Comment on: 'The entropy cost of protein association'.

An understanding of the origin of the entropy loss involved in association processes (protein‐protein or protein‐ligand) is important for an interpretation of the many equilibria that play a role in biology. Although the total entropy change on association can be determined by measuring the transition temperature and associated enthalpy change (Privalov et al., 1995), interpretation of the results is hindered by the fact that there are several partially cancelling contributions [Tidor and Karplus, 1993 (TK)]. The loss of three translations and three (two for a linear molecule) rotations is entropically unfavorable. This can be offset by favorable entropic contributions, including those that arise from the hydrophobic effect, changes in the protonation states, solvent and counter ion release, as well as the presence of six (five for a linear molecule) new vibrational degrees of freedom. Because the translational and rotational contributions to the entropy of binding appear to be very simple to calculate, they have often been discussed. However, there are no direct measurements of their contribution to binding and theoretical estimates vary by more than an order of magnitude. A recent paper by Tamura and Privalov (TP) (1997) purports to make a direct measurement of the translational/rotational entropy contribution to the formation of a dimer of the Streptomyces subtilisin inhibitor (SSI). TP studied the temperature-induced unfolding of the wild-type dimer and of a mutant (D83C), with a disulfide cross-link between the subunits, by differential scanning calorimetry; the measurements were made as a function of concentration. When normalized to the same temperature, the entropy of unfolding of the native SSI dimer adjusted t oa1M standard state was found to be approximately the same as that of the cross-linked mutant dimer; the measured difference was ‐5 6 4 cal/mol·K. This result is very different from that obtained in an earlier calorimetric study of the same system by Tamura et al. (1994), who found a value of about ‐100 cal/mol·K for the difference of the entropies of unfolding of the two systems. Although the difference between the two sets of results is of some concern, TP have given cogent argument to support their measurements (TP, 1997; Privalov, private communication). We do not feel we have the expertise to comment on this question and consider only the interpretation of the TP result in what follows.