A portion of the mercuric ion reductase enzyme, with electron transfer co-factors shown in blue and an internal pair of cysteine residues in red. A critical aspect of the hypothesized mechanism for the enzyme is the flexibility of the c-terminal helix and loop, which allows an outer pair of cysteines to bind mercuric ion from the surroundings and pass it to the inner pair of cysteines for oxidation. The crystal structure c-terminal helix and loop are shown as the white ribbon structure, with the outer pair of cysteins shaded green. One structure from a molecular dynamics simulation is shown as the cyan ribbon, illustrating the flexibility of this region.

Peter Kollman, University of California, San Francisco

Research Objectives
(1) To study the early stage of the folding processes of small proteins. (2) To identify important folding intermediates by limited folding simulations using locally enhanced sampling (LES). (3) To refine and evaluate the free energy of structure predictions for small proteins.

Computational Approach
We use the AMBER molecular mechanics simulation program suite and Gaussian quantum mechanical calculation packages.

Accomplishments
We have significantly improved both the single-CPU performance and the scalability of the molecular dynamics code in AMBER. Overall, we have achieved a factor of 6 speedup over the existing code and significantly improved the scalability.

We completed a series of microsecond-scale molecular dynamics simulations on small proteins, including a full microsecond and two 200 ns simulations on the villin headpiece and four 200 nsec simulations on BBA1 that started from fully unfolded states. These simulations marked the beginning of direct simulations of the folding process with detailed all-atom representations of both protein and solvent that may help us to achieve a full elucidation of protein folding mechanisms.

From these simulations, we identified a highly native-like marginally stable folding intermediate as well as other compact intermediate states whose radii of gyration (i.e., size) are comparable to or smaller than that of the native state. This suggests the existence of multiple compact intermediate states that may play roles in the folding process and supports the notion that the barrier separating the folded native state and the unfolded (or partially folded) non-native states may be entropic.

On the other hand, all four folding simulations on BBA1 yielded a similar structure in which the helical secondary structure formed early and was maintained throughout the remainder of the simulations. This observation suggests that the folding process of this protein may follow a simple secondary-tertiary mechanism in which stable secondary helical structures form in the early stages and the completion of the folding process is marked by the formation of the tertiary contacts between these secondary structures. This scenario is distinct from that of the villin headpiece, in which the secondary structures were only partially formed even when the tertiary contacts started to form. Taken together, the results suggest diverse folding mechanisms.

Significance
Elucidation of the mechanism of protein folding has remained a scientific challenge for decades. Molecular dynamics simulation with full representation of solvent possesses a unique advantage to study protein folding due to their atomic level resolution and accuracy.

Publications
Y. Duan and P. A. Kollman, "Pathways to a protein folding intermediate observed in a 1-microsecond simulation in aqueous solution," Science 282, 740 (1998).

C. Simmerling, J. L. Miller, and P. A. Kollman, "Combined locally enhanced sampling and Particle Mesh Ewald as a strategy to locate the experimental structure of a nonhelical nucleic acid," J. Am. Chem. Soc. 120, 7149 (1998).

Y. Duan and P. A. Kollman, "Toward a realistic simulation of the folding process of small proteins," J. Phys. Chem. (submitted).