# The Hydrophobic Effect: Why Proteins Fold

Drop a globular protein into water and it will fold into one specific three-dimensional shape. Nothing in the polypeptide chain is pulling it into that shape from the outside. The instruction is already in the sequence, and much of the driving force comes from the water itself. Nonpolar side chains disrupt the hydrogen-bonding network of water, and the system responds by burying those side chains in the protein interior. That tendency of nonpolar groups to associate in water is the hydrophobic effect.

You will meet this concept constantly. It explains why soluble proteins have hydrophobic cores and polar surfaces, why membrane proteins are built inside out, why some proteins unfold when you cool them, and why hydrophobic interaction chromatography works. It also shows up in every hydropathy plot and every discussion of protein stability, so getting the thermodynamics right matters more than memorizing a slogan about "oil and water."

## Quick Answer

- The hydrophobic effect is the tendency of nonpolar molecules and groups to associate in water, driven by the release of ordered water from their surfaces into bulk solution [1].
- The release is favorable because free water molecules are more disordered than water trapped against a nonpolar surface, so the entropy of the system increases, consistent with the second law of thermodynamics [1].
- In an aqueous environment, protein folding is driven by the strong tendency of hydrophobic residues to be excluded from water [2].
- Hydrophobic free energy scales with nonpolar surface area, so burying more nonpolar surface releases more of the ordered water and contributes more stabilization [4].
- The classic one-line interpretation: water pushes nonpolar groups together because that arrangement lets water be more disordered, not because the nonpolar groups attract each other strongly [5]. Privalov disputes this and argues that van der Waals attraction between nonpolar groups is the favorable term [9].

## What the Hydrophobic Effect Actually Is

Water molecules hydrogen bond to each other in a constantly rearranging network. A nonpolar molecule cannot donate or accept hydrogen bonds, so the water molecules adjacent to it lose some of that freedom. They become more ordered, forming cage-like arrangements around the nonpolar surface [1]. This is an unfavorable change in entropy for the water.

When two nonpolar molecules come together, part of that ordered water is released into bulk solution [1]. Freed water molecules tumble and hydrogen bond in many more ways than they could while pinned against a nonpolar surface, so the entropy of the water rises. In this textbook picture, that entropy increase is what makes association favorable near room temperature.

This is why the hydrophobic effect is often described as an organizing force based on repulsion by the solvent instead of attractive forces at the site of organization [5]. Tanford made this point explicitly: the effect assembles cell and organelle membranes, but the absence of strong attractive forces between the associating groups is what keeps membranes fluid and deformable [5]. Protein folding is different in one respect. It also involves directed polar bonds, including hydrogen bonds and ionic interactions, which make a folded protein more rigid than a membrane [5].

The classic source for the idea that a hydrophobic factor drives folding is Kauzmann's 1959 review [3]. Kauzmann explained the low solubility of hydrocarbons through hydration shells, possibly clathrate-like water [4]. That explanation was not universally accepted at the time because of skepticism about clathrate hydration shells. Baldwin restated it in 2014 with a dynamic hydration shell formed by van der Waals attraction between hydrocarbon carbons and water oxygens, and concluded that the decrease in entropy upon hydration is the root cause of hydrophobicity, probably from extensive ordering of water molecules in the shell [4].

## How the Effect Builds a Folded Protein

In a folded soluble protein, hydrophobic residues are concentrated in the interior and polar residues are enriched on the surface. Myoglobin is the textbook case: its interior consists almost entirely of nonpolar residues such as leucine, valine, methionine and phenylalanine, while its outside carries both polar and nonpolar residues [2]. That interior region is the hydrophobic core.

Secondary structure follows the same logic. Many alpha helices and beta strands are amphipathic: one face is hydrophobic and points into the protein interior, the other is more polar and points into solution [2]. An amphipathic helix can therefore sit at an interface, with one face buried and one face hydrated.

Membrane proteins invert this arrangement. Porins and similar membrane proteins are "inside out" relative to soluble proteins: their outside is largely hydrophobic, contacting lipid alkane chains, while the center contains many charged and polar residues [2]. The same physical rule produces the opposite architecture because the solvent environment is different.

The free energy contribution scales with nonpolar surface area [4]. That relationship is the reason burial is a useful proxy for hydrophobic stabilization: the more nonpolar surface you hide from water, the more ordered water you release, and the larger the favorable contribution.

## Worked Example

Two quick calculations on ubiquitin (PDB 1UBQ) show how these ideas look in real data.

**Hydropathy.** The sequence from the PDB file has 76 residues:

```
MQIFVKTLTGKTITLEVEPSDTIENVKAKIQDKEGIPPDQQRLIFAGKQLEDGRTLSDYNIQKESTLHLVLRLRGG
```

GRAVY is the sum of the Kyte-Doolittle hydropathy values of all residues divided by the number of residues [11][10]:

$$
\text{GRAVY} = \frac{\sum_{i=1}^{N} h_i}{N}
$$

where $h_i$ is the Kyte-Doolittle value of residue $i$ and $N$ is the number of residues. For ubiquitin, GRAVY = -0.4895. A 19-residue sliding-window profile ranges from -1.726 to +0.347, so ubiquitin has no strongly hydrophobic stretch. For contrast, the hypothetical peptides LIVFLLAAIVGLLLA and KDEKRNQEDKSRE give GRAVY +3.200 (sum 48.0 over 15 residues) and -3.538 (sum -46.0 over 13).

**Burial.** Remove waters, run Biopython `ShrakeRupley().compute(model, level='R')`, and divide each residue's SASA by the theoretical maximum of Tien et al. 2013 (Biopython `residue_sasa_scales['Wilke']`, for example Ile 197, Leu 201, Val 174, Phe 240, Met 224, Lys 236, Arg 274, Asp 193, Glu 223 square Angstrom). Calling residues with relative accessibility below 0.20 buried: 17 of 23 Ile/Leu/Val/Phe/Met (73.9%) are buried versus 2 of 22 Asp/Glu/Lys/Arg (9.1%; Asp21 0.17 and Lys27 0.06). At a 0.25 cutoff the nonpolar figure rises to 20 of 23 (87.0%). Fully buried examples include Ile3 0.00, Val5 0.00, Ile23 0.00, Leu56 0.00, Leu67 0.00, Val26 0.01, Ile30 0.01 and Leu43 0.01. The most exposed charged residues are Arg74 0.77, Asp32 0.70 and Glu16 0.67. Computed with Biopython 1.88.

The pattern matches the model: nonpolar side chains are buried, charged side chains stay exposed.

## Reading Hydropathy and Burial Numbers

The Kyte-Doolittle scale assigns each residue a hydropathicity value. Nonpolar residues score positive: Ile 4.5, Val 4.2, Leu 3.8, Phe 2.8, Cys 2.5, Met 1.9, Ala 1.8. The remaining residues score negative: Gly -0.4, Thr -0.7, Ser -0.8, Trp -0.9, Tyr -1.3, Pro -1.6, His -3.2, Glu -3.5, Gln -3.5, Asp -3.5, Asn -3.5, Lys -3.9, Arg -4.5 [10].

GRAVY is the sum of those values divided by the number of residues; positive values indicate a hydrophobic sequence [11][10]. A windowed profile, not a single average, is what you use to look for local hydrophobic segments.

| Quantity | What it measures | What it does not tell you |
|---|---|---|
| GRAVY | Average hydropathy of the whole sequence | Whether the protein folds, or where the core is |
| Windowed hydropathy | Local hydrophobic character along the sequence | Membrane topology without further evidence |
| Relative SASA | Fraction of a residue's surface exposed in a structure | Why that residue ended up there |
| Hydrophobic free energy | Energetic contribution from nonpolar surface burial | The total folding free energy on its own |

A positive GRAVY is a hint, not a verdict. GRAVY is a crude sequence average and does not predict folding. Window size and cutoff choices change the picture, so state the ones you used.

## The Hydrophobic Effect Is Not a Bond

The name invites a wrong picture. The hydrophobic effect is not a bond between nonpolar groups, and it is not a force that nonpolar groups exert on each other across space. In the textbook picture it is a solvent-driven tendency that emerges from water's behavior, although Privalov argues that van der Waals contact between nonpolar groups supplies the favorable term [9].

It is also not the same as hydrophobicity of a single molecule. Chandler emphasized that the hydrophobic effect is multifaceted and depends on whether hydrophobic molecules are individually hydrated or driven to assemble into larger structures [7]. His review distinguishes small and large length scales, with a crossover around 1 nm: for small solutes the solvation free energy scales with solute volume, for large ones with surface area [7]. The underlying theory paper is Lum, Chandler and Weeks 1999, and the scaling picture is still an active area.

The temperature dependence is another place where the simple story breaks. The hydrophobic interaction changes from entropy-driven at 22 degrees C to enthalpy-driven at about 113 degrees C [8]. Baldwin combined calorimetric and solubility data for six liquid hydrocarbons transferred to water and found that the entropy of transfer extrapolates to zero at a strikingly similar temperature for all six, Ts = 112.8 +/- 2.2 degrees C [8]. His model applies where the heat capacity change on transfer is constant, and it predicts that plots of the specific entropy change on unfolding versus temperature nearly intersect close to 113 degrees C, as Privalov observed [8]. The major share of the large enthalpy change seen when proteins unfold at high temperature comes from the hydrophobic interaction [8].

## Common Mistakes

- **Treating the hydrophobic effect as an attractive force between nonpolar groups.** In Tanford's description it is a solvent-driven effect based on repulsion by water, not a direct attraction at the site of association [5]. Start from the entropy of the water, but note that Privalov assigns the favorable term to van der Waals contact between nonpolar groups [9].
- **Saying the hydrophobic effect is always entropy-driven.** That holds near room temperature. It becomes enthalpy-driven near 113 degrees C [8]. State the temperature range when you make the claim.
- **Assuming a positive GRAVY means the protein is membrane-bound.** GRAVY is a whole-sequence average and does not predict folding or localization. Use a windowed profile and additional evidence.
- **Using a single relative-accessibility cutoff as if it were standard.** The 0.20 cutoff is a convention. Results depend on the cutoff, the SASA algorithm and the normalization scale. Report the cutoff and the scale you used.
- **Confusing the hydrophobic core with a rigid, static structure.** The core is a region enriched in nonpolar residues, not a solid block. Side chains inside it still move.
- **Forgetting that folding also involves polar interactions.** Hydrogen bonds and ionic interactions contribute to the folded state and make proteins more rigid than membranes [5]. The hydrophobic effect is a major driver, not the only term in the free energy balance.

## Limitations

The molecular explanation of the hydrophobic effect is still debated. The textbook ordered-water-cage and entropy picture, associated with Kauzmann and restated by Baldwin in 2014, contrasts with Privalov's view that the favorable force is van der Waals contact between nonpolar groups [4][9]. Privalov argued that the favorable contribution to the hydrophobic interaction is van der Waals attraction between nonpolar groups, not hydration of those groups, revising the conventional view [9]. Present both pictures when you teach or write about this.

The entropy-driven character is a room-temperature simplification. Baldwin's analysis shows the crossover to enthalpy-driven behavior near 113 degrees C [8], so any statement that the hydrophobic effect "is entropic" needs a temperature qualifier.

The Chandler 1 nm crossover and the volume-versus-area scaling come from a summary of the Nature article page, not a verbatim quote, and the underlying theory paper is Lum, Chandler and Weeks 1999. Treat the exact crossover value as approximate.

Cold denaturation is a real consequence of the temperature dependence. It is a general phenomenon of globular proteins caused by the strongly temperature-dependent interaction of nonpolar groups with water [9]. Privalov reported that hydration of nonpolar groups is thermodynamically favorable (negative Gibbs energy) and becomes more favorable as temperature falls, so at low enough temperature the compact chain unfolds and exposes its internal nonpolar groups [9]. This is the opposite of what a simple "hydrophobic effect drives folding" statement would predict, and it is a good check on whether you actually understand the thermodynamics.

Finally, hydropathy window sizes and cutoffs are conventions, not physical constants. GRAVY is a crude sequence average and does not predict folding.

## Frequently Asked Questions

### What is the hydrophobic effect in one sentence?

It is the tendency of nonpolar molecules and groups to associate in water because doing so releases ordered water from their surfaces into bulk solution, increasing the entropy of the water [1]. The association is favorable because free water is more disordered than water trapped against a nonpolar surface [1].

### Why is the hydrophobic effect described as entropy-driven?

Water molecules next to a nonpolar surface become more ordered, forming cage-like arrangements [1]. When nonpolar surfaces associate, some of that ordered water is released into bulk solution, and the released water is more disordered than it was when confined [1]. That entropy increase is the favorable term near room temperature.

### How do hydrophobic interactions in proteins produce a hydrophobic core?

In an aqueous environment, protein folding is driven by the strong tendency of hydrophobic residues to be excluded from water [2]. Nonpolar side chains end up packed in the interior, as seen in myoglobin, whose interior consists almost entirely of nonpolar residues such as leucine, valine, methionine and phenylalanine [2]. Amphipathic helices and strands follow the same rule, with their hydrophobic faces pointing inward [2].

### What does amphipathic mean for a helix or strand?

It means one face of the secondary structure element is hydrophobic and the other is more polar [2]. The hydrophobic face points into the protein interior, and the polar face points into solution [2]. This arrangement lets the same element satisfy both the hydrophobic core and the aqueous environment.

### Why do proteins fold at all, if the hydrophobic effect is just water ordering?

Folding is driven by the strong tendency of hydrophobic residues to be excluded from water [2], but it also involves directed polar bonds that stabilize the folded state and make it more rigid than a membrane [5]. The balance of forces is reviewed in Dill's 1990 Biochemistry review [6]. The hydrophobic effect is a major term, not the whole free energy balance.

## References

1. [Berg et al. Biochemistry 8th ed., Section 1.3 Concepts from chemistry (hydrophobic effect)](https://digfir-published.macmillanusa.com/berg8e/berg8e_ch01_4.html)
2. [Berg et al. Biochemistry 8th ed., Section 2.4 Tertiary structure: nonpolar cores](https://digfir-published.macmillanusa.com/berg8e/berg8e_ch02_5.html)
3. [Kauzmann 1959. Some factors in the interpretation of protein denaturation. Adv Protein Chem](https://doi.org/10.1016/S0065-3233(08)60608-7)
4. [Baldwin 2014. Dynamic hydration shell restores Kauzmann's 1959 explanation of how the hydrophobic factor drives protein folding. PNAS 111:13052](https://doi.org/10.1073/pnas.1414556111)
5. [Tanford 1978. The hydrophobic effect and the organization of living matter. Science 200:1012](https://doi.org/10.1126/science.653353)
6. [Dill 1990. Dominant forces in protein folding. Biochemistry 29:7133](https://doi.org/10.1021/bi00483a001)
7. [Chandler 2005. Interfaces and the driving force of hydrophobic assembly. Nature 437:640](https://doi.org/10.1038/nature04162)
8. [Baldwin 1986. Temperature dependence of the hydrophobic interaction in protein folding. PNAS 83:8069](https://doi.org/10.1073/pnas.83.21.8069)
9. [Privalov 1990. Cold denaturation of proteins. Crit Rev Biochem Mol Biol 25:281](https://doi.org/10.3109/10409239009090612)
10. [ExPASy ProtScale: Kyte & Doolittle hydropathicity scale (J Mol Biol 1982, 157:105)](https://web.expasy.org/protscale/pscale/Hphob.Doolittle.html)
11. [ExPASy ProtParam documentation: GRAVY definition](https://web.expasy.org/protparam/protparam-doc.html)

## Related Articles

- [Protein Folding](/blog/guides/protein-folding)
- [Tertiary Structure of Protein: The 3D Fold](/blog/guides/tertiary-structure-of-protein)
- [Hydrophobic Interaction Chromatography for Protein Purification](/knowledge/molecular-biology/hydrophobic-interaction-chromatography-for-protein-purification-resin-selection-and-salt-optimizatio)
- [Transmembrane Protein](/blog/guides/transmembrane-protein)
- [Levinthal's Paradox and Anfinsen's Experiment: How Proteins Find Their Fold](/blog/research-skills/levinthal-paradox-and-anfinsen-experiment)
- [How to Read a Helical Wheel and a Hydropathy Plot (Amphipathic Helices and Transmembrane Segments)](/blog/research-skills/helical-wheel-and-hydropathy-plot)