Structure and mechanism of action of nonnucleoside inhibitors of HIV-1 reverse transcriptase: strategies to combat drug resistance

Abstract

In the past few years, drug research has focused on three HIV-1 enzymes, reverse transcriptase (RT), protease, and integrase. In the case of RT, a number of potent inhibitors have been discovered. These can be classified into two distinct groups, nucleoside analogs and nonnucleoside inhibitors; however, mutations in RT have allowed the virus to develop resistance to all of the known drugs. In order to better understand the interactions between amino acid residues in the protein and nonnucleoside inhibitors, a computer model of the nonnucleoside inhibitor binding pocket of RT has been developed, using a subset of amino acid residues surrounding the pocket. The results of molecular mechanics minimizations of three RT/nonnucleoside inhibitor complexes showed that the resultant total energies of complexation (binding energies) correlated with EC 50 values if and only if the calculations were carried out using coordinates from the cognate complex while allowing for adjustments of the protein relative to the inhibitor. If a model was constructed using only the crystal data of one particular RT/inhibitor complex (RT/8-Cl TIBO), the calculations did not correctly order the other inhibitors. The difficulty in devising such a “generic” model nonnucleoside binding site in HIV-1 RT is likely due to the inherent flexibility of the enzyme. A comparison of the structure(s) of HIV-1 RT in complexes with different nonnucleoside inhibitors shows that the enzyme readily adapts to the shape of each inhibitor upon complexation. In contrast to the side-chain residues of HIV protease, the amino acid residues surrounding the binding pocket in RT adopt geometries that are unique to each bound inhibitor, adopting positions that make tight van der Waals contacts. Accompanying these changes at the site where the inhibitor binds are alterations in the geometry of the nearby polymerase active site. These changes can be conveniently monitored by measuring the increase in the distance between residue G231 in the RT primer grip region and aspartyl residues (D110, D185, and D186) in the polymerase active site. The magnitude of the change in this distance correlates inversely with inhibitor EC 50, suggesting a possible mechanism of action of the drugs. Calculations using a site where various amino acids residues were changed to simulate mutations in RT that induce resistance to the nonnucleoside inhibitors revealed that a combination of less favorable inhibitor–protein interactions and slight geometry changes in the polymerase active site are responsible for the decreased effectiveness of the inhibitors against mutant RTs. The modeling results are discussed with regard to both the mechanism of inhibition as well as application of these insights to strategies for the design of better nonnucleoside inhibitors.