Computational Studies of the HIV-1 Protease Dimer Interface.

Abstract

HIV-1 protease (HIVp) is one of four major drug targets to prevent propagation of the infectious HIV virion. Currently, all ten marketed HIVp drugs are inhibitors that target the HIVp active site. However, these drug therapies provide selective pressure resulting in mutations of the protease that escape drug efficacy. Consequently, the development of inhibitors of HIVp that have new modes of action is necessary. The dimer interface is an attractive target due to its highly conserved nature and its importance in forming an active enzyme. Until now, all dissociative inhibitors were created by mimicking residues at the dimer interface, and resulted in several non-drug-like compounds. However, we created several receptor-based pharmacophore models of the HIVp dimer interface, using ensembles of multiple protein structures (MPS). The MPS method was used to map the dimer interface with a series of small-molecule probes – methanol, ethane, and benzene. The maps were translated into pharmacophore models which were used to filter in silico, three-dimensional library of small molecules. The MPS method identified several novel small-molecule inhibitors capable of inhibiting dimerization, with several compounds characterized with less than 50 uM-level affinity. In the clinically relevant multi-drug resistant form of HIVp, these compounds maintained dissociative inhibition with nearly identical inhibition rates. Zhang-Poorman kinetic analysis verified the small molecules inhibit HIVp in a dissociative manner. In addition to creating novel inhibitors, we modeled the protein-ligand interaction of known dissociative inhibitors using Langevin Dynamics. Ten, 10-ns simulations were initiated based on the hypothetical mechanism of ligand binding, but the dynamics simulations showed that the complex was unstable. Although the simulations did not result in a clear mechanism for protein-ligand binding, of the known dimerization inhibitors, they did demonstrate the entropic penalty of the proposed binding mechanism is unfavorable. Finally, we propose to use hydrogen/deuterium exchange (HDX) - mass spectrometry techniques to obtain new structural information and further characterize the molecular recognition between HIVp and dimer inhibitors. HDX can provide the first structural evidence defining the mechanism of HIVp dissociative inhibition by small molecules. HDX could be broadly applicable for a range of active-site and allosteric inhibitors.