Role of Conformational Changes in G Protein-Coupled Receptor Activation

Citation

Scott, Caitlin Eileen (2014) Role of Conformational Changes in G Protein-Coupled Receptor Activation. Dissertation (Ph.D.), California Institute of Technology. doi:10.7907/Z94747VG. https://resolver.caltech.edu/CaltechTHESIS:05312014-002701930

Abstract

Transmembrane signal transduction is achieved by activation of G protein-coupled receptors (GPCRs) like the human cannabinoid type 1 (CB1) receptor, the human cannabinoid type 2 (CB2) receptor, and the human mu-opioid receptor. These receptors exist in the membrane in an ensemble of conformations each of which might bind to different signaling molecules and cause different physiological effects. Understanding the structural basis of their activation will eventually help us in designing drugs that target these receptors with potentially minimal undesirable side effects. CB1 is of particular interest because it is located in the central nervous system and modulates hunger, making it an attractive anti-obesity drug target. In this receptor, mutating a single residue, threonine 210, to isoleucine in the third transmembrane (TM3) domain makes it far more active than the wild-type (WT) receptor, whereas mutating it to alanine makes it fully inactive. CB1 is difficult to model because it has a small sequence identity with the receptors that have been crystallized. We used the first principles-based GEnSeMBLE method to predict 3D structures of these receptors representing the fully inactive to highly constitutively active states. With this software, we quickly found a set of low energy receptor conformations by sampling trillions of helix orientations. Differences in the intracellular surface explain experimental differences in activation for the CB1 receptor and its mutants. These predictions were validated by designing double mutants that were expected to switch the inactive T210A to WT levels of activation and expected to switch the very active L207A to T210A levels of activation. These predictions were first verified computationally then experimentally with GTPgammaS assays. The accuracy of our predictions indicate that the GEnSeMBLE method is a useful procedure for predicting GPCR structures at various activation states. Known inverse agonists were docked to these predicted CB1 receptor structures, and the resulting complexes were inserted into a solvated lipid bilayer for 50 ns of NPT molecular dynamics with NAMD software. The inverse agonist preferentially binds to a pre-activated CB1 state, but during MD, traits of the inactive structure start to form suggesting that the ligand induces conformational changes.

Thesis Files