The Hydrophobic Effect and the Protein Core

Modern protein engineering and drug design routinely exploit the hydrophobic effect as a quantitative design parameter, not just a qualitative concept. Structure-based drug design uses buried nonpolar surface area between a ligand and its target pocket as a first-pass estimate of binding affinity, because the same $\Delta G\approx\sigma\Delta A$ scaling that describes protein core burial also governs protein–ligand and protein–protein binding interfaces; hydrophobic "hot spots" at binding interfaces are systematically the residues whose mutation to alanine most destabilizes binding (alanine-scanning mutagenesis directly measures the cavity-plus-transfer penalty described above). Protein stabilization engineering for industrial enzymes and therapeutic antibodies frequently targets core-packing defects — filling small cavities with slightly larger hydrophobic residues, or replacing surface-exposed hydrophobic patches that risk aggregation — using exactly the surface-area/packing-density logic developed here. Computational structure prediction and design tools (Rosetta, AlphaFold-adjacent scoring functions, molecular dynamics force fields) all encode an implicit or explicit hydrophobic solvation term, typically parameterized directly from the atomic solvation parameter σ and from experimentally measured ΔCp values, because getting the hydrophobic core right is the single largest determinant of whether a predicted or designed fold is physically stable. In membrane biology, the same principle runs in reverse: transmembrane helices bury their hydrophobic surface not in a protein core but in the lipid bilayer interior, and the same $\Delta G_{tr}$ scaling with nonpolar surface area (this time favoring the lipid environment over water) underlies hydropathy-plot algorithms used to predict transmembrane segments from sequence alone. Finally, understanding of cold denaturation and pressure denaturation — both explained through the temperature and pressure dependence of hydrophobic hydration free energy — is central to formulating stable biologics and to understanding piezophilic (high-pressure-adapted) deep-sea organisms.

Ask the tutor: How does the hydrophobic effect combine with backbone hydrogen bonding to determine which secondary structures form the protein's tertiary fold?

Water is a small, highly cohesive molecule that forms on average close to four hydrogen bonds per molecule in the bulk liquid, each roughly 1.9–2.1 Å (O···H) with bond enthalpies around 10–25 kJ/mol depending on geometry. This network is dynamic — bonds break and reform on picosecond timescales — but it is essentially unbroken in the bulk: every water molecule can find hydrogen-bonding partners in essentially any direction. When a nonpolar solute is introduced, water molecules at the interface cannot hydrogen-bond across the solute's surface. They compensate by reorienting to maximize hydrogen bonding among themselves, tangential to the solute surface, forming a partial cage — a "clathrate-like" shell reminiscent of the ordered hydrate crystals formed by methane and other small gases at low temperature. This reorientation is not free: it costs orientational and translational entropy, because the number of ways the shell waters can arrange while satisfying both their own hydrogen-bonding needs and the geometric constraint of the solute surface is much smaller than in the bulk. Crucially, the enthalpy of this hydration shell is often not unfavorable — shell waters can actually have as many or more hydrogen bonds as bulk water, just more constrained in orientation. This is why dissolving a small hydrocarbon like methane in water is enthalpically nearly neutral or even mildly exothermic at room temperature, yet strongly entropically unfavorable — the classic thermodynamic fingerprint of hydrophobic hydration. The free energy cost of solvation, ΔG = ΔH − TΔS, is dominated by the −TΔS term because ΔS is large and negative for the solute entering water (ordering the shell), which means ΔS is large and positive for the reverse process — the nonpolar group leaving water to be buried in a protein interior or an oil phase. The second essential feature is the heat capacity anomaly. Because ordered hydration-shell water has a different (lower) heat capacity contribution than bulk water, and because the degree of order in the shell itself is temperature-dependent, both ΔH and ΔS for hydrophobic hydration vary strongly and roughly linearly with temperature — producing a large, positive ΔCp for the process of dissolving a nonpolar group in water (and hence a large negative ΔCp for burial/folding, since burial is the reverse of solvation). This ΔCp is the single most reliable calorimetric signature that a given contribution to protein stability is hydrophobic in origin, because polar and electrostatic interactions contribute comparatively little to ΔCp.

The third piece of the picture is what the buried interior actually looks like once assembled. Protein cores are not oil droplets with residual liquid-like disorder; they are packed nearly as densely as a crystal of small organic molecules, with packing coefficients (fraction of volume actually occupied by atoms, using van der Waals radii) around 0.72–0.75, essentially indistinguishable from the packing density of crystalline amino acids and comparable to close-packed spheres (0.74 for face-centered cubic packing). This density is required for van der Waals contacts to contribute meaningfully to stability and explains why core mutations that change side-chain volume (e.g., Leu→Ala, creating a cavity) are often strongly destabilizing even though both residues are nonpolar — packing defects cost stability quite apart from any change in hydrophobicity per se.

The free energy of transferring a nonpolar solute from water to a nonpolar (oil-like or protein-interior) environment is

where:

• $\Delta G_{tr}$ — standard free energy of transfer (water → nonpolar phase), negative for favorable burial
• $\Delta H_{tr}$ — enthalpy of transfer, small in magnitude near room temperature
• $T$ — absolute temperature (K)
• $\Delta S_{tr}$ — entropy of transfer, large and positive near room temperature (water released from an ordered shell to the bulk)

Reading: at 25 °C, $\Delta G_{tr}$ for small hydrocarbons is dominated by the $-T\Delta S_{tr}$ term; $\Delta H_{tr}$ is often close to zero or even positive (unfavorable), meaning the whole driving force for burial is entropic at this temperature. The temperature dependence of $\Delta H_{tr}$ and $\Delta S_{tr}$ is governed by a large, essentially temperature-independent heat capacity change:

where:

• $\Delta C_{p,tr}$ — heat capacity change on transfer (large, positive, for water → oil; hence large negative for oil/protein-interior → water, i.e., for unfolding)
• the near-constancy of $\Delta C_{p,tr}$ over 0–100 °C is the empirical basis for integrating it to get $\Delta H(T)$ and $\Delta S(T)$

Integrating from a reference temperature $T_h$ (the temperature at which $\Delta H_{tr}=0$) gives:

where:

• $T_h$ — temperature at which the transfer enthalpy vanishes (empirically near 22 °C for many hydrocarbons)
• $T_s$ — temperature at which the transfer entropy vanishes (empirically near 112–140 °C for many hydrocarbons, always higher than $T_h$)

Reading: because $T_s \gg T_h$, there is a wide temperature window (around physiological temperature) in which $\Delta H_{tr}$ is small but $\Delta S_{tr}$ is still large and positive — exactly the regime in which hydrophobic burial is entropy-driven. As $T \to T_s$, $\Delta S_{tr}\to 0$ and the driving force becomes purely enthalpic; this is the microscopic reason protein hydrophobic stability itself is temperature-dependent and why proteins can undergo cold denaturation. Empirically, the transfer free energy for burying nonpolar surface correlates linearly with the buried accessible surface area (ASA):

where:

• $\sigma$ — atomic solvation parameter, empirically about $-25$ cal/(mol·Å²) (favorable, i.e., negative sign for burial) for aliphatic/aromatic carbon transferring from water to a nonpolar liquid or protein-like environment
• $\Delta A_{np}$ — nonpolar surface area removed from water contact upon folding/burial (Ų)
• $b$ — small intercept, often taken as zero in first approximation

Reading: this linear relation is the working currency of the field — it lets you predict the stability contribution of burying a given side chain, or the destabilization from a core cavity-forming mutation, directly from a change in buried surface area.

Example 1 — Foundational: transfer free energy of a simple hydrocarbon, water → cyclohexane. Experimental data for the transfer of ethane from water to a nonpolar solvent (a classic model system for the hydrophobic effect, per Tanford and later calorimetric work) near 25 °C give approximately:

Compute $\Delta G_{tr}$ at $T = 298\ \text{K}$:

This is favorable overall ($\Delta G_{tr}<0$, ethane prefers the nonpolar phase) despite an unfavorable (positive) enthalpy — the entropy term alone, $-T\Delta S_{tr}=-7.45$ kcal/mol, more than compensates. This is the textbook demonstration that hydrophobic partitioning can be enthalpically uphill and still occur, because it is entropically driven. Limiting-case check: if $\Delta S_{tr}$ were exactly zero (no water reorganization at all, e.g. an artificial solute that does not perturb water structure), $\Delta G_{tr}$ would simply equal $\Delta H_{tr}=+2.0$ kcal/mol — unfavorable, and ethane would prefer to stay in water. The favorable partitioning we actually observe is entirely a consequence of the nonzero, positive $\Delta S_{tr}$, confirming the entropic origin of the effect in this system.

Example 2 — Applied: burial free energy of a phenylalanine side chain from ASA data. Phenylalanine's fully exposed side-chain ASA is about 200 Ų (aromatic ring plus CH₂ linker); in a typical folded protein core it becomes almost completely buried, losing contact with water. Using $\sigma = -25$ cal/(mol·Å²):

Compare this to the total net stability of a typical small globular protein, $\Delta G_{fold}\approx -8$ to $-12$ kcal/mol below the unfolded state. A single buried phenylalanine can in principle contribute nearly half the entire net folding stability — which is exactly why single core mutations (Phe→Ala, Phe→Gly) at buried positions are so often severely destabilizing or fold-disrupting, and why aromatic and branched aliphatic residues are so strongly conserved at core positions across evolutionarily related folds. Limiting-case check: if the same Phe residue were instead placed at a fully solvent-exposed surface position in the folded state (as many are), $\Delta A_{np}\to 0$ and $\Delta G_{tr}\to 0$ — surface aromatic residues contribute essentially nothing to hydrophobic stabilization, consistent with the empirical observation that surface hydrophobic patches (unless involved in specific binding interfaces) do not correlate with increased protein stability.

Example 3 — Exam/competition level: dissecting a calorimetric unfolding curve into hydrophobic and polar contributions. A differential scanning calorimetry (DSC) experiment on a small globular protein (100 residues) gives, at the melting temperature $T_m = 333\ \text{K}$ (60 °C):

(a) Estimate $\Delta H_{unfold}$ at physiological temperature, 310 K (37 °C), using the Kirchhoff relation:

(b) Estimate $\Delta S_{unfold}$ at 310 K. At $T_m$, by definition $\Delta G_{unfold}(T_m)=0$, so $\Delta S_{unfold}(T_m) = \Delta H_{unfold}(T_m)/T_m = 90/333 = 0.270$ kcal/(mol·K). Extrapolate using the heat capacity integral:

(c) Compute $\Delta G_{unfold}$ at 310 K:

Since unfolding is unfavorable by +5.4 kcal/mol at 37 °C, folding is favorable by $\Delta G_{fold}=-5.4$ kcal/mol at physiological temperature — a modest but entirely typical net stability for a small globular protein, arising as a small difference between a large favorable hydrophobic entropy term and large unfavorable conformational-entropy and polar-desolvation terms buried inside $\Delta H_{unfold}$ and $\Delta S_{unfold}$. Limiting-case check: note that the large positive $\Delta C_{p,unfold}=1.2$ kcal/(mol·K) is itself diagnostic — a value this large (typical experimental range 0.5–2 kcal/(mol·K) for globular proteins of this size) could not arise from polar or electrostatic interactions alone, whose heat capacity contributions are an order of magnitude smaller; it is essentially entirely attributable to exposure of the hydrophobic core upon unfolding, directly linking the calorimetric number to the physical picture of clathrate-shell formation developed above.

Walter Kauzmann (1916–2009) is the figure most closely associated with formalizing the hydrophobic effect as the dominant force in protein folding. Trained as a physical chemist at Princeton, Kauzmann worked in the 1950s at a moment when the physical basis of protein stability was almost entirely unknown — the double helix had just been solved, but nobody had a convincing account of why a linear chain of amino acids would spontaneously and reproducibly collapse into a unique three-dimensional structure. Kauzmann's landmark 1959 review, drawing on earlier solubility work by Kauzmann's contemporaries and on the classic thermodynamic studies of hydrocarbon solubility by researchers including Henry S. Frank and Marjorie Evans (whose 1945 "iceberg" model of water structuring around nonpolar solutes anticipated the modern clathrate picture), argued forcefully that burying nonpolar side chains away from water — not hydrogen bonding, which he showed could be satisfied about equally well in unfolded and folded states — was the dominant thermodynamic driving force for folding. This was a genuinely counterintuitive claim at the time, since hydrogen bonds are visually salient in structural models (Pauling's α-helix and β-sheet had just been proposed) and it was natural to assume they were also energetically dominant. Kauzmann's insight — that the entropy of water, not any direct bond within the protein, supplies most of the folding free energy — reoriented the entire field toward thinking about solvent thermodynamics as central to macromolecular structure, a viewpoint later reinforced quantitatively by Charles Tanford's systematic measurements of amino acid side-chain transfer free energies (the hydrophobicity scales still used today) and by decades of high-precision calorimetry establishing the ΔCp signature discussed above.