3D-QSAR, CoMFA, and CoMSIA of new phenyloxazolidinones derivatives as potent HIV-1 protease inhibitors

Abstract

As a fundamental factor in acquired immunodeficiency syndrome (AIDS) therapy, it has been shown that HIV-1 protease inhibitors to be an important in restraining HIV-1. In the present study, the three-dimensional quantitative structure–activity relationship (3D-QSAR) modeling was conducted on a series of phenyloxazolidinone derivatives. The comparative molecular field analysis (CoMFA), comparative molecular similarity indices analysis (CoMSIA), performed on the training set of 51 compounds, and the optimum PLS model on CoMFA/CoMSIA descriptors showed “leave-one-out” cross-validation correlation coefficients (Q ²) of 0.772 and 0.720 as well as the non-cross-validated correlation coefficients (R ²_ncv ) of 0.966 and 0.963, respectively. Furthermore, the satisfactory results, based on the bootstrapping analysis and tenfold cross-validation, suggest the highly statistical significance of the optimal model. The statistical parameters from the models indicate that the data are well fitted and have high predictive ability. Moreover, the resulting 3D CoMFA/CoMSIA contour maps provide useful guidance for designing highly active inhibitors. The external predictive capability of the established model was evaluated by an external test set of 16 compounds, resulting in the predicted correlation coefficients (R ²_pred ) of 0.9848 and 0.9291, respectively. It is recommended that the bulky electron-donating substituents at the regions A and D can increase the biological activities of the inhibitors. It is expected that the developed model could provide some useful information for the future synthesis of highly potent HIV-1 protease inhibitors.