One of the first APC substrates to be identified was its own activator, Cdc20, hinting at the existence of autoregulation [21
]. Initial reports suggested that Cdc20 behaved similarly to other APC substrates, being targeted in part via a different activator (Cdh1) through Cdc20’s D-box [22
]. Interestingly, we show here that, unlike other APC substrates, Cdc20 is largely targeted for destruction by the APC through an autoubiquitination mechanism that occurs when Cdc20 is bound to the APC as an activator. Importantly, this mechanism appears to be regulated throughout the cell cycle, and may be influenced by the presence or absence of substrates.
The observation that Cdc20 turnover was only partially reduced in conditional APC mutants led some authors to speculate that the residual turnover observed might be mediated by an APC-independent pathway. Our work in a strain that permits the deletion of the APC11 gene shows that in unperturbed cells, Cdc20 is turned over solely by the APC. This discrepancy is likely due to the fact that conditional alleles may not be completely null for APC activity, whereas deletion of the gene encoding the catalytic subunit (APC11) eliminates activity completely.
has long been assumed to be the APC complex that targets Cdc20 for destruction (Cdh1trans
, ) [17
]. However, deletion of APC11
leads to much greater steady-state levels of Cdc20 than deletion of CDH1
, suggesting the existence of other APC-mediated mechanisms [25
]. This suggests two obvious models for turnover. First, Cdc20 bound to the APC as an activator could recognize another molecule of Cdc20 leading to ubiquitination of the substrate Cdc20 (Cdc20trans
, ). In this case, the substrate Cdc20 should behave similarly to other Cdc20 substrates. Alternatively, Cdc20 may bind to the APC as an activator and this binding alone may be sufficient for autoubiquitination (Cdc20cis
, ). To evaluate the relative contributions of the three possible modes of Cdc20 turnover, we generated strains in which only one mechanism of turnover was possible and performed in vitro
experiments with similar perturbations. These experiments strongly suggested that Cdc20cis
is the predominant form of Cdc20 turnover.
Previous work showed that Cdc20 not only recruits substrates to the APC, but also serves to activate the APC, since its presence was also required for the ubiquitination of the APC substrate Nek2A, which can bind the APC independently of an activator [9
]. Importantly, these results suggested that an N-terminal fragment of Cdc20 containing the C-box was sufficient to activate the APC toward Nek2A, and that the C-box was required for this activation [9
]. Interestingly, we find that a Cdc20 C-box mutant, which does not support viability and is unable to drive Dbf4 turnover in vivo [25
](), is still targeted for turnover by the APC, although its turnover is compromised. This result suggests that the C-box is not absolutely required for APC activity, but is specifically required for stimulating APC activity towards other APC substrates, potentially by properly orientating either the substrate and or the catalytic arm of the APC so substrate ubiquitination can occur. Interestingly, deletion of the C-terminal IR domain, which does not result in a growth defect, has a significant effect on Cdc20 turnover, slightly greater than the defect seen for the lethal C-box mutant. The IR domain has been shown to interact with Cdc27, the terminal subunit of the TPR arm [13
]. The non-essential nature of the IR-Cdc27 interaction could suggest that it is an intermediate in the reaction mechanism when Cdc20 is particularly susceptible to autoubiquitination. Consistent with this observation, this interaction is not required for the processive ubiquitination of other APC substrates [35
]. However, the lack of affinity provided by the Cdc27-IR interaction is compensated by an interaction between the activator, substrate, and Doc1 on the APC core. However, autoubiquitination does not use Doc1, possibly, making the affinity provided by the Cdc27-IR interaction more important.
The discovery that Cdc20 is targeted for turnover by Cdh1, which is itself cell-cycle regulated, suggested a mechanism by which Cdc20’s cyclical expression could be achieved. Work from the Cross lab [26
] and from experiments presented here suggests that Cdh1 may contribute to but is not necessary for Cdc20’s cell cycle oscillation. However, previous work [21
] suggested that oscillation in Cdc20 levels is also achieved by transcriptional regulation. CDC20
is a member of the CLB2
cluster of genes [36
], whose transcription is under the control of Fkh2/Ndd1 [37
]. The observation that Cdc20 levels still oscillate in cells that express CDC20
under a constitutive promoter (TEF1p
) in the absence of Cdh1 implies an additional cell cycle regulated mechanism. This is strong evidence that regulation of the Cdc20cis
mechanism we observe is sufficient to drive the oscillatory behavior of Cdc20 through out the cell cycle.
Previous work has shown that phosphorylation of the TPR subunits (Cdc27, Cdc16, and Cdc23) by the Cyclin-Dependent Kinase (CDK) increases the affinity of Cdc20 for the APC [39
]. It is possible that these phosphorylations are regulating the Cdc20cis
mechanism. However, these phosphorylations promote Cdc20 binding to the APC, and occur when CDK activity is high. If these phosphorylations are promoting the Cdc20cis
turnover during the cell cycle, we would expect to see the lowest Cdc20 levels when CDK activity is highest. However, we observe that the lowest Cdc20 levels occur during the G1 phase of the cell cycle, when CDK activity is lowest. Alternatively, phosphorylation of the TPR proteins may cause Cdc20 to bind in a slightly different position on the APC, which may inhibit the Cdc20cis
These data suggest the following model. The APC is hyperphosphorylated in early mitosis, which increases its affinity for Cdc20. As APCCdc20
runs out of substrates, Cdc20 begins to autoubiquitinate, constituting the majority of the late mitotic turnover. This model for the regulation of Cdc20 stability by the presence of substrates () is similar to that put forth for the ubiquitin conjugase Ube2C [40
]. As cells exit mitosis, APC becomes dephosphorylated and Cdh1 becomes active, thus removing residual Cdc20. Additionally, our model for substrate inhibition of Cdc20 turnover may explain why it is advantageous for the cell to have Cdc20 binding to the APC strongly enhanced by the presence of substrates [35
]. In this way, Cdc20 would be unlikely to be prematurely degraded when substrates are present.
Interestingly, Cdc20 turnover has been shown to increase in the presence of spindle poisons. This turnover is dependent on an intact Spindle Assembly Checkpoint (SAC) [28
]. The exact mechanism for this turnover is unknown, but it will be interesting to determine the mechanism for Cdc20 turnover during SAC activation.