Embargo until
Journal Title
Journal ISSN
Volume Title
Johns Hopkins University
For over fifty years, the unfolded state of proteins had been thought to be featureless and random. Experiments by Tanford and Flory confirmed that unfolded proteins possessed the same dimensions as those predicted of a random flight chain in good solvent. In the late eighties and early nineties, however, researchers began to notice structural trends in unfolded proteins. Some experiments showed that the unfolded state was very similar to the native state, while others indicated a conformational preference for the polyproline II helix in unfolded proteins. As a result, a paradox developed. How can unfolded proteins be both random and nonrandom at the same time? Current experiments and most theoretical simulations cannot characterize the unfolded state in high detail, so we have used the simplified hard sphere model of Richards to address this question. By modeling proteins as hard spheres, we can not only determine what interactions are important in the unfolded state of proteins, but we can address the paradox directly by investigating whether nonrandom behavior is in conflict with random coil statistics. Our simulations identify hundreds of disfavored conformations in short peptides, each of which proves that unfolded proteins are not at all random. Some interactions are important for the folded state of proteins as well. For example, we find that an α-helix cannot be followed directly by a β-strand because of steric considerations. The interactions outlined here limit the conformational possibilities of an unfolded protein far beyond what would be expected for a random coil. For a 100-residue protein, we find that approximately 9 orders of magnitude of conformational freedom are lost because of iii local chain organization alone. Furthermore, we show that the existence of this organization is compatible with random coil statistics. Although our simulations cannot settle the controversy surrounding the unfolded state, we can conclude that new methods of characterizing the unfolded state are needed. Since unfolded proteins are not random coils, the methods developed for describing random coils cannot adequately describe the complexities of this diverse structural ensemble.