Ligand Binding\nand Functional Effect of Novel Bicyclic\nα5 GABA<sub>A</sub> Receptor Negative Allosteric Modulators

The significant importance of GABA<sub>A</sub> receptors\nin the\ntreatment of central nervous system (CNS) disorders has been known\nfor a long time. However, only in recent years have experimental protein\nstructures been published that can open the door to understanding\nprotein–ligand interactions and may effectively help the rational\ndrug design for the future. In our previous work (Szabó, G. J. Med. Chem. 2022, 65(11), 7876), where a promising selective\nα5-GABA<sub>A</sub> negative allosteric modulator (NAM) was\ndeveloped containing the 3-(4-fluorophenyl)-5-methyl-1,2-oxazole headgroup,\nwe noticed a switch-like effect of a single nitrogen atom for the\nreceptor function in some derivatives having a dihydro-naphthyridinone\nor dihydro-isoquinolinone moiety. Here, we focused on this chemotype,\nand a small set of compounds were designed to investigate ligand–receptor\ninteractions experimentally and through computational methods. Elaborated\ncompounds were tested against GABA<sub>A</sub> α1 and α5\nsubunit-containing receptors, and binding affinities and functional\nactivities were measured. Starting from the published experimental\nstructure of an engineered, homopentameric, basmisanil-binding GABA<sub>A</sub> receptor-like construct consisting of modified α5 subunits\nand an α1-containing GABA<sub>A</sub> structure, we created\na new model of the ligand binding site at the α5/γ2 interface.\nUsing this model, the measured ligand affinities were able to be reproduced\nwell by free energy perturbation (FEP) calculations. In addition,\ncalculations were able to explain the obtained structure–activity\nrelationships, among others, the switch-like effect of the aromatic\nnitrogen position in the dihydro-naphthyridinone motif for the functional\ncharacter, and suggest different binding poses for the ligands presenting\nsilent versus negative allosteric effects in this set (SAMs vs. NAMs,\nrespectively). We believe that our results can help design α5\nselective GABA<sub>A</sub> negative allosteric modulators and better\nunderstand the GABA<sub>A</sub> receptor.