Multiple Molecular Dynamics Simulations and Free-Energy Predictions Uncover the Susceptibility of Variants of HIV-1 Protease against Inhibitors Darunavir and KNI-1657

Abstract

HIV-1 protease (PR) is considered to be the main targets of anti-AIDS drug design because of its role in the proteolytic processing of viral polyproteins. However, the emergence of drug-resistant HIV has become a major problem in the therapy of HIV-1-infected patients. Focused on the complexes of wild type (WT) PR and two mutant PRs (V32I/L33F/I54M/V82I and V32I/L33F/I54M/I84 V) with inhibitors Darunavir (DRV) and KNI-1657 (KNI), respectively, we have conducted research on the conformational dynamics and the resistance mechanism caused by residue mutations through multiple molecular dynamics (MD) simulations combined with an energy (MM-PBSA and solvated interaction energy (SIE)) prediction. The results indicate that mutated residues of PR alter the distance between flap regions and catalytic sites, the volume of the inner catalytic site, and the curling degree of the flap tips, thereby affecting DRV and KNI inhibitor binding to PR. These mutated residues reduced the binding affinity of the two mutant PRs to DRV, resulting in drug resistance, whereas the two mutant PRs increase the binding affinity with KNI, indicating they enhance the sensitivity to KNI. Compared with the WT PR, the changes in van der Waals interaction and electrostatic interaction in the two variant PRs play a vital part in the binding of PR with DRV and KNI. These results may supply valuable guidance for the design of anti-AIDS drugs targeting PR.

Using deep learning and molecular dynamics simulations to unravel the regulation mechanism of peptides as noncompetitive inhibitor of xanthine oxidase

“…The MM-PBSA method as applied to small molecule binding is an end-point method estimating the binding free-energy difference between the protein–ligand complex 53 – 56 . The single-trajectory approach is favored for its straightforward implementation and cancellation of covalent energy errors as conformations for the complex and separated receptor and ligand are based on shared configurations from MD simulations.…”

Machine learning‐guided design of potent darunavir analogs targeting HIV‐1 proteases: A computational approach for antiretroviral drug discovery

“…20 Therefore, the application of ML-based techniques to design novel analogs, utilizing insights gained from available reported compound datasets accessed from a public chemical database, combined with several in silico methods in structure-based drug design (SBDD). This is challenged to prove the potency of designed DRV analogs against wild-type (WT) 21 and mutants of PR, [21][22][23] presenting a potent and rapid approach to mitigating the viral load of HIV patients and curbing the acquired immune deficiency syndrome (AIDS) epidemic.…”

Molecular recognition between volatile molecules and odorant binding proteins 7 by homology modeling, molecular docking and molecular dynamics simulation

“…Therefore, in our work, we first used the homology modeling method to construct the 3D structure of SfruOBP7, then used the molecule docking method [24][25][26][27] to obtain complex models of SfruOBP7 and different chemical volatile molecules (Fig. 1) and finally carried out MD simulations and various analysis approaches [28][29][30] for these complexes. The interaction between SfruOBP7 and different chemical volatile molecules is elucidated in detail; at the same time, the key factors of their binding are identified.…”