3D‐QSAR and Docking Study of the Binding Mode of Steroids to Progesterone Receptor in Active Site

Abstract

AbstractAutodock 3.0 was applied to the search for the binding mode of a set of steroids with progesterone receptor (PR). The results of docking investigation show that oxygen atom of 3‐keto forms three hydrogen bonds with the residues Gln725 and Arg766 respectively. This is consistent with the reported work of S. P. Williams et al. (Nature, 1998, 393, 392–396). Moreover, we have studied in particular the most affinitive compound, steroid with a para alkylthio substituent of benzene at position 11. We find that there is a deep cavity around this substituent and a wide cavity around the substituent on position 17. This result could help us to select substituents to improve the PR‐binding affinity of steroids. The most probable aligned conformers were selected by a docking‐guided conformation selection method, and all‐orientation and all‐placement searches were applied to optimize CoMFA and CoMSIA models. Stable and highly predictive 3D‐QSAR models were built with cross‐validated coefficients q2 0.773 and 0.699 respectively for CoMFA and CoMSIA. The predictive ability of these models was validated using a test set. The results agree with those obtained from the docking investigations.