Computer-assisted assessment of potentially useful non-peptide HIV-1 protease inhibitors

Abstract

Quantitative structure–activity relationship (QSAR) studies were recently performed to model the bioactivities of two different series of non-peptide HIV-1 protease inhibitors. The sum of the substructures of these two compound classes giving rise to new actives can cause synergistic effects on bioactivities and enhanced pharmacokinetic parameters. Therefore, the two congeneric series were joined and a MIA-QSAR model was built and used to estimate the biological activities of new compounds derived from the miscellany of substructures of the most active compounds of both series. The QSAR model was validated through leave-one-out cross-validation and external validation, and its robustness attested by means of a Y-randomization test. One of the proposed compounds was very promising and, therefore, submitted to ADME evaluation, demonstrating improved properties in comparison to the existing compounds. Docking studies demonstrated the high affinity of the novel compound towards HIV-1 protease, especially due to interactions with catalytic Asp dyad, in agreement with the expected trend obtained by QSAR for the proposed compounds and by the experimental data of the most active ligands.