Biomedical Engineering Reference
In-Depth Information
24
Hybrid Models; the QM/MM
Approach
It all started with a paper (Warshel and Levitt 1976) in the
Journal of Molecular Biology
,
which by and large is not read by chemists. Once again, I will let the two authors give you
their ideas in their own words.
A general method for detailed study of enzymic reactions is presented. The method considers
the complete enzyme-substrate complex together with the surrounding solvent and evaluates
all the different quantum mechanical and classical energy factors that can affect the reaction
pathway. These factors include the quantummechanical energies associatedwith bond cleavage
and charge redistribution of the substrate and the classical energies of steric and electrostatic
interactions between the substrate and the enzyme. The electrostatic polarization of the enzyme
atoms and the orientation of the dipoles of the surrounding water molecules is simulated
by a microscopic dielectric model. The solvation energy resulting from this polarization is
considerable and must be included in any realistic calculation of chemical reactions involving
anything more than an isolated atom
in vacuo
. Without it, acidic groups can never become
ionized and the charge distribution on the substrate will not be reasonable. The same dielectric
model can also be used to study the reaction of the substrate in solution. In this way, the reaction
is solution can be compared with the enzymic reaction...
What the authors did was to combine a molecular mechanics (MM) potential for the
solvent with an early (MINDO/2) quantum mechanical model for the solute. By 1998 such
hybrid methods had become sufficiently important to justify anAmerican Chemical Society
symposium (Gao and Thompson 1998).
The early hybrid models such as that of Warshel and Levitt are often referred to as
'QM/MM'. Note that there is no mention of geometry optimization, and that attention
focuses on the solute/solvent problem.
