Zusammenfassung
The present work is concerned with the theoretical investigation of enantioselective ester hydrolysis as catalysed by the enzyme lipase A of Bacillus subtilis. A model consisting of enzyme, substrate, and solvent is described at the atomistic level with a combined quantum-mechanical (QM) and molecular-mechanical (MM) approach. Density-functional theory (DFT) is used for the QM part and the Charmm22 force field for the MM part of the model.
According to kinetic investigations of lipases and serine hydrolases, the discrimination of enantiomers occurs during the acylation steps of the reaction. It is commonly accepted that this step proceeds via an instable tetrahedral intermediate, whose existence has not yet been proven experimentally for lipases. One part of this work is therefore concerned with the theoretical characterisation of this intermediate. To this end, Charmm22 force field parameters were derived on the basis of ab initio data.
The enantioselectivity of the lipase-catalysed ester hydrolysis was analysed exemplarily for the chiral substrate 1-(2-naphthyl)-ethyl-acetate. Different binding modes of the substrate in the active site were identified initially using MM-based molecular dynamics (MD) simulations. Starting from the tetrahedral intermediate, both covalent steps of the acylation reaction (from the Michaelis complex via the intermediate to the acylenzyme) were treated at the QM/MM level using two different methodological approaches. On the one hand, reaction paths and stationary points on the QM/MM potential energy surfaces were determined by geometry optimisation, from which activation energies, ∆E‡, of the acylation reaction were derived. On the other hand, the free energies of activation, ∆G‡, were computed using the umbrella sampling method. The quantitative analysis of enantioselectivity was performed with a stochastic kinetic model of the kinetic resolution experiment, which accounts for the competition between the enantiomers. The barriers and rate constants computed at the QM/MM level served as input parameters for the kinetic model.
The results mirror the complexity of the enzymatic reaction. It is found that the substrate can bind in different orientations and that the potential-energy curves for the acylation reaction are strongly dependent on the binding modes and the starting geometries. The tetrahedral intermediate is often a shallow minimum, and is even nonexisting in some reaction paths, so that technically sound results can only be obtained using umbrella sampling simulations. In this manner the experimentally determined enantiopreference for the R-enantiomer is reproduced, however, with a too low E value. The QM/MM investigations are therefore able to provide qualitative insights into the mechanism, but not quantitative predictions of enantioselectivity.
