LAUSR

As proteins perform most cellular functions, quantitative understanding of protein energetics is required to gain control of biological phenomena. Accurate models of native proteins can be obtained experimentally but the lack of equally fine models of unfolded ensembles impedes the calculation of protein folding energetics from first principles. Here we show that an atomistic unfolded ensemble model, consisting on a few dozen conformations built from a protein sequence, can be used in conjunction with an X-ray structure of its native state to calculate accurately by difference the changes in enthalpy and in heat capacity of the polypeptide upon folding. The calculation is done using Molecular Dynamics simulations and popular force fields and water models and, for the two model proteins studied (barnase and SNase), the results agree within error or are very close to their experimentally determined properties. The enthalpy sampling of the unfolded ensemble is done through short 2-ns simulations that do not significantly modify the representative distribution of Rg of the starting conformations. The impressive accuracy obtained opens the possibility to investigate quantitatively systems or phenomena not amenable to experiment, and paves the way for addressing the calculation of protein conformational stability (i.e. the change in Gibbs energy upon folding), a central goal of Structural Biology. So far, these calculated enthalpy and heat capacity changes, combined with the experimentally determined melting temperatures of the corresponding protein, allow to reproduce the stability curves of both barnase and SNase.