|Home | About | Journals | Submit | Contact Us | Français|
Cytokinesis occurs just as chromosomes complete segregation and reform nuclei. It has been proposed that cyclin/Cdk kinase inhibits cytokinesis until exit from mitosis; however, the timer of cytokinesis has not been experimentally defined. Whereas expression of a stable version of Drosophila cyclin B blocks cytokinesis along with numerous events of mitotic exit, stable cyclin B3 allows cytokinesis even though it blocks late events of mitotic exit . We examined the interface between mitotic cyclin destruction and the timing of cytokinesis.
In embryonic mitosis 14, the cytokinesis furrow appeared 60 s after the metaphase/anaphase transition and closed 90 s later during telophase. In cyclin B or cyclin B3 mutant cells, the cytokinesis furrow appeared at an earlier stage of mitosis. Expression of stable cyclin B3 delayed and prolonged furrow invagination; nonetheless, cytokinesis completed during the extended mitosis. Reduced function of Pebble, a Rho GEF required for cytokinesis, also delayed and slowed furrow invagination, but incomplete furrows were aborted at the time of mitotic exit. In functional and genetic tests, cyclin B and cyclin B3 inhibited Pebble contributions to cytokinesis.
Temporal coordination of mitotic events involves inhibition of cytokinesis by cyclin B and cyclin B3 and punctual relief of the inhibition by destruction of these cyclins. Both cyclins inhibit Pebble-dependent activation of cytokinesis, whereas cyclin B can inhibit cytokinesis by additional modes. Stable cyclin B3 also blocks the later return to interphase that otherwise appears to impose a deadline for the completion of cytokinesis.
Cells don’t simply exit mitosis: the chromosomes are faithfully segregated, the mitotic spindle is transformed and ultimately disassembled, nuclei and other organelles are rebuilt, and the cell pinches into two in the process of cytokinesis. Little is known about the regulation of these events. In particular, cytokinesis must be properly coordinated with spindle function for accurate chromosome partitioning into daughter cells.
Activation of protein degradation by the Anaphase Promoting Complex (APC) is essential for mitotic exit, and one important pathway of its action has been identified [2-5]. APC directs metaphase destruction of securin, a protein that inhibits separase. Separase, a specific endoprotease, then cleaves cohesin, a protein that holds sister chromatids together. The discovery of this pathway for the release of sister chromatids superseded an earlier supposition that destruction of cyclins would reverse cyclin/Cdk activation to initiate exit from mitosis. Nonetheless, destruction of mitotic cyclins does contribute to mitotic exit.
Stable versions of cyclin B in metazoans or Clb2 in S. cerevisiae failed to block transition to anaphase, but they did block final mitotic exit [1, 2, 6-17]. Thus, cyclin destruction controls steps such as spindle disassembly and cytokinesis, and recent results suggest contributions beyond those seen in early work, which tested only one member of the family of mitotic cyclins. Three evolutionarily conserved classes of cyclins — a cyclin A, cyclin B, and cyclin B3 — contribute to mitosis in Drosophila, whereas S. cerevisiae has six mitotic cyclins (CLB1—6). In systems from yeast to frogs, different cyclin types have overlapping but nonidentical functions [8, 12, 18-21]. The mitotic cyclins of Drosophila are degraded in succession: cyclin A prior to metaphase/anaphase; cyclin B at the transition to anaphase; and cyclin B3 early during anaphase [12, 20, 22-24]. Additionally, stabilization of each of these cyclins blocks diminishing subsets of mitotic exit events as if the destruction schedule orders and times these events [1, 12]. In this regard, stabilization of cyclin B blocked cytokinesis, whereas stabilization of cyclin B3 did not .
The onset of cytokinesis is marked by the assembly of a cortical actin ring, a step thought to be regulated by the small GTP binding protein RhoA [25-27]. Drosophila pebble (pbl), which encodes a RhoA-specific exchange factor, is required for cytokinesis . pbl mutant embryos show no obvious defect during the first 13 mitotic cycles, which occur in a syncytial cytoplasm without cytokinesis, but show failures of cytokinesis 14 [28, 29]. Immunostaining reveals intriguing changes in Pebble abundance and localization; Pebble declines in early mitosis, then intensifies cortically and subsequently focuses to the equatorial cortex as cytokinesis furrowing occurs . Pebble staining persists along the furrow as it deepens and then migrates into the newly formed telophase nuclei. The requirement for Pebble in cytokinesis and the cortical presence of the protein exclusively when it might be needed suggests that regulation of its abundance and location might define a window of opportunity for cytokinesis.
We describe the roles that cyclin B and cyclin B3 play in timing cytokinesis and in inhibiting Pebble-dependent contributions to cytokinesis.
As a foundation for our analysis, we examined the normal progress to cytokinesis in live and fixed Drosophila embryos during cell cycle 14 (Figures 1A—1A”). Metaphase cells exhibited fan-shaped arrays of microtubles organized into a broad spindle, short astral microtubules, and a cortex that stained brightly with phalloidin, but they showed no sign of cytokinesis (cell 1). During anaphase, microtubule bundles became parallel as the spindle narrowed, astral microtubules reached the equatorial cell cortex, and the cytokinesis furrow appeared at this position (cells 2 and 3) just as the elongating spindle reached 11 μm (Figure 1B). The furrow deepened during anaphase and telophase, and cytokinesis was completed before the midbody depolymerized (cells 4 and 5). Finally, cortical actin staining declined to interphase levels.
Real-time detection of a plasma membrane protein fused to GFP (see Movie 1, available with the Supplementary Material online) or a HistoneH2-GFP fusion (our unpublished data) revealed membrane and chromosome dynamics, respectively. The cytokinesis furrow appeared about 60 s after the metaphase/anaphase transition and 30 s after the beginning of cell elongation (anaphase B). Moreover, cytokinesis was completed about 90 s after its initiation (Figure 1C). Because the furrow was not initiated until after cyclin degradation [12, 20, 22-24], we asked if cyclin B and/or cyclin B3 activities prevent earlier onset of cytokinesis initiation.
Mitotic cyclins B and B3 are partially redundant, and each is individually dispensable for many mitoses in Drosophila [19, 20]. Nonetheless, it was suggested that coordination of mitotic-exit events was disturbed in cells deficient in cyclin B . We tested whether loss of cyclin B or cyclin B3 affected cytokinesis timing during mitosis of cell cycle 16, a stage of embryogenesis with a negligible residue of maternal cyclin products [19, 20]. Wild-type cells that have not yet separated their daughter complements of DNA beyond a threshold degree can be reliably predicted to lack a cytokinesis furrow (Figures 2A and 2D). In contrast, cells below this threshold show well-established cytokinesis furrows in cyclin B and cyclin B3 mutant embryos (Figures 2B, 2C, 2E, and 2F; Figure S1, available with this article online). Furthermore, as expected for precocious cytokinesis in the cyclin B and cyclin B3 mutants, the proportion of anaphase cells that had not yet begun cytokinesis was reduced (Figures 2D—2F). It can also be noted that an increased proportion of anaphase cells are in the process of cytokinesis in the mutants (Figures 2E and 2F), suggesting that furrow ingression is slowed, perhaps as a secondary consequence of early initiation and interference by the as-yet-immature spindle. In any case, cyclin B and cyclin B3 are required to delay cytokinesis until chromosomes separation is well on its way.
We estimated the length of delay due to cyclin B and cyclin B3 that would account for the change in the coordination of cytokinesis and progress of chromosome separation that we observed in the mutants. Using real-time analysis of histone-GFP in wild-type embryos during cycle 16, we have developed a conversion table relating chromosome separation to time. The degree of chromosome separation when cytokinesis furrows are first detected in cyclin mutants versus controls (Figures 2D—2F) corresponds to an advance of roughly 25 and 15 s in cyclin B and cyclin B3 mutants, respectively, if one assumes that all of the change in coordination is due to delay in cytokinesis. If it is mitotic destruction that ordinarily relieves cyclin inhibition of cytokinesis in the wild-type, stabilization of the cyclins ought to extend the period of inhibition and thus further delay cytokinesis.
Studies in numerous organisms have shown that stabilization of cyclin B blocks mitotic exit events, including cytokinesis [1, 6, 7, 9, 10, 12-16]. In Drosophila the arrest occurs with incomplete chromosome separation [1, 12] and robust cortical phalloidin staining, the latter indicating F-actin accumulation (Figures 3A and 3A’). In marked contrast, stabilization of Drosophila cyclin B3 blocked exit from mitosis without blocking cytokinesis , as if destruction of cyclin B, but not cyclin B3, governed the onset of cytokinesis. However, intermediates in cytokinesis were frequently seen in embryos expressing cyclin B3 (Figures 3B and 3B’ versus Figures 1A and 1A’), suggesting that cytokinesis progressed slowly. This was confirmed in live records following the induction of stable cyclin B3 in embryos marked with histone H2-GFP (Movie 2, available with this article online) or a membrane-GFP marker (Movie 3, available with this article online). The time between metaphase/anaphase and visible furrowing increased from 1 min in the control to 3 min (Movie 2, available with this article online), and furrow ingression was slower (360 versus 90 s). The delay was observed consistently, although its duration ranged from 30 (see Movie 3) to 300 s (our unpublished data). Similarly, the rate of furrow ingression varied but often took more than 5 min (e.g., Movies 2 and 3, available with this article online).
Although retarded by stable cyclin B3, cytokinesis continues apparently to completion. In fixed preparations, some cells with completed furrows appeared to be escaping the arrest; they had reduced phospho-histone staining and a midbody-like structure (Figures 3B and 3B’, cell 7), but other cells retained robust phosphohistone staining, condensed chromosomes, and obvious spindles and lacked a nuclear envelope (Figures 3B, 3B’, 3E, 3E’, 3F, and 3F’ and our unpublished data). Thus, cyclin B3/Cdk1 levels sufficient to block many features of the exit from mitosis slowed but did not block cytokinesis. Consequently, whereas the previous work suggested an absolute distinction between the ability of cyclin B to inhibit cytokinesis and cyclin B3’s ability to do so, we now conclude that they both have this ability but that only cyclin B gives a complete block.
To explore the regulatory interface between cyclin destruction and initiation of cytokinesis, we examined the influence of cyclin B and cyclin B3 on Pebble, an upstream activator of cytokinesis.
Because Pebble activation of RhoA is likely to be a key upstream step in reorganizing cortical actin to initiate cytokinesis, we expected pbl mutants to lack cytokinesis furrows. Surprisingly, in cells of pbl2 mutant embryos, we frequently detected furrow initiation and ingression during mitosis 14 (our unpublished data), and, to a lesser extent, in mitosis 15 (Figures 4A—4C’). When cells in the final steps of mitosis were ordered according to the degree of spindle disassembly, chromosome decondensation and reformation of the nuclear envelope, it appeared that ingressed but incomplete furrows were in the process of regression (Figures 4B—4C’). Furthermore, abortion of incomplete furrows appears to be the source of the binucleate cells that typify interphase 15 pbl mutant embryos. These findings indicate that either Pebble is not absolutely essential for furrow initiation or the mutation does not fully remove Pebble function.
Staining for Pebble in pbl2 mutant embryos reveals substantial staining in cycle 15 nuclei (see Figure S2, available with this article online). Because the pbl2 allele drastically truncates the protein and should eliminate Pebble epitopes , we infer that maternally encoded Pebble persists into cycle 15. Consequently, the phenotype at cytokinesis 14 appears to be due to a reduction in function, rather than the absence of function.
We used RNA interference (RNAi) to further probe the requirement for Pebble. Analysis of fixed embryos showed that pbl RNAi mimicked the pbl mutant phenotype (and produced binucleate cells) while only reducing, not eliminating, Pebble staining (Figure S2, available with this article online). In real-time images of uninjected and control injected cycle-14 embryos, cytokinesis furrows appeared 30 s after cells began to elongate (taken as initiation of anaphase B), whereas cytokinesis furrows appeared more than 100 s after the onset of cell elongation in embryos injected with pbl dsRNAs (Movies 1, 4, and 5, available with this article online). Moreover, cytokinesis furrows aborted either soon after cytokinesis initiation (Movie 4) or when furrows had deeply ingressed (Figures 4D—4I and Movie 5). Thus, pbl RNAi delays the appearance of the cytokinesis furrow and slows furrow ingression. As cells exit mitosis, nuclear envelopes reform, and the still incomplete furrows abort to produce binucleate cells. The abortion of furrows delayed by the reduction of Pebble function contrasts with the successful completion of the dramatically delayed furrows seen in cyclin B3-expressing cells.
Hypothesizing that extending mitosis might allow completion of the slow cytokinesis events in pbl2 mutants, we expressed stable cyclin B3 in the mutant embryos. Rather than permitting completion, the combination gave an arrest where cytokinesis furrows were absent (or exceptional) (Figures 4J and 4J’). Moreover, late spindles were broad and short (Figures 4J and 4J’ and our unpublished data). We conclude that stable cyclin B3 and the pbl2 mutant act in synergy to inhibit cytokinesis initiation and mitotic progression.
These results suggested that cyclin B3 (and possibly cyclin B) inhibits a residual activity of the Pebble pathway or suppresses a parallel pathway that acts in conjunction with Pebble. To further test interactions with Pebble, we developed a genetic approach to assay whether cyclin B or cyclin B3 mutations modify the cytokinesis defects induced by the expression of a dominant-negative Pebble construct.
A deletion of pebble sequences encoding a conserved DH (Dbl Homology) domain created a dominant-negative construct (PblΔDH), which induced rough eyes when expressed in the eye under the control of GAL4 . For our studies, we used a similarly sensitized background, but we did this in the context of wing development, where we could show that the phenotype was associated with a cytokinesis defect. The wing-specific driver MS1096-Gal4  in conjunction with UAS-PblΔDH produced a “multiple wing hair” phenotype (Figures 5A—5E). The severity of the phenotype, which was scored according to cell density, frequency of cells with multiple wing hairs, and the multiplicity of hairs, increased with temperature consistent with temperature influences on GAL4 (Figures 5B—5D) and with the effective dose of the X chromosome driver, which is higher in hemizygous males than in heterozygous females (our unpublished data).
To test whether wing expression of UAS-PblΔDH produced a defect in cell division, we stained pupal wings with Hoechst just before eclosion of the adults. In control wings, each hair was associated with a single nucleus, consistent with studies showing that each cell initiates one, and only one, hair (Figure 5F) . In contrast to the orderly array of nuclei in the control, PblΔDH-expressing wings showed clusters of nuclei (Figure 5G). These nuclei were heterogeneous in size, consistent with descriptions of nuclear fusion after cytokinesis failures in pbl mutants. The wing territories in which we found clusters of nuclei corresponded to the territories showing multiple wing hairs, and tufts composed of many hairs were associated with clusters having more or bigger nuclei. We assessed the time of action of PblΔDH by using temperature shift experiments (from 18°C to 30°C). No enhancement of the phenotype was observed when the shift was done after the last pupal wing cell divisions (our unpublished data). We conclude that cytokinesis failure occurs in wings expressing PblΔDH, and we used the reduced cell density and the multiple wing hair phenotype as a readout of cytokinesis failures.
We next tested whether the severity of the wing phenotype is sensitive to the dose of endogenous genes known or thought to function in the Pebble pathway. Removal of one endogenous copy of pbl, RhoA, or zipper (gene encoding a myosin II heavy chain) strongly enhanced the PblΔDH phenotype (Figures 5H—5J, compared to to5B).5B). We used an overproduction strategy to test for modification by the Drosophila Rho-Associated Kinase (Drok), a likely target of RhoA action . Expression of a catalytically activated, truncated form of Drok (Drok-CAT, ), strongly suppressed the multiple wing hair and the cell density PblΔDH phenotypes in one third of the flies (Figure 5K, compared to to5B).5B). In the other flies, the Drok-CAT transgene appeared to have hyper-activated the Pebble pathway; it mimicked overexpression of a wild-type pbl transgene (reduced and folded wings; our unpublished data)—as expected for a positive factor in the Pebble pathway. Altogether, these genetic interactions are consistent with the notion that Pebble acts in a cascade involving RhoA, Drok, and myosin II to promote cytokinesis and to prevent the development of multiple wing hairs (see the Discussion).
Although pleiotropic regulators such as cyclin B and cyclin B3 are likely to have multiple inputs into cytokinesis, we can probe their involvement in regulating a specific pathway by partially crippling that pathway and thereby making the system uniquely sensitive to further changes in the activity of the pathway. Because our studies of cytokinesis in the embryo suggested a synergy in the inhibition of cytokinesis by reduction in pbl function and expression of stable cyclin B3, we posited that cyclin B3 inhibits pbl function. To test this further, we asked whether reduction of cyclin B3 function would modify the wing phenotype due to PblΔDH expression. Because the cyclin B3 mutant is zygotically viable, we were able to test cycB3 homozygotes as well as heterozygotes. The removal of one or two copies of cyclin B3, a proposed inhibitor of pbl, ought to improve pbl function and hence suppress the actions of the dominant-negative construct. We compared the severity of the PblΔDH phenotype in flies having two, one, or zero functional copies of cyclin B3 by counting the frequency of cells with single hairs and multiple hairs. Both criteria showed a dose-dependent suppression of PblΔDH by reduction of cyclin B3 (Figures 6A—6C). Similar analysis using a cyclin B mutation showed that it also acts as a dose-dependent suppressor of PblΔDH and that the strength of suppression is similar to that of cyclin B3 (Figures 6D—F). Mutation in cyclin A did not modify this phenotype, suggesting that suppression of PblΔDH is a specific feature of cyclin B and B3 mutations (our unpublished data). We conclude from these data that both cyclin B and cyclin B3 genetically inhibit the contribution of the Pebble pathway to cytokinesis.
Successful segregation of chromosomes to the spindle poles is futile unless cytokinesis acts at the right instant and position to reliably trap full genetic complements in each daughter cell. The onset of anaphase is triggered by targeted protein degradation, which is initiated upon activation of the APC, but it is not known to what extent the subsequent events run free according to their own internal schedule, progress under constraints that couple different events, or are timed by independent inputs from central regulatory circuits. While several post-anaphase events are blocked by a nondegradable form of cyclin B [1, 2, 6, 7, 9-17, 33], it was unclear whether cyclin B destruction, which occurs at the transition to anaphase, has a timing input into the later event of cytokinesis or whether it is simply a necessary precondition. Here, we probe the regulatory coupling between cyclin destruction and cytokinesis in Drosophila. We detail the timing of the processes, demonstrate that destruction of cyclin B and cyclin B3 influences the timing of cytokinesis, and provide evidence that one pathway of cyclin action involves inhibition of Pebble contributions to the initiation of cytokinesis.
Signals from the spindle contribute to positioning of the cytokinesis plane [27, 34]. In order to guide the process, these signals must precede cytokinesis and ought to interface with signals timing the initiation of cytokinesis. Whether the spindle signals emanate from the astral microtubules, the central spindle, or both is still a matter of debate [34, 35]. Drosophila embryonic spindles develop peripheral bundles of microtubules that extend during early anaphase to reach the cortex at the future furrow position (Figure 1A’, cell 2). The proximity to the site of furrowing might make these fibers particularly efficient in delivering a spindle signal to initiate cytokinesis . Other spindle changes, such as the maturation of the midbody, occur in concert with the onset and progression of cytokinesis.
Observations showing that cyclin stabilization can block cytokinesis drew our attention to a possible role of cyclin destruction in timing cytokinesis. Importantly, cyclin destruction ordinarily precedes cytokinesis. The onset of cyclin B destruction precedes the onset of furrowing by about 60 s, and immunohistochemical staining of cyclin B disappears prior to cytokinesis initiation [11, 12, 22, 24]. Cyclin B3 destruction occurs slightly after cyclin B, but it is also over prior to the onset of cytokinesis .
We used three complementary approaches to test the influence of cyclin B and cyclin B3 destruction on the timing of cytokinesis. In the most direct of these, we demonstrated that genetic elimination of cyclin B or cyclin B3 led to cytokinesis furrow initiation at an earlier stage of mitosis (Figure 2). In a second genetic approach, we partially crippled a pathway activating cytokinesis and showed that removing doses of cyclin B or cyclin B3 alleviated the phenotype as if each encoded cytokinesis inhibitors of similar potency (Figure 6). In the third approach, we examined the consequences of persistence of function. As previously reported, stabilized cyclin B but not stabilized cyclin B3 blocked cytokinesis . However, we document that stable cyclin B3 retards the onset and slows the progression of cytokinesis; these results agree with the mutant analysis in showing that both cyclin B and cyclin B3 suppress precocious cytokinesis (Figure 3).
Inhibition of cytokinesis by cyclin B has been observed in numerous systems. In contrast, there are few studies of cyclin B3 other than those in Drosophila, and the continued, albeit slowed, cytokinesis during a mitotic arrest has not been reported in other systems. Even though the detailed isotype specificities of the Drosophila cyclin B3 might differ in other phyla, the conservation of the A, B, and B3 classes of cyclin and their relative timing of degradation  leads us to suspect that our findings will, at least in part, be generalizable.
Together, our results show that Drosophila cyclin B and cyclin B3 can inhibit cytokinesis, that their normal function defers cytokinesis, and that their prompt destruction is required for timely advance to cytokinesis. However, more detailed considerations are required to address whether this regulatory input from the cyclins reflects a direct timing control of cytokinesis or whether the actions of cyclins on cytokinesis might be mediated indirectly via the inhibition of a distinct mitotic event that is a necessary prerequisite for cytokinesis.
Stabilization of cyclin B inhibits events that precede the onset of cytokinesis. For example, stable cyclin B inhibits the transition to anaphase B  and blocks the dissociation of passenger proteins from centromeres [14, 36, 37], events thought to be required for cytokinesis. Consequently, indirect routes of inhibition of cytokinesis are likely to contribute to the very effective block induced by stable cyclin B. We propose that stable cyclin B has both direct and indirect actions that inhibit cytokinesis (Figure 7).
Mitosis progresses normally in the presence of stable cyclin B3 up to the time at which cytokinesis would usually initiate in mid-anaphase B. Stable cyclin B3 alters or slows some steps of spindle maturation that are coincident with cytokinesis. The peripheral microtubules that contact the cortex at the midzone do not develop fully, and the midbody does not develop the highly compacted structure that characterizes the late spindle in Drosophila. Although these effects of stable cyclin B3 might influence progression of already initiated cytokinesis, one of the earliest and perhaps most direct effects of cyclin B3 stabilization is to delay the onset of cytokinesis (Figure 7).
We were attracted to the pbl gene product as a candidate target for the regulation of cytokinesis because it is specifically required for cytokinesis [28, 29] and its action as a RhoA GEF implied early action in the pathways promoting actin assembly into the ring that presages the furrow [27, 38]. Thus, onset of cytokinesis might be triggered by activating Pebble. We showed, by using two different approaches, that mitotic cyclins influence Pebble action.
A cytokinesis furrow is still evident in embryos homozygous for pbl2, but it is delayed, slowed, and eventually aborted (Figure 4; Movies 4 and 5, available with this article online). Persistence of maternal Pebble suggested that residual furrowing was due to persisting Pebble function. Furrowing was also delayed and slowed upon the expression of stable cyclin B3 (Figure 3; also Movies 2 and 3, available with this article online). Importantly, furrows were suppressed when stable cyclin B3 was expressed in pbl2 (Figure 4). This indicates that both Pebble and cyclin impinge on early steps in cytokinesis.
We developed a genetic test to assess the activity of Pebble promotion of cytokinesis. Expression of a dominant-negative version of Pebble (PblΔDH) in wing discs partially suppressed Pebble function and led to cytokinesis failures, a reduction in cell density, and a multiple wing hair phenotype (Figure 5). The severity of this phenotype was very sensitive to small changes in the activity of the Pebble pathway; such changes include those caused by inactivation of just one of the two endogenous alleles of pbl or by inactivation of one allele of other genes in the Pebble pathway. Our findings support a pathway of sequential activation of RhoA, Rho-Associated Kinase (Drok), and myosin II leading to cytokinesis (Figure 7) [27, 38]. Like mutations in Pebble pathway functions, cyclin B and cyclin B3 mutations dominantly modified the PblΔDH-induced phenotypes (Figure 6). The cyclin mutations suppressed the phenotype, as expected for reduction of inhibitors. A similar study showed that a deficiency encompassing the cyclin B gene suppressed an eye phenotype induced by PblΔDH (R. Saint, personal communication). These findings show that cyclin B and cyclin B3 inhibit a step or pathway with regulatory connections to Pebble action on cytokinesis.
The genetic interaction experiments, although implying a regulatory connection between the cyclins and Pebble function, do not define how direct this connection might be. It had previously been suggested that cyclin/Cdk1 directly inhibits a downstream component, myosin II [39, 40]. However, tests of this idea have been negative [16, 41-43]. This led us to consider inhibition of the pathway at more upstream levels, perhaps at the level of Pebble itself.
Consistent with a limiting and regulatory role, Pebble appears to be absent or at low levels early during mitosis and to accumulate cortically and then concentrate in the region of the furrow at the time of cytokinesis . Furthermore, the level of Pebble early in mitosis is likely to be inadequate to promote cytokinesis given that RNAi treatment of embryos or mutation of pbl provoked cytokinesis defects while leaving detectable amounts of protein (Figure S2 and Movies 4 and 5, available with this article online). From this we suggest that the apparent rise in Pebble levels as anaphase progresses is significant and contributes to the initiation of cytokinesis. We found that expression of stable cyclin B or stable cyclin B3 suppressed the rise in Pebble immunostaining (our unpublished data). We propose that cyclin/Cdk regulation of Pebble accumulation controls the onset of cytokinesis, but we note that the pathway includes numerous additional regulation opportunities that might be used to create a robust program.
In addition to supporting the model that cyclin B3 degradation times the onset of cytokinesis, our results with stable cyclin B3 suggest that termination of the window of opportunity for cytokinesis is also influenced by cyclin B3. We found that both reduction of Pebble function and expression of stable cyclin B3 delayed and slowed cytokinesis, but the outcome differed. When Pebble function was reduced, furrow ingression terminated and the furrow receded as the cell exited mitosis with two nuclei. In contrast, cytokinesis was completed in the presence of stable cyclin B3 after a long slow course. Apparently, stable cyclin B3, while inhibiting the progression of cytokinesis, also promotes successful cytokinesis by maintaining the cell in a state conducive to cytokinesis. Perhaps this opposing positive effect of stable B3 is due to its ability to prolong the mitotic state and block nuclear membrane formation, which is usually followed by nuclear recruitment of Pebble away for its site of action at the furrow.
Our results show that inhibition of cytokinesis by cyclin B and cyclin B3 normally prevents precocious cytokinesis and that mitotic destruction of these cyclins contributes to timely cytokinesis. Although stable cyclin B inhibits cytokinesis by additional modes, both cyclins B and B3 inhibit activation of cytokinesis by the Pebble pathway.
We propose that an as-yet-undefined regulatory interface between cyclin activity and Pebble function connects the central cell cycle machinery to cytokinesis to provide temporal regulation. In addition, we suggest that cyclin B3 can delay a deadline for completing cytokinesis that might ordinarily be imposed by nuclear envelope reformation and nuclear recruitment of Pebble away from its site of action.
We used the wild-type strain Sevelen and the following mutant alleles: for cyclin B, cycB2 ; for cyclin B3, cycB32 ; for pebble, pbl2 ; for RhoA, RhoA72O ; and for zipper, zip1 . The G289 plasma membrane GFP fusion strain was described in . Stable cyclin B3 or Cdc2 with stable cyclins B were expressed under a heat shock promoter induced during G2 of embryonic cycle 14, as described in  together (or not) with histone H2-GFP (2AvD-GFP ). The transgene expressing the plasma membrane GFP fusion was combined with the transgene encoding the heat-shock-inducible stable cyclin B3, and the resulting strain was used to acquire the supplementary data in Movie 3. We used the driver MS1096  at the indicated temperature for wing expression of a truncated dominant-negative version of Pebble (UAS-PblΔDH497-5495B, UAS-PblΔDH497-5496A)  and a catalytically active form of the Drosophila Rho-Associated Kinase Drok (UAS-Drok-CAT) .
Embryos were fixed as in , devitellinized by hand, and stained as described in . The following primary antibodies/stains were used: anti-phospho-histone H3 (Upstate Biotechnologies, 1:800), anti-α-tubulin (Sigma, 1:400), anti-Pebble (, 1:800), phalloidin-Rhodamine (Molecular Probes, 1:1000), and Hoechst 33258 (Molecular Probes, 1 μg/ml). Secondary antibodies (except for the Donkey anti-Rat IgG (H+L) Cy3 conjugates [Jackson ImmunoResearch Laboratories, 1:600]), imaging, and videomicroscopy procedures were described in . For nuclear staining of wing cells, adult flies were removed from their pupal cases 1–2 hr before eclosion, and wings were fixed in 37% formaldehyde for 10 min, incubated with Hoechst 33258 (1 μg/ml) for 1 hr, and then quickly washed and mounted for imaging.
Right wings from adult flies of the appropriate genotype were removed and dehydrated in 100% ethanol for 1 hr and mounted in ethanol:glycerol 1:1. Pictures were taken from the dorsal surface between wing veins III and IV in a 166 μm × 140 μm rectangle centered above the posterior crossvein. For quantification, at least 15 independent right wings of each genotype were analyzed.
dsRNAs covering the first 714 base pairs of the Pebble coding sequence  were polymerized by the T7 RNA polymerase with the Ribomax kit (Promega). We used a PCR product flanked with two T7 promoter sequences (written in uppercase, see below) as a template for in vitro transcription. This PCR product was amplified from pBlueScriptSKII-Pblwtcl1a  with the following primers: TAATACGACTCACTATAGGGAGAatggaaatggagaccattgaagag and TAATACGACTCACTATAGGGAGAgcaaacgcctccgttttctaaaag. Pebble dsRNAs were annealed during cooling at room temperature after denaturation for 1 min at 94°C and adjusted to 2 mg/ml in injection buffer (5 mM KCl, 0.1 mM Na Phosphate buffer [pH 7.8]). Control dsRNAs were made for part of the coding sequence of the bacterial LacI from a PCR product amplified with the following primers: TAATACGACTCACTATAGGGAGAtctgaccagacacccatcaac and TAATAC GACTCACTATAGGGAGAgctttccagtcgggaaacctg. Twenty-minute Sevelen or G289 embryo collections were injected with control or Pebble dsRNAs immediately after egg deposition, at the posterior or the anterior pole, by standard injection protocols, and processed for videomicroscopy as described in .
Figure S1. Endogenous Cyclin B and Cyclin B3 Influence the Onset of Cytokinesis
(A-C) As in panels (A)—(C) of Figure 2 in the main text, these panels show mitosis 16 in wild-type, cyclin B mutant, and cyclin B3 mutant embryos, respectively. Phalloidin staining is shown in red, tubulin in green, and DNA in blue, and these are repeated in isolated grayscale images in the prime, double-prime, and triple-prime panels, respectively. The scale bar, 2.5 μm.
Figure S2. Pebble Levels in Cycle 15 Cells from Wild-Type and pbl2 Mutant Embryos or Wild-Type Embryos Injected with Pebble dsRNAs
(A-C) Pebble levels detected by indirect immunofluorescence on cycle 15 cells from wild-type embryos (A), homozygous pbl2 mutant embryos (B), or wild-type embryos injected with Pebble dsRNAs after egg deposition (C). (D) We assayed background immunofluorescence by treating wild-type embryos as in (A) but without Pebble primary antibodies. Images of (A)—(D) embryos were taken with the same settings and processed in parallel. The scale bar represents 5 μm.
Both RNA interference and the pbl2 mutation reduced, but did not completely abolish, the Pebble staining in cycle 15 embryonic cells. Because the allele truncates the protein before the epitopes recognized by the Pebble antibodies, we conclude that only partial reduction of Pebble function is sufficient to induce cytokinesis defects.
We thank H. Bellen, W. Chia, L.Luo, and R. Saint for sharing reagents and for personal communications, and D. Parry and G. Hickson for critical reading of the manuscript. A.E. was supported by the European Molecular Biology Organization and by the Human Frontier Science Program Organization long-term fellowships. This work was funded by National Institutes of Health grant #GM37193 to P.H.O’F.
Supplementary Material Two additional figures as well as five movies are available with this article online at http://images.cellpress.com/supmat/supmatin.htm.