The meaning of hydrophobicity.

K. P. Murphy et al. (1, 2, 3) recently proposed what would appear to be a profound change in our way of thinking about hydrophobic interactions. On the basis of their elegant and careful experiments on small molecules (4) and on proteins (5), they propose that "for globular proteins hydrophobic interactions lead to destabilization" (1) and that "the hydration effect of nonpolar solutes stabilizes the dissolved state and thus in itself cannot be regarded as a cause of their hydrophobicity" (2). These statements would appear to conflict with the current dogma that hydration of nonpolar solutes is opposed by the entropy of water ordering at 25°C and that protein folding is driven by hydrophobicity. It appears, however, that the statements of Murphy et al. illuminate a specific aspect of nonpolar solvation and do not contradict the traditional dogma. The apparent contradiction can be traced to the use by Murphy et al. of a novel meaning for the terms "hydrophobic interactions" and "hydration effect," defined by a specific hypothetical process they describe. Several times in the past there have been disagreements about the meaning of the terms "hydrophobicity," "hydrophobic effect," and "hydrophobic hydration" (6). There appear to be at least three distinct meanings ofthese terms: (i) "Hydrophobic" has been used to refer to any transfer of a nonpolar solute to any aqueous solution. (ii) Altematively, it has been used more specifically to refer to transfers ofnonpolar solutes into aqueous solution when a particular characteristic temperature dependence is observed, as noted below. These two meanings describe experimental observations and make no reference to a particular molecular interpretation. (iii) "Hydrophobicity" has also been used to refer to particular molecular models, generally involving the ordering ofwater molecules around the nonpolar solute. Whereas the common usage of the term now appears to be definition (ii) (7), Murphy et al. (1) appear to have used a particular variant of definition (iii). Hydrophobicity is as good a term as any with which to describe the unusual temperature dependence ofthe solvation ofnonpolar solutes in water (ii). The unusual feature is that nonpolar solvation in water is strongly opposed by entropy at around 25°C and has a large positive heat capacity change. Simpler solvation processes are instead opposed by enthalpy over a wide range oftemperatures and have small heat capacity changes. This is the feature of nonpolar solvation first identified by Butler (8) and by Frank and Evans (9) that merits special terminology. Definition (i) needs no special term because it describes otherwise ordinary solution processes. There are also two problems with using meaning (iii) to define hydrophobicity: the molecular mechanism is not yet fully understood, and "water ordering" is an appropriate description of the entropic explosion of nonpolar solutes at 25°C, but not over a broader temperature range (2). Hence I believe the most sensible use of"hydrophobicity" is in situation (ii); it would therefore be simply an operational definition of an experimental observation. The apparent contradiction can be traced to a particular, novel meaning that Murphy et al. attribute to "hydrophobic interactions" and "hydration effect." According to meaning (ii), "hydrophobicity" describes a complete transfer process as measured in partitioning experiments: (a) removal of the solute from the pure medium, with the breaking of solute-solute bonds; (b) closing the cavity therein; (c) creating a cavity in water; and (d) making the solute-water bonds. This is an experimentally measurable process. Murphy et al., however, define "hydrophobic interactions" and "hydration effect" in terms of the free energy of creating a cavity in water minus the corresponding free energy at a special temperature, T = T where the transfer is least favored. They obtain this hydration free energy by subtracting the transfer enthalpy at T, from the total transfer free energy. [Murphy et al. state that the transfer enthalpy at Ts equals the vaporization enthalpy, but that is not completely correct; rather it equals the oilwater transfer enthalpy at T. and includes the enthalpy of solute-water interactions, (d), in addition to (a) and (b).] Using this hypothetical "compact gas" reference state (2), Murphy et al. make the point that, relative to T7,, the opening of the water cavity at other temperatures is more energetically favored than it is entropically disfavored. In short, Murphy et al. express the free energy of transfer as (2):