CELL CYCLE INHIBITION AND DRUG DESIGN: CRYSTAL STRUCTURES OF CDK2-INHIBITOR COMPLEXES
M. Noble1 , N. Brown1, J. Endicott1, L. Johnson1, A. Lawrie1, J. Tucker1, P. Tunnah1, A. Calvert2, N. Curtin2, B. Golding2, R. Griffin2, D. Newell2
1Laboratory
of Molecular Biophysics, South Parks Road, Oxford OX1 3QU, United
Kingdom
2Cancer Research Unit and Department of Chemistry,
Newcastle University, United Kingdom
Keywords:
Cyclin dependent kinase, cancer therapy, inhibitor design
Orderly progression through the eukaryotic cell cycle is directed by a family of protein serine/threonine kinases called the cyclin dependent kinases (CDKs). Inappropriately regulated, CDKs can promote cellular proliferation, which may lead to the development of cancer. Our aim is to develop potent and selective CDK inhibitors, which would be invaluable tools for cell cycle studies, as well as having considerable potential as cancer therapeutics. We have determined structures for CDK2 in complex with both the potent but non-specific protein kinase inhibitor staurosporine (Ki ~ 2nM)[1], and with a family of inhibitors based upon O6-methylguanine (unpublished). Members of this family are as potent as olomoucine against certain members of the CDK family (Ki ~ 2µM). The 12 structures of inhibited CDK2 which we have solved in this work have allowed us to analyse the available determinants of affinity in CDK2 inhibitor binding, as well as to explore the plasticity of the active site and of the binding mode of a family of closely related inhibitors. This analysis will be discussed.
Protein kinase inhibitors for which the binding mode has been structurally characterised compete with ATP to bind in the cleft between the N- and C-terminal domains of the kinase catalytic core. In order to map the nature of this binding cleft, a novel protocol has been developed to analyse the surface of a protein. In a manner analogous to the program GRASP [2], the molecular surface of the target protein is calculated and triangulated [3]. The surfacing probe co-ordinates corresponding to each vertex of the triangulated surface are used as positions at which to evaluate an energetic potential. GRID [4] is used to evaluate this potential, considering the probe to have a chemical nature chosen by the user from the wide range of probe types for which GRID has been calibrated. Probe potentials are subsequently used to colour the protein surface, which can be viewed in a molecular graphics program. The result is an interactive and intuitive depiction of the relative preference of different regions of the protein surface for each of the chemical probe types for which the GRID calculation is performed.