Role for cyclin-dependent kinase 2 in mitosis exit

Mitosis requires cyclin-dependent kinase (cdk) 1-cyclin B activity [1]. Exit from mitosis depends on the inactivation of the complex by the degradation of cyclin B [2]. Cdk2 is also active during mitosis [3, 4]. In Xenopus egg extracts, cdk2 is primarily in complex with cyclin E, which is stable [5]. At the end of mitosis, downregulation of cdk2-cyclin E activity is accompanied by inhibitory phosphorylation of cdk2 [6]. Here, we show that cdk2-cyclin E activity maintains cdk1-cyclin B during mitosis. At mitosis exit, cdk2 is inactivated prior to cdk1. The loss of cdk2 activity follows and depends upon an increase in protein kinase A (PKA) activity. Prematurely inactivating cdk2 advances the time of cyclin B degradation and cdk1 inactivation. Blocking PKA, instead, stabilizes cdk2 activity and inhibits cyclin B degradation and cdk1 inactivation. The stabilization of cdk1-cyclin B is also induced by a mutant cdk2-cyclin E complex that is resistant to inhibitory phosphorylation. P21-Cip1, which inhibits both wild-type and mutant cdk2-cyclin E, reverses mitotic arrest under either condition. Our findings indicate that the proteolysis-independent downregulation of cdk2 activity at the end of mitosis depends on PKA and is required to activate the proteolysis cascade that leads to mitosis exit.