Active site dynamics and combined quantum mechanics/molecular mechanics (QM/MM) modelling of a HIV-1 reverse transcriptase/DNA/dTTP complex

We have investigated the structure and dynamics of the HIV-1 reverse transcriptase (HIV-RT) active site, by modelling the active conformation of the HIV-1 RT/DNA/deoxythymidine triphosphate (dTTP) ternary complex. This has included molecular dynamics simulations with the CHARMM27 force field, and combined quantum mechanics/molecular mechanics (QM/MM) calculations. Three different ternary systems were studied to investigate the effects of different protonation states of the dTTP substrate (a deprotonated and two different mono-protonated triphosphate forms of dTTP at the active site), and the effects of different possible protonation state of potentially catalytic aspartate residues (Asp185 and Asp186) were tested. Several potentially important hydrogen-bonding interactions with amino acids and bound water molecules in the deoxyribonucleoside triphosphate (dNTP) binding pocket were examined. The model of the deprotonated form of the dTTP substrate with the two aspartates in their charged (basic) form seemed to be the most stable and its orientation was in good agreement with X-ray crystallographic structure. In addition, two different semiempirical (AM1/CHARMM and PM3/CHARMM) QM/MM methods were tested for the HIV-RT system, in structural optimizations. Both methods provided conformations of the triphosphate moiety in either fully deprotonated or mono-protonated forms, which agreed well with the experimental structure of dTTP. The only significant difference between the AM1/CHARMM and PM3/CHARMM minimized structures is that the PM3/CHARMM Pα–O3′ optimized distance (important for nucleotide addition) is longer by 0.66 Å in the deprotonated system but shorter by 0.37 Å in the mono-protonated triphosphate system as compared with those obtained from AM1/CHARMM minimized structure. The obtained results suggest that both of these QM/MM methods, and the stochastic boundary molecular dynamics approach applied in this work, can give reasonable results for modelling the catalytically active complex of this important enzyme. The results provide insight into the structure and interactions of the active site of this important enzyme, with implications for its mechanism, which may be useful in inhibitor design.