Molecular Simulation and Systems Biology - Systems Biology: Networks, Models, and Applications - page 66

Figure 3.1 Images from a molecular dynamics simulation of the protein 1BBL
in explicit solvent. The left-hand panel shows the folded 1BBL protein [82]
(40 amino acids, 618 atoms) surrounded by 15,000 water molecules. The
central panel is a detail view showing a shell of water molecules within 5 Å of
the folded protein. The right-hand image shows a representative extended
(unfolded) structure of 1BBL. The protein and two chloride counterions are
shown as spheres, while water molecules are shown in a stick representation.
This and other molecular graphics in this chapter were generated with the
PyMOL program [83] unless otherwise noted.
use molecular simulations to calculate the lateral diffusion coefficient
for an integral membrane protein in order to understand how rapidly
such proteins move around the cell surface. Molecular simulation
methods can also be used to estimate association and dissociation
rate constants for protein-protein complexes involved in a signaling
cascade. Also, the rates of enzymatic processes relevant for systems
biology, such as phosphorylation, can be calculated from coupled
quantum mechanical/molecular mechanical (QM/MM) simulations.
QM/MM simulations use classical models in combination with a
quantum mechanical treatment of a subregion of the system.
The most probable way molecular simulations will be used to
provide parameters for simplified models is through the calculation of
free energy differences between different states of a system. Significant
advances have been made in calculating absolute protein-ligand bind-
ing free energies [1] and in estimating biomolecular binding free
energies from simulation [2-4]. New techniques [5] promise the possi-
bility of simulating a single model protein-DNA complex and using
those data to predict the optimum binding site for the transcription
factor in question. Clearly, these free energy calculations can be used
to supply missing parameters for models of biological systems, but
they also show promise in helping to elucidate the underlying
processes that must be modeled, for example, by predicting transcrip-
tion factor binding sites to help decode gene regulatory networks.
At a more detailed level, molecular simulations can be used to
develop intermediate coarse-grained models of biological systems,
where collections of tens of atoms are represented by a single “particle.”