Cyclins and Cyclin-Dependent Kinases

A cyclin-dependent kinase is a protein kinase — an enzyme that transfers the terminal phosphate of ATP onto serine or threonine residues of target proteins — that requires binding to a regulatory partner, a cyclin, before it can catalyze this reaction. The catalytic subunit (CDK) is present at roughly constant levels throughout the cell cycle; what varies, often more than tenfold across a cycle, is the concentration of its cyclin partner. Because CDK activity is essentially proportional to the amount of cyclin available to bind it, the cell converts a time-varying protein concentration into a time-varying enzymatic activity. Different cyclins are synthesized and destroyed at different points in the cycle and prefer different CDK partners, so distinct cyclin–CDK complexes peak sequentially: cyclin D–CDK4/6 in G1, cyclin E–CDK2 at the G1/S boundary, cyclin A–CDK2 during S phase, and cyclin B–CDK1 at G2/M. Each complex phosphorylates a partially overlapping but distinct set of substrates, so each phase transition is executed by a molecularly distinct switch. The name "cyclin" reflects the discovery that these proteins accumulate steadily through interphase and then are destroyed abruptly at a specific point in mitosis or at the G1/S transition — a cyclic sawtooth pattern of accumulation followed by catastrophic degradation. This is not passive decay: cyclins are actively tagged with chains of the small protein ubiquitin by dedicated E3 ubiquitin ligase complexes (the SCF complex and the anaphase-promoting complex/cyclosome, APC/C), marking them for destruction by the proteasome. Because degradation is enzymatic and switch-like rather than gradual dilution, CDK inactivation at the end of a phase is abrupt and effectively irreversible, giving the cell cycle its directionality — cells do not slide backward from mitosis into S phase. A useful precision to hold onto: "activity" of a cyclin–CDK complex is not solely set by cyclin abundance. Full activity requires three coincident events — cyclin binding, phosphorylation of a conserved threonine in the kinase's activation loop by CDK-activating kinase (CAK, itself a CDK–cyclin complex, CDK7–cyclin H), and removal of inhibitory phosphates from two residues near the ATP-binding site by Cdc25 phosphatases. In addition, small inhibitory proteins called CDK inhibitors (CKIs, e.g., p21, p27, and the INK4 family including p16) can bind cyclin–CDK complexes and block activity outright, providing another layer of control that will matter greatly for Cell Cycle Checkpoints and Their Logic.

Step 1 — Cyclin synthesis and CDK binding. Transcription of cyclin genes is turned on at specific points of the cycle (for example, cyclin B mRNA and protein accumulate through S and G2). As newly made cyclin protein accumulates, it binds monomeric CDK molecules that are already present in the cell. Cyclin binding alone induces a conformational change: a flexible loop of the CDK called the T-loop (activation loop), which in the unbound kinase blocks the substrate-binding groove, is partially displaced. This gives partial, low-level kinase activity.

Step 2 — CAK phosphorylation. The CDK-activating kinase (CDK7 paired with cyclin H, part of the general transcription factor TFIIH complex) phosphorylates a conserved threonine within the T-loop (Thr160 in human CDK1/CDK2 numbering). This phosphorylation locks the T-loop out of the substrate groove and aligns catalytic residues, boosting kinase activity roughly 100- to 300-fold over the cyclin-bound but unphosphorylated state.

Step 3 — Inhibitory phosphorylation and its removal (the Wee1/Cdc25 switch). In parallel, a kinase called Wee1 phosphorylates two residues near the ATP-binding pocket of CDK1 (Thr14 and Tyr15 in humans), which sterically blocks ATP from being positioned correctly and keeps the complex inactive even though cyclin is bound and the T-loop is phosphorylated. This is the crucial "primed but paused" state: cyclin B–CDK1 accumulates through G2 while held inactive by Wee1 phosphorylation. Mitotic entry requires the phosphatase Cdc25 to strip these inhibitory phosphates. Because active CDK1 itself phosphorylates and activates Cdc25 (positive feedback) while phosphorylating and inhibiting Wee1 (double-negative feedback that behaves as a positive loop), a small increase in Cdc25 activity triggers an explosive, switch-like, nearly all-or-none activation of the entire cyclin B–CDK1 pool — not a gradual ramp. This bistable feedback circuit is what makes mitotic entry a sudden event rather than a slow slide.

Step 4 — Substrate phosphorylation reshapes the cell. Active cyclin B–CDK1 phosphorylates hundreds of substrates: nuclear lamins (triggering nuclear envelope breakdown), condensin subunits (chromosome condensation), microtubule-associated proteins (spindle assembly), and Golgi/ER structural proteins (organelle fragmentation for equal partitioning). Each substrate has a specific phosphorylation threshold and consensus motif, so the same rising kinase activity triggers different events in sequence as its level crosses different thresholds — explaining why nuclear envelope breakdown, chromosome condensation, and spindle formation occur in a defined order even though driven by the same enzyme.

Step 5 — Destruction of the cyclin, and reset. Once chromosomes are correctly attached to the spindle (monitored by the spindle-assembly checkpoint, foreshadowing Cell Cycle Checkpoints and Their Logic), the APC/C ubiquitin ligase is activated. It attaches polyubiquitin chains to cyclin B (and to securin, releasing separase to cleave cohesin and trigger anaphase), targeting cyclin B for proteasomal destruction. CDK1 activity collapses within minutes. Dephosphorylation of CDK1 substrates by phosphatases (notably PP1 and PP2A-B55) then reverses all the mitotic changes — the nuclear envelope reforms, chromosomes decondense, the cell divides. The same logic, with different molecular players, governs G1/S: cyclin D–CDK4/6 phosphorylates the retinoblastoma protein (Rb), partially relieving its repression of E2F transcription factors; E2F then drives transcription of cyclin E, whose complex with CDK2 hyperphosphorylates Rb further in a positive feedback loop, fully activating E2F and committing the cell to S phase — the molecular basis of the restriction point.

Discovery of cyclin — Tim Hunt, sea urchin embryos, 1983. Studying protein synthesis in fertilized sea urchin eggs at the Marine Biological Laboratory in Woods Hole, Tim Hunt labeled newly made proteins with radioactive methionine and ran samples on gels at successive time points after fertilization. Most proteins' bands stayed constant. One band, however, appeared, grew steadily in intensity, and then abruptly vanished — precisely at each cell division. This happened again in the next cycle, and the next, tracking division with uncanny periodicity. Hunt named this protein cyclin. Because its destruction coincided exactly with cleavage, and because the periodicity repeated autonomously through many division cycles, this was strong evidence that a synthesis-then-destruction cycle of a specific protein was mechanistically tied to progression through mitosis, not merely correlated with it.

Maturation-promoting factor (MPF) — Yoshio Masui and Clement Markert, frog oocytes, 1971. Working independently, Masui and Markert showed that cytoplasm from a mature (meiotically arrested) Xenopus oocyte, injected into an immature oocyte, was sufficient to induce that immature oocyte to undergo maturation (germinal vesicle breakdown, entry into meiotic M phase) — even in the absence of new protein synthesis in the recipient at the moment of injection. This demonstrated a diffusible, transferable cytoplasmic activity, MPF, capable of driving a cell into M phase. Crucially, MPF activity itself was later shown to oscillate — high in M phase, low in interphase — and this oscillation persisted even in cell-free extracts, showing the clock was biochemical, intrinsic to the cytoplasm, and not dependent on nuclear or chromosomal signals.

Genetic identification of CDK1 — Paul Nurse, fission yeast, late 1970s–1987. Using Schizosaccharomyces pombe, Paul Nurse and colleagues isolated temperature-sensitive cell-division-cycle (cdc) mutants. One gene, cdc2, was special: loss-of-function mutations arrested cells before mitosis, while a different gain-of-function mutant allele (cdc2-1w / wee1 pathway mutants) caused cells to divide at abnormally small size, entering mitosis prematurely. This showed cdc2 was rate-limiting for the G2/M transition in both directions — too little activity blocks mitosis, too much triggers it early — precisely what is expected of a master regulatory switch rather than a passive bystander. Nurse then showed, by expressing the human homolog (CDC2, now called CDK1) in cdc2-mutant yeast and rescuing the defect, that the gene was functionally conserved from yeast to humans — a striking demonstration that the mitotic control mechanism is one of the most deeply conserved systems in eukaryotic biology. Nurse, Hunt, and Leland Hartwell (who had independently identified budding yeast cdc genes controlling "Start," the G1 commitment point) shared the 2001 Nobel Prize in Physiology or Medicine for these discoveries.

Convergence — MPF is cyclin B bound to CDK1. Biochemical purification of MPF from frog eggs in the late 1980s (work from James Maller's and Marc Kirschner's laboratories) showed that the purified activity consisted of two components: a catalytic kinase subunit identical to Nurse's cdc2/CDK1, and a regulatory subunit identical to Hunt's cyclin B. This unified three independent experimental traditions — biochemistry of oocyte maturation, radiolabeling of oscillating embryonic proteins, and yeast genetics of cell-cycle mutants — into a single molecular model: MPF = cyclin B–CDK1, and its oscillating activity is the engine of mitotic entry and exit.

Example 1 — Interpreting a Wee1 loss-of-function phenotype. In fission yeast, wild-type cells divide at a characteristic length of about 14 μm. A wee1⁻ mutant divides at roughly 9 μm — smaller cells enter mitosis prematurely because inhibitory phosphorylation of CDK1 is lost, so cyclin B–CDK1 becomes active as soon as enough cyclin B has accumulated, without waiting for the normal Wee1-imposed delay (hence the name "wee"). Now consider a double mutant, wee1⁻ cdc25⁻, where Cdc25 (the phosphatase that removes the inhibitory phosphates) is also nonfunctional. Reasoning: Wee1 adds inhibitory phosphate; Cdc25 removes it. If Wee1 is absent, no inhibitory phosphate is added in the first place, so the requirement for Cdc25 to remove it becomes moot. The double mutant is therefore predicted — and observed — to resemble the wee1⁻ single mutant (small cells, premature division), because epistasis places Wee1's phosphorylation upstream of Cdc25's action on the very same regulatory sites: removing the phosphate-adding enzyme masks the phenotype of losing the phosphate-removing enzyme. This kind of epistasis analysis is exactly how Nurse's group ordered the components of the CDK1 switch before any biochemistry was done.

Example 2 — Quantitative reasoning about cyclin destruction and timing. Suppose in a cultured mammalian cell cyclin B protein accumulates during G2 with an effective net synthesis rate that produces a doubling of cyclin B level every 25 minutes (synthesis exceeding basal turnover), and mitotic entry requires cyclin B–CDK1 activity to cross a threshold corresponding to roughly 8-fold the starting G2 level. Starting from a baseline at G2 onset, how long until threshold is reached, and what happens after APC/C activation? Using doublings: 8-fold = 2³, so 3 doubling times are needed. 3 × 25 minutes = 75 minutes to reach threshold — consistent with typical G2 durations of roughly 1–4 hours observed in many mammalian cell types. Once the spindle-assembly checkpoint is satisfied and APC/C-Cdc20 is activated, cyclin B is polyubiquitinated and degraded with a half-life of only a few minutes (often under 10 minutes in vivo), far faster than the ~25-minute doubling time of synthesis. This asymmetry — slow, gradual accumulation over an hour or more versus catastrophic degradation within minutes — is exactly what produces the sawtooth cyclin profile and the sharp, one-way transition out of mitosis rather than a symmetric rise-and-fall.

Paul Nurse trained as a geneticist and turned to fission yeast, Schizosaccharomyces pombe, as a genetically tractable model in which cell size and division timing could be measured with precision under a microscope. His systematic screens for temperature-sensitive cdc mutants that arrested at defined points of the cell cycle, followed by careful epistasis and complementation analysis, pinned down cdc2 as the master controller of the G2/M transition — years before the corresponding human protein or the biochemistry of cyclin binding were known. His later demonstration that human CDK1 could substitute for yeast cdc2 in a cdc2-null yeast strain was a landmark proof of deep evolutionary conservation, showing that the molecular logic of cell-cycle control long predates the divergence of fungi and animals. Nurse shared the 2001 Nobel Prize in Physiology or Medicine with Leland Hartwell and Tim Hunt, and has continued to work on cell-cycle control, cell size regulation, and science policy, including as President of the Royal Society and founding director of the Francis Crick Institute in London.

The cyclin–CDK circuit is now a major target in oncology. Because cancer cells frequently disable the brakes on this system (loss of the CDK inhibitor p16, amplification of cyclin D or CDK4, loss of functional Rb), restraining CDK activity pharmacologically restores some control over proliferation. CDK4/6 inhibitors — palbociclib, ribociclib, and abemaciclib — block phosphorylation of Rb by cyclin D–CDK4/6, holding hormone-receptor-positive, HER2-negative breast cancer cells in G1, and are now combined with endocrine therapy as standard first-line treatment for metastatic disease, extending median progression-free survival substantially in randomized trials. More broadly, CDK inhibitors are in active development or use for other cancers, since aberrant CDK activation is close to a universal hallmark of unchecked proliferation. In basic and translational research, monitoring cyclin levels (particularly cyclin B1 and cyclin E) by immunohistochemistry or flow cytometry is used to assess proliferative index in tumor biopsies and to time experiments in synchronized cell populations. In reproductive biology, understanding MPF and meiotic cyclin–CDK regulation informs assisted reproduction protocols, since oocyte maturation arrest and release depend on exactly the cyclin B–CDK1 oscillation first characterized in frog eggs. In agriculture and plant biology, cyclin–CDK homologs (plants have their own CDK and cyclin families) regulate meristem growth and are targets for engineering growth rate and organ size in crops. And because this pathway is so deeply conserved, model organisms as distant as fission yeast continue to be used to dissect the basic wiring of cell-cycle control, with direct relevance to human disease — a direct legacy of the experiments described above, and the conceptual foundation for understanding Cell Cycle Checkpoints and Their Logic and The Restriction Point and Commitment to Division.

Ask the tutor: How does the spindle-assembly checkpoint physically prevent APC/C from triggering cyclin B destruction until every chromosome is correctly attached?