Cyclin-dependent kinases belong to a family of heterodimeric Ser/Thr protein kinases, consisting of two subunits: a catalytic subunit (a kinase) and an activating subunit (a cyclin). Cdks are regulated at multiple levels, including the binding of activating cyclin subunits, activating and inhibitory phosphorylations and small protein inhibitors. Crystallographic studies of Cdk2/cyclin A complex shed some light on how these regulators control the activity of the kinase. Most of them work by inducing conformational changes in and around the catalytic cleft of the kinase, which alters the affinity of substrates, the rate of phosphorylation or both.
The catalytic ATP-binding cleft is built between the carboxy-terminal and amino-terminal lobes of the kinase. Cdk substrates make the contact mainly with the surface of the carboxy-terminal lobe, interacting with the residues around the ATP-binding cleft. The amino-terminal lobe contains a highly conserved PSTAIRE alpha-helix that interacts with cyclin. In the inactive kinase molecule, some amino-acid residues within the active site are oriented so that the phosphates of incoming ATP are not available for the phosphorylation reaction. Cyclin binding to the PSTAIRE element reorients these residues so that the ATP now fits properly in the cleft. In addition, cyclin binding moves a flexible loop adjacent to the active site of Cdks, called the T-loop, away from the catalytic site entrance. This allows access of the substrates to the ATP. To fully activate the kinase, the T-loop must undergo phosphorylation on a conserved residue T160 (hence the name - T-loop). This phosphorylation causes an additional conformational change in the T-loop that completes the re-organization of the substrate-binding site of the Cdk/cyclin complex.
The fully active Cdk/cyclin complex can be turned off by the binding of the small protein inhibitors of Cip family and INK families. The Cip family protein p27 was shown to inhibit Cdk-cyclin complexes by changing the shape of the catalytic cleft. It also inserts a helix inside the ATP binding site thereby displacing ATP. The INK family protein p16 causes disassociation of cyclin by binding both N and C-lobes of the kinase. This binding does not overlap with cyclin-binding site, but causes allosteric changes in the molecule that propagates to the active site and causes cyclin to fall off.
Cdk/cyclin complexes are also inhibited by phosphorylation of Cdk on conserved T14 and Y15 residues, situated within the active site of the kinase. Y15 phosphorylation of Cdk2/cyclinA was shown to sterically block the access of its substrates to the catalytic site. At the same time, Y15 phosphorylation disorients bound ATP, rendering its conformation non-productive. The effect of the T14 phosphorylation on the conformation of the active site is less clear, although functional studies show that pT14 alone is less inhibitory than pY15 alone or pT14 and pY15 combined. Interestingly, phosphorylation of these residues occurs only in the presence of a bound cyclin.
The crystallographic studies cited above were done primarily on Cdk2/cyclin A complex, while Cdk1/Cyclin B complex has been notoriously difficult to crystallize. Given the high sequence similarity within the family of Cdk proteins, it is reasonable to expect that other Cdks would undergo similar conformational changes upon cyclin binding and phosphorylation. Truncated cyclin B was co-crystallized with Cdk2, and the structure of this complex was almost identical in structure to Cdk2/cyclin A.
Cyclin-dependent kinases are activated by binding to the proteins of the cyclin family. The name “cyclins” reflects the periodicity in their protein levels throughout the cell cycle. Cyclins were first discovered in early embryos of marine invertebrates (sea urchin and clam). There it was shown that two proteins, subsequently named cyclins A and B, accumulated before mitosis and rapidly disappeared thereafter. The rounds of synthesis and degradation happened once per cell cycle, hence the name – cyclins. Using embryonic extracts, cyclins were shown to co-immunoprecipitate with Cdc2 kinase activity. The fission yeast cyclin Cdc13 was also shown to associate with active Cdc2 kinase. In starfish oocytes, the Cdc2 kinase was found to associate with cyclin B1 with 1:1 stoichiometry, forming a catalytically active heterodimer. Together, this evidence led to the idea that that cyclins could be activating subunits of Cdc2 kinase, and numerous subsequent studies showed that that this indeed is the case.