Though normally cytoplasmic (left), cyclin B (red) targeted to the nucleus (right) can start S phase.


In mammalian cells the work of the cell cycle is divided between two workhorses: Cdk2/cyclin E for S phase and Cdk1/cyclin B for mitosis. Now, Jonathan Moore, Jane Kirk, and Tim Hunt (Cancer Research UK London Research Institute, London, UK) show that this apparent specificity is achieved by limiting access to substrates. By denying entrance to the nucleus, cells prevent Cdk1–cyclin B from jumpstarting S phase at inopportune times.

In frog egg extracts, S phase is induced in nuclei by Cdk2–cyclin E, but not by Cdk1–cyclin B. This difference has often been construed as specificity in cyclin substrate preferences, but the new results show that cyclin/Cdk pairs are in fact surprisingly promiscuous enzymes. Hunt's group simply replaced the nuclear export signal from cyclin B with a nuclear localization signal and found that this altered Cdk1–cyclin B promoted both DNA replication and mitosis.“I'm a biochemist,” says Hunt. “I tend to think in terms of specificity between substrate and enzyme. So it was a shock to find the different [CDKs] might not discriminate their substrates.” Vertebrate cells apparently avoid the danger that cyclin B might initiate S phase via its cytoplasmic localization and low levels during G1. In yeast, a single cyclin can, under some circumstances, promote both S phase and mitosis. Yeast may not have as severe a need to restrict cyclin activities because the G2 to M transition is less clear than in higher eukaryotes. “Maybe it's okay to start mitosis early in yeast because chromosomes are still able to replicate as they are set on the metaphase plate,” Hunt suggests. ▪


Moore, J., et al.