Introduction

The developing Drosophila eye has provided an extremely powerful model system for the study of pathways that control cellular differentiation. In particular, the role of the EGF receptor pathway in initiating photoreceptor differentiation has been analyzed in great detail. Following activation of receptor tyrosine kinases (RTKs), newly activated MAPK has been shown to phosphorylate Yan and pointed (Pnt), two ETS-domain transcription factors that together regulate a common set of critical target genes (Lai and Rubin, 1992; Brunner et al., 1994; O'Neill et al., 1994). A third ETS-transcription factor, Tramtrack (Ttk), functions downstream of Yan and Pnt to repress genes required for R7 differentiation (Lai et al., 1996). In undifferentiated cells, Pnt is found in an inactive form whereas Yan and Ttk actively repress their target genes (Lai and Rubin, 1992; Chang et al., 1995; Dickson et al., 1995). Phosphorylation of Yan blocks its ability to repress transcription and targets it for nuclear export and subsequent ubiquitin-dependent degradation (Rebay and Rubin, 1995). In contrast, phosphorylation of Pnt stimulates its ability to activate transcription of YAN/PNT target genes. One such transcriptional target is Phyllopod (Phyl). When expressed, Phyl and seven in absentia (Sina) form a complex that targets Ttk for degradation (Carthew and Rubin, 1990; Li et al., 1997; Tang et al., 1997). Therefore, prior to RTK activation, Yan and Ttk repress genes required for differentiation. Following activation, Yan and Ttk are targeted for degradation, allowing the initiation of the R7 cell fate.

Recently, Zipursky and co-workers have shown that Ebi is needed for EGFR-regulated differentiation in Drosophila (Dong et al., 1999). Mutant alleles of ebi genetically interact with components of the EGFR pathway, and ebi mutant embryos possess many of the phenotypes previously described for egfr mutant embryos. Studies of Ebi function in the eye revealed that it is required for R7 differentiation. Ebi has been proposed to facilitate neuronal differentiation by promoting the down-regulation of TTK88 (Dong et al., 1999).

Elimination of EGFR function in the developing eye disk has demonstrated that this RTK is also required for normal control of cell proliferation. Indeed, EGFR gain-of-function mutations cause excess proliferation and tissue growth in the disk (Baker et al., 1992; Dominguez et al., 1998). Since cell cycle exit and differentiation must be coordinated in order for the developmental program to proceed, it makes biological sense that the pathways controlling differentiation may also participate in the events that culminate in cell cycle withdrawl. Precisely how the EGFR pathway controls cell proliferation in some settings, and stimulates differentiation in others, remains unclear.

In Drosophila, S phase entry is severely reduced in both cyclin E and de2f mutant embryos (Knoblich et al., 1994; Duronio et al., 1995). Conversely, the ectopic expression of either cyclin E or dE2F/dDP, in either the embryo or imaginal disks, is sufficient to drive cells into S phase (Duronio and O'Farrell, 1994; Knoblich et al., 1994; Richardson et al., 1995; Asano et al., 1996; Brook et al., 1996). During normal development, the activities of cyclin E and dE2F/dDP are constrained, in part, by Dacapo (Dap) and the retinoblastoma homolog, Rbf. In several cell types Dap, a Drosophila homolog of the p21/p27 CDK inhibitors, is important for timely cell cycle exit. During embryogenesis, a pulse of Dap expression is needed for epidermal cells to introduce a G₁ phase following mitosis of cell cycle 16 (de Nooij et al., 1996; Lane et al., 1996). As a consequence, dap mutants fail to arrest the cell cycle and instead continue through a 17th cell cycle before arresting in G₁ of cell cycle 18. In contrast to dap mutants, Rbf-deficient embryos briefly arrest following mitosis of cell cycle 16, but fail to maintain the arrest and instead re-enter the cell cycle (Du and Dyson, 1999). These results suggest that Dap expression initiates the G₁ arrest by directly inhibiting cyclin E/Cdc2c, whereas Rbf functions to maintain the G₁ phase, possibly by limiting the expression of E2F targets such as cyclin E.

Here we describe observations that link ebi to cell cycle control and provide a different perspective on Ebi function to that described by Dong et al. (1999). We show that ebi mutations genetically interact with a number of key cell cycle regulators and demonstrate a requirement for Ebi in cell cycle exit during the developmental program. Although the inability to exit the cell cycle correlates with a reduction in the capacity to differentiate, our results suggest that an inability to degrade Ttk88 is not responsible for the cell cycle phenotypes of the ebi mutants. These results show that the loss of Ebi function perturbs both cell cycle exit and the onset of differentiation, indicating that Ebi is ideally situated to regulate these two processes coordinately.

Results

ebi alleles are dominant enhancers of the GMR-dE2F-dDP-p35 rough eye phenotype

Previously, we have used a dE2F/dDP overexpression phenotype in the Drosophila eye to screen randomly generated EMS and X-ray-induced mutations for alleles that are modifiers of E2F activity (Staehling-Hampton et al., 1999). As the initial screen was not saturated, we used the same GMR-dE2F-dDP-p35 chromosome to screen a P-element collection for additional alleles that are important for this E2F-dependent phenotype. From this screen, the P-element P[w+; LacZ]K16213 was isolated as a dominant enhancer of the GMR-dE2F-dDP-p35 phenotype (Figure 1). K16213 enhanced the general roughness of the eye, increasing the irregular arrangement of the ommatidial facets relative to the GMR-dE2F-dDP-p35 phenotype alone (Figure 1), and generating a large number of bristle duplications. K16213 is a lethal insertion that maps to position 21C on the left arm of chromosome II. K16213 was tested against mutant alleles that had been mapped to the 21C region and a complementation group of three alleles was found that failed to complement the P-element (CC1, CC3 and CC4). No escapers were observed from any of the possible trans-heterozygous combinations of K16213, CC1, CC3 and CC4, indicating that the mutated gene is essential for Drosophila development. Each of the three EMS alleles enhanced the GMR-dE2F-dDP-p35 to a similar degree to that observed for the P-element (data not shown).

Excision of the P-element from K16213 reverted the lethality associated with this chromosome and eliminated its ability to modify the GMR-dE2F-dDP-p35 phenotype (data not shown). Database searches using K16213 flanking sequence identified a putative open reading frame (ORF) that encodes a 700 amino acid protein with a C-terminal WD-repeat domain and a divergent N-terminal F-box domain possessing many, but not all of the residues present in bona fide F-box proteins such as Cdc4 (Figure 2). Sequencing of this ORF from flies carrying the CC1, CC3 and CC4 alleles identified distinct mutations corresponding to CC1 (L16Q), CC3 (S602F) and CC4 (an 11 bp deletion at amino acid 313) (Figure 2). During the course of our investigation, Dong et al. (1999) reported the isolation of mutations in this gene as dominant enhancers of the roughex mutation. Dong et al. (1999) have named this gene Ebi, and we, therefore, refer to the alleles described here as ebi^K16213, ebi^CC1, ebi^CC3 and ebi^CC4 (Figure 2).

Mutant alleles of ebi dominantly suppress the GMR-p21 rough eye and restore the second mitotic wave

As we had isolated ebi alleles as enhancers of a dE2F-dependent phenotype that is characterized by increased cell proliferation, we examined whether Ebi gene dosage could suppress eye phenotypes caused by a reduction in cell proliferation. de Nooij and Hariharan (1995) have shown that expression of the human cyclin-dependent kinase inhibitor p21 strongly inhibits S phase entry in the developing eye disk, giving a rough eye phenotype characterized by reduced numbers of pigment cells and fused ommatidia. The GMR-p21 phenotype is dominantly suppressed by mutations in Ebi (Figure 3). The number of ommatidia in the pGMR-p21/ebi adult eye (Figure 3C and F) are similar to wild type and, other than minor bristle defects, the pGMR-p21/ebi adult eye is morphologically normal (Figure 3; data not shown).

Wild-type eye disks exhibit a band of S phases, posterior to the morphogenetic furrow that constitutes the second mitotic wave (Wolff and Ready, 1993). This wave of S phase is required to generate the full complement of cells that will be needed for subsequent differentiation events. GMR driven expression of p21 abolishes the second mitotic wave of S phases, causing the dramatic reduction in the number of cells observed in the adult eye (de Nooij and Hariharan, 1995; Figure 3). To examine the effect of reducing the dosage of Ebi on this process, 3rd instar larval eye disks were isolated from the various genotypes and S phase cells labeled by BrdU incorporation. Consistent with the adult phenotype, the second mitotic wave of S phase in pGMR-p21/ebi disks is partially restored and in some cases returns to levels similar to that observed in the wild type (Figure 3). Thus, the normal function of Ebi is necessary in these G₁ phase cells for p21 to block S phase entry efficiently.

ebi mutants display ectopic S phases in the PNS and CNS at stage 13/14 of embryogenesis

The ability of ebi alleles to enhance an eye phenotype resulting from increased cell proliferation, and to suppress a phenotype resulting from reduced cell proliferation raised the possibility that Ebi might act to limit proliferation during normal development. To test this we examined whether loss of Ebi function would lead to inappropriate S phase entry. Stocks carrying the ebi alleles were intercrossed to generate trans-heterozygous mutants, and BrdU incorporation was used to determine the patterns of embryonic S phases (see Materials and methods). Following germ band retraction in wild-type embryos, cells within the peripheral nervous system (PNS) have completed S phase of cell cycle 16 and arrest in G₁ of cell cycle 17. In contrast, cells within the fore and hindgut enter a period of endo-reduplication. In ebi mutants, the endocycles in the gut (consistent with stage 13 embryos) occur as normal (Figure 4). However, cells within the PNS fail to exit the cell cycle and continue into a 17th S phase (Figure 4). This PNS phenotype is very similar to that observed in dap mutant embryos, although ebi mutants do not display the ectopic S phases seen in the epidermis of dap mutant embryos (de Nooij et al., 1996; Lane et al., 1996). In addition to ectopic S phases in the PNS, ebi mutants display aberrant S phase entry within the central nervous system (CNS) (Figure 4). The ectopic S phases in the CNS are clearly visible as an expansion of the BrdU incorporation in the CNS of the mutant embryos relative to the incorporation observed in the wild type (Figure 4). The cell cycle defects observed in the ebi mutant could arise through a defect in cell cycle withdrawal or might reflect a delay in the normal cell cycle program. It is not possible from the above observations to distinguish definitively between these two possible models. However, if the extra S phases we see in the ebi mutants occur due to a delay in the S phase program we would expect to see a reduction in the number of S phases in the mutant at earlier stages in development. To address this possibility we performed BrdU pulse labeling experiments on wild-type and ebi mutant embryos to analyze S phases at developmental times prior to stage 13/14. Consistent with a defect in cell cycle exit at later stages, we find no evidence of a reduction in the number of S phases in the ebi mutant at developmental times prior to stage 13/14, when compared with the wild type (data not shown). Thus, it seems most likely that the additional S phases observed in the ebi mutant arise from a defect in cell cycle exit and not from a delay in the S phase program.

The neuronal-specific antibody Mab22C10 (Zipursky et al., 1984) was used to analyze the integrity of neuronal lineages in ebi mutants. Consistent with the cell cycle defects described, we observe a number of defects in neuronal differentiation including a reduction in peripheral neuron staining and an absence of a large number of Mab22C10-positive cells within the CNS (Figure 5). This phenotype has also been observed in phyllopod mutants (Chang et al., 1995) and in embryos that over-express Ttk69 and Ttk88 from heat shock transgenes (Guo et al., 1995).

One possible interpretation of the experiments described above is that Ebi has a function that limits S phase entry, and that the defects in neuronal differentiation are an indirect effect of a failure to arrest the cell cycle completely. The observations of Dong et al. (1999), showing that Ebi regulates Ttk88 degradation, suggest an alternative interpretation. If the failure to degrade Ttk88 interferes with neuronal differentiation, then the additional S phases described above might reflect the continued proliferation of cells that have failed to differentiate, rather than a direct role in cell cycle control. A third possibility is that Ebi function is important for both cell cycle exit and for the onset of differentiation. We first looked for biochemical evidence to support the idea that Ebi promotes Ttk88 degradation and then, in order to distinguish between these models, we examined whether simply blocking neuronal differentiation by elevating Ttk88 was sufficient to cause the increase in S phase cells that occurs when Ebi levels are reduced.

Ebi promotes the degradation of Ttk88

We found three lines of evidence that link Ebi to Sina and Pyllopod-dependent degradation of Ttk88. The first line of evidence comes from transient transfection experiments. Previously, Li et al. (1997) have demonstrated that Sina and Phyl are able to target Ttk88 for ubiquitin-dependent degradation when expressed in transient transfection experiments in S2 cells. S2 cells contain high levels of endogenous Ebi (data not shown). To interfere with the activity of this protein, we expressed an N-terminal fragment (EbiN) that was previously used as a dominant-negative mutant to interfere with Ttk88 degradation in the eye disk (Dong et al., 1999). Cells were co-transfected with pIZT vectors (Invitrogen; see Materials and methods) constitutively expressing either Cat or EbiN from the OpIE2 promotor, in the presence of metallothionine-inducible vectors containing Phyl, Sina and Ttk88-myc (Tang et al., 1997). Following transfection and copper induction, cells were metabolically labeled with [³⁵S]methionine and then chased with cold methionine for various time points to monitor Ttk88 degradation (Li et al., 1997). As has been previously reported, in the presence of Sina and Phyl, Ttk88 is degraded with a half life of 25 min, a process that can be blocked by the proteosome inhibitor MG132 (Li et al., 1997; Figure 6). We observed that Ttk88 degradation was blocked in cells expressing the EbiN dominant-negative from pIZT, whereas a control pIZT vector that constitutively expressed Cat gave no effect (Figure 6A).

The second line of evidence stems from an in vitro degradation assay. To investigate whether Ebi is linked toTtk88 degradation we attempted to reconstitute Ttk88 degradation in vitro. Initial experiments showed that Ttk88 was not degraded when co-expressed with Sina and Phyl in rabbit reticulocyte lysate (data not shown). In an attempt to stimulate the degradation activity, we incubated GST (glutathione S-transferase) or GST–Ebi beads together with ubiquitin, an ATP regeneration system and Sina– Phyl–Ttk co-expressed in reticulocyte lysate. Neither GST nor GST–Ebi was able to promote Ttk88 degradation in this setting (Figure 6B). We speculated that Ebi may function as a bridge between the Sina–Phyl–Ttk88 complex and an activity required for ubiquitylation and subsequent degradation of Ttk88. In order to provide this missing activity, we pre-incubated GST or GST–Ebi beads with S2 extracts (now referred to as loaded beads) prior to performing the degradation reaction. Loaded GST–Ebi beads were unable to degrade Ttk88 when expressed alone. However, when Sina and Phyl were co-translated with Ttk88, loaded GST–Ebi beads were able to promote degradation of TtK88 by a mechanism that was blocked by the proteosome inhibitor, LLnL, whereas loaded GST beads could not (Figure 6B and C). The need for Sina and Phyl for Ttk degradation mirrors their requirement in vivo (Li et al., 1997; Tang et al., 1997). In vitro-translated dE2F, dDP or cyclin E was not degraded in any of the experiments described (Figure 6C; data not shown). In order to map the region of Ebi that associates with the activity required for Ttk88 degradation, we incubated loaded GST–EbiN and GST–EbiC beads and then performed the degradation assay. Interestingly, neither half of Ebi alone was capable of targeting Ttk88 for degradation (data not shown). These data suggest that the full-length Ebi protein may act to bring the Sina–Phyl–Ttk88 complex and a ubiquitylation activity into close proximity.

The results of the in vitro degradation assay suggested that Ebi might physically associate with Ttk88, Sina and Phyl. This was tested using the in vitro translated Sina, Phyl and Ttk88, and the GST–Ebi fusion proteins described above. GST–Ebi was found to interact strongly with Sina and Phyl, and weakly with Ttk88, when these proteins were expressed and assayed individually (Figure 7A). The GST control showed no association (Figure 7A). Interestingly, when Sina, Phyl and Ttk88 are co-expressed in the same lysate, GST–Ebi is able to pull down Ttk88 with a much higher affinity compared with when Ttk88 is expressed and bound alone (Figure 7B). This suggests that the Ebi–Ttk interaction is indirect and may require Sina and Phyl to facilitate association. To define the region of Ebi required for Sina, Phyl and Ttk association, we tested GST–EbiN (N-terminal domain) and GST–EbiC (C-terminal WD-repeat domain) fusion proteins for binding. GST–EbiN does not associate with any of the proteins, whereas GST–EbiC is able to bind to all three proteins, although with a slightly reduced affinity when compared with the full-length protein (Figure 7A). Since the -propeller structure formed by WD-repeat domains provides a surface for many protein–protein interactions, it is possible that this domain in Ebi provides a scaffold for Sina, Phyl and Ttk88 association. To provide further evidence for these interactions, we transiently co-transfected pIZT-Ebi with pIZT-V5His-Sina or pIZT-V5His-Phyl, or all three constructs into S2 cells. Sina and Phyl (His₆- and V5-tagged) were purified from lysates from the various transfected populations by Ni–NTA agarose chromatography. The beads were then subjected to western blotting using a monoclonal antibody to Ebi. Ebi was found to co-precipitate with Sina and Phyl but was not observed in the untransfected control (Figure 7C).

Increased Ttk levels do not phenocopy the effects of ebi alleles on S phase entry

If stabilization of Ttk88 is responsible for the cell cycle phenotypes associated with Ebi, we expected to observe a consistent pattern of genetic interactions between GMR-p21 or GMR-E2F-DP-p35 and mutations in the EGFR pathway that function to regulate Ttk88 levels. However, unlike mutations in Ebi, we found that loss-of-function alleles of egfr, gap1, raf1, ras1, mapk, yan, sina, phyl and ttk had no strong effect on either the GMR-p21 or the GMR-E2F-DP-p35 phenotype (see Materials and methods for details). Mutations in the Ets-domain transcription factor Pointed (pnt) enhanced the GMR-E2F-DP-p35 phenotype but failed to modify the GMR-p21 phenotype (Staehling-Hampton et al., 1999). While it is possible that Ebi is the only dosage-sensitive component of the EGFR pathway whose levels effect cell proliferation, these results suggested that Ebi might have an activity that is independent of Ttk88 degradation.

To test directly whether stabilization of Ttk88 is likely to be responsible for the changes in cell cycle control caused by ebi mutant alleles, we examined the effects of elevating the levels of Ttk88. As previously described, ectopic expression of Ttk88 was induced by heat shock for 30 min at 39°C in embryos carrying a hs-Ttk88 (Guo et al., 1995; see Materials and methods for details). Following recovery, embryos were either pulsed with BrdU or aged for immunohistochemistry. As previously reported, the ectopic expression of Ttk88 from a heat shock-regulated transgene is sufficient to disrupt neuronal differentiation in stage 13–14 embryos (Figure 8; Guo et al., 1995). Unlike the ebi mutant embryos, BrdU incorporation demonstrated that the block to differentiation in hs-Ttk88 embryos does not result in a failure to exit the cell cycle in the PNS and CNS (Figure 8). Furthermore, expression of Ttk88 from a GMR transgene inhibits differentiation in the eye disk (Li et al., 1997; Tang et al., 1997; data not shown). However, the GMR-Ttk88 transgene failed to suppress the GMR-p21 eye phenotype and was unable to restore S phases in the GMR-p21 eye disk (Figure 9). In contrast, halving the dosage of Ebi or expressing cyclin E from the GMR promoter was sufficient to restore the second mitotic wave of S phases (Figure 9). We conclude that increasing Ttk88 protein to a level where differentiation is perturbed in either the embryo or the eye disk is insufficient to promote S phase entry in either of the situations where ebi mutations gave this effect (Figures 8 and 9). We infer that Ebi must have a second function, independent of Ttk88 degradation, that is important for regulating cell cycle exit.

Discussion

Modifier screens provide a relatively unbiased way to identify genes that are functionally important for a biological process. Mutant alleles of ebi were first isolated by Dong et al. (1999) as enhancers of a roughex (rux) mutation. Although Rux is known to act as a cyclin-dependent kinase inhibitor (Thomas et al., 1994, 1997; Dong et al., 1997; Foley et al, 1999), Dong et al. (1999) noted that ebi mutations enhanced the differentiation defects of the rux phenotype rather than the earlier cell cycle defects. Consistent with this interpretation they observed that ebi mutations dominantly suppress differentiation phenotypes induced by constitutive activation of the EGFR pathway, and that Ebi regulates Ttk88 protein levels in the eye imaginal disk. Ttk88 is an ETS-domain transcription factor that functions to repress genes required for photoreceptor differentiation. The idea that Ebi normally functions during neuronal differentiation is further supported by the observation that the expression of neuronal markers is reduced in ebi mutant embryos.

In this study, we have independently isolated mutant alleles of ebi as dominant enhancers of a rough eye phenotype induced by elevated dE2F activity. We find that ebi alleles are dominant suppressors of the GMR-p21 phenotype, and that this suppression occurs because halving the gene dosage of Ebi reduces the ability of the p21 CDK inhibitor to block the second mitotic wave. The second mitotic wave represents the synchronous division of cells that have not yet become committed to a cell fate; therefore, neither the p21-induced arrest, nor its suppression by ebi alleles appears to be tied to neuronal differentiation. These results suggest that the function provided by Ebi is important for some aspects of cell cycle control. Consistent with this, we observed inappropriate BrdU incorporation in the PNS and ectopic BrdU incorporation in the CNS of ebi mutant embryos.

Although cell cycle exit and the induction of differentiation are tightly linked processes, it is evident that the role of Ebi in cell cycle control is distinct from its previously reported role in the regulation of Ttk88. Ectopic expression of Ttk88 is able to disrupt neuronal differentiation in both the embryo and the eye disk, yet Ttk88 was insufficient to promote S phase entry or to suppress the p21-induced cell cycle arrest. Thus Ebi appears to have at least two distinct functions; Ebi promotes EGFR-induced down-regulation of Ttk88, and independently it also promotes G₁ arrest in certain cell types. In the light of these results, we suggest that, in some cell types, Ebi could serve to coordinate cell exit with the onset of differentiation.

What is the molecular basis of Ebi function? Closely related homologs of Ebi are found in diverse species. Two regions of these proteins that are most highly conserved, and that define this group of proteins, include an N-terminal domain that resembles an F-box sequence and the C-terminal domain that contains WD40-repeats. The conservation of a divergent F-box suggests that Ebi homologs may target associated proteins for degradation. If so, Ttk88 appears to be the most likely substrate. The expression of an N-terminal fragment of Ebi blocks the Sina–Phyllopod-induced degradation of Ttk88 in tissue culture cells and in eye imaginal disks, and the C-terminus of Ebi can physically interact with Sina–Phyl–Ttk88. Moreover, in vitro assays show that Ebi-associated proteins can promote the degradation of Ttk88. Since F-box proteins can target multiple substrates for degradation [for examples see Cdc4 (Drury et al., 1997; Feldman et al., 1997) and Slimb (Jiang and Struhl, 1998)], this would provide a simple model to explain how Ebi could independently regulate multiple processes. The results described here could be explained simply if Ebi targets Ttk88 and a positive cell cycle regulator such as dE2F or cyclin E (Clurman et al., 1996; Marti et al., 1999). However, we have been unable to detect a physical association between Ebi and either cyclin E or dE2F/DP using the Drosophila proteins, or their human homologs, in either in vitro assays or by transient transfection. In addition, GST–Ebi beads failed to stimulate the degradation of cyclin E, dE2F or dDP in the in vitro degradation assays described above. While the genetic interactions suggested that cyclin E and dE2F might be the most obvious candidates, it is clear that a more systematic approach is needed to identify a suitable cell cycle regulator.

Several observations suggest that Ebi may not act simply as a component of the conventional SCF complex. The N-terminal sequences of Ebi proteins contain many, but not all of the residues that are conserved in F-box proteins that interact with Skp1 (Figure 2). Drosophila Ebi binds very weakly to GST–Skp1, but we have been unable to recover Ebi–Skp1 complexes from transfected cells. The Drosophila genome contains at least nine genes that are related to Skp1 and it is possible that Ebi interacts with a Skp-like protein. It is intriguing that, in addition to the connections between Ebi and Ttk88, Ebi homologs have also been physically linked with transcriptional repressors in two other species. Sif2, the yeast ortholog of Ebi, has been found to interact with Sir4, a protein implicated in transcriptional silencing (Cockell et al., 1998). The recent purification of the SMRT repressor complex revealed that it contains TBL1, a human homolog of Ebi (Guenther et al., 2000). While the role of the Ebi-related proteins in these complexes is uncertain, these studies raise the possibility that Ebi's role in G₁ arrest may be attributable to its association with repressor complexes. One attractive model, which combines all of these studies, is the idea that the divergent F-box in the N-terminus of Ebi recruits a specialized Skp–Cul complex to chromatin in order to selectively ubiquitylate repressor proteins. Clearly, further work is needed to test this and other models of Ebi function.

Materials and methods

Genetics and Drosophila strains

Flies were cultured as previously described (Staehling-Hampton et al., 1999). ebi alleles (CC1, CC3 and CC4) were isolated following EMS mutagenesis by virtue of their failure to complement a deficiency that removes the 21C region. GMR-p21 flies were a gift from I.Hariharan. GMR-dE2F-dDP-p35 flies were described previously (Staehling-Hampton et al., 1999). Sina alleles were kindly provided by A.Travers and M.Freeman. All other fly stocks used were obtained from the Bloomington Indiana stock center. GMR-p21 genetic interaction tests: genetic interactions were performed as described previously (Staehling-Hampton et al., 1999). The following mutations in members of the EGFR and sevenless pathways were tested for their ability to modify the GMR-p21 and GMR-dE2F-dDP-p35 phenotypes: E(sev)3C^E2F (ras hypomorph); E(sev)3C^elb (ras hypomorph); Gap1^A13D; Yan^p; aop¹(yan allele); Egfr E1 (ellipse); sev^d2; rl^x162; raf^c110; rl^sem; sina²; ttk¹. The four Ebi alleles (ebi^K16213, ebi^CC1, ebi^CC3 and ebi^CC4) were balanced over a Cyo-wglacZ balancer. Trans-heterozygous ebi mutant embryos were identified by performing LacZ antibody staining in combination with either anti-BrdU or Mab22C10 antibodies. ebi mutants were unambiguously identified by a lack of the characteristic wingless pattern of LacZ expression.

Plasmids and molecular biology

Full-length Ebi cDNA was obtained from ATCC (Clone #: LD13813). Plasmid construction was performed by PCR using the following primers: (A) 5'-TGAGTTTTTCCAGCGACG-3' plus 5' EcoRI or BamHI linkers; (B) 5'-TCAGAATTTGCGAGGTCC-3' plus 5' HindIII or SpeI linkers; (C) 5'-GAACAAGAGCTGCGGAGTC-3' plus 5' HindIII or SpeI linkers; (D) 5'-GGGACTCCCGCAGCTCTTGTTC-3' plus 5' EcoRI or BamHI linkers. Full-length Ebi: PCR, primers A + B; EbiN: PCR, primers A + C; EbiC: PCR, primers B + D. GST–Ebi fusions were constructed in the pGEX-2T vector (Pharmacia Biotech) and were expressed in JM109 cells for 3 h at 30°C following induction. His₆-tagged Ebi fusion constructs were made in pQE-30 (Qiagen). S2 transfection constructs were made using the pIZT insect select system (Invitrogen). Protein expression and purification was performed following the manufacturer's instructions. Full-length His₆–Ebi was used to immunize mice for monoclonal antibody production. Hybridomas were screened against GST–EbiN and GST–EbiC fusion proteins by ELISA. Positive clones were screened for their ability to recognize endogenous Ebi on westerns from S2 and embryo extracts, to immunoprecipitate endogenous Ebi, to recognize Ebi in immunofluorescence in S2 cells and in immunohistochemistry in embryos. In vitro transcription/translation (pBS-SK+) and pIZT-His₆-V5 (Invitrogen) constructs for Ttk88-myc, Sina and Phyl were derived from pRmHa3 constructs kindly provided by A.Tang and G.Rubin (Tang et al. 1997).

BrdU analysis of eye imaginal disks

Isolation, preparation and analysis of eye imaginal disks were performed as described previously (Staehling-Hampton et al., 1999). Briefly, eye disks from third instar larvae were dissected in Schneider's media and then incubated in BrdU (0.5 mg/ml in Schneider's) for 1 h. Disks were washed in Schneider's, washed in phosphate-buffered saline (PBS), and then fixed for 30 min in 4% formaldehyde in PEM buffer (100 mM PIPES pH 7.0, 2 mM EGTA and 1 mM MgSO₄). After fixation the disks were washed in PBS and 0.3% Triton X-100 for 10 min, PBS and 0.6% Triton X-100 for 30 min, and washed in methanol for 30 min. Disks were then re-hydrated through a methanol/PBS series and transferred to PBS, 0.3% Triton X-100, 2 M HCl for 30 min. The disks were washed twice in PBS for 10 min and then blocked for 30 min in PBS, 0.3% Triton X-100, 10% NGS. After blocking, the disks were incubated overnight at 4°C in a 1:100 dilution of mouse anti-BrdU antibody (Becton Dickinson) in PBS, 0.3% Triton X-100, 10% NGS. The disks were then washed in PBS and 0.3% Triton X-100, and the blocking step repeated. Secondary antibody incubations were performed at room temperature for 2 h using a 1:100 dilution of HRP-goat anti-mouse (Bio-Rad) in PBS, 0.3% Triton X-100, 10% NGS. The disks were washed in PBS and developed using diaminobenzidene (DAB).

Embryo analysis

For BrdU staining, staged embryos were harvested in 0.7% NaCl and 0.01% Triton X-100, washed extensively in dH₂O, then dechorionated in 50% Chlorox for 2 min. Embryos were permeabilized in octane for 6 min and then incubated in 1 mg/ml BrdU (Sigma) in Schneider's medium for 30 min. Embryos were fixed in an equal volume of 4% formaldehyde and heptane. BrdU incorporation was detected using a mouse anti-BrdU antibody (Becton Dickinson, 1:100). For hs-Ttk88 expression, embryos were collected, aged for 6 h at 25°C, then heat shocked for 30 min at 39°C. Following a 1 h recovery period at 25°C, embryos were either aged to stage 16 for immunohistochemistry with Mab22C10 antibody or pulsed with BrdU to determine the pattern of S phases. Wild-type flies were heat shocked in parallel for control. Antibody staining was performed as previously described (Guo et al., 1995).

Binding assays

Radiolabeled in vitro translation products (TNT) were synthesized using the TNT-coupled transcription/translation system as per the manufacturer's instructions (Promega). Five hundred nanograms of the GST fusion protein on Sepharose beads were pre-incubated for 10 min at room temperature with 200 l of Z' buffer [25 mM HEPES pH 7.5, 12.5 mM MgCl₂, 20% (v/v) glycerol, 0.1% NP-40, 150 mM KCl] with 1 l of 200 mM dithiothreitol (DTT), 6 l of 5 mg/ml bovine serum albumin and 50 g/ml ethidium bromide. Two microliters of IVT product were added and the reaction incubated, with gentle rocking for 1 h at room temperature. Matrices were washed three times with 1 ml of NETN buffer (20 mM Tris pH 8.0, 100 mM NaCl, 1 mM EDTA, 0.5% NP-40) without ethidium bromide prior to electrophoresis and autoradiography.

S2 cells were transiently transfected with pIZT-Ebi, pIZT-His₆-V5-Sina and/or pIZT-His₆-V5-Phyl constructs. Seventy-two hours post-transfection, cells were harvested and lysed in ELB250 [50 mM HEPES pH 7.0, 250 mM NaCl, 0.5% NP-40, 0.5 mM EDTA, 1 mM phenylmethylsulfonyl fluoride (PMSF), 0.5 mM NaF, 0.1 mM Na₂VO₃, 0.5 mM DTT and protease inhibitors] prior to clarification of extracts using a 42K spin. Extracts were incubated with Ni–NTA agarose beads for 30 min at 4°C before washing three times with ELB250. Proteins associated with the beads were subjected to electrophoresis and then western blotting with V5 (Invitrogen) and Ebi antibodies.

Pulse–chase experiments

Experiments were performed as previously described by Li et al. (1997). Briefly, S2 cells were transfected by CellFectin (Gibco-BRL) as per the manufacturer's instructions. Forty-eight hours post-transfection, cells were washed in Graces-M (methionine-free media) supplemented with 10% dialysed FBS (Gibco-BRL) and 0.7 mM CuSO₄. Cells were then incubated with 100 Ci of [³⁵S]methionine (Amersham) in 0.5 ml Graces-M with 10% FBS and 0.7 mM CuSO₄ for 4 h. The medium was then replaced with Graces plus 10% FBS, 0.7 mM CuSO₄ and 3 mg/ml methionine, and then cells were incubated for the indicated periods of time. Cells were harvested as per Li et al. (1997). Ttk88myc was then immunoprecipitated using 9E10 antibodies before SDS–PAGE and autoradiography.

In vitro degradation assays

Five hundred nanograms of GST fusion protein on Sepharose beads were incubated with S2 extracts for 1 h at 4°C. Matrices were washed three times with 1 ml of ELB250 before performing the degradation reaction. Reactions consisted of the GST fusion protein (with or without pre-incubation in S2 extracts), radiolabeled protein (1 l of radiolabeled Ttk88–myc, Sina, Phyl co-expressed in the same TNT reaction, or 0.25 l of Ttk88–myc expressed alone, per reaction), 6 g of ubiquitin, 1.5 l of 10 ATP regeneration system (20 mM HEPES pH 7.0, 10 mM ATP, 10 mM magnesium acetate, 300 mM creatine phosphate, 0.5 mg/ml creatine phosphokinase) and 1.5 l of 10 reaction buffer (40 mM magnesium acetate, 10 mM DTT and 1 mM PMSF). The proteolysis inhibitor LLnL (Sigma), dissolved in dimethylsulfoxide (DMSO), was added to the reactions indicated to a final concentration of 50 M. DMSO was added to reactions without LLnL as a control. The total reaction volume was 15 l, not including the bead volume. Reactions were incubated at 30°C for 2 h prior to SDS–PAGE on 8% gels followed by fluorography and autoradiography.

Acknowledgements

The authors thank Yian-Rong Chen for expert technical assistance, T.Laverty for sending the P-element collection, I.Hariharan for the GMR-p21 stocks, A.Travers and M.Freeman for Sina alleles, A.Tang and G.Rubin for Ttk88-myc, Sina and Phyl expression constructs. S.B. was supported by post-doctoral fellowships from EMBO and Human Frontiers Science Project. K.S.-H. received a postdoctoral fellowship from the American Cancer Society. This work was supported by NIH grant number GM53203 to N.D.

References

Asano M, Nevins JR and Wharton RP (1996) Ectopic E2F expression induces S phaseand apoptosis in Drosophila imaginal discs. Genes Dev, 10, 1422–1432. | PubMed | ISI | ChemPort |

Bai C, Sen P, Hofmann K, Ma L, Goebl M, Harper JW and Elledge SJ (1996) Skp1 connects cell cycle regulators to the ubiquitin proteolysis machinery through a novel motif, the F-box. Cell, 86, 263–274. | Article | PubMed | ISI | ChemPort |

Baker N and Rubin GM (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. | PubMed | ISI | ChemPort |

Brook A, Xie JE, Du W and Dyson N (1996). Requirements for dE2F function in proliferating cells and in post-mitotic differentiating cells. EMBO J, 15, 3676–3683. | PubMed | ISI | ChemPort |

Brunner D, Ducker K, Oellers N, Hafen E, Scholz H and Klambt C (1994) The ETS domain protein Pointed-P2 is a target of MAP kinase in the sevenless signal transduction pathway. Nature, 370, 386–389. | Article | PubMed | ISI | ChemPort |

Carthew RW and Rubin GM (1990) seven in absentia, a gene required for specification of R7 cell fate in the Drosophila eye. Cell, 63, 561–577. | Article | PubMed | ISI | ChemPort |

Chang HC, Solomon NM, Wasserman DA, Karim FD, Therrien M, Rubin GM and Wolff T (1995) phyllopod functions in fate determination of a subset of photoreceptors in Drosophila. Cell, 80, 463–472. | PubMed | ISI | ChemPort |

Clurman BE, Sheaff RJ, Thress K, Groudine M and Roberts JM (1996) Turnover of cyclin E by the ubiquitin-proteasome pathway is regulated by cdk2 binding and cyclin phosphorylation. Genes Dev, 10, 1979–1990. | PubMed | ISI | ChemPort |

Cockell M, Renauld H, Watt P and Gasser SM (1998) Sif2p interacts with Sir4p amino-terminal domain and antagonizes telomeric silencing in yeast. Curr Biol, 8, 787–790. | PubMed | ISI | ChemPort |

de Nooij JC and Hariharan IK, (1995). Uncoupling cell fate determination from patterned cell division in the Drosophila eye. Science, 270, 983–985. | PubMed | ChemPort |

de Nooij JC, Letendre MA and Hariharan I (1996) A cyclin-dependent kinase inhibitor, Dacapo, is necessary for timely exit from the cell cycle during Drosophila embryogenesis. Cell, 87, 1237–1247. | Article | PubMed | ISI | ChemPort |

Dickson BJ, Dominguez M, Van der Straten A and Hafen E (1995) Control of Drosophila photoreceptor cell fates by Phyllopod, a novel nuclear protein acting downstream of the Raf kinase. Cell, 80, 453–462. | PubMed | ISI | ChemPort |

Dominguez M, Wasserman JD and Freeman M (1998) Multiple function of the EGF receptor in Drosophila eye development. Curr Biol, 8, 1039–1048. | PubMed | ISI | ChemPort |

Dong X, Zavitz KH, Thomas BJ, Lin M, Campbell S and Zipursky SL (1997) Control of G₁ in the developing Drosophila eye: rca1 regulates Cyclin A. Genes Dev, 11, 94–105. | PubMed | ISI | ChemPort |

Dong X, Tsuda L, Zavitz KH, Lin M, Li S, Carthew RW and Zipursky SL (1999) ebi regulates epidermal growth factor receptor signaling pathways in Drosophila. Genes Dev, 13, 954–965. | PubMed | ISI | ChemPort |

Drury LS, Perkins G and Diffley JF (1997) The Cdc4/34/53 pathway targets Cdc6p for proteolysis in budding yeast. EMBO J, 16, 5966–5976. | Article | PubMed | ISI | ChemPort |

Du W and Dyson N (1999) The role of RBF in the introduction of G₁ regulation during Drosophila embryogenesis. EMBO J, 18, 916–925. | Article | PubMed | ISI | ChemPort |

Duronio RJ and O'Farrell PH (1994) Developmental control of a G₁–S transcriptional program in Drosophila. Development, 120, 1503–1515. | PubMed | ISI | ChemPort |

Duronio RJ, O'Farrell PH, Xie JE, Brook A and Dyson N (1995) The transcription factor E2F is required for S phase during Drosophila embryogenesis. Genes Dev, 9, 1445–1455. | PubMed | ISI | ChemPort |

Feldman RM, Correll CC, Kaplan KB and Deshaies RJ (1997) A complex of Cdc4p, Skp1p and Cdc53p/cullin catalyzes ubiquitination of the phosphorylated CDK inhibitor Sic1p. Cell, 91, 221–230. | Article | PubMed | ISI | ChemPort |

Foley E, O'Farrell PH and Sprenger F (1999) Rux is a cyclin-dependent kinase inhibitor (CKI) specific for mitotic cyclin–Cdk complexes. Curr Biol, 9, 1392–1402. | Article | PubMed | ISI | ChemPort |

Guenther MG, Lane WS, Fischle W, Verdin E, Lazar MA and Shiekhattar R (2000) A core SMRT corepressor complex containing HDAC3 and TBL1, a WD40-repeat protein linked to deafness. Genes Dev, 14, 1048–1057. | PubMed | ISI | ChemPort |

Guo M, Bier E, Jan LY and Jan YN (1995) tramtrack acts downstream of numb to specify distinct cell fates during asymmetric cell divisions in the Drosophila PNS. Neuron, 14, 913–925. | PubMed | ISI | ChemPort |

Jiang J and Struhl G (1998) Regulation of the Hedgehog and wingless signaling pathways by the F-box/WD40 repeat protein Slimb. Nature, 391, 493–496. | Article | PubMed | ISI | ChemPort |

Knoblich JA, Sauer K, Jones L, Richardson H, Saint R and Lehner CF (1994) Cyclin E controls S phase progression and its down-regulation during Drosophila embryogenesis is required for the arrest of proliferation. Cell, 77, 107–120. | Article | PubMed | ISI | ChemPort |

Lai ZC and Rubin GM (1992) Negative control of photoreceptor development in Drosophila by the product of the yan gene, an ETS domain protein. Cell, 70, 609–620. | Article | PubMed | ISI | ChemPort |

Lai ZC, Harrison SD, Karim F, Li Y and Rubin GM (1996) Loss of tramtrack gene activity results in ectopic R7 cell formation, even in a sina mutant background. Proc Natl Acad Sci USA, 93, 5025–5030. | Article | PubMed | ChemPort |

Lane ME, Sauer K, Wallace K, Jan NY, Lehner CF and Vaessin H (1996) Dacapo, a cyclin-dependent kinase inhibitor, stops cell proliferation during Drosophila development. Cell, 87, 1225–1235. | Article | PubMed | ISI | ChemPort |

Li S, Li Y, Carthew RW and Lai ZC (1997) Photoreceptor cell differentiation requires regulated proteolysis of the transcriptional repressor Tramtrack. Cell, 90, 469–478. | Article | PubMed | ISI | ChemPort |

Marti A, Wirbelauer C, Scheffner M and Krek W (1999) Interaction between ubiquitin-protein ligase SCF-Skp2 and E2F-1 underlies the regulation of E2F-1 degradation. Nature Cell Biol, 1, 14–19.

O'Neill EM, Rebay I, Tijan R and Rubin GM (1994) The activities of two Ets-related transcription factors required for Drosophila eye development are modulated by the Ras/MAPK pathway. Cell, 78, 137–147. | Article | PubMed | ChemPort |

Rebay I and Rubin GM (1995). Yan functions as a general inhibitor of differentiation and is negatively regulated by activation of the Ras1/MAPK pathway. Cell, 81, 857–866. | Article | PubMed | ISI | ChemPort |

Richardson H, O'Keefe LV, Marty T and Saint R (1995) Ectopic cyclin E expression induces premature entry into S phase and disrupts pattern formation in the Drosophila eye imaginal disc. Development, 121, 3371–3379. | PubMed | ISI | ChemPort |

Staehling-Hampton K, Ciampa PJ, Brook A and Dyson N (1999) A genetic screen for modifiers of E2F in Drosophila melanogaster. Genetics, 153, 275–287. | PubMed | ISI | ChemPort |

Tang AH, Neufeld TP, Kwan E and Rubin GM (1997) PHYL acts to down regulate TTK88, a transcriptional repressor of neuronal cell fates, by a SINA-dependent mechanim. Cell, 90, 459–467. | Article | PubMed | ISI | ChemPort |

Thomas BJ, Gunning DA, Cho J and Zipursky L (1994) Cell cycle progression in the developing Drosophila eye: roughex encodes a novel protein required for the establishment of G₁. Cell, 77, 1003–1014. | Article | PubMed | ISI | ChemPort |

Thomas BJ, Zavitz KH, Dong X, Lane ME, Weigmann K, Finley RL, Brent R, Lehner CF and Zipursky SL (1997) Roughex down-regulates G₂ cyclins in G₁. Genes Dev, 11, 1289–1298. | PubMed | ISI | ChemPort |

Wolff T and Ready DF (1993) Pattern formation in the Drosophila retina. In Bate,M. and Martinez Arias,A. (eds), The Development of Drosophila melanogaster Cold Spring Harbor Laboratory Press, Cold Spring H