Modeling the archetype cysteine protease reaction using dispersion corrected density functional methods in ONIOM-type hybrid QM/MM calculations; the proteolytic reaction of papain

Abstract

A proteolytic reaction of papain with a simple peptide model substrate N-methylacetamide has been studied. Our aim was twofold: (i) we proposed a plausible reaction mechanism with the aid of potential energy surface scans and second geometrical derivatives calculated at the stationary points, and (ii) we investigated the applicability of the dispersion corrected density functional methods in comparison with the popular hybrid generalized gradient approximations (GGA) method (B3LYP) without such a correction in the QM/MM calculations for this particular problem. In the resting state of papain the ion pair and neutral forms of the Cys–His catalytic dyad have approximately the same energy and they are separated by only a small barrier. Zero point vibrational energy correction shifted this equilibrium slightly to the neutral form. On the other hand, the electrostatic solvation free energy corrections, calculated using the Poisson–Boltzmann method for the structures sampled from molecular dynamics simulation trajectories, resulted in a more stable ion-pair form. All methods we applied predicted at least a two elementary step acylation process via a zwitterionic tetrahedral intermediate. Using dispersion corrected DFT methods the thioester S–C bond formation and the proton transfer from histidine occur in the same elementary step, although not synchronously. The proton transfer lags behind (or at least does not precede) the S–C bond formation. The predicted transition state corresponds mainly to the S–C bond formation while the proton is still on the histidine Nδ atom. In contrast, the B3LYP method using larger basis sets predicts a transition state in which the S–C bond is almost fully formed and the transition state can be mainly featured by the Nδ(histidine) to N(amid) proton transfer. Considerably lower activation energy was predicted (especially by the B3LYP method) for the next amide bond breaking elementary step of acyl–enzyme formation. Deacylation appeared to be a single elementary step process in all the methods we applied.