Abstract
G-protein coupled receptors (GPCRs) represent a significant target class for pharmaceutical therapies. However, to date, only about 10% of druggable GPCRs have had their structures characterized at atomic resolution. Further, because of the flexibility of GPCRs, alternative conformations remain to be modeled, even after an experimental structure is available. Thus, computational modeling of GPCRs is a crucial component for understanding biological function and to aid development of new therapeutics. Previous single- and multi-template homology modeling protocols in Rosetta often generated non-native-like conformations of transmembrane α-helices and/or extracellular loops. Here we present a new Rosetta protocol for modeling GPCRs that is improved in two critical ways: Firstly, it uses a blended sequence- and structure-based alignment that now accounts for structure conservation in extracellular loops. Secondly, by merging multiple template structures into one comparative model, the best possible template for every region of a target GPCR can be used expanding the conformational space sampled in a meaningful way. This new method allows for accurate modeling of receptors using templates as low as 20% sequence identity, which accounts for nearly the entire druggable space of GPCRs. A model database of all non-odorant GPCRs is made available at www.rosettagpcr.org.
Author Summary Structure-based drug discovery is among the new technologies driving the development of next generation therapeutics. Inherent to this process is the availability of a protein structure for virtual screening. The most heavily drugged protein family, G-protein coupled receptors (GPCRs), however suffers from a lack of experimental structures that could hinder drug development. Technical challenges prevent the determination of every protein structure, so we turn to computational modeling to predict the structures of the remaining proteins. Again, traditional techniques fail due to the high divergence of this family. Here, we build on available methods specifically for the challenge of modeling GPCRs. This new method outperforms other methods and allows for the ability to accurately model nearly 90% of the entire GPCR family. We therefore generate a model database of all GPCRs (www.rosettagpcr.org) for use in future drug development.
