Exploration of the structural requirements of HIV-protease inhibitors using pharmacophore, virtual screening and molecular docking approaches for lead identification.

Pharmacoinformatics approaches are widely used in the field of drug discovery as it saves time, investment and animal sacrifice. In the present study, pharmacore-based virtual screening was adopted to identify potential HIV-protease ligands as anti-HIV agents. Pharmacophore is the 3D orientation and spatial arrangement of functional groups that are critical for binding at the active site cavity. Virtual screening retrieves potential hit molecules from databases based on imposed criteria. A set of 30 compounds were selected with inhibition constant as training set from 129 compounds of dataset set and subsequently the pharmacophore model was developed. The selected best model consists of hydrogen bond acceptor and donor, hydrophobic and aromatic ring, features critical for HIV-protease inhibitors. The model exhibits high correlation (R=0.933), less rmsd (1.014), high cross validated correlation coefficient (Q(2)=0.872) among the ten models examined and validated by Fischer's randomization test at 95% confidence level. The acceptable parameters of test set prediction, such as R(pred)(2)=0.768 and r(m(test))(2)=0.711 suggested that external predictivity of the model was significant. The pharmacophore model was used to perform a virtual screening employing the NCI database. Initial hits were sorted using a number of parameters and finally seven compounds were proposed as potential HIV-protease molecules. One potential HIV-protease ligand is reportedly confirmed as an active agent for anti-HIV screening, validating the current approach. It can be postulated that the pharmacophore model facilitates the selection of novel scaffold of HIV-protease inhibitors and can also allow the design of new chemical entities.