|Home | About | Journals | Submit | Contact Us | Français|
In budding yeast, the protein phosphatase Cdc14 coordinates late mitotic events and triggers exit from mitosis. During early anaphase, Cdc14 is activated by the FEAR network, but how signaling through the FEAR network occurs is poorly understood.
We find that the FEAR network component Spo12 is phosphorylated on S118. This phosphorylation is essential for Spo12 function and is restricted to early anaphase, when the FEAR network is active. The anaphase specific phosphorylation of Spo12 requires mitotic CDKs and depends on the FEAR network components Separase and Slk19. Furthermore, we find that CDC14 is required to maintain Spo12 in the dephosphorylated state prior to anaphase.
Our results show that anaphase-specific phosphorylation of Spo12 is essential for FEAR network function and raise the interesting possibility that Cdc14 itself helps to prevent the FEAR network from being prematurely activated.
Exit from mitosis requires the inactivation of mitotic cyclin-dependent kinases (CDKs). In budding yeast, the conserved protein phosphatase Cdc14 induces CDK inactivation by targeting mitotic Clb cyclins for degradation . Throughout most of the cell cycle, Cdc14 is bound to its inhibitor Cfi1/Net1 in the nucleolus . During anaphase, two signaling networks induce the release of Cdc14 from Cfi1/Net1 in the nucleolus allowing Cdc14 to dephosphorylate substrates in the nucleus and cytoplasm. These are the Cdc Fourteen Early Anaphase Release (FEAR) network and the Mitotic Exit Network (MEN) (reviewed in ). The FEAR network promotes the release of Cdc14 from Cfi1/Net1 during early anaphase, whereas the MEN functions later during anaphase and maintains Cdc14 in its released state.
Cdc14 released by the FEAR network contributes to spindle stability, allows for segregation of the rDNA array, and activates the MEN [4–8]. These functions of Cdc14 promote cell survival, as 20% of FEAR network mutant cells lose viability during anaphase . The pathway is comprised of Separase, Esp1 in budding yeast, the protease that triggers chromosome segregation; Slk19, an Esp1-associated factor; Cdc5, the Polo-like kinase; and Spo12, a protein of unknown function. How the FEAR network promotes the transient release of Cdc14 from the nucleolus during early anaphase is only beginning to be understood. At the metaphase – anaphase transition, Separase/Esp1 becomes activated and functions together with Slk19 to down-regulate the protein phosphatase PP2A associated with its targeting subunit Cdc55 [9, 10]. This allows Clb-CDKs to phosphorylate Cfi1/Net1, which is necessary for Cdc14 release from the inhibitor [9, 11]. However, Cfi1/Net1 phosphorylation is not sufficient to maintain Cdc14 in the released state because in MEN mutants although Cfi1/Net1 remains phosphorylated, Cdc14 returns into the nucleolus [2, 11–13]. How CDC5 and SPO12 function within the FEAR network is not clear, but epistasis analysis placed SPO12 and its homolog BNS1 parallel to ESP1 and SLK19, and CDC5 downstream of, or in parallel to, the protease and its binding partner .
We characterized the role of Spo12 in the FEAR network. We show that Spo12 is phosphorylated on S118 and that this phosphorylation is restricted to anaphase. Therefore, this S118 phosphorylation is a marker that mirrors FEAR network activity. Additionally, this phosphorylation event is mediated by mitotic CDKs and depends on ESP1 and SLK19, which places SPO12 downstream of the Esp1-Slk19 complex in the FEAR network. We also show that CDC14 is needed to maintain Spo12 in the dephosphorylated state before anaphase. Our results not only shed light onto how signaling through the FEAR network occurs, but also raise the interesting possibility that Cdc14 is active while sequestered in the nucleolus.
A previous study showed that Spo12 phosphorylation predominantly occurs on two conserved serine residues, serine 118 (S118) and serine 125 (S125) . To determine the functional importance of this phosphorylation, we replaced endogenous SPO12 with SPO12 phospho-mutants containing alanine substitutions for either serine (spo12-S118A or spo12-S125A) or both serines (spo12-SSAA), and determined that these mutations did not affect protein stability or localization (Figure 1A, data not shown). However, the spo12-S118A, spo12-S125A and spo12-SSAA mutants behaved like spo12Δ mutants under all conditions examined. First, spo12-SSAA cells displayed an average mitotic exit delay of 13 minutes (SD 5.7 minutes), which correlates with the 10–15 minute delay that has been previously observed in spo12Δ mutants (Figure 1B, data not shown, ). Second, like spo12Δ cells, spo12 phospho-mutants exhibited synthetic lethality when combined with a deletion of the MEN activator LTE1 (Figure 1C). Third, when induced to undergo meiosis, the spo12 phospho-mutant strains, like spo12Δ cells, formed only two spores instead of four (Figure 1D, ). Our results indicate that the phosphorylation of Spo12 on S118 and S125 is necessary for its function.
To determine whether phosphorylation of Spo12 was cell cycle regulated, antibodies were raised against peptides containing phosphorylated S118 or S125 (see Experimental Procedures). Only antibodies against the peptide containing phospho S118 (henceforth pS118 antibody) exhibited phospho-specificity in Western blots. The pS118 antibody recognized overexpressed wild-type Spo12 and Spo12-S125A (though with somewhat decreased efficiency than wild-type protein), but not Spo12-S118A (Figure 1E). The phosphospecificity of the antibody was further confirmed by treatment of Spo12 immunoprecipitates with alkaline phosphatase, after which the pS118 antibody no longer recognized Spo12 (Figure S1). Interestingly, phosphorylation on S118 was not constant throughout the cell cycle, but restricted to anaphase (Figure 1E, F).
We next examined S118 phosphorylation in cells expressing SPO12 from its native promoter. Spo12 protein levels fluctuate during the cell cycle and are maximal during anaphase (Figure 1A, ), making it difficult to discern whether the increase in pS118 signal in anaphase was due to increased phosphorylation or simply reflected higher protein levels. Therefore, we developed a method to probe more equal amounts of Spo12 for S118 phosphorylation across the cell cycle. We first immunoprecipitated Spo12–13Myc and determined the amount of Spo12 in each time point (Figure S2A, top panel). Based on these measurements, the amount of loaded Spo12-immunoprecipitate was adjusted so that equal amounts of Spo12 were present across time points in the experimental Western blot (Figure S2A, middle panel). This normalization of total Spo12 levels allowed us to examine Spo12 phosphorylation independently of Spo12 protein levels across the cell cycle. Although this strategy worked most of the time, in some experiments Spo12 levels were too low in early stages of the cell cycle and we were not able to compensate for these differences by adjusting the loading (see for example Figures 2C, ,3E,3E, ,5E).5E). Using this method, we found that S118 phosphorylation fluctuates during the cell cycle. S118 phosphorylation was absent during G1, S phase and metaphase, but high levels were present during early anaphase (Figure 1G, H). As cells exited mitosis, S118 phosphorylation declined again. Our results demonstrate that Spo12 phosphorylation on S118 is restricted to anaphase.
We next wished to identify the mechanisms that restrict S118 phosphorylation to anaphase. To determine whether the MEN regulates S118 phosphorylation, we arrested cdc15-2 cells in metaphase using the microtubule depolymerizing drug nocodazole at the permissive temperature (23°C) and released them from the block into the restrictive temperature (37°C). This allowed cells to progress synchronously into anaphase before arresting prior to exit from mitosis due to failure to release Cdc14 in late anaphase (Figure 2B). In wild-type cells, S118 phosphorylation appeared in anaphase and declined upon exit from mitosis (Figure 2A, B). In cdc15-2 cells, S118 phosphorylation also appeared in anaphase, but remained high in the late anaphase arrest (Figure 2A, ,3A).3A). Similar results were obtained with other MEN mutants (dbf2-2 and tem1-3) and in cells defective in the MEN and FEAR network component Cdc5 (Figure 2C, D, data not shown [13, 18–20]). Our results show that the MEN is not required for S118 phosphorylation, but that MEN activity is needed for the loss of phosphorylation of this residue.
To determine whether FEAR network components were needed for Spo12 phosphorylation during anaphase, we examined S118 phosphorylation in various FEAR network mutants. S118 phosphorylation was not observed in slk19Δ cdc15-2 or slk19Δ cells (Figures 3A, B, data not shown), indicating that SLK19 is required for S118 phosphorylation. To determine whether S118 phosphorylation required ESP1 function, we examined the consequences of inactivating the protease by using the esp1-1 allele. We deleted MAD1 in esp1-1 cdc15-2 cells to avoid possible effects of an active spindle assembly checkpoint (reviewed in ). In cdc15-2 mad1Δ cells, S118 phosphorylation occurred as cells entered anaphase (Figure 3C, D). Inactivation of ESP1 prevents anaphase spindle elongation because cells cannot resolve the linkages between sister chromatids (reviewed in ). In this genetic background, cells arrest with metaphase-like spindles, but other aspects of cell cycle progression continue to occur . S118 phosphorylation was barely detectable in cdc15-2 esp1-1 mad1Δ cells (Figure 3C, D). In contrast, neither deletion nor overexpression of FOB1 affected S118 phosphorylation (Figure S3A, B), which is consistent with previous observations demonstrating that FOB1 functions downstream of SPO12 in the FEAR network . Our data indicate that ESP1 and SLK19, but not FOB1 are required for S118 phosphorylation during anaphase.
Cdc55, a regulatory subunit of PP2A, antagonizes Clb-CDK dependent phosphorylation of Cfi1/Net1, thereby preventing the release of Cdc14 from the nucleolus during early anaphase . Esp1 and Slk19 activity in early anaphase is thought to antagonize Cdc55-PP2A, thereby allowing Clb-CDKs to phosphorylate Cfi1/Net1 specifically during anaphase [9, 11]. Having established that Esp1 and Slk19 regulate S118 phosphorylation of Spo12, and that this phosphorylation, like that of Cfi1/Net1, occurs only during anaphase, we were surprised to find that deletion of CDC55 did not affect either the extent or kinetics of S118 phosphorylation (data not shown). More importantly, inactivation of CDC55 did not restore S118 phosphorylation to esp1-1 cells (Figure 3E, F). Additionally, FEAR-network induced phosphorylation of Cfi1/Net1 by CDKs is not required for Spo12-S118 phosphorylation, as cells carrying a CFI1/NET1 allele in which the 6 Clb-CDK sites were changed to residues that cannot be phosphorylated (net1-6CDK; ) did not affect phosphorylation of S118 (data not shown). Our results indicate that CDC55 does not regulate S118 phosphorylation, nor is phosphorylation of Cfi1/Net1 by Clb-CDKs necessary for S118 phosphorylation.
Our previous studies suggested that SPO12 functions in a pathway parallel to ESP1 and SLK19 because the deletion of SPO12 and its homolog BNS1 (Bypasses Need for Spo12) enhanced the Cdc14 nucleolar release defect of esp1-1 and slk19Δ mutants . In this original epistasis analysis, we deleted BNS1 in addition to SPO12 because it is homologous to SPO12, and it had been shown to suppress the phenotypes associated with deleting SPO12 in meiosis . In contrast to our epistasis analysis, the finding that S118 phosphorylation is essential for SPO12 function and is dependent on ESP1 and SLK19 indicate that SPO12 functions downstream of ESP1 and SLK19. To address whether this discrepancy could be due to the fact that both BNS1 and SPO12 were deleted in the original epistasis analysis, we examined Cdc14 localization in slk19Δ, spo12Δ, and slk19Δ spo12Δ mutants in the presence of the intact BNS1 gene. This analysis revealed that inactivation of SPO12 only marginally, if at all, enhanced the Cdc14 release defect of slk19Δ mutants (Figure 4). Our results indicate that SPO12 functions downstream of SLK19 and that a parallel contribution of SPO12 is only observed when its homolog BNS1 is deleted.
Which protein kinase mediates S118 phosphorylation during anaphase? Mitotic-CDKs are good candidates, as they mediate phosphorylation of Cfi1/Net1 in the nucleolus specifically during early anaphase . Furthermore, S118 and S125 of Spo12 each contain the minimal CDK consensus phosphorylation motif of a proline in the +1 position. To examine the consequences of CDK inactivation on S118 phosphorylation, we employed the cdc28-as1 allele, which can be specifically inhibited with the adenosine analog 1-NM-PP1 . Using the temperature sensitive cdc14-3 allele, cells were arrested in anaphase when S118 phosphorylation levels are high. Arrested cells were then treated with the cdc28as1 inhibitor. Within 30 minutes of inhibitor addition, S118 phosphorylation was lost in cells carrying the cdc28-as1 allele (Figure 5A, B). In contrast, other protein kinases that regulate mitotic progression such as the MEN kinases Cdc15 and Dbf2, Cdc5, the Aurora B kinase Ipl1, and the MAP kinase Hog1 were not needed for S118 phosphorylation (Figure 2, data not shown). We conclude that, like for Cfi1/Net1, the anaphase specific phosphorylation of Spo12 depends on CDKs.
During exit from mitosis, S118 phosphorylation is lost. The APC/CCdh1 targets Spo12 for proteasomal degradation during exit from mitosis , raising the possibility that degradation of Spo12 eliminates S118-phosphorylated Spo12 from cells. If degradation was the sole mechanism for removing S118-phosphorylated Spo12, S118 phosphorylation should be maintained in cdh1Δ cells. However, this was not the case. Upon release from a cdc15-2 anaphase arrest, Spo12 phosphorylation was lost in the absence of CDH1, despite the persistence of the Spo12 protein (Figure 5C, D). Thus, the primary cause of the loss of S118 phosphorylation during mitotic exit is likely dephosphorylation.
Cdc14 activation and release occurs at the time of S118 dephosphorylation. We therefore examined whether CDC14 was required for S118 dephosphorylation. Wild-type and cdc14-3 cells were synchronized in G1 and then released into the cell cycle at the restrictive temperature. In wild-type cells, S118 phosphorylation was restricted to anaphase (Figure 5E, F). However, in cdc14-3 cells, S118 phosphorylation occurred prior to anaphase. In fact, phosphorylation was detected as early as 50 minutes after release, even before cells entered metaphase, and this phosphorylation remained high throughout the duration of the experiment (Figures 5E, F). Phosphorylation was not detected in cdc14-3 cells carrying the spo12-S118A allele (Figure S4A, B) indicating that it was indeed S118 phosphorylation that occurred early in cdc14-3 cells. The fact that Esp1 is not active prior to the metaphase – anaphase transition because it is bound to its inhibitor Pds1 , also suggests that ESP1 is not needed for S118 phosphorylation when CDC14 is inactive.
To determine whether CDC14 contributes to Spo12 dephosphorylation prior to anaphase, we inactivated CDC14 in cells that were arrested in G1 by pheromone treatment. Due to prolonged pheromone treatment, a subset of cdc14-3 cells, and to a lesser extent wild-type cells, escape from the G1 arrest. Under these conditions, phosphorylated S118 is observed in cdc14-3 cells (Figure 5G). This phosphorylation depends on CDC28, as inactivation of CDC28 results in a rapid loss of S118 phosphorylation (Figure 5H). Inactivation of CDC14 in nocodazole arrested cells also lead to S118 phosphorylation (data not shown). Cdc14 was also able to dephosphorylate S118 in vitro. Recombinant wild-type Cdc14, but not catalytically inactive phosphatase, dephosphorylated S118 when added to Spo12 immunoprecipitates (Figure S4C).
If Cdc14 antagonizes Spo12 phosphorylation, which is required to promote early anaphase release of Cdc14, CDC14 ought to inhibit its own release from the nucleolus. Consistent with this idea we find that two temperature sensitive proteins, cdc14-1 and cdc14-3 are released from the nucleolus as soon as the protein is inactivated by temperature shift (Figure 6A, B). Spo12 localization was not affected in these mutants (data not shown). A Cdc14 protein (cdc14C283S/R289A) that is catalytically inactive and fails to bind substrates does not localize to the nucleolus in any cell cycle stage either (Figure 6C). Together, our results indicate that CDC14 is required throughout the cell cycle to maintain S118 in the dephosphorylated state and may even directly dephosphorylate S118. Our data further imply that in the absence of CDC14 and CDC28, other phosphatases are able to dephosphorylate S118.
Our results implicate Cdc14 in keeping S118 in the dephosphorylated state until anaphase thereby helping to prevent its own nucleolar release. Prior to anaphase Cdc14 is sequestered in the nucleolus by Cfi1/Net1, a competitive inhibitor of Cdc14 in vitro (Ki =3 nM) . However, it is possible that a fraction of Cdc14 is highly mobile during these cell cycle stages and can shuttle in and out of the nucleolus. To test this possibility we compared the mobility of nucleolar localized Cdc14 with that of released Cdc14. We determined fluorescence recovery after photobleaching (FRAP) of Cdc14-GFP. After Cdc14-GFP was photobleached in half of the nucleolus, we did not detect any recovery of signal within the time of the experiment (>10 minutes; Figure 6A, B; n=4) indicating that Cdc14 was exceedingly immobile when located in the nucleolus. In contrast, Cdc14 was highly mobile when present on the spindle pole body, one place where the phosphatase localizes to after its release from the nucleolus. The mean half time of recovery of the Cdc14-GFP signal at the spindle pole body was 12.5 seconds (SD 6.37 sec) (Figure 6C, D; n=4). These results indicate that when Cdc14 is located in the nucleolus the protein is tightly anchored, arguing against the possibility that a mobile fraction of Cdc14 was responsible for dephosphorylating Spo12 during early stages of the cell cycle. Our results support a model in which nucleolar Cdc14 functions throughout most of the cell cycle to counteract CDK-mediated phosphorylation of S118, until the onset of anaphase, when the FEAR network either modifies Spo12 or Cdc14 or upregulates Clb-CDKs in such a way that Cdc14 can no longer effectively promote the dephosphorylation of Spo12.
Spo12 promotes the release of Cdc14 from the nucleolus in part by antagonizing the Cdc14 release inhibitor Fob1 and also by additional mechanisms (Figure S5, ). The C-terminus of Spo12 fulfills all of Spo12’s FEAR network functions and contains a conserved domain with two phosphorylation sites, serine 118 and serine 125 . Using a phospho-specific antibody directed against S118 on Spo12, we found that phosphorylation of this residue occurs during early anaphase, concomitant with the release of Cdc14 from the nucleolus mediated by the FEAR network, and that this phosphorylation is then lost as cells exit mitosis. This, together with the observation that S118 is essential for Spo12 function, indicates that S118 phosphorylation is required for mediating the FEAR network-dependent release of Cdc14 from the nucleolus. Therefore, S118 phosphorylation is the first modification on a FEAR network component that both mirrors FEAR network activity and is required for FEAR network function. It is not known whether S125 is phosphorylated in vivo. Given that the S125 site is located only seven residues away from S118, also resembles a CDK site, and is essential for Spo12 function, this is likely to be the case.
Whether or not S118 phosphorylation is sufficient to promote the transient release of Cdc14 from the nucleolus during early anaphase remains to be determined. We were unable to test this possibility because SPO12 mutants that mimic constitutive phosphorylation through mutation of either S118 or S125 to glutamic or aspartic acid led to a loss of SPO12 function (BNT, unpublished observations). However, we believe that Spo12 phosphorylation is not likely to be sufficient because CDKs also phosphorylate Cfi1/Net1, which is required for FEAR-network mediated release of Cdc14 from the nucleolus .
Several lines of evidence indicate that Clb-CDKs are responsible for phosphorylating Spo12 on S118. The residues surrounding S118 resemble a CDK consensus sequence and S118 phosphorylation in vivo depended on CDK activity. Currently it is unclear whether Clb-CDKs phosphorylate nuclear Spo12, which then translocates into the nucleolus to promote Cdc14 release, or whether Clb-CDKs phosphorylate Spo12 in both compartments. The pS118 antibody did not detect pS118-Spo12 in indirect immunofluorescence analyses.
CDK activity alone is not sufficient for S118 phosphorylation. Mitotic CDKs are activated in G2, yet S118 is not phosphorylated until anaphase entry. As Esp1 and Slk19 are important for S118 phosphorylation, these two proteins may be required by mitotic CDKs to phosphorylate S118 or could be preventing dephosphorylation. To do this, Esp1 and Slk19 could target Clb-CDKs to the nucleolus or transiently down-regulate Cdc14 activity. The observation that ESP1 and SLK19 are only needed for S118 phosphorylation when CDC14 is functional is consistent with both possibilities.
Precedent for anaphase specific phosphorylation of Cdc14 regulators by Clb-CDKs exists. Cfi1/Net1 was previously shown to be phosphorylated by CDKs on T212 only during anaphase and this also depends on SLK19 . Therefore, Spo12-S118 phosphorylation and Cfi1/Net1-T212 phosphorylation may be co-regulated. Yet, despite some similarities in their regulation, the mechanisms whereby mitotic CDK bring about these phosphorylation events only during anaphase appear to be different. In the case of Cfi1/Net1 phosphorylation, Esp1 and Slk19 are thought to down-regulate PP2A via Cdc55 at the onset of anaphase, thereby causing a transient net increase in Clb-CDK phosphorylation . However, deletion of CDC55 did not alter the phosphorylation pattern of S118 in Spo12, nor did it restore S118 phosphorylation in an esp1-1 mutant. Furthermore, deletion of CDC55 rescues the lethality of spo12Δ lte1Δ double mutants , indicating that CDC55 functions in parallel to SPO12. Thus, although Spo12 and Cfi1/Net1 are found together at the rDNA and are phosphorylated by the same kinase at the same time, their mode of regulation appears to differ, as one includes downregulation of Cdc55 while the other does not. It will be important to characterize this difference further.
Several lines of evidence indicate that SPO12 functions largely downstream of ESP1 and SLK19. Spo12-S118 phosphorylation depends on ESP1 and SLK19. Furthermore, overexpression of SPO12 can rescue the Cdc14 release defects of slk19Δ cells, whereas overexpression of ESP1 cannot rescue the Cdc14 release defects in either slk19Δ cells or spo12Δ cells . Finally, examination of Cdc14 release from the nucleolus in slk19Δ, spo12Δ, and slk19Δ spo12Δ cells revealed no significant additive effects on the release of Cdc14 from the nucleolus in the double mutant (Figure 4). Thus, SPO12 functions largely downstream of ESP1 and SLK19, although a minor SLK19 independent effect on Cdc14 release from the nucleolus exists that is shared with its homolog BNS1.
Esp1 and Slk19 do not solely function through Spo12 to bring about the release of Cdc14 from the nucleolus. Loss of either ESP1 or SLK19 function has a much more severe effect on Cdc14 nucleolar release than loss of SPO12 function [13, 14]. ESP1 and SLK19 also promote CDK-dependent phosphorylation of Cfi1/Net1 by antagonizing Cdc55 (Figure S6). Thus, Esp1, through Slk19, directs Clb-CDKs to phosphorylate multiple components of the RENT complex. This in turn, leads to the partial disassembly of the RENT complex and release of Cdc14 from its inhibitor (Figure S6). In addition, Cdc5 phosphorylation of one or more components of the FEAR network are necessary to bring about this event [14, 29].
Spo12 levels are low, but not absent, during G1 and S phase (Figure 1A, ). Our results show that CDC14 is required to maintain this pool of Spo12 in the dephosphorylated state. In the absence of CDC14, S118 phosphorylation occurs prematurely as soon as Clb-CDKs accumulate in cells. These results indicate that Cdc14 functions to prevent S118 phosphorylation in stages of the cell cycle when the phosphatase is localized in the nucleolus in wild-type cells. Given that Cdc14 is highly immobile when located in the nucleolus, the most likely explanation is that Cdc14 functions in the nucleolus to bring about dephosphorylation of S118. Cdc14 may dephosphorylate S118 directly (S118 can be dephosphorylated by Cdc14 in vitro), though we cannot exclude a more complicated scenario in which Cdc14 activates another phosphatase.
Irrespective of how Cdc14 promotes Spo12 dephosphorylation, the observation that CDC14 maintains S118 in the dephosphorylated state while bound to Cfi1/Net1 in the nucleolus, is at odds with the observation that Cfi1/Net1 functions as a competitive inhibitor of Cdc14 in vitro with high affinity . The recent discovery of Tof2, which directly binds to nucleolar Cdc14 and enhances its phosphatase activity in vitro, could explain this apparent discrepancy . Cdc14 and Tof2 preferentially associate with NTS1, whereas Cfi1/Net1 predominantly associates with NTS2 regions of the rDNA [15, 31, 32]. It is thus possible that a fraction of Cdc14, bound to Tof2 is actually active in the nucleolus. Besides dephosphorylating S118 within Spo12, Cdc14 may also maintain Cfi1/Net1 in the dephosphorylated state. Like S118 in Spo12, Cfi1/Net1 is highly phosphorylated when cells carrying a temperature sensitive cdc14-1 allele are grown at the restrictive temperature .
Our analysis of S118 phosphorylation revealed another interesting result. In MEN mutants, Cdc14 returns to the nucleolus after being briefly released, yet S118 phosphorylation is maintained in these mutants. Why is Cdc14 resequestered after FEAR network-mediated release no longer capable of promoting dephosphorylation of S118? It is unlikely that high mitotic CDK activity out-competes Cdc14 because in late anaphase mitotic CDK activity has declined by at least 50% from its maximal levels in metaphase . Could the localization of Spo12 and Cdc14 be changed in such a way that Cdc14 can no longer dephosphorylate S118? Spo12 localization is not altered in MEN mutants, as the protein is found in both the nucleus and nucleolus (BNT, unpublished observations). We also tested the possibility that Cdc14 localization within the rDNA repeat is changed, such that it may be prevented from dephosphorylating S118. However, chromatin immunoprecipitation analyses showed that the distribution of Cdc14 within NTS1 and NTS2 did not differ between wild-type and MEN mutants (Figure S7). Perhaps phosphorylated Spo12 is shielded from Cdc14 during anaphase or Cdc14 is prevented from dephosphorylating the protein. It is also possible that despite an overall decrease in Clb-CDK levels, kinase activity could remain high in the nucleolus. Addressing these possibilities will be essential if we are to understand the molecular mechanisms of Cdc14 control.
Finally, our results reveal a novel function of Cdc14 in controlling its own release from the nucleolus and promoting mitotic exit. Previous studies showed that during late anaphase Cdc14 promotes its own release by promoting MEN activity . Our studies indicate that during early anaphase Cdc14 inhibits its own release by antagonizing FEAR network activity. Why would Cdc14 first inhibit its mitotic exit-promoting function but later promote it? By inhibiting its own release during early stages of anaphase, Cdc14 could set a threshold for activation of its mitotic exit-promoting function and ensure that nucleolar release of the phosphatase only occurs when Clb-CDK levels are extremely high. This system would provide a simple mechanism to guarantee that initiating exit from mitosis only occurs once high Clb-CDK levels have been reached and hence cells have progressed well into mitosis. During later stages of anaphase, Cdc14 promotes its mitotic exit function by activating the MEN, a pathway that can promote the release of Cdc14 from the nucleolus in the absence of high Clb-CDK activity. In this way, Cdc14 ensures that it remains released in the face of declining Clb-CDK levels. Our analysis of Spo12 phosphorylation therefore, has not only revealed an unexpected function of Cdc14 outside of anaphase, but also points toward Cdc14 itself having a central role in controlling its own mitotic exit promoting activity.
All strains are isogenic with W303 (A2587) and are listed in Supplemental Table 1. Unless stated otherwise, gene deletions were constructed using the one-step PCR based gene replacement method . The Spo12 phospho-mutant strains were described previously . Yeast strain A21809 contains plasmid A1705 (pCDC14-C283S/R289A-1HA in YCPlac22 background) that was cloned by replacing the PflMI-AflII fragment of Cdc14 in plasmid A487 (pCDC14-1HA in YCPlac22 background) with the PflMI-AflII fragment from pYL55 (pCDC14-C283S/R289A), which was a gift from Harry Charbonneau.
Growth conditions for individual experiments are described within the Figure legends. When unspecified, cells were grown in yeast extract peptone (YEP) plus 2% glucose at 25°C. To obtain a G1 arrest, 5μg/ml alpha-factor was used as described previously . To prevent escape from the arrest, 2.5μg/ml alpha-factor was added to all cultures again 1.5 hours later. To arrest cells in metaphase, cells were first arrested in G1 using alpha-factor, then released into media containing nocodazole for 90 minutes at a final concentration of 15 μg/ml. After release into fresh medium containing 1% DMSO, intact spindles were not detected until approximately 30 minutes after release from the block. 1-NM-PP1 inhibitor was used at indicated concentrations to inhibit Cdc28-as1 . For the sporulation experiment, diploid cells were grown to saturation in YPD (YEP + 2% glucose) for 24 hours, diluted into YPA (YEP + 2% KAc) at OD600 = 0.3 and grown overnight. Cells were then washed with water and resuspended in SPO medium (0.3% KAc [pH = 7.0]) at OD600 = 1.9 at 25°C to induce sporulation.
Indirect in situ immunofluorescence on whole cells was carried out as previously described . Unless otherwise indicated, 100 cells were scored per strain per time-point. Rat anti-tubulin antibodies (Oxford Biotechnology) and anti-rat-FITC antibodies (Jackson Immunoresearch) were used at a 1:200 dilution. Other antibodies were used at the following concentrations: mouse anti-HA.11 (Covance) at 1:150, anti-mouse-Cy3 (Jackson Immunoresearch) at 1:1000, goat anti-Cdc14 (Santa Cruz) at 1:300, anti-goat-Cy3 (Jackson Immunoresearch) at 1:500.
Western blot analysis was performed as previously described . Mouse anti-Vph1 antibodies (Molecular Probes) were used at 1:2000, mouse anti-Pgk1 antibodies (molecular Probes) were used at 1:25000, mouse anti-Myc antibody (Covance) at a 1:1000 dilution, and anti-mouse horseradish peroxidase-conjugated antibodies (Amersham) were used at 1:2000. Anti-GST (Amersham) antibody was used at 1:2000 and donkey anti-goat-HRP (Santa Cruz) was used at 1:4000.
Antibodies were custom-made by Abgent Technologies against phospho-S118 Spo12 using peptide QLQQRFA(pS)PTDRLVSC. The anti-phospho S118 antibody was shown to be phospho-specific by Western blot (Figure 1 and S1, S2).
Spo12-13Myc was immunoprecipitated using a 1:2000 dilution of 9E10 anti-Myc antibody (Convance) as previously described . The amount of Spo12 in each sample was determined by probing a Western blot with anti-Myc-HRP antibody (Invitrogen) at 1:4000 for 2 hours. For measuring Spo12 phosphorylation, blots were blocked for 8 hours at room temperature in 5% BSA in TBS-T, then incubated with 1:500 of the rabbit phospho-S118 antibody in 1.5% BSA in TBS-T overnight at 4°C. Blots were washed five times with TBS-T and incubated with a goat anti-rabbit-HRP antibody (Jackson Immunoresearch) at a 1:2000 dilution for 1.5 hours at 25°C and then again washed five times with TBS-T. A non-specific band was seen in Western blots probed with the Spo12 phospho-antibody (noted with an asterisk when present). To determine total Spo12 protein levels, blots were blocked for 30 minutes in 3% milk in PBS-T and then incubated with anti-Myc-HRP antibody (Invitrogen) in 1% milk/1% BSA in PBS-T at a concentration of 1:4000 for 2 hours at room temperature.
FRAP experiments were conducted at 25°C. Details regarding instrumentation and data analysis were as described .
We are grateful to Pedro Carvalho and David Pellman for help with FRAP experiments. We also appreciate the generous gifts of a Cfi1/Net1-6Cdk mutant strain from Ray Deshaies and Cdc14 plasmids from Harry Charbonneau. Additionally, we would like to thank Frank Solomon, and Frank Stegmeier, and members of the Amon lab for critical reading of the manuscript, as well as Leon Chan and Elçin Ünal for technical assistance. This research was supported by a National Institute of Health grant GM 056800 to A.A., who is also an investigator of the Howard Hughes Medical Institute.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.