Computer simulations of biomolecular structure and dynamics using a vibrational empirical potential energy function

Abstract

The main theoretical methods which are currently being applied to the determination of protein structure and fluctuations, enzyme-substrate, enzyme-inhibitor or protein-drug interactions are Monte Carlo simulations, molecular dynamics, minimization techniques and normal mode analysis. Using these methods it is possible to understand the behavior of complex biological macromolecules in terms of fundamental molecular forces at a level which is generally inaccessible to experimental techniques (X-rays, NMR) alone. At the most fundamental level, prediction of physical properties for molecular systems involves the direct solution of the Schrodinger equation for the nuclear and electronic degrees of freedom. Such studies are referred to as ab initio calculations. They rapidly cannot be used for systems containing more than 15–20 atoms heavier than hydrogen. To investigate molecules with up to one hundred atoms, it is necessary to invoke various additional approximations and to use simplified molecular hamiltonians with parameters extracted from experiments (semi-empirical quantum mechanical calculations). For macromolecular systems the so called Molecular Mechanics method has to be used. In this approach the Born-Oppenheimer approximation is applied to solve for the electronic energies at fixed nuclear positions and to treat the electronic energies as the potential energy field for the nuclear motions. Then an analytic empirical potential energy function may be used to approximate the way in which the molecular energy changes with the coordinates of the atoms. The empirical fit of the potential energy surface is called the force field.