Functional studies based on patch clamp

Functional studies based on patch-clamp electrophysiology provide additional information on the active state of GlyR α1. Current-voltage measurements from single muscle metabolism in outside-out patches at symmetric 145-mM chloride concentration yielded a unitary conductance of 86 pS for the main open state and a range from 18 to 64 pS for a series of less populated states (Bormann et al., 1993). Quantification of the reversal potential by whole-cell recordings upon substitution of the external solution with diluted NaCl or a non-permeant organic anion salt revealed a strong selectivity for chloride with a permeability ratio of Cl− versus Na+ of ∼25 (Keramidas et al., 2002). Finally, systematic analyses of channel permeability to large polyatomic anions in activated GlyR provided estimates of the pore dimensions under physiological conditions (Bormann et al., 1987). In these experiments, the permeability sequence chloride > formate > bicarbonate > acetate, with phosphate and propionate being not measurably permeant, is consistent with a minimum pore diameter of 5.3 Å (Rundström et al., 1994). Altogether, these measurements provide stringent “functional” restraints that should be reproduced by any meaningful representation of the physiologically active state of GlyR α1; note that the pore diameter at position −2′ in the wide-open and the semi-open cryo-EM structures of the zebrafish GlyR α1 is 8.4 Å and 4.8 Å, respectively. Finally, computational approaches have been also developed to aid in the functional annotation of ion-channel structures. The most popular ones, including the software HOLE (Smart et al., 1996), assess the functional state of the channel (i.e., open or closed) from static measurements of the physical dimensions of the pore. However, since pore dewetting may result in channel closure even when the pore may geometrically accommodate water and ions, a.k.a. hydrophobic gating (Beckstein and Sansom, 2006), these approaches may lead to incorrect conclusions. Assuming that pore hydration can be used as a proxy for ion conductance, a three-level dynamic annotation has been recently proposed to obtain functional insights beyond static examinations (Trick et al., 2016). Application of this methodology in the context of GlyR α1 predicted the wide-open structure to be representative of the active state, and the semi-open structure as ion impermeable and possibly desensitized (Trick et al., 2016). However, a joint X-ray crystallography and simulation analysis of GLIC and GlyR last year challenged the physiological significance of the wide-open architecture, providing evidence in favor of the narrower-pore structure (i.e., semi-open) as being representative of the open state (Gonzalez-Gutierrez et al., 2017). In this communication, we reanalyze the cryo-EM structures of the zebrafish GlyR α1 by an original combination of all-atom molecular dynamics (MD) simulations, computational electrophysiology, and polyatomic anion permeation simulations to compare with patch-clamp experiments. Our analysis indicates that none of the currently available open-channel structures of GlyR is a true representation of the active state. Rather, antigenic determinant suggests that a distinct open-channel form captured by MD during the relaxation of the wide-open structure, which is stable, chloride conducting, and selective to small polyatomic anions in simulation, provides a closer representation of the physiologically active state of GlyR.
Results
Discussion pLGICs mediate synaptic transmission by switching between discrete states in response to ligand-binding events. Providing an atomistic description of the relevant physiological states is crucial for the development of new pharmacological strategies. Here, we explore the structure of the physiologically active state of the GlyR α1 by MD and validate the simulation results by computational electrophysiology and polyatomic anion permeation simulations. Our analysis suggests that none of the currently available open-channel structures of GlyR is a true representation of the active state; the wide-open glycine-bound structure being five times as conductive and non-selective to large polyatomic anions, and the semi-open glycine/ivermectin bound structure being non-permeable to chloride at physiological conditions. By contrast, the open-channel state captured by MD during the gentle relaxation of the wide-open structure (MD-open) is stable in a native membrane environment, and ion-conductive and selective in qualitative agreement with patch-clamp electrophysiology. Based on these results, we conclude that the MD-open structure provides the closest representation to date of the physiologically active state of GlyR α1. Coordinates of the MD-open structure are available for download in PDB format as a 3D molecular model. Despite the radical conclusion on the wide-open and the semi-open structures of GlyR α1, note that our observations are consistent with recent simulation analyses by others (Trick et al., 2016, Gonzalez-Gutierrez et al., 2017).