A Structural Study of G protein Selectivity in the Muscarinic Acetylcholine Receptor Family

Abstract

G protein coupled receptors (GPCRs) are integral membrane proteins that act as sensory organs for the cell by binding to extracellular signals that activate the receptor for intracellular signal transduction. GPCRs primarily signal through G proteins, which are trimeric proteins named by the distinct signaling pathway they activate: Gαq, Gαi/o, Gαs, and Gα12/13. These receptors evolved to selectively activate one G protein type but can weakly couple to other G proteins. The mechanism of this selectivity remains unknown and increased understanding would allow for rational drug design targeting a single binding partner for a desired physiological effect and eliminating side effects. Our study investigates the structural determinants of G protein selectivity in GPCRs using the Muscarinic acetylcholine receptor family for biophysical modelling. Muscarinic acetylcholine receptors are important pharmacological targets because of their function in modulating signaling pathways implicated in diseases like Alzheimer's. This family contains five members (M1-M5) that are all activated by the same acetylcholine neurotransmitter. These receptors share high structure and sequence homology, however the M1, M3 and M5 receptors selectively bind to the Gαq/11 protein and the M2 and M4 receptors selectively bind to the Gαi/o protein. To investigate this selectivity, we modeled the M1 and M2 receptor bound to each G protein and performed all-atom molecular dynamics to relax the four complex systems for 2 µs. Representative snapshots of the four complexes were then used for structural comparison and binding free energy calculations. Thermodynamically, the M1 and M2 receptors showed a stronger affinity for their cognate binding partners, Gαq and Gαi respectively, this is consistent with experimental studies on the M3 receptor. Structurally, the M1 complexes displayed a larger variety in G protein orientation compared to the M2 complexes which can be explained by the larger G protein binding cavity, suggesting more dynamic and flexible binding. Additionally, we discovered a residue contact between transmembrane helices 5 and 6 that was only present in the cognate complexes (M1:Gαq and M2:Gαi) and similar strong interactions exclusive to each cognate complex. These contacts along with prominent residue interactions in the non-cognate complexes are optimal candidates for further mutation and kinetic studies. Overall, our data supports the previously proposed bar-code hypothesis for G protein selectivity, where a combination of contacts appear to be responsible for G protein binding selectivities.