Modeling bond correlations in denatured proteins and polypeptides

Bond-orientational correlations for finite-length homopolypeptides and a selected group of denatured proteins are obtained by numerical simulations using a polypeptide model with a potential of mean force. These correlations characterize the stiffness of the polypeptide backbone and are generally described by either an exponential or a power-law decay in the asymptotic limit. However, for finite length polypeptides and unfolded proteins the correlations significantly deviate from either single exponential or power-law behavior. A heuristic model is developed to analyze the correlations of homopolypeptides, which depends on the chain length and the side-chain properties. The model contains power-law and multi-exponential terms, the latter which could be interpreted as local persistence lengths. In the asymptotic limit, the model reduces to the expected power-law behavior. Simulations of denatured proteins show that the power-law behavior of the correlations is significantly suppressed and only the multi-exponential term is needed to model the correlations. In addition, average persistence lengths (ranging from 2.0 to 2.5 nm) are obtained from the correlations by fitting single exponentials and shown to be in general agreement with experiments, which also assume single exponential decay. © 2016 Wiley Periodicals, Inc. Biopolymers 105: 312-323, 2016.