The stereotyped pattern of veins in the Drosophila wing is generated in response to local EGF signalling. Mutations in the rhomboid (rho) gene, which encodes a sevenpass membrane protein required to enhance signalling transmitted by the EGF receptor (Egfr), inhibit vein development and disrupt the vein pattern. By contrast, net mutations produce ectopic veins in intervein regions. We have cloned the net gene and show that it encodes a basic HLH protein that probably acts as a transcriptional repressor. net and rho are expressed in mutually exclusive patterns during the development of the wing imaginal disc. Lack of net activity causes rho expression to expand, and vice versa. Furthermore, ectopic expression of net or rho results in their mutual repression and thus suppresses vein formation or generates tube-like wings composed of vein-like tissue. Egfr signalling and net exert mutually antagonising activities during the specification of vein versus intervein fate. While Egfr signalling represses net transcription, net exhibits a two-tiered control by repressing rho transcription and interfering with Egfr signalling downstream of Rho. Our results further suggest that net is required to maintain intervein development by restricting Egfr signalling, which promotes vein development, to the Net-free vein regions of the wing disc.

Wing primordia originate in the Drosophila embryo from cells straddling the parasegmental compartment boundary of the second thoracic segment, and are subdivided into anterior and posterior compartments by the engrailed (en) gene, which is active in the posterior compartment (Lawrence and Morata, 1976; Kornberg et al., 1985; Hidalgo, 1994). During larval stages, pattern formation along the anteroposterior axis of the wing disc is mediated by posterior hedgehog (Hh) signalling, which activates in adjacent anterior cells the decapentaplegic (dpp) gene whose secreted TGFβ-like product acts as morphogen for both compartments (Tabata and Kornberg, 1994; Tabata et al., 1995; Guillen et al., 1995; Lecuit et al., 1996; Nellen et al., 1996; Hidalgo, 1998). While this aspect of wing pattern formation has attracted considerable interest as a model system for the analysis of cell communication and the signalling pathways underlying pattern formation (Díaz-Benjumea and Cohen, 1993), comparatively little is known about the genes and mechanisms that are turned on by the molecular prepatterns in the wing disc and establish the differentiated structures within the wing blade.

Adult wing blades consist of two planar epithelial cell layers that originate from two distinct cell populations after a further subdivision of the wing disc into a dorsal and ventral compartment. Wings show a stereotyped pattern of five longitudinal and two transverse veins, separating distinct sectors of intervein regions. Genetic and molecular analysis has revealed that vein versus intervein development is initiated in larval wing discs by the activity of prepattern genes and refined by multiple tiers of intercellular signalling in prepupal discs and pupal wings (García-Bellido, 1977; García-Bellido and de Celis, 1992; Sturtevant et al., 1993, 1994, 1997; de Celis et al., 1995, 1996, 1997; Sturtevant and Bier, 1995; Yu et al., 1996; Milán et al., 1997; Biehs et al., 1998; Roch et al., 1998; Guichard et al., 1999). Vein and intervein cells differ in several aspects. Vein cells are small, densely packed, and covered by a thick, darkly pigmented cuticle, whereas intervein cells are large, flat, and covered by a thin, transparent and lightly pigmented cuticle. In contrast to dorsal and ventral vein cells, which form tubular structures filled with haemolymph, intervein cells on opposite sides of the wing are tightly connected through a dense extracellular matrix. Finally, vein cells remain alive throughout adulthood, whereas intervein cells die upon eclosion of the fly.

Because of the stereotyped wing pattern, subtle wing defects can be easily identified and therefore many mutants affecting the architecture or pattern of the wing have been isolated. One of these mutants, net, discovered by Bridges in 1931 (Lindsley and Zimm, 1992), displays ectopic wing veins and often blisters on the wing surface, which indicates that the cell layers of opposing wing surfaces are not properly connected in the intervein regions (compare Fig. 4A with 4B). net was shown to interact genetically with Notch (N) (García-Bellido and de Celis, 1992). N belongs to the class of neurogenic genes and encodes a transmembrane receptor that, like all products of neurogenic genes, serves to single out a cell from a group of equivalent cells by a signalling process termed lateral inhibition (Artavanis-Tsakonas et al., 1983, 1995; Campos-Ortega, 1997). In the wing, N signalling is necessary to restrict vein cell development from seven- to eight-cell wide domains of vein-competent provein cells to rows of two to three cells in width (García-Bellido and de Celis, 1992; Sturtevant and Bier, 1995). This process depends also on the rhomboid (rho) gene (de Celis et al., 1997), which encodes a sevenpass membrane protein (Bier et al., 1990) necessary to promote the extracellular activation of Egfr, e.g. by the EGF-related ligand Vein (Vn) (Schnepp et al., 1996; Simcox et al., 1996; Simcox, 1997), in a process that further depends on the transmembrane protein Star (S) (Sturtevant et al., 1993; Guichard et al., 1999).

The rho gene is initially expressed in a series of discrete stripes corresponding to the longitudinal vein primordia of third instar wing discs and is subsequently maintained in differentiating vein cells throughout pupal development (Sturtevant et al., 1993; Sturtevant and Bier, 1995). Since all other known members of the Egfr signalling pathway are expressed uniformly, localized rho activity is thought to provide the decisive clue for vein formation and to maintain vein cell fate. This conclusion is based on several observations: (1) absence of rho activity causes loss of veins, whereas ectopic rho activity results in the formation of ectopic veins in intervein regions (Sturtevant et al., 1993; Noll et al., 1994; Sturtevant and Bier, 1995); (2) in pupal wing discs of N mutants, rho expression expands into the intervein regions and causes a broadening of the veins (Sturtevant and Bier, 1995; de Celis et al., 1997); and (3) in Nts; net1 double mutant wing discs, rho expression covers entire intervein regions (Sturtevant et al., 1997). These results further demonstrate that N signalling restricts rho expression and thus Egfr signalling to cells initiating vein development and suggest that net activity is required to inhibit vein-promoting rho activity in intervein regions (de Celis et al., 1997).

Here we show that net encodes a novel basic helix-loop-helix (bHLH) protein that may act as a transcriptional repressor. In wing discs, net is expressed in intervein regions, in a pattern complementary to that of rho. Ectopic expression of rho completely represses net transcription, whereas net expression expands into primordial vein regions in the absence of rho activity. Conversely, ectopic net expression represses rho transcription in vein primodia, whereas rho expression expands into intervein regions in net deficient wing discs. Thus, the transcription patterns of the two genes are complementary and their activities are mutually exclusive. Our results suggest that net is repressed by Rho-dependent Egfr signalling while its product in turn functions as negative regulator of rho and the Egfr signalling pathway. Our results imply that Net prevents vein development by the inhibition of Egfr signalling in intervein regions and thereby maintains intervein development.

General procedures

Standard procedures, such as isolation and Southern blot analysis of genomic DNA, construction and screening of genomic libraries, chromosomal walking, isolation and northern blot analysis of poly(A)+ RNA and PCR, were carried out essentially as described (Frei et al., 1985; Kilchherr et al., 1986; Sambrook et al., 1989; Fu and Noll, 1997).

Isolation of genomic clones

To clone the net locus, a chromosomal walk was initiated from the genomic insert BPal-1 of a λ phage, which included the distal breakpoint of Df(2L)al (Schneitz et al., 1993), and extended for about 240 kb until it overlapped with genomic clones harbouring the lethal(2)giant larvae locus at the left telomere of the second chromosome (kindly provided by B. Mechler). The distal breakpoint of Df(2L)al is thus located about 300 kb from the telomere. To isolate the breakpoints of an inversion uncovering net and u-shaped (ush), a genomic library of the corresponding mutant stock In(2L)TE99(Z)XR93/CyO (generously provided by Pascal Heitzler) was prepared in λ DASH II and screened with the most proximal EcoRI fragment of Y38-4 (Fig. 1A). A breakpoint clone, XR93-11, which included the distal exon and 3′ downstream region of net, about 8 kb of TE99, and approximately 1 kb of the 5′ promoter region of the ush locus (Cubadda et al., 1997), and a wild-type clone of net from the CyO balancer chromosome, XR93-8 (Fig. 1A), were isolated.

Fig. 1.

Isolation and molecular characterization of the net locus. (A) Chromosomal walk covering the net locus. Three telomeric deficiencies Df(2L)U314, Df(2L)PM4, and Df(2L)net18 are mapped (filled boxes delimit the regions including the proximal breakpoints) with respect to a genomic EcoRI map (two EcoRI fragments whose order has not been determined are separated by a short vertical line) of the chromosomal region between 21A2 and 21B1. Below the EcoRI map, the distal portion of a 240 kb chromosomal walk (Y34-3 to Y40-8 are inserts isolated from a KrSB1/CyO library in EMBL4; XR93-8 is a wild-type DNA insert isolated from a In(2L)TE99(Z)XR93/CyO library in λ DASH II) is shown, which overlaps at its proximal end with a chromosomal walk including the region uncovered by Df(2L)al (Schneitz et al., 1993) and at its distal end with a 40 kb chromosomal walk including l(2)gl and the left telomere (Mechler et al.,1985). (B) Transcriptional organization of the net gene and map of mutations disrupting net. The locations of the B104 insertion netK1, the 8 kb insertion netK7 (mapped to the 2.0 kb HindIII fragment), netPM50, as well as of the small net1 duplication and of the distal breakpoint of the inversion In(2L)TE99(Z)XR93 uncovering net and ush (personal communication and kind gift from P. Heitzler), are shown with respect to an enlarged restriction map of the net gene (as defined by the three fragments indicated at the bottom, RI, RIIa and RIIb, that were tested as transgenes for rescue of the net mutant phenotype) and its intron/exon structure below (open reading frame in black). The netPM50 allele is a lost P-element insertion that was kindly provided by C. Caggese. H (HindIII), N (NheI), P(PstI), R (EcoRI), S (SalI), X (XhoI).

Fig. 1.

Isolation and molecular characterization of the net locus. (A) Chromosomal walk covering the net locus. Three telomeric deficiencies Df(2L)U314, Df(2L)PM4, and Df(2L)net18 are mapped (filled boxes delimit the regions including the proximal breakpoints) with respect to a genomic EcoRI map (two EcoRI fragments whose order has not been determined are separated by a short vertical line) of the chromosomal region between 21A2 and 21B1. Below the EcoRI map, the distal portion of a 240 kb chromosomal walk (Y34-3 to Y40-8 are inserts isolated from a KrSB1/CyO library in EMBL4; XR93-8 is a wild-type DNA insert isolated from a In(2L)TE99(Z)XR93/CyO library in λ DASH II) is shown, which overlaps at its proximal end with a chromosomal walk including the region uncovered by Df(2L)al (Schneitz et al., 1993) and at its distal end with a 40 kb chromosomal walk including l(2)gl and the left telomere (Mechler et al.,1985). (B) Transcriptional organization of the net gene and map of mutations disrupting net. The locations of the B104 insertion netK1, the 8 kb insertion netK7 (mapped to the 2.0 kb HindIII fragment), netPM50, as well as of the small net1 duplication and of the distal breakpoint of the inversion In(2L)TE99(Z)XR93 uncovering net and ush (personal communication and kind gift from P. Heitzler), are shown with respect to an enlarged restriction map of the net gene (as defined by the three fragments indicated at the bottom, RI, RIIa and RIIb, that were tested as transgenes for rescue of the net mutant phenotype) and its intron/exon structure below (open reading frame in black). The netPM50 allele is a lost P-element insertion that was kindly provided by C. Caggese. H (HindIII), N (NheI), P(PstI), R (EcoRI), S (SalI), X (XhoI).

Mapping of net rearrangements

The breakpoint of the inversion In(2L)TE99(Z)XR93 was mapped to the net locus by Southern blot analysis, and a clone that included the proximal breakpoint was isolated form a genomic library (s. above). In addition, the position of the breakpoint was determined precisely by DNA sequencing. The breakpoints of the terminal deficiencies Df(2L)PM4, Df(2L)net18, and Df(2L)U314 were mapped by whole-genome Southern analysis. The deviations from wild type in netK1 and net1 were mapped to the 0.58kb SalI-HindIII fragment and that of netK7 to the 2.0 kb HindIII fragment (Figs 1B and 2A) by Southern blot analysis. The precise insertion sites of netK1, net1, and netPM50 were determined by nested PCR with the genomic DNA of these mutants and subsequent DNA sequencing.

Isolation of cDNA clones and 5′ RACE

Nine net-cDNA clones were isolated from 7.5×105 phages of a 4-8 hour embryonic cDNA library in λ UNI-ZAP XR (Schneitz et al., 1993) by screening it with a 10.3 kb EcoRI fragment from XR93-8. To clone a cDNA that includes the transcriptional start site of the net transcript, the 5′ RACE technique was applied to poly(A)+ RNA from 4-8 hour old embryos and from third instar larvae, using the 5′ Amplifinder RACE kit and following the instructions of Clontech. The RACE products of RNAs from both stages extended the longest net-cDNA, netce2, by the same length of 19 base pairs (Fig. 2A). DNA sequencing and comparison with the wild-type genomic DNA sequence revealed that the netce1 cDNA had a frame shift mutation that was due to the insertion of an additional A into the string of seven As at positions 941-947 (Fig. 2A). Hence the 0.58 kb SalI-HindIII fragment was exchanged in netce1 by the corresponding genomic fragment from XR93-8 to generate the corrected net-cDNA clone, netce1-ΔA.

Fig. 2.

cDNA and protein sequence of the net gene and sequence comparison of the Net bHLH domain with those of other (b)HLH transcription factors. (A) Sequence of net exons and Net protein. The nucleotide and putative amino acid sequences derived from netce1, 5′ RACE, and genomic DNA sequences are shown. The bHLH domain of the Net protein is boxed and prolines of the Pro-rich region (amino acids 100 to 164) are shown in bold letters. The positions of the 2.75 kb intron, the B104 netK1 insertion (the preceding 5 bp are the duplicated target site), the 244 bp net1 duplication (the preceding 244 bp are duplicated), and the P-element insertion netPM50 (the preceding 10 bp were duplicated in this lost allele) are indicated by vertical lines below triangles. The 5′ end of the sequenced net-cDNA, netce1, is marked by a vertical arrow. The last nucleotides of two cDNAs whose transcript was cleaved after the underlined canonical poly(A) addition signal AATAAA are encircled, while the third encircled poly(A) addition site, found in all other cDNAs, occurs after a putative noncanonical AAATA signal (also underlined). Restriction sites shown in Fig. 1B are indicated. The GenBank Accession number for the nucleotide and amino acid sequence of the net gene and its mutant alleles is AF303350. (B) Net is a bHLH protein of class A. The Net bHLH domain is aligned with those of class A (NDF1 to Da), class B (Myc to Sim), and class C bHLH proteins (H to E(spl)mβ), and with those of non-basic HLH proteins (Emc to ID2). These four classes have been defined by Ohsako et al. (1994). Class A proteins, which include the ASC products (Campuzano et al., 1985), bind to CA(C/G)CTG target sites, act as transcriptional activators, and lack an arginine in position 13 of the basic region, which is present in class B and C proteins (marked by a black circle). Class B proteins bind to CA(C/T)GTG target sites, but do not bind to class A sites, a property that is based on the specific interaction of an arginine residue at position 13 in the basic region of class B bHLH domains with the G4 nucleotide of class B DNA binding sites (Ferre-D’Amare et al., 1993). Class C (or proline-bHLH) proteins, which have a proline residue at position 6 of their basic region (indicated by an open circle), act as transcriptional repressors and include proteins encoded by hairy (h) (Ohsako et al., 1994; Barolo and Levine, 1997) and the Enhancer of split complex (Delidakis and Artavanis-Tsakonas, 1992; Knust et al., 1992) as well as their mammalian homologue HES (Ishibashi et al., 1994). Class C proteins also have an arginine at position 13 of their bHLH domain and bind to both class B and class C (CACG(C/A)G) sites, whereas class B proteins bind only weakly to class C sites. Class C proteins further include a C-terminal WRPW motif known to interact with the co-repressor groucho. Members of the last class lack a basic region and hence cannot bind to DNA by themselves, but are able to form heterodimers with class A or class B bHLH proteins and hence act as dominant-negative repressors by dimerization with bHLH activator proteins. Proteins have been aligned to maximize number of amino acid identities (black background) and homologies (grey background) that have been conserved within a given class.

Fig. 2.

cDNA and protein sequence of the net gene and sequence comparison of the Net bHLH domain with those of other (b)HLH transcription factors. (A) Sequence of net exons and Net protein. The nucleotide and putative amino acid sequences derived from netce1, 5′ RACE, and genomic DNA sequences are shown. The bHLH domain of the Net protein is boxed and prolines of the Pro-rich region (amino acids 100 to 164) are shown in bold letters. The positions of the 2.75 kb intron, the B104 netK1 insertion (the preceding 5 bp are the duplicated target site), the 244 bp net1 duplication (the preceding 244 bp are duplicated), and the P-element insertion netPM50 (the preceding 10 bp were duplicated in this lost allele) are indicated by vertical lines below triangles. The 5′ end of the sequenced net-cDNA, netce1, is marked by a vertical arrow. The last nucleotides of two cDNAs whose transcript was cleaved after the underlined canonical poly(A) addition signal AATAAA are encircled, while the third encircled poly(A) addition site, found in all other cDNAs, occurs after a putative noncanonical AAATA signal (also underlined). Restriction sites shown in Fig. 1B are indicated. The GenBank Accession number for the nucleotide and amino acid sequence of the net gene and its mutant alleles is AF303350. (B) Net is a bHLH protein of class A. The Net bHLH domain is aligned with those of class A (NDF1 to Da), class B (Myc to Sim), and class C bHLH proteins (H to E(spl)mβ), and with those of non-basic HLH proteins (Emc to ID2). These four classes have been defined by Ohsako et al. (1994). Class A proteins, which include the ASC products (Campuzano et al., 1985), bind to CA(C/G)CTG target sites, act as transcriptional activators, and lack an arginine in position 13 of the basic region, which is present in class B and C proteins (marked by a black circle). Class B proteins bind to CA(C/T)GTG target sites, but do not bind to class A sites, a property that is based on the specific interaction of an arginine residue at position 13 in the basic region of class B bHLH domains with the G4 nucleotide of class B DNA binding sites (Ferre-D’Amare et al., 1993). Class C (or proline-bHLH) proteins, which have a proline residue at position 6 of their basic region (indicated by an open circle), act as transcriptional repressors and include proteins encoded by hairy (h) (Ohsako et al., 1994; Barolo and Levine, 1997) and the Enhancer of split complex (Delidakis and Artavanis-Tsakonas, 1992; Knust et al., 1992) as well as their mammalian homologue HES (Ishibashi et al., 1994). Class C proteins also have an arginine at position 13 of their bHLH domain and bind to both class B and class C (CACG(C/A)G) sites, whereas class B proteins bind only weakly to class C sites. Class C proteins further include a C-terminal WRPW motif known to interact with the co-repressor groucho. Members of the last class lack a basic region and hence cannot bind to DNA by themselves, but are able to form heterodimers with class A or class B bHLH proteins and hence act as dominant-negative repressors by dimerization with bHLH activator proteins. Proteins have been aligned to maximize number of amino acid identities (black background) and homologies (grey background) that have been conserved within a given class.

DNA sequencing and homology searches

DNA sequencing was carried out on both strands of the net-cDNA netce1 (Fig. 2A) and the corresponding genomic DNA, including both intron-exon boundaries and about 2.2 kb of the intron, with a DNA sequencer model 373A using dye terminators (Applied Biosystems). Searches of the data bases for similarities were performed with the WWW-based BLAST program provided by the National Center for Biotechnology Information (NCBI) of the NIH (Bethesda, MD; Altschul et al., 1990) and enhanced with BEAUTY (Baylor College of Medicine, Houston, TX; Worley et al., 1995). Sequence alignments were made using the Gelassemble and Pileup modules of the Genetics Computer Group Wisconsin Package (GCG).

In situ hybridization to wing discs and pupal wings

In situ hybridization to imaginal discs and pupal wings with DIG-labeled probes was carried out essentially as described (Tautz and Pfeifle, 1989), using a DIG DNA-labelling and detection kit (Boehringer) and following the manufacturer’s application manual. Prepupae and pupae were staged. The outer case of prepupae was cut open, and prepupae were fixed in 4% formaldehyde for at least 60 minutes at room temperature or at 4°C overnight. For the preparation of pupal wings, the brown case and the transparent pupal cuticle surrounding the wings were removed, and the pupal wings were pulled out and fixed in 4% formaldehyde (or 4% paraformaldehyde for RNA probes) for 1 hour at room temperature or at 4°C overnight. DIG-labeled probes for net RNA consisted of the entire 2.2 kb insert of netce1-ΔA and for rho RNA of a 2.5 kb EcoRI fragment of a rho-cDNA (a gift from E. Bier). DIG-labeled antisense RNA probes were produced as run-off-transcripts for net from the netce1-ΔA template cleaved by EcoRI and for rho from the 2.5 kb rho-cDNA clone digested with HindIII. No signal was detectable in wing discs or pupal wings with a sense RNA probe for net. A knirps (kni) RNA-labeled probe was kindly supplied by C. Wolf.

Rescue of net mutants

To rescue the net phenotype, the 10.3 kb EcoRI fragment of XR93-8 was cloned into the EcoRI site of the P-element vector pW6, which carries the miniwhite gene as marker (Klemenz et al., 1987), to produce the construct RI (Fig. 1B). In addition, RIIa and RIIb were constructed by cloning the 3.3 kb NheI-XhoI and 3.9 kb NheI-SalI fragments of phage Y38-4, respectively, with the 8.8 kb EcoRI-NheI fragment of XR93-8 (Fig. 1B) into the EcoRI/XhoI cleaved polylinker of pW6. Each construct was injected together with pUChspΔ2-3 P-element helper plasmid (D. Rio, unpublished), carrying the transposase gene, into 30-60 min old w1118 or y w embryos and w+ transformants were selected (Rubin and Spradling, 1982). Transformed lines whose net transgene was not located on the second chromosome were crossed to w1118; net1 flies, their progeny crossed inter se, and the w+ F2 progeny screened for the absence of the net phenotype.

Construction and activation of UAS-net

The UAS-net transgene was constructed by cloning the EcoRI-XhoI insert of netce1-ΔA into the P-element vector pUAST (Brand and Perrimon, 1993), and transformed lines carrying the UAS-net transgene were obtained as described above. The lines 4U16-3.5, 8U14-7.1, and 9U5-1 were crossed to the GAL4-enhancer trap lines MS1096, MS209, C765, 10 (gifts from K. Basler), hGAL4, enGAL4, ptcGAL4, dppGAL4, 69BGAL4 and 71BGAL4 (all obtained from the Drosophila stock center in Bloomington). The progeny was tested for ectopic net expression with DIG-labeled net-cDNA probes. In addition, the GAL4-lines were crossed to UAS-lacZ (a gift from K. Basler) and histochemically stained for β-galactosidase activity (Boehringer) or with an anti-β-galactosidase antibody (Cappel). All these GAL4 lines strongly drive expression of net in the wing disc as well as in other tissues. Since MS1096 produced the strongest vein suppression in wings when combined with a UAS-net, only results obtained with this driver line are reported. Combination of a rhoGAL4 line (gift from C. Klämbt) with UAS-net caused embryonic lethality and hence could not be tested for vein suppression.

Mounting of fly wings

Wings were removed from flies stored in 70% ethanol, transferred directly into a drop of Advantage Permanent Mounting Media (Innovex), covered with a cover slip and photographed under bright-field optics through a compound microscope.

Drosophila strains

Df(2L)net18/CyO, Df(2L)U314/CyO, netK7/CyO, and netK1 stocks were kindly provided by B. Mechler. Df(2L)PM4/CyO and netPM50 stocks, and the In(2L)TE99(Z)XR93/CyO stock were generous gifts from C. Caggese and P. Heitzler. The UAS-rho line was kindly provided by C. Klämbt; GAL4-enhancer trap lines MS1096, MS209, C765, 10 and the UAS-lacZ line were supplied by K. Basler; and UAS-kni and UAS-knrl by R. Schuh; px72 by S. Hayashi. hGAL4, enGAL4, ptcGal4, dppGAL4, 69BGAL4, 71BGAL4 and UAS-Dras1V12; net1, px1, ri, w1118 and y w stocks, as well as the enhancer trap lines bsP1292 (synonym: bsl(2)03267) and bsl(2)k07909 were obtained from the Bloomington and Umea stock centres.

Molecular identification and characterization of the net locus

Cloning of the net gene was initiated by positional cloning. We started a chromosomal walk from the distal breakpoint of the deficiency Df(2L)al (Schneitz et al., 1993), which complements net. The walk was extended distally by screening genomic λ phage libraries for overlapping inserts until the lethal(2) giant larvae locus at the left telomere of the second chromosome was reached (Mechler et al., 1985). To delimit the net gene within the cloned genomic DNA, we took advantage of chromosomal aberrations including terminal deficiencies that either complement or uncover net. As illustrated in Fig. 1A, the proximal breakpoint of the deficiency Df(2L)U314, which complements net, is located distal to the proximal breakpoints of the larger deficiencies Df(2L)PM4 and Df(2L)net18, both of which uncover net. Hence, at least part of the net gene is included in the 30-40 kb region between the proximal breakpoints of Df(2L)U314 and Df(2L)PM4. To identify net gene sequences, we analysed this region in the spontaneous mutants net1, netK1, netK7 by Southern blot analysis and cloned the distal breakpoint of the X-ray induced inversion In(2L)TE99(Z)XR93, previously shown to uncover ush and net (P. Heitzler, personal communication).

Molecular lesions associated with the mutant alleles netK1 and net1 affected the same 0.58 kb SalI-HindIII DNA fragment (Fig. 1B) and suggest that it includes functional sequences of the net gene. Using a DNA probe overlapping with this fragment, we isolated nine cDNAs derived from a single 2.2 kb mRNA expressed in embryos, larvae and pupae. To confirm that this transcript is derived from the net gene, we attempted to rescue net mutants to a wild-type phenotype by P-element-mediated germline transformation. The first rescue construct, RI, which extended about 0.4 kb beyond the net inversion breakpoint of In(2L)TE99(Z)XR93 in the downstream region of the 2.2 kb transcript (Fig. 1B), showed no rescue of the net wing phenotype. A rescue construct including additional downstream sequences, RIIa, rescued about 45% or 80% of the flies to a wild-type phenotype when present as one or two copies. Finally, the net wing phenotype was rescued in all flies harbouring a single transgene, RIIb, which included an additional 0.54 kb of downstream sequences (Fig. 1B). These results suggest that the 2.2 kb mRNA encodes the Net protein and that downstream sequences of the gene are necessary to provide proper net activity in wing discs.

The net gene encodes a novel bHLH protein

Sequence analysis of the net-cDNAs and mapping the transcriptional start site of net by 5′ RACE showed that the longest cDNA is nearly full-length, missing only 19 bp at its 5′ end, and that net codes for a single 2.185 kb mRNA (Fig. 2A). Sequencing the corresponding genomic DNA revealed that the net transcript consists of two exons that are separated by a 2.75 kb intron and include one large open reading frame of 1095 bp. All isolated net-cDNAs terminate at one of three different positions within the 3′ untranslated region of the transcript and include a short polyA track at their 3′ ends. As most of them end at the same position shortly after a putative polyA addition signal that deviates from the canonical sequence, it appears that this signal is functional (Fig. 2A).

The open reading frame encodes a polypeptide of 365 amino acids including a basic helix-loop-helix (bHLH) domain, which suggests that the Net protein is a DNA-binding transcription factor. The large superfamily of bHLH transcriptional regulators has been subdivided into three different classes (Ohsako et al., 1994). Members that belong to different classes bind to similar but distinct DNA recognition sites, thought to reflect characteristic differences in amino acids at specific positions of their bHLH domains. The Net protein is most closely related to bHLH proteins of class A, encoded by proneural genes like atonal (ato) and those of the Achaete-scute Complex (ASC) (Fig. 2B). However, in contrast to class A proteins, which act as activators, Net contains a proline-rich domain in its N-terminal moiety (20% Pro between amino acids 100 and 164; Fig. 2A), which is characteristic of several transcriptional repressors (Cowell, 1994). Diagnostic domains or motifs other than the bHLH and proline-rich domains have not been found.

For a more detailed analysis of the molecular lesions in mutant net alleles, we amplified genomic DNA from homozygous netK1 and net1 flies by PCR. Sequencing revealed that netK1 was an insertion, associated with a 5 bp target site duplication, of the B104 transposable element at position 1041 of the spliced net transcript, whereas net1 was a small duplication after nucleotide 1105 of the preceding 244 bp (Fig. 2A). The B104 insertion interrupted the net open reading frame by several inframe stop codons and thus produced a Net protein truncated just before its bHLH domain (Fig. 2A). The 244 bp net1 duplication also generated several inframe stop codons, truncating Net in the basic region of the bHLH domain (Fig. 2A). Assuming that bHLH-dependent DNA binding is essential for Net function, we conclude that both netK1 and net1 are lack-of-function alleles. These findings and the complete rescue by the RIIb transgene of the net mutant phenotype demonstrate unequivocally that the 2.2 kb transcript encodes the Net functions.

net expression in intervein regions of the developing wing disc and pupal wing

Northern blot analysis showed that the 2.2 kb net transcript was expressed during all stages of the Drosophila life cycle (not shown). In situ hybridization of a net cDNA probe to whole-mount embryos (not shown) and imaginal discs revealed complex expression patterns. As, during development, the putative net lack-of-function alleles cause discernible defects only in the wing, the specific expression of net in the embryo may serve a redundant function. Therefore, the net function was analysed only during wing development.

In wing discs, net expression was first observed in early third instar larvae. net transcripts were confined throughout the third instar to prospective intervein sectors and to a narrow region straddling the wing margin (Fig. 3A), but were excluded from primordial wing veins expressing rho (Fig. 3B; Sturtevant et al., 1993). The complementary expression patterns of net and rho were maintained until the pupal wings had developed and expression of rho was being restricted from seven- to eight-cell wide stripes of vein-competent provein cells to the narrow two to three cell wide stripes of future vein cells (Fig. 3C; Sturtevant et al., 1994) in P1 pupae (Yu et al., 1996). Shortly before this P1 pupal stage, net expression was enhanced along the wing margin and in regions flanking proveins (Fig. 3D). During P1 and the subsequent P2 pupal stage, when crossveins are recognisable, net was expressed in the intervein regions, but continued to be repressed in the broad provein stripes (Fig. 3E). After 30 hours APF (after puparium formation) at 22°C, net remained repressed in proveins while its expression in intervein regions was considerably reduced (not shown). Thus, net and rho expression appeared to be mutually exclusive during the entire development of the wing although net expression did not expand when rho expression was reduced to the narrow stripes of vein cells in P1 pupae.

Fig. 3.

Expression of net and rho mRNA in late third instar wing discs and pupal wings of wild-type and net mutants. (A-E) Expression patterns of net and rho in developing wild-type wings. Expression patterns of net (A,D,E) and rho (B,C) transcripts were analysed by in situ hybridization of DIG-labeled probes to late third instar wing discs (A,B) and staged pupal wings (C-E). net transcripts appear in intervein regions of the wing pouch (A) and are excluded from vein primordia (B) where rho transcripts accumulate in wing discs (the wing margin (M) and the positions of L2 to L5 are indicated). Expression of net along the wing margin (A) is flanked on both sides by rho expression (B). Note that absence of net RNA from L2 is unclear only because L2 is out of focus in (A). Pupal wings at 28-30 hours APF (at 22°C) have refined their rho expression from proveins to the narrow veins (C) shortly before expression in crossveins appears. At an earlier time (24-28 hours APF at 22°C), pupal wings express net along the boundary regions between veins and interveins (most prominently along L1 and L3-L6, more weakly along L2) and along the wing margin, but net is repressed in proveins and the wing margin (D). At a slightly later time (28-32 hours APF at 22°C), net continues to be expressed weakly in all intervein regions, but remains repressed in all proveins (E). (F-H) net (G) and rho (F,H) expression patterns in late third instar wing discs (G,H) and pupal wings (F) of In(2L)TE99(Z)XR93/Df(2L)net18 animals. Note that no expression of net is detectable in the entire wing pouch of these animals (G, label marked by asterisk is located in ventral pleura and hinge region below the plane of focus), but net is expressed ectopically in notum (n), dorsal hinge (h), and ventral pleura and hinge regions (arrow and asterisk). Pupal wing blades of these net mutants also completely lack staining for the DIG-labelled net probe (not shown). rho expression expands into intervein regions of wing discs (H), with the exception of that between L3 and L4 (arrow). In pupal wings (26-30 hours APF at 22°C), rho expression is confined to veins in proximal and medial regions, but still expands into intervein regions in the distal parts of the wing except between L3 and L4 (e.g. into a region delimited by the broken line in F). (I) Endogenous and ectopic expression of net under the control of MS1096 in late third instar MS1096/+; UAS-net/+ wing disc. Note that net is ubiquitously expressed in the wing pouch with the exception of the wing margin (M). The same pattern is observed for rho transcripts, when rho is expressed under the control of MS1096. Arrows in C,D,F mark tracheae. In C-F,H, the position of longitudinal veins L1 to L6 are indicated. Wing discs are shown with their ventral side to the left and anterior upwards (A,B) or with their dorsal side upwards and anterior to the left (G-I). Pupal wings are shown with anterior upwards and proximal to the left.

Fig. 3.

Expression of net and rho mRNA in late third instar wing discs and pupal wings of wild-type and net mutants. (A-E) Expression patterns of net and rho in developing wild-type wings. Expression patterns of net (A,D,E) and rho (B,C) transcripts were analysed by in situ hybridization of DIG-labeled probes to late third instar wing discs (A,B) and staged pupal wings (C-E). net transcripts appear in intervein regions of the wing pouch (A) and are excluded from vein primordia (B) where rho transcripts accumulate in wing discs (the wing margin (M) and the positions of L2 to L5 are indicated). Expression of net along the wing margin (A) is flanked on both sides by rho expression (B). Note that absence of net RNA from L2 is unclear only because L2 is out of focus in (A). Pupal wings at 28-30 hours APF (at 22°C) have refined their rho expression from proveins to the narrow veins (C) shortly before expression in crossveins appears. At an earlier time (24-28 hours APF at 22°C), pupal wings express net along the boundary regions between veins and interveins (most prominently along L1 and L3-L6, more weakly along L2) and along the wing margin, but net is repressed in proveins and the wing margin (D). At a slightly later time (28-32 hours APF at 22°C), net continues to be expressed weakly in all intervein regions, but remains repressed in all proveins (E). (F-H) net (G) and rho (F,H) expression patterns in late third instar wing discs (G,H) and pupal wings (F) of In(2L)TE99(Z)XR93/Df(2L)net18 animals. Note that no expression of net is detectable in the entire wing pouch of these animals (G, label marked by asterisk is located in ventral pleura and hinge region below the plane of focus), but net is expressed ectopically in notum (n), dorsal hinge (h), and ventral pleura and hinge regions (arrow and asterisk). Pupal wing blades of these net mutants also completely lack staining for the DIG-labelled net probe (not shown). rho expression expands into intervein regions of wing discs (H), with the exception of that between L3 and L4 (arrow). In pupal wings (26-30 hours APF at 22°C), rho expression is confined to veins in proximal and medial regions, but still expands into intervein regions in the distal parts of the wing except between L3 and L4 (e.g. into a region delimited by the broken line in F). (I) Endogenous and ectopic expression of net under the control of MS1096 in late third instar MS1096/+; UAS-net/+ wing disc. Note that net is ubiquitously expressed in the wing pouch with the exception of the wing margin (M). The same pattern is observed for rho transcripts, when rho is expressed under the control of MS1096. Arrows in C,D,F mark tracheae. In C-F,H, the position of longitudinal veins L1 to L6 are indicated. Wing discs are shown with their ventral side to the left and anterior upwards (A,B) or with their dorsal side upwards and anterior to the left (G-I). Pupal wings are shown with anterior upwards and proximal to the left.

net activity represses rho transcription

In homozygous net1 wing discs, rho is ectopically expressed in broad domains, alternating with areas devoid of rho expression (Sturtevant and Bier, 1995; Sturtevant et al., 1997). To determine if this pattern of rho expression in wing discs reflects the null phenotype of net mutants, we examined rho expression in In(2L)TE99(Z)XR93/Df(2L)net18 wing discs, which fail to express net in the wing pouch (Fig. 3G). These discs displayed the same rho expression during larval and pupal stages as described for net1 mutants, i.e. rho expanded into all intervein regions with the exception of the sector between veins L3 and L4, but the effect was most pronounced in the distal intervein regions B (between L2 and L3) and D (between L4 and L5) (Fig. 3F,H). Hence, it appears that net represses rho in intervein regions of the wild-type wing, except in the region between L3 and L4. These effects of net mutations on rho expression are consistent with the null phenotype in the wing blade of In(2L)TE99(Z)XR93/Df(2L)net18 flies, which show ectopic veins most frequently in the distal intervein sectors B and D (Fig. 4B). This net null phenotype is indistinguishable from that of net1 and netK1 mutants (not shown), which confirms our assumption that the bHLH domain of Net is indispensable for its function.

To test if Net acts as a repressor of rho transcription, we expressed net ectopically in wing discs by means of the GAL4/UAS system (Brand and Perrimon, 1993). When activated by the GAL4 driver line MS1096, a GAL4 P-element insertion at the Beadex locus (Rørth, 1996; Milán et al., 1998), UAS-net was expressed nearly ubiquitously in the wing pouch (Fig. 3I) and widely repressed rho transcription (Fig. 5A), which displays a pattern similar to that in homozygous rhove discs (Sturtevant et al., 1993; D. B., H.-P. L., H. J. and M. N., unpublished). Consistent with this effect on rho expression, the wing phenotype of MS1096/Y; UAS-net flies was almost identical to that of homozygous rhove flies (Fig. 4D): the most distal part of L3 and the distal half of L5 were missing, while the distal part of L4 was also absent or interrupted by gaps (Fig. 4C). In contrast, L2 was wild type (Fig. 4C) except in occasional flies, in which its most distal part was missing, although rho expression was clearly absent from L2 proveins (Fig. 5A). In addition, wings in which net was ectopically expressed during development were smaller (Fig. 4C). Interestingly, when the dosage of MS1096 and UAS-net was reduced to one copy in net1 females, the net1 phenotype was rescued (Fig. 4E). The suppression of distal L5 formation was independent of net1 (not shown), while L4 was no longer suppressed at the reduced dosage. The absence of the anterior crossvein (Fig. 4C,E) was not caused by the ectopic net expression, but induced by the MS1096 driver in the absence of its UAS-net target (Fig. 4F; Milán et al., 1998).

Fig. 4.

Wing phenotypes of net gain-of-function and rho loss-of-function mutants are similar, and the dramatic wing phenotype caused by ubiquitous rho activity is suppressed by ubiquitous co-expression of net. (A) Wild-type wing with designations of longitudinal veins (L1-L6), anterior (acv) and posterior crossveins (pcv), and intervein sectors (A-E). (B) Wing of a net null mutant, In(2L)TE99(Z)XR93/Df(2L)net18, in which ectopic veins and plexi are most prominent in distal portions of sectors B and D and slightly less abundant in sectors A and E. This phenotype is indistinguishable from that of homozygous net1 and netK1 mutants. (C,D) Wing of a MS1096/Y; UAS-net/UAS-net male (C) is compared to that of a rhove mutant (D). Two doses of the UAS-net gain-of-function transgene, expressed ectopically under the control of MS1096, suppress the distal portions of L3-L5 and produce a wing phenotype very similar to that of the rho loss-of-function mutant rhove. (E,F) Wing of a MS1096/+; net1; UAS-net/+ female (E) is compared to that of the GAL4 driver line MS1096/MS1096 (F). One dose of the net gain-of-function transgene still suppresses the distal part of L5, is able to completely suppress the net null phenotype of net1 and the weak vein phenotype of the MS1096 driver line (except its loss of the anterior crossvein), which lacks the anterior crossvein, affects the posterior crossvein, and displays few ectopic veins (arrow in F; Milán et al., 1998), but hardly suppresses the distal half of L4 as observed for two doses of the UAS-net transgene (C). (G,H) Wings of MS1096/+; UAS-rho/+ (G) and of MS1096/+; UAS-net/UAS-rho mutants (H) are compared. The dramatic wing phenotype resulting from the ubiquitous expression of UAS-rho under the control of MS1096, which transforms the entire wing blade into vein-like tissue of dorsal and ventral epithelia that fail to adhere (G), is suppressed to give rise to a nearly wild-type wing phenotype (H; only deltas between L3 and L4 (arrow) remain), when UAS-net is co-expressed with UAS-rho during wing development. All wings are oriented with anterior upwards and proximal to the left.

Fig. 4.

Wing phenotypes of net gain-of-function and rho loss-of-function mutants are similar, and the dramatic wing phenotype caused by ubiquitous rho activity is suppressed by ubiquitous co-expression of net. (A) Wild-type wing with designations of longitudinal veins (L1-L6), anterior (acv) and posterior crossveins (pcv), and intervein sectors (A-E). (B) Wing of a net null mutant, In(2L)TE99(Z)XR93/Df(2L)net18, in which ectopic veins and plexi are most prominent in distal portions of sectors B and D and slightly less abundant in sectors A and E. This phenotype is indistinguishable from that of homozygous net1 and netK1 mutants. (C,D) Wing of a MS1096/Y; UAS-net/UAS-net male (C) is compared to that of a rhove mutant (D). Two doses of the UAS-net gain-of-function transgene, expressed ectopically under the control of MS1096, suppress the distal portions of L3-L5 and produce a wing phenotype very similar to that of the rho loss-of-function mutant rhove. (E,F) Wing of a MS1096/+; net1; UAS-net/+ female (E) is compared to that of the GAL4 driver line MS1096/MS1096 (F). One dose of the net gain-of-function transgene still suppresses the distal part of L5, is able to completely suppress the net null phenotype of net1 and the weak vein phenotype of the MS1096 driver line (except its loss of the anterior crossvein), which lacks the anterior crossvein, affects the posterior crossvein, and displays few ectopic veins (arrow in F; Milán et al., 1998), but hardly suppresses the distal half of L4 as observed for two doses of the UAS-net transgene (C). (G,H) Wings of MS1096/+; UAS-rho/+ (G) and of MS1096/+; UAS-net/UAS-rho mutants (H) are compared. The dramatic wing phenotype resulting from the ubiquitous expression of UAS-rho under the control of MS1096, which transforms the entire wing blade into vein-like tissue of dorsal and ventral epithelia that fail to adhere (G), is suppressed to give rise to a nearly wild-type wing phenotype (H; only deltas between L3 and L4 (arrow) remain), when UAS-net is co-expressed with UAS-rho during wing development. All wings are oriented with anterior upwards and proximal to the left.

Fig. 5.

Opposite effects on net and rho expression in mutant wing discs expressing net or rho ectopically, and ectopic expression of net in the absence of rho activity. Expression patterns of rho (A,D) and net (B,C,E,F) transcripts were analysed by in situ hybridization of DIG-labelled probes to wing discs of MS1096/+; UAS-net/UAS-net (A), MS1096/+; UAS-rho/+ (B), MS1096/+; UAS-Dras1V12/+ (C,D) and rhove (E) late third instar larvae, and to a rhove pupal wing (F). (A) rho expression is repressed by the ectopic expression of net in the wing pouch except in two stripes flanking the wing margin (M). (B-D) net expression is repressed in the entire wing pouch by the ectopic expression of rho (B) or the constitutive activation of Ras1 (C), which activates rho ubiquitously in the wing pouch (D) and generates hyperplastic discs (C,D). (E,F) In rhove animals that lack rho activity in most of the wing pouch (rho expression similar to pattern shown in A), net is expressed ectopically in all vein primordia of third instar wing discs, although not at the same level as in intervein regions (E; positions of L2 to L5 are indicated), and remains expressed ectopically 26-28 hours APF (at 22°C) in the distal portions of pupal wings (F; distal to broken line; arrow marks trachea), where vein development is suppressed (Fig. 4D). All wing discs are shown with their dorsal side upwards and anterior to the left (A-E), the pupal wing is shown with anterior to the left and proximal upwards (F).

Fig. 5.

Opposite effects on net and rho expression in mutant wing discs expressing net or rho ectopically, and ectopic expression of net in the absence of rho activity. Expression patterns of rho (A,D) and net (B,C,E,F) transcripts were analysed by in situ hybridization of DIG-labelled probes to wing discs of MS1096/+; UAS-net/UAS-net (A), MS1096/+; UAS-rho/+ (B), MS1096/+; UAS-Dras1V12/+ (C,D) and rhove (E) late third instar larvae, and to a rhove pupal wing (F). (A) rho expression is repressed by the ectopic expression of net in the wing pouch except in two stripes flanking the wing margin (M). (B-D) net expression is repressed in the entire wing pouch by the ectopic expression of rho (B) or the constitutive activation of Ras1 (C), which activates rho ubiquitously in the wing pouch (D) and generates hyperplastic discs (C,D). (E,F) In rhove animals that lack rho activity in most of the wing pouch (rho expression similar to pattern shown in A), net is expressed ectopically in all vein primordia of third instar wing discs, although not at the same level as in intervein regions (E; positions of L2 to L5 are indicated), and remains expressed ectopically 26-28 hours APF (at 22°C) in the distal portions of pupal wings (F; distal to broken line; arrow marks trachea), where vein development is suppressed (Fig. 4D). All wing discs are shown with their dorsal side upwards and anterior to the left (A-E), the pupal wing is shown with anterior to the left and proximal upwards (F).

We conclude that Net is able to repress the transcription of rho in the wing pouch, but net activity is crucial for the repression of rho in wild-type discs mainly in the distal intervein sectors A, B, D and E. Repression of rho by ectopic Net has no effect on the proximal portions of wing veins and, in most flies, on L2, which is consistent with the phenotype of rhove mutants, demonstrating that vein formation completely depends on rho activity only in the distal portions of longitudinal veins L3 to L5 (Sturtevant et al., 1993; Fig. 4D).

Rho and Ras activities prevent net transcription

Ectopic expression of rho under the control of a heat-inducible promoter during late third instar or early pupal development causes broadening of veins and ectopic vein formation, whereas absence or reduction of rho activity, or impaired Egfr-dependent signalling, generates the complementary phenotype, i.e. loss of vein structures (Sturtevant et al., 1993; Fig. 4D). To test if the development of ectopic veins is linked to the repression of net, net transcription was examined in wing discs after ectopic activation of UAS-rho by MS1096. Indeed, in such discs, which express rho strongly in almost the entire disc (not shown, but the pattern is identical to that of net transcripts in Fig. 3I), no net transcripts were detectable (Fig. 5B), which demonstrates that ectopic rho expression represses net transcription. It appears that in these wing discs all intervein regions develop as veins, since the resulting adult wings have a tube-like appearance and consist entirely of vein-like tissue (Fig. 4G; Guichard et al., 1999). This phenotype is much stronger than that of net null mutants (Fig. 4B). Ectopic Rho protein suppresses intervein fate in regions where such a fate is independent of net expression. Hence, we assume that in these regions rho represses intervein-promoting and vein-suppression genes that are different from net, and thus activates vein-promoting genes. These genes may show partial redundancy with net functions in regions in which their expression overlaps with that of net.

To investigate further whether rho mediates its repression of net through activated components of the Egfr signalling pathway, constitutively active Ras protein was expressed in MS1096/+; UAS-Dras1V12/+ wing discs. As evident from Fig. 5C, constitutive Ras activity produced overgrown wing discs that fail to express net. Although the adult wing phenotype could not be observed in these flies because they die as early pupae, ectopic activation of the Ras/mitogen-activated protein kinase (MAPK) signalling pathway in wing discs has been shown to give rise to ectopic veins (Karim and Rubin, 1998). These results illustrate that ectopic expression of rho or activated components of the Egfr signalling pathway represses the transcription of net and presumably of additional vein-suppression genes, and suggest that their repression is a prerequisite for vein formation.

Wing discs in which Ras is constitutively active accumulated high levels of rho transcripts in intervein regions (Fig. 5D). In contrast, rho expression was disrupted in wing discs and pupal wings that are homozygous for rhove (Sturtevant et al., 1993; D. B., H.-P. L., H. J. and M. N., unpublished, but pattern similar to that in Fig. 5A). In such discs, net transcription expanded into the vein primordia of third instar wing discs (compare Fig. 5E with Fig. 3A) and remained expressed ectopically in the distal portions of L3 to L5 of pupal wings (Fig. 5F), where vein development was suppressed (Fig. 4D). This result is consistent with the observation that ectopic expression of net was able to repress vein formation only in regions (Fig. 4C) where vein formation depended on rho expression (Fig. 4D). These findings show that net and rho transcripts accumulate in mutually exclusive patterns in developing wings, and specify intervein and vein precursor cells, respectively. Thus, net expression is negatively regulated by Egfr signalling and, in turn, represses rho expression.

Finally, UAS-net and UAS-rho were co-expressed ubiquitously under the control of MS1096. Surprisingly, while ubiquitous expression of either net or rho in developing wings generated curved wings with suppressed distal L5 (Fig. 4E) or tube-like wings composed of vein-like tissue (Fig. 4G), ubiquitous co-expression of net and rho resulted in a nearly wild-type wing blade (Fig. 4H). Since the ubiquitous expression of Rho, which is under the control of MS1096, is independent of Net, this result indicates that Net represses, in addition to rho, other vein-promoting genes downstream of Rho in Egfr signalling. If this were not the case and the only function of Net was to repress the endogenous rho gene, we would expect the same phenotype after ubiquitous co-expression of Net and Rho as after ubiquitous expression of Rho (Fig. 4G) rather than the nearly wild-type wing observed (Fig. 4H). Therefore, Net suppresses vein development in interveins through a control that consists of at least two tiers, one repressing rho, the other interfering with the activation of vein-promoting genes downstream of Rho-enhanced Egfr signalling. Curiously, although ubiquitous expression of Net is able to suppress the effect of ubiquitously co-expressed Rho in intervein regions, it is unable to repress vein development in veins. It follows that differences between future vein and intervein regions exist that are sufficient to correctly determine vein versus intervein fates in the presence of an excess of both Net and Rho. Our findings further suggest a considerable redundancy in the gene networks participating in the specification of vein versus intervein fates. This conclusion is also borne out by the observation that if both Rho and Net are reduced or lost in rhovenet1 double mutants, veins are less severely truncated than in wings of rhove single mutants (Díaz-Benjumea and García-Bellido, 1990a; Sturtevant and Bier, 1995).

Mutual repression of net and rho in wing vein patterning

The specification of vein versus intervein fate in the Drosophila wing disc crucially depends on the restriction of Rho expression to the primordial veins (Wasserman and Freeman, 1997; Bier, 1998). Rho enhances Egfr signalling, which generates elevated levels of activated MAPK essential for proper vein development and the suppression of intervein fate in vein cells (Gabay et al., 1997; Guichard et al., 1999). In the absence of Egfr signalling, either in clones mutant for the Drosophila Egfr (Díaz-Benjumea and García-Bellido, 1990b) or in wings of rhovevn1 flies (Díaz-Benjumea and García-Bellido, 1990a; Sturtevant and Bier, 1995), veins fail to develop and assume an intervein fate.

Our results show that a counterpart of rho is net whose transcription is confined to intervein sectors of third instar wing discs. Absence of net activity in third instar wing discs results in derepression of rho in all intervein regions except sector C and initiates the formation of ectopic veins. Conversely, in the absence of rho activity, net is ectopically expressed in vein primordia of wing discs, as well as in distal veins L3 to L5 of pupal wings where differentiation of veins is suppressed. Thus, while net is repressed in vein cells by high Egfr signalling dependent on Rho, Net protein suppresses vein fate in intervein cells by repressing rho transcription (Sturtevant and Bier, 1995). Repression of net in veins by high levels of Rho-dependent Egfr signalling is crucial for vein development, as ectopic expression of net is able to repress rho transcription and suppress vein fate. However, Egfr signalling in the absence of Rho is sufficient to initiate normal vein development in the proximal and anterior portions of the wing. Consistent with this finding, repression of rho by ectopic Net prevents vein formation only in regions where it depends on Rho.

Ubiquitous expression of rho or activated Ras represses net transcription in the entire wing disc and promotes vein development throughout the wing. Such flies have tube-like wings composed mainly, if not exclusively, of vein cells (Guichard et al., 1999). Since this phenotype is much stronger than that of net null mutants, it follows (1) that additional factors that repress rho must be present in intervein regions, and (2) that ectopic Rho is able to suppress intervein fate in regions where such a fate is independent of net expression, and hence that in these regions rho is able to repress intervein-promoting and vein-suppression genes different from net. The first conclusion is also supported by the notion that rho does not expand into the intervein region between L3 and L4 in net mutants. The simplest explanation for these observations is that rho expression is controlled by a set of separate silencers responding to different repressors in different, perhaps overlapping intervein regions and that Net acts as one of the repressors of rho.

net encodes a novel member of the bHLH protein family

Net is a member of the bHLH transcription factor family (Horimoto et al., 1994; Murre et al., 1994; Phillips, 1994), which has been subdivided into three structurally and functionally distinct classes (Ohsako et al., 1994). Since Net contains a valine residue in position 13 of the basic region of its bHLH domain, one might surmise that it is a class A protein (Fig. 2B) and acts as a transcriptional activator. This assumption is consistent with the fact that Net has no class C-type repressor domains like the orange domain or the C-terminal WRPW-motif (Dawson et al., 1995; Fisher et al., 1996), which associates with the co-repressor Groucho (Paroush et al., 1994; Alifragis et al., 1997; Jimenez et al., 1997). However, Net includes a proline-rich domain in its N-terminal moiety that might function as a repressor domain as shown for non-bHLH transcriptional regulators such as Even-skipped, Female-specific 1 and RGM1 (Cowell, 1994; Hanna-Rose and Hansen, 1996). Thus, despite the similarity of the Net bHLH domain to that of class A bHLH proteins, Net may act as a transcriptional repressor. This supposition is supported by the observation that Net represses rho expression during wing development. Although we cannot exclude that Net activates a gene whose product represses rho, we found that Net was also able to repress vein-promoting genes in the presence of activator proteins activated by ectopic Rho through enhanced Egfr signalling in interveins (Fig. 4H; see below). The simplest explanation of these results is that Net is a transcriptional repressor and thus exceptional among class A proteins.

Tiers of vein versus intervein fate specification

Any attempt to understand the role of net in the patterning of wing veins must consider its function in the context of the complex process of vein versus intervein specification, which occurs in several tiers of intimately linked gene regulatory circuits (Biehs et al., 1998; Lunde et al., 1998). The first distinction between vein and intervein anlagen is made during the third larval instar when the wing pouch is subdivided along the anteroposterior axis into sectors by prepattern genes activating morphogenetic signals, such as Hh and Dpp, whose threshold concentrations determine the sector boundaries. Longitudinal veins are first specified by the activity of vein-organising genes that are activated in narrow stripes along these boundaries by short range signals, while intervein genes, like N and net, are thought to be activated through cues of prepattern genes (Biehs et al., 1998). In the second tier of regulation, Egfr signalling plays a decisive role. Accordingly, one of the first genes to be activated by vein-organising genes in third instar wing discs is rho (Sturtevant et al., 1997). During late third instar and early prepupal stages, the membrane proteins Rho and S enhance signalling of the ubiquitously expressed Egfr, activated by the neuregulin-like protein Vn (Schnepp et al., 1996), and possibly additional ligands, in longitudinal veins and adjacent provein cells (Guichard et al., 1999). High Egfr signalling activates Delta (Dl) as well as rho. Because Rho is able to stimulate Egfr signalling in adjacent cells, rho and Dl expression continue to expand (Sturtevant et al., 1994; de Celis et al., 1997). Net is required to prevent the expansion of rho (and Dl) expression beyond the proveins into the intervein regions (Sturtevant and Bier, 1995). Our results suggest that Net might act in this process as a repressor of rho by binding to its wing enhancer, which is deleted in the rhove allele (Sturtevant et al., 1993). Conversely, high Egfr signalling stimulated by Rho is essential in proveins to repress net, as net transcription is no longer confined to intervein cells in rhove mutant wings.

Dl expression in proveins initiates the third tier in the regulation of wing vein patterning. At the same time, Rho-dependent Egfr signalling begins to downregulate egfr mRNA in proveins (Sturtevant et al., 1994), yet MAPK activity remains high up to the early P2 pupal stage when crossveins become visible (Guichard et al., 1999), presumably because of the continued stimulation of Egfr signalling through Rho. During this third tier of regulation, Dl activates N signalling in the lateral provein cells flanking the narrow central stripes of vein cells (de Celis et al., 1997). N expression, which has been activated in a pattern complementary to that of Dl in third instar wing discs, is now enhanced by a positive feedback loop of N signalling in lateral provein cells. In addition, since N signalling represses rho transcription through the proline-bHLH protein E(spl)mβ, expression of rho is again confined to the narrow stripes of vein cells by the early P2 pupal stage (de Celis et al., 1997). This regulatory loop of N signalling appears to be analogous to that operating during lateral inhibition in neurogenesis, by which a neural precursor cell, stochastically selected from a group of proneural cells, prevents its neighbours from adopting a neural fate (Artavanis-Tsakonas and Simpson, 1991). However, in contrast to lateral inhibition, vein cells are not selected stochastically from provein cells, as longitudinal veins are straight, which suggests that expression of rho is always restricted to the central cells of the provein stripe. This is achieved by a fourth tier of vein fate regulation through Dpp signalling (Yu et al., 1996). Expression of Dpp is activated by high Egfr signalling in provein cells at the same time as Dl expression. The effect of Dpp, however, is confined to the middle of the provein stripes by the secreted product of the short gastrulation (sog) gene, an antagonist of Dpp expressed in the adjacent intervein regions. If we assume that Dpp signalling inhibits expression of N, N signalling would be blocked only in the central cells of proveins, which are thus determined to become vein cells. Consistent with this last tier of regulation, veins in dpp clones are no longer straight, but follow an irregular path confined to the wider provein stripe (Yu et al., 1996). Dpp signalling maintains both its own expression and that of Rho in vein cells and determines their differentiation (Yu et al., 1996).

Finally, the refinement of proveins to veins further depends on less well understood mutually inductive processes, in which predetermined vein cells of the dorsal wing surface signal to the underlying ventral cells to maintain vein cell identity (García-Bellido, 1977; García-Bellido and de Celis, 1992; Milán et al., 1997).

Redundancy in wing vein patterning

Vein cells are not only specified for their specific fate, but are also prevented from adopting the alternative intervein fate. This is achieved at all regulatory tiers by Rho-stimulated Egfr signalling, which represses, initially in proveins and later in veins, not only net, as shown here, but also blistered (bs; Roch et al., 1998). bs encodes a transcription factor, the Drosophila homologue of the mammalian serum response factor (DSRF; Montagne et al., 1996), which is instrumental in the specification of intervein fate (Fristrom et al., 1994; Roch et al., 1998). Like net and rho, activities of bs and rho repress each other and hence are mutually exclusive and expressed in complementary patterns of intervein and vein territories. However, whereas net and rho mutually depend on each other already from the onset of vein versus intervein development in third instar wing discs, bs and rho begin to restrict each other’s expression only during prepupal development (Roch et al., 1998). Similar to net and rho, this mutual repression is not apparent in the proximal parts of the wing because of a redundancy inherent in this regulation. Although rho and net (or bs) are able to repress each other when expressed ectopically, no ectopic expression occurs in the absence of one of these genes in the proximal wing regions.

The activities of Rho and Net are not restricted to the mutual repression of their genes. This conclusion follows from the surprising observation that their ubiquitous co-expression in the developing wing generates a nearly wild-type wing. Since ubiquitous expression of Rho produces wings composed entirely of vein-like tissue, Net is able to repress vein-promoting genes in intervein regions, despite the presence of an activating signal stimulated by the ubiquitous Rho, while the situation is reversed in veins where Net is unable to repress vein-promoting genes in the presence of Net-independent Rho expression. Similarly, Rho-dependent signalling represses intervein-promoting genes, like bs, in veins, despite the presence of Net, but does not have this capability in interveins in the presence of Rho-independent Net expression. It follows that even in the absence of endogenous Net and Rho vein and intervein primordia are in different states upon which ubiquitous Net and Rho can act to correctly determine their fates, which implies a considerable redundancy in the specification of vein versus intervein fate.

Because of the necessity for the concomitant suppression of the alternative fate when vein or intervein fates are specified, a system evolved in which Rho determined vein development by repressing net and later bs, which in turn specify intervein fate by repressing vein development in intervein regions. Such a balanced system is intrinsically labile unless it is stabilised through feedback loops (see above). Multiple feedback loops operate at all tiers of vein fate regulation, and tiers are closely linked and overlap in time, which further enhances the stability of the system because it generates redundancy. We propose that the functions of Net and Bs are partially redundant because they both repress rho in intervein regions during overlapping, though not identical, developmental periods. Thus, while Net represses rho in all intervein sectors of third instar wing discs except sector C, Bs begins to repress rho in these regions only in early prepupal wings (Roch et al., 1998). In view of this hypothesis, it might be less surprising that the wing phenotype of net null mutants is much weaker than that resulting from ubiquitous expression of rho, which also represses net completely, but converts almost the entire wing into vein material (Guichard et al., 1999). We assume that the lack of Net function in net wing discs is partially compensated by the activation of bs, whose product represses rho in most of the intervein regions during the prepupal and pupal stage. This assumption is consistent with our observations (Fig. 3F) and with the earlier finding that bs null mutants exhibit a wing phenotype very similar to that resulting from ubiquitous expression of rho in the developing wing (Fristrom et al., 1994; Roch et al., 1998). The rhove-like phenotype obtained after ubiquitous expression of Net during wing development is largely explained by the ability of Net to repress rho. The partial redundancy of net and bs functions in wing discs is supported by experiments in which ubiquitous expression of Net was still able to suppress the strong ectopic vein formation phenotype of bs2/bsP1292 mutants (not shown, but phenotype is indistinguishable from that produced in a net1 mutant background shown in Fig. 4E). In addition, bs expression is reduced in distal portions of net1 wing discs and hence appears to depend partially on Net (Biehs et al., 1998), a finding that is consistent with our observation that LacZ expression of the bs enhancer trap line bsP1292 is ectopically activated and enhanced after ectopic expression of Net in MS1096/+; UAS-net/bsP1292 wing discs (D. B., H.-P. L., H. J. and M. N., unpublished).

Since net and plexus (px) mutant alleles interact genetically and exhibit indistinguishable genetic behaviours and wing phenotypes (Díaz-Benjumea and García-Bellido, 1990), one might surmise that they are part of the same developmental pathway. Indeed, px, which encodes a ubiquitously expressed nuclear matrix protein, like Net is required for the repression of rho (Matakatsu et al., 1999). Our finding that net expression is normal in px1 or px72 mutant wing discs (D. B., H.-P. L., H. J. and M. N., unpublished) is unable to support a role for Px in the regulation of net activity. However, although px72 is the strongest known px allele, at present no px null alleles are available (García-Bellido, 1977; Matakatsu et al., 1999) to subject this possibility to a conclusive test.

The partial redundancy of net and bs might have evolved in more advanced insects like Diptera by the acquisition of vein-suppressing genes that reduced the much larger number of wing veins characteristic of more primitive insects. It has been proposed that such cryptic paraveins are still detectable in extravein mutants of Drosophila that display a strong tendency to misexpress rho and form ectopic veins between and parallel to L1 to L6 (Biehs et al., 1998). According to this hypothesis, net may have evolved to suppress paraveins P2 and P6, while hairy (h) acquired the ability to suppress paraveins P4 and P5 (Biehs et al., 1998). It is attractive to speculate that the paravein-suppressing function of Net may have evolved as an ability to repress rho in paraveins, but not in vein primordia where net was not expressed, and thus entailed a general vein-suppressing function, which initially may have been shared with bs. While bs later lost its early vein-suppressing function, its prepupal vein-suppressing function and pupal intervein differentiation function were retained, which resulted in a temporal overlap of rho repression by Net and Bs and thus led to a temporal redundancy of net and bs functions. It is interesting to note that the rho repressing functions of net and h as well as that of E(spl)mβ (de Celis et al., 1997) all encode bHLH proteins, a property that is entirely consistent with a general hypothesis of how gene networks evolve (Noll, 1993).

We thank P. Heitzler, C. Caggese, B. Mechler, K. Basler, C. Klämbt, R. Schuh, and the Umea and Bloomington stock centres for fly stocks; E. Bier for a rho-cDNA; and C. Wolf for a kni probe. We thank W. Boll and T. Müller for DNA sequencing; T. Gutjahr for help in the generation of transgenic flies and extensive computational analyses; and F. Ochsenbein and T. Gutjahr for expert artwork. We are grateful to E. Frei for crucial advice and discussions leading to the identification of net and comments on the manuscript. This work has been supported by a fellowship of the Deutsche Forschungsgesellschaft (to H.-P. L.), a grant from the Max-Planck Society (to H. J.), grant 31-40874.94 from the Swiss National Science Foundation (to M. N.) and by the Kanton Zürich.

Alifragis
,
P.
,
Poortinga
,
G.
,
Parkhurst
,
S. M.
and
Delidakis
,
C.
(
1997
).
A network of interacting transcriptional regulators involved in Drosophila neural fate specification revealed by the yeast two-hybrid system
.
Proc. Natl. Acad. Sci. USA
94
,
13099
13104
.
Altschul
,
S. F.
,
Gish
,
W.
,
Miller
,
W.
,
Myers
,
E. W.
and
Lipman
,
D. J.
(
1990
).
Basic local alignment search tool
.
J. Mol. Biol
.
215
,
403
410
.
Artavanis-Tsakonas
,
S.
and
Simpson
,
P.
(
1991
).
Choosing a cell fate: a view from the Notch locus
.
Trends Genet
.
7
,
403
408
.
Artavanis-Tsakonas
,
S.
,
Muskavitch
,
M. A. T.
and
Yedvobnick
,
B.
(
1983
).
Molecular cloning of Notch, a locus affecting neurogenesis in Drosophila melanogaster
.
Proc. Natl. Acad. Sci. USA
80
,
1977
1981
.
Artavanis-Tsakonas
,
S.
,
Matsuno
,
K.
and
Fortini
,
M. E.
(
1995
).
Notch signaling
.
Science
268
,
225
232
.
Barolo
,
S.
and
Levine
,
M.
(
1997
).
hairy mediates dominant repression in the Drosophila embryo
.
EMBO J
.
16
,
2883
2891
.
Biehs
,
B.
,
Sturtevant
,
M. A.
and
Bier
,
E.
(
1998
).
Boundaries in the Drosophila wing imaginal disc organize vein-specific genetic programs
.
Development
125
,
4245
4257
.
Bier
,
E.
(
1998
).
Localized activation of RTK/MAPK pathways during Drosophila development
.
BioEssays
20
,
189
194
.
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
,
190
203
.
Brand
,
A. H.
and
Perrimon
,
N.
(
1993
).
Targeted gene expression as a means of altering cell fates and generating dominant phenotypes
.
Development
118
,
401
415
.
Campos-Ortega
,
J. A.
(
1997
).
Neurogenesis in Drosophila: an historical perspective and some prospects
.
Perspect. Dev. Neurobiol
.
4
,
267
271
.
Campuzano
,
S.
,
Carramolino
,
L.
,
Cabrera
,
C. V.
,
Ruíz-Gómez
,
M.
,
Villares
,
R.
,
Boronat
,
A.
and
Modolell
,
J.
(
1985
).
Molecular genetics of the achaete-scute gene complex of D. melanogaster
.
Cell
40
,
327
338
.
Cowell
,
I. G.
(
1994
).
Repression versus activation in the control of gene transcription
.
Trends Biochem. Sci
.
19
,
38
42
.
Cubadda
,
Y.
,
Heitzler
,
P.
,
Ray
,
R. P.
,
Bourouis
,
M.
,
Ramain
,
P.
,
Gelbart
,
W.
,
Simpson
,
P.
and
Haenlin
,
M.
(
1997
).
u-shaped encodes a zinc finger protein that regulates the proneural genes achaete and scute during the formation of bristles in Drosophila
.
Genes Dev
.
11
,
3083
3095
.
Dawson
,
S. R.
,
Turner
,
D. L.
,
Weintraub
,
H.
and
Parkhurst
,
S. M.
(
1995
).
Specificity for the hairy/enhancer of split basic helix-loop-helix (bHLH) proteins maps outside the bHLH domain and suggests two separable modes of transcriptional repression
.
Mol. Cell. Biol
.
15
,
6923
6931
.
de Celis
,
J. F.
,
Llimargas
,
M.
and
Casanova
,
J.
(
1995
).
ventral veinless, the gene encoding the Cf1a transcription factor, links positional information and cell differentiation during embryonic and imaginal development in Drosophila melanogaster
.
Development
121
,
3405
3416
.
de Celis
,
J. F.
,
de Celis
,
J.
,
Ligoxygakis
,
P.
,
Preiss
,
A.
,
Delidakis
,
C.
and
Bray
,
S.
(
1996
).
Functional relationship between Notch, Su(H) and the bHLH genes of the E(spl) complex: the E(spl) genes mediate only a subset of Notch activities during imaginal development
.
Development
122
,
2719
2728
.
de Celis
,
J. F.
,
Bray
,
S.
and
García-Bellido
,
A.
(
1997
).
Notch signalling regulates veinlet expression and establishes boundaries between veins and interveins in the Drosophila wing
.
Development
124
,
1919
1928
.
Delidakis
,
C.
and
Artavanis-Tsakonas
,
S.
(
1992
).
The Enhancer of split [E(spl)] locus of Drosophila encodes seven independent helix-loop-helix proteins
.
Proc. Natl. Acad. Sci. USA
89
,
8731
8735
.
Díaz-Benjumea
,
F. J.
and
García-Bellido
,
A.
(
1990a
).
Genetic analysis of the wing vein pattern of Drosophila
.
Roux’s Arch. Dev. Biol
.
198
,
336
354
.
Díaz-Benjumea
,
F. J.
and
García-Bellido
,
A.
(
1990b
).
Behavior of cell mutant for an EGF receptor homologue of Drosophila in genetic mosaics
.
Proc. R. Soc. London Biol. Sci
.
242
,
36
44
.
Díaz-Benjumea
,
F. J.
and
Cohen
,
S. M.
(
1993
).
Interaction between dorsal and ventral cells in the imaginal disc directs wing development in Drosophila
.
Cell
75
,
741
752
.
Ferre-D’Amare
,
A. R.
,
Prendergast
,
G. C.
,
Ziff
,
E. B.
and
Burley
,
S. K.
(
1993
).
Recognition by Max of its cognate DNA through a dimeric b/HLH/Z domain
.
Nature
363
,
38
45
.
Fisher
,
A. L.
,
Ohsako
,
S.
and
Caudy
,
M.
(
1996
).
The WRPW motif of the hairy-related basic helix-loop-helix repressor proteins acts as a 4-amino-acid transcription repression and protein-protein interaction domain
.
Mol. Cell. Biol
.
16
,
2670
2677
.
Frei
,
E.
,
Baumgartner
,
S.
,
Edström
,
J.-E.
and
Noll
,
M.
(
1985
).
Cloning of the extra sex combs gene of Drosophila and its identification by P-element-mediated gene transfer
.
EMBO J
.
4
,
979
987
.
Fristrom
,
D.
,
Gotwals
,
P.
,
Eaton
,
S.
,
Kornberg
,
T. B.
,
Sturtevant
,
M.
,
Bier
,
E.
and
Fristrom
,
J. W.
(
1994
).
blistered: a gene required for vein/intervein formation in wings of Drosophila
.
Development
120
,
2661
2671
.
Fu
,
W.
and
Noll
,
M.
(
1997
).
The Pax2 homolog sparkling is required for development of cone and pigment cells in the Drosophila eye
.
Genes Dev
.
11
,
2066
2078
.
Gabay
,
L.
,
Seger
,
R.
and
Shilo
,
B.-Z.
(
1997
).
In situ activation pattern of Drosophila EGF receptor pathway during development
.
Science
277
,
1103
1106
.
García-Bellido
,
A.
(
1977
).
Inductive mechanisms in the process of wing vein formation in Drosophila
.
W. Roux’s Arch. Dev. Biol
.
182
,
93
106
.
García-Bellido
,
A.
and
de Celis
,
J. F.
(
1992
).
Developmental genetics of the venation pattern of Drosophila
.
Annu. Rev. Genet
.
26
,
277
304
.
Guichard
,
A.
,
Biehs
,
B.
,
Sturtevant
,
M. A.
,
Wickline
,
L.
,
Chacko
,
J.
,
Howard
,
K.
and
Bier
,
E.
(
1999
).
rhomboid and Star interact synergistically to promote EGFR/MAPK signaling during Drosophila wing vein development
.
Development
126
,
2663
2676
.
Guillen
,
I.
,
Mullor
,
J. L.
,
Capdevila
,
J.
,
Sanchez-Herrero
,
E.
,
Morata
,
G.
and
Guerrero
,
I.
(
1995
).
The function of engrailed and the specification of Drosophila wing pattern
.
Development
121
,
3447
3456
.
Hanna-Rose
,
W.
and
Hansen
,
U.
(
1996
).
Active repression mechanisms of eukaryotic transcription repressors
.
Trends Genet
.
12
,
229
234
.
Hidalgo
,
A.
(
1994
).
Three distinct roles for the engrailed gene in Drosophila wing development
.
Curr. Biol
.
4
,
1087
1098
.
Hidalgo
,
A.
(
1998
).
Growth and patterning from the engrailed interface
.
Int. J. Dev. Biol
.
42
,
317
324
.
Horimoto
,
K.
,
Yamamoto
,
H.
,
Yanagi
,
K.
,
Ohshima
,
K.
and
Otsuka
,
J.
(
1994
).
A simple procedure for assigning a sequence motif with an obscure pattern: application to the basic/helix-loop-helix motif
.
Protein Eng
.
7
,
1433
1440
.
Ishibashi
,
M.
,
Moriyoshi
,
K.
,
Sasai
,
Y.
,
Shiota
,
K.
,
Nakanishi
,
S.
and
Kageyama
,
R.
(
1994
).
Persistent expression of helix-loop-helix factor HES-1 prevents mammalian neural differentiation in the central nervous system
.
EMBO J
.
13
,
1799
1805
.
Jimenez
,
G.
,
Paroush
,
Z.
and
Ish-Horowicz
,
D.
(
1997
).
Groucho acts as a corepressor for a subset of negative regulators, including Hairy and Engrailed
.
Genes Dev
.
11
,
3072
3082
.
Karim
,
F. D.
and
Rubin
,
G. M.
(
1998
).
Ectopic expression of activated Ras1 induces hyperplastic growth and increased cell death in Drosophila imaginal tissues
.
Development
125
,
1
9
.
Kilchherr
,
F.
,
Baumgartner
,
S.
,
Bopp
,
D.
,
Frei
,
E.
and
Noll
,
M.
(
1986
).
Isolation of the paired gene of Drosophila and its spatial expression during early embryogenesis
.
Nature
321
,
493
499
.
Klemenz
,
R.
,
Weber
,
U.
and
Gehring
,
W. J.
(
1987
).
The white gene as a marker in a new P-element vector for gene transfer in Drosophila
.
Nucleic Acids Res
.
15
,
3947
3959
.
Knust
,
E.
,
Schrons
,
H.
,
Grawe
,
F.
and
Campos-Ortega
,
J. A.
(
1992
).
Seven genes of the Enhancer of split complex of Drosophila melanogaster encode helix-loop-helix proteins
.
Genetics
132
,
505
518
.
Kornberg
,
T.
,
Sidén
,
I.
,
O’Farrell
,
P.
and
Simon
,
M.
(
1985
).
The engrailed locus of Drosophila: in situ localization of transcripts reveals compartment-specific expression
.
Cell
40
,
45
53
.
Lawrence
,
P. A.
and
Morata
,
G.
(
1976
).
Compartments in the wing of Drosophila: A study of the engrailed gene
.
Dev. Biol
.
50
,
321
337
.
Lecuit
,
T.
,
Brook
,
J.
,
Ng
,
M.
,
Calleja
,
M.
,
Sun
,
H.
and
Cohen
,
S. M.
(
1996
).
Two distinct mechanisms for long-range patterning by Decapentaplegic in the Drosophila wing
.
Nature
381
,
387
393
.
Lindsley
,
D. L.
and
Zimm
,
G. G.
(
1992
).
The Genome of
Drosophila melanogaster. San Diego
:
Academic Press
.
Lunde
,
K.
,
Biehs
,
B.
,
Nauber
,
U.
and
Bier
,
E.
(
1998
).
The knirps and knirps-related genes organize development of the second wing vein in Drosophila
.
Development
125
,
4145
4154
.
Matakatsu
,
H.
,
Tadokoro
,
R.
,
Gamo
,
S.
and
Hayashi
,
S.
(
1999
).
Repression of the wing vein development in Drosophila by the nuclear matrix protein Plexus
.
Development
126
,
5207
5216
.
Mechler
,
B. M.
,
McGinnis
,
W.
and
Gehring
,
W. J.
(
1985
).
Molecular cloning of lethal(2)giant larvae, a recessive oncogene of Drosophila melanogaster
.
EMBO J
.
4
,
1551
1557
.
Milán
,
M.
,
Baonza
,
A.
and
García-Bellido
,
A.
(
1997
).
Wing surface interactions in venation patterning in Drosophila
.
Mech. Dev
.
67
,
203
213
.
Milán
,
M.
,
Díaz-Benjumea
,
F. J.
and
Cohen
,
S. M.
(
1998
).
Beadex encodes an LMO protein that regulates Apterous LIM-homeodomain activity in Drosophila wing development: a model for LMO oncogene function. Genes Dev
.
12
,
2912
2920
.
Montagne
,
J.
,
Groppe
,
J.
,
Guillemin
,
K.
,
Krasnow
,
M. A.
,
Gehring
,
W. J.
and
Affolter
,
M.
(
1996
).
The Drosophila Serum Response Factor gene is required for the formation of intervein tissue of the wing and is allelic to blistered
.
Development
122
,
2589
2597
.
Murre
,
C.
,
Bain
,
G.
,
van Dijk
,
M. A.
,
Engel
,
I.
,
Furnari
,
B. A.
,
Massari
,
M. E.
,
Matthews
,
J. R.
,
Quong
,
M. W.
,
Rivera
,
R. R.
and
Stuiver
,
M. H.
(
1994
).
Structure and function of helix-loop-helix proteins
.
Biochim. Biophys. Acta
1218
,
129
135
.
Nellen
,
D.
,
Burke
,
R.
,
Struhl
,
G.
and
Basler
,
K.
(
1996
).
Direct and long-range action of a DPP morphogen gradient
.
Cell
85
,
357
368
.
Noll
,
M.
(
1993
).
Evolution and role of Pax genes
.
Curr. Opin. Genet. Dev
.
3
,
595
605
.
Noll
,
R.
,
Sturtevant
,
M. A.
,
Gollapudi
,
R. R.
and
Bier
,
E.
(
1994
).
New funtions of the Drosophila rhomboid gene during embryonic and adult development are revealed by a novel genetic method, enhancer piracy
.
Development
120
,
2329
2338
.
Ohsako
,
S.
,
Hyer
,
J.
,
Panganiban
,
G.
,
Oliver
,
I.
and
Caudy
,
M.
(
1994
).
Hairy function as a DNA-binding helix-loop-helix repressor of Drosophila sensory organ formation
.
Genes Dev
.
8
,
2743
2755
.
Paroush
,
Z.
,
Finley
,
R. L.
, Jr
,
Kidd
,
T.
,
Wainwright
,
S. M.
,
Ingham
,
P. W.
,
Brent
,
R.
and
Ish-Horowicz
,
D.
(
1994
).
Groucho is required for Drosophila neurogenesis, segmentation, and sex determination and interacts directly with Hairy-related bHLH proteins
.
Cell
79
,
805
815
.
Phillips
,
S. E.
(
1994
).
Built by association: structure and function of helix-loop-helix DNA-binding proteins
.
Structure
2
,
1
4
.
Roch
,
F.
,
Baonza
,
A.
,
Martín-Blanco
,
E.
and
García-Bellido
,
A.
(
1998
).
Genetic interactions and cell behaviour in blistered mutants during proliferation and differentiation of the Drosophila wing
.
Development
125
,
1823
1832
.
Rørth
,
P.
(
1996
).
A modular misexpression screen in Drosophila detecting tissue-specific phenotypes
.
Proc. Natl. Acad. Sci. USA
93
,
12418
12422
.
Rubin
,
G. M.
and
Spradling
,
A. C.
(
1982
).
Genetic transformation of Drosophila with transposable element vectors
.
Science
218
,
348
353
.
Sambrook
,
J.
,
Fritsch
,
E. F.
and
Maniatis
,
T.
(
1989
).
Molecular Cloning: A Laboratory Manual
.
Cold Spring Harbor, NY
:
Cold Spring Harbor Laboratory Press
.
Schneitz
,
K.
,
Spielmann
,
P.
and
Noll
,
M.
(
1993
).
Molecular genetics of aristaless, a prd-type homeo box gene involved in the morphogenesis of proximal and distal pattern elements in a subset of appendages in Drosophila
.
Genes Dev
.
7
,
114
129
.
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
.
Simcox
,
A.
(
1997
).
Differential requirement for EGF-like ligands in Drosophila wing development
.
Mech. Dev
.
62
,
41
50
.
Simcox
,
A. A.
,
Grumbling
,
G.
,
Schnepp
,
B.
,
Bennington-Mathias
,
C.
,
Hersperger
,
E.
and
Shearn
,
A.
(
1996
).
Molecular, phenotypic, and expression analysis of vein, a gene required for growth of the Drosophila wing disc
.
Dev. Biol
.
177
,
475
489
.
Sturtevant
,
M. A.
and
Bier
,
E.
(
1995
).
Analysis of the genetic hierarchy guiding wing vein development in Drosophila
.
Development
121
,
785
801
.
Sturtevant
,
M. A.
,
Roark
,
M.
and
Bier
,
E.
(
1993
).
The Drosophila rhomboid gene mediates the localized formation of wing veins and interacts genetically with components of the EGF-R signaling pathway
.
Genes Dev
.
7
,
961
973
.
Sturtevant
,
M. A.
,
O’Neill
,
J. W.
and
Bier
,
E.
(
1994
).
Down-regulation of Drosophila Egf-r mRNA levels following hyperactivated receptor signaling
.
Development
120
,
2593
2600
.
Sturtevant
,
M. A.
,
Biehs
,
B.
,
Marin
,
E.
and
Bier
,
E.
(
1997
).
The spalt gene links the A/P compartment boundary to a linear adult structure in the Drosophila wing
.
Development
124
,
21
32
.
Tabata
,
T.
and
Kornberg
,
T. B.
(
1994
).
Hedgehog is a signaling protein with a key role in patterning Drosophila imaginal discs
.
Cell
76
,
89
102
.
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
.
Tautz
,
D.
and
Pfeifle
,
C.
(
1989
).
A non-radioactive in situ hybridization method for the localization of specific RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback
.
Chromosoma
98
,
81
85
.
Wasserman
,
J. D.
and
Freeman
,
M.
(
1997
).
Control of EGF receptor activation in Drosophila
.
Trends Cell Biol
.
7
,
431
436
.
Worley
,
K. C.
,
Wiese
,
B. A.
and
Smith
,
R. F.
(
1995
).
BEAUTY: an enhanced BLAST-based search tool that integrates multiple biological information resources into sequence similarity search results
.
Genome Res
.
5
,
173
184
.
Yu
,
K.
,
Sturtevant
,
M. A.
,
Biehs
,
B.
,
François
,
V.
,
Padgett
,
R. W.
,
Blackman
,
R. K.
and
Bier
,
E.
(
1996
).
The Drosophila decapentaplegic and short gastrulation genes function antagonistically during adult wing vein development
.
Development
122
,
4033
4044
.