Research Objectives
Our goal is the complete characterization of the folding pathways of proteins, which reach their unique folded state from among an astronomically large number of possible conformations, and an understanding of their equilibrium phase diagrams.
Computational Approach
We use Monte Carlo and molecular dynamics techniques to directly simulate the equilibrium and kinetic properties of protein-like heteropolymers. These codes utilize the massively parallel processing capabilities of the T3E in two ways. First, our free energy calculations are parallelized to allow for efficient determinations of the entire phase diagram of the polymer. Second, since each folding event is microscopically different, folding pathways must be characterized statistically; the ability to simulate tens of millions of folding events in parallel is critical to the success of this project.
Accomplishments
By direct evaluation of the free energy surface of a model protein we have demonstrated the existence of a third phase of proteins, the "molten globule." The molten globule is a distinct liquid-like state that appears in the phase diagram of proteins separated by first order phase transitions from the unfolded and folded ("native") states.
For the first time, we have completely and explicitly characterized the folding pathway of a protein-like heteropolymer 48 residues long. Using a novel computational method that relies heavily on the parallel processing capabilities of the T3E, we unambiguously determined the ensemble of transition state conformations that govern the rate-limiting step of folding, and showed that folding proceeds through a molten-globule-like intermediate.
Significance
Protein folding is one of the great unsolved problems of modern biophysics. A better understanding of the protein folding problem has potential biological, biomedical, and industrial applications. An understanding of the mechanisms by which proteins naturally achieve their functional folded states can be expected to yield insights into the prediction of these folded structures directly from amino acid sequences, and to permit the design of medically useful proteins. The molten-globule state we have studied is implicated in the mechanisms of folding as well as the aggregation of improperly folded proteins, both of which have medical relevance; for example, the improper aggregation of proteins is responsible for several diseases, including Alzheimer's.
Publications
Pande, V. S., and D. S. Rokhsar. N.d. Is the molten globule a third phase of proteins? Proceedings of the National Academy of Sciences of the U.S.A., in press.
Pande, V. S., A. Y. Grosberg, T. Tanaka, and D. S. Rokhsar. N.d. Protein folding pathways: is a "new view" needed? Current Opinions Structural Biology, in press.
Pande, V. S., and D. S. Rokhsar. N.d. Transition states and intermediates in a lattice model for protein folding. Nature: Structural Biology, in preparation.
Monte Carlo simulation of a folding event. Each frame displays the average position of a
48-mer chain during a 10^4 iteration time window. The color of each bead represents the
variance of the position of the bead during this time interval, with yellow/green indicating
large fluctuations and blue indicating small fluctuations. The entire folding event takes 8 x
10^5 iterations.