Wing vein formation in Drosophila melanogaster: Hairless is involved in the cross-talk between Notch and EGF signaling pathways

Mechanisms of Development 115 (2002) 3–14 www.elsevier.com/locate/modo

Wing vein formation in Drosophila melanogaster: Hairless is involved in the cross-talk between Notch and EGF signaling pathways Bernd Johannes, Anette Preiss* Institut fu¨r Genetik (240), Universita¨t Hohenheim, Garbenstrasse 30, 70593 Stuttgart, Germany Received 8 August 2001; received in revised form 1 March 2002; accepted 12 March 2002

Abstract Wing vein development in Drosophila is controlled by different morphogenetic pathways, including Notch. Hairless (H) antagonizes Notch target gene activation by binding to the Notch signal transducer Suppressor of Hairless [Su(H)]. Accordingly, overexpression of H phenocopies reduction of Notch activity. Deletion of the Su(H)-binding domain in H–C2 results in loss of H activity. However, overexpression of H–C2 induces formation of ectopic veins. In a screen for genetic modifiers of this phenotype, we have identified several genes involved in Notch and epidermal growth factor (EGF) signaling. Most notably veinlet, an activator of EGF signaling, acts downstream of H– C2. H–C2 positively regulates veinlet maybe through inhibition of inter-vein determinants in agreement with a model, whereby Notch and EGF signaling pathways cross-regulate vein pre-patterning. q 2002 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Blistered; Cross-talk; Decapentaplegic; Drosophila epidermal growth factor receptor; Dpp; Drosophila; Epidermal growth factor pathway; Enhancers; Genetic interaction; Hairless; H–C2; Masking; Mutation; net; Notch pathway; Overexpression; Gene regulation; Phenocopy; rhomboid; Screen; Serum response factor; Suppressor of Hairless binding domain; Suppressors; Suppressor of Hairless; veinlet; Wing development; Wing veins

1. Introduction Wings of the adult Drosophila fly are an excellent model system for a genetic analysis of cross-talk between different morphogenetic signaling pathways. Especially the establishment of the normal venation pattern, which is a multi step process involving the precisely coordinated activity of epidermal growth factor (EGF), Notch and dpp signaling cascades, is extremely sensitive to genetic change (Lindsley and Zimm, 1992; reviewed in de Celis, 1998). Early wing vein specification is under the control of the EGF signaling pathway which is required for the adoption and maintenance of pro-vein cell fate during larval and early pupal development (Sturtevant et al., 1993; Roch et al., 1998; Guichard et al., 1999; see Bier, 1998 and de Celis, 1998 for reviews). Pro-vein specification is tightly linked to the expression of veinlet (ve) (Sturtevant et al., 1993; Sturtevant and Bier, 1995) which is involved in the processing and activation of spitz (spi), the ligand of the Drosophila EGF receptor (DER) (reviewed in Kla¨mbt, 2002). Expression of ve within presumptive pro-vein cells starts already in third instar larval stages in the wing imaginal disc and predicts * Corresponding author. Tel.: 149-711-459-2206; fax: 149-711-4592211. E-mail address: [email protected] (A. Preiss).

presumptive vein territories. Therefore, ve is the earliest marker for vein fate (Sturtevant et al., 1993; Sturtevant and Bier, 1995). At the same time, a down-regulation of blistered (bs), which encodes the Drosophila homologue of the mammalian serum response factor (SRF), is observed (Montagne et al., 1996; Roch et al., 1998). bs is required to specify inter-vein fate in the presumptive inter-vein regions of the developing wing (Fristrom et al., 1994; Roch et al., 1998). This mutual exclusive regulation of bs and ve is guided by pre-pattern gene activity during larval development, which establish boundaries that provide positional cues for proper vein placement (Sturtevant and Bier, 1995; Sturtevant et al., 1997; Biehs et al., 1998; Lunde et al., 1998). A number of genes are involved in the regulation of ve, most notably net which inhibits ve expression outside of the pro-vein areas (Sturtevant and Bier, 1995; Brentrup et al., 2000). Later during pupal development, when bs and ve become independent of pre-patterning genes and interdependent of each other, the pro-vein pattern is further refined and finally translated into the adult veins proper (Biehs et al., 1998; Roch et al., 1998; see de Celis, 1998, for review). During pupal development, vein fate is maintained by the activity of the dpp signaling pathway which acts epistatic to EGF signaling at this stage (de Celis, 1997; Diaz-Benjumea et al., 1989). It has been proposed that dpp signals consoli-

0925-4773/02/$ - see front matter q 2002 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(02)00083-7

4

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

date and finally conclude vein fate under the guidance of EGF activity (Diaz-Benjumea and Garcia-Bellido, 1990; Sturtevant and Bier, 1995; Yu et al., 1996; de Celis, 1997). During this stage, pro-vein fate is restricted to yield the final vein width in a process of lateral inhibition, governed by the Notch signaling pathway (ArtavanisTsakonas et al., 1999). Once the Notch receptor is activated through its ligand Delta (Dl), vein competent cells are forced into inter-vein fate. In more detail, it is thought that the intracellular domain of Notch (N intra) is cleaved upon receptor activation and, together with the DNA-binding protein Suppressor of Hairless [Su(H)], leads to transcriptional activation of effector genes (Struhl and Adachi, 1998; Artavanis-Tsakonas et al., 1999). In the context of vein width refinement, mb , a member of the Enhancer of split complex [E(spl)-C], is an important target gene. mb encodes a transcriptional repressor of the bHLH family which antagonizes the activity of vein-promoting genes and turns down ve expression (de Celis et al., 1997; Huppert et al., 1997). Hairless (H) has been shown to act as antagonist of Notch signaling by directly binding to Su(H), thereby inhibiting activation of downstream effector genes (Brou et al., 1994; Maier et al., 1999). In accordance with its antagonistic role, H overexpression interferes with several Notch-dependent developmental events, including vein width refinement (Maier et al., 1992, 1997). In a systematic structure–function analysis, various H deletion-constructs were engineered and tested for residual activity in vivo (Maier et al., 1997). In the construct H–C2, the presumptive Su(H)-binding domain was removed. As a consequence, H–C2 protein has completely lost the ability to bind to Su(H) protein and to interfere with Su(H)-dependent developmental processes like bristle development, wing margin specification or vein width refinement in vivo (Maier et al., 1997; Marquart et al., 1999; Nagel et al., 2001). Apart from the internal deletion, the H–C2 protein compares to the wild type with respect to antibody recognition, apparent molecular weight, subcellular distribution, stability as well as biochemical interactions with other H partner proteins (Maier et al., 1997, 1999; unpublished observations). With regard to endogenous activity, the H–C2 lines are rather weak compared to the full length H constructs, however, after heat shock induction, expressivity is similar (Maier et al., 1997). Surprisingly, overexpression of H–C2 induces lethality like the wild type construct and in addition, leads to the induction of ectopic vein material in certain inter-vein regions of the wing (Maier et al., 1997; Nagel et al., 2000; this work). Keeping in mind that H itself is not a transcriptional regulator but rather functions through protein–protein interactions with different protein targets (Morel et al., 2001 and references therein), these phenotypes cannot be simply explained by altered activation of Notch target genes. Rather, they might uncover a currently unknown Su(H) independent activity of H involving different protein(s)

maybe outside of the Notch signaling cascade. In order to understand this phenomenon, we analyzed genetically the involvement of H–C2 in the process of vein establishment in comparison to wild type H. The data presented in this work are in agreement with a model whereby H, apart from antagonizing Notch signaling, positively regulates EGF signaling during the process of wing vein formation. 2. Results 2.1. Pheno-critical period of vein induction by hs–H–C2 Overexpression of hs–H–C2 induced ectopic veins only between day 5 and 6 after egg deposition. Phenotypes varied significantly and were arranged into a phenotypic series of five classes (Fig. 1). Using precisely synchronized cultures, the pheno-critical period was restricted to pre-pupal and early pupal developmental stage, starting approximately at the larval–prepupal transition (Fig. 2C). Induction of H–C2 during mid- to late-third instar larval stages did not result in ectopic vein formation even with prolonged and, in their consequence, semi-lethal heat shocks (see also Maier et al., 1997). Ectopic venation was not randomly distributed and certain regions of the wing blade were more sensitive than others (summarized in Fig. 2B). Three spots proved sensitive also in the control animals (asterisks in Fig. 1) and were subsequently excluded from further analysis. In order to distinguish between temporal and/or sensitivity differences, we partitioned the wing into distinct inter-vein sectors A–F (Fig. 2A) and scored the appearance of extra veinlets over time for each sector independently (Fig. 2C). As is apparent from this representation, the six different sectors respond with a similar temporal profile, but with different sensitivities. The less sensitive sectors A and C reveal two pheno-critical periods, which might, however, be a consequence of the time convolution of the data. In summary, the main impact of H–C2 on the wing venation process occurs in the pre-pupal and early pupal stages of development. 2.2. H and H–C2 synergize with regard to vein formation As our previous work shows, the H–C2 protein is unable to bind to Su(H) and has lost basically all of H wild type activity (Maier et al., 1997; Marquart et al., 1999): H–C2 is unable to rescue the haplo-insufficient H loss-of-function phenotype and does not cause the typical H gain-of-function bristle phenotypes (Maier et al., 1997). Thus, H–C2 venation phenotypes might either uncover a Su(H) independent function of H or an unrelated, novel activity. Although overexpression of H wild type constructs causes only little ectopic vein material (Fig. 3D) (Maier et al., 1992; 1997), H and H–C2 proteins are able to synergize upon overexpression. Class 3–5 overexpression phenotypes (Fig. 1) are only achieved with two copies of the hs–H–C2 construct (Fig. 3A), whereas a single copy gives reproducibly only a class 1

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

5

phenotype, even after double heat shock (Fig. 3B). A class 5 phenotype can, however, easily be restored by addition of one copy of a full length hs–H construct, i.e. flies carrying two copies of hs–H–C2 and flies carrying one copy of hs– H–C2 plus one copy hs–H show the same phenotype upon induction for 1 h at 378C (compare Fig. 3A,C). We conclude that the vein inductive property of H–C2 is a native function of H, and that H is able to partially substitute for H–C2 in this process. 2.3. Candidate screen for modifiers of the H–C2 wing venation phenotype If indeed the vein inductive property of H–C2 uncovers a Su(H) independent activity of H, the question arises as to what other targets H might act on. In order to identify such putative targets, we set up a candidate screen for dominant modifiers, concentrating on the three main signaling pathways which normally contribute to wing vein development, Notch, EGF and dpp signaling cascades (see Tables 1–4 and Section 4 for further details). In the doubly heterozygotes, H–C2 was induced during the pheno-critical period with double heat shocks to make up for the weak phenotypes caused by a single hs–H–C2 copy (Fig. 3B,C). Classification of the phenotypic levels was according to Fig. 1. To account for inaccuracies, the numerical modification index was converted to four general terms: strongly positive (s 1) with enhancement of more than two phenotypic levels, weakly positive (w 1) with enhancement of more than one

Fig. 1. Phenotypic series of H–C2 effects on wing venation. Wing venation phenotypes resulting from overexpression of H–C2 were arranged in a phenotypic series and classified accordingly. Genotypes are homozygous w 1118; hs–H–C2. Class 0, wing with wild type phenotype (no heat shock). Asterisks mark positions where small veinlets may arise after heat shock in a w 1118 or in a y 1 w 67c23 background. Class 1, ectopic veins are observed in the distal territory of the costal cell between L1 and L2. Class 2, in addition to ectopic veins in the costal cell, extra veinlets arise close to L5 in the marginal cell. Class 3, typified by branching of ectopic veins and vein dots between L4 and L5. Class 4, typically, the posterior cross-vein bifurcates and detaches from L5, which sends long branches into the marginal and second posterior cells. Class 5, a massive network of ectopic veins is observed in the above fields. In some cases, additional ectopic veinlets are seen in the distal area of the subcostal cell or bifurcations of the anterior cross-vein.

Fig. 2. Pheno-critical period of H–C2 vein-promoting activity. (A) Schematic representation of the sectors in which the phenotypic effects of H–C2 overexpression were monitored. (B) Overview of zones sensitive for H–C2 effects. Highest sensitivity is shown in red, intermediate sensitivity in orange and least sensitivity in yellow. No ectopic veins were observed in the areas of gray shading. (C) Pheno-critical period individually determined for sectors A–F (see legend). Pupae from synchronized cultures were heat shocked at the given times and wings of the adults scored for phenotypes. Percentage of mutant wings was determined independently for each sector A–F.

6

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

Notch ligands, mutations affecting the signal transduction machinery as well as Notch target genes. Furthermore, other neurogenic mutations as well as several proneural members were included in the screen (Table 1). We found that reduction of single doses of Dl or components of the E(spl) complex including E(spl)mb , enhanced the H–C2 wing phenotype in a dominant manner. Most of the other components of the Notch signaling pathway behaved largely neutral in this screen. Notably, reduction of the Su(H) gene dose did not considerably alter H–C2 effects. Interestingly, N loss of function alleles acted as strong negative modifiers, whereas N Ax alleles enhanced the response to H–C2 overexpression. Components of the EGF signaling pathway, screened for Table 1 Modification of H–C2 effects by Notch signaling components Group

Loci and genotypes

Notch

Ser Dl N

dx Su(H)

H E(spl)

mam

Fig. 3. Synergy between wild type H and H–C2. (A) Two copies of hs–H– C2, induced for 1 h at 378C during the pheno-critical period, result in a typical class 5 phenotype. (B) A single copy of hs–H–C2, induced for 2 £ 1 h at 378C, gives only a class 1 phenotype. (C) A single copy of each, hs–H plus hs–H–C2, induced for 1 h at 378C results in a phenotype indistinguishable from two copies of hs–H–C2. (D) Two copies of hs–H induced for 1 h at 378C result in ectopic veinlets and branching of L2, L4 and L5 veins close to the wing margin.

Ser rx82/1 Dl FX3/1 w 1118 N 55e11 FRT101/w 1118 y p w a N 5419/w 1118 w a N 264-10 rb 1/w 1118 N Ax-E2 sn 3/w 1118; e 11/1 N nd-1 rb/1 N nd-3 rb/1 N spl-1/1 w p dx 24/w 1118 Su(H) AR9 pr 1/1 Adh uf3 Su(H) PI3 cn/1 Su(H) S5 pr 1 H wa/1 e 4 E(spl) E75 Tb p/1 ry gro E48 tx/1 st boss 44 tx/1 Df(3R)X10/1 Df(3R)P11/1 Df(3R)BX22/1 mam 04615/1 mam IL115/1

n s1 w2 w2 w2 w1 (w 2) w2 n n n n n n (w 2) (w 1) w2 w1 w1 n (w 2) n

Neurogenic

bib kuz neu

cn bw bib ID05/1 kuz k01403 neur IF65/1

n n (w 2)

ASC 1 bHLH

sc ac ASC

sc 1 cho 1/w 1118 y p ac 1 w a Df(1)sc B57 w a/1 y p ac 1 sc 1pn 1 w e59/w 1118 ru 1 h 25 th 1 st 1 cu 1 sr e s ca 1/1 ru 1 h 26 th 1 st 1 cu 1 sr e s ca 1/1 emc D st 1 in 1 kni ri-1 p p Dr Mio/1 emc 04322/1 wg Sp da 1 pr 1 cn 1/1 da II134/1

n n w2 w2 n n w2 (w 1) n n

h emc

phenotypic level, neutral (n) for an enhancement of up to one phenotypic level. Given reduced viability of certain mutant combinations, we defined weak negative (w 2) for a suppression of more than half a phenotypic level in at least two repeated assays. The results are summarized in Tables 1–4. In a first set of experiments, combinations with Notchfamily members and relatives were analyzed. These included mutations in the Notch receptor itself and in

Class a

da

a Classes: n, neutral, enhancement up to one and suppression less than a half phenotypic level; w 2, weak negative, reduction of H–C2 effects by approximately 0.5 phenotypic levels; w 1, weak positive, enhancement of more than one phenotypic level; s 1, strong positive, enhancement of more than two phenotypic levels; Brackets indicate a tendency which was obvious but not strong enough to qualify as modifier according to our scheme.

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14 Table 2 Modification of H–C2 effects by EGF signaling components Group

Loci and genotypes

EGF

spi S ve vn ve vn DER

Drk Sos 14-3-3-e rl pnt EGF-related

a

aop ast ec ed px

Class a

spi 1 cn 1 bw 1 sp 1/1 spi s3547/1 S S4 cn 1 bw 1 sp 1/1 S k09530/1 ve 1/1 ve rho–lac1/1 vn 1/1 ve 1 vn 1/1 cn Egfr flb bw sp/1 Egfr k05115/1 Bc Egfr Elp/1 drk k02401/1 sos k05224/1 14-3-3-e j2B10/1 rl 1 sp/1 pnt 07825/1 st e pnt 1/1

n n w2 (w 2) w2 w2 w2 w2 n n w1 n n n n n n

aop 1 cn bw sp/1 ast 1 dpp d-ho/1 br 1 w e ec 1 rb 1 t 4/1 ed 1 dp O2 cl 1/1 px 1 sp p/1 rl 1 px 1 sp p/1

(w 2) n n n w1 n

For classification, see Table 1.

dominant interactions with H–C2, included ligand and its processing, the receptor itself, components of the signal transduction cascade as well as related genes (see Table 2). As expected, ve and to a lesser degree Star (S) mutations dominantly reduced the amount of extra vein material. Both genes are essential for vein formation (Sturtevant et al., 1993; Sturtevant and Bier, 1995; Guichard et al., 1999), and thus, reduction of their activity was expected to antagonize the H–C2 vein-promoting effect. By lowering the ve gene dose, the suppression was almost complete except for some ectopic vein material in the region of the anterior cross-vein (sector A in Fig. 2B; data not shown), a region

which is not affected by the homozygous ve 1 (see Fig. 4G; Sturtevant et al., 1993; Sturtevant and Bier, 1995). Table 3 summarizes the effects of mutations in dpp signaling pathway components on the vein-promoting activity of H–C2. These are the dpp ligand, its receptor, and downstream signaling components. Both, dpp S4 and sax HB mutations caused a weak suppression which was unambiguous but considerably weaker as the dominant suppression by ve 1 (Table 3). As has been shown before that dpp acts epistatic to EGF signaling during wing vein development (de Celis, 1997), we propose that the observed weak interactions are not direct but rather reflect late events during dpp-dependent vein consolidation. A number of other mutations were tested that either have wing venation phenotypes on their own or have been associated to either pathway or otherwise to wing vein formation (Table 4). The role of many of these genes is little understood, however, some have been characterized by genetic or molecular means in more detail (Lindsley and Zimm, 1992). Some rather strong interactions were noted, in particular with net, bs and lace. net has been proposed to act as a pre-pattern component that might restrict ve expression to the presumptive pro-vein areas (Sturtevant et al., 1993; Brentrup et al., 2000). The gene bs encodes the Drosophila homologue of the SRF which acts as a selector of inter-vein Table 4 Modification of H–C2 effects by miscellaneous factors Group

Loci and genotype

Others

sca vvl net bs chl nw shn ab lace dsr hv cv cv-d cv-c rn det cno

Table 3 Modification of H–C2 effects by dpp signaling components Group

Loci and genotypes

dpp

dpp

n (w 2) – n

tkv

tkv strII/1

n

sax

Wo

sax /1 sax HB/1

n (w 2)

dally

dally 06464 ry 506/1

n

Mad a

Class a

dpp s1/1 dpp s4/1 dpp s1/dpp s4 dpp d8/1

12

Mad b pr/1

For classification, see Table 1.

7

th l(2)k07918 l(2)35Fc l(3)j1E4 l(3)L6540 l(3)S057302 flr btl nub

n a

For classification, see Table 1.

Class a b p Adh n2 pr 1 cn 2 sca 1/1 vvl ZM/1 net 1 or 1 sp p/1 bs P1292/1 chl 1 l(2)bw 1 bw 2b mr 2/1 chl 1 nw 2/1 cn 1 shn 1 bw 1 sp 1/1 ab 2 ix 2 bw 1 sp 2/1 ab k02807/1 Adh n7 lace 2 cn 1/1 dsr 1 hv 1/1 y p cv 1 v 1 f 1/1 cv-d 1/1 cv-c 1 Sb sbd-2 Ubx bxd-113/1 rn 3p p Ubx bx-1 sr 1 e s/1 det 1 st p cno 3 e s/1 st cno ts e/1 th j5C8/1 l(2)k07918 k07918/1 l(2)35Fc k11403/1 l(3)j1E4 j1E4/1 l(3)L6540 L6540/1 l(3)S057302 S057302/1 flr 3/1 btl 1/1 nub 1 b noc Sco lt 1 stw 3/1

n n s1 s1 n n n n n n s1 n (w 2) n (w 2) n n n w2 n n n (w 1) w1 (w 1) n n n n

8

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

fate and is required to establish inter-vein identity even in the absence of vein inductive cues (Fristrom et al., 1994; Montagne et al., 1996). lace has recently been proposed to be involved in glycolipid synthesis (Adachi-Yamada et al., 1999). No other dominant interactions were noted (Table 4). 2.4. Dominant modifiers of H–C2

Fig. 4. Epistasis experiments between H–C2, net and ve. (A) Heterozygous net 1/1 flies have wild type wings. (B) Homozygous net 1 mutation causes the typical net-like venation most notably in the costal, the subcostal and the second posterior cell. (C) Loss of one net gene dose strongly enhances the phenotypic effects of a single hs–H–C2 copy that compares to a double copy (compare with Fig. 3B,C). The genotype is net 1/1, [hs–H–C2]/1 heat shocked for 1 h at 378C during the pheno-critical period. (D) Ectopic venation can be dramatically enhanced in the homozygous net background. The genotype is net 1/net 1; [hs–H–C2]/1, heat shocked for 2 £ 1 h at 378C during the pheno-critical period. (E) A single copy of hs-H heat shocked for 2 £ 1 h at 378C typically induces small deltas at the distal tips of L2 and L5. (F) If one copy of net is lost, hs–H induces a strong net-like venation indistinguishable from H–C2; the genotype is net 1/1; [hs–H]/1 heat shocked for 2 £ 1 h at 378C. (G) Homozygous ve 1 flies typically lack distal veins. (H) This phenotype is unchanged despite the overexpression of hs– H–C2. Sometimes, ectopic veinlets close to the anterior cross-vein are present, a territory beyond ve control. The genotype is [hs–H–C2]/[hs– H–C2]; ve 1/ve 1, heat shocked for 1 h at 378C during the pheno-critical period. (I) Overexpression of UAS–H–C2 within the posterior compartment using the en–Gal4 driver line results in a nearly complete transformation of inter-vein into vein material posterior to L4. The arrow points to ectopic bristles within this vein field. Please note the bifurcated anterior cross-vein (arrowhead). The genotype is [en–Gal4]/[UAS–H–C2]. (J) If the flies are mutant for ve at the same time, no ectopic veins are formed, however, the ectopic bristles remain (arrow). The genotype is [en–Gal4]/[UAS–H– C2];ve 1/ve 1. (K) Likewise ectopic expression of wild type H within the posterior compartment results in small vein patches and irregular vein thickening. No ectopic bristles are induced. As expected from disruption of Notch signaling, the wing margin is affected. The genotype is [en–Gal4]/ 1; [UAS–H]/1. (L) The small ectopic vein patches disappear in a ve mutant background. Loss of distal veins seems more pronounced. The genotype is [en–Gal4]/1; ve 1/ve 1 [UAS–H].

Our screen revealed a small number of strong dominant modifiers of H–C2 vein-promoting activity: two repressors, N and ve, and a small number of enhancers, Dl, lace, net and bs. We note that reduction of bs function appears to destabilize inter-vein fate decision in general, thereby sensitizing inter-vein tissue to a variety of pro-vein signals. For example, even simple heat treatment is sufficient to induce ectopic venation in this genetic background. We conclude that it is the sensitized background rather than direct interactions between bs and H–C2 which uncovers the vein-promoting activity of H–C2. We therefore analyzed the epistatic relationship between known pro-vein and inter-vein factors ve and net with H–C2. In agreement with the essential role of ve in the establishment of vein fate, homozygous ve 1 mutant completely suppressed ectopic vein induction through hs–H–C2, except for some small veinlets in sectors A and B (Fig. 4G,H). No ectopic veins were visible in the distal wing blade, where the ve 1 phenotype is apparent, even after strong overexpression (double heat shock with two copies hs–H–C2) (compare Fig. 4G,H). The respective hs–H–C2 control gave a fully penetrant ectopic venation phenotype under the same conditions. This demonstrates that H–C2 strictly depends on ve for the induction of veins and suggests that overexpression of H–C2 might somehow result in the ectopic activation of ve. In accordance with the proposed role of net as negative regulator of ve, net 1 mutations cause extensive extra veins (Sturtevant et al., 1993; Brentrup et al., 2000). This phenotype comprises nearly all aspects of hs–H–C2 overexpression in a wild type background (Fig. 4A,B). Unexpectedly, overexpression of H–C2 enhanced the net phenotype considerably: not only did the heterozygous net 1 mutants resemble the homozygotes, but the homozygotes developed massive patches of vein tissue and extensive blistering of the wing (Fig. 4C,D). Moreover, net 1; hs–H–C2 homozygotes were semi-lethal at 258C and the stock was only viable at 188C. Because net 1 is a complete null allele (Brentrup et al., 2000), this result excludes the simple model that H–C2 promotes vein development by inhibition of net activity. Instead, H–C2 acts independent of net either as veinpromoting factor, e.g. by activation of ve, or by repression of other negative regulators of ve that act in addition to net. The latter seems more plausible with regard to normal H function. In the heterozygous net 1 background which is phenotypically wild type (Fig. 4A) but sensitized for ectopic vein formation, the vein-promoting activity of H is revealed: overexpression of full length H in net 1 heterozygotes

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

resulted in ectopic venation which was a perfect phenocopy of the H–C2 effects (Fig. 4F; compare with Figs. 1, 3 and 4C). Apparently, wild type H has a vein-promoting activity which can be likewise explained by antagonizing a negative regulator of ve. 2.5. H–C2 functions upstream of ve and causes upregulation of ve within inter-vein areas As induction of ectopic vein material by H–C2 is extremely sensitive to developmental time, we reassessed the epistatic relationship with ve 1 by continuously overexpressing H–C2 or H with the aid of the Gal4/UAS system in a wild type and a ve 1 background. Prolonged overexpression of H–C2 with en–Gal4 (Fig. 4I) or Bx MS1096–Gal4 (not shown) driver lines resulted in conversion of most of the

9

distal inter-vein areas to vein tissue. Only the region between L3 and L4 proved resistant. Furthermore, we also observed induction of microchaetae in this area (Fig. 4I, arrow). Apart from this conversion, extensive tissue loss was noted while the wing margin itself remained intact. Overexpression of H–C2 in the ve 1 background showed both tissue loss and microchaetae on the wing blade (Fig. 4J, arrow). Again, no induction of ectopic vein material was observed in the more distal regions (Fig. 4J) in support of the notion that H–C2 acts upstream of ve. We propose that H–C2 promotes vein induction by up-regulating ve activity. Moreover, processes independent of ve are influenced by H– C2, which finally lead to the induction of ectopic bristles on the wing blade and to tissue loss. Overexpression of H in the same experimental set-up was unable to convert inter-veintissue into vein material apart from few ectopic veinlets. At the same time, H caused wing tissue and margin loss accompanied by broadened wing veins (Fig. 4K,L). These are the known hallmarks of an impaired Notch signaling (Shellenbarger and Mohler, 1978; Siren and Portin, 1989; de Celis and Bray, 1997) As vein determination depends on the balance between ve and bs activity (Roch et al., 1998), we examined the influence of H or H–C2 on wild type bs or ve gene expression. Full length H and H–C2 expression was ectopically induced from UAS-constructs in the posterior compartment via the en–Gal4 driver. ve expression was monitored with either enhancer trap line, ve rho–lac1 or ve X81(Bier et al., 1990; Free-

Fig. 5. Regulation of ve is altered by H–C2. The consequences of ectopic expression of H–C2 within the posterior compartment on expression of either ve and bs was compared with that in the anterior compartment or the wild type control at different times APF. Activity of ve was visualized with the enhancer trap line ve rho–lac1 in panels (A–C) and ve X81 in panels D and E. In all panels, gene activity of bs is shown in red (anti-SRF-antibody staining) and of ve in green (anti-b -galactosidase staining). (A) In the wild type, the expression patterns are complementary with ve marking the provein areas of the developing veins and bs the cells within the inter-vein regions, shown at early pre-pupal (,4 h APF, A) and pupal stages (,28 h APF, A 0 ). (B) Overexpression of full length H within the posterior compartment results in broadened pro-vein territories positive for ve (arrow). In the pre-pupal wing (,8 h APF, B) the broadening is apparent and becomes more pronounced in the pupal wing (,24 h APF, B 0 ). The genotype is [en– Gal4/1; [UAS–H]/ve rho lac1. (C) Overexpression of H–C2 within the posterior compartment does not considerably broaden the ve positive territories but leads to ectopic induction of ve expression within distal inter-vein areas (arrow). Whereas the pre-pupal wing shows only slight alterations (,6 h APF, C), vein fate induction is conspicuous in the pupal wing (,32 h APF, C 0 ). During this stage, bs and ve regulation is still complementary, but there are numerous cells where expression slightly overlaps (overlay appears yellow). Note the ability of H–C2 to downregulate bs without ve-induction at the anterior cross-vein (arrowhead). The genotype is [en–Gal4/[UAS–H– C2]; 1/ve rho–lac1. (D) Activity of the enhancer trap line ve X81 in early prepupal wings (,6 h APF) of the wild type. At this stage, the reporting power for ve activity has already faded except for some weak staining in L3 and L5 and at the posterior margin. (E) Overexpression of H–C2 within the posterior compartment elicits strong ectopic activity at the posterior margin (arrow) at around 6 h APF. This ectopic expression is the first sign of misregulation of ve by H–C2. The genotype is [en–Gal4/[UAS–H– C2]; 1 /ve X81.

10

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

man et al., 1992). Expression of ve and bs is mutually exclusive in vein and inter-vein tissue, respectively (Fig. 5A,A 0 ) (Montagne et al., 1996; Roch et al., 1998). Overexpression of H did not alter this complementary expression pattern, however, caused strong ve expression within and confined to pro-vein areas (Fig. 5B,B 0 ; arrow). This is in contrast to the effects of H–C2 overexpression where ectopic expression of ve rho–lac1 was very prominent in pupal wings already in early pupal stages (Fig. 5C) and remained strongly activated at least until 36 h after pupal formation (APF) (Fig. 5C 0 , arrow). Earlier wing development was analyzed using the ve X81–lacZ reporter. In wing discs from late third instar larvae, no patterning defects apart from tissue loss were observed. However, deviations from wild type became apparent already in the pre-pupal wing discs as early as 4 h APF (not shown). At 6 h APF, a strong and reliable ectopic staining near the wing margin, where pro-vein 5 develops, was observed (compare Fig. 5D with E, arrow) that later spread into the adjacent inter-vein field (Fig. 5C,C 0 ). Overall, ectopic ve expression reliably predicted the pattern of ectopic venation caused by overexpression of H–C2. This is in agreement with the hypothesis that ve regulation is a target of H–C2 activity.

unable to temporally resolve formation of ectopic veins (mostly cross-veins) from loss of veins (mostly longitudinals) (see Fig. 6C): at the same time and in the same wing, ectopic vein material was induced in the more proximal regions, and longitudinal veins were removed from the distal regions of the wing. These observations might also explain the counter-intuitive behavior of the Notch gain-of-

2.6. Pro-vein activity of Notch The simplest model to explain the differential effects of H and H–C2 on the regulation of ve expression would be the assumption of a vein inductive role of Notch in pre-pupal and early pupal development in agreement with the biphasic developmental role of Notch e.g. during eye development (Baker and Yu, 1997; Nagel and Preiss, 1999). There, Notch first promotes proneural fate before restricting it to single photoreceptor precursor cells in the course of lateral inhibition. By binding to Su(H), H would limit such a veinpromoting activity of Notch at an early inductive phase. Because H–C2 is unable to bind to Su(H), an assumed inductive Notch signal would be able to pass and thus, set the stage for ectopic veins. To test this assumption, we expressed an activated Notch receptor (N intra) under heat shock control during the H–C2 pheno-critical period (Fig. 2C). As hs–N intra overexpression at 378 proved extremely lethal to larvae and pupae alike, we performed the induction at lower temperatures of 34–358. Under these conditions, hs–N intra is able to induce ectopic veinlets, mostly of cross-vein character, in all the regions where H–C2 is also able to induce ectopic vein material (Fig. 6A–D; Rebay et al., 1993). We noted slightly different sensitivities, as area A was more sensitive, whereas area B never developed ectopic veinlets (see Fig. 2A,B). The overall vein-promoting effect of hs–N intra at 348C was not as strong as that of hs–H–C2 at 378C albeit clearly visible. These results demonstrate that Notch signaling is able to exert a positive influence on wing vein specification during early pupal stages, closely followed by the well characterized vein suppressing activity of Notch signaling (Rebay et al., 1993). Thus, we were

Fig. 6. Vein-promoting activity of Notch. Two copies of the hs–N intra construct were induced for 1 h at 34–358C at different time points after egg laying (AEL). Wings of emerging adults were analyzed in comparison to wild type control, where ectopic venation is never observed under these conditions. (A) Induction at 128–132 h AEL causes frequently the development of ectopic veinlets running from the posterior cross-vein through the distal inter-vein area. There are additional vein spots near L5 (arrows). (B) Ectopic vein spots between L4 and L5 as well as ectopic cross-veins (arrows) are observed after induction at 140–144 h AEL. (C) At 152–156 h AEL, the activity of N intra begins to erase distal parts of the longitudinal veins while it still induces the formation of ectopic cross-veins in the proximal regions (arrow). (D) Overexpression of two copies of hs–H–C2 for 1 h at 34–358C at 128–132 h AEL results in phenotypes comparable to the above with respect to ectopic vein formation, except for the bifurcation of L2 which was never observed after overexpression of N intra.

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

function and loss-of-function alleles, N Ax-E2 and N 55e11, during the modifier screen (Table 1). However, as described before in detail (e.g. Shellenbarger and Mohler, 1978; de Celis and Bray, 1997 and references therein), disruption of Notch signaling during the H–C2 pheno-critical period using the N ts1 allele or by generating N 55e11 mutant cell clones neither caused vein nor cross-vein loss nor a reduction in vein width (not shown). Both results as well as those from other authors strongly suggest that the presumptive Notch pro-vein activity is not critical for proper vein development in the pre-pupal and early pupal wing. However, ongoing Notch signaling at normal levels seems to be absolutely required for H–C2 to exert its inductive provein activity. 3. Discussion 3.1. A developmental switch during the H–C2 pheno-critical period The onset of pupariation is a major developmental switch, where expression of many genes as well as their developmental effects changes dramatically. This is, for example, observed in the regulation of ve and bs from third instar larval stage to early pupal stage: whereas the activation of both genes depends on pre-pattern genes like net early on, their regulation becomes inter-dependent and mutually exclusive about 4 h after puparium formation (Roch et al., 1998). We interpret the abrupt onset of H–C2 vein-promoting activity accordingly. Maybe, H–C2 responds to or influences the activity of other factors which only become available at that time and play a role in the promotion or repression of vein development. This is reflected by the onset of the positive influence of H–C2 on ve expression at around 4 h after pupariation. Thus, the H–C2 pheno-critical period might reflect a developmental switch for a requirement of H activity for vein fate decisions. 3.2. H–C2 and the Notch signaling pathway Involvement of the Notch signaling pathway in the refinement of proper vein width is well established (de Celis et al., 1997; Huppert et al., 1997). Current models suggest that during this lateral inhibition process, Dl acts as inhibitor of vein formation by directing cells, neighboring presumptive vein cells, into the inter-vein fate (de Celis et al., 1997). This model is in line with our observation that Dl mutants act as enhancers of H–C2 vein promotion. In Dl mutants, the threshold for vein fate is lowered as determined vein cells are less likely to be driven back into inter-vein fate. The interrelationship of H–C2 and Notch signaling during vein formation is, however, not restricted to the process of vein width refinement. Overexpression of N intra promotes early vein formation and may thus be setting the stage for provein development within inter-vein areas. Despite the fact that Notch activity is not necessary for pro-vein specifica-

11

tion itself, it is required for the vein-promoting activity of H–C2 as Notch mutations act as strong dominant suppressors of H–C2 effects. In agreement, N Ax-E2, a hyper-activated allele of Notch (Brennan et al., 1997; de Celis and Bray, 2000 and references therein), acts as a weak enhancer of H– C2 and it is even possible that this effect is initially stronger but then obliterated through enhanced lateral inhibition. 3.3. H–C2 and the EGF signaling pathway Reduction of the ve gene dose resulted in a very pronounced, dominant repression of the H–C2 phenotype despite the fact that the allele ve 1 has no dominant visible phenotype. This suggests that ve plays a crucial role for H– C2 to exert its inductive effects. Interestingly, dosage reduction of either, the Drosophila EGF receptor or the MAPK rolled (rl), had no dominant influence on the H–C2 ectopic venation phenotype (Table 2). The former result was unexpected and suggests that the Drosophila EGF-receptor itself is not rate limiting in this process. This notion is in line with the observation that also ve 1 is fully recessive in combination with loss-of-function alleles for the Drosophila EGFreceptor (data not shown). The allele rl 1 is a mild hypomorph (Lindsley and Zimm, 1992) and the reduction of MAPK activity might not be strong enough to influence the H–C2 phenotype. Together with the results of full epistasis of ve over H–C2, these data suggest that neither the Drosophila EGF-receptor nor the EGF signal transduction cascade are influenced by overexpression of H–C2. 3.4. H–C2 acts on the regulation of ve Overexpression of H results in the extension of ve expression all over the pro-vein area, whereas that of H–C2 induces in addition ve outside the pro-veins also within the inter-vein fields. Thus, both act positively on ve regulation but H–C2 is clearly different from H with respect to the apparent conversion of presumptive inter-vein- to pro-vein cell fates. Pro-vein activity is a normal aspect of H wild type function which is uncovered in a sensitized background: halving the gene dose of net or bs might result in a subtle increase of ve activity which can then be pushed by H above the threshold for pro-vein fate (Fig. 4F, and data not shown). Although our results clearly demonstrate that H and H–C2 act positively on the regulation of ve, we cannot conclude that this regulation is direct. Rather H might act negatively on the output of vein repressing factors. Since H has the capability to interact with a number of different proteins, overexpression of either H or H–C2 could influence stoichiometry of complexes or availability of factors involved in ve regulation. Two such factors are encoded by net and bs which have overlapping activity with regard to the negative regulation of ve, however, show a remarkably different temporal activity profile in that net acts during larval stage and bs at the transition of larval to pre-pupal stage (Sturtevant and Bier, 1995; Montagne et al., 1996; Roch et al., 1998; Brentrup et al., 2000). Thus, bs appears the

12

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

more likely target of H activity supported by the fact that H– C2 can promote vein induction even in the absence of net and that bs repression is observed at the anterior cross-vein without simultaneous ve induction (Fig. 5C 0 ). Whether this inhibition is direct or via as yet unknown factors and pathways requires further studies. 3.5. Dual activity of H during vein development As outlined above, both H and H–C2 have the ability to up-regulate ve. However, in contrast to H, H–C2 overexpression is capable of overriding inter-vein fate and thus inducing ve expression within inter-vein territories. Several lines of evidence suggest that this is a normal facet of H function. First, we find synergistic effects of combined overexpression of H and H–C2. Second, in a sensitized genetic background of reduced copies of inter-vein specifying genes like bs and net, H itself possesses vein-promoting activity just like H–C2. These results can be explained if one assumes a dual, independent activity for H, which is depicted in Fig. 7: on one hand, H might up-regulate ve, for example, by interfering with inter-vein specifying factors, an activity retained by H–C2. On the other hand, H, by virtue of binding to Su(H), antagonizes Notch-depen-

dent processes such as an early vein fate-promoting activity and subsequently vein width refinement. This activity is presumably lost in H–C2 due to the lack of the Su(H)-binding domain. Unfortunately, our attempts to test this hypothesis directly failed because we were unable to generate Su(H) mutant clones in the background of H–C2 (data not shown). A dual role of H in suppressing inductive Notch signaling and enhancing ve activity would explain why the wild type protein is unable to induce ectopic venation except in a sensitized background, where ve activity has already reached a critical threshold by the reduction of its negative regulators. In contrast, H–C2 can no longer antagonize Notch activity, but might still promote vein formation by interfering with ve suppression. Altogether, our genetic data may be taken as an example for a link between Notch and EGF signaling which in the context of vein formation appears to involve the activity of H influencing both, Notch signaling and, via ve, EGF signaling as well. 4. Experimental procedures 4.1. Fly work Flies were cultured on standard cornmeal food at 258C except for Gal4/UAS crosses which were raised at 178C. All stocks and alleles used in this paper are described by Lindsley and Zimm (1992), on the BDGP fruitfly or Flybase server, respectively, with the following exceptions: UAS– C2, the H–C2 deletion construct under UAS control (Maier et al., 1997; Schreiber, 2000); H wa (Bernd Johannes, unpublished), Df (3R) E(spl) X10 and Df(3R) E(spl) P11 (Wurmbach, 1998). Dominant interactions were studied after crossing virgins of respective genetic background to hs–H–C2 homozygous males. Parents were removed from the vials on day 5. After further 3 days, a double heat shock was applied as described below. Eclosing flies were collected in roughly 12 h intervals. Animals shocked between the fifth and the sixth day after egg laying were analyzed and scored. Selected combinations were set up three times independently to assess the accuracy of the classification scheme. 4.2. Phenotypic analysis

Fig. 7. Two different activities of H on vein fate establishment. Pro-vein fate (dark grey) requires the activity of ve, whereas inter-vein fate determination and specification (light gray) involves negative regulators (box) like net and bs. This negative regulation of ve might be inhibited by H and H–C2 alike, thereby allowing presumptive inter-vein cells to respond to pro-vein inductive cues leading in consequence to ve activation. Thus, overexpression of either H or H–C2 primes these cells for a fate change. One of these hypothetical cues might be provided by early Notch signaling. Notch signaling is, however, antagonized by wild type H protein. Therefore, overexpression of H does not result in ectopic vein formation because inductive Notch signals are inhibited. At later stages, Notch opposes vein formation by repressing ve activity: in the process of lateral inhibition, vein fate is restricted from pro-vein territories to the final vein width. Again, this process is antagonized by H.

Heatshock was induced by pulsed exposure to 37–388C for 1 h in a hybridization oven followed by a regeneration time of 1 h at 258C. Double heat shock was two times 1 h with a regeneration lag of 1 h at room temperature to prevent extensive larval and pupal lethality. The critical period for induction of the H–C2 wing phenotype was roughly determined by shocking a mixed culture and analyzing eclosing flies at daily intervals. For a detailed analysis of the pheno-critical time window, flies were allowed to lay eggs for 1 h and the synchronized culture was shocked at the indicated time of development. We took great care to exclude over- and underpopulated cultures

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

as these conditions are reported to cause severe deviations from the normal developmental time (Ashburner, 1989). Each experiment was paralleled with a wild type (Oregon R;w 1118 or y 1 w 67c23) and a negative control, containing only the mutant in a wild type or y 1 w 67c23 background. Flies selected for a detailed phenotypic analysis were transferred to a humid chamber without food and allowed to age for another 24 h, then dehydrated in ethanol (96%) for at least 4 h. For each mutant background, at least 60 wings were classified according to Fig. 1 and the average phenotypic level was calculated. The degree of modification was defined as the phenotypic difference between the mutant/ wild-type vs. the mutant/H–C2 combination. In summary, an overall reliability of ^1 phenotypic levels was achieved. Reduced viability of certain mutant backgrounds resulted in a considerable but erratic suppression of the H–C2 phenotype thereby reducing the probability for the identification of negative modifiers. Wings were dissected in ethanol and mounted in Euparal without drying. They were analyzed with a Zeiss Axiophot and photographs taken with Agfa Pan 25 black/white film. Pictures were scanned and processed using Adobe Photoshop or Corel Photopaint and assembled with Corel Draw. 4.3. Expression analyses For antibody stainings, third instar and early pre-pupal wing disks were dissected in ice-cold phosphate buffered saline (PBS) and subsequently fixed for 20–30 min in 4% paraformaldehyde in PBS at room temperature on a rocking tray. Pupae were freed from their pupal case and cut into two halves with fine scissors, refixed for additional 30 min while rotating vigorously and then washed once in PBS containing 0.1% Tween (PBT). The wings were excised from the cuticular envelope and refixed again for 20 min. Beyond this point, disks and pupal wings were treated the same. Fixative was removed by washing 3 £ 15 min with PBT and pre-incubated for 15–30 min with 3% normal goat serum in PBT. Afterwards, primary antibodies directed against bs (1:900; gift from M. Affolter) and b-galactosidase (1:2000; Promega) were added and incubated overnight at 48C. Secondary anti-mouse-DTAF (1:250) and anti-rat-Cy3 (1:150) (Jackson Lab) were incubated overnight at 48C. Washes were in PBT; tissue was mounted in Vectashield (Vector Lab). Images were collected with a confocal microscope (Bio-Rad MRC1024) and pictures processed and assembled using Corel Photopaint and Corel Draw. Acknowledgements We are indebted to S. Schreiber for providing the UAS– C2 construct and to A.C. Nagel and D. Maier for moral and technical support and helpful discussions. We are grateful to M. Affolter, M. Freeman, G. Struhl and U. Walldorf for fly lines and antibodies, respectively. We thank the Garcia-

13

Bellido and the Nu¨ sslein-Volhardt labs as well as the Stock Centers of Mid America, Umea˚ and Bloomington for sending us numerous fly stocks. We thank I. Wech and W. Staiber for technical hints and helping hands.

References Adachi-Yamada, T., Gotoh, T., Sugimura, I., Tateno, M., Nishida, Y., Onuki, T., Date, H., 1999. De novo synthesis of sphingolipids is required for cell survival by down-regulating c-Jun N-terminal kinase in Drosophila imaginal discs. Mol. Cell. Biol. 19, 7276–7286. Artavanis-Tsakonas, S., Rand, M.D., Lake, R.J., 1999. Notch signaling: cell fate control and signal integration in development. Science 5415, 770– 776. Ashburner, M., 1989. Drosophila, a laboratory handbook, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Baker, N.E., Yu, S.Y., 1997. Proneural function of neurogenic genes in the developing Drosophila eye. Curr. Biol. 7, 122–132. Biehs, B., Sturtevant, M.A., 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., 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. Brennan, K., Tateson, R., Lewis, K., Martinez-Arias, A., 1997. A functional analysis of Notch mutations in Drosophila. Genetics 147, 177–188. Brentrup, D., Lerch, H.-P., Ja¨ ckle, H., Noll, M., 2000. Regulation of Drosophila wing vein patterning: net encodes a bHLH protein repressing rhomboid and is repressed by rhomboid dependent Egfr signalling. Development 127, 4729–4741. Brou, C., Logeat, F., Lecourtois, M., Vandekerckhove, J., Kourilsky, P., Schweisguth, F., Israe¨ l, A., 1994. Inhibition of the DNA-binding activity of Drosophila Suppressor of Hairless and of its human homolog, KBF2/RBP-J kappa, by direct protein–protein interaction with Drosophila Hairless. Genes Dev. 8, 2491–2503. de Celis, J.F., 1997. Expression and function of decapentaplegic and thick veins during the differentiation of the veins in the Drosophila wing. Development 124, 1007–1018. de Celis, J.F., 1998. Positioning and differentiation of veins in the Drosophila wing. Int. J. Dev. Biol. 42, 335–343. de Celis, J.F., Bray, S., 1997. Feed-back mechanisms affecting Notch activation at the dorsoventral boundary in the Drosophila wing. Development 124, 3241–3251. de Celis, J.F., Bray, S., 2000. The Abruptex domain of Notch regulates negative interactions between Notch, its ligands Delta, Serrate and Fringe. Development 127, 1291–1302. de Celis, J.F., Bray, S., Garcia-Bellido, A., 1997. Notch signalling regulates veinlet expression and establishes boundaries between veins and interveins in the Drosophila wing. Development 124, 1919–1928. Diaz-Benjumea, F.J., Garcia-Bellido, A., 1990. Genetic analysis of the wing vein pattern of Drosophila. Roux Arch. Dev. Biol. 198, 336–354. Diaz-Benjumea, F.J., Gonzales-Gaitan, M.A.F., Garcia-Bellido, A., 1989. Developmental genetics of the wing vein pattern of Drosophila. Genome 31, 612–619. Freeman, M., Kimmel, B.E., Rubin, G.M., 1992. Identifying targets of the rough homeobox gene of Drosophila: evidence that rhomboid functions in eye development. Development 116, 335–346. Fristrom, D., Gotwals, P., Eaton, S., Kornberg, T.B., Sturtevant, M., Bier, E., Fristrom, J.W., 1994. blistered: a gene required for vein/intervein formation in wings of Drosophila. Development 120, 2661–2671. Guichard, A., Biehs, B., Sturtevant, M.A., Wickline, L., Chacko, J., Howard, K., Bier, E., 1999. rhomboid and Star interact synergistically

14

B. Johannes, A. Preiss / Mechanisms of Development 115 (2002) 3–14

to promote EGFR/MAPK signaling during Drosophila wing vein development. Development 126, 2663–2676. Huppert, S.S., Jacobsen, T.L., Muskavitch, M.A.T., 1997. Feedback regulation is central to Delta–Notch signalling required for Drosophila wing vein morphogenesis. Development 124, 3283–3291. Kla¨ mbt, C., 2002. EGF receptor signalling: roles of Star and Rhomboid revealed. Curr. Biol. 12, R21–R23. Lindsley, D.L., Zimm, G.G., 1992. The genome of Drosophila melanogaster, Academic Press, San Diego, CA. Lunde, K., Biehs, B., Nauber, U., Bier, E., 1998. The knirps and knirpsrelated genes organize development of the second wing vein in Drosophila. Development 125, 4145–4154. Maier, D., Stumm, G., Kuhn, K., Preiss, A., 1992. Hairless, a Drosophila gene involved in neural development, encodes a novel, serine rich protein. Mech. Dev. 38, 143–156. Maier, D., Marquart, J.B., Thompson-Fontaine, A., Beck, I., Wurmbach, E., Preiss, A., 1997. In vivo structure–function analysis of Drosophila Hairless. Mech. Dev. 67, 97–106. Maier, D., Nagel, A.C., Johannes, B., Preiss, A., 1999. Subcellular localization of Hairless protein shows a major focus of activity within the nucleus. Mech. Dev. 89, 195–199. Marquart, J., Alexief-Damianof, C., Preiss, A., Maier, D., 1999. Rapid divergence in the course of Drosophila evolution reveals structural important domains of the Notch antagonist Hairless. Dev. Genes Evol. 209, 155–164. Montagne, J., Groppe, J., Guillemin, K., Krasnow, M.A., Gehring, W.J., 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. Morel, V., Lecourtois, M., Massiani, O., Maier, D., Preiss, A., Schweisguth, F., 2001. Transcriptional repression by Suppressor of Hairless involves the binding of a Hairless–dCtBP complex in Drosophila. Curr. Biol. 11, 789–792. Nagel, A.C., Preiss, A., 1999. Notch spl is deficient for inductive processes in the eye, and E(spl) D enhances split by interfering with proneural activity. Dev. Biol. 208, 406–415.

Nagel, A.C., Maier, D., Preiss, A., 2000. Su(H) independent activity of Hairless during mechano-sensory organ formation in Drosophila. Mech. Dev. 94, 3–12. Nagel, A.C., Wech, I., Preiss, A., 2001. scalloped and strawberry notch are target genes of Notch signaling in the context of wing margin formation. Mech. Dev. 109, 241–251. Rebay, I., Fehon, R.G., Artavanis-Tsakonas, S., 1993. Specific truncations of Drosophila Notch define dominant activated and dominant negative forms of the receptor. Cell 74, 319–329. Roch, F., Baonza, A., Martin-Blanco, E., Garcia-Bellido, A., 1998. Genetic interactions and cell behaviour in blistered mutants during proliferation and differentiation of the Drosophila wing. Development 125, 1823– 1832. Schreiber, S., 2000. PhD thesis, Uni Hohenheim. Shellenbarger, D.L., Mohler, J.D., 1978. Temperature-sensitive periods and autonomy of pleiotropic effects of l(1)N ts1, a conditional Notch lethal in Drosophila. Dev. Biol. 62, 432–446. Siren, M., Portin, P., 1989. Interaction of Hairless, Delta, Enhancer of split and Notch genes of Drosophila melanogaster as expressed in adult morphology. Genet. Res. 53, 23–26. Struhl, G., Adachi, A., 1998. Nuclear access and action of Notch in vivo. Cell 93, 649–660. Sturtevant, M.A., Bier, E., 1995. Analysis of the genetic hierarchy guiding wing vein development in Drosophila. Development 121, 785–801. Sturtevant, M.A., Roark, M., 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., Biehs, B., Marin, E., 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. Wurmbach, E., 1998. PhD thesis, Uni Hohenheim. Yu, K., Sturtevant, M.A., Biehs, B., Franc¸ ois, V., Padgett, R.W., Blackman, R.K., Bier, E., 1996. The Drosophila decapentaplegic and short gastrulation genes function antagonistically during adult wing vein development. Development 122, 4033–4044.