression nearly fills the region where Serrate is not expressed. Occasional islands where neither gene is expressed are always located at the anterior of the veinlet (rhom) stripe (arrow). They may reflect a long lasting repressive influence from Wingless expressed early. For the sake of completeness, we mention that extensive cell death occurs in segment polarity mutants (Pazdera et al., 1998) and this may contribute to the ultimate expression pattern. (E) Expression of β-galactosidase (brown) and Serrate (purple) in a stage 12 hedgehog mutant embryo carrying a ftz-lacZ transgene. The posterior edge of the expanded Serrate domain aligns with the anterior edge of the ftz-lacZ domain. This is known to correspond to the parasegment boundary and the anterior edge of the Engrailed domain in alternate segments (Lawrence and Johnston, 1989). Schematic representation of this result is shown in the diagram. F: Expression of β-galactosidase (brown) and Serrate (purple) in a stage 12 wingless mutant embryo carrying a ftz-lacZ transgene. The posterior edge of the expanded Serrate domain just falls short of the anterior edge of the ftz-lacZ domain (white arrow). This suggests the presence of residual repressive activity originating from the posterior compartment. Diagram shows schematically the expression of all relevant genes: veinlet (rhom) expression flanks the domain of extinct Engrailed expression (Sanson et al., 1999) and Serrate expansion is more limited than in a hedgehog mutant (Wiellette and McGinnis, 1999).

Fig. 4.

Expression of Serrate is negatively regulated by Wingless, Hedgehog and possibly Engrailed. (A) Expression of Serrate in the wild-type at stage 11 (detected with an RNA probe). In the ventral region, Serrate is expressed in segmental stripes. Lateral expression (at the edge of the picture) corresponds to the tracheal pits. (B) Overactivation of the Hedgehog pathway by uniform expression of Ci (with armadillo-Gal4) eradicates ventral expression. Expression in the tracheal pits is unaffected. (C) Repression of Serrate is also seen when the Wingless pathway is uniformly activated as in embryos carrying armadillo-Gal4 and UAS-ArmadilloS10 (an activated form of Armadillo; Pai et al., 1997). (D) Stage 13 hh embryos stained with a Serrate RNA probe in red and a veinlet (rhom) probe in purple. Serrate expression is clearly expanded compared to the wild-type and veinlet (rhom) expression nearly fills the region where Serrate is not expressed. Occasional islands where neither gene is expressed are always located at the anterior of the veinlet (rhom) stripe (arrow). They may reflect a long lasting repressive influence from Wingless expressed early. For the sake of completeness, we mention that extensive cell death occurs in segment polarity mutants (Pazdera et al., 1998) and this may contribute to the ultimate expression pattern. (E) Expression of β-galactosidase (brown) and Serrate (purple) in a stage 12 hedgehog mutant embryo carrying a ftz-lacZ transgene. The posterior edge of the expanded Serrate domain aligns with the anterior edge of the ftz-lacZ domain. This is known to correspond to the parasegment boundary and the anterior edge of the Engrailed domain in alternate segments (Lawrence and Johnston, 1989). Schematic representation of this result is shown in the diagram. F: Expression of β-galactosidase (brown) and Serrate (purple) in a stage 12 wingless mutant embryo carrying a ftz-lacZ transgene. The posterior edge of the expanded Serrate domain just falls short of the anterior edge of the ftz-lacZ domain (white arrow). This suggests the presence of residual repressive activity originating from the posterior compartment. Diagram shows schematically the expression of all relevant genes: veinlet (rhom) expression flanks the domain of extinct Engrailed expression (Sanson et al., 1999) and Serrate expansion is more limited than in a hedgehog mutant (Wiellette and McGinnis, 1999).

Fig. 5.

Late removal of hedgehog expression leads to a novel denticle pattern. (A) Expression of hedgehog in a stage 11 Ci[repressor] embryo (engrailed-Gal4 UAS-Ci[repressor). Expression is no longer detectable in the segmented ectoderm. This is consistent with the finding that Ci is a repressor of hedgehog expression in imaginal disks (Dominguez et al., 1996). Expression of engrailed is normal in such embryos (not shown). (B) Expression of veinlet (red) and Serrate (black) in an embryo of the same genotype as in A. The domain of veinlet (rhom) expression is reduced by one cell at the anterior and this is replaced by ectopic Serrate expression. (C) Denticle pattern of a Ci[repressor] larva (as in A,B) carrying a veinlet-nuclear lacZ enhancer trap gene. Staining with X-gal shows that the veinlet (rhom)-expressing cells give rise to type 4 denticles. Note that veinlet (rhom) expression is uniform in the two rows of expressing cells. Polarity reversals are clearly seen at the veinlet (rhom)/Serrate interfaces (ectopic and endogenous). This denticle phenotype is partially rescued by adding back extra Hedgehog from a transgene (not shown). This suggests that loss of hedgehog transcription is the principal cause of the phenotype but we cannot exclude the possibility that Ci[repressor] represses an additional factor.

Fig. 5.

Late removal of hedgehog expression leads to a novel denticle pattern. (A) Expression of hedgehog in a stage 11 Ci[repressor] embryo (engrailed-Gal4 UAS-Ci[repressor). Expression is no longer detectable in the segmented ectoderm. This is consistent with the finding that Ci is a repressor of hedgehog expression in imaginal disks (Dominguez et al., 1996). Expression of engrailed is normal in such embryos (not shown). (B) Expression of veinlet (red) and Serrate (black) in an embryo of the same genotype as in A. The domain of veinlet (rhom) expression is reduced by one cell at the anterior and this is replaced by ectopic Serrate expression. (C) Denticle pattern of a Ci[repressor] larva (as in A,B) carrying a veinlet-nuclear lacZ enhancer trap gene. Staining with X-gal shows that the veinlet (rhom)-expressing cells give rise to type 4 denticles. Note that veinlet (rhom) expression is uniform in the two rows of expressing cells. Polarity reversals are clearly seen at the veinlet (rhom)/Serrate interfaces (ectopic and endogenous). This denticle phenotype is partially rescued by adding back extra Hedgehog from a transgene (not shown). This suggests that loss of hedgehog transcription is the principal cause of the phenotype but we cannot exclude the possibility that Ci[repressor] represses an additional factor.

Fig. 6.

Correlation between gene expression and denticle pattern in three mutant situations. (A) Expression of Serrate (black) and veinlet (red) in a stage 13 wingless mutant. An endogenous Serrate/veinlet (rhom) interface is marked by a blue arrowhead. Expression of Serrate has expanded posteriorly and induced an ectopic veinlet (rhom) stripe there. Ectopic activation of veinlet (rhom) occurs because Wingless is no longer present to counter the induction of veinlet (rhom) by Serrate. An ectopic interface (white arrowhead) forms as a result. (B) High magnification view of a wg cuticle pattern. Different types of ordered denticle types can be seen. The presumed position of an endogenous veinlet (rhom)/Serrate interface is shown with a blue arrowhead as in A. (C) Expression of Engrailed (brown) and veinlet (purple) in a stage 13 patched mutant. Segmental grooves form at both the endogenous and the ectopic engrailed (hedgehog)/veinlet (rhom) interfaces. Both grooves are deeper than wild-type grooves. Many cells flanking the grooves are below the plane of focus and cannot be seen as a result. (D) Cuticle pattern of a patched mutant. The engrailed (hedgehog)/veinlet (rhom) interfaces correspond to polarity reversal at the anterior and the posterior of each belt. Although these reversals are not always perfect, they are completely penetrant. Cells in the middle of each belt make unusually shaped denticles (notice the round tip, red arrow). This (and also the deep grooves) may be a consequence of the overactivation of the Egfr pathway since veinlet (rhom) expression is stronger in patched mutants than in wild type siblings. (E) Expression of Serrate (red) and veinlet (black) in a wgen double mutant. Expression of Serrate is widespread except in segmentally repeated holes. These holes are lined with veinlet (rhom)-expressing cells. (F) Cuticle preparation of a wgen double mutant. Segmentally repeated denticle whorls are visible. Inset shows a whorl in side view, demonstrating the ‘bulging out’ effect.

Fig. 6.

Correlation between gene expression and denticle pattern in three mutant situations. (A) Expression of Serrate (black) and veinlet (red) in a stage 13 wingless mutant. An endogenous Serrate/veinlet (rhom) interface is marked by a blue arrowhead. Expression of Serrate has expanded posteriorly and induced an ectopic veinlet (rhom) stripe there. Ectopic activation of veinlet (rhom) occurs because Wingless is no longer present to counter the induction of veinlet (rhom) by Serrate. An ectopic interface (white arrowhead) forms as a result. (B) High magnification view of a wg cuticle pattern. Different types of ordered denticle types can be seen. The presumed position of an endogenous veinlet (rhom)/Serrate interface is shown with a blue arrowhead as in A. (C) Expression of Engrailed (brown) and veinlet (purple) in a stage 13 patched mutant. Segmental grooves form at both the endogenous and the ectopic engrailed (hedgehog)/veinlet (rhom) interfaces. Both grooves are deeper than wild-type grooves. Many cells flanking the grooves are below the plane of focus and cannot be seen as a result. (D) Cuticle pattern of a patched mutant. The engrailed (hedgehog)/veinlet (rhom) interfaces correspond to polarity reversal at the anterior and the posterior of each belt. Although these reversals are not always perfect, they are completely penetrant. Cells in the middle of each belt make unusually shaped denticles (notice the round tip, red arrow). This (and also the deep grooves) may be a consequence of the overactivation of the Egfr pathway since veinlet (rhom) expression is stronger in patched mutants than in wild type siblings. (E) Expression of Serrate (red) and veinlet (black) in a wgen double mutant. Expression of Serrate is widespread except in segmentally repeated holes. These holes are lined with veinlet (rhom)-expressing cells. (F) Cuticle preparation of a wgen double mutant. Segmentally repeated denticle whorls are visible. Inset shows a whorl in side view, demonstrating the ‘bulging out’ effect.

The ftz-LacZ transgene encodes a nuclear localised GFP-β-galactosidase fusion protein under the control of the complete fushi tarazu promoter. Details of the construction can be obtained upon request.

Embryo preparations

Except when anti-actin was used, immunofluorescence was done according to standard protocols (e.g. Vincent and O’Farrell, 1992) using AlexaTM fluorescent conjugates (Alexa 488 and Alexa 592, Molecular probes). For anti-actin staining (Fig. 1C), embryos were manually devitellinised to avoid methanol treatment, which destabilises the cytoskeleton. Antibodies used were rabbit anti-β-galactosidase (Cappell), mouse anti-β-galactosidase (Promega), rabbit anti-Engrailed (gift from C. H. Girdham and P. H. O’Farrell, UCSF), mouse anti-Fasciclin 3 (mAb 7G10 from Developmental Studies Hybridoma Bank), mouse anti-actin (mAb JLA20 from Developmental Studies Hybridoma Bank).

For RNA single and double in situ hybridisation, embryos were fixed and hybridised with digoxigenin-or fluorescein-labelled single-stranded RNA probe as described by Jowett (1997) except that fixed embryos were kept in 100% methanol and no proteinase K treatment was performed. The following cDNAs were used: Serrate (gift from E. Knust, Düsseldorf University, Germany), veinlet (rhom) (gift from J. F. de Celis, Cambridge University), engrailed (Poole et al., 1985) and hedgehog (gift from M. van den Heuvel, Oxford University).

For double labelling with an antibody and an RNA probe, the RNA in situ was performed first with the hybridization step performed at 58°C. After histochemical staining, embryos were fixed in 4% formaldehyde for 30 minutes and standard immunostaining was subsequently performed.

β-galactosidase activity was detected in cuticle preparations as described previously by Sanson et al. (1999).

For visualisation of the cuticle pattern, 24-hour-old embryos were mounted in Hoyer’s and photographed in DIC or phase-contrast microscopy.

Regulation of veinlet (rhom) expression

Experiments in this paper are concerned with gene expression and pattern formation in the ventralmost region of the Drosophila ectoderm. In the wild type, segmental expression of veinlet (rhom) commences at late stage 11 and is clearly established at stage 12 (not shown). We find that the structure of veinlet (rhom) stripes is subtler than previously described (Bier et al., 1990; Szüts et al., 1997; Wiellette and McGinnis, 1999). In situ hybridization with a veinlet (rhom) probe reveals expression in twinned substripes: a strong substripe at the posterior and a weaker one at the anterior (Fig. 1B). The substripes are separated by a single row of cells that do not stain. These ‘middle cells’ do (or did) nevertheless express veinlet (rhom) at low level since, in veinlet-lacZ embryos, they contain weak β-galactosidase activity (see below). To map more precisely the fine structure of the veinlet (rhom) domain onto specific cell fates, we used the finding that actin bundles serve as a template for denticles and thus anticipate the mature denticle pattern (Dickinson and Thatcher, 1997; Edwards et al., 1997). veinlet-lacZ embryos were triple stained with anti-actin to locate prospective denticles, anti-Fasciclin 3 to visualize cell boundaries and anti-β-galactosidase to detect veinlet (rhom) expression (Fig. 1C). From such preparations, we find that the second row of denticles (counting from the anterior) is secreted by the most anterior cells of each veinlet (rhom) stripe while the most posterior veinlet (rhom)-expressing cells make the fourth row of actin bundles. Squeezed between these two rows are the middle cells, which produce the third row of actin bundles. These weakly staining cells can still be detected at hatching and are seen to underlie type 3 denticles as expected (arrowheads in Fig. 1D). We now address how the twinned structure of the veinlet (rhom) stripes arises.

The anterior veinlet (rhom) substripe abuts the posterior compartment, the domain of engrailed expression and the source of Hedgehog in each segment (Fig. 2A). This component of veinlet (rhom) expression could therefore be activated by Hedgehog signaling. We found that, indeed, ectopic activation of the Hedgehog pathway by overexpression of the transcription factor Cubitus interruptus (Ci) leads to additional veinlet (rhom) expression. Embryos carrying the ubiquitous armadillo-Gal4 driver and the UAS-Ci responder have expanded domains of veinlet (rhom) expression (Fig. 2B). Expansion is towards the posterior since the expanded stripes are still bounded by segmental grooves at the anterior. The domain where veinlet (rhom) is not activated corresponds to the engrailed-expressing cells and the zone of Wingless influence. This suggests that Engrailed itself could possibly repress veinlet (rhom). Repression by Wingless signaling has already been shown (Sanson et al., 1999; Gritzan et al., 1999). As expected, therefore, uniform Ci expression in a wingless mutant leads to widespread veinlet (rhom) expression (Fig. 2C). This confirms that activation of the Hedgehog pathway is sufficient to induce veinlet (rhom) expression except where the Wingless pathway is active (and also where Engrailed is present). However, veinlet (rhom) expression is still detected in the ectoderm of hedgehog mutants (Sanson et al., 1999) suggesting that an additional signal operates during normal development.

One such additional regulator of veinlet (rhom) expression is Serrate (Wiellette and McGinnis, 1999), a protein expressed behind each veinlet (rhom) domain right against the posterior substripe (Fig. 3A). In the absence of Serrate, veinlet (rhom) stripes become thinner with expression being lost at the posterior (Wiellette and McGinnis, 1999). Since Serrate is a known ligand of Notch, we asked whether ectopic activation of the Notch pathway is sufficient to activate veinlet (rhom) expression. We expressed Nintra (a ligand-independent form of Notch, Lieber et al., 1993; Struhl et al., 1993) uniformly and assayed veinlet (rhom) expression. Expression is increased, especially in the middle cells and twinned stripes are no longer resolved (compare Fig. 3B,C with A). However, the domain of veinlet (rhom) expression is still bounded by Engrailed at the anterior and Serrate at the posterior (Fig. 3B,C). Again, we presume that veinlet (rhom) is not activated everywhere because of the presence of antagonistic signals such as Wingless and possibly other factors (see below). Nevertheless, the Notch pathway has a positive effect on veinlet (rhom) expression implying that Notch may contribute to normal veinlet (rhom) expression. Most likely, expression at the posterior of each veinlet (rhom) substripe is activated by Serrate produced in posterior adjacent cells (see also Wiellette and McGinnis, 1999). Since expression of Serrate is detected earlier than that of veinlet (rhom), around early stage 11 (not shown), it appears that Serrate defines the posterior edge of veinlet (rhom) stripes. Note that veinlet (rhom) is not expressed in Serrate-expressing cells and that, therefore, activation of veinlet (rhom) by Serrate is strictly non-cell autonomous.

In hedgehog mutants, Serrate expression is still present and even expanded (Wiellette and McGinnis, 1999 and see below) and this could account for the remaining veinlet (rhom) expression in these embryos. If this were the case, one would expect that veinlet (rhom) would not be expressed in the absence of both hedgehog and Serrate. We therefore assayed veinlet (rhom) expression in hhser double mutants. No expression is detectable in the epidermis (Fig. 3D), confirming that Hedgehog and Serrate both contribute to the activation of veinlet (rhom) expression, Hedgehog at the anterior and Serrate at the posterior. Thus, veinlet (rhom) expression is positively regulated in a position-dependent manner by two distinct signals.

Regulation of Serrate expression

As we show above, Serrate contributes to veinlet (rhom) regulation. Since we would like to obtain a complete picture of how Wingless and Hedgehog (the two proteins secreted at the parasegment border) regulate veinlet (rhom) expression, we decided to investigate the effect of these two signaling pathways on Serrate expression. In the ventral ectoderm, Serrate is expressed in stripes (2-to 3-cell wide) located at the posterior of each veinlet (rhom) domain (Fig. 3A). Overactivation of either the Hedgehog or the Wingless pathway (by overexpressing the downstream effectors Ci or Armadillo) completely abrogates this expression (Fig. 4A-C). This result shows that, at least in an overexpression assay, signaling by Hedgehog or Wingless represses Serrate expression. We now test the validity of these regulatory relationships in mutant embryos.

Consistent with the repressive activity of overactivated Hedgehog signaling, Serrate expression is expanded in hedgehog mutants and only thin bands of non-expression remain (Wiellette and McGinnis, 1999; Gritzan et al., 1999). Double staining shows that these thin bands express veinlet (rhom) (Fig. 4D). Thus veinlet (rhom) expression flanks the expanded domain of Serrate expression and is most certainly activated by Serrate since hhser double mutants no longer express veinlet (rhom) (see above).

We mapped the expansion of serrate expression in hedgehog mutants using ftz-lacZ as a permanent marker of the parasegment boundary, which lines the anterior edge of the posterior compartment. With this landmark, we can see that, in hh, Serrate expression expands posteriorward all the way to the parasegment boundary (Fig. 4E). Expression of Serrate also expands anteriorly since only a thin band of non-expressing cells remain. We conclude therefore that both the anterior and the posterior edge of the Serrate domain are affected in hedgehog mutants. The cells that do not express Serrate (and express veinlet) in hh correspond to those that expressed Engrailed previously (in hh embryos, Engrailed decays progressively; Bejsovec and Wieschaus, 1993) and, therefore, Engrailed expression may have a repressive effect on the Serrate promoter.

In wingless mutants, Serrate expression also expands, again consistent with the repressive effect of Wingless. Expansion occurs to a lesser extent than in hedgehog mutants (Wiellette and McGinnis, 1999; Gritzan et al., 1999). Using the same ftz-lacZ transgene, we find that in wg embryos, the repressed state is maintained in stripes that include the domain of extinct Engrailed expression and flanking cells (Fig. 4F). These flanking cells are those that were exposed to Hedgehog during early development (hedgehog is briefly expressed in wingless mutants, Lee et al., 1992) and this may explain why they do not express Serrate (see Discussion).

Since expression of wingless and hedgehog is mutually dependent, the expansion of Serrate expression in either mutant could be the compounded outcome of losing both activities. (The overexpression experiments suggest that both do repress Serrate). Therefore, the relative contributions of each mutation in Serrate deregulation are not easily distinguished. We suggest that posterior expansion of Serrate expression in hh is due indirectly to the loss of Wingless expression because Hedgehog is not believed to act beyond the wingless-expressing cells (Fietz et al., 1995). Therefore, the range of Wingless defines the posterior edge of each Serrate stripe. Conversely, Wingless signaling is not operative at the posterior of Engrailed stripes (Sanson et al., 1999) and, therefore, it must be Hedgehog that prevents Serrate expression from creeping anteriorward. In other words, Hedgehog appears to specify the anterior limit of the Serrate domain. This implies that Hedgehog acts over three cell diameters, slightly more than is expected from its known range of action in embryos. This issue will be addressed in the Discussion.

Overall, it appears that Serrate expression is primarily regulated negatively and that its pattern is delimited by Wingless, Hedgehog and Engrailed.

Correlating gene expression with denticle patterns

In Serrate mutants, veinlet (rhom) expression wanes at the posterior of each stripe (Wiellette and McGinnis, 1999). By analogy, one might predict that, in hedgehog mutants, veinlet (rhom) expression would decay at the anterior of each stripe since Hedgehog activates the anterior veinlet (rhom) substripe. However, the phenotype of hedgehog mutants is not so straightforward because of the loss of wingless expression in these embryos. As shown above, loss of wingless results in the expansion of the Serrate domain and this in turn, leads to the activation of veinlet (rhom) expression in ectopic cells (see below). In order to assess the proximal role of Hedgehog, we sought a way to block its expression around stage 11 when the denticle pattern begins to be specified. This was achieved by expressing a truncated form of Ci (CiN/Zn, Hepker et al., 1997), which represses the expression of hedgehog (Fig. 5A). We hence refer to the UAS-CiN/Zn transgene as UAS-Ci[repressor]. In embryos of the genotype engrailed-Gal4 UAS-Ci[repressor] (which we call Ci[repressor] embryos), hedgehog is expressed normally until early stage 11 and then fades away. In these embryos, veinlet (rhom) expression decays at the anterior of each stripe as predicted and, in its place, Serrate expression appears. The latter observation confirms the negative effect of Hedgehog on Serrate expression. In addition, veinlet (rhom) expression is upregulated in the middle cells. This, we think, is because the middle cells are now in contact with Serrate-expressing cells and have a fully activated Notch pathway. The final outcome is that veinlet (rhom) is expressed in 2-cell-wide stripes at uniformly high intensity (Fig. 5B,C).

Importantly, Ci[repressor] embryos display a denticle pattern unlike any previously seen in mutants (Fig. 5C). In the middle of each belt, one clearly recognises a mirror-image 5, 4, 4, 5 pattern. This is flanked by small denticles pointing away from the belt. As seen with a veinlet-lacZ transgene, the type 4 denticles in these embryos correspond to the domain of veinlet (rhom) expression. Thus the ectopic Serrate/veinlet (rhom) interface corresponds to the formation of denticle types 5/4 at the anterior of each belt while types 4/5 still form at the normal veinlet (rhom)/Serrate interface. Note that this interface (whether endogenous or ectopic) is characterized by a polarity reversal. The identity of the small anterior-pointing denticles at the anterior of each belt is not readily apparent. On the basis of size and morphology, they could be either of type 1 or type 6. We suggest that they are of type 1 because, like type 1 denticles in the wild type, they express Engrailed. This was shown by mapping Engrailed expression in Ci[repressor] larvae with a lacZ reporter (not shown). Finally, it seems natural to assign the type 6 label to the small denticles at the posterior of Ci[repressor] denticle belts since gene expression there is the same as in the wild type.

In the previous paragraph, we argued that a veinlet (rhom)/Serrate interface corresponds to denticle types 4/5 with its attendant polarity reversal while Engrailed expression corresponds to type 1 (as suggested by Dougan and DiNardo, 1992; Lawrence et al., 1996). These correlations apply to both Ci[repressor] and wild-type embryos. Further inspection of the wild-type situation suggests additional relationships. First, the interface between cells expressing hedgehog (and engrailed) and veinlet (rhom) corresponds to types 1 and 2 and is also accompanied by a polarity reversal. Second, type 3 denticles are made in the middle of the veinlet (rhom) domain and may correspond to normal activation of the Egfr pathway without influence from another pathway. Finally, activation of the Wingless pathway is always associated with the absence of denticles (Bejsovec and Martinez Arias, 1991; Sanson et al., 1999; Payre et al., 1999).

We now ask if the relationships outlined above have any predictive value in various mutant conditions. We start with wingless mutants because they have been described as making only type 5 denticles (Bejsovec and Wieschaus, 1993) and yet the striped expression pattern of veinlet (rhom) and Serrate in these embryos (Fig. 6A) suggests that more pattern elements should form. We re-examined wg cuticle preparation at high magnification and found reproducible, ordered arrays of denticles of more than one type (see also Gritzan et al., 1999). In particular, types 4 and 5 are recognisable by their size, shape and polarity (Fig. 6B). This arrangement is expected from the veinlet (rhom)/Serrate interfaces seen at earlier embryonic stages (Fig. 6A). Since hedgehog (and engrailed) are no longer expressed in wingless mutants, no hedgehog (engrailed)/veinlet (rhom) interfaces exist and, consistent with this, denticles types 1 and 2 are not recognisable.

We now turn to a second mutation. The phenotype of patched mutants has been described as a mirror-image duplication with denticle types 1, 2, 3, 2, 1 in sequence (Martinez-Arias et al., 1988). Although this is broadly accurate, we find the middle denticles difficult to type. What is clear is the presence of 1/2 and 2/1 interfaces (Fig. 6D). Those are easy to interpret in light of the expression of all the relevant genes. First, no Serrate is expressed in these embryos (Wiellette and McGinnis, 1999). This is expected since loss of patched function leads to ubiquitous activation of the Hedgehog pathway (Ingham et al., 1991) and this, as we have shown, leads to repression of Serrate. As a consequence, no Serrate/veinlet (rhom) interface exist and therefore, no denticle types 4/5 form. However, ectopic Engrailed stripes appear (Martinez-Arias et al., 1988) and hence new hedgehog (engrailed)/veinlet (rhom) interfaces are created (Fig. 6C) and ectopic 2/1 types form at the posterior of each belt.

Finally, we look at wgen double mutants. Although these embryos are lacking much pattern (Lawrence et al., 1996), remaining whorls of polarity reversals are readily seen (Fig. 6F). These whorls are anticipated by holes in the otherwise uniform expression of Serrate (Fig. 6E). We suggest that these holes follow, at least in part, from a long-lasting repressive effect of Hedgehog, which is probably expressed fleetingly in these mutants (Bejsovec and Wieschaus, 1993). In these mutants, expression of veinlet (rhom) is in rings that are inscribed within the Serrate holes (Fig. 6E). This is to be expected since Serrate activates veinlet (rhom) expression non-autonomously. Note that, although one would intuitively assume that the rings of veinlet (rhom) expression are induced by an activity originating from within (such as factor X in Gritzan et al., 1999), it is the surrounding cells that produce the activator. As seen in Fig. 6F, the circular veinlet (rhom)/Serrate interfaces lead to local polarity reversal. We would like to suggest that type 5 denticles surround each circle while type 4 denticles line its inner face and type 3 fill the interior in response to Egfr signaling. However, although the shape of the various denticles is broadly consistent with this suggestion, we acknowledge that there is a subjective element to denticle typing. Note that, in wgen embryos, most denticles point towards the midline. This could be due to influences from the dorsoventral patterning system. We note in passing that denticle whorls bulge out from the epithelium (Fig. 6F, inset). This may reflect local changes in cell adhesion at veinlet (rhom)/Serrate interfaces. We are currently investigating this further.

A final layer in the segmentation cascade

In this paper, we show that veinlet (rhom) and Serrate are segmentation genes acting downstream of segment polarity genes and thus form an additional layer in the segmentation cascade initiated by gap and pair-rule genes. We have worked out how the localized expression of Wingless and Hedgehog at the parasegment boundary leads to expression of Serrate and veinlet (rhom) at specific positions within the segmental pattern (Fig. 7A-C). Both signaling pathways repress Serrate expression. Since both pathways are believed to activate transcription, we imagine that they activate the expression of a repressor of Serrate. In addition, Serrate may also be negatively regulated by the transcriptional repressor Engrailed. In contrast to Serrate, veinlet (rhom) is regulated both positively and negatively: it is repressed by Wingless (Sanson et al., 1999; Gritzan et al., 1999) and activated by Hedgehog (Fig. 2).

Fig. 7.

From Wingless and Hedgehog to the denticle pattern. (A) Summary of the regulatory interactions between hedgehog, wingless, Serrate and veinlet (rhom). (B) Spatial regulation of Serrate expression by Wingless and Hedgehog signaling and Engrailed. Dotted lines emanating from the Hedgehog cells reflect an early action of Hedgehog whose range may require cell movement and proliferation as described in the text. (C) Spatial regulation of veinlet (rhom) expression by Wingless, Hedgehog and Serrate signaling. Dotted arrows emanating from the Hedgehog cells indicate weak Hedgehog action beyond one cell diameter. (D) Apparent correlations between gene expression and specific cuticular structures. Further experiments are needed to confirm the correlations in parentheses.

Fig. 7.

From Wingless and Hedgehog to the denticle pattern. (A) Summary of the regulatory interactions between hedgehog, wingless, Serrate and veinlet (rhom). (B) Spatial regulation of Serrate expression by Wingless and Hedgehog signaling and Engrailed. Dotted lines emanating from the Hedgehog cells reflect an early action of Hedgehog whose range may require cell movement and proliferation as described in the text. (C) Spatial regulation of veinlet (rhom) expression by Wingless, Hedgehog and Serrate signaling. Dotted arrows emanating from the Hedgehog cells indicate weak Hedgehog action beyond one cell diameter. (D) Apparent correlations between gene expression and specific cuticular structures. Further experiments are needed to confirm the correlations in parentheses.

In addition to this vertical flow of information, regulatory interactions also exist between veinlet (rhom) and Serrate. At the least, Serrate activates veinlet (rhom) expression by way of the Notch pathway (Fig. 3). This effect is purely non-cell autonomous. In contrast, Serrate appears to repress veinlet (rhom) in a cell autonomous manner (indeed, in cells where it is expressed, Serrate represses the Notch pathway; Micchelli et al., 1997). However, it is also possible that whichever mechanism activates Serrate expression also represses veinlet (rhom) expression. This would explain why the expression of Serrate and veinlet (rhom) is always mutually exclusive.

The regulatory interactions summarised above are sufficient to explain the spatial pattern of both Serrate (Fig. 7B) and veinlet (rhom) (Fig. 7C) expression. Non-autonomous repression of Serrate by Wingless and Hedgehog ensures that Serrate is expressed in stripes. As we have shown, the spread of Wingless towards the anterior defines the posterior edge of the domain of Serrate expression. Similarly, the anterior edge of the Serrate domain appears to be specified over three cell diameters by Hedgehog (Fig. 7B) slightly further than expected since Hedgehog is thought to act only over 1-2 cells in Drosophila embryos (Fietz et al., 1995). Expression of veinlet (rhom) is activated by two different signals, Hedgehog at the anterior and Serrate at the posterior. Although Hedgehog signaling is symmetrical, it does not activate veinlet (rhom) expression anteriorly (blocked red arrow in Fig. 7C) because there, Wingless represses veinlet (rhom) expression. Likewise, Serrate activates veinlet (rhom) expression but only on one side because of unilateral repression by Wingless (blocked green arrow in Fig. 7C).

These interactions display a clear temporal hierarchy. The secreted molecules Hedgehog and Wingless are expressed first and where they do not reach, Serrate expression is subsequently allowed. At stage 11, Hedgehog and Serrate activates veinlet (rhom) expression in separate cells. Ultimately, this chain of interactions results in detailed patterns of gene expression.

Correlating signaling pathways with denticle type

Mapping the expression pattern of various genes onto the denticle pattern suggests simple correlations, which are listed in the Results and summarised in Fig. 7D.

These correlations have allowed us to see pattern where it was previously thought there was none, as in wingless mutants (Fig. 6B). We now believe that wingless mutants make denticle type 3, 4 and 5 and not exclusively type 5 as suggested by Bejsovec and Wieschaus (1993). The correlations provide a guide to understand various phenotypes such as those of patched mutants and wgen double mutants. In wgen double mutants, the correlation between gene expression and denticle type/polarity is particularly evident. As we have shown, expression of veinlet (rhom) is in circles surrounded by Serrate expression and this correlates with polarity reversal in the cuticle. Non-uniform gene expression shows that these embryos have more pattern than previously noted (Lawrence et al., 1996). For such embryos to be truly unpatterned, they would have to express Serrate uniformly as well as not express veinlet (rhom). This may occur in wgenhh triple mutants since they may not contain any repressor of Serrate. We presume that the converse situation (Serrate ‘off’ and veinlet (rhom) ‘on’ everywhere) would also lead to unpatterned embryos. We predict that this situation would prevail in wg ptcen triple mutants.

Although the correlations have good predictive value (Fig. 6), they suffer from several limitations. First, denticle shape does not necessarily reflect an integer value. Indeed, unambiguous typing is not always possible and exact denticle shapes vary from segment to segment. Second, causal relationships between the activation of a particular signaling pathway and a given denticle type still remain to be investigated. We expect that the various signaling pathways control cytoskeletal behavior, which in turn affects denticle shape and cell polarity. Local polarity reversals indicate that individual cells are able to locate the source of a particular signal, suggesting that subcellular signaling complexes control the cytoskeleton directly. Third, we cannot exclude the involvement of additional regulators. In particular, it is possible that redundant regulators of the Notch and Egfr pathway contribute to the choice of denticle type. These could include Vein (another Egfr ligand; Schnepp et al., 1996), Delta (a Notch ligand; Fehon et al., 1990) or possibly Fringe. vein is not required for embryogenesis (Schnepp et al., 1996) suggesting that it does not play an important role if any. Possible contributions from Delta to denticle patterning are not readily assessed because of Delta’s earlier action in neurogenesis (Lehmann et al., 1983). We are currently attempting to overcome this problem.

The role of morphogens in the denticle pattern

Our results show that no single morphogen organises the denticle pattern: patterning arises, at least initially, from the combined actions of Wingless and Hedgehog. We now discuss in turn to what extent either of these two factors have attributes of a morphogen in this system.

Wingless is clearly not involved in the specification of denticle types (or diversity) across each belt since it does not act in this region of the epidermis. If it did, veinlet (rhom) and Serrate would not be expressed because, as we have shown, they are both repressed by Wingless. Nevertheless, Wingless acts at a distance, over 3-to 5-cell diameters to set the boundaries of the Serrate expression domain and thus establishes conditions for subsequent juxtacrine signaling. Long-range Wingless action is also required for the asymmetric action of Serrate: Serrate does not activate veinlet (rhom) expression posteriorly because of the presence of Wingless there, 3-to 5-cell diameters from the site of wingless transcription. In this sense, Wingless modulates, at a distance, the outcome of local signaling. In neither of these activities is there evidence for concentration-dependent signaling. However, one cannot formally exclude the possibility that the specification of type 6 denticle requires low-level Wingless. Furthermore, the suggestion that Wingless is not a morphogen in the embryonic epidermis is at odds with studies of the first thoracic segment where various levels of Wingless signalling lead to the specification of distinct cuticular structures (Lawrence et al., 1996). Re-assessment of these phenotypes with early molecular markers might tell whether or not Wingless acts directly in a concentration-dependent manner in the embryonic epidermis.

The situation with Hedgehog is clearer since it has qualitatively distinct effects over a narrow strip of cells (Fig. 7B,C). It activates veinlet (rhom) expression in adjoining posterior cells while its repressive effect on Serrate expression extends over three cell diameters. This suggests that, at high level, Hedgehog activates veinlet (rhom) (near the source) while at both low and high levels it repress Serrate expression (further away from the source). In this sense, Hedgehog qualifies as a morphogen (as originally suggested by Heemskerk and DiNardo, 1994). Whether differential responses at different distances from the Hedgehog source reflect true concentration dependence remains to be assessed. We note here that the repressive effect of Hedgehog on Serrate expression might take place early in development since, in wingless mutants, hedgehog expression decays around stage 10 (Lee et al., 1992) and yet Serrate expression is still confined at the anterior (Fig. 4F). We suggest that early Hedgehog has a repressive effect on Serrate expression that lasts at least until stage 11, when veinlet (rhom) expression commences. It is therefore conceivable that the 3-cell-wide domain where Serrate is repressed at stage 11 originates by cell proliferation from a single row of cells that abut the Hedgehog source at early embryonic stages. According to this scenario, the effects of Hedgehog on Serrate and veinlet (rhom) expression would both be occurring over one cell diameter. The apparent difference in range would reflect the difference in timing between these two effects and the intervening proliferation. We are currently testing this model by assessing the activity of a membrane-tethered form of Hedgehog.

To sum up, in the bald area of abdominal segments, one cell type forms in response to one signaling pathway while within denticle belts, a rich pattern of cell types arise from juxtacrine cell interactions initiated by the activation of distinct signaling pathways. Some of these pathways are controlled by the localised expression of segment polarity genes such as wingless and hedgehog while others are regulated by downstream genes like veinlet (rhom) and Serrate. Because wingless and hedgehog are expressed first, they are effectively at the top of the hierarchy and the knock-on effects of losing hedgehog or wingless function explain the “organiser activity” of the parasegment boundary. Interestingly, the denticle pattern arises from the combined action of Wingless and Hedgehog originating from the parasegment boundaries of adjacent segments and therefore, two parasegment boundaries are needed to provide the signals that pattern a single denticle belt.

We thank Gary Struhl for pointing out an inconsistency in an early version of the Discussion. We are also grateful to Steve DiNardo, Peter Lawrence, Steve Marygold, Alex Gould and Niall Armes for discussions and comments on the manuscript. Friendly exchanges of information with Steve DiNardo are acknowledged. M. L. is the recipient of a post-doctoral fellowship from the European Molecular Biology Organization while C. A. and J.-P. V. are supported by the UK Medical Research Council.

Alexandre
,
C.
,
Jacinto
,
A.
and
Ingham
,
P. W.
(
1996
).
Transcriptional activation of hedgehog target genes in Drosophila is mediated directly by the Cubitus interruptus protein, a member of the GLI family of zinc finger DNA-binding proteins
.
Genes Dev
.
10
,
2003
2013
.
Baker
,
N. E.
(
1988
).
Embryonic and imaginal requirements for wingless, a segment-polarity gene in Drosophila
.
Dev. Biol
.
125
,
96
108
.
Bejsovec
,
A.
and
Martinez Arias
,
A.
(
1991
).
Roles of wingless in patterning the larval epidermis of Drosophila
.
Development
113
,
471
485
.
Bejsovec
,
A.
and
Wieschaus
,
E.
(
1993
).
Segment polarity gene interactions modulate epidermal patterning in Drosophila embryos
.
Development
119
,
501
517
.
Bier
,
E.
,
Jan
,
L. Y.
and
Jan
,
Y. N.
(
1990
).
rhomboid, a gene required for dorsoventral axis establishment and peripheral nervous system development in Drosophila melanogaster
.
Genes Dev
.
4
,
680
681
.
Dickinson
,
W. J.
and
Thatcher
,
J. W.
(
1997
).
Morphogenesis of denticles and hairs in Drosophila embryos: involvement of Actin-associated proteins that also affect adult structures
.
Cell Motility Cytoskel
.
38
,
9
21
.
DiNardo
,
S.
,
Sher
,
E.
,
Heemskerk-Jongens
,
J.
,
Kassis
,
J. A.
and
O’Farrell
,
P. H.
(
1988
).
Two-tiered regulation of spatially patterned engrailed gene expression during Drosophila embryogenesis
.
Nature
332
,
604
609
.
DiNardo
,
S.
,
Heemskerk
,
J.
,
Dougan
,
S.
and
O’Farrell
,
P. H.
(
1994
).
The making of a maggot: Patterning the Drosophila embryonic epidermis
.
Curr. Opin. Genet. Dev
.
4
,
529
534
.
Dominguez
,
M.
,
Brunner
,
M.
,
Hafen
,
E.
and
Basler
,
K.
(
1996
).
Sending and receiving the Hedgehog signal: Control by the Drosophila Gli protein Cubitus interruptus
.
Science
272
,
1621
1625
.
Dougan
,
S.
and
DiNardo
,
S.
(
1992
).
Drosophila wingless generates cell type diversity among engrailed expressing cells
.
Nature
360
,
347
350
.
Edwards
,
K. A.
,
Demsky
,
M.
,
Montague
,
R. A.
,
Weymouth
,
N.
and
Kiehart
,
D. P.
(
1997
).
GFP-Moesin illuminates actin cytoskeleton dynamics in living tissue and demonstrates cell shape changes during morphogenesis in Drosophila
.
Dev. Biol
.
191
,
103
117
.
Fehon
,
R. G.
,
Kooh
,
P. J.
,
Rebay
,
I.
,
Regan
,
C. L.
,
Xu
,
T.
,
Muskavitch
,
M. A.
and
Artavanis-Tsakonas
,
S.
(
1990
).
Molecular interactions between the protein products of the neurogenic loci notch and delta, two EGF-homologous genes in Drosophila
.
Cell
61
,
523
534
.
Fietz
,
M. J.
,
Jacinto
,
A.
,
Taylor
,
A. M.
,
Alexandre
,
C.
and
Ingham
,
P. W.
(
1995
).
Secretion of the amino-terminal fragment of the Hedgehog protein is necessary and sufficient for hedgehog signaling in Drosophila
.
Curr. Biol
.
5
,
643
650
.
Freeman
,
M.
,
Kimmel
,
B. E.
and
Rubin
,
G. M.
(
1992
).
Identifying targets of the rough homeobox gene of Drosophila: evidence that rhomboid functions in eye development
.
Development
116
,
335
346
.
Garcia-Bellido
,
A.
,
Ripoll
,
P.
and
Morata
,
G.
(
1973
).
Developmental compartmentalisation of the wing disc of Drosophila
.
Nature New Biol
.
245
,
251
253
.
Gritzan
,
U.
,
Hatini
,
V.
and
DiNardo
,
S.
(
1999
).
Mutual antagonism between signals secreted by adjacent Wingless and Engrailed cells leads to specification of complementary regions of the Drosophila parasegment
.
Development
126
,
4107
4115
Heemskerk
,
J.
and
DiNardo
,
S.
(
1994
).
Drosophila hedgehog acts as a morphogen in cellular patterning
.
Cell
76
,
449
460
.
Hepker
,
J.
,
Wang
,
Q. T.
,
Motzny
,
C. K.
,
Holmgren
,
R.
and
Orenic
,
T. V.
(
1997
).
Drosophila Cubitus interruptus forms a negative feedback loop with patched and regulates expression of hedgehog target genes
.
Development
124
,
549
558
.
Hooper
,
J. E.
and
Scott
,
M. P.
(
1989
).
The Drosophila patched gene encodes a putative membrane protein required for segmental patterning
.
Cell
59
,
751
765
.
Ingham
,
P. W.
(
1988
).
The molecular genetics of embryonic pattern formation in Drosophila
.
Nature
335
,
25
34
.
Ingham
,
P. W.
,
Taylor
,
A. M.
and
Nakano
,
Y.
(
1991
).
Role of the Drosophila patched gene in positional signaling
.
Nature
353
,
184
187
.
Jowett
,
T.
(
1997
).
Tissue In Situ Hybridization
.
Heidelberg
:
Wiley/Spektrum
.
Lawrence
,
P. A.
and
Johnston
,
P.
(
1989
).
Pattern formation in the Drosophila embryo: allocation of cells to parasegments by even-skipped and fushi tarazu
.
Development
105
,
761
767
.
Lawrence
,
P. A.
,
Sanson
,
B.
and
Vincent
,
J. P.
(
1996
).
tCompartments, wingless and engrailed: patterning the ventral epidermis of Drosophila embryos
.
Development
122
,
4095
4103
.
Lawrence
,
P. A.
and
Struhl
,
G.
(
1996
).
tMorphogens, compartments, and pattern: Lessons from Drosophila?
Cell
85
,
951
961
.
Lee
,
J. J.
,
Von Kessler
,
D.
,
Parks
,
S.
and
Beachy
,
P. A.
(
1992
).
Secretion and localized transcription suggest a role in positional signaling for products of the segmentation gene hedgehog
.
Cell
71
,
33
50
.
Lehmann
,
R.
,
Jimenez
,
F.
,
Dietrich
,
U.
and
Campos-Ortega
,
J. A.
(
1983
).
On the phenotype and development of mutants of early neurogenesis in Drosophila melanogaster
.
Roux Arch. Dev. Biol
.
192
,
62
74
.
Lieber
,
T.
,
Kidd
,
S.
,
Alcamo
,
E.
,
Corbin
,
V.
and
Young
,
M. W.
(
1993
).
Antineurogenic phenotypes induced by truncated Notch proteins indicate a role in signal transduction and may point to a novel function for Notch in nuclei
.
Genes Dev
.
7
,
1949
1965
.
Martinez-Arias
,
A.
and
Lawrence
,
P. A.
(
1985
).
Parasegments and compartments in the Drosophila embryo
.
Nature
313
,
639
642
.
Martinez-Arias
,
A.
,
Baker
,
N. E.
and
Ingham
,
P. W.
(
1988
).
Role of segment polarity genes in the definition and maintenance of cell states in the Drosophila embryo
.
Development
103
,
157
170
.
Micchelli
,
C. A.
,
Rulifson
,
E. J.
and
Blair
,
S. S.
(
1997
).
The function and regulation of cut expression on the wing margin of Drosophila: Notch, wingless and a dominant negative role for Delta and Serrate
.
Development
124
,
1485
1495
.
Nellen
,
D.
,
Burke
,
R.
,
Struhl
,
G.
and
Basler
,
K.
(
1996
).
Direct and long-range action of a Dpp morphogen gradient
.
Cell
85
,
357
368
.
O’Keefe
,
L.
,
Dougan
,
S. T.
,
Gabay
,
L.
,
Raz
,
E.
,
Shilo
,
B. Z.
and
DiNardo
,
S.
(
1997
).
Spitz and Wingless, emanating from distinct borders, cooperate to establish cell fate across the Engrailed domain in the Drosophila epidermis
.
Development
124
,
4837
4845
.
Pai
,
L.M.
,
Orsulic
,
S.
,
Bejsovec
,
A.
and
Peifer
,
M.
(
1997
).
Negative regulation of Armadillo, a Wingless effector in Drosophila
.
Development
124
,
2255
2266
.
Payre
,
F.
,
Vincent
,
A.
and
Carreno
,
S.
(
1999
).
ovo/svb integrates Wingless and DER pathways to control epidermis differentiation
.
Nature
400
,
271
275
.
Pazdera
,
T.
,
Janardhan
,
P.
and
Minden
,
J.
(
1997
).
Patterned epidermal cell death in wild-type and segment polarity mutant Drosophila embryos
.
Development
125
,
3427
3436
.
Pignoni
,
F.
and
Zipursky
,
S. L.
(
1997
).
Induction of Drosophila eye development by decapentaplegic
.
Development
124
,
271
278
.
Poole
,
S. J.
,
Kauvar
,
L. M.
,
Drees
,
B.
and
Kornberg
,
T.
(
1985
).
The engrailed locus of Drosophila: structural analysis of an embryonic transcript
.
Cell
40
,
37
43
.
Rutledge
,
B. J.
,
Zhang
,
K.
,
Bier
,
E.
,
Jan
,
Y. N.
and
Perrimon
,
N.
(
1992
).
The Drosophila spitz gene encodes a putative EGF-like growth factor involved in dorsal-ventral axis formation and neurogenesis
.
Genes Dev
.
6
,
1503
1517
.
Sanson
,
B.
,
White
,
P.
and
Vincent
,
J. P.
(
1996
).
Uncoupling Cadherin-based adhesion from Wingless signaling in Drosophila
.
Nature
383
,
627
630
.
Sanson
,
B.
,
Alexandre
,
C.
,
Fascetti
,
N.
and
Vincent
,
J.-P.
(
1999
).
Engrailed and Hedgehog make the range of Wingless asymmetric in Drosophila embryos
.
Cell
98
,
207
216
.
Schnepp
,
B.
,
Grumbling
,
G.
,
Donaldson
,
T.
and
Simcox
,
A.
(
1996
).
Vein is a novel component in the Drosophila epidermal growth factor receptor pathway with similarity to the neuregulins
.
Genes Dev
.
10
,
2302
2313
.
Schweitzer
,
R.
,
Shaharabany
,
M.
,
Seger
,
R.
and
Shilo
,
B. Z.
(
1995
).
Secreted Spitz triggers the DER signaling pathway and is a limiting component in embryonic ventral ectoderm determination
.
Genes Dev
.
9
,
1518
1529
.
Singer
,
M. A.
,
Penton
,
A.
,
Twombly
,
V.
,
Hoffmann
,
F. M.
and
Gelbart
,
W. M.
(
1997
).
Signaling through both type I DPP receptors is required for anterior-posterior patterning of the entire Drosophila wing
.
Development
124
,
79
89
.
Struhl
,
G.
,
Fitzgerald
,
K.
and
Greenwald
,
I.
(
1993
).
Intrinsic activity of the Lin-12 and Notch intracellular domains in vivo
.
Cell
74
,
331
345
.
Szüts
,
D.
,
Freeman
,
M.
and
Bienz
,
M.
(
1997
).
Antagonism between EGFR and Wingless signaling in the larval cuticle of Drosophila
.
Development
124
,
3209
3219
.
Tabata
,
T.
,
Schwartz
,
C.
,
Gustavson
,
E.
,
Ali
,
Z.
and
Kornberg
,
T. B.
(
1995
).
Creating a Drosophila wing de novo, the role of engrailed and the compartment border hypothesis
.
Development
121
,
3359
3369
.
Vincent
,
J. P.
and
O’Farrell
,
P. H.
(
1992
).
The state of engrailed expression is not clonally transmitted during early Drosophila development
.
Cell
68
,
923
931
.
Wiellette
,
E. L.
and
McGinnis
,
W.
(
1999
).
Hox genes differentially regulate Serrate to generate segment-specific structures
.
Development
126
,
1985
1995
.