Ligand-Optimized Homology Models of D₁ and D₂ Dopamine Receptors: Application for Virtual Screening

Recent breakthroughs in crystallographic studies of G protein-coupled receptors (GPCRs), together with continuous progress in molecular modeling methods, have opened new perspectives for structure-based drug discovery. A crucial enhancement in this area was development of induced fit docking procedures that allow optimization of binding pocket conformation guided by the features of its active ligands. In the course of our research program aimed at discovery of novel antipsychotic agents, our attention focused on dopaminergic D2 and D1 receptors (D2R and D1R). Thus, we decided to investigate whether the availability of a novel structure of the closely related D3 receptor and application of induced fit docking procedures for binding pocket refinement would permit the building of models of D2R and D1R that facilitate a successful virtual screening (VS). Here, we provide an in-depth description of the modeling procedure and the discussion of the results of a VS benchmark we performed to compare efficiency of the ligand-optimized receptors in comparison with the regular homology models. We observed that application of the ligand-optimized models significantly improved the VS performance both in terms of BEDROC (0.325 vs 0.182 for D1R and 0.383 vs 0.301 for D2R) as well as EF1% (20 vs 11 for D1R and 18 vs 10 for D2R). In contrast, no improvement was observed for the performance of a D2R model built on the D3R template, when compared with that derived from the structure of the previously published and more evolutionary distant β2 adrenergic receptor. The comparison of results for receptors built according to various protocols and templates revealed that the most significant factor for the receptor performance was a proper selection of "tool ligand" used in induced fit docking procedure. Taken together, our results suggest that the described homology modeling procedure could be a viable tool for structure-based GPCR ligand design, even for the targets for which only a relatively distant structural template is available.