Hybrid Quantum Mechanical and Molecular Mechanical Studies of the Reaction Mechanism of Lactate and Malate Dehydrogenase

Principal Investigator

Jill E. Gready

Division of Biochemistry and Molecular Biology

John Curtin School of Medical Research

Enzymes are proteins which are responsible for virtually all chemical reactions in cells. They bind reactants (substrates) with a high degree of specificity, and chemically transform them with a phenomenal rate acceleration compared to the analogous reaction in solution. It is of fundamental interest to characterize the unique structural and energetic properties of the enzyme active site which enable this binding and catalysis. This project focuses on the reaction mechanism of two important glycolytic enzymes, lactate dehydrogenase (LDH) and malate dehydrogenase (MDH). LDH converts pyruvate to L-lactate in the presence of the cofactor NADH; similarly, MDH converts oxaloacetate to L-malate. The relatively new hybrid quantum mechanical and molecular mechanical (QM/MM) technique is employed in this study. The computational cost of modelling a chemical reaction in an enzyme is minimized by partitioning the system: the active site and the substrate are treated quantum mechanically, while the protein environment is simulated using classical molecular mechanics. The two regions are coupled via approximate, parameterized interactions. The QM/MM approach thus takes advantage of the chemical accuracy of quantum mechanics for the active site, while including the influential environmental effects using the computationally less expensive MM. Few enzyme systems have been studied with this method, and much work remains to improve the protocol, parameterization, and reliability of the technique. Thus the aim of this research is to develop protocols and simulation conditions for hybrid QM/MM calculations of LDH and MDH in an effort to determine the enzymatic reaction mechanism.


Rebecca K. Schmidt

Division of Biochemistry and Molecular Biology John Curtin School of Medical Research


v53 - VPP, PC

What are the results to date and the future of the work?

This project has been started by setting up and refining conditions for molecular dynamics (MD) simulations of three systems:

(1) LDH/NADH/oxamate (original X-ray crystallographic structure)

(2) LDH/NADH/pyruvate (reactant complex)

(3) LDH/NAD+/lactate (product complex)

- Appendix A

The purpose of these MD simulations is two-fold. Firstly, they will provide a relaxed starting structure for future QM/MM calculations. Secondly, these calculations are less computationally costly than QM/MM simulations, so it is best to optimize the protocol for the MD region using these simulations. One of the central tasks undertaken since September 1996 was determining the parameters used in the empirical energy function. Standard AMBER and AMBER/OPLS force field values are used for the protein. Parameters for the cofactor NADH/NAD+ are obtained from previous work in the Gready group, with some modifications made to match current protein force fields. For the substrates (pyruvate and lactate) and the inhibitor (oxamate), a series of ab initio calculations were required to determine partial charges for MD simulations. Surprisingly, the isoelectronic molecules pyruvate and oxamate display different conformational preferences. This difference impacts on the choice of torsional parameters for the MD simulations, and may well be significant in the reaction mechanism, so further work is progressing on addressing this matter.

Using these parameters, a series of MD calculations of the three systems have been performed, and the following issues were explored:

· comparison of AMBER and AMBER/OPLS force fields - effect of varying the cutoff distance for nonbonded interactions - comparison of restraining or not restraining the exterior region of protein - effectiveness of different protocols for minimization of initial crystallographic structure before MD begins - refinement of definition of restrained portion of protein and of placement of solvating water

· studies of the validity of including counterions in protein to balance charged amino acids.

Many more of these MD simulations are required to optimize the protocol and then finally obtain the best initial starting structure for the QM/MM calculations. Once a suitably relaxed structure is obtained, the development of QM/MM protocol will begin. These calculations will be even more computationally demanding, but many will have to be performed in order to optimize conditions unique to the QM/MM method (i.e. effectiveness of geometry optimization methods; effect of different partitioning of the system into QM and MM regions; and QM/MM interaction parameters).

What computational techniques are used?

The GAUSSIAN94 package is used to perform ab initio quantum mechanical calculations. Amber 4.1 is used to perform standard molecular dynamics (MD) simulations and to analyze results, such as determining root mean square (rms) deviations from the crystallographic structure or measuring time histories of torsions and interatomic distances. The QM/MM calculations will employ the locally developed MOPS program as well as the ChemShell package, currently being ported and modified as part of the Fujitsu Area 3 project.