Inhibitory phosphorylation of Cdk1 controls entry into mitosis in all eukaryotic cells, yet the signals that control addition and removal of inhibitory phosphorylation are poorly understood. Previous work found that Zds1/2 are required for normal control of Cdk1 inhibitory phosphorylation; however, it was unclear whether they work in signaling mechanisms that directly control Cdk1 inhibitory phosphorylation or whether their inactivation affects Cdk1 phosphorylation indirectly by causing a failure in an early cell cycle event that triggers a checkpoint response. We therefore examined the functions of Zds1/2 to more clearly define their role in entry into mitosis.
We discovered that Zds1 forms a tight stoichiometric complex with PP2ACdc55
via an association with the Cdc55 subunit. This discovery complements and extends previous work that detected an association between Zds1/2 and PP2ACdc55
in coimmunoprecipitation assays, and it suggests that the primary role of Zds1/2 is to mediate the function or regulation of PP2ACdc55
(Gavin et al.
; Collins et al.
; Queralt and Uhlmann, 2008
; Yasutis et al.
). Because PP2ACdc55
acts directly on Mih1, Zds1/2 is placed immediately upstream of Mih1 (Pal et al.
). Surprisingly, we found that Zds1/2 are required for PP2ACdc55
-dependent dephosphorylation of Mih1 but not for PP2ACdc55
-dependent dephosphorylation of Swe1, which demonstrates that they target PP2ACdc55
to Mih1. Zds1 undergoes complex cell cycle–dependent changes in phosphorylation, which suggests that Zds1/2 may be targets of upstream signals that control entry into mitosis. Together, these observations suggest a model in which Zds1/2 control Cdk1 inhibitory phosphorylation by relaying upstream signals that control entry into mitosis to Mih1 via PP2ACdc55
. In this model, inactivation of Zds1/2 causes a mitotic arrest because the signal to remove Cdk1 inhibitory phosphorylation is not properly relayed. Although we favor this model, we cannot rule out the possibility that inactivation of Zds1/2 triggers a checkpoint arrest by causing defects in early cell-cycle events that could not be detected by our assays. It is also possible that Zds1/2 regulate entry into mitosis via additional mechanisms that are independent of PP2ACdc55
Our understanding of the function of Zds1/2 is currently limited because we do not yet have comprehensive data on the functional consequences of dephosphorylation of Mih1 by PP2ACdc55
-Zds1/2. To rigorously determine the functional significance of Mih1 dephosphorylation it will be necessary to assay the activity of differently phosphorylated forms of Mih1 and to map and mutate the relevant phosphorylation sites, which has not yet been possible. Several observations suggest, however, that Yck1/2-dependent phosphorylation inhibits Mih1 activity and that dephosphorylation leads to increased Mih1 activity. First, Yck1/2-dependent hyperphosphorylation of Mih1 occurs during interphase, when Mih1 activity should be suppressed, and PP2ACdc55
-Zds1/2 dephosphorylates Mih1 during entry into mitosis, when Mih1 activity must be increased (Pal et al.
). Second, cdc55Δ
causes a failure in Mih1 dephosphorylation, increased Cdk1 inhibitory phosphorylation, and severe defects in entry into mitosis that are rescued by swe1Δ
(Minshull et al.
; Yang et al.
; Pal et al.
). Finally, zds1Δ zds2Δ
causes a complete failure in Mih1 dephosphorylation and severe defects in entry into mitosis, even though PP2ACdc55
retains activity against Swe1. The facts that Zds1/2 play a specific role in targeting PP2ACdc55
to Mih1 and that loss of Zds1/2 causes severe defects in entry into mitosis strengthen the evidence that removal of Yck1/2-dependent Mih1 phosphorylation promotes entry into mitosis. Inactivation of Yck1/2 causes a failure in Mih1 hyperphosphorylation but does not cause premature entry into mitosis, as would be predicted if Yck1/2 inhibited Mih1 (Pal et al.
). Preliminary experiments suggest, however, that Yck1/2 may also be required for inactivation of Swe1, which complicates interpretation of the phenotype caused by loss of Yck1/2 (S. Anastasia and D. Kellogg, unpublished data).
The hyperphosphorylation of Mih1 early in interphase contrasts with Xenopus Cdc25, which is not hyperphosphorylated during interphase. A possible explanation is that hyperphosphorylation of Mih1 early in the cell cycle reflects an added layer of checkpoint control that does not operate in Xenopus egg extracts. In budding yeast, Wee1 and Cdc25 respond to checkpoint signals that coordinate entry into mitosis with bud growth or morphogenesis, whereas in early Xenopus embryos, cell division occurs without accompanying growth or morphogenesis. Hyperphosphorylation of Mih1 early in the cell cycle may therefore inhibit Mih1 until bud growth or morphogenesis is complete. In this model, Zds1/2 would play a critical role in relaying checkpoint signals that trigger dephosphorylation of Mih1 when checkpoint conditions have been satisfied. Interestingly, inhibition of PP2A in interphase Xenopus extracts leads to extensive and quantitative hyperphosphorylation of Cdc25 in the absence of mitotic Cdk1 activity (Kumagai and Dunphy, 1992
; Izumi and Maller, 1995
). Although the identity of the kinase is unknown, this observation suggests that the potential for similar checkpoint signaling may exist in vertebrate cells and that the checkpoint is turned off in Xenopus oocytes.
Previous work reached the conclusion that Zds1/2 inhibit the activity of PP2ACdc55
against Net1 (Queralt and Uhlmann, 2008
). Interestingly, the behavior of Swe1 in cells that lack Zds1/2 is also consistent with an inhibitory role for Zds1/2. In this case, the delayed phosphorylation of Swe1 and accumulation of Swe1 in an intermediate phosphorylation form could be explained by a role for Zds1/2 in inhibiting the activity of PP2ACdc55
against Swe1. Because PP2ACdc55
opposes phosphorylation of Swe1 by Cdk1/Clb2, hyperactivity of PP2ACdc55
in the absence of Zds1/2 would lead to decreased phosphorylation of Swe1, which is consistent with the observed behavior of Swe1 in zds1Δ zds2Δ cdk1-Y19F
cells. These observations suggest the possibility that Zds1/2 are capable of playing both activating and inhibitory roles in the regulation of PP2ACdc55
. Cell cycle–dependent changes in phosphorylation of Zds1/2 could determine whether they play activating or inhibitory roles. An interesting possibility is that a change in the phosphorylation state of Zds1/2 during entry into mitosis leads to inhibition of PP2ACdc55
activity against Swe1 and to stimulation of PP2ACdc55
activity against Mih1. This would simultaneously promote full hyperphosphorylation and inactivation of Swe1, as well as dephosphorylation and activation of Mih1. In this model, mechanisms that regulate PP2ACdc55
could promote mitotic progression via coordinated inhibition of Swe1 and activation of Mih1. This could be achieved in a relatively simple manner. For example, a change in Zds1/2 phosphorylation could target PP2ACdc55
to Mih1, thereby pulling active PP2ACdc55
away from Swe1.
An intriguing finding was that Zds1 becomes quantitatively hyperphosphorylated when it is dissociated from the PP2ACdc55
complex. This finding suggests that Zds1 is continuously being dephosphorylated by associated PP2ACdc55
and that its phosphorylation state is determined in a highly dynamic manner by opposing kinase and phosphatase activities. Theoretical considerations have shown that a kinase and phosphatase acting on the same target protein can generate switchlike behavior (Goldbeter and Koshland, 1981
; Ferrell, 1996
). If the kinase and phosphatase operate at or near their maximal velocities, the target protein can exhibit large switchlike changes in its phosphorylation state in response to small changes in enzyme velocity. This behavior is referred to as “zero-order ultrasensitivity.” The fact that PP2ACdc55
is associated with Zds1/2 suggests that it is acting at its maximal velocity, which fulfills one requirement for a zero-order ultrasensitivity. Thus the PP2ACdc55
-Zds1/2 complex may contribute to switchlike activation of Cdk1 during entry into mitosis.
The discovery that Zds1 undergoes cell cycle–dependent changes in phosphorylation raises the possibility that regulation of PP2ACdc55
plays an important role in the mechanisms that trigger entry into mitosis. An interesting model is that phosphorylation of Zds1/2 by Cdk1/Cln2 early in the cell cycle inhibits their ability to target dephosphorylation of Mih1. Because Cdk1/Cln2 drives bud growth, this would ensure that entry into mitosis does not occur while bud growth is ongoing (McCusker et al.
). The fact that multiple signals feed into Zds1/2 suggests that they could be an integration point for checkpoint signals that report on the status of cell size or morphogenesis. Zds1 is localized to the growing bud, so it is well positioned to relay these kinds of signals (Bi and Pringle, 1996
). An important goal for future work will be to fully define the signals that control Zds1/2, as this will likely lead to a better understanding of the checkpoint signals that control entry into mitosis.