Ligand discrimination during virtual screening of the CB1 cannabinoid receptor crystal structures following cross-docking and microsecond molecular dynamics simulations†

Abstract

The therapeutic potential of the CB1 cannabinoid receptor remains underexploited with only a few synthetic ligands on the market. The crystal structures of both the inactive and active-state CB1 receptor have recently been solved, allowing for unprecedented opportunities in structure-based drug discovery applications such as virtual screening. In this study, we have investigated the virtual screening performance of the active and inactive-state CB1 crystal structures and their ability to discriminate between agonist and inverse agonist/antagonist ligands. The ligands of inactive and active-state CB1 receptor crystal structures were then swapped via cross-docking and the resulting structures were subjected to microsecond molecular dynamics (MD) simulations, followed by virtual screening of the MD-extracted structures. The original crystal structures were found to be biased towards ligands matching their activation state during virtual screening. MD simulations of the cross-docked CB1 structures resulted in a minor shift of receptor conformation towards the inactive state for the active-state CB1 structure complexed with the inverse agonist taranabant. Effects on virtual screening were more pronounced, as MD simulations of the cross-docked receptor–ligand complexes reversed the ligand bias in virtual screening observed with the original crystal structures. The simulations also produced receptor conformations that outperformed the crystal structures in virtual screening and in predicting the binding pose of the cognate ligand. The findings of this study highlight the potential of cross-docking and MD simulations to reverse the ligand bias of crystal structures, which may be useful when the crystal structure of only one activation state is available.

Abstract

Ligands of inactive and active-state CB1 receptor crystal structures were swapped and virtual screening performance assessed after molecular dynamics simulations.