Protein kinases as targets for virtual ligand screening

CAVASOTTO, CLAUDIO N.; TOTROV, MAXIM; ABAGYAN, RUBEN A.

New Orleans, LA

Conferencia; 225th National Meeting of the American-Chemical-Society; 2003

Protein kinases are drug targets of the utmost importance since they have been implicated in the development and progression of many diseases, including several types of cancer. However, a complicating factor in computational drug design is receptor rearrangement upon ligand binding, evidenced by the difficulty to cross-dock correctly native ligands from different receptor conformations. We report here flexible ligand docking tests using the ICM software on different protein kinase families. Using smoothed potential maps to represent the receptor, our methodology was able to place the ligand within accuracy of 1.5 Å in many cross-docking experiments. We also address the problem of receptor flexibility in those cases in which the cross-docking accuracy is higher than 2.0 Å. We screened a virtual chemical library of 1,000 randomly selected compounds with similar molecular weight to the known binders. The program?s ability to identify the putative ligands within the top ranking compounds is discussed and enrichment factors reported.