In Silico Methods for the Discovery of Orthosteric GABA Receptor Compounds

The GABA receptor (GABA-R) is a heterodimeric class C G protein-coupled receptor comprised of the GABA and GABA subunits. The endogenous orthosteric agonist γ-amino-butyric acid (GABA) binds within the extracellular Venus flytrap (VFT) domain of the GABA subunit. The receptor is associated with numerous neurological and neuropsychiatric disorders including learning and memory deficits, depression and anxiety, addiction and epilepsy, and is an interesting target for new drug development. Ligand- and structure-based virtual screening (VS) are used to identify hits in preclinical drug discovery. In the present study, we have evaluated classical ligand-based in silico methods, fingerprinting and pharmacophore mapping and structure-based in silico methods, structure-based pharmacophores, docking and scoring, and linear interaction approximation (LIA) for their aptitude to identify orthosteric GABA-R compounds. Our results show that the limited number of active compounds and their high structural similarity complicate the use of ligand-based methods. However, by combining ligand-based methods with different structure-based methods active compounds were identified in front of DUDE-E decoys and the number of false positives was reduced, indicating that novel orthosteric GABA-R compounds may be identified by a combination of ligand-based and structure-based in silico methods.