Molecular Dynamics (MD) applied to complex systems such as proteins that have energetic barriers separating different configurations may suffer from slow sampling of the configuration space. Therefore, new methods are needed to improve the ability of exploring conformational space. One way to accelerate the exploration of configuration space is with various Hamiltonian replica exchange methods (HREM) whereby multiple systems differing in their Hamiltonians are run by MD in parallel and, periodically, attempts at exchanging them is made according to a Monte Carlo rule that maintains the Boltzmann distribution.

The HREM (implemented in the CUKMODY MD code designed for the efficient simulations of solvated proteins) is first used to study the conformational states of the zwitterionic form of the pentapeptide Met-enkephalin. There is a competition between open forms of the peptide driven by polar solvation of the terminal ammonium and carboxylate groups and closed forms driven by their salt-bridge formation. Normal MD started from an open state does not sample closed conformations. A small number of HREM systems were found to be sufficient to sample closed and open states a sufficient number of times to obtain potentials of mean force along various reaction coordinates. A Principal Component Analysis (PCA) shows that the first two principal modes capture more than one-half of the HREM generated fluctuations. The first mode corresponds to the end-to-end distance fluctuations and shows that the closed zwitterionic state is the predominant species. The second mode describes the presence of two conformations of similar end-to-end distance that differ in the values of neighboring psi and phi dihedral angles, because such psi/phi compensation can still produce the same end-to-end distance.

HPPK (6-Hydroxymethyl-7,8-dihydropterin pyrophosphokinase) catalyzes the transfer of pyrophosphate from ATP to HP (6-hydroxymethyl-7,8-dihydropterin). This first reaction in the folate biosynthetic pathway is an important target for potential antimicrobial agents. The conformations of the HP binding pockets of E. coli HPPK and Y. pestis HPPK are studied by HREM methods. Root mean square fluctuation (RMSF) calculated based on backbone atoms shows that loop 2 (which is close to the HP binding pocket) of YpHPPK has a much larger flexibility than that of loop 2 of EcHPPK. By using clustering methods, we find that EcHPPK and YpHPPK have residue conformations around the HP binding pocket that are are close to but distinct from those found in the crystal structures. There are near-closed conformations that have different HP binding pocket shapes in EcHPPK and YpHPPK that have potential for discriminating among ligands.

The conformational space of ATP binding to HPPK and the unbinding process of ATP from HPPK is studied using a restraint MD method and a targeted reweighting scheme (implemented in CUKMODY). ATP remains remarkably stable in its binding pocket when HPPK is driven, by using the restraint method, towards its open form. When the ATP is induced to leave its binding pocket by using a targeted reweighting method, it uses a special path that preserves the hydrogen bonds and salt bridges existing in previous stages along the path.