Hydrophobic interaction

The Structural Basis of the Hydrophobic Interaction

Many crucially important processes take place in water. Processes that are central to not only much of chemical processing, but also in biological processes. Yet our understanding of the role water plays remains poorly understood.

If we want to understand the influence of water on processes at the molecular level, we need to be able to study structures at that level. For crystals, this is relatively straightforward: the fact that we can arrange molecules - however complex - in a regular repeating arrangement in a crystal means we can exploit the power of crystallography in solving structures from simple crystals to proteins and viruses. Where - as in a liquid - we have no crystalline order, we have to use other techniques.

The ability to probe experimentally the local structural environment of hydrophobic solutes is a relatively recent development. Neutron scattering with isotopic substitution, first established by Enderby and coworkers from the 1960s (see for example Enderby J.E. and Neilson G.W. in Water: A Comprehensive Treatise, ed. Franks F. 6, 1 (1979)). The relatively large difference in the neutron scattering lengths of hydrogen and deuterium has enabled detailed structural studies of aqueous systems. In appropriate binary molecular mixtures, this can allow us direct access to the three intermolecular pair correlations of structural importance, namely the solute-solute, solute-solvent and solvent-solvent correlations. In addition, the subsequent development of efficient and reliable computational methods for the extraction of intermolecular orientational correlation functions from the neutron data has made feasible the interpretation of neutron scattering data on more complex aqueous systems in terms of detailed and experimentally consistent structural models.

We have been particularly interested in the way in which non-polar groups on molecules interact with each other in water - the so-called hydrophobic interaction, in which non-polar groups such as alkyl groups tend to come together in aqueous solution. This effect is certainly of great importance in maintaining the structural integrity of proteins and biological membranes. How it works, however, remains controversial.

Since the studies on the thermodynamics of non-polar solvation by Frank and Evans in 1945 [J. Chem. Phys. 13 507-532 (1945)], the standard model for the hydration of non-polar solute species has emphasised the enhancement of the solvent water structure hydrating the solute. The exact nature of this structural ordering has remained unclear, but that the non-polar solvation process results in a net loss of entropy is not disputed. This entropic mechanism for association of non-polar moieties subsequently became central to models of protein folding and stability following the seminal work of Kauzmann in 1959 [Adv. Prot. Chem. 14 1-63 (1959)], who argued that a structural interpretation of the thermodynamic data could be developed by considering the hydration of the polar and non-polar molecular groups that constitute proteins.

Over the past few years, we have been using both neutron scattering and EXAFS spectroscopy to try to understand the structural nature of the hydrophobic interaction. For the neutron work, which has been carried out at both the Institut Laue-Langevin, Grenoble, France, and the ISIS Pulsed Spallation Neutron Source at the Rutherford Appleton Laboratory in the UK, we have focussed on alcohol-water systems.

These are an important class of solvent media that involve a range of intermolecular interactions. The thermodynamics of these solutions at low alcohol concentrations show that the non-polar alkyl residues dominate the solution properties (see for example Franks F. and Desnoyers J.E. Alcohol-Water Mixtures Revisited Water Sci. Revs. 1,171-232 (1985)).

Initial work on methanol-water solutions focussed on the nature of the hydration shell of the non-polar methyl group (Soper and Finney 1993). This work showed for the first time that the water molecules in the hydration shell tended to be oriented with their dipole moments oriented tangentially to a shell centred on the methyl carbon, as would be expected for the 'standard model' of hydrophobic hydration, which argues for a cage-like, or 'clathrate' structure around the non-polar group (see figure 1). This work did however demonstrate that this hydration structure was considerably more disordered than suggested by the standard model. (It is after all a liquid!)

Figure 1

Using isotope substitution on the hydrogens of the water molecules, we were also able to compare the structure of the water in the hydration shell with that in pure water. Here, the 'standard model' would have us believe that the hydration shell water is 'more ordered' (whatever that means) than is the water in the bulk. As shown in figure 2, which shows the structure of the water from the hydrogen's viewpoint both in the hydration shell and in pure water, there is no significant difference. An unexpected surprise. This result has since been confirmed in a number of other systems.

Figure 2: The hydrogen-hydrogen correlation function for a 1:9 ratio methanol/water mixture (line) compared to the same function for pure water (circles). The intramolecular peak at r ~ 1.5 Å, the hydrogen bond peak at r ~ 2.3 Å, and a third characteristic peak near r ~ 33.8 Å are all clearly visible in both cases.

What then is the source of the entropic contribution to the hydrophobic driving force if it isn't an ordering of the water in the hydration shell, as suggested in the standard model? We think we are closing in on what may be at least part of the answer. This comes from some very recent work on tertiary-butanol - water solutions.

Tertiary butanol, being the largest fully water miscible monohydric alcohol, is a particularly excellent model for investigations of hydrophobic hydration and associated effects. In earlier investigations of this system, we found a preference for intermolecular association of the alcohol molecules through their non-polar groups rather than through hydrogen-bonding interactions. Moreover, as found in the earlier methanol - water work, the water molecules that directly solvate the non-polar regions of the alcohol molecule preferentially adopt configurations in which their dipole moments are oriented tangentially to the non-polar surfaces.

Our most recent work first of all confirms the structure of the first hydration shell is as was found previously, reflecting the intrinsic nearly-tetrahedral geometry of a hydrogen bonded water network. However, beyond the first shell of solvent water neighbours, differences begin to appear. In going from pure water to the t-butanol solution, there appears a small but significant reduction in the radial position of the second shell of neighbours. Beyond the first neighbour shell, the solvent becomes slightly more densely packed in the solution.

The focus of the majority of earlier structural work on the structures of solutions containing non-polar groups has been on the first hydration shell. In a range of systems the conclusion has been drawn that there is no measurable perturbation relative to bulk water for the water in the first shell, raising serious questions concerning the standard interpretation of the structural basis of the entropic driving force of the hydrophobic interaction. The finding of these most recent results of a small yet significant change in the solvent water network beyond the first neighbour shell is both consistent with these earlier results whilst additionally telling us precisely where a change in the ordering of water, such as that postulated in the standard model, can be found.

These results have major implications for our understanding of non-polar hydration phenomena, and of hydrophobic interactions in particular. The previously dominant view that hydrophobic processes are governed by perturbation of the near-neighbour hydration structure would thus appear to be an insufficient basis for understanding or modelling these effects. The work presented here underlines the subtle nature of the perturbation of water structure that has important consequences for both chemical processes and biological systems.

See also a recent Nature Science Update write-up of some of our ESRF-based X-ray work.

Some selected references

Soper A.K. and Finney J.L. Hydration of Methanol in Aqueous Solution Phys. Rev. Lett. 71, 4346-4349 (1993)

Finney J.L., and Soper A.K. Solvent structure and perturbations in solutions of chemical and biological importance. Chem. Soc. Revs. 1994, 1 - 10, (1994)

Turner J., Soper A.K. and Finney J.L. Water Structure in Aqueous Solutions of Tetramethylammonium Chloride. Mol. Phys. 77, 411-429 (1995)

Bowron D.T., Filipponi A., Lobban C. and Finney J.L. Structural determination of the hydrophobic hydration shell of Kr". Phys. Rev. Letts, 79, 1293-1296 (1997)

Bowron D.T., Finney J.L., and Soper A.K. Structural investigation of solute-solute interactions in aqueous solutions of tertiary butanol. J. Phys. Chem. 102, 3551-3563, (1998)

Bowron D.T., Finney J.L., and Soper A.K. The structure of pure tertiary butanol, Mol. Phys. 93, 531-543 (1998)

Bowron D.T., Filipponi A., Lobban C. and Finney J.L. Temperature induced disordering of the hydrophobic hydration shell of Kr and Xe. Chem. Phys. Lett. 293, 33-37 (1998)

Bowron D.T., Filipponi A., Roberts, M.A., and Finney J.L, Hydrophobic hydration and the formation of a clathrate hydrate. Phys. Rev. Letts 81, 4164-4167 (1998).

Bowron D.T., Finney J.L. and Soper A.K. Structural Investigation of Solute-Solute Interactions in Aqueous Solutions of Tertiary Butanol J. Phys. Chem. 102, 3551-3563 (1998)

J.L. Finney, "The Structural Basis of the Hydrophobic Interaction". In 'Hydration Processe in Biology', ed. M.-C. Bellissent-Funel, IOS Press 1999. pp 115-124.

Other interesting relevant work

De Jong P.J.K., Wilson J.E., Neilson G.W. and Buckingham A.D. Hydrophobic Hydration of Methane Mol. Phys. 91, 99-103 (1997)