The tumour suppressor protein p21(WAF1) plays a central role in regulating eukaryotic cell-cycle progression. Through its association with G1- and S-phase CDK complexes it regulates activation of the retinoblastoma protein (pRb) and E2F transcription factors. Recognition of CDK/cyclin complexes by p21 occurs, at least in part, through a protein-protein interaction with a binding groove on the cyclin subunit. The same groove has been shown to be involved in the recruitment of macromolecular CDK substrates, including pRb and E2F. Blocking of this recruitment site therefore prevents recognition and subsequent phosphorylation of CDK substrates and offers a therapeutic approach towards restoration of p21-like tumour suppression. Starting from the C-terminal cyclin-binding domain of p21 we have identified the minimal and optimized bioactive (152)HAKRRLIF(159) peptide sequence with respect to CDK protein kinase inhibition where pRb is the substrate. The phosphorylation of histone H1, however, which does not contain a recognizable cyclin-binding motif, was unaffected. Detailed structure-activity relationship investigations revealed that the determinants within this sequence are residues Arg(155), Leu(157) and Phe(159) and more completely define the composition of the cyclin-binding motif. A marked increase in potency was obtained upon replacement of the native Ser(153) with an Ala residue in the context of short synthetic peptide inhibitors and significantly, this mutation resulted in comparable affinity with CDK2/cyclin A as does the full-length recombinant p21 (which has CDK2 and cyclin A binding sites). Peptides derived from various proteins known to interact with cyclins were compared for potency and selectivity. A molecular model of the complex between the cyclin groove and the HAKRRLIF peptide was constructed. This model accounts for the observed peptide structure-activity relationships, including the potency enhancement of the LIF sequence occupying the hydrophobic pocket. Furthermore, it provides generic insights into molecular interactions governing cyclin groove recognition and lays the foundation for the development of peptidomimetic inhibitors of CDKs.