Rational Discovery of GSK3-Beta Modulators Aided by Protein Pocket Prediction and High-Throughput Molecular Docking
Abstract
Over the last three decades, computer-aided drug design (CADD) methods have attracted increasing attention of medicinal chemists especially due to their potential to penetrate into the molecular level of drugs' mechanism of action. So far, CADD techniques have significantly contributed to rational development of more than two hundred novel drugs.
Brute force of super-computers has enabled chemists to screen virtual ligand libraries of millions of chemical structures in affordable time span while indicating which compounds should be prioritized in further preclinical research and which can be eliminated a priori. A prominent position in CADD is held by structure-based methods that analyze 3D structures of biological targets to find optimal small binding molecules modulating the target's bioactivity.
In the current work, we have performed a protein binding pocket screening on an X-ray model of human Glycogen Synthase Kinase 3 beta (GSK3 beta) employing an algorithm based on Voronoi tessellation. The found binding sites were analyzed and compared with the results of surface screening of the GSK3 beta model by molecular docking based calculations.
Finally, the revealed binding sites were exploited in a structure-based virtual screening supported by pleasingly parallelized calculations on a peta-flops-scale supercomputer. The most promising GSK3 beta modulators resulting from the in silico screening have been proposed for in vitro testing.