- Kaplan, Anat Levit;
- Confair, Danielle N;
- Kim, Kuglae;
- Barros-Álvarez, Ximena;
- Rodriguiz, Ramona M;
- Yang, Ying;
- Kweon, Oh Sang;
- Che, Tao;
- McCorvy, John D;
- Kamber, David N;
- Phelan, James P;
- Martins, Luan Carvalho;
- Pogorelov, Vladimir M;
- DiBerto, Jeffrey F;
- Slocum, Samuel T;
- Huang, Xi-Ping;
- Kumar, Jain Manish;
- Robertson, Michael J;
- Panova, Ouliana;
- Seven, Alpay B;
- Wetsel, Autumn Q;
- Wetsel, William C;
- Irwin, John J;
- Skiniotis, Georgios;
- Shoichet, Brian K;
- Roth, Bryan L;
- Ellman, Jonathan A
There is considerable interest in screening ultralarge chemical libraries for ligand discovery, both empirically and computationally1-4. Efforts have focused on readily synthesizable molecules, inevitably leaving many chemotypes unexplored. Here we investigate structure-based docking of a bespoke virtual library of tetrahydropyridines-a scaffold that is poorly sampled by a general billion-molecule virtual library but is well suited to many aminergic G-protein-coupled receptors. Using three inputs, each with diverse available derivatives, a one pot C-H alkenylation, electrocyclization and reduction provides the tetrahydropyridine core with up to six sites of derivatization5-7. Docking a virtual library of 75 million tetrahydropyridines against a model of the serotonin 5-HT2A receptor (5-HT2AR) led to the synthesis and testing of 17 initial molecules. Four of these molecules had low-micromolar activities against either the 5-HT2A or the 5-HT2B receptors. Structure-based optimization led to the 5-HT2AR agonists (R)-69 and (R)-70, with half-maximal effective concentration values of 41 nM and 110 nM, respectively, and unusual signalling kinetics that differ from psychedelic 5-HT2AR agonists. Cryo-electron microscopy structural analysis confirmed the predicted binding mode to 5-HT2AR. The favourable physical properties of these new agonists conferred high brain permeability, enabling mouse behavioural assays. Notably, neither had psychedelic activity, in contrast to classic 5-HT2AR agonists, whereas both had potent antidepressant activity in mouse models and had the same efficacy as antidepressants such as fluoxetine at as low as 1/40th of the dose. Prospects for using bespoke virtual libraries to sample pharmacologically relevant chemical space will be considered.