3D QSAR Analysis of a Set of Pneumocystis carinii DHFR Inhibitors through Pharmacophore Generation Approach

A ligand-based pharmacophore model using Catalyst HypoGen algorithm was developed for set of pyrimidine and quinazoline analogues as pcDHFR inhibitors with an aim to obtain rational hypothetical image of the primary chemical features responsible for activity and this pharmacophore model was used as an in silico screening tool to retrieve novel and potential inhibitors against pcDHFR from various databases. The best pharmacophore model for selective pcDHFR inhibitors (Hypo-1) was obtained through a Cat-Scramble validation process. The best pharmacophore model (Hypo-1) for pcDHFR inhibitors consisting of one hydrogen bond acceptor lipid (HBAl), three hydrophobic (HY) and one ring aromatic (RA) features, with highest correlation coefficient (0.94), cost difference (45.1), low RMS (0.72), as well as it shows a high goodness of fit and predictive factor. Hydrophobic interactions are essential for ligand pharmacophore interaction. Pharmacophore models have been validated toward a test set containing 8 molecules. To further evaluate the model external test set comprising of known pcDHFR inhibitors were mapped on to developed pharmacophoric model which also showed five point mapping and estimated values in close range to actual values. The models were used for screening chemical data base. The validated pharmacophore model (Hypo-1) was used as a 3D query for virtual screening to retrieve potential inhibitors from the Maybridge and National Cancer Institute (NCI) databases. This resulted in identification of three druggable structurally diverse potent lead compounds. The results of our study will act as a valuable tool for retrieving potent compounds with desired biological activities and designing novel selective pcDHFR inhibitors.