Metadynamics simulations leveraged by statistical analyses and artificial intelligence-based tools to inform the discovery of G protein-coupled receptor ligands

Abstract

G Protein-Coupled Receptors (GPCRs) are a large family of membrane proteins with pluridimensional signaling profiles. They undergo ligand-specific conformational changes, which in turn lead to the differential activation of intracellular signaling proteins and the consequent triggering of a variety of biological responses. This conformational plasticity directly impacts our understanding of GPCR signaling and therapeutic implications, as do ligand-specific kinetic differences in GPCR-induced transducer activation/coupling or GPCR-transducer complex stability. High-resolution experimental structures of ligand-bound GPCRs in the presence or absence of interacting transducers provide important, yet limited, insights into the highly dynamic process of ligand-induced activation or inhibition of these receptors. We and others have complemented these studies with computational strategies aimed at characterizing increasingly accurate metastable conformations of GPCRs using a combination of metadynamics simulations, state-of-the-art algorithms for statistical analyses of simulation data, and artificial intelligence-based tools. This minireview provides an overview of these approaches as well as lessons learned from them towards the identification of conformational states that may be difficult or even impossible to characterize experimentally and yet important to discover new GPCR ligands.

Keywords: GPCRs (G protein-coupled receptors); enhanced sampling; machine learning; metadynamics; molecular dynamics simulation.