Molecular Modeling Strategies to Probe Stereoselectivity of HIV Protease Inhibitors

Abstract

HIV being the etiologic agent of AIDS remains one of the standstill challenges all over the world from the past few decades. During HIV replication, HIV protease plays a critical role in splitting a polyprotein precursor (gag and gag-pol) into smaller subunits to produce the infectious virions. Inhibitors of HIV protease evade the viral replication by blocking the cleavage of polyprotein precursors in the life of HIV. Recently, it has been demonstrated that viral protease possesses a two-fold C2 symmetry in its structure which opened a new era for the development of stereoselective drugs with improved efficacy. Furthermore, current guidelines of U.S. Food and Drug Administration (FDA) also demonstrated that use of stereoselective drugs is more efficacious as compared to racemates. However, until now there was no report available for the molecular interaction profile of stereoselective drugs with HIV Protease. Therefore, it necessitates a detailed knowledge about 3D structural features important for the inhibition of HIV protease and molecular basis of stereoselectivity of its inhibitors. In this project stereoselectivity of known HIV protease inhibitors has been explored as well as correlated with respective ligand-protein interaction pattern by using both structure and ligand based computational approaches. Molecular docking results delineated a hypothesis of HIV protease inhibitors within the binding cavity of HIV. Our docking results suggests that difference in inhibitory potency of stereoisomeric inhibitors is might be due to the difference in their interaction pattern within the binding cavity of HIV protease. In order to predict the anti-HIV properties of new chemical entities, various ligand based pharmacophore were also generated. However, the final model comprises of four different pharmacophoric features including two hydrophobic, one acceptor and one aromatic feature. The finally selected model showed 100% sensitivity, 75% specificity and 82% (Mathew’s Correlation Coefficient) MCC values. Finally selected model also showed 100% sensitivity, 66% specificity and 78% MCC values with the external dataset of commercially available FDA approved drugs against HIV. Both the ligand-protein interaction pattern and 18 pharmacophore model collectively aid in understanding the stereoselectivity in drugs which might help in the future forecasting of new chemical entities against HIV. Additionally, the difference in the biological activities of stereoisomers might have different clinical significance, metabolic rate and thus, the use of enantiopure compounds can be more advantageous over racemates.