Ab initio
calculations, which involve a quantum mechanical treatment of the electrons,
are used to study reactions of small molecules in the gas phase. These
methods are expensive for small chemical systems and intractable for larger
ones, so they are not useful for reactions that occur in condensed phases
or involve large molecules. Unfortunately, this includes most reactions
of interest, in particular those that occur in enzymes.
A method of growing interest for treating reactions in enzymes is based on the hybrid approach of treating atoms near the active site (where reaction occurs) quantum mechanically, while treating distant atoms with a more approximate, classical or "molecular mechanical" method. Such methods are termed QM/MM. QM/MM methods have been applied to reactions in solution and in enzymes. Of ultimate interest to me is application of such methods to an enzyme currently of great interest, HIV-1 protease, pictured below, and in fact I have published an initial feasibility study (see publications below). It has become clear that the treatment of the interface between the QM and MM regions in these hybrid methods is critical, and promising new strategies for the interface have been proposed. I am currently involved in testing these strategies on model systems of small molecules to establish their accuracy. When we have a good method, we will proceed to reactions in enzymes.