MOLECULAR MODELING AND MOLECULAR MECHANICS 167
fi eld.^19 The CHARMM force fi eld (more information at http://yuri.harvard.
edu/ ) was used in a study of the temperature dependence of both the structure
and the internal dynamics of (carbonmonoxy)myoglobin.^20 The geometry
about the iron center was assumed to be octahedral. Analysis of the move-
ments of the iron center with respect to the heme group indicated that the
largest - amplitude motions were perpendicular to the heme plane.
A major problem in modeling of large biomolecules arises from the fl exibil-
ity of proteins and DNA with corresponding numbers of adoptable geometries.
The experimental system modeled is therefore only one possible representa-
tion of the many possible geometries. A second diffi culty is that the accuracy
of molecular mechanics models of biomolecules is substantially lower than
that of small molecules. Large numbers of independent parameters are needed,
but there are few experimentally known structures and these are often of low
precision. Electrostatic considerations and solvent effects cause further limita-
tions. The real value in molecular modeling of macromolecular systems, as
stated by the reference 3a authors, emerges when the models make predictions
that can be tested experimentally. Qualitatively, the models can be used to
visualize molecules whose structures are not accessible by any other means.
In Part 3 of reference 3a, the authors give much needed advice on develop-
ing a force fi eld, taking into account bond length deformation, valence angle
deformation, torsion angle deformation, out - of - plane deformation, van der
Waals and electrostatic interactions, and hydrogen bonding interactions. All
of these parameters are interrelated, and modifi cation of one must lead to
further testing of all. They reiterate that the force fi eld parameter set must
model the molecule under consideration as accurately as possible and that
results of the calculation should be compared to experimental data whenever
this is available. The authors then discuss carrying out the calculation, fi rst
listing the important considerations. These include: having an adequate start-
ing model, choosing an appropriate energy minimization method, and consid-
ering the probability that there are many possible energy minima.
4.3.4 A Molecular Modeling Descriptive Example,
Ubiquinol/cytochrome c oxidoreductase – cytochrome bc 1 complex, a protein
found in electron transport chain of mitochrondrial membranes, bacteria, and
chloroplasts, is the subject of Section 7.6. X - ray crystallographic structures of
the native protein, and those with inhibitors in place, are discussed in Section
7.6.2. One set of crystallographic studies, carried out by the E. A. Berry group,^21
(PDB: 1BCC, native, and 3BCC, stigmatellin bound) showed substantial move-
ment of the soluble head of the Rieske iron – sulfur protein (ISP), containing
an [Fe 2 S 2 ] cluster, between reaction domains in cytochrome b and cytochrome
c 1 protein subunits. In a 1999 publication, results of steered molecular dynam-
ics (SMD) simulations of the ISP soluble domain of native and stigmatellin
bound cytochrome bc 1 were reported.^22 The SMD technique overcomes a time -
scale limitation (of a few nanoseconds) for molecular dynamics simulations of