Mapping pH-Dependent State Transitions of a Pentameric Ligand-gated Ion Channel through Markov State Modeling

Abstract

Pentameric ligand-gated ion channels are of great importance as intercellular communicators that mediate electrochemical signaling in neurons and other excitable cells. Recently, representative structures in this protein family have been determined through X-ray crystallography or cryo-electron microscopy, confirming the conservation of several topological features. However, details of the gating transition between open and closed states remain unclear, which is hampering efforts towards further mechanistic modeling and drug design. The bacterial homolog, GLIC from Gloeobacter violaceus, is gated by changes in pH and has the advantage of being associated with structural and functional data covering multiple functional states. Here, we present Markov state models of the gating conformational transitions of GLIC by using hundreds of microsecond-long molecular dynamics simulations. These have been started from different initial conditions along a transition path and at different pH values, thus allowing sampling of the whole conformational landscape between open and resting states. From these models, we could confirm the experimental result that the channels prefer a closed conformation at neutral pH, while a small fraction of channels open at low pH. In contrast, simulations of the proton-sensitized mutant, I233T, yielded a clear shift of the free energy profile towards the open state, evident at both pH values. This observation agrees with our electrophysiology recordings of both wild-type and mutant GLIC, demonstrating that simulations are able to predict non-trivial shifts in activation profiles. Furthermore, these results offer insight into the fraction of channels opened upon activation, and provide a basis for mechanistic modeling of intermediate states along the gating pathway.