The Enhancer of split complex of Drosophila melanogaster harbors three classes of Notch responsive genes

Mechanisms of Development 80 (1999) 171–180

The Enhancer of split complex of Drosophila melanogaster harbors three classes of Notch responsive genes Elisa Wurmbach, Irmgard Wech, Anette Preiss* Universita¨t Hohenheim, Institut fu¨r Genetik (240), 70593 Stuttgart, Germany Received 21 August 1998; accepted 6 November 1998

Abstract Many cell fate decisions in higher animals are based on intercellular communication governed by the Notch signaling pathway. Developmental signals received by the Notch receptor cause Suppressor of Hairless (Su(H)) mediated transcription of target genes. In Drosophila, the majority of Notch target genes known so far is located in the Enhancer of split complex (E(spl)-C), encoding small basic helix-loop-helix (bHLH) proteins that presumably act as transcriptional repressors. Here we show that the E(spl)-C contains three additional Notch responsive, non-bHLH genes: m4 and ma are structurally related, whilst m2 encodes a novel protein. All three genes depend on Su(H) for initiation and/or maintenance of transcription. The two other non-bHLH genes within the locus, m1 and m6, are unrelated to the Notch pathway: m1 might code for a protease inhibitor of the Kazal family, and m6 for a novel peptide.  1999 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Drosophila; Enhancer of split; Hairy-enhancer of split like genes (mammalian) (HES); Gene regulation; m4/ma; Notch pathway; Notch targets; Suppressor of Hairless

1. Introduction During the development of metazoans, cell type specification of many tissues requires local cell-cell communications. The Notch signaling pathway, named after the highly conserved group of trans-membrane receptors both in vertebrates and invertebrates, is instrumental in this type of cell fate choices (reviewed in Artavanis-Tsakonas et al., 1995). In Drosophila, the Notch pathway accounts for the differentiation of a very broad range of tissues from all three germ layers, the ectoderm (e.g. neuronal vs. epidermal cell fate), the mesoderm (e.g. muscle pioneer vs. fusion cell fate) and the endoderm (e.g. interstitial vs. principle larval midgut cell fate) (Corbin et al., 1991; Campos-Ortega, 1993; Tepass and Hartenstein, 1995). The general paradigm for Notch activity appears to postpone cellular differentiation until subsequent developmental signals are received (ArtavanisTsakonas et al., 1995). For example, in a process called ‘lateral inhibition’, single cells or small groups of cells of * Corresponding author. Tel.: +49-711-459-2206; fax: +49-711-4592211; e-mail: [email protected]

primary fate are selected amongst the many of the same potential. Due to the Notch signal, the latter cannot follow this developmental path and eventually differentiate with a secondary fate. Several players of the Notch pathway involved in lateral inhibition have been identified (reviewed in Artavanis-Tsakonas et al., 1995). Notch ligands are presented by the presumptive primary cell and activate the Notch pathway in their direct neighbors. Once the Notch receptor is activated, the signal is mediated by Su(H) which is involved in the transcriptional activation of Notch target genes (see Kopan and Turner, 1996 and references therein). Notch target genes include the HES (hairy-enhancer of split like genes) class of genes encoding small proteins with a basic helixloop-helix (bHLH) motif that presumably act as transcriptional repressors (Jennings et al., 1994; de Celis et al., 1996; Lee, 1997). Other Notch controlled processes, for example inductive events during the morphogenesis of the imaginal wing, involve structurally different classes of Notch target genes (Diaz-Benjumea and Cohen, 1995; Micchelli et al., 1997). In vertebrates, the family of HES proteins comprises four

0925-4773/99/$ - see front matter  1999 Elsevier Science Ireland Ltd. All rights reserved. PII: S09 25-4773(98)002 12-3

172

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

members (Lee, 1997). In Drosophila there are seven members clustered within the E(spl) complex (E(spl)-C), m3, m5, m7, m8, mb, mg and md, generically named E(spl) bHLH genes or proteins, respectively (Delidakis and ArtavanisTsakonas, 1992; Knust et al., 1992). Two other HES-like Drosophila genes, hairy and deadpan, are located elsewhere in the genome and are both Notch independent (Rushlow et al., 1989; Bier et al., 1992). Within the E(spl)-C, there are five other genes, m1, m2, m4, m6 and ma, that are located in the center of the complex, intermingled only by the bHLH genes m3 and m5 (Schrons et al., 1992) (Fig. 1). With the exception of m4, nothing is known about their molecular structure. Sequence analysis of m4 has been uninformative towards its possible function since it encodes a novel protein (Kla¨mbt et al., 1989). Interestingly, m4 is transcribed in response to Notch receptor activation like the bHLH genes and is dependent on the Su(H) signal transducer (Bailey and Posakony, 1995). In accordance, the spatio-temporal distribution of m4 transcripts is indistinguishable of some E(spl) bHLH transcripts (Knust et al., 1992; Singson et al., 1994; de Celis et al., 1996). Based on these observations we addressed the question, whether other genes within the E(spl) locus also belong to the Notch signaling pathway. Here we report that two more

genes, ma and m2, are Notch responsive. Structurally, ma is very similar to m4, and both genes depend on Su(H) for transcriptional activation like the E(spl) bHLH genes. The other gene, m2 encodes a novel protein. Despite its responsiveness to Notch receptor activation, transcriptional activation of m2 is not strictly dependent on Su(H) which appears essential for maintenance of transcription. The remaining two genes, m6 and m1, do not belong to Notch pathway genes: they encode a novel peptide and a prospective protease inhibitor of the Kazal family, respectively. Altogether, these data indicate that the E(spl)-C harbors several, structurally divergent classes of Notch target genes.

2. Results 2.1. Sequence analysis of the non-bHLH genes in the E(spl) locus The E(spl) locus contains five non-bHLH genes, m1, m2, m4, m6 and ma that are split by the bHLH genes m3 and m5 (Fig. 1). Only m4 has been characterized at the molecular level and its novel sequence is uninformative regarding its function (Kla¨mbt et al., 1989). Our molecular analysis of the

Fig. 1. The central E(spl) locus. (A) Schematic drawing of the central part of the E(spl)-C covering the transcription units mb to m7; E(spl) bHLH genes are boxed. The arrows indicate the orientation of transcription of the respective genes. Hybridization probes are indicated as hatched bars. The orientation is proximal to distal from left to right; the scale above the map is given in kilo bases (kb). Restriction sites are: H, HindIII; R, EcoRI; S, SalI; X, XhoI; P, PstI (selected site). Polymorphic sites are marked with an asterisk. (B) Transcripts of ma, m1, m2 and m6 presented at a larger scale. Presumptive TATA-box (T) and poly-adenylation (pA) sites are indicated. In m1 and ma, the latter sites differ from the optimal consensus at the fifth position (T and C instead of A, respectively. Translated segments are cross-hatched. Translation initiation site according to Cavener (1987) is optimal for all the genes except for m6. The 5′ region of m2 contains a single, degenerate Su(H) binding site at − 686 [S]; that of ma contains two perfect sites at − 341 and at − 761 (on the opposite strand), and two divergent sites at − 624 and at − 634. There are two Brd-boxes [B] in the ma 3′-UTR at + 517 and + 630 as well as two K-boxes [K] are + 540 and at + 613. Single K-boxes are present in the trailers of m2 at + 889 and of m6 at + 1056. See text for further details. Divergent Su(H) binding sites are marked with an asterisk.

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

four other non-bHLH genes now reveals that ma is structurally related with m4 (Fig. 2). The similarity of Ma and M4 is 43% and the identity is 36% and extends over the entire length of both proteins. Moreover, both show a presumptive PKC-phosphorylation site at similar positions. Fig. 2 shows an alignment of the two proteins. Leviten et al. (1997) reported a similarity between Bearded (Brd) and M4 sharing 16% of the amino acids. Indeed, the two regions of similarity noted by the authors are also well conserved between M4 and Ma (Fig. 2). Furthermore, in transcripts of both Brd and m4 genes, three common regulatory sequence motifs within the 3′ untranslated regions (3′-UTR) are found, the ‘Brd box’, the ‘GY box’ and the ‘K box’ which are supposed to confer transcript instability (Leviten et al., 1997; Lai et al., 1998). Like in m4, the sequence motif of the Brd box is found twice in the 3′-UTR of ma mRNA at similar positions (+517 and +630) but no GY box. None of the other four non-bHLH E(spl)-C genes contains either Brd or GY box (Fig. 1). The K box appears to be more common. It is found twice in the 3′-UTR of ma (+540 and +613) and each once in that of m2 (+889) and of m6 (+1056) (Fig. 1), which is within the same range of E(spl) bHLH genes. It should be noted that this motif is also detected within the ma 5′ region (−521) which is rather unusual (Lai et al., 1998). The two genes m2 and m6 encode novel proteins. The presumptive M2 protein comprises 218 amino acids containing a short opa repeat and a conspicuous serine/asparagine-motif ([S3NXNS3]2; X = hydrophobic). M6 is a rather short polypeptide of 70 amino acids without detectable motifs (Fig. 3). Sequence analysis of m1 indicates that it encodes a serine protease inhibitor of the Kazal family, based on the similarity in topological arrangement of presumptive disulfide bridges (Laskowski and Kato, 1980; Kumazaki and Ishii, 1990; Rupp et al., 1991). As shown

173

in Fig. 3, M1 could contain two reactive sites. Kazal inhibitors are often stored in zymogen granules and are secreted with the zymogens in the pancreatic juice (Laskowski and Kato, 1980). 2.2. Expression analysis of the non-bHLH genes in the E(spl) locus Because of the high degree of structural similarity between ma and m4, the expression patterns of both genes during embryonic development and in imaginal discs were compared. As shown in Fig. 4A, embryonic expression patterns are nearly indistinguishable, and appear very similar to those of E(spl) bHLH genes, particularly of m5, m7 and m8 (Knust et al., 1992). Characteristic for this group of genes is the early mesectodermal expression which appears shortly before the onset of gastrulation. Later on, transcripts of ma and m4 are detected in the neuro-ectoderm as well as in the mesoderm in a highly dynamic pattern in many stages of embryogenesis. In imaginal discs, the expression domains of ma and m4 are similar to those described for different E(spl) bHLH genes (Fig. 4) (Singson et al., 1994; de Celis et al., 1996). Transcripts of m4 accumulate primarily within presumptive proneural clusters of eye-antennal, wing and leg discs, a pattern remarkably similar to that of m8 or m7 expression (Fig. 4B) (Singson et al., 1994; de Celis et al., 1996). However, ma transcripts are detected in a pattern matching very closely that of mb expression. In the eye disc, ma is expressed not only within but also posterior to the morphogenetic furrow. In the wing pouch, staining of presumptive intervein regions and wing margin is apparent. In the leg disc as well as in the notal part of the wing disc, a more general expression is observed with highest concentration in areas encompassing proneural clusters (Fig. 4B) (de Celis et al., 1996). The expression patterns suggest that

Fig. 2. Protein sequence comparisons of M4 and Ma. Alignment of predicted protein sequences of Ma and M4. Regions of high similarity are boxed, gaps are indicated by a dash. A possible PKC phosphorylation site is printed in bold. The dark bar marks a presumptive basic-amphipatic helix, originally identified by Leviten et al., 1997) in a comparison of Brd with M4. Both, M4 and Ma might also contain such a structural element, which is however, remarkably different: the basic amino acids of the M4 helix are flanked by acidic amino acids (E, glutamate); the basic face of the Ma helix is much wider because it is regularly disrupted by glycines. The second region of high similarity observed between Brd and M4/Ma is marked with a striped bar.

174

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

Fig. 3. Sequences of M1, M2 and M6. Predicted protein sequences of M1, M2 and M6 are shown. The M1 protein shows similarity to serine protease inhibitors of the Kazal family; two presumptive reactive sites are underlined. Both, M2 and M6 show no similarities to known proteins. M2 contains a conspicuous regular array of serine residues ([S3NXNS3]2; X = hydrophobic; underlined) that might contain several phosphorylation sites and a polyglutamine stretch (dotted underlined).

both genes are under the same regulatory control as are the different E(spl) bHLH genes and thus, might serve a role in Notch mediated cell differentiation as well. Surprisingly, also m2 transcripts accumulate in a pattern reminiscent of the transcript distribution of E(spl) bHLH genes although there are no structural similarities with either the bHLH or the m4/ma genes (Figs 3, and 5). At first, m2 transcripts are detected in a very dynamic pattern in the neuro-ectoderm of stage 9 embryos (Fig. 5A). At stage 10/11 the transcripts accumulate at high levels in the presumptive mesoderm (Fig. 5B), however, disappear quickly with the onset of germ band retraction (Fig. 5C). Imaginal disc expression is rather weak. In eye disc, m2 transcripts are observed close to as well as posterior of the morphogenetic furrow. In the wing disc areas of proneural clusters stain weakly, as does the dorso-ventral boundary and vein/

intervein regions (Fig. 5G,H). These patterns are typical of E(spl) bHLH gene expression (de Celis et al., 1996). Therefore, m2 appears to be regulated in a similar manner like E(spl) bHLH and m4/ma genes and could also serve as Notch target gene. Unlike the other E(spl)-C genes, the m6 gene is expressed within neuronal cells in the embryo. Transcription starts at low level during early stage 9 (not shown). Later in development, m6 mRNA accumulates within the developing central nervous system (Fig. 5E). After germ band retraction, the transcripts are detected in the brain, the ventral nerve chord and the presumptive peripheral nervous system as well (Fig. 5F). The m6 gene is weakly expressed in imaginal tissues. Specific staining appears in the notal anlagen of the wing disc (Fig. 5I). As deduced from the expression patterns m6 might play a role in the differentiation of neuronal tis-

Fig. 4. Comparison of m4/ma expression patterns. (A) Transcript accumulation of ma and m4 genes during embryogenesis is shown; embryos are at stage 6, 8, 11 and 12 from left to right, respectively (according to Campos-Ortega and Hartenstein, 1985. Please note early mesectodermal expression during gastrulation (stages 6 through 7), the neuro-ectodermal (stage 8/9) and the mesodermal (stage 10/11) accumulation of the respective transcripts. During germ band retraction (stage 12) expression vanishes rapidly and remains strongest in thoracic clusters and the procephalic region. (B) Expression of ma and m4 genes in the eye-antennal disc, the wing disc and the leg disc (from left to right) is shown. In the eye disc, these transcripts accumulate in a stripe of cells close to the morphogenetic furrow which is typical of E(spl) bHLH genes. In addition, ma gene expression is also detected in the differentiating eye field posterior to the furrow. Weak staining is also observed in the second antennal segment in areas, where presumptive sensory organs develop. In the wing and the leg disc, transcripts of ma accumulate predominantly in areas encompassing proneural clusters. In the wing pouch, expression is observed in intervein regions and along the presumptive wing margin. Thus, the ma expression domains match very well those of mb. In contrast, m4 transcripts accumulate nearly exclusively in proneural cluster regions of the wing and the leg disc, yielding a pattern similar to that of m8 or m7 expression. Fig. 5. Expression patterns of m2, m1 and m6. In situ hybridizations of the respective probes to wild type embryos and discs are shown as indicated. Transcript accumulation in embryos of the given stages of (A) m2, stage 8 (dorsal view), (B) m2, stage 10/11, (C) m2, stage 14, (D) m1, stage 13/14, (E) m6, stage 10, (F) m6, stage 14 as well as in imaginal discs of (G) m2, eye-antennal disc, (H) m2, wing disc and (I) m6, wing disc, are shown. Embryos are oriented with anterior to the left and dorsal up; lateral views are shown (except in A. Staging of embryos is according to (Campos-Ortega and Hartenstein, 1985).

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

Fig. 4

Fig. 5

175

176

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

Fig. 6. Notch ectopic expression induces gene activity. An activated Notch receptor (Nintra) was ectopically expressed in a pair rule pattern in developing embryos. As a consequence, all E(spl)-C genes except m1 and m6 are transcriptionally activated, as revealed by in situ hybridization with gene specific probes as indicated. This shows that the three non-bHLH gene, ma, m4 and m2 are also Notch target genes. In order to visualize endogenous gene expression as well, older embryos are shown for m6 and m1.

sue, unlike the predicted function of E(spl) bHLH genes (Artavanis-Tsakonas et al., 1995). An unrelated pattern is observed for m1, where transcripts accumulate transiently at the fusion sites of anterior and posterior midgut and very specifically to high levels in the proventriculus of the embryo (Fig. 5D and Fig. 6). Assuming that m1 gene products might have a role of Kazal inhibitors, expression would be expected in the digestive tract and maybe also in the proventriculus (Strasburger, 1932; Laskowski and Kato, 1980). 2.3. Regulation of E(spl)-C gene expression by Notch pathway activation It has been shown previously that transcriptional expression of some E(spl)-C genes encoding bHLH transcription factors is activated in response to Notch signaling (Jennings et al., 1994, 1995; Bailey and Posakony, 1995; Lecourtois and Schweisguth, 1995). In order to elucidate, whether amongst the newly characterized genes were further Notch responsive genes, we systematically compared all bHLH and non-bHLH genes in the E(spl) locus for their Notch responsiveness. Truncation of the extracellular domain of Notch results in a constitutively activated Notch receptor (see Artavanis-Tsakonas et al., 1995 and references therein). Such an activated Notch receptor (Nintra) was expressed in a zebra pattern employing the Gal4/UAS system (Brand and Perrimon, 1993). These

embryos were subsequently hybridized with probes specific for the particular genes of the E(spl) locus. If the transcription of the respective gene is dependent on the Notch signal, it should be expressed in a striped pattern. As shown in Fig. 6, all genes within the E(spl) locus except m1 and m6 are activated in seven stripes, indicating that these genes are regulated by the Notch pathway during embryogenesis. From this we conclude that the E(spl)-C comprises three different classes of Notch responsive genes, the 7 bHLH genes, ma/m4 and m2. 2.4. Su(H) is the transmitter of Notch signaling to ma, m4 and m2 The transcription factor Su(H) seems to provide a direct link between the activated Notch receptor and Notch target genes. In accordance, the promoter regions of the E(spl)-C genes md, mg, m4, m5 and m8 contain respective Su(H) protein binding sites (Bailey and Posakony, 1995; Lecourtois and Schweisguth, 1995; Eastman et al., 1997). Here we investigated, whether this is also true for ma and m2. We generated embryos homozygous mutant for Su(H) derived from germ-line clones. These Su(H) mutant embryos were hybridized in situ with probes for m2, ma, and m4 (as control), respectively, and compared with the heterozygous siblings that bear one Su(H) zygotic copy and, thus, develop normally (Lecourtois and Schweisguth, 1995). As shown in

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

Fig. 7, ma and m4 transcripts are nearly absent in the neuroectoderm of the mutant embryos. This result is very similar to that observed for E(spl) bHLH genes (Jennings et al., 1994; Jennings et al., 1995). In accordance, the ma promoter region contains two putative Su(H) binding sites that fit exactly the consensus sequence [(C/T)GTG(G/A)GAA(C/ A)], and two potential sites that differ in their last position (T instead of C/A at position 9) (Fig. 1), a variation observed before for a genuine Su(H) binding site in the md promoter (Eastman et al., 1997). Similarly, the promoter region of m4 contains three perfect Su(H) binding sites (Bailey and Posakony, 1995). In contrast to m4 and ma, a considerable amount of m2 transcripts is found in Su(H) mutant embryos especially in the dorsal ectoderm, although at a reduced level (Fig. 7). Apparently, Su(H) is not strictly required for the initiation of m2 transcription, but seems to augment it. Only at later stages of development, m2 transcripts disappear suggesting that Su(H) is essential for the maintenance of m2 transcription. The m2 promoter region contains a single divergent Su(H) binding site which differs at two positions, the fifth and the ninth, from the consensus (Fig. 1). Similarly divergent Su(H) binding sites have been shown to bind the Su(H) protein in vitro: for example the site mg5 differs at three positions, the first, seventh and ninth, from the consensus and is bound by Su(H) protein expressed in S2 cell culture (Eastman et al., 1997). This, together with the observation that m2 expression is weaker in the Su(H) mutant embryos suggests that this sequence is a bona fide Su(H) binding site.

Fig. 7. Dependence of transcription on Su(H). Su(H) mutant embryos derived from germ line clones were hybridized with ma, m4 and m2 probes as indicated. In comparison to embryos fertilized with wild type sperm (right column), mutant embryos (left column) show little signal when probed with m4 or ma. In contrast to this, m2 transcripts are well detected in both classes of embryos of similar age (stage 8/9. As shown in the lowest row, m2 expression disappears during later development (stage 11) in the mutant embryos to completely vanish by stage 13.

177

3. Discussion 3.1. The E(spl)-C contains three different classes of Notch responsive genes The work presented here extends our current understanding on the structure and function of the E(spl) gene complex. Initially, the E(spl) locus was defined by genetic means particularly in the context of neurogenesis. Loss of Notch signaling during this process results in a misrouting of presumptive epidermal cells into the neural pathway and, as a consequence, embryos die from a hypertrophied nervous system – the ‘neurogenic phenotype’ (Lehmann et al., 1983). Later it was found that only the concomitant loss of several genes in the E(spl) locus resulted in such a neurogenic phenotype (Schrons et al., 1992). Although mutations in the only vital gene of the locus, l(3) groucho (gro) also resulted in an enlarged nervous system, this phenotype was rather weak and occurred with low penetrance, indicating that gro could not account solely for E(spl) function (Preiss et al., 1988). Instead, other genes within the vicinity had to be important for Notch signaling. Subsequent molecular analyses revealed that the region contains seven structurally related genes that all encode presumptive transcriptional regulators of the bHLH class which are expressed in very similar patterns in embryos (Kla¨mbt et al., 1989; Delidakis and Artavanis-Tsakonas, 1992; Knust et al., 1992). These genes seem functionally overlapping explaining the failure to identify individual mutations based on phenotype. The similar embryonic expression suggested a common mechanism of transcriptional regulation. Indeed, it was shown, that E(spl) bHLH genes are transcriptionally activated in response to the Notch signal involving Su(H) as signal transducer (Jennings et al., 1994; Jennings et al., 1995; Bailey and Posakony, 1995; Lecourtois and Schweisguth, 1995; Eastman et al., 1997). In conclusion, E(spl) bHLH genes serve as Notch targets, thereby implementing the Notch signal during neurogenesis. Thus, these genes were considered equivalent to E(spl)-C. Our work now reveals, that by the definition of Notch target genes, the E(spl)-C has to be extended to three further genes, m4, ma and m2. These three genes are structurally unrelated to the bHLH genes and thus, represent new classes of Notch target genes. The high degree of similarity between m4 and ma suggests that the respective gene products have similar functions and might, at least in part, be able to replace each other. This might explain why mutations in either gene have gone undetected. In contrast, m2 is structurally very different and represents up to now a single member of its kind within the complex. Nevertheless, so far no mutations in m2 have been isolated (Ziemer et al., 1988; A.P., I.W. and B. Johannes, unpublished). Thus, m2 appears non-essential by genetic means and its function might be either insignificant or compensated by other E(spl)-C gene products.

178

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

3.2. m4 and ma constitute a new gene family Based on their structural similarity, m4 and ma seem to constitute a novel gene class which might be grouped together with Brd to a new gene family (Leviten et al., 1997). Brd has a role in the promotion of adult sensory organ development of Drosophila and, therefore, acts contrary to Notch signaling which operates by preventing neural commitment. Genetic interactions suggest a functional connection between Brd and the Notch pathway (Leviten and Posakony, 1996). The mode of Brd function remains still elusive. The most prominent structural feature, a presumptive basic amphipatic a-helical domain is largely conserved between all three proteins. However, significant differences regarding the basic portion of the helix might account for different functional qualities of the respective proteins (Fig. 2). This is supported by the notion that Brd protein is slightly basic overall (pI 8.65) whilst M4 and Ma are rather acidic (pI of 4.7 and 4.6, respectively). Furthermore, Brd lacks the highly conserved C-terminal domain of M4/Ma. The structural similarities between M4/Ma and Brd might reflect a common biochemical mode of action. Yet, as expected for targets of Notch signaling, M4/Ma might still exert opposite functions of Brd, i.e. repression of neural fate instead of its promotion (as discussed in Leviten et al., 1997). However, it is also conceivable that M4/ Ma provide antagonistic activities which are set free in response to the activated Notch receptor, maybe to balance or to silence the signaling cascade. A further similarity between Brd and m4/ma genes, specific sequence motifs within the 3′-UTR, is common to all E(spl) bHLH genes but mb (Lai et al., 1998; this work). Two of these motifs, the Brd box and the K box, have been associated with transcript destabilization suggesting that the mRNAs of all these genes are short lived (Lai et al., 1998). Even though the significance of these elements is not yet fully understood, they might contribute to control protein levels of the individual genes to secure a given stoichiometry amongst the respective gene products. 3.3. Evolution of the E(spl)-C Only a limited number of Notch target genes are known up to date and the vast majority, ten genes, is located within the E(spl)-C (Artavanis-Tsakonas et al., 1995; Kopan and Turner, 1996; this work). These ten genes represent three structurally different gene classes and two gene families that have members elsewhere in the genome. The E(spl) bHLH genes belong to the HES gene family together with hairy and deadpan. Although both, hairy and deadpan are involved in neuronal specification as well, their roles are clearly separable from that of E(spl) bHLH genes (Rushlow et al., 1989; Bier et al., 1992). In contrast, E(spl) bHLH gene activity appears largely overlapping (Schrons et al., 1992). We do not yet know, whether Brd and m4/ma exert similar or opposite functions. However, genetically m4/ma seem to

be redundant as well. Functional redundancy amongst these family members located within the E(spl) locus is not surprising assuming that they have arisen by gene duplication in the past. Less expected is the high degree of conservation during evolution indicating that this arrangement is fairly old (Maier et al., 1993). The other two genes located within the E(spl)-C, m1 and m6 appear unrelated to developmental mechanisms involving the Notch signaling pathway. Their place within the complex might be accidental, for example due to a transposition event. Alternatively, they might have lost their Notch responsiveness secondarily and gained new functions. In any case, the present arrangement of the E(spl) gene region appears advantageous to the fitness of the fly. Why would this arrangement be supported by evolutionary selection? Especially during imaginal stages, E(spl)-C genes are transcribed in remarkably different, albeit overlapping patterns. Thus, E(spl)-C gene expression cannot be solely regulated through Notch signaling but rather, additional factors appear to contribute to tissue specific regulation (de Celis et al., 1996). Based on the expression patterns in the wing disc, E(spl)-C genes can be roughly grouped in three classes: one expressed in nearly all proneural clusters comprising m8, m7 and m4; a second expressed also in the presumptive wing blade comprising ma and mb and the third active only in a small subgroup of proneural clusters comprising mg and md. Two genes, m2 and m3 are expressed rather weakly, and m5 not at all in the discs (de Celis et al., 1996; this work). This classification is very coarse, not regarding subtle differences. However, it highlights the point that the nearest neighbors show the most similar expression patterns. It might be that the genes within each group share regulatory elements which are targets for the above postulated spatially active regulatory factors. This can be easily envisaged for the ma/mb pair. These two genes are transcribed in the opposite direction, separated by approximately 5 kb that might harbor such shared elements (Fig. 1). The explanation appears more difficult for the other two gene groups. Both neighboring gene pairs, m7/m8 as well as mg/md, are transcribed in the same orientation making enhancer sharing less likely. Furthermore, m4, m7 and m8 span together a rather large region of more than 15 kb. However, regulatory elements with such long range effects, acting beyond independently regulated genes, are not without precedence in a gene complex (Gorman and Kaufman, 1995) and could indeed explain why these genes are grouped together and why this arrangement has been conserved in the course of evolution. In summary, the most significant finding of this work is the fact that the E(spl)-C comprises three structurally different classes of Notch responsive genes. This finding was completely unexpected and points to a surprising complexity in the response of a given cell to Notch receptor activation. Clearly, much remains to be learned about the function of these genes, in particular the two new classes, m4/ma and m2. Only a few examples of Notch target genes besides the

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

E(spl) bHLH genes are known to date, and they work primarily during imaginal development (Artavanis-Tsakonas et al., 1995; Kopan and Turner, 1996). Apart from the E(spl) bHLH genes, only m4 was known to be expressed also during embryogenesis in response to Notch signaling. Our work now places m4 together with ma within a new class of genes that are targets of the Notch signaling pathway. It will be interesting to find out, whether there are additional members of this class in the Drosophila genome and what their role within this pathway might be.

4. Materials and methods 4.1. Expression analysis Ectopic Notch expression in a zebra pattern in the embryo was induced with the GAL4/UAS-system (Brand and Perrimon, 1993), by crossing UAS-Notchintra and prd-GAL4 flies (obtained from S. Bray and C. Delidakis, respectively). Su(H) germline clones were generated with the Su(H) alleles SF8 and AR9 as described in Lecourtois and Schweisguth (1995) employing the FLP-ovoD technique developed by Chou and Perrimon (1992). In situ hybridizations were performed according to the protocol of Tautz and Pfeifle (1989) with modifications taken from Lehner and O’Farell (1990). 4.2. Molecular analysis A cDNA library made of RNA from 3–12-hour-old embryos (Poole et al., 1985) was screened according to standard protocols (Sambrook et al., 1989). The following genomic fragments spanning the respective genes were used for the screening (Fig. 1): a 3.7 kb EcoRI fragment for ma (+13.5/ + 17.2); a 2.9 kb HindIII/SalI fragment for m1 (+9.2/ + 12.1); a 2.9 kb HindIII/Sau3A fragment for m2 (+5.4/ + 8.3); and a 1.9 kb XhoI/PstI fragment for m6 (−8/ − 9.9) (coordinates are according to Preiss et al. (1988) (Fig. 1). For sequencing, cDNA inserts and fragments thereof were subcloned into Bluescript vectors (Stratagene). Sequencing was performed by chain termination; reactions were separated on an ALF sequencer according to the procedure of the manufacturer (Pharmacia). Both strands were sequenced, gaps closed by the use of walking primers. Computer analyses were carried out on genius.embnet.dkfz-Heidelberg using the program HUSAR 4.0 which is based on GCG software (University of Wisconsin, USA; Devereux et al., 1984).

Acknowledgements We are grateful to S. Bray, C. Delidakis and F. Schweisguth for providing numerous fly stocks. Special thanks go to A. Ephrussi for providing the logistics and infrastructure for digital image recordings and to A. Jenny and B. Johannes

179

for computer aid. We are indebted to S. Bray, A. Jenny and A. Nagel for helpful discussions and critical reading of the manuscript. The respective sequences have been deposited in the EMBL databank under the accession numbers m1: AJ010167, m2: AJ010168, m6: AJ012204, ma: AJ011140. This work was supported by a grant from the Deutsche Forschungsgemeinschaft (DFG PR 533/1-1) to AP. EW was a Montrepos fellow of the Universita¨t Hohenheim. References Artavanis-Tsakonas, S., Matsuno, K., Fortini, M.E., 1995. Notch signaling. Science 268, 225–232. Bailey, A.M., Posakony, J.W., 1995. Suppressor of Hairless directly activates transcription of Enhancer of split complex genes in response to Notch receptor activity. Genes Dev. 9, 2609–2622. Bier, E., Vaessin, H., Younger-Shepherd, S., Jan, L.Y., Jan, Y.N., 1992. deadpan, an essential pan-neural gene in Drosophila, encodes a helixloop-helix protein similar to the hairy gene product. Genes Dev. 6, 2137–2151. Brand, A.H., 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., 1993. The Development of Drosophila melanogaster. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Campos-Ortega, J.A., Hartenstein, V., 1985. The Embryonic Development of Drosophila melanogaster. Springer, New York. Cavener, D.R., 1987. Comparison of the consensus sequence flanking translational start sites in Drosophila and vertebrates. Nucl. Acids Res. 15, 1353–1361. Chou, T.B., Perrimon, N., 1992. Use of a yeast site-specific recombinase to produce female germline chimeras in Drosophila. Genetics 131, 643– 653. Corbin, V., Michelson, A.M., Abmayr, S.M., Neel, V., Alcamo, E., Maniatis, T., Young, M.W., 1991. A role for the Drosophila neurogenic genes in mesoderm differentiation. Cell 67, 311–323. de Celis, J.F., de Celis, J., Ligoxygakis, P., Preiss, A., Delidakis, C., Bray, S., 1996. Functional relationships 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. Delidakis, C., Artavanis-Tsakonas, S., 1992. The Enhancer of split [E(spl)] locus of Drosophila encodes seven independent helix-loophelix proteins. Proc. Natl. Acad. Sci. USA 89, 8731–8735. Devereux, J., Haeberli, P., Smithies, O., 1984. A comprehensive set of sequence analysis programs for the VAX. Nucl. Acids Res. 12, 387– 395. Diaz-Benjumea, F.J., Cohen, S., 1995. Serrate signals through Notch to establish a wingless-dependent organizer at the dorsal/ventral compartment boundary of the Drosophila wing. Development 121, 4215– 4225. Eastman, D.S., Slee, R., Skoufus, E., Bangalore, L., Bray, S., Delidakis, C., 1997. Synergy between Suppressor of Hairless and Notch in regulation of Enhancer of split mg and md expression. Mol. Cell. Biol. 17, 5620– 5628. Gorman, M.J., Kaufman, T.C., 1995. Genetic analysis of embryonic cisacting regulatory elements of the Drosophila homeotic gene Sex combs reduced. Genetics 140, 557–572. Jennings, B., Celis, J., Delidakis, C., Preiss, A., Bray, S., 1995. Role of Notch and achaete-scute complex in the expression of Enhancer of split bHLH proteins. Development 121, 3745–3752. Jennings, B., Preiss, A., Delidakis, C., Bray, S., 1994. The Notch signalling pathway is required for Enhancer of split bHLH protein expression

180

E. Wurmbach et al. / Mechanisms of Development 80 (1999) 171–180

during neurogenesis in the Drosophila embryo. Development 120, 3537–3548. Kla¨mbt, C., Knust, E., Tietze, K., Campos-Ortega, J.A., 1989. Closely related transcripts encoded by the neurogenic gene complex Enhancer of split of Drosophila melanogaster. EMBO J 8, 203–210. Knust, E., Schrons, H., Grawe, F., 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. Kopan, R., Turner, D.L., 1996. The Notch pathway: democracy and aristocracy in the selection of cell fate. Curr. Opin. Neurobiol. 6, 594–601. Kumazaki, T., Ishii, S., 1990. Disulfide bridge structure of ascidian trypsin inhibitor I: similarity to Kazal-type inhibitors. J. Biochem. (Tokyo) 107, 414–419. Lai, E.C., Burks, C., Posakony, J.A., 1998. The K box, a conserved 3′UTR sequence motif, negatively regulates accumulation of Enhancer of split complex transcripts. Development 125, 4077–4088. Laskowski, M. Jr., Kato, I., 1980. Protein inhibitors of proteinases. Annu. Rev. Biochem. 49, 593–626. Lecourtois, M., Schweisguth, F., 1995. The neurogenic Suppressor of Hairless DNA-binding protein mediates the transcriptional activation of the Enhancer of split complex genes triggered by Notch signaling. Genes Dev. 9, 2598–2608. Lee, J.E., 1997. Basic helix-loop-helix genes in neural development. Curr. Opin. Neurobiol. 7, 13–20. Lehmann, R., Jimenez, F., Dietrich, U., Campos-Ortega, J.A., 1983. On the phenotype and development of mutants of early neurogenesis in Drosophila melanogaster. Roux Arch. Dev. Biol. 192, 62–74. Lehner, C.F., O’Farell, P.H., 1990. The roles of Drosophila cyclins A and B in mitotic control. Cell 61, 535–547. Leviten, M., Lai, E., Posakony, J., 1997. The Drosophila gene Bearded encodes a novel small protein and shares 3′-UTR sequence motifs with multiple Enhancer of split complex genes. Development 124, 4039– 4051. Leviten, M.W., Posakony, J.W., 1996. Gain-of-function alleles of Bearded interfere with alternative cell fate decisions in Drosophila adult sensory organ development. Dev. Biol. 176, 264–283. Maier, D., Marte, B.M., Scha¨fer, W., Yu, Y., Preiss, A., 1993. Drosophila evolution challenges postulated redundancy in the E(spl) gene complex. Proc. Natl. Acad. Sci. USA 90, 5464–5468.

Micchelli, C.A., Rulifson, E.J., Blair, S.S., 1997. The function and regulation of cut expression on the wing margin of Drosophila: Notch, Wingless and a dominant negative role for Delta and Serrate. Development 124, 1485–1495. Poole, S.J., Kauvar, L.M., Drees, B., Kornberg, T., 1985. The engrailed locus of Drosophila: structural analysis of an embryonic transcript. Cell 40, 37–43. Preiss, A., Hartley, D.A., Artavanis-Tsakonas, S., 1988. The molecular genetics of Enhancer of split, a gene required for embryonic neural development in Drosophila. EMBO J 7, 3917–3927. Rupp, F., Payan, D.G., Magill-Solc, C., Cowan, D.M., Scheller, R.H., 1991. Structure and expression of a rat agrin. Neuron 6, 811–823. Rushlow, C.A., Hogan, A., Pinchin, S.M., Howe, K.R., Lardelli, M.T., IshHorowicz, D., 1989. The Drosophila hairy protein acts in both segmentation and bristle patterning and shows homology to N-myc. EMBO J. 8, 3095–3103. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York. Schrons, H., Knust, E., Campos-Ortega, J.A., 1992. The Enhancer of split complex and adjacent genes in the 96F region of Drosophila melanogaster are required for segregation of neural and epidermal progenitor cells. Genetics 132, 481–503. Singson, A., Leviten, M.W., Bang, A.G., Hua, X.H., Posakony, J.W., 1994. Direct downstream targets of proneural activators in the imaginal disc include genes involved in lateral inhibitory signaling. Genes Dev. 8, 2058–2071. Strasburger, M., 1932. Bau, Funktion und Variabilita¨t des Darmtraktus von Drosophila melanogaster Meigen. Z. wiss. Zool. 140, 539–649. Tautz, D., 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. Tepass, U., Hartenstein, V., 1995. Neurogenic and proneural genes control cell fate specification in the Drosophila endoderm. Development 121, 393–405. Ziemer, A., Tietze, K., Knust, E., Campos-Ortega, J.A., 1988. Genetic analysis of Enhancer of split, a locus involved in neurogenesis in Drosophila melanogaster. Genetics 119, 63–74.