N&V Final.indd

The cell division cycle is a fundamental and highly complex process that lies at the core of every proliferating cell. Progression of the eukaryotic cell through four phases — G1, S, G2 and M — ensures that genetic information is first faithfully replicated (during S phase) and then equally distributed to two daughter cells (during M phase). The function of G1 and G2 is to prepare the cell for DNA replication and mitosis, respectively. A plethora of intraand extracellular cues impinges on the cell-cycle machinery, which integrates and translates this information into signals affecting cell proliferation, cell-cycle arrest, or differentiation. The eukaryotic cell cycle is controlled by cyclin-dependent kinases (CDKs) that must associate with a regulatory cyclin subunit to be active1. In lower organisms such as yeast, cell-cycle progression is controlled by only one CDK, called Cdk1 or Cdc2, in complex with different cyclins, whereas in mammals combinations of different CDKs and cyclins need to be sequentially activated to drive the cell cycle. The simplified but dogmatically accepted view has been that in mammalian cells, progression through G1 phase is driven by the activities of Cdk4 and Cdk6, which associate with D-type cyclins. Entry into S phase and initiation of DNA replication require the activity of Cdk2, which is successively activated by E-type cyclins in late G1 and in S phase, and by A-type cyclins in S and G2 phase. Cdk2 activity is antagonized by the action of p21 and p27, two CDK inhibitors that bind to both the CDK and cyclin moiety, and thereby interfere with the ability of the cyclin–CDK complex to phosphorylate target substrates. Finally, entry into M phase requires the activity of Cdk1 (Cdc2), which associates with both Aand B-type cyclins. The ingrained concept that Cdk1 controls the G2/M transition and initiation of mitosis, whereas Cdk2 regulates entry into S phase and the onset of DNA replication, was challenged two years ago when two independent groups reported the successful generation of Cdk2−/− mice2,3. Mice lacking Cdk2 were found to be viable, developed normally, and were only slightly smaller than wild-type mice. The main difference in phenotype that these mice showed was male and female sterility due to meiotic defects. Although synchronized Cdk2−/− mouse embryonic fibroblasts (MEFs) enter S phase with delayed kinetics compared with wildtype MEFs, no major aberrations were found in somatic cells. These results were surprising, as Cdk2 was thought to be indispensable for promoting the G1/S transition, particularly given its exclusivity as a CDK-binding partner for cyclins E1 and E2 (with the exception of Cdk3, which is not expressed in the mouse strains analysed). This conundrum gave rise to the question: which kinase compensates for the loss of Cdk2 activity? A report by Aleem et al.4 on page 831 of this issue shows convincingly that Cdk1 can functionally substitute for Cdk2 in a Cdk2−/− background. The authors initially set out to investigate the genetic interaction between Cdk2 and p27, because Cdk2 is generally considered to be a major target of p27. They assumed that phenotypic defects in p27−/− mice should, at least partially, be suppressed in p27−/− Cdk2−/− mice. However, the researchers found very similar defects in both p27−/− and double-knockout mice, such as pituitary intermediate lobe hyperplasia and adenomas, enlarged thymi due to increased thymocyte proliferation, female sterility (double-knockout mice exhibit sterility in both sexes), and increased body size (at least in males). Concomitant work by Martín et al.5 shows very similar results and corroborates the idea that Cdk2 may not be the only target of p27 and the closely related p21, as p21−/− Cdk2−/− mice were also analysed in this study. A hint as to the nature of the kinase came from the finding that loss of both p27 and Cdk2, or p27 alone, significantly increased the proportion Cdk1