In this report, we have examined the regulation of the mitotic inducer Cdc25C by 14-3-3 proteins during the cell cycle in Xenopus egg extracts. We have observed that the inactive hypophosphorylated form of Cdc25 present during interphase is quantitatively associated with two 14-3-3 proteins (p28 and p31). p31 is most similar to the ε form of 14-3-3 and appears to be the major partner of Cdc25: approximately 86% of Cdc25 is bound to 14-3-3ε during interphase. The other binding partner (p28) most closely resembles the 14-3-3ζ protein. It appears that most, if not all, of the inactive Cdc25 in Xenopus extracts is bound to either the ε or ζ form of 14-3-3. Furthermore, both the ε and ζ homodimers would each be present in an approximately fivefold molar excess over Cdc25 in egg extracts. Thus, the abundance of 14-3-3 proteins and the stoichiometry of their interaction with Cdc25 could account for the suppression of endogenous Cdc25 in Xenopus egg extracts during interphase.
A variety of observations strongly suggest that these 14-3-3 proteins act as negative regulators of Cdc25. First, the binding of 14-3-3 proteins to Cdc25 is highly regulated during the cell cycle. 14-3-3 proteins bind only to the interphase form of Cdc25 that displays weak activity toward the Cdc2–cyclin B complex. In contrast, the mitotic form of Cdc25 that can efficiently dephosphorylate Cdc2–cyclin B contains no detectable 14-3-3 protein. Another argument that 14-3-3 proteins negatively regulate Cdc25 is that a mutant of Cdc25 containing a single amino acid change that completely abrogates 14-3-3 binding shows a strongly enhanced ability to induce mitosis. This Cdc25-S287A mutant can compromise the checkpoints involving unreplicated and damaged DNA. In the case of the replication checkpoint, Xenopus
egg extracts containing unreplicated DNA and the Cdc25-S287A mutant enter mitosis efficiently even though DNA synthesis has been completely abolished by the treatment with aphidicolin. Significantly, in such extracts, mitosis takes place at about 90 min after activation of the extract, which corresponds to the time that extracts lacking aphidicolin would normally enter mitosis. It appears that the presence of the Cdc25-S287A mutant largely abolishes the responsiveness of Xenopus
egg extracts to unreplicated/damaged DNA. Recently, Peng et al. (1997)
have shown that a mutant of human Cdc25C (S216A) with a defective 14-3-3 binding site overrides G2
checkpoint controls in human cells. Thus, the regulation of Cdc25 by binding of 14-3-3 proteins appears to be a conserved mechanism of checkpoint control in vertebrates.
The molecular mechanism by which 14-3-3 proteins suppress the action of Cdc25 remains to be established. Because 14-3-3 proteins are known to form dimers (Aitken, 1996
), it is plausible that the binding of 14-3-3 may serve to oligomerize Cdc25 in egg extracts during interphase. As described herein, the binding of 14-3-3 appears to result in only a modest suppression (less than twofold) of the ability of Cdc25 to dephosphorylate a recombinant Cdc2–cyclin B complex, but we cannot be certain that these in vitro assays faithfully recapitulate the conditions found in vivo. Notwithstanding this caveat, it is conceivable that such a small effect on Cdc25 activity could tip the balance between the competing actions of Cdc25 and Wee1/Myt1 so that the Tyr-15 and Thr-14 dephosphorylation of Cdc2 could not proceed as long as 14-3-3 proteins remain bound, but other possibilities must also be considered. For example, the binding of 14-3-3 could preclude the interaction of Cdc25 with positive regulators. At least two kinases, including Cdc2–cyclin B and Plx1, phosphorylate Cdc25 in its N-terminal regulatory domain and stimulate its phosphatase activity at mitosis (Izumi and Maller, 1995
; Kumagai and Dunphy, 1996
). Perhaps binding of 14-3-3 could hinder the ability of these kinases to carry out their stimulatory phosphorylations. Alternatively, 14-3-3 proteins could enhance the ability of Cdc25 to interact with negative regulators such as the PP2A-like, Cdc25-inhibitory phosphatase (Kumagai and Dunphy, 1992
; Clarke et al., 1993
Another type of explanation for the function of 14-3-3 would be that these proteins might preclude the ability of Cdc25 to interact physically with the Cdc2–cyclin B complex. For example, 14-3-3 proteins could directly affect the recognition of Cdc2–cyclin B by Cdc25 or could indirectly prevent this interaction by keeping Cdc25 at an intracellular location where it would not have access to Cdc2–cyclin B. Xenopus
Cdc25 contains an excellent putative bipartite nuclear localization signal with the sequence KR
(the two basic clusters are underlined). Intriguingly, this putative nuclear localization sequence resides at amino acid residues 298 to 315 in Cdc25 and thus lies in close proximity to the 14-3-3 binding site around Ser-287, raising the possibility that binding of 14-3-3 could influence the intracellular localization of Cdc25. The localization of Cdc25C varies somewhat depending on the cell type. In human and fission yeast cells, Cdc25C is a nuclear protein during G2
-phase (Millar et al., 1991
; Girard et al., 1992
). In hamster cells, Cdc25C is cytoplasmic during G2
and enters the nucleus at about the beginning of mitosis (Seki et al., 1992
). At this time, it is not known whether the intracellular localization of Cdc25C plays a causal role in mitotic entry.
Recent studies have implicated Chk1 as the kinase that phosphorylates Ser-216 in the 14-3-3 binding site of human Cdc25C (Peng et al., 1997
; Sanchez et al., 1997
). A Xenopus
homologue of Chk1 has not been described, but clearly it will be valuable to ask whether a putative Chk1 homologue can phosphorylate Ser-287 of Xenopus
Cdc25 to allow the binding of 14-3-3ε and 14-3-3ζ. Interestingly, the binding of 14-3-3ε and -ζ to Xenopus
Cdc25 is similar whether or not the replication/damage checkpoint has been activated. This observation suggests that a kinase that phosphorylates Ser-287 is active in the absence of a checkpoint-triggered delay of mitosis. Thus, the function of Chk1 could be to maintain this critical phosphorylation of Cdc25 past the time at which mitosis would normally occur in an extract lacking unreplicated or damaged DNA.
In previous studies, we have analyzed the activities of the various enzymes controlling the tyrosine phosphorylation of Cdc2 in the absence and presence of unreplicated DNA (Kumagai and Dunphy, 1992
; Mueller et al., 1995a
). These studies have indicated that both Wee1 and Myt1 are highly active during interphase and that their kinase activities toward Cdc2–cyclin B as measured in vitro are not detectably altered by the presence of unreplicated DNA. In the case of Cdc25, its phosphatase activity is maintained in the same inactive state in the presence and absence of unreplicated DNA. It is apparent that this inactive form of Cdc25 is complexed with 14-3-3 proteins and that imposition of the replication checkpoint would keep Cdc25 in this state. As a consequence, the dephosphorylation of Tyr-15 and Thr-14 on Cdc2 would be precluded. A similar mechanism appears to account for the interphase arrest of egg extracts containing UV-damaged DNA. According to this scheme, the inhibitory phosphorylation of Cdc2 would be the ultimate target of the unreplicated and damaged DNA checkpoints in Xenopus
egg extracts, as is the case in fission yeast and humans (Enoch and Nurse, 1990
; Lundgren et al., 1991
; Jin et al., 1996
; Blasina et al., 1997
; Furnari et al., 1997
; O’Connell et al., 1997
; Peng et al., 1997
; Rhind et al., 1997
; Sanchez et al., 1997
In conclusion, we have found that 14-3-3 proteins negatively regulate the ability of Cdc25 to induce mitosis in Xenopus egg extracts. When this negative regulatory system is compromised by a mutation that prevents Cdc25 from interacting with 14-3-3, the unreplicated and damaged DNA checkpoints are unable to operate properly. In the future, it will be important to elucidate the molecular mechanisms controlling the association of 14-3-3 proteins with Cdc25 and how these regulatory mechanisms are modulated by unreplicated and damaged DNA.