Cyclin-dependent kinases (Cdks) control progression through the eukaryotic cell cycle. Cdks are serine and threonine kinases, and their actions are dependent on associations with their activating subunits, cyclins. Cyclin abundance is regulated by protein synthesis and degradation; the activity of Cdks is therefore regulated to a large degree by the presence of different cyclins. In the budding yeast Saccharomyces cerevisiae, a single Cdk, Cdc28 (which is equivalent to Cdk1 in other organisms), associates with multiple cyclins to regulate the cell cycle. By contrast, animal cells possess multiple cell-cycle-regulatory Cdks, which are each regulated by multiple cyclins. The rationale for possessing multiple cyclins is not fully understood, although the evolution of multiple cyclins has been partially characterized (BOX 1).Multiple cyclins are probably advantageous because they allow for flexible control of the cell cycle. Different cyclins are independently regulated transcriptionally and post-transcriptionally, providing regulatory flex-ibility at the level of input. Also, cyclins possess over-lapping, but distinct, functional activities, allowing further refinement of control and probably the timely and irreversible occurrence of cell-cycle transitions. In this review, we focus on the budding yeast cyclins that activate Cdc28 as a model for cyclin specificity. In some cases, this information will probably be directly transfer able to other eukaryotic systems thanks to the close conservation of the molecules involved; in other cases, the yeast system might provide general principles even when the details differ.Cyclin function during the cell cycleCyclin specificity can be deduced from a genetic requirement for a specific subset of cyclins for a cell-cycle event to occur (FIG. 1). For example, advancement through G1 phase of the cell cycle, which involves bud emergence, spindle pole body (SPB) duplication and the activation of subsequently expressed cyclins, requires at least one of the G1-phase cyclins — Cln1, Cln2 or Cln3. In the absence of CLN1–3, G1 arrest occurs. Following Cln function, efficient initiation of DNA replication and progression through S phase requires the early-expressed B-type cyclins Clb5 and Clb6. In their absence, the B-type cyclins Clb1–4 will drive a late initiation of DNA replication. The B-type cyclins Clb1–4 are required for mitotic events, such as spindle morphogenesis; these cyclins also prevent mitotic exit and cytokinesis, and therefore, mitotic cyclin activity must be downregulated for cell division to be completed. As noted above, a simplification in budding yeast is that all of these cyclins bind to and activate the same Cdk subunit, Cdc28. In instances when this was tested, all of the cyclin-dependent cell-cycle steps required Cdc28 activity as well.All of the G1 and B-type cyclins can bind to and activate Cdc28. The levels of all the cyclins have been quantified on the scale of molecules per cell1, and they were found to range from a few hundred to several thousand molecules per cell. These numbers do not take into account possible variations in the specific activity of different cyclin–Cdc28 complexes; indeed, Clb5 has been shown to have a lower ability to activate Cdc28 than Clb2, and this difference has been proposed to contribute to the ordering of cell-cycle events2.Mechanisms for cyclin specificityCyclin specificity can be achieved in various ways: cyclins are expressed or are present at stable levels at different times; they are differentially sensitive to cell-cycle-regulated inhibitors; they are differentially Laboratory of Yeast Molecular Genetics, Rockefeller University, 1230 York Avenue, New York, New York 10021, USA.Correspondence to F.R.C. e-mail: [email protected]:10.1038/nrm2105Spindle pole body (SPB). The yeast equivalent of the centrosome, which nucleates microtubules, including those that will form the spindle.Multiple levels of cyclin specificity in cell-cycle controlJoanna Bloom and Frederick R. CrossCyclins regulate the cell cycle by binding to and activating cyclin-dependent kinases (Cdks). Phosphorylation of specific targets by cyclin–Cdk complexes sets in motion different processes that drive the cell cycle in a timely manner. In budding yeast, a single Cdk is activated by multiple cyclins. The ability of these cyclins to target specific proteins and to initiate different cell-cycle events might, in some cases, reflect the timing of the expression of the cyclins; in others, it might reflect intrinsic properties of the cyclins that render them better suited to target particular proteins.REVIEWSNATURE REVIEWS | MOLECULAR CELL BIOLOGY VOLUME 8 | FEBRUARY 2007 | 149© 2007 Nature Publishing GroupAnaphase promoting complex(APC). A multicomponent ubiquitin ligase that targets proteins for degradation by the proteasome.restricted to specific subcellular locations; or they bind specifically to only some phosphorylation targets. In some cases, intrinsic cyclin specificity has been traced to specific modular sequences in the cyclin protein (for example, nuclear localization sequences (NLS), destruction boxes that regulate proteolysis or hydro-phobic patches that can mediate interactions with sub-strates). However, the transfer of cyclin specificity by the interchange of modules has been, at best, incompletely demonstrated.Cyclin-specific targeting of Cdk activity is unlikely to be essential. In budding yeast, ectopically expressed Clb1 can rescue a strain in which all other B-type cyc-lins have been deleted3. Although cells that carry only the B-type cyclins Clb1 and Clb2 (clb3–6∆ cells) are not viable, early expression of Clb2 from the CLB5 promoter coupled with the deletion of the gene that encodes Swe1, an inhibitor of early-expressed Clb2–Cdc28 (see below), can rescue the viability of clb3–6∆ cells4. Also, the overexpression of CLB5 or the deletion of the gene that encodes Sic1, an inhibitor of B-type cyclin activity, rescues cells that lack all three G1 cyclins5,6. Similarly, in fission yeast, deletion of the cyclins cig1, cig2 and puc1 allows DNA replication and mitosis to be driven by a single cyclin, Cdc13 (REF. 7). These data indicate that there is significant overlap in the ability of cyclins to target different substrates. A mutant Cdc28 was identi-fied with partial cyclin-independent kinase activity, and it can bypass the genetic requirement for G1 cyclins, but not the