Thermodynamics of Protein Folding

Introduction

Why does a long polypeptide adopt a specific functional shape instead of remaining a shapeless coil? Folding is a thermodynamic competition between ensembles of conformations in a particular solvent. Native-state contacts favour compact structures, while an unfolded chain has many more accessible arrangements and can interact extensively with water. The winning state is often only modestly lower in Gibbs free energy. This marginal stability makes proteins responsive to mutations, temperature, pH and binding partners.

Core explanation

For a simple two-state description U ⇌ N, define ΔG fold = G N − G U at a given temperature and solution composition. If ΔG fold is negative, the native ensemble is favoured at equilibrium. The relation ΔG fold = −RT ln(K fold), with K fold = [N]/[U], connects a free-energy difference to a population ratio under appropriate dilute-solution assumptions. Some proteins require intermediate or multiple-state descriptions, so the two-state model is a useful approximation rather than a universal law.

At constant pressure and temperature, ΔG = ΔH − TΔS. Folding can gain favourable enthalpic contributions from close packing, hydrogen bonds, electrostatics and other interactions. It also reduces the chain's conformational entropy because far fewer backbone and side-chain arrangements remain. Solvent changes complicate the picture: burying nonpolar surface changes water structure and can favour folding, while desolvating polar groups can be costly unless they gain new partners. Calling folding purely “enthalpy-driven” or purely “entropy-driven” omits these opposing terms.

The unfolded state is not one fully extended structure. It is a large ensemble with transient local structure and fluctuating contacts. The native state is likewise an ensemble of related conformations rather than a single rigid coordinate set. Thermodynamic stability concerns the free-energy difference between these ensembles. A protein can contain many favourable contacts but still have a small net ΔG because the unfolded chain retains favourable interactions and much greater conformational freedom.

Temperature changes both the −TΔS term and the underlying hydration and heat-capacity contributions. Heating often leads to denaturation, but the simplistic claim that heat “breaks peptide bonds” is wrong; ordinary thermal unfolding usually disrupts noncovalent organization while the primary sequence remains intact. Some proteins can also unfold at low temperature under suitable conditions because hydrophobic hydration has a nontrivial temperature dependence. A detailed stability curve therefore cannot always be predicted from one constant ΔH and ΔS over a wide range.

Chemical denaturants such as urea or guanidinium salts can shift the equilibrium toward unfolded states. Experiments may monitor a spectroscopic signal versus denaturant concentration or temperature and fit a model to estimate stability. A midpoint is where populations of two assumed states are equal, not where every molecule is halfway folded. Mutations can shift ΔG fold by altering both native and unfolded ensembles; a residue found in the folded structure does not reveal the mutation's full thermodynamic effect by itself.

Step-by-step reasoning

For a folding problem, declare whether ΔG refers to folding or unfolding because signs reverse. Identify the temperature, solvent, pH and ligand conditions. Separate structural explanations into native-state contacts, unfolded-state entropy and solvent contributions. If a population ratio is given, calculate ΔG from the equilibrium expression and check the sign. Finally state whether the two-state approximation is justified by the evidence.

Visual explanation

Draw two free-energy wells: a broad unfolded well representing many conformations and a narrower native well at slightly lower G. A small vertical difference marks marginal thermodynamic stability; a barrier between wells represents folding kinetics, which is a separate question. Beneath the drawing, write ΔG fold = G N − G U = ΔH fold − TΔS fold and show that both the chain and water contribute to ΔS fold.

Real-world analogy

A carefully arranged stack of objects may be held by many contacts but lose the freedom available to scattered objects. Whether stacking is favoured depends on both contact benefits and lost arrangements. Protein folding similarly balances contact energy against entropy, though the molecular solvent and continuous conformational motion make its calculation more subtle than a stack of objects.

Real-world example

An enzyme can function well at its normal temperature but lose activity after heating. A likely explanation is that an increased unfolded population disrupts the active-site geometry. Cooling might restore activity if unfolding is reversible and aggregation or chemical damage has not occurred. If unfolded molecules aggregate, however, cooling need not recover the original protein even when its isolated native state would be thermodynamically favoured.

Why?

Why can one mutation markedly change the folded fraction of a protein with thousands of atomic contacts? The stability is the difference between large, competing contributions, often a relatively small net value. Altering packing, hydration or entropy by one residue can shift that difference enough to change equilibrium populations.

Common misconception

“More native-state hydrogen bonds always means more stable protein.” A new bond may merely replace a hydrogen bond to water, strain the structure, or alter unfolded-state interactions. Stability depends on ΔG between whole ensembles, not a count of contacts in one static model.

Worked example

At 298 K, suppose a two-state protein has K fold = [N]/[U] = 9. Then ΔG fold = −RT ln 9 ≈ −(8.314 J mol⁻¹ K⁻¹)(298 K)(2.197) ≈ −5.44 kJ mol⁻¹. The folded fraction is 9/(9+1) = 90%. A negative free energy therefore need not mean 100% of molecules are folded. The calculation assumes equilibrium and that N and U capture the relevant states.

Quick check

1. Does ordinary protein denaturation necessarily cleave the peptide backbone? Answer: No. It usually disrupts native noncovalent structure; peptide-bond hydrolysis or other chemical damage is a distinct process.

Exam focus

Specify your ΔG sign convention and use equilibrium populations correctly. Explain both favourable contacts and the entropy cost of constraining the chain, including solvent effects. Distinguish the equilibrium preference for a fold from the rate at which the chain reaches it.

Advanced insight

The Gibbs energy of a protein is a property of an ensemble, not one model structure. Ligand binding can stabilise a particular conformation by changing its chemical potential and thereby shift a folding or conformational equilibrium. This thermodynamic linkage underlies many allosteric effects and explains why stability measured without a ligand may differ from stability inside a cell.

Summary

Protein folding is favoured when the native ensemble has lower Gibbs energy than the unfolded ensemble under the specified conditions. Contact formation and solvent changes compete with loss of chain entropy. The net stability can be modest, making proteins sensitive to environment and sequence while leaving peptide bonds intact during ordinary unfolding.

Practice questions

1. A protein has [N]/[U] = 1 at a thermal midpoint. What is ΔG fold in a two-state equilibrium model at that temperature? Answer: ΔG fold = −RT ln 1 = 0. The native and unfolded populations are equal by the model's definition; individual molecules are not necessarily half-folded. 2. Explain why burying a polar side chain without a hydrogen-bond partner can destabilise a fold. Answer: The polar group loses favourable interactions with water on burial but gains no compensating internal partner. The resulting desolvation cost can outweigh other local packing gains. 3. A denatured enzyme fails to refold when cooled. Does this prove its native state is thermodynamically unstable at the lower temperature? Answer: No. Aggregation, misfolded kinetic traps or irreversible chemical changes may block recovery. Equilibrium stability of an isolated chain and the kinetics of reaching that state are different questions.