Inhibitors of Glycogen Synthase Kinase-3beta and Cyclin-Dependent Kinases Modelled by 3D-QSAR Using a Novel Alignment Method Based on Electrostatic Potentials.

Mahindra Makhija ¹, Erik Helmerhorst²
¹M.Makhija@exchange.curtin.edu.au, Western Australian Biomedical Research Instittue, School of Biomedcial Sciences, Curtin University of Technology, Bentley, WA 6102, Australia; ²E.Helmerhorst@curtin.edu.au, Western Australian Biomedical Research Instittue, School of Biomedcial Sciences, Curtin University of Technology, Bentley, WA 6102, Australia

Abstract: Glycogen synthase kinase-3 (GSK-3) is a ubiquitous, multifunctional serine/threonine kinase that regulates numerous signalling pathways in cells. Overactivity of GSK-3 is linked to the pathology of several diseases including Alzheimer’s disease and non-insulin-dependent diabetes and, therefore, there is an immense interest in the development of specific inhibitors of this enzyme. Paullones and aloisines are known inhibitors of GSK-3? and of the cyclin dependent kinases (CDK). In this study, a comparative molecular similarity indices analysis (CoMSIA) was performed using a training set of 56 paullones with known in-vitro biological activity toward GSK-3?, CDK1 and CDK5. This training set was aligned using atom-fitting and a novel molecular electrostatic potential (MEP) approach (Makhija, M. T. and Kulkarni, V. M. Journal of Computer-Aided Molecular Design 2001, 15, 961 – 978). The predictive ability of the resultant model then was evaluated using a test set of 12 aloisines, which are structurally unrelated to the paullone inhibitors. The final model for GSK-3? using a MEP based alignment showed a cross-validated r2 of 0.686 and predictive r2 of 0.668 demonstrating its value in designing novel inhibitors. Similarly, the final models for CDK1 and CDK5 also showed considerable internal and external predictability. The contour maps for GSK-3?, CDK1 and CDK5 describing the steric, electrostatic, hydrophobic, hydrogen bond donor and acceptor interactions also will be invaluable in designing new selective inhibitors of GSK-3? for the treatment of Alzheimer’s disease.