Expression of the cyclin-dependent kinase inhibitor Dacapo is regulated by Cyclin E

Mechanisms of Development 97 (2000) 73±83

www.elsevier.com/locate/modo

Expression of the cyclin-dependent kinase inhibitor Dacapo is regulated by Cyclin E Joriene C. de Nooij, Karolina H. Graber, Iswar K. Hariharan* Massachusetts General Hospital Cancer Center, Building 149, 13th Street, Charlestown, MA 02129, USA Received 13 April 2000; received in revised form 24 July 2000; accepted 24 July 2000

Abstract The Cip/Kip family of cyclin-dependent kinase inhibitors (CKIs) has been implicated in mediating cell cycle arrest prior to terminal differentiation. In many instances, increased expression of CKIs immediately precedes mitotic arrest. However, the mechanism that activates CKI expression in cells that are about to stop dividing has remained elusive. Here we have addressed this issue by investigating the expression pattern of dacapo, a Cip/Kip CKI in Drosophila. We show that the accumulation of dacapo RNA and protein requires Cyclin E and that increased expression of Cyclin E can induce dacapo expression. We also show that the oscillation of the Cyclin E and Dacapo proteins are tightly coupled during ovarian endocycles. Our results argue for a mechanism where Cyclin E/Cdk activity induces Dacapo expression but only within certain windows that are permissive for dacapo expression. q 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Drosophila; Cell cycle; Cyclin-dependent kinase inhibitor; Dacapo; Cyclin E; Embryo; Imaginal disc; Oogenesis

1. Introduction A common feature of almost all cells that undergo terminal differentiation is that they exit from the mitotic cell cycle prior to differentiation. The mechanisms that activate exit from the cell cycle, are most likely controlled by the overall developmental program. Although it remains largely unknown how developmental cues impinge on the regulatory controls of cell proliferation, it is conceivable that a single common mechanism may exist that shuts off the cell cycle in all situations where cells terminally differentiate. One class of proteins that has been implicated in regulating terminal cell cycle arrest is the mammalian Cip/Kip family of cyclin-dependent kinase inhibitors (CKIs). The CKIs of this family, p21 Cip1, p27 Kip1 and p57 Kip2, can inhibit the activity of the cyclin-dependent kinases (cdk's), most notably Cdk4, Cdk6 and Cdk2, which mediate the progression through the G1-phase of the cell cycle and the transition into S-phase (reviewed in Sherr and Roberts, 1995). Activation of these cdk's depends on the association with a cyclin subunit, Cyclin D (for Cdk4 and Cdk6) or Cyclin E (Cdk2), as well as several phosphorylation and dephosphorylation events (Morgan, 1995). The level of Cdk activity in G1 determines whether a new round of cell division is initiated. Hence, a * Corresponding author. Tel.: 11-617-724-9534; fax: 11-617-726-7808. E-mail address: [email protected] (I.K. Hariharan).

decrease in the activity of these cdk's, either by a decrease in cyclin abundance, by inhibitory phosphorylation, or in particular, by the activity of CKIs, can result in cell cycle exit and a termination of the cell division program. Consistent with a presumed function in regulating cell cycle exit is the expression pattern of the Cip/Kip CKIs. In a number of situations, both in tissue culture cells and in vivo, accumulation of these CKIs correlates strongly with the onset of terminal differentiation (Halevy et al., 1995; Matsuoka et al., 1995; Parker et al., 1995; Polyak et al., 1994). To address the signi®cance of these proteins during development, mice de®cient for each of the three CKIs, p21, p27 and p57, have been generated. Although mice de®cient for p21 cip1 do not display any developmental defects, in p27 kip1 and p57 kip2 de®cient mice, abnormalities in cell number and organ size are observed. In many tissues in the p27 and p57 mutant mice, cell proliferation continues inappropriately and cell differentiation is often signi®cantly delayed (Casaccia-Bonne®l et al., 1997; Durand et al., 1998; Fero et al., 1996; Kiyokawa et al., 1996; Nakayama et al., 1996; Yan et al., 1997; Zhang et al., 1997). In each case, however, cells eventually exit from the cell cycle, which suggests that CKIs may not be critical for cell cycle exit per se, but rather may govern the timing of cell cycle exit. A role for CKIs in regulating the timing of cell cycle exit has also been demonstrated by the analysis of the mutant

0925-4773/00/$ - see front matter q 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S 0925-477 3(00)00435-4

74

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

phenotype of the Drosophila Cip/Kip homolog, dacapo (dap) (de Nooij et al., 1996; Lane et al., 1996). In Drosophila embryos, imminent exit from the cell cycle is characterized by an abrupt downregulation of Cyclin E and a brief pulse of Dacapo expression. In dap mutant embryos, most cells in the epidermis and cells in the peripheral nervous system fail to exit from the mitotic cycle at the appropriate time and instead undergo one extra mitotic division. However, following this additional cycle, the mutant cells do arrest mitotically and hence it is likely that additional mechanisms operate to achieve an exit from the mitotic cycle. A likely scenario is that the downregulation of positive cell cycle regulators such as Cyclin E may function in combination with the activation of dap expression to ensure a timely cell cycle arrest. A crucial question that remains to be addressed is: how are the upregulation of dap and the downregulation of cyclin E both precisely coupled to the last cell cycle in the developmental program of differentiating cells? In this study we have sought to gain further insight into the mechanisms that coordinate cell cycle exit with the onset of differentiation by investigating the regulation of Dap expression in vivo. We show that Cyclin E is an important regulator of Dap levels. Alterations in the level of Cyclin E affect the normal pattern of Dap expression in embryos, in imaginal discs, and in the endocycling nurse cells in the ovary. Furthermore, our data indicate that Cyclin E regulates dap at both transcriptional and post-transcriptional levels, and that dap is also regulated by inputs other than Cyclin E.

2. Results 2.1. Expression of Dacapo is severely reduced in cyclin E mutant embryos In a number of different tissues, expression of Dap occurs precisely during the last mitotic division that cells undergo before they terminally differentiate. For instance, in the embryonic epidermis, a rapid accumulation of dap RNA and Dap protein is only detected following S-phase 16, just before these cells arrest in the G1-phase of cycle 17 (Lane et al., 1996; de Nooij et al., 1996). Exit from the cell cycle also requires alterations in the levels of Cyclins and the activity of Cyclin-Cdk complexes (Richardson et al., 1993; Knoblich et al., 1994; Sigrist and Lehner, 1997). Is the induction of dap directly regulated by the developmental cues that dictate cell cycle exit? Or alternatively, is the induction of dap RNA and protein a response to an alteration in the activity of one of the other cell cycle regulators? To help distinguish between these possibilities, we looked at whether alterations in the levels of one of the known cell cycle regulators would affect the normal pattern of Dap expression in different embryonic tissues. We therefore examined embryos that were mutant for the G1-S regulators, cyclin E, E2F1 and DP and for the regulators that mediate the G2-M transition, string (stg), cyclin A, rca1,

cyclin B and cdc2. With the notable exception of cyclin E mutants, no obvious abnormalities in the Dap expression pattern could be observed in any of these mutants (Fig. 1I,J; data not shown). Analysis of cyclin E (cycE) mutant embryos, however, showed signi®cant differences in the expression pattern of Dap. Dap protein levels are severely reduced in the cells of the PNS and are almost absent in the cells of the CNS of embryos homozygous or hemizygous for the cycE null allele cycE Ar95 (Fig. 1A,F). The level of dap RNA is also reduced in the cells of the PNS and the CNS (Fig. 1G,H). This suggests that Cyclin E may regulate dap expression at least in part at the transcriptional level. Immunostaining of cycE mutant embryos with an anti-Elav antibody (Robinow and White, 1991) does not show signi®cant abnormalities in the cellularity of the PNS and CNS, suggesting that the loss of dap expression is not a secondary consequence of a failure to form these tissues (data not shown). Interestingly, no reduction in Dap expression levels is observed in the epidermis of cycE mutant embryos (Fig. 1A,D). However, cycE mutants have been shown to complete all 16 epidermal divisions normally, possibly due to the activity of residual maternal supplies of Cyclin E (Knoblich et al., 1994). It is therefore likely that the residual supply of Cyclin E may also be suf®cient to trigger Dap expression in the epidermis. Alternatively, the apparently normal expression of Dap in the cycE mutant epidermal cells may re¯ect a difference between the mechanisms that regulate Dap expression in the epidermis and in the nervous system. These results also show that normal Dap expression is not contingent upon progression through the cell cycle. Embryos mutant for string, which encodes a Cdc25-like phosphatase, arrest in the G2 phase of cycle 14. Despite the absence of S-phase progression and cell division in string mutants, the onset of Dap expression occurs at the appropriate developmental stage in both the epidermis and the cells of the nervous system (Fig. 1I,J). These observations also show that Dap expression is not normally contingent upon cell cycle progression or on reaching cycle 16 in the epidermis. The reduced level of dap RNA in cycE mutant embryos seems to indicate a role for Cyclin E in the transcriptional regulation of dap. A potential mechanism for the transcriptional regulation of dap by Cyclin E could involve the E2F transcription factor. It has been demonstrated previously that E2F can mediate the activation of a `G1-S transcriptional program' that, in the CNS, is dependent upon cyclin E function. E2F can also activate stg expression during G2 (Duronio and O'Farrell, 1995; Duronio et al., 1995; Neufeld et al., 1998). However, while E2F1 and DP mutant embryos clearly show a reduction in the RNA levels of the E2F responsive genes PCNA and RNR2 (Duronio et al., 1995; Royzman et al., 1997), Dap expression appears to be normal in both E2F and DP mutant embryos (data not shown). Cyclin E is also expressed normally in the CNS in E2F

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

75

Fig. 1. Dacapo expression is severely reduced in cyclin E mutant embryos. Panels (A±F) show the pattern of Dap expression in wildtype (A±C) and cycE mutant (D±F) embryos as assessed by staining with a Dap monoclonal antibody (A±F). Normal levels of Dap protein are observed in the epidermal cells in stage 11±12 cycE mutants (compare A and D). In later stage cycE mutant embryos, a signi®cant reduction in Dap protein is observed in the cells of the PNS (E) and CNS (F). Panels (G,H) show dap RNA expression in the CNS detected by in situ hybridization. Embryos mutant for string (I,J) show Dap protein expression that is appropriate for the developmental stage. Anterior is to the left in all panels. The embryos in (C,F,G,H) are viewed from the ventral side. All other panels show lateral views.

and DP mutant embryos (Duronio and O'Farrell, 1975; Royzman et al., 1997). Thus, although we cannot rule out a redundant function for E2F in regulating dap expression, our experiments do not provide any evidence for E2F as a regulator of dap expression. Taken together, our results show that when Cyclin E activity is abolished in the nervous system, Dap expression is no longer observed thus indicating a requirement for Cyclin E activity in regulating Dap expression. This activity of Cyclin E does not appear to involve the activity of the E2F transcription factor. Moreover, Dap expression in either the epidermis or the nervous system is not contingent upon cell cycle progression. 2.2. Ectopic expression of Cyclin E can induce Dap expression The reduction in Dap levels in cycE mutant embryos

indicates a requirement for Cyclin E in regulating Dap expression. We therefore wanted to determine if Cyclin E is suf®cient to induce Dap expression and examined whether ectopic overexpression of Cyclin E in third instar larval eye and wing imaginal discs was able to induce an ectopic expression of Dap. In wildtype eye-imaginal discs, low levels of Dap protein can be detected only at the posterior edge of the morphogenetic furrow (MF), coinciding with the time that the ®rst ®ve photoreceptor cells (R8, R2±5) exit from the cell cycle and commit to their individual cell fates (Wolff and Ready, 1991; Fig. 2A,C). Occasionally, Dap expressing cells can be detected in the region just posterior to the second mitotic wave, presumably indicating exit from the cell cycle of the cells that have undergone this additional division. In third instar larval wing discs, Dap expression is virtually undetectable. Consistent with a requirement for cyclin E in inducing the normal pattern of Dap expression, ectopic expression of

76

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

Fig. 2. Ectopic expression of Cyclin E results in an ectopic induction of Dacapo. Panels (A±D) show eye-imaginal discs stained with an anti-Dap monoclonal antibody. (C,D) are higher magni®cations of (A,B), respectively. Wildtype eye discs (A,C) show only moderate expression of Dap in photoreceptor precursors immediately posterior to the MF (indicated by arrowhead). In sevcycE eye-discs (B,D), high levels of Dap protein are readily detectable in Sev expressing cells, R3, R4, R1, R6, and R7. Arrows indicate a pair of R3 and R4 photoreceptor cell nuclei. In panels (E,F), dap RNA in situs on wildtype (E) and sevcycE (F) eye discs indicates a small increase in the dap RNA levels in response to ectopic expression of Cyclin E. Panels (G±J) show wing imaginal discs in which an enGAL4 driver drives expression in the posterior compartment of cyclin E (G,H,J) or E2F1 (I) from UAScycE and UASE2F1 transgenes, respectively. Expression of cycE results in a signi®cant induction of Dap protein (G) and dap RNA (H) in the posterior wing disc compartment. A similar increase in dap RNA levels is observed when E2F1 is ectopically expressed (I). A 2.7kb genomic fragment of the dap promoter linked to a lacZ reporter shows increased bgalactosidase activity in the posterior compartment of the wing disc in response to elevated levels of Cyclin E (J). Anterior is to the left in all panels.

Cyclin E in either eye or wing discs results in an ectopic induction of Dap protein. In eye-imaginal discs, a cycE transgene under the control of the sevenless (sev) promoter (sevcycE) signi®cantly increases Dap levels in the sev pattern, i.e. strong Dap expression is observed in the nuclei of the developing R3, R4, R1, R6 and R7 photoreceptor cells (Fig. 2B,D). In contrast, little induction of Dap is observed in the cone cells where the sevenless promoter is also active. These cells have been in a post-mitotic state for a longer period before the promoter is active and may no longer be competent to express Dap in response to Cyclin E. Similar results were obtained when Cyclin E was expressed using the GMR promoter (targeting cycE expression to all cells posterior to the MF; data not shown). In analogous experiments in wing imaginal discs, utilizing the GAL4/ UAS system (Brand and Perrimon, 1993), expression of Cyclin E in the posterior compartment using an engrailed GAL4 (enGAL4) driver results in a dramatic increase in the expression of the Dap protein (Fig. 2G). Thus in both eye

and wing imaginal discs, Cyclin E can induce ectopic expression of Dap. To determine whether the induction of Dap is due to an increase in dap RNA levels, we examined RNA levels of dap in whole-mount preparations by in situ hybridizations in eye and wing discs in which Cyclin E was ectopically expressed. In contrast to the dramatic increase in Dap protein, relatively modest increases in the levels of dap RNA are observed in both the eye and wing discs (Fig. 2E,F,H). Nonetheless, these results are consistent with the notion that Cyclin E regulates dap RNA levels, and furthermore, suggests that dap transcription may be responsive to Cyclin E levels. To test whether dap regulatory elements are responsive to Cyclin E levels, we have generated several lines of transgenic ¯ies that contain a 2.7 kb fragment of the dap promoter region linked to a b-galactosidase reporter. Immunohistochemical analysis of wing discs from ¯ies that contain the pCasdap2.7kb-lacZ transgene either

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

alone, or in the presence of the enGAL4 driver, showed no signi®cant lacZ staining in the posterior compartment of the wing disc. In the presence of both the enGAL4 driver and a UAScycE transgene, increased b-galactosidase activity is clearly detected in the region where Cyclin E is expressed (Fig. 2J). Thus Cyclin E can activate the expression of a reporter gene under the control of dap regulatory elements. Despite our earlier ®ndings that E2F1 was not required for dap expression, ectopic overexpression of E2F1 does result in a small increase in the level of dap RNA (Fig. 2I). This might indicate that E2F is a redundant component of the

77

mechanism by which Cyclin E activates dap expression. Alternatively, E2F may in¯uence dap levels indirectly, since it has been shown previously that overexpression of E2F1 in wing discs results in an increased level of cycE (Neufeld et al., 1998). Thus, ectopic overexpression of Cyclin E can cause a signi®cant induction of Dap protein. While we also do detect an increase in the level of dap RNA, this increase is modest and seems unlikely to fully account for the dramatic increase in the Dap protein level that is observed in response to Cyclin E. Hence, both transcriptional, and post-transcriptional mechanisms may function in mediating

Fig. 3. Dap and Cyclin E proteins oscillate during oogenesis. Panel (A) shows a schematic of the different stages from oogenesis and is adapted from Keyes and Spradling (1997). Panel (B) shows Dap protein expression. The inset shows Dap expression in the germarium at the completion of the mitotic cyst cell divisions. Oscillating expression in individual nurse cell nuclei is revealed by dark staining in some nuclei and light or absent staining of others. Consistently high expression is observed in the oocyte nucleus (ON). Panel (C) shows dap RNA detected by an in situ hybridization. Accumulation in the developing oocyte and low uniform levels in the nurse cells are observed. Increased levels are observed in Stage 10 egg chambers. Panels (D,E) show a double staining experiment where egg chambers are stained with an anti-Cyclin E type II antibody (kindly provided by H. Richardson; panel D) and anti-Dap (panel E). While the patterns are similar, the arrowheads indicate nuclei which show a relative difference in expression suggesting that the oscillations are slightly out of phase.

78

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

the increased expression of Dap protein in response to Cyclin E expression. 2.3. Cyclin E is required for Dap oscillation in the endocycling nurse cells Since Cyclin E appears to be a requirement for the expression of the Dap protein, we examined Dap expression in the context of oscillating levels of Cyclin E in the nurse cell nuclei in the ovary. Oogenesis normally starts with four mitotic germ cell divisions which generate a 16-cell cyst. Surrounded by somatically-derived follicle cells, each cyst forms an individual egg chamber which ultimately gives rise to a mature egg (reviewed by Spradling, 1993; Fig. 3A). One of the 16-germ cell nuclei arrests in the prophase of meiosis I, becomes transcriptionally silent and is speci®ed as the oocyte nucleus (Williamson and Lehmann, 1996; Su et al., 1998). The remaining 15 cells, the nurse cells, proceed through a series of endocycles. In wildtype ovaries, Dap expression is ®rst detected in the germ cell nuclei in the germarium at a time when the four mitotic cyst cell divisions are about to be completed (region 2A of oogenesis; inset in Fig. 3B). In the maturing egg chambers Dap is detected in the endocycling nurse cell

nuclei at levels which, within an individual egg chamber, vary from very high to undetectable (Fig. 3B). This is likely to re¯ect an oscillation in the level of Dap protein during the endocycles, as has been postulated for Cyclin E (Lilly and Spradling, 1996). In contrast to the nurse cell nuclei, high levels of Dap are observed in the oocyte nucleus throughout oogenesis. Consistent with the immunostaining, dap RNA can be observed in the germarium early in oogenesis, and subsequently, high levels of dap RNA are found in the future oocyte throughout oogenesis (Fig. 3C). However, the pattern of dap RNA expression in the nurse cells differs signi®cantly from the pattern of Dap protein. In contrast to the dramatic oscillation of Dap protein levels in the nurse cell nuclei, no oscillation of dap RNA is evident. Only a uniform low level of RNA can be detected in the endocycling nurse cells. Although relative differences may exist but may be beyond our level of detection, this difference in the expression pattern between dap RNA and protein suggests that post-transcriptional mechanisms may regulate the levels of Dap protein in the individual nurse cell nuclei. The expression pattern of Dap protein is reminiscent of the expression pattern of Cyclin E in the ovary, i.e. a strong

Fig. 4. Reduced Dap oscillation in cycE fs(2)01672 mutant ovaries. Panel (A) shows Dap protein expression in wildtype egg chambers showing darkly staining nuclei interspersed with nuclei that do not stain. Panel (B) shows expression in cycE fs(2)01672 egg chambers where the majority of nuclei show intermediate levels of staining.

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

oscillation in the nurse cell nuclei, and a persistently high level of protein in the oocyte nucleus (Lilly and Spradling, 1996). Confocal images of ovaries double labeled with antibodies against both Cyclin E and Dap show that the expression of Cyclin E and Dap are largely overlapping (Fig. 3D,E). However, some of the nurse cell nuclei in the individual egg chambers do differ in the relative levels of Cyclin E and Dap protein (arrowed in Fig. 3D,E), suggesting that the oscillations of Cyclin E and Dap are slightly out of phase. To assess the regulatory relationship between Cyclin E and Dap in the nurse cells more directly, we examined Dap expression in ovaries obtained from females homozygous for the cycE fs(2)01672 allele. This allele of cycE speci®cally perturbs the oscillation of Cyclin E in the nurse cell nuclei and hence ovaries from homozygous cycE fs(2)01672 females show a severe reduction in the amplitude of Cyclin E oscillation (Lilly and Spradling, 1996). Interestingly, these mutant ovaries also show a strong reduction in Dap oscillation, and most nurse cell nuclei show intermediate levels of Dap expression; nuclei with high or undetectable levels of Dap were almost never observed in mutant egg chambers (Fig. 4). Thus in the nuclei of the endocycling nurse cells, expression of Cyclin E and Dap appear to be tightly linked. Dap protein levels oscillate strongly and these oscillations are dampened by a reduction in the extent of Cyclin E oscillation. Since dap RNA levels in the nurse cells seem to remain relatively constant during oogenesis, it appears likely, that in these cells, Cyclin E controls the level of Dap protein predominantly by post-transcriptional mechanisms. 3. Discussion Inhibitors of cyclin dependent kinases (CKIs) appear to be important in determining when a cell makes the developmental transition from being a proliferating precursor cell into a terminally differentiated post-mitotic cell (Elledge et al., 1996; Zavitz and Zipursky, 1998). In many instances, expression of the CKIs immediately precedes exit from the cell cycle and studies in mice, Drosophila, and C. elegans have demonstrated that loss of CKI activity is associated with a delayed exit from the mitotic cell cycle in embryonic tissues (Hong et al., 1998; de Nooij et al., 1996; Lane et al., 1996; reviewed in Zavitz and Zipursky, 1998). The failure to exit from the cell cycle at the appropriate time can lead to increased cell numbers and abnormalities in differentiation. To further understand the mechanisms that control the precise timing of cell cycle exit, we have sought to characterize the regulation of Dap expression in Drosophila. Any mechanism that allows activation of Dap expression at exactly the time that cells are about to complete their individual division program is likely to integrate information from both the proliferation status as well as the differentiation program of these cells. We show here how the

79

mechanism responsible for activating Dap expression may meet both these requirements. Dap expression appears to be activated by a component of the cell cycle machinery, Cyclin E. However, developmental signals appear to restrict the ability of Cyclin E to activate Dap expression to a period during the ®nal division cycle that cells undergo prior to mitotic quiescence. 3.1. Cyclin E is an important regulator of Dap expression Our conclusion that Cyclin E is an important regulator of Dap expression is based on three independent observations. Firstly, cyclin E mutant embryos show a severe reduction in the levels of Dap in the PNS and CNS. Secondly, ectopic expression of Cyclin E can induce ectopic expression of Dap. Thirdly, ovaries obtained from cycE fs(2)01672 homozygous females which have a reduction in the oscillation of Cyclin E levels in the endocycling nurse cell nuclei also show a diminution in the oscillation of Dap levels. Our experiments also indicate that Dap expression is not contingent upon cell cycle progression or on completing a speci®c number of cell cycles. Even though Dap is normally expressed in the epidermis during the G2-phase of cycle 16, Dap expression is unperturbed in both the epidermis and the PNS in mutants for stg, which arrest in the G2 phase of cycle 14 (Edgar and O'Farrell, 1989; Edgar and O'Farrell, 1990). These observations argue that the activation of Dap expression by Cyclin E is not a result of Cyclin E induced cell proliferation or S-phase entry (Richardson et al., 1995), since Dap expression can occur normally in the absence of cell cycle progression. In each of the three situations described above, Cyclin E appears to function as a positive regulator of Dap expression. This was a surprising ®nding in light of previous observations that Dap expression is normally initiated as Cyclin E levels decline (e.g. cycle 16 in the epidermis). Moreover, there also appear to be situations where Cyclin E levels are very high but appear unable to induce expression of Dap. For instance, in the early embryo, dap expression cannot be detected prior to cycle 16 while the levels and activity of Cyclin E/Cdk2 kinase complexes are constitutively high throughout all preceding cell divisions (Sauer et al., 1995). Similarly, Cyclin E is expressed in the eye-imaginal disc in asynchronously cycling precursor cells anterior to the morphogenetic furrow but does not induce signi®cant Dap expression. Taken together, it seems likely that Cyclin E activates Dap expression, but only within permissive windows de®ned by developmental cues. The expression of Cyclin E (endogenous or ectopically expressed) leads to the accumulation of Dap protein only in restricted groups of cells, often cells that are either about to exit from the mitotic cycle or cells that have recently completed mitotic exit. One potential mechanism by which Cyclin E and developmental signals could act together in coordinating the exact onset of Dap expression is schematically summarized in Fig. 5.

80

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

Fig. 5. A model for the regulation of Dap levels by Cyclin E. Prior to cell cycle exit when precursor cells are generated developmental signals prevent the expression of Dacapo despite the presence of cyclin E (phase I). Inhibition of Dap expression could be due to a lack of other essential transcriptional activators or due to the activity of a transcriptional repressor. Close to the time that cells are about to complete their division program and differentiate, developmental constraints are relieved and Cyclin E is able to induce expression of dap (phase II). A rapid accumulation of Dap may further be achieved by promoting Dap translation. In addition, high Cyclin E activity may also promote Dap degradation (phase IIa), ensuring that signi®cant accumulation of Dap is restricted to the interval when Cyclin E levels begin to decline (phase IIb). Once Cyclin E levels have completely declined, dap expression is no longer induced (phase III).

During the generation of precursor cells, expression of Dap fails to be activated or may be repressed, this despite the presence of Cyclin E that drives these cycles (phase I in Fig. 5). However, once the cell division program is about to be completed, developmental constraints are relieved and now Cyclin E is able to activate Dacapo expression (phase II). The initial levels of Dap may remain at low levels due to a relative high rate of degradation (phase IIa; see below), and only once Cyclin E levels start to decline in anticipation of cell cycle exit, a rapid accumulation of Dap is allowed (phase IIb). When the level of Cyclin E declines even further, Dap expression can no longer be induced (phase III). 3.2. Mechanisms by which cycE may regulate Dap accumulation The activation of Dap expression by Cyclin E within the permissive interval (phase II in Fig. 5) is likely to result both from effects on dap transcription as well as the regulation of Dap protein levels by post-transcriptional mechanisms. In cycE mutant embryos, dap RNA levels are signi®cantly reduced. Moreover, ectopic expression of Cyclin E in imaginal discs leads to a detectable increase in dap RNA levels. Our observation that a 2.7 kb genomic fragment containing dap regulatory elements is responsive to increases in the level of Cyclin E further supports the notion that Cyclin E can activate dap transcription. Whether the regulation at the transcriptional level involves the E2F transcription factor remains unclear. While Dap expression is normal in both E2F1 and DP mutant embryos (Duronio et al., 1995; Royz-

man et al., 1997), which indicates that E2F is not essential for the onset of Dap expression, overexpression of E2F1 in wing discs is able to cause a small increase in the level of dap RNA. Thus a redundant function for E2F1 in the regulation of dap expression cannot be excluded. The recent discovery of a second E2F-like gene in Drosophila, E2F2, for which no mutants are currently available (Sawado et al., 1998), may explain a redundant function for E2F1. The regulation of dap at the transcriptional level seems unlikely to account fully for the dynamic expression pattern of Dap protein. In the situations where we overexpressed Cyclin E, the increase in Dap protein is much more dramatic than the increase in dap RNA. Likewise, in the ovary, dramatic oscillations in Dap levels are observed without appreciable ¯uctuations in the levels of dap RNA. Thus, in these situations, Cyclin E seems to regulate dap at a post-transcriptional level. This is perhaps reminiscent of the type of regulation observed for the mammalian CKI p27 kip1. Accumulation of p27 has been shown to depend largely on an increased translation of p27 mRNA (Hengst and Reed, 1994). Although mammalian Cyclin E has, so far, not been implicated in the translational regulation of p27 or any of the other CKIs, our observations suggest that such a mechanism could potentially operate in mammalian cells. Another mode of post-transcriptional regulation that may involve Cyclin E activity is a regulation at the level of protein stability. Degradation of the yeast CKI Far1 and mammalian p27 Kip1 protein has been shown to be accelerated by the phosphorylation of these inhibitors by CDC28/ CLN2 and Cdk2/Cyclin E complexes, respectively. Only upon phosphorylation by the Cdk complex are these CKIs

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

targeted to the ubiquitin-proteasome pathway (Pagano et al., 1995; Henchoz et al., 1997; Sheaff et al., 1997). Like Far1 and p27 Kip1, Dap contains a consensus Cdk-phosphorylation site (Ser 205) and we have previously shown that Dap can be phosphorylated by mammalian Cdk2/Cyclin E complexes in vitro (de Nooij et al., 1996). To test whether Dap degradation is regulated by cdk phosphorylation, we generated ¯ies that express a dap transgene lacking the putative cdk phosphorylation site (UASdap Ala205). So far we have been unable to observe a signi®cant difference in the half-life of the mutant protein in vivo. However, overexpression of the UASdap Ala205 transgene in wing imaginal discs consistently elicits more severe phenotypes than with overexpression of a wildtype dap transgene (data not shown). Thus, Dap degradation may indeed be enhanced by Cdk phosphorylation in vivo. The strong oscillation of Dap protein levels in the nurse cell nuclei in the ovary could potentially be regulated in part by Cyclin E/cdk2-mediated Dap degradation. In support of such a mechanism is the observation that Dap protein fails to disappear completely in the nurse cell nuclei of females which are mutant for the female sterile cyclin E allele (cycE fs(2)01672) where Cyclin E levels are reduced. A role for Cyclin E in Dap degradation may `sharpen' the onset of the peak of Dap expression as cells exit from the cell cycle (Fig. 5). When dap is ®rst transcribed, high levels of cyclin E may also promote Dap protein degradation thus keeping Dap protein accumulation at a low level (phase IIa in Fig. 5). However, once Cyclin E levels start to decline in anticipation of cell cycle exit, Dap is no longer degraded, which results in a rapid accumulation of Dap protein (phase IIb). Soon afterwards, a decline in dap transcription and translation results in a fall in Dap protein levels. 3.3. Developmental factors controlling Dap expression In addition to its regulation by Cyclin E, we have shown that Dap expression is also dependent upon other factors. It is likely that, in addition to cyclin E responsive elements, the dap promoter contains multiple regulatory elements that direct the expression of dap to speci®c tissues at speci®c times during development. For the stg gene, which regulates entry into mitosis 14, an extensive array of such elements has been identi®ed (Edgar et al., 1994; Lehman et al., 1999). In addition, the presence of tissue speci®c repressors could prevent precocious expression of Dap expression even in the presence of Cyclin E activity. For instance, it seems very likely that the lack of Dap expression in the embryonic epidermis prior to cycle 16, despite high levels of Cyclin E, is due to the activity of a repressor of dap transcription. A detailed analysis of the dap promoter will help us understand how these different mechanisms cooperate to regulate dap transcription. 3.4. Conclusions Our experiments show that cyclin E is important in regulating the expression of dap and that it does so via both

81

transcriptional and post-transcriptional mechanisms. The ®nding that dap can be regulated in a variety of ways differs from the CKIs in vertebrates where each of the CKIs appears to be regulated predominantly by either transcriptional or post-transcriptional mechanisms. The CKIs p21Cip1 and p57Kip2 both appear to be regulated primarily at the level of transcription whereas p27Kip1 is primarily regulated at the translational level and at the level of protein degradation. The reasons for why different CKIs are regulated via different mechanisms are not understood but it may be that different modes of regulation are required in speci®c biological situations. Dap is the only CKI that has been identi®ed in Drosophila to date and appears to function in a wide variety of situations. For instance, Dap appears to function during endoreplication, perhaps to transiently inactivate Cyclin E, and is also expressed in the pole cells during their long period of mitotic quiescence. In addition, in the cells of the PNS and CNS Dap appears to function to facilitate a permanent exit from the cell cycle. In order to accommodate these diverse functions, dap may have acquired a more complex mode of regulation during its evolution. The ability of Cyclin E to regulate dap expression in several ways may re¯ect a necessity to modulate with precision the relative levels of a cyclin and its inhibitor in a variety of biological situations. 4. Experimental procedures 4.1. Drosophila strains and cultures Flies were grown on a standard cornmeal medium at 258C. Stocks and alleles used in experiments were stg 7B (Edgar and O'Farrell, 1989); cycE Ar95, Df(2)191 and cycE fs(2)01672 (Knoblich et al., 1994; Lilly and Spradling, 1996); cycA neo114 (Lehner and O'Farrell, 1990); Df(2)95A (Knoblich and Lehner, 1993); rca1 (Dong et al., 1997); dE2F 91 and dE2F 155 (Duronio et al., 1995) and dDP a2 and dDP a4 (Royzman et al., 1997). The enGAL4 driver (Glise and Noselli, 1997) was obtained from the Bloomington Stock Center, the UAScycE and sevcycE lines from C. Lehner, and the GMRdE2F/dDP and UASE2F1 transgenic lines from N. Dyson (Du et al., 1996; Neufeld et al., 1998). 4.2. Molecular biology pCasper2.7dap-lacZ was constructed by cloning a BamHI-SalI fragment into pCaSpeRbgal (Thummel et al., 1988) by standard methods (Sambrook et al., 1989). This fragment includes the region 3075 bp to 398 bp upstream of the ATG initiation codon. 4.3. Immunohistochemistry and immuno¯uorescence Monoclonal antibodies to Dap were raised against a full length Histidine-tagged Dap protein, expressed using a pQIA vector (Qiagen). Antibodies were generated with the

82

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83

help of Chidi Ngwu and the Cancer Center Monoclonal Antibody Facility according to procedures described in Harlow and Lane (1988). Immunostaining of whole-mount embryos and eye-imaginal discs was performed as described (Patel, 1994; Tomlinson and Ready, 1987). Ovaries were dissected in PBS, and ®xed in PLP (0.01 M NaIO4, 0.075 M Lysine, 0.037 M NaPO4, pH 7.2) containing 2% formaldehyde, for 30 min at room temperature (RT). Following washes in PBS and incubation in PTN (10% NGS, 0.3% Triton X-100 in PBS), antibody staining was performed overnight at 48C. Staining with secondary antibody (3 h at RT) was either followed by DAB/peroxidase detection or directly analysed by confocal microscopy. Antibodies were used at the following dilutions, anti-Dap (NP1) at 1:4, rat anti-Elav at 1:10 (Robinow and White, 1991), anti-b-galactosidase (Cappel) at 1:1000, HRP-conjugated anti-mouse IgG (Biorad) at 1:200, Cy3-conjugated anti-mouse IgG (Jackson) at 1:50, and FITC-conjugated anti-rat (Jackson) at 1:50. Both the mouse monoclonal antibody (8B10; used at 1:5), and a rat polyclonal antiserum (used at 1:200) directed against Cyclin E type I and type II, respectively, were kindly provided by H. Richardson. RNA in situ hybridizations and dap RNA probes used were as described (de Nooij et al., 1996; Tautz and Pfei¯e, 1989). Acknowledgements We greatly appreciate the generous gifts of antibodies, cDNAs and ¯y stocks by Helena Richardson, Terry OrrWeaver, Mary Lilly, Alan Spradling, Larry Zipursky, Bruce Edgar, Nick Dyson, and Christian Lehner. Thanks to Chide Ngwu and the Cancer Center monoclonal antibody facility for help in generating Dacapo monoclonal antibodies and to Yimen Ge for help with confocal microscopy. We thank members of the Hariharan and Dyson labs for discussions and Mary Lilly for advice on oogenesis. This work was supported in part by a grant from the National Eye Institute (EY11632) to I.K.H. References Brand, A.H., Perrimon, N., 1993. Targeted gene expression as a means of altering cell fates and generating dominant phenotypes. Development 118, 401±415. Casaccia-Bonne®l, P., Tikoo, R., Kiyokawa, H., Fredrich Jr, V., Chao, M.V., Koff, A., 1997. Oligodendrocyte precursor differentiation is perturbed in the absence of the cyclin-dependent kinase inhibitor p27Kip1. Genes Dev. 11, 2335±2346. de Nooij, J.C., Letendre, M.A., Hariharan, I.K., 1996. A cyclin-dependent kinase inhibitor. Dacapo, is necessary for timely exit from the cell cycle during Drosophila embryogenesis. Cell 87, 1237±1247. Dong, X., Zavitz, K.H., Thomas, B.J., Lin, M., Campbell, S., Zipursky, S.L., 1997. Control of G1 in the developing Drosophila eye: rca1 regulates Cyclin A. Genes Dev. 11, 94±105. Du, W., Xie, J.-E., Dyson, N., 1996. Ectopic expression of dE2F and dDP induces cell proliferation and death in the Drosophila eye. EMBO J. 15, 3684±3692.

Durand, B., Fero, M.L., Roberts, J.M., Raff, M.C., 1998. p27Kip1 alters the response of cells to mitogen and is part of a cell-intrinsic timer that arrests the cell cycle and initiates differentiation. Curr. Biol. 8, 431±440. Duronio, R.J., O'Farrell, P.H., 1995. Developmental control of the G1 to S transition in Drosophila: cyclin E is a limiting downstream target of E2F. Genes Dev. 9, 1456±1468. Duronio, R.J., O'Farrell, P.H., Xie, J.-E., Brook, A., Dyson, N., 1995. The transcription factor E2F is required for S phase during Drosophila embryogenesis. Genes Dev. 9, 1445±1455. Edgar, B.A., O'Farrell, P.H., 1989. Genetic control of cell division patterns in the Drosophila embryo. Cell 57, 177±187. Edgar, B.A., O'Farrell, P.H., 1990. The three postblastoderm cell cycles of Drosophila embryogenesis are regulated in G2 by string. Cell 62, 469± 480. Edgar, B.A., Lehman, D.A., O'Farrell, P.H., 1994. Transcriptional regulation of string (cdc25): a link between developmental programming and the cell cycle. Development 120, 3131±3143. Elledge, S.J., Winston, J., Harper, J.W., 1996. A question of balance: the role of cyclin-dependent kinase inhibitors in development and tumorigenesis. Trends Cell Biol. 10, 388. Fero, M.L., Rivkin, M., Tasch, M., Porter, P., Carow, C.E., Firpo, E., Polyak, K., Tsai, L.-H., Broudy, V., Perlmutter, R.M., Kaushansky, K., Roberts, J.M., 1996. A syndrome of multiorgan hyperplasia with features of gigantism, tumorigenesis, and female sterility in p27kip1de®cient mice. Cell 85, 733±744. Glise, B., Noselli, S., 1997. Coupling of Jun amino-terminal kinase and Decapentaplegic signaling pathways in Drosophila morphogenesis. Genes Dev. 11, 1738±1747. Halevy, O., Novitch, B.G., Spicer, D.B., Skapek, S.X., Rhee, J., Hannon, G.J., Beach, D., Lasser, A.B., 1995. Correlation of terminal cell cycle arrest of skeletal muscle with induction of p21 by MyoD. Science 267, 1018±1021. Harlow, E., Lane, D., 1988. Antibodies: a laboratory manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Henchoz, S., Chi, Y., Catarin, B., Herskowitz, I., Deshaies, R.J., Peter, M., 1997. Phosphorylation- and ubiquitin-dependent degradation of the cyclin dependent kinase inhibitor Far1p in budding yeast. Genes Dev. 11, 3046±3060. Hengst, L., Reed, S.I., 1994. Translational control of p27Kip1 accumulation during the cell cycle. Science 271, 1861±1864. Hong, Y., Roy, R., Ambros, V., 1998. Developmental regulation of a cyclin-dependent kinase inhibitor controls postembryonic cell cycle progression in C. elegans. Development 125, 3585±3597. Keyes, L.N., Spradling, A.C., 1997. The Drosophila gene fs(2)cup interacts with otu to de®ne a cytoplasmic pathway required for the structure and function of germ-line chromosomes. Development 124, 1419±1431. Kiyokawa, H., Kineman, R.D., Manova-Todorova, K.O., Soares, V.C., Hoffman, E.S., Ono, M., Khanam, D., Hayday, A.C., Frohman, L.A., Koff, A., 1996. Enhanced growth of mice lacking the cyclin-dependent kinase inhibitor function of p27kip1. Cell 85, 721±732. Knoblich, J.A., Lehner, C.F., 1993. Synergistic action of Drosophila cyclins A and B during the G2-M transition. EMBO J. 12, 65±74. Knoblich, J.A., Sauer, K., Jones, L., Richardson, H., Saint, R., Lehner, C.F., 1994. Cyclin E controls S phase progression and its downregulation during Drosophila embryogenesis is required for the arrest of cell proliferation. Cell 77, 107±120. Lane, M.E., Sauer, K., Wallace, K., Jan, Y.N., Lehner, C.F., Vaessin, H., 1996. Dacapo, a cyclin-dependent kinase inhibitor, stops cell proliferation during Drosophila development. Cell 87, 1225±1235. Lehman, D.A., Patterson, B., Johnston, L.A., Balzer, T., Britton, J.S., Saint, R., Edgar, B.A., 1999. Cis-regulatory elements of the mitotic regulator, string/Cdc25. Development 126, 1793±1803. Lehner, C.F., O'Farrell, P.H., 1990. The roles of Drosophila cyclins A and B in mitotic control. Cell 61, 535±547. Lilly, M.A., Spradling, A.C., 1996. The Drosophila endocycle is controlled by Cyclin E and lacks a checkpoint ensuring S-phase completion. Genes Dev. 10, 2514±2526.

J.C. de Nooij et al. / Mechanisms of Development 97 (2000) 73±83 Matsuoka, S., Edwards, M.C., Bai, C., Parker, S., Zhang, P., Baldini, A., Harper, J.W., Elledge, S.J., 1995. p57KIP2, a structurally distinct member of the p21CIP1 Cdk inhibitor family, is a candidate tumor suppressor gene. Genes Dev. 9, 650±662. Morgan, D.O., 1995. Principles of CDK regulation. Nature 374, 131±134. Nakayama, K., Ishida, N., Shirane, M., Inomata, A., Inoue, T., Shishido, N., Horii, I., Loh, D.Y., Nakayama, K., 1996. Mice lacking p27Kip1 display increased body size, multiple organ hyperplasia, retinal dysplasia and pituitary tumors. Cell 85, 707±720. Neufeld, T.A., de la Cruz, A.F.A., Johnston, L.A., Edgar, B.A., 1998. Coordination of growth and cell division in the Drosophila wing. Cell 93, 1183±1193. Pagano, M., Tam, S.W., Theodoras, A.M., Beer-Romero, P., Del Sal, G., Chau, V., Yew, P.R., Draetta, G.F., Rolfe, M., 1995. Role of the ubiquitin-proteasome pathway in regulating the abundance of the cyclindependent kinase inhibitor p27. Science 269, 682±685. Parker, S.B., Eichele, G., Zhang, P., Rawls, A., Sands, A.T., Bradley, A., Olson, E.N., Harper, W.J., Elledge, S.J., 1995. p53-independent expression of p21CIP1 in muscle and other terminally differentiating cells. Science 267, 1024±1027. Patel, N.H., 1994. Imaging neuronal subsets and other cell types in wholemount Drosophila embryos and larvae using antibody probes. In: Goldstein, L.S.B., Fyrberg, E.A. (Eds.). Drosophila melanogaster: Practical Uses in Cell and Molecular Biology, Academic Press, San Diego, CA, pp. 445±487. Polyak, K., Kato, J.-y., Solomon, M.J., Sherr, C.J., Massague, J., Roberts, J.M., Koff, A., 1994. p27Kip1, a cyclin-Cdk inhibitor, links transforming growth factor-b and contact inhibition to cell cycle arrest. Genes Dev. 8, 9±22. Richardson, H.E., O'Keefe, L.V., Reed, S.I., Saint, R., 1993. A Drosophila G1-speci®c cyclin E homolog exhibits different modes of expression during embryogenesis. Development 119, 673±690. Richardson, H., O'Keefe, L.V., Marty, T., 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. Robinow, S., White, K., 1991. Characterization and spatial distribution of the ELAV protein during Drosophila melanogaster development. J. Neurobiol. 22, 443±461. Royzman, I., Whittaker, A.J., Orr-Weaver, T.L., 1997. Mutations in Drosophila DP and E2F distinguish G1-S progression from an associated transcriptional program. Genes Dev. 11, 1999±2011. Sambrook, J., Fritsch, E.F., Maniatis, T., 1989. Molecular Cloning: A Laboratory Manual, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY. Sauer, K., Knoblich, J.A., Richardson, H., Lehner, C.F., 1995. Distinct

83

modes of cyclin E/cdc2c kinase regulation and S-phase control in mitotic and endoreduplication cycles of Drosophila embryogenesis. Genes Dev. 9, 1327±1339. Sawado, T., Yamaguchi, M., Nishimoto, Y., Ohno, K., Sakaguchi, K., Matsukage, A., 1998. E2F2, a novel E2F-family transcription factor in Drosophila melanogaster. Biochem. Biophys. Res. Commun. 251, 409±415. Sheaff, R.J., Groudine, M., Gordon, M., Roberts, J.M., Clurman, B.E., 1997. Cyclin E-CDK2 is a regulator of p27Kip1. Genes Dev. 11, 1464±1478. Sherr, C.J., Roberts, J.M., 1995. Inhibitors of mammalian G1 cyclin-dependent kinases. Genes Dev. 9, 1149±1163. Sigrist, S.J., Lehner, C.F., 1997. Drosophila ®zzy-related down-regulates mitotic cyclins and is required for cell proliferation arrest and entry into endocycles. Cell 90, 671±681. Spradling, A.C., 1993. Developmental genetics of oogenesis. In: Bate, M., Arias, A.M. (Eds.). The Development of Drosophila melanogaster, Cold Spring Harbor Laboratory Press, Cold Spring Harbor, NY, pp. 1±70. Su, T.T., Campbell, S.D., O'Farrell, P.H., 1998. The cell cycle program in germ cells of the Drosophila embryo. Dev. Biol. 196, 160±170. Tautz, D., Pfei¯e, C., 1989. A non-radioactive in situ hybridization method for the localization of speci®c RNAs in Drosophila embryos reveals translational control of the segmentation gene hunchback. Chromosoma 98, 81±85. Thummel, C.S., Boulet, A.M., Lipshitz, H.D., 1988. Vectors for Drosophila P-element-mediated transformation and tissue culture transfection. Gene 74, 445±456. Tomlinson, A., Ready, D.F., 1987. Cell fate in the Drosophila ommatidium. Dev. Biol. 120, 366±376. Williamson, A., Lehmann, R., 1996. Germ cell development in Drosophila. Annu. Rev. Cell Dev. Biol. 12, 365±391. Wolff, T., Ready, D.F., 1991. The beginning of pattern formation in the Drosophila compound eye: the morphogenetic furrow and the second mitotic wave. Development 113, 841±850. Yan, Y., Frisen, J., Lee, M.-H., Massague, J., Barbacid, M., 1997. Ablation of the CDK inhibitor p57kip2 results in increased apoptosis and delayed differentiation during mouse development. Genes Dev. 11, 973±983. Zavitz, K.H., Zipursky, S.L., 1998. Controlling cell proliferation in differentiating tissues: genetic analysis of the negative regulators of G1-S phase progression. Curr. Opin. Cell. Biol. 9, 773±781. Zhang, P., Liegeois, N.J., Wong, C., Finegold, M., Hou, H., Thompson, J.C., Silverman, A., Harper, J.W., DePinho, R.A., Elledge, S.J., 1997. Altered cell differentiation and proliferation in mice lacking p57KIP2 indicates a role in Beckwith-Wiedemann syndrome. Nature 387, 151± 158.