(362c) Factors Affecting the Cooperativity of Protein Folding Transitions

One of the most sought-after goals of computational biology is the à priori prediction of the folded state of a protein from knowledge of only the primary structure. Related to this goal is the determination of kinetic pathways of protein folding. Molecular simulation has played a prominent role in each of these endeavors, but doubt is often cast on the ability of simulation to capture the true biophysics involved. Experimentally, such small proteins usually fold through a two-state mechanism, which fact has often been used to determine how well the particular protein model and forcefield captures true protein behavior. If a specific model and forcefield yield results that demonstrate a folding mechanism involving only two states, more confidence is placed in the results. When results do not show two-state folding, the model and/or forcefield is often cited as the reason. Even when the simulation produces the expected two-state folding mechanism, criticism is still made as to the degree of cooperativity of the transition. Said criticism is one of the main arguments against the use of coarse-grain models which have become popular in recent years. One difference between the simulation protocols used to investigate protein folding and the experimental procedures is the number of protein molecules used in each. The vast majority of all protein folding studies using simulation involve only one protein molecule due to computational limitations. Contrastingly, experiments measure the aggregate of the folding behavior of a large number of protein molecules--the thermodynamic limit. The purpose of this study was to investigate how this discrepancy reveals itself in simulation results. The hypothesis of this work is that simulated proteins that exhibit non-two-state folding on a single molecule level will exhibit two-state folding at higher protein concentrations. This hypothesis is tested by simulating systems of proteins with multiple interacting molecules in the simulation box and comparing the results to a system with only a single molecule. The results indicate that among the transitions seen in the folding pathway, some are true cooperative transitions and others are only seen because the simulation is done on the single-molecule level. Moreover, as the number of molecules increases, the cooperativity of the ?true? transitions is enhanced while that of the ?single-molecule? transitions become less distinct. Overall, the results indicate that several of the criticisms laid against simulation of protein folding, as well as those against coarse-grain models, are misplaced.