On the applicability of homology models of G protein-coupled receptors to computer-aided drug discovery

COMP 252

Santiago Vilar, vilarvarelas@niddk.nih.gov, Giulio Ferino, ferinog@niddk.nih.gov, and Stefano Costanzi, stefanoc@mail.nih.gov. Laboratory of Biological Modeling, NIDDK, National Institutes of Health, 12A Center Drive Rm. 4003, Bethesda, MD 20892-5621
We recently built and compared to the corresponding crystal structure molecular models of the β2-adrenergic receptor (β2-AR) in complex with the inverse agonist Carazolol. The homology models were built using rhodopsin as template and docking was performed through an induced fit procedure that allows flexibility to the ligands and the receptor binding pocket. Here, to probe if homology modeling of G protein-coupled receptors (GPCRs) can generate structures applicable to computer-aided drug discovery, we subjected the crystal structure and our homology models of the β2-AR to a simulated virtual screening. In particular, 67 β2-AR ligands were dispersed within over 50,000 diverse molecules selected from the ZINC lead-like database, and the capacity of prioritizing the active over the inactive compounds through molecular docking was assessed by plotting ROC curves. Our results demonstrate that GPCR homology models are indeed applicable to virtual screenings, in some cases almost as effectively as the experimental structure.

Poster Session
6:00 PM-8:00 PM, Tuesday, August 18, 2009 Walter E. Washington Convention Center -- Ballroom A, Poster

Division of Computers in Chemistry

The 238th ACS National Meeting, Washington, DC, August 16-20, 2009