GPCR are integral membrane proteins involved in cell signaling in a variety of cell types which makes them highly sought after targets for drug development. While GPCR are attractive drug targets, a significant number remain orphans; GPCR having unknown endogenous ligands. Ligand identification efforts have various starting points with computational methods becoming increasingly popular. Homology modeling is a widely used computational tool that generates models of proteins with unknown structures via sequence commonality to proteins of similar structure and function. Traditionally, homology modeling template selection (protein of known structure) is based on global sequence identity, however better models may result from templates with highly localized sequence identity. For example, high sequence identity localized within the ligand binding pocket may produce better models to examine protein-ligand interactions. This in silico benchmark study examined the performance of global versus local template selection methods for 6 crystallized class A GCPR (CXCR4, FFAR1, NOP, P2Y12, OPRK, and muscarinic-M1) with the long term goal of optimizing ligand identification efforts in orphan GPCR. Global, 61-residue weighted, and 8-residue weighted all-atom and alpha carbon RMSD of models versus crystal structures were calculated. The 61 residues used are common to GPCR ligand binding pockets and the 8 residues selected have strong interactions in 70% of class A GPCR ligand binding. Overall, 4 of the 6 (3 of the 6) local template models gave better 61-residue weighted RMSD (8-residue weighted RMSD) than the global template models. Of the local template models with worse correlations, the RMSD were similar to that of the global template models. These data suggest that locally weighted templates may provide better models of GPCR ligand binding pocket. RMSD of ligand poses suggest that locally selected template models better mimic the crystallographic ligand position. Additional evaluations of ligand docking performance are ongoing to verify these results.

Homology model template selection benchmarking: Global versus local similatiry measures

256th National Meeting (2018)

The official archive and repository for all ACS Meetings & Expositions abstracts and select posters or presentations.