Library Docking for Cannabinoid-2 Receptor Ligands

Cannabinoid receptors are both therapeutically attractive and are interesting model systems for structure-based methods. Here we investigated topical questions in library docking using the CB2 receptor. While a CB1R docking campaign found potent but nonselective ligands, here subtype selective ligands were found by targeting polar residues. Hit rates and hit affinities improved with library size, but docking against active and inactive receptor states did not reliably bias toward agonists or antagonists. Cryo-EM structures of two of the new agonists superposed well on the docking predictions. Structure-based optimization led to 10- to 140-fold improvements within three series, consistent with well-behaved ligands. Hit rates with an explicit 2.6 billion molecule library resembled those of an implied 11 billion molecule library from a building-block method, supporting the latter's ability to explore this space, though higher affinities were discovered from the explicit set. Implications for future studies are considered.