Buried surface area, conformational entropy, and protein stability

In this paper, how to calculate the loss of conformational entropy owing to protein folding from data on protein stability, the disulfide bonding pattern and the buried surface area is suggested. The average values of the initiation constant and entropy loss per residue for the helix‐coil transition, calculated according to the suggested approach, are in a good agreement with available data. It is shown that the loss of conformational entropy for proteins and protein fragments, thus calculated, can be quite accurately approximated by three straight lines, each representing the loss of conformational entropy in a given range of molecular weights. The first line corresponds to the formation of α helices, the second to proteins with molecular weights below 10,000, and the third to proteins with molecular weights above this value. The standard error varied from 1.3 kcal/mol, for the first region, to between ∼3 and 5 kcal/mol for the third. Use of these approximating linear functions is proposed for the calculation of protein stabilities from x‐ray data. Such calculations, performed for more than 20 proteins and their fragments, have led to at least qualitative agreement between the calculated stabilities and available data for more than 90% of the cases. While quantitative predictions are at the limit of the method's accuracy, predictions of the probabilities of proteins or their fragments to be stable are expected to give a correct answer in ∼95% of the cases.