The Many Faces of Structure-Based Potentials: From Protein Folding Landscapes to Structural Characterization of Complex Biomolecules - Computational Modeling of Biological Systems - page 44

simulations. Constricted conformations like polypeptide slipknots, found in coarse-
grained models, are shown to be sterically possible with the AA geometry [ 43 ].
Lastly, the AA geometry allows a clear way to add perturbative nonnative chemical
effects like hydrogen bonding [ 3 ] and partial charges.

4 Applications

SBMs are being applied to diverse problems, and in the remaining sections we
describe a representative sample of how perfectly funneled SBMs are currently in
use. In each case, the SBM can be constructed and implemented using SMOG and
GROMACS. In Sects. 4.1 - 4.3 , molecular dynamics is used to describe a system
at thermodynamic equilibrium. In this case, it is necessary to adequately sample
configuration space until the quantities of interest have converged. Finally, in
Sect. 4.4 molecular dynamics is used to find deep energetic minima in perturbed
structure-based potentials for molecular modeling applications.

4.1 Folding

4.1.1 Protein Folding

The most-established application of SBM is to the study of protein folding.
Determining the TSE, the shape and size of free energy barriers, and the existence of
folding intermediates are all topics of interest. Figure 5 shows the result of AA SBM
folding simulations for two of the most thoroughly studied proteins, chymotrypsin
inhibitor-2 (CI2) and the SH3 domain. These two proteins are two-state folders,
meaning the protein only populates two basins spanned by a cooperative transition.
Figure 5 a,d shows representative traces of Q versus time during constant tem-
perature molecular dynamics near folding temperature T F . T F is the temperature
such that the folding and unfolding basins are equally populated. Simulations are
performed at T F because it maximizes the sampling rate of the folding transition. T F
is determined by running simulations at high and low temperatures, and iteratively
converging on a temperature where both folding and unfolding is observed. Q is
defined as the fraction of native residue pairs with at least one atom-atom contact
within 1.2 times its native separation. Alternative definition of Q, such as the
fraction of atom-atom contacts formed, may shift the locations of basins in the
resulting free energy landscape, but will preserve the heights of any barriers.
Q traces from long molecular dynamics trajectories at various temperatures can
be combined using weighted histogram analysis (WHAM) [ 31 ], to obtain an optimal
density of states. The density of states can then be used to extrapolate F.Q/ at any
temperature (Fig. 5 b, e). Always, care must be taken to ensure that the trajectories