Origin of Heat Capacity Changes in a “Nonclassical” Hydrophobic Interaction

The early work of Edsall illustrated the effect of nonpolar groups in raising the apparent heat capacities of solutes in aqueous solution. Subsequently, the hydrophobic effect has been well established as the result of the formation of structured water around such groups. While the nature of this structuring remains a topic of debate, it is universally accepted that these water molecules possess a higher heat capacity and a lower entropy than bulk water. Consequently, hydrophobic interactions, in which nonpolar surfaces are shielded from bulk water, are classically characterized by a favourable entropic binding signature together with a negative change in heat capacity at constant pressure (DCp). [3–5,7] Nonetheless, the foundations of the latter in solvent reorganisation are not universally accepted. Recently, we observed a paradoxical enthalpy-driven thermodynamic binding signature in studies on a model ligand– protein interaction, namely the association of small hydrophobic ligands within the hydrophobic binding pocket of recombinant mouse major urinary protein (rMUP). This binding signature has been observed in a number of hydrophobic molecular interactions (reviewed by Meyer et al.) and has led to the concept of the “nonclassical” hydrophobic interaction. However, the thermodynamic relationship between the “nonclassical” and “classical” hydrophobic interaction has remained obscure, especially since the former typically exhibits the negative change in heat capacity of the latter ; this suggests that the molecular basis of each lies in solvent reorganization. In the case of rMUP we found that the binding pocket is suboptimally hydrated, a phenomenon that is increasingly being ACHTUNGTRENNUNGreported in other proteins. Under these circumstances, there exists an imbalance between solute–solute dispersion interactions following the association, versus solute–solvent dispersion interactions that exist prior to the association. Moreover, the favourable entropic contribution that results from the expulsion of solvent water molecules is small, and leads to a thermodynamic binding signature that is enthalpy driven. However, a negative change in heat capacity on binding (DCbp) is nonetheless observed in rMUP–ligand interactions, and is of similar or greater magnitude to that observed in typical ligand–protein interactions (Table 1). The serendipitous discovery that rMUP binds members of the primary aliphatic alcohol series as surrogate ligands permits a rigorous quantitative assessment of the solvation contribution to DCp, since the hydration thermodynamics of these ligands are very well characterized. We measured the temperature dependence of the standard enthalpy of binding of hexan-1-ol, heptan-1-ol and octan-1-ol to rMUP using conventional isothermal titration calorimetry (ITC) experiments. A linear fit of these data (Figure 1) result-