Developmental Biology 326 (2009) 212–223

Contents lists available at ScienceDirect

Developmental Biology j o u r n a l h o m e p a g e : w w w. e l s e v i e r. c o m / d e v e l o p m e n t a l b i o l o g y

SOCS36E specifically interferes with Sevenless signaling during Drosophila eye development Isabel Almudi a, Hugo Stocker b, Ernst Hafen b, Montserrat Corominas a, Florenci Serras a,⁎ a b

Departament de Genètica, Facultat de Biologia and Institut de Biomedicina (IBUB), Universitat de Barcelona, Diagonal 645, 08028 Barcelona, Spain Institute of Molecular Systems Biology (IMSB), ETH Zürich, Wolfgang-Pauli-Strasse 16, 8093 Zürich, Switzerland

a r t i c l e

i n f o

Article history: Received for publication 19 July 2008 Revised 17 November 2008 Accepted 17 November 2008 Available online 30 November 2008 Keywords: Sevenless Socs36E Photoreceptor Cell specification Signal transduction

a b s t r a c t During the development of multicellular organisms the fate of individual cells is specified with great precision and reproducibility. Although classical genetic approaches led to the identification of many of the signaling pathways contributing to cell fate specification, they have provided little insight into the mechanisms that ensure robustness and reproducibility. We have used the specification of the R7 photoreceptor cells controlled by the Sevenless receptor tyrosine kinase (Sev) pathway to screen for modulators of pathway activity and to uncover the mechanisms underlying the robustness of cell fate decisions. Here we provide genetic evidence that the Drosophila SOCS36E adaptor protein containing an SH2 domain and a SOCS box acts as an attenuator of Sev signaling. Overexpression of Socs36E strongly suppresses the specification of extra R7 photoreceptor cells in response to constitutive activation of Sev, and loss of Socs36E function suppresses the loss of R7 cells when Sev activity is impaired. In a wild-type background, however, loss and gain of Socs36E function exhibits little effect on R7 specification. We also show that SH2 domain of SOCS36E is essential for this function in inhibiting Sev action and that Socs36E expression is suppressed by high Sev pathway activity. In our model, only the cell able to activate high levels of receptor tyrosine kinase signaling will repress SOCS36E expression, reduce the negative effect on Sev signaling and allow this cell to differentiate into R7. In contrast, the remaining cells fail to receive high signaling, and thus maintain high levels of SOCS36E. This represses residual Sev activity and blocks R7 development. Therefore, Socs36E constitutes a novel partially redundant feedback mechanism that contributes to the robustness of R7 specification. The SOCS family of adaptor proteins may have evolved as modulators of specific signaling pathways that contribute to the robustness and precision of cell fate specification. © 2008 Elsevier Inc. All rights reserved.

Introduction Cell-to-cell signaling is commonly regarded as the most important mechanism to drive cell specification. Many signaling pathways and their associated signal transduction molecules have been genetically and biochemically characterized. However, little is known about how these pathways are regulated to ensure the robustness and reproducibility of cell fate decisions. Signaling events must be regulated in space and time to activate specific genetic programs at the right place and moment in order to avoid wrong developmental decisions. Very likely a network of regulatory molecules will limit the range and duration of signaling activity that drives cell specification. Feedback mechanisms and specific signaling inhibitors of receptor activation are examples that could account for such a network of controlling molecules (Freeman and Gurdon, 2002). The development of the fly retina has been a key model to isolate and identify molecules involved in signaling between cells. The specification of photoreceptor cells occurs in the third instar eye ⁎ Corresponding author: Fax: +34 93 4034420. E-mail address: [email protected] (F. Serras). 0012-1606/$ – see front matter © 2008 Elsevier Inc. All rights reserved. doi:10.1016/j.ydbio.2008.11.014

imaginal disc with great precision to ensure proper function of the visual system. The eight photoreceptors of each ommatidium are recruited in a stepwise fashion by local cell interactions. High activity of the Ras/MAPK pathway controlled by the EGF receptor is important for the specification of R1–R7 photoreceptor cells (Freeman, 1996). But in the case of R7 specification an additional burst of Ras/MAPK controlled by the Sevenless (Sev) receptor tyrosine kinase is required (Freeman, 1996; Simon et al., 1991). The Sev receptor is expressed in nine cells in each ommatidial cluster, the precursors of the R1/R6, R3/ R4, R7 photoreceptors and four cone cells, known as the Sev equivalence group (Tomlinson et al., 1987). However, only the R7 precursor will activate the Sev receptor by the Bride of sevenless (Boss) ligand, which is expressed in the adjacent R8 cells (Reinke and Zipursky, 1988). Sev null alleles lack R7 cells in all ommatidia (Tomlinson and Ready, 1986) and the precursors that normally should differentiate into R7 will now trigger the non neural cone cell fate. Conversely, the constitutive activation of Sev induces additional cells of the Sev equivalence group to adopt an R7 photoreceptor cell fate (Basler et al., 1991). The tight spatial and temporal regulation of Ras/ MAPK activity is therefore essential to ensure the specification of the precise number and arrangement of photoreceptor cells.

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

In search for novel negative regulators of the Sev pathway we performed an EP based genetic screen for suppressors of constitutively activated sev transgene and isolated Suppressor of cytokine signaling 36E (Socs36E) as a gene involved in Sev signaling modulation. SOCS proteins are conserved from flies to mammals and were initially identified as inhibitors of cytokine signaling pathways by acting through a negative feedback loop involving the inhibition of Janus kinase activity (JAK/STAT signaling pathway) (Endo et al., 1997; Hilton et al., 1998; Starr et al., 1997; Yoshimura et al., 1995). Socs36E encodes the Drosophila homolog to the mammalian SOCS-5 (Callus and Mathey-Prevot, 2002; Karsten et al., 2002) and, like other members of the SOCS protein family, contains a SH2 domain flanked by a variable N-terminal domain and a conserved C-terminal domain, termed the SOCS box. The SH2 domain binds phosphorylated tyrosine residues, whereas the SOCS box participates in an ubiquitin ligase complex to promote the degradation of target proteins (Zhang et al., 1999). In addition to the JAK/STAT pathway, SOCS proteins may also regulate signaling pathways, including receptor tyrosine kinases (Baetz et al., 2004; Callus and Mathey-Prevot, 2002; Kario et al., 2005; Krebs and Hilton, 2003; Rawlings et al., 2004). However, little is known about how SOCS36E can modulate those receptors in physiological conditions. Because cells can respond to Sev only for a restricted period of ommatidial recruitment (Basler and Hafen, 1989a,b; Bowtell et al., 1989) and because only the R7 precursor activates Sev, we decided to explore whether SOCS36E is involved in the regulation of Sev activity in the equivalence group to single out the R7. We provide genetic evidence that SOCS36E specifically acts as attenuator of Sev signaling in cells that express the Sev receptor but do not differentiate into R7. Moreover, SOCS36E targets the Sev receptor through its SH2 domain to block signaling transduction. Additionally, high levels of Ras/MAPK repress Socs36E in precursors of R7. Our results show that SOCS36E constitutes a novel partially redundant feedback mechanism that contributes to the robustness of R7 specification. Materials and methods Fly constructs and mutants The insertion site of the EP(34-120y+) line was determined by plasmid rescue after EcoRI or XbaI digestion of genomic DNA. EP(34120y+) is inserted 2 kbp upstream of the Socs36E open reading frame. The activated sev construct sevS11 (Basler et al., 1991) and recombinant flies sev-Gal4 sevS11/TM3 were used for the screening and genetic interactions. The following fly strains were used: UAS-Socs36E, UASSocs36E-SH2⁎, UAS-Socs36E-ΔSB (Callus and Mathey-Prevot, 2002), UAS-Socs44A (Rawlings et al., 2004), and the hypomorphic allele sevd2 (Δ22) which consists of a null sevd2 allele partially rescued by a constitutive sevΔ22 construct (Bohmann et al., 1994). Activation of Egfr was achieved using the gain-of-function alleles Elp B1, Elp 1, or ectopically activation of the following constructs: UAS-torDEGFR (Dominguez et al., 1998) or UAS-λtop (Queenan et al., 1997); sevRasV12 (Fortini et al., 1992) and sev-RafTorY9 (Dickson et al., 1996) for activation of the pathway; UAS-IR against Socs36E was used to induce RNAi (VDRC stock center); Df(2L)Exel7070 (Bloomington Stock Center) is a deficiency that uncovers a region including Socs36E; the EP line DrkEP(2)2477 (Bloomington stock center) to ectopically activate Drk; the JAK/STAT alleles osupd1, hopc111, Stat92E06346 (all from Bloomington) and UAS-upd (from Hou, S); sev-Nnucl and sev-Nact (Fortini et al., 1993) and sev-lz (Flores et al., 2000) transgenes were used to ectopically express Notch and lozenge respectively under the sev enhancer; GMRGal4 (Hay et al., 1994) and sev-Gal4 were used to drive expression of transgenes. CantonS was used as wild type. All crosses and fly culture were done in standard fly medium. When single UAS transgenes had to be compared to double transgenes, a UAS-GFP construct was added to avoid effects of titration of the Gal4.

213

P-element excision The P{EPgy2}Socs36EEY06665 flies carry an insertion in the second exon of Socs36E. These flies were crossed to Δ2-3 transposase to excise the inserted element and to obtain a loss-of-function mutation of Socs36E. Genomic DNA extraction and PCR were performed with a pair of primers flanking the insertion point (upper primer: gccggcggaagtgcgtcag; lower primer: cagcgtgggcggtgtgga) in order to check whether those excisions resulted in complete or partial removal of the inserted element. The partial or imprecise excisions were sequenced using the same primers to molecularly define the deletions obtained. Whole-mount in situ hybridization In situ hybridization using digoxigenin-labeled antisense RNA probes was carried out using standard protocols. DIG-labeled riboprobes for Socs36E were synthesized using a complete cDNA clone from DGC (SD04308), sequenced using primers from the SP6 and T7 promoters and linearized with EcoRI for antisense probe, and SalI for the sense probe. To test the specificity of the riboprobe we used apGal4 to drive expression either UAS-Socs36E or UAS-Socs44A in the dorsal compartment of the wing disc, allowing us to compare the dorsal with the ventral domain of the same disc. Antisense probe strongly hybridized in the ap domain only when the Socs36E transgene was activated (data not shown). Scanning electron microscopy and histology Flies were dehydrated in 25, 50, 70, 90, 95 and 100% ethanol for 24 h each to prepare samples for scanning electron microscopy (SEM). To get rid of accumulated debris in the eyes, flies were sonicated for 30 s in an ultrasound bath followed by a final change of 100% ethanol. Flies were critical-point dried and coated with gold to be examined in a Hitachi S-2300 scanning microscope. Adult flies were fixed, dehydrated and embedded in Spurr's medium. Semithin sections were obtained and stained with methylene blue for analysis under a Leica DMLB microscope. For each genotypic combination, ommatidia from 3 to 5 different eyes were counted in blind analysis. A Chi square test on contingency tables was performed. This allowed us to see significant differences in number of R7 per ommatidium comparing control versus experimental flies. Antibody production and western blot analysis 5′ cDNA of Socs36E was inserted into a pPRO-EX-HTa expression vector (Invitrogen) to produce a fusion protein with 6 ×His residues. A histidine-tagged protein of ∼ 48 kDa from bacterial extract was purified using His-Select Nickel Affinity Gel (Sigma). The purified SOCS36E protein was injected into rabbits and rats to generate polyclonal antibodies. To test the SOCS36E antibody, total protein extracts were obtained from embryos (0–24 h) by homogenizing 50 µl of dechorionated embryos in standard loading buffer. The extracts were processed and analyzed using SDS-PAGE and Western blot transfer standard protocols. Immunodetection was performed using rabbit antiSOCS36E antibody (1:3000) and detected with goat anti-rabbit peroxidase (1:3000) secondary antibody with EZ-ECL system (Biological industries Ltd., Kibbutz Beit Haemek, Israel). For immunohistochemistry, rat or rabbit anti-SOCS36E were used 1:500 in blocking buffer and incubated overnight at 4 °C. Immunohistochemistry and bioimaging The P(GawB)NP5170-5-1 line (Drosophila Genetic Resource Center) has inserted a Gal4 sequence in the first intron of the Socs36E gene, which promotes the expression of UAS-GFP to trace cells expressing Socs36E.

214

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

Antibody staining on imaginal discs was done using a standard protocol. Primary antibodies used in this study were: mouse anti-Elav (1:100 Developmental Studies Hybridoma Bank, DSHB), mouse antiProspero (1:4, DSHB), mouse anti-Cut (1:100, DSHB) and mouse antiRough (1:100, DSHB). Secondary antibodies were obtained from Jackson Immuno Research anti-mouse-Rhodamine Red (1:200) and anti-mouse-Cy5 (1:200). To stain the actin cytoskeleton, Rhodaminecoupled phalloidin (Molecular Probes) was used at 1:40 dilution for 30 min after disc fixation. Double fluorochrome-labeled samples were analyzed and captured using an Olympus confocal microscope. Images were processed using NIH ImageJ software. Final artwork was processed using Adobe Photoshop 7.0 software. To analyze differences in Socs36E expression between pre-R7 and pre-cone cells, we compared pixel intensity between the GFP mean pixilation (green channel) in Pros positive cells (red channel) from Socs36E-Gal4; UAS-GFP (n = 2 discs) and Socs36E-Gal4;UAS-GFP sevS11 (n = 5 discs) eye imaginal discs. Measurements were done using RGB Measure plugin from NIH ImageJ software. Green/Red mean pixel ratios were calculated in order to normalize the data from different cells. These ratios were grouped in two independent populations: preR7 cells and pre-cone cells for each genotype. Data were analyzed using two independent non-parametric statistical tests, the Mann– Whitney and the Kolmogorov–Smirnov tests, to investigate whether there were differences between those populations. Northern blot analysis Total RNA was extracted from larvae by using Trizol (GIBCO/BRL) and mRNA purified with polyA-TRACT system (Promega). The Socs36E probe

used for this analysis was a 667 bp PCR product (exons 3–4). The probe was labeled with [32P]dCTP by random priming. Hybridization was carried out for 1 h at 68 °C and after washing membrane was exposed to film at −80 °C. Results Socs36E overexpression suppresses activated Sevenless Overexpression of an N-terminally truncated version of the Sev receptor (sevS11) in the Sev equivalence group results in the formation of supernumerary R7 cells in each ommatidium and a concomitant disturbance of the hexagonal lattice of the ommatidial units. The degree of roughness of the eye surface is a sensitive measure for the number of extra R7 cells and indicates the levels of Sev pathway activity during the specification of photoreceptor cells. In search for novel negative regulators of the Sev signaling pathway, we aimed at identifying genes whose expression is capable of suppressing the sevS11 rough eye phenotype. To this end, we performed a UAS (upstream activating sequences)-Gal4-based EP (enhancer/promoter) screen (Rorth, 1996; Rorth et al., 1998) to randomly express genes together with the sevS11 mutation (Sese et al., 2006). Fly lines containing random insertions of a double-headed EP transposable element carrying 3′ UAS and 5′ UAS sites, which permit the transcription of genes flanking the insertion in response to Gal4, were crossed to flies carrying both the sevS11 construct and the sevGal4 driver. The EP(34-120y+) line strongly suppressed the sevS11 rough eye phenotype and excision of the 5′ UAS site by cre-loxP mediated recombination yielded a single headed EP(y−) element that also suppressed the sevS11 phenotype (Figs. 1A–C). This suppression

Fig. 1. Socs36E overexpression suppresses sevS11. (A) Double-headed EP transposable element EP(34-120y+) showing the 5′ and 3′ UAS sites (black arrowheads) and LoxP sites. The 3′ UAS drives the expression of Socs36E gene. (B) Rough eye phenotype due to the expression of an activated Sev receptor (sevS11). (C) Suppression of the sevS11 phenotype by the expression of EP(34-120y−) under the control of the sev-Gal4 driver. (D–E). The suppression of the irregular pattern of ommatidia and the presence of extra R7 cells per ommatidium are observed in semithin sections of eyes of sevS11 flies (D) caused by the activation of UAS-Socs36E under the sev-Gal4 driver (E). (F) Histogram showing the distribution profiles of number of R7 photoreceptors per ommatidium (D) and (E). Statistical analysis reveals significant differences between the genotypes (p b 0.001). (G) Activation of Socs36E by sev-Gal4 in a wild-type background results in a loss of R7 photoreceptors in some ommatidia. R7 rhabdomeres are colored in green.

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

was not due to the EP insertion site as there was no suppression without the Gal4 driver (Fig. 1B). Plasmid rescue revealed the Suppressor of cytokine signaling 36E (Socs36E) as the gene activated by the insertion of EP(34-120y+) (Fig. 1A). Indeed, the expression of a UASSocs36E transgene under the control of sev-Gal4 resulted in a suppression of the sevS11 rough eye as well in a suppression of the extra R7 cells (p b 0.001; Figs. 1D–F). Moreover, overexpression of Socs36E by sev-Gal4 in a wild-type background was able to interfere with R7 specification. The resulting eyes lacked R7 cells in 7% of the ommatidia (n = 228) and exhibited a rough appearance (Fig. 1G). Thus, overexpression of Socs36E interferes with the specification of R7 cells.

215

R7 per ommatidium. Eyes from flies in which Socs36EEY11 was placed over the deficiency Df(2L)Exel7070 that uncovers the Socs36E locus exhibited extra R7 cells in 2.7% (n = 333) of ommatidia with extra R7 (not shown). We reproduced the same phenotype by expressing a Socs36E RNAi construct (Dietzl et al., 2007) under the control of GMR-Gal4 or sev-Gal4 (Fig. 2F). Moreover, under conditions in which Sev activity is limiting, the loss of R7 cells was partially suppressed by a loss of Socs36E function. Flies carrying the hypomorphic allele sevd2(Δ22) lack R7 cells in 10% of the ommatidia (n = 268). Homozygosity for Socs36EEY11 in sevd2(Δ22) eyes resulted in missing R7 cells in only 2% of the ommatidia (n =532). From these results we conclude that Socs36E is a partially redundant negative regulator of R7 cell specification.

Socs36E is required for R7 inhibition Socs36E is expressed in the developing ommatidial clusters To get insight into the function of Socs36E we used the transposable element P{EPgy2}Socs36EEY06665 inserted into the second exon of Socs36E (Bellen et al., 2004). Flies homozygous for this insertion are viable and exhibit normal eyes and photoreceptor organization. We generated an imprecise excision of this transposable element to disrupt the gene. Sequencing the resulting excision allele, hereafter termed Socs36EEY11, showed an insertion of 500 bp that corresponds to the truncated transposable element in the second exon of the Socs36E gene (Fig. 2A). Although Northern blot analysis demonstrated the presence of a transcript of higher molecular weight, a stop codon is located at the residue 179, which should give rise to a truncated protein without the SH2 and SOCS box domains (Figs. 2A, B). In fact, Western blot analysis using polyclonal antibody against SOCS36E showed that the 68 kDa band, which corresponds to the predicted molecular weight, was missing in the mutant flies (Fig. 2B). Because the Socs36EEY11 mutant was homozygously viable, we conclude that loss of Socs36E function is dispensable for viability and fertility. However, eyes of Socs36EEY11 mutant flies were rough (Figs. 2C, D). In semithin sections of mutant eyes, we observed the presence of extra R7 photoreceptors in 2% (n = 310) of the ommatidia examined (Fig. 2E) in contrast to wild type flies, which always develop only one

It has been reported previously that Socs36E is expressed in the morphogenetic furrow, a signaling center of the eye disc (Karsten et al., 2002). We have re-examined the expression of Socs36E using a riboprobe synthesized from a complete cDNA clone (SD04308) and found that Socs36E is expressed throughout the eye disc with high levels in the developing ommatidia, in addition to the morphogenetic furrow. Within the ommatidial cluster the accumulation of Socs36E mRNA was surprisingly high in some cells and low or even undetectable in others (Figs. 3A–C), suggesting that Socs36E expression is tightly regulated during eye development, particularly during the later stages of ommatidial development behind the morphogenetic furrow. An intriguing possibility to explain the heterogeneity in Socs36E mRNA distribution is that cells of the Sev equivalence group express high levels of Socs36E to prevent the specification of multiple R7 cells, whereas the R7 precursor cell exhibits low levels of Socs36E thus permitting high Ras/MAPK signaling. To explore this possibility, we used a Socs36E-Gal4 line to drive expression of the UAS-GFP reporter (hereafter Socs36E-GFP). Those discs were stained for multiple markers to identify the cells that expressed Socs36E. The expression

Fig. 2. A loss-of-function mutation of Socs36E results in extra R7 photoreceptor cells. (A) Diagram of 500 bp insertion in the Socs36EEY11 line. The sequence in blue corresponds to the P-element sequence, whereas the sequence in pink corresponds to the Socs36E gene. The 21 nucleotides resulting from the imprecise excision are represented in black. Note the stop codon (red) with an asterisk. (B) Upper panel: Northern blot shows a 3500 bp transcript in wild type (wt) and 4000 bp transcript in the Socs36EEY11 line; lower panel: Western blot showing the absence of the 68 kDa protein in Socs36EEY11. (C–D) Wild-type and Socs36EEY11 eyes, respectively. (E) Semithin section from Socs36EEY11 eyes shows some ommatidia with extra R7 photoreceptors. (F) The same phenotype is reproduced in semithin sections from a GMR-Gal4; UAS-RNAi-Socs36E eye.

216

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

Fig. 3. Socs36E expression pattern in the third instar eye imaginal disc. (A–C) Socs36E in situ hybridization; antisense (A) and sense riboprobe (B) reveal high levels of Socs36E expression posterior to the morphogenetic furrow (black arrowhead) and in the developing ommatidia. (C) Detail from the posterior part of the eye imaginal disc. Note that some cells in the ommatidia are expressing high levels of Socs36E. (D–D″) Some photoreceptor precursor cells, labeled with anti-Elav (red), are expressing high levels of Socs36E (GFP). (E–E″) Detail from developing ommatidia marked with anti-Elav (red), and Socs36E (GFP). (F–F″) Anti-Rough (Ro) staining (red) reveals that R3 and R4 precursor cells express high levels of Socs36E, whereas R2 and R5 precursor cells do not express it. The adjacent R1 and R6 express intermediate levels of Socs36E. (G–G″) Cone cells, labeled with anti-Cut (red), also express Socs36E. (H–H″) R7 precursor cells, stained with anti-Prospero (red), do not express Socs36E. (I–I″′) Localization of anti-SOCS36E is restricted to the ommatidial clusters and morphogenetic furrow (arrowhead). The specified neural cells of the clusters were counterstained with anti-Elav and the ommatidial organization and morphogenetic furrow with phalloidin. (J–J″) Detail of SOCS36E localization in photoreceptors. Upper panels: tangential sections; lower panels: optical cross section (white diagonal line) through outer photoreceptors. (K–K″) Detail of SOCS36E localization in R7 photoreceptors. Upper panels: tangential sections; lower panels: optical cross section (white diagonal line) through R7 photoreceptors co-stained with anti-Pros. Scale bars: 10 µm.

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

of Socs36E-GFP was highly concentrated at the posterior region of the third instar disc, though low levels of GFP were also found in the rest of the disc (Figs. 3D–D″). As for Socs36E mRNA, the distribution of Socs36E-GFP was heterogeneous. Immunolabeling with the panneural marker Elav revealed some photoreceptors with high levels of GFP, some with low and some others that lacked GFP (Figs. 3E–E″) altogether. Using the anti-Rough (Ro) antibody as a specific marker for R2, R3, R4 and R5 precursors, we established that the R3 and R4 photoreceptor cells expressed Socs36E-GFP strongly, whereas R2 and R5 did not (Figs. 3F–F″). In optical sections, the photoreceptors R1 and R6, which are located adjacent to R2 and R5, respectively, also showed intermediate levels of GFP. Another photoreceptor lacking GFP was identified as R7 by staining with an antibody against Prospero (Pros) (Figs. 3H–H″). Anti-Prospero also stains the non-neural cone cells, which can be distinguished in a different confocal plane, or more specifically with the cone cell marker Cut. Using both markers we were able to show localization of strong Socs36E-GFP activity in cone cells (Figs. 3G–G″). Thus, the levels of Socs36E expression closely correlated with the fate of the cells within the R7 equivalence group. Cells whose fate is not determined by Sev, such as R3 and R4, or cells that do not become R7 because Sev is not activated, such as the cone cells, express high levels of Socs36E. On the other hand, no Socs36E expression is observed in the R7 cells whose fate depends on the activation of the Sev pathway. We also used anti-SOCS36E to evaluate the pattern of expression in the eye. We found that SOCS36E localized in some photoreceptors (R3, R4, R1 and R6) and cone cells but was missing or poorly expressed in R7, R2 and R5 (Figs. 3I–K), indicating that the pattern assessed with the antibody is very similar to that observed with the Socs36E-GFP reporter. As already found for the Socs36E mRNA the protein was detected in the morphogenetic furrow, suggesting that the expression of the reporter line is only partially reproducing the endogenous expression. This difference could be due to an enhancer element specific for the ommatidia development targeted with Socs36E-GFP. In later stages of ommatidial development, when all photoreceptors are already specified at the late third instar or early pupae, both SOCS36E protein and the Socs36E-GFP expression, were localized in all cells including R7 (not shown). This suggests that SOCS36E could also act in R7 when high levels of Ras are not longer required. Because the localization of Socs36E-GFP is sharper than the anti-SOCS36E (compare Fig. 3D with Fig 3I), in further experiments we have used that reporter line to monitor Socs36E expression.

217

Socs36E expression is repressed by high Sev signaling To explore whether high levels of Sev activity are sufficient to repress Socs36E expression, we recombined sevS11 with Socs36E-Gal4 UAS-GFP and checked for high Pros expression as a nuclear marker of R7 cells (Figs. 4A–B). Indeed, eye discs of sevS11 larvae had an increased number of cells with low levels of Socs36E compared with discs of wild-type larvae. Moreover, many of the cells that had low or no Socs36E expressed Pros, suggesting that Socs36E expression is repressed in ectopic R7 cells. We quantified the Socs36E-GFP expression by measuring pixel intensity of Pros positive cells for the red (Pros) and green (GFP) channels (see Materials and methods). In wild-type eye imaginal discs, GFP intensity in cone cells was six fold higher than in R7 precursors (R7 median = 0.051; cone cell median = 0.386; p b 0.001). In sevS11 discs, few cone cells still present in ommatidia showed pixel intensities similar to wild-type cone cells, whereas the extra R7 cells of sevS11 ommatidia showed a significantly lower GFP intensity (R7 median = 0.130; cone cells median = 0.345; p b 0.001). In addition, we reproduced the same observation by using flies with activated Ras protein under the sev promoter (sev-RasV12). In sev-RasV12 eye imaginal discs all Pros positive endogenous and extra R7 cells lacked GFP expression (Figs. 4C–D). SOCS36E genetically interacts with Sev Drk, the fly ortholog to the mammalian Grb2, is the adaptor protein that links Sev to Sos and its downstream effectors Ras and MAPK (Olivier et al., 1993; Simon et al., 1993). Drk/Grb2 contains an SH2 domain flanked on each side by an SH3 domain. We hypothesized that SOCS36E suppressed Sev signaling activity by antagonizing Drk for binding to the activated Sev receptor through the SH2 domain. To test this hypothesis, we determined whether the suppression of the sevS11 induced multiple-R7 phenotype by Socs36E overexpression is counteracted by overexpression of drk. Overexpression of UAS-DrkEP(2)2477 (Kraut et al., 2001) under the control of sev-Gal4 resulted in flies with rough eyes and further enhanced the rough eye phenotype of sevS11 flies. Only 20% (n = 51) of ommatidia contained three or less R7 cells (Fig. 5A). In contrast, 95% (n = 338) of the ommatidia of sevS11 flies overexpressing Socs36E contained three or less R7 cells (Fig. 5B). Eyes of sevS11 flies co-expressing UAS-Socs36E and UAS-DrkEP(2)2477 under sev-Gal4 still showed some rough eye phenotype and the number of ommatidia with ≤3 R7 cells was 48% (n = 185, Figs. 5C–D). This result

Fig. 4. Repression of Socs36E by high levels of Ras. (A–A″) Ectopic and endogenous R7 precursor cells (high levels of Pros: red) of the sevS11 eye disc do not express Socs36E (GFP) (example arrowhead). In this confocal plane some cone cells are still detected by low levels of Pros protein and they express Socs36E (arrow). (B–B″) Detail of the previous image. (C–C″) Ectopic and endogenous R7 precursor cells (Pros: red) of the sev-RasV12 eye disc do not express Socs36E (GFP). (D–D′) Detail of a sev-RasV12 eye imaginal disc. Note the lack of Socs36E (GFP). Scale bars: 10 µm.

218

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

Fig. 5. Drk overexpression reduces sevS11 suppression by Socs36E. (A) Rough eye due to the overexpression of Drk by sev-Gal4 driver in a sevS11 background. (B) Suppression of sevS11 phenotype by overexpression of Socs36E driven by sev-Gal4. (C) Expression of both Drk and Socs36E transgenes in a sevS11 background produces an intermediate rough eye phenotype. (D) Distribution profiles of number of R7 in each ommatidium for the genotypes represented.

indicates that in the presence of high levels of Drk, SOCS36E is unable to maintain the same suppression of sevS11. To test the role of the two functional domains of SOCS36E we analyzed mutations that have been tested in the mammalian homolog SOCS-5 (Kario et al., 2005; Nicholson et al., 2005). The SH2 mutant did not co-immunoprecipitate with the EGF receptor, demonstrating that the interaction EGFR-SOCS-5 is mediated by the SH2 domain. Mutations in both the SH2 domain and the SOCS box ablate EGFR degradation. Therefore, SOCS36E may interact with the Sev receptor through the SH2 domain in Drosophila eye development. We used transgenes carrying mutations for these domains and compared their abilities to suppress the sevS11 rough eye phenotype when overexpressed by sev-Gal4 to the effect of the wild-type Socs36E construct. For the SH2 mutant (SH2⁎) the conserved arginine (Arg500) was substituted for a lysine to cripple the ability to bind phosphorylated tyrosine residues. For the SOCS box mutant (ΔSB) the conserved SOCS box was deleted entirely (Callus and Mathey-Prevot, 2002). We found that UAS-Socs36E-SH2⁎ ommatidia did not show a significant reduction of R7 cells in comparison to controls that lacked the sev-Gal4 driver (p = 0.2621; Figs. 6A–C). In contrast, flies carrying sev-Gal4 activated UAS-Socs36E-ΔSB resulted in a significant reduction of the number of R7 cells in comparison to controls (p b 0.001; Figs. 6D–F). 80% of the ommatidia contained three to seven R7 photoreceptors per ommatidium in controls carrying sevS11 and inactive UAS-Socs36EΔSB. When UAS-Socs36E-ΔSB construct was activated, nearly 80% of ommatidia contained only one to three R7 cells. These results indicate that the SH2 domain is essential for the Sevenless–SOCS36E interaction. It is worth noting that in these conditions the suppression of UAS-Socs36E-ΔSB is less prominent than with the wild-type UASSocs36E transgene. Specificity of Socs36E interaction with Sev The specificity of this novel function of Socs36E was also tested in different genetic backgrounds. It has been described that mammalian

SOCS-5 co-immunoprecipitates with EGFR (Kario et al., 2005; Nicholson et al., 2005). Weak suppression of Egfr has also been reported in the wing disc after Socs36E overexpression (Callus and Mathey-Prevot, 2002; Rawlings et al., 2004). We observed that the rough eye phenotype generated by the gain-of-function ElpB1 and Elp1 alleles (Baker and Rubin, 1989, 1992) or by the constitutively activated UAS-λtop construct (Queenan et al., 1997) or the UAS-torDEGFR constructs (Dominguez et al., 1998) was not suppressed after expression of Socs36E (Fig. 1 supplemental data; Figs. 7A–B). We quantified the number of R7 cells and found that UAS-torDEGFR under sev-Gal4 driver produced ectopic R7's in 20% of ommatidia (n = 250) whereas 14% (n = 210) when co-expressed with Socs36E (Fig. 7C). The statistical analysis confirmed that the differences between both genotypes for the number of R7 cells per ommatidium, were not significant (p = 0.198; Fig. 7C). The R7 determination also relies on other factors in addition to Sev. We hypothesized that if SOCS36E acts specifically downstream of Sev, the generation of R7's by those other factors should not be suppressed by activation of SOCS36E. For example, Lozenge (lz) is a transcription factor required for R7 specification (Daga et al., 1996). Ectopic expression of lz under the sev promoter transforms R3/R4 photoreceptors into R7's (Flores et al., 2000), but activation of UAS-Socs36E with sev-Gal4 in sev-lz flies did not rescue the R3/R4 transformation (p = 0.159; Figs. 7D–F). Also, constitutively activated alleles of Notch (sev-Nnucl and sev-Nact) in the sev equivalence group converts non-R7 precursors into R7 photoreceptors in a Sev independent manner (Figs. 7G and 7J) (Cooper and Bray, 2000; Fortini et al., 1993). We found that UAS-Socs36E in these cells was not able to rescue the activated Notch rough eye phenotype neither the extra R7 cells (p = 0.237 for sev-Nnucl and p = 0.015 for sev-Nact; Figs. 7G–L). The specificity of Socs36E as suppressor of sev was further assayed by overexpressing activated versions of other members of the signaling pathway, Ras and Raf. Expression of RasV12 and RafTorY9 under the control of sev regulatory sequences produces a severe rough eye phenotype (Fig. 2 Supplemental data), mainly due to the

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

219

Fig. 6. The SH2 domain of Socs36E is essential for suppression of sev. (A) Rough eye due to the sevS11 without the activation of the UAS-Socs36E-SH2⁎ transgene. Semithin section reveals the presence of several R7 cells (green) in each ommatidium. (B) The expression of UAS-Socs36E with the SH2 domain mutated does not suppress the sevS11 phenotype. (C) Histogram showing the distribution profiles of the number of R7 photoreceptors per ommatidium for each genotype represented. Statistical analysis reveals no significant differences between the genotypes (p = 0.2621). (D) SevS11 phenotype is not suppressed without the UAS-Socs36E-ΔSB activity. (E) Expression of UAS-Socs36E-ΔSB under the control of sev-Gal4 in a sevS11 background rescues the regular pattern of ommatidia and reduces the number of R7 in each ommatidium. (F) Histogram quantifying the number of R7 cells per ommatidium with and without UAS-Socs36E-ΔSB activity resulting in significant differences between them (p b 0.001).

transformation of cone cells into R7 photoreceptors (Dickson et al., 1992; Fortini et al., 1992). When we co-expressed these transgenes with UAS-Socs36E, we found weak or no suppression of the rough eye phenotype. It is likely that the weak suppression resulted from the suppression of the endogenous Sev kinase and suggests that the interaction may occur upstream of Ras/Raf. Since Socs36E has been described as a negative regulator of the JAK/STAT pathway, we also examined whether the suppression of R7 cell specification by Socs36E was due to downregulation of the JAK/ STAT signaling pathway rather than a direct interaction with Sev signaling. Flies carrying a loss-of-function allele of the JAK/STAT ligand upd did not suppress the sevS11 rough eye phenotype (Fig. 3 Supplemental data). Furthermore, similar results were obtained using other loss-of-function alleles of the JAK/STAT pathway such as the receptor interacting protein hopscotch (hop) and the downstream transcription factor Stat92E, as they did not affect the rough eye phenotype (Figs. 3C and D Supplemental data). Moreover, it has been shown that loss-of-function mosaic clones for JAK/STAT alleles showed a regular array of ommatidia containing a wild-type complement of correctly differentiated cells, albeit ommatidial polarity was inverted (Zeidler et al., 1999), indicating that the JAK/ STAT pathway is not required for cell fate specification or differentiation in ommatidial cells. In addition, we also over-

expressed another member of the fly SOCS family, the Socs44A gene, which also contains a SH2 domain and is more similar to mammalian SOCS-6 and SOCS-7 (Rawlings et al., 2004). No suppression of sevS11 was found (Fig. 4 Supplemental data), suggesting that the interaction between SOCS36E and Sev is highly specific, being in agreement with multiple functions of SOCS proteins (Rawlings et al., 2004). Discussion We have demonstrated that SOCS36E is a partially redundant Sev attenuator that is specific for this receptor in the developing eye. Two features characterize the function of SOCS36E in this context. First, Socs36E expression is repressed by high Sev activity in the precursor of R7. Second, SOCS36E serves to attenuate receptor activity by the SH2 domain. Attenuation of the Sev activity was proven in vivo, however, by scoring the reduction of R7 cells in sev overexpression, or even in the wild type. The slight increase of R7 cells in Socs36E loss-of-function conditions strongly supports the notion that SOCS36E keeps the signaling below a threshold. Our observations are the first evidence of SOCS36E acting on Sev signaling in the eye photoreceptor specification. Socs36E expression is high in cells that express Sev but avoid R7 differentiation, and absent in pre-R7's. Thus we hypothesize that

220

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

Socs36E contributes to maintaining low levels of Ras/MAPK activity in cells that may respond to Sev activation and that repression of Socs36E is required to permit high levels of Ras/MAPK signaling to single out the R7. Indeed the low or absent levels of Socs36E in extra R7 precursor cells of either sevS11 or sev-RasV12 eyes, demonstrates that high Ras/MAPK activity represses, directly or indirectly, Socs36E expression in the developing eye. Thus, within the Sev equivalence group there would be two different cellular scenarios involving SOCS36E function and gene expression. First, a population of cells (R1, R3, R4, R6 and cone cells) in which the negative regulator dampens the response, prevents inadequate activation of the signal pathway and thus represents a feedback loop. One of the fundamental features that account for robustness in systems is the existence of feedback regulation (Freeman, 2000). Negative feedback loops constitute a central mechanism by which systems attain robustness, as they comprise a major stabilizing role in complex circuits (Amit et al., 2007). Second, a strong signal is reinforced in R7 by suppression of the Socs36E expression. The R7 specification requires high Ras activity, which is normally achieved by the activation of both EGFR and Sev, but EGFR activity is not sufficient to downregulate Socs36E. The precursors of R3 and R4 require EGFR signaling for their specification and still contain high levels of Socs36E. Yet we have shown that high levels of Ras activity, using sev-RasV12, remove SOCS36E in cells that should normally express it. This argues that EGFR and Sev receptors achieve different levels of Ras activity. Alternatively, it could be that in R7, Ras activity is higher because Sev and EGFR are activated and it is this joint activity that results in the suppression of Socs36E expression. The spatial and temporal requirements of sev expression have been previously studied (Basler and Hafen, 1989a,b; Bowtell et al., 1989). Thus Sev function is only required during a brief, defined period for the R7 initiation and subsequently is dispensable for differentiation. It is precisely during this period that SOCS36E is absent in the R7 but present in the rest of the equivalence group. Activation of receptor tyrosine kinases such as Sev and EGFR triggers receptor oligomerization and autophosphorylation on tyrosine residues. These phosphotyrosines serve as docking sites for SH2 domain containing adaptor proteins that link the receptor to downstream effectors such as Ras. In mammals, the SH2 domain adaptor protein Grb2 links receptors to the Ras/MAPK cascade by recruiting Son-of-sevenless (Sos) to the membrane (Rozakis-Adcock et al., 1993). As stated above, the rescue of the ectopic activation of Sev receptor by missexpression of Socs36E is almost complete. The intermediate phenotype found when both drk and Socs36E are coexpressed suggests that these proteins antagonize each other's ability to modulate Sev signaling. This could be achieved either by competing for the same docking site or by binding to different docking sites or phosphorylated adaptor proteins. In any case, the absence of rescue in sevS11 eyes when we overexpressed a Socs36E transgene with the SH2 domain mutated demonstrates that the SH2 domain is essential for inhibiting Sev action. The SH2 domain is not only involved in receptor association but also in its degradation since SOCS proteins use their SH2 domain for the recognition of substrates (Kario et al., 2005). Thus, very likely the physiological degradation of the target protein will be fully accomplished in the presence of a functional SOCS box in addition to SH2. It has been proposed that signal attenuation could function by conjugating ubiquitin to activated tyrosine kinase receptors, thereby promoting receptor endocytosis and lysosomal degradation (Marmor and Yarden,

2004). Indeed, SOCS36E could bind phosphorylated Tyr motifs and recruit the receptor towards the degradation machinery when Sev is wrongly activated in any of the non-R7 cells of the equivalence group. We do not have any evidence for a reduction in Sev protein levels by antibody staining. However, since only the activated receptor would be degraded in non-R7 cells, this minor fraction would not be detected with an antibody that recognizes the entire Sev protein pool. In addition to negatively regulate the JAK/STAT pathway, the mammalian ortholog SOCS-5 also suppresses EGFR signaling in mammalian cells (Kario et al., 2005; Nicholson et al., 2005). In flies, not only the JAK/STAT pathway, but the also the EGFR is repressed by overexpression of Socs36E during wing venation. However Socs36E only weakly suppresses Drosophila wing vein specification (Callus and Mathey-Prevot, 2002; Rawlings et al., 2004). Overexpression of wild type Socs36E and mutated SOCS Box transgenes results in absence of the anterior crossvein, a portion of vein tissue that links two longitudinal veins (Callus and Mathey-Prevot, 2002). Also, in heterozygous flies for EGFR loss-of-function alleles, the overexpression of Socs36E results in occasional thinning of a segment of vein L4 and appearance of ectopic vein material. As we describe here, overexpression of Socs36E is not able to rescue the eye roughness and extra R7's due to gain-of-function alleles of Egfr. It may be that a weak suppression of the EGFR in the eye escaped our observations. But even so, it would contrast with the solid suppression of the activated form of Sev after overexpression of Socs36E and with the observation that also in hypomorphic sev the loss of Socs36E results in a significant recovery of the normal number of R7's. We speculate that a mechanistic basis for the specificity in the eye for Sev versus EGFR would be that the SH2 domain of SOCS36E has a higher binding affinity for Sev phosphorylated tyrosine residues than for other tyrosine kinase receptors. The specificity of SOCS36E for Sev has also been analyzed for other factors involved in R7 specification. If SOCS36E would be simply downstream of R7 determination, then its expression should suppress R7 specification in different contexts. The transcription factor Lz is normally expressed in R1/R6 and later in R7 (Flores et al., 1998). In R7 precursors, transcription of prospero is activated by Ras and also requires activation by Lz (Xu et al., 2000). Notch activation in the presumptive R7 cell constitutes a signal required, in combination to Sev, to specify the R7 fate (Tomlinson and Struhl, 2001). Ectopic activation of either lz or N in the Sev equivalence group results in R3/4 or R1/6 transformations into R7, respectively. However, those transformations were not suppressed by the activation of SOCS36E. This observation strengthens our view that SOCS36E interferes specifically with the tyrosine kinase receptor and not simply with the R7 cell fate. SOCS44, which also contains the similar modular architecture as mammalian SOCS and shows greatest sequence similarity to SOCS-6 and SOCS-7 has been reported to have opposite effects on EGFR signaling in the wing venation as it enhances EGFR signaling (Rawlings et al. 2004). The observation that this member of the SOCS family does not suppress Sev provides additional support for the specific function of SOCS36E. This is in agreement with the proposed multiple functions of the Drosophila SOCS proteins (Rawlings et al., 2004). As stated above, the loss-of-function of Socs36E results in a low expressivity phenotype indicative of redundancy. In structural terms, robust systems can be enhanced if there are multiple means to achieve

Fig. 7. Overexpression of Socs36E does not suppress Egfr, lozenge and Notch gain-of-function alleles. (A) UAS-torDEGFR; sev-Gal4 line shows strong rough eyes with ectopic R7 photoreceptor cells, detected by semithin sections. (B) Socs36E overexpression under the sev-Gal4 driver does not suppress rough eye phenotype. (C) Histogram reveals the distribution of number of R7 per ommatidium for each genotype represented (p = 0.198). (D) Expression of lz under sev promoter gives rise to transformation of R3 and R4 photoreceptors to R7's, observed in semithin sections. (E) Overexpression of Socs36E with sev-Gal4 does not rescue the sev-lz phenotype. (F) Histogram showing distribution profiles of the number of R7 per ommatidium for (D) and (E) (p = 0.1596). (G) sev-Nnucl line shows rough eyes and multiple R7 per ommatidium. (H) Socs36E overexpression under sev-Gal4 driver does not suppress the sev-Nnucl phenotype. (I) Distribution profiles of number of R7 per ommatidium of (G) and (H) (p = 0.2367). (J) Rough eye and semithin section from a sevNact fly. (K) Activation of UAS-Socs36E with sev-Gal4 does not rescue the sev-Nact phenotype. (L) Distribution profiles of number of R7 per ommatidium of (J) and (K) (p = 0.0148).

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

221

222

I. Almudi et al. / Developmental Biology 326 (2009) 212–223

a specific function because failure of one of them can be rescued by others (Kitano, 2004). Besides, this redundancy, which could be due to factors with similar functions, could also explain why Socs36E was not detected in other screenings of the pathway based on loss-of-function mutations (Gaul et al., 1992; Karim et al., 1996). The structural robustness of the pathway is essential to ensure the R7 driven UV vision of the fly (Harris et al., 1976). Disruption of Sev signaling results in cone cell lineage instead the R7 lineage (Harris et al., 1976; Tomlinson and Ready, 1986). Thus, we propose the action of SOCS36E as a model in which mechanisms of feedback regulation will enable fail-safe performance in the face of perturbations that could result in wrong developmental decisions. Acknowledgments We thank M. Morey, M. Sesé, D. Haro, L. Serra and J. Casanova for their help and discussions. We are grateful to the Berkeley Drosophila Genome Project for providing the Socs36E cDNA clone (SD04308), to S. Hou, B. Mathey-Prevot, D. Harrison, M. Freeman, S. ArtavanisTsakonas, the Bloomington Stock center, the VDRC stock center and DGRC for fly stocks. We thank the Serveis Científico Tècnics of the University of Barcelona (SCT-UB) for their help and advice in scanning electron microscopy. We also acknowledge the Developmental Studies Hybridoma Bank for antibodies. I.A. is a recipient of a fellowship from Universitat de Barcelona. This work was supported by BFU2004-04732 and BFU2006-07334/BMC grants from the Ministerio de Educación y Ciencia of Spain. Appendix A. Supplementary data Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.ydbio.2008.11.014. References Amit, I., Wides, R., Yarden, Y., 2007. Evolvable signaling networks of receptor tyrosine kinases: relevance of robustness to malignancy and to cancer therapy. Mol. Syst. Biol. 3, 151. Baetz, A., Frey, M., Heeg, K., Dalpke, A.H., 2004. Suppressor of cytokine signaling (SOCS) proteins indirectly regulate toll-like receptor signaling in innate immune cells. J. Biol. Chem. 279, 54708–54715. Baker, N.E., Rubin, G.M., 1989. Effect on eye development of dominant mutations in Drosophila homologue of the EGF receptor. Nature 340, 150–153. Baker, N.E., Rubin, G.M., 1992. Ellipse mutations in the Drosophila homologue of the EGF receptor affect pattern formation, cell division, and cell death in eye imaginal discs. Dev. Biol. 150, 381–396. Basler, K., Hafen, E., 1989a. Dynamics of Drosophila eye development and temporal requirements of sevenless expression. Development 107, 723–731. Basler, K., Hafen, E., 1989b. Ubiquitous expression of sevenless: position-dependent specification of cell fate. Science 243, 931–934. Basler, K., Christen, B., Hafen, E., 1991. Ligand-independent activation of the Sevenless receptor tyrosine kinase changes the fate of cells in the developing Drosophila eye. Cell 64, 1069–1081. Bellen, H.J., Levis, R.W., Liao, G., He, Y., Carlson, J.W., Tsang, G., Evans-Holm, M., Hiesinger, P.R., Schulze, K.L., Rubin, G.M., Hoskins, R.A., Spradling, A.C., 2004. The BDGP gene disruption project: single transposon insertions associated with 40% of Drosophila genes. Genetics 167, 761–781. Bohmann, D., Ellis, M.C., Staszewski, L.M., Mlodzik, M., 1994. Drosophila Jun mediates Ras-dependent photoreceptor determination. Cell 78, 973–986. Bowtell, D.D., Simon, M.A., Rubin, G.M., 1989. Ommatidia in the developing Drosophila eye require and can respond to sevenless for only a restricted period. Cell 56, 931–936. Callus, B.A., Mathey-Prevot, B., 2002. SOCS36E, a novel Drosophila SOCS protein, suppresses JAK/STAT and EGF-R signalling in the imaginal wing disc. Oncogene 21, 4812–4821. Cooper, M.T., Bray, S.J., 2000. R7 photoreceptor specification requires Notch activity. Curr. Biol. 10, 1507–1510. Daga, A., Karlovich, C.A., Dumstrei, K., Banerjee, U., 1996. Patterning of cells in the Drosophila eye by Lozenge, which shares homologous domains with AML1. Genes Dev. 10, 1194–1205. Dickson, B., Sprenger, F., Morrison, D., Hafen, E., 1992. Raf functions downstream of Ras1 in the Sevenless signal transduction pathway. Nature 360, 600–603. Dickson, B.J., van der Straten, A., Dominguez, M., Hafen, E., 1996. Mutations modulating Raf signaling in Drosophila eye development. Genetics 142, 163–171.

Dietzl, G., Chen, D., Schnorrer, F., Su, K.C., Barinova, Y., Fellner, M., Gasser, B., Kinsey, K., Oppel, S., Scheiblauer, S., Couto, A., Marra, V., Keleman, K., Dickson, B.J., 2007. A genome-wide transgenic RNAi library for conditional gene inactivation in Drosophila. Nature 448, 151–156. Dominguez, M., Wasserman, J.D., Freeman, M., 1998. Multiple functions of the EGF receptor in Drosophila eye development. Curr. Biol. 8, 1039–1048. Endo, T.A., Masuhara, M., Yokouchi, M., Suzuki, R., Sakamoto, H., Mitsui, K., Matsumoto, A., Tanimura, S., Ohtsubo, M., Misawa, H., Miyazaki, T., Leonor, N., Taniguchi, T., Fujita, T., Kanakura, Y., Komiya, S., Yoshimura, A., 1997. A new protein containing an SH2 domain that inhibits JAK kinases. Nature 387, 921–924. Flores, G.V., Daga, A., Kalhor, H.R., Banerjee, U., 1998. Lozenge is expressed in pluripotent precursor cells and patterns multiple cell types in the Drosophila eye through the control of cell-specific transcription factors. Development 125, 3681–3687. Flores, G.V., Duan, H., Yan, H., Nagaraj, R., Fu, W., Zou, Y., Noll, M., Banerjee, U., 2000. Combinatorial signaling in the specification of unique cell fates. Cell 103, 75–85. Fortini, M.E., Simon, M.A., Rubin, G.M., 1992. Signalling by the sevenless protein tyrosine kinase is mimicked by Ras1 activation. Nature 355, 559–561. Fortini, M.E., Rebay, I., Caron, L.A., Artavanis-Tsakonas, S., 1993. An activated Notch receptor blocks cell-fate commitment in the developing Drosophila eye. Nature 365, 555–557. Freeman, M., 1996. Reiterative use of the EGF receptor triggers differentiation of all cell types in the Drosophila eye. Cell 87, 651–660. Freeman, M., 2000. Feedback control of intercellular signalling in development. Nature 408, 313–319. Freeman, M., Gurdon, J.B., 2002. Regulatory principles of developmental signaling. Annu. Rev. Cell Dev. Biol. 18, 515–539. Gaul, U., Mardon, G., Rubin, G.M., 1992. A putative Ras GTPase activating protein acts as a negative regulator of signaling by the Sevenless receptor tyrosine kinase. Cell 68, 1007–1019. Harris, W.A., Stark, W.S., Walker, J.A., 1976. Genetic dissection of the photoreceptor system in the compound eye of Drosophila melanogaster. J. Physiol. 256, 415–439. Hay, B.A., Wolff, T., Rubin, G.M., 1994. Expression of baculovirus P35 prevents cell death in Drosophila. Development 120, 2121–2129. Hilton, D.J., Richardson, R.T., Alexander, W.S., Viney, E.M., Willson, T.A., Sprigg, N.S., Starr, R., Nicholson, S.E., Metcalf, D., Nicola, N.A., 1998. Twenty proteins containing a C-terminal SOCS box form five structural classes. Proc. Natl. Acad. Sci. U. S. A. 95, 114–119. Karim, F.D., Chang, H.C., Therrien, M., Wassarman, D.A., Laverty, T., Rubin, G.M., 1996. A screen for genes that function downstream of Ras1 during Drosophila eye development. Genetics 143, 315–329. Kario, E., Marmor, M.D., Adamsky, K., Citri, A., Amit, I., Amariglio, N., Rechavi, G., Yarden, Y., 2005. Suppressors of cytokine signaling 4 and 5 regulate epidermal growth factor receptor signaling. J. Biol. Chem. 280, 7038–7048. Karsten, P., Hader, S., Zeidler, M.P., 2002. Cloning and expression of Drosophila SOCS36E and its potential regulation by the JAK/STAT pathway. Mech. Dev. 117, 343–346. Kitano, H., 2004. Biological robustness. Nat. Rev. Genet. 5, 826–837. Kraut, R., Menon, K., Zinn, K., 2001. A gain-of-function screen for genes controlling motor axon guidance and synaptogenesis in Drosophila. Curr. Biol. 11, 417–430. Krebs, D.L., Hilton, D.J., 2003. A new role for SOCS in insulin action. Suppressor of cytokine signaling. Sci. STKE 2003, PE6. Marmor, M.D., Yarden, Y., 2004. Role of protein ubiquitylation in regulating endocytosis of receptor tyrosine kinases. Oncogene 23, 2057–2070. Nicholson, S.E., Metcalf, D., Sprigg, N.S., Columbus, R., Walker, F., Silva, A., Cary, D., Willson, T.A., Zhang, J.G., Hilton, D.J., Alexander, W.S., Nicola, N.A., 2005. Suppressor of cytokine signaling (SOCS)-5 is a potential negative regulator of epidermal growth factor signaling. Proc. Natl. Acad. Sci. U. S. A. 102, 2328–2333. Olivier, J.P., Raabe, T., Henkemeyer, M., Dickson, B., Mbamalu, G., Margolis, B., Schlessinger, J., Hafen, E., Pawson, T., 1993. A Drosophila SH2-SH3 adaptor protein implicated in coupling the Sevenless tyrosine kinase to an activator of Ras guanine nucleotide exchange, Sos. Cell 73, 179–191. Queenan, A.M., Ghabrial, A., Schupbach, T., 1997. Ectopic activation of torpedo/Egfr, a Drosophila receptor tyrosine kinase, dorsalizes both the eggshell and the embryo. Development 124, 3871–3880. Rawlings, J.S., Rennebeck, G., Harrison, S.M., Xi, R., Harrison, D.A., 2004. Two Drosophila suppressors of cytokine signaling (SOCS) differentially regulate JAK and EGFR pathway activities. BMC Cell Biol. 5, 38. Reinke, R., Zipursky, S.L., 1988. Cell–cell interaction in the Drosophila retina: the bride of sevenless gene is required in photoreceptor cell R8 for R7 cell development. Cell 55, 321–330. Rorth, P., 1996. A modular misexpression screen in Drosophila detecting tissue-specific phenotypes. Proc. Natl. Acad. Sci. U. S. A. 93, 12418–12422. Rorth, P., Szabo, K., Bailey, A., Laverty, T., Rehm, J., Rubin, G.M., Weigmann, K., Milan, M., Benes, V., Ansorge, W., Cohen, S.M., 1998. Systematic gain-of-function genetics in Drosophila. Development 125, 1049–1057. Rozakis-Adcock, M., Fernley, R., Wade, J., Pawson, T., Bowtell, D., 1993. The SH2 and SH3 domains of mammalian Grb2 couple the EGF receptor to the Ras activator mSos1. Nature 363, 83–85. Sese, M., Corominas, M., Stocker, H., Heino, T.I., Hafen, E., Serras, F., 2006. The Cdi/ TESK1 kinase is required for Sevenless signaling and epithelial organization in the Drosophila eye. J. Cell. Sci. 119, 5047–5056. Simon, M.A., Bowtell, D.D., Dodson, G.S., Laverty, T.R., Rubin, G.M., 1991. Ras1 and a putative guanine nucleotide exchange factor perform crucial steps in signaling by the sevenless protein tyrosine kinase. Cell 67, 701–716. Simon, M.A., Dodson, G.S., Rubin, G.M., 1993. An SH3-SH2-SH3 protein is required for p21Ras1 activation and binds to Sevenless and Sos proteins in vitro. Cell 73, 169–177.

I. Almudi et al. / Developmental Biology 326 (2009) 212–223 Starr, R., Willson, T.A., Viney, E.M., Murray, L.J., Rayner, J.R., Jenkins, B.J., Gonda, T.J., Alexander, W.S., Metcalf, D., Nicola, N.A., Hilton, D.J., 1997. A family of cytokineinducible inhibitors of signalling. Nature 387, 917–921. Tomlinson, A., Ready, D.F., 1986. Sevenless: A cell-specific homeotic mutation of the Drosophila eye. Science 231, 400–402. Tomlinson, A., Struhl, G., 2001. Delta/Notch and Boss/Sevenless signals act combinatorially to specify the Drosophila R7 photoreceptor. Mol. Cell. 7, 487–495. Tomlinson, A., Bowtell, D.D., Hafen, E., Rubin, G.M., 1987. Localization of the Sevenless protein, a putative receptor for positional information, in the eye imaginal disc of Drosophila. Cell 51, 143–150. Xu, C., Kauffmann, R.C., Zhang, J., Kladny, S., Carthew, R.W., 2000. Overlapping activators and repressors delimit transcriptional response to receptor tyrosine kinase signals in the Drosophila eye. Cell 103, 87–97.

223

Yoshimura, A., Ohkubo, T., Kiguchi, T., Jenkins, N.A., Gilbert, D.J., Copeland, N.G., Hara, T., Miyajima, A., 1995. A novel cytokine-inducible gene CIS encodes an SH2-containing protein that binds to tyrosine-phosphorylated interleukin 3 and erythropoietin receptors. EMBO J. 14, 2816–2826. Zeidler, M.P., Perrimon, N., Strutt, D.I., 1999. Polarity determination in the Drosophila eye: a novel role for unpaired and JAK/STAT signaling. Genes Dev. 13, 1342–1353. Zhang, J.G., Farley, A., Nicholson, S.E., Willson, T.A., Zugaro, L.M., Simpson, R.J., Moritz, R.L., Cary, D., Richardson, R., Hausmann, G., Kile, B.J., Kent, S.B., Alexander, W.S., Metcalf, D., Hilton, D.J., Nicola, N.A., Baca, M., 1999. The conserved SOCS box motif in suppressors of cytokine signaling binds to elongins B and C and may couple bound proteins to proteasomal degradation. Proc. Natl. Acad. Sci. U. S. A. 96, 2071–2076.