Identification of highly active natural thyroid hormone receptor agonists by pharmacophore-based virtual screening

Thyroid hormone receptor (TR) is an important target for the treatment of metabolic diseases. The X-ray crystallographic data for the TR complexed with different ligands were employed to generate feature-based pharmacophore models of the active site of TR receptor. The derived hypothesis was then used to find novel hit compounds through an in silico virtual screening on the Universal Natural Products Database (UNPD). The binding mode and action mechanism of the hit compounds were further investigated by molecular docking and molecular dynamics studies, then compounds possessing similar binding site with the crystal ligand were subjected to binding activity assay. Finally, UNPD19665 and UNPD184785 were proved to be more active than crystal TR ligands, with a binding activity value of 9.82 nM and 12.62 nM, respectively. Communicated by Ramaswamy H. Sarma.