Automated Identification of Cholesterol Interaction Sites on G-Protein Coupled Receptors by Coarse-Grained Simulation

Abstract

Extensive experimental evidence suggests that cholesterol influences the function of G-protein coupled receptors (GPCRs), perhaps by direct and specific interactions with the protein surface. Interaction motifs (cholesterol recognition amino acid consensus motif and cholesterol consensus motif) have been identified by bioinformatic analysis and inspection of crystal structures. Here we report an approach based on analysis of molecular simulation data to identify other putative interaction sites. In unbiased simulations of GPCRs in cholesterol containing membranes, cholesterol samples the protein surface. This data is then analyzed to identify localized, specific, and reproducible interaction sites. The result of the calculation is a ranking of membrane-facing residues in terms of a cholesterol interaction “score,” which may then be used to guide experimental tests based on mutagenesis. Results based on coarse-grained “Martini” simulations of the A2A adenosine receptor are presented, which predict that cholesterol is most likely to bind in three distinct areas on the surface of the protein.