Protein Folding, Stability, and Denaturation - Biochemistry

SECTION 1

Historical Context & Motivation

The question of how a linear chain of amino acids spontaneously adopts a precise three-dimensional shape ranks among the most profound problems in molecular biology. For much of the twentieth century, biochemists assumed that protein structure was somehow templated by other cellular machinery, but a series of elegant experiments revealed that the amino acid sequence alone contains all the information necessary for a protein to fold into its native conformation. Understanding protein folding is not merely an academic exercise: misfolded proteins underlie diseases ranging from Alzheimer's to cystic fibrosis, and rational drug design depends on knowing precisely how a target protein is shaped. The historical arc of this field spans thermodynamic reasoning, kinetic paradoxes, and computational breakthroughs that continue to reshape modern biochemistry.

1957 • Myoglobin Structure Solved

John Kendrew and Max Perutz used X-ray crystallography to determine the first atomic-resolution protein structure—sperm whale myoglobin—demonstrating that proteins possess defined, reproducible three-dimensional architectures.

1961 • Anfinsen's Refolding Experiment

Christian Anfinsen demonstrated that denatured ribonuclease A spontaneously refolds to its catalytically active form, establishing the thermodynamic hypothesis—the native structure corresponds to the global free-energy minimum dictated by the amino acid sequence.

1968 • Levinthal's Paradox

Cyrus Levinthal calculated that a random search through all possible conformations would take longer than the age of the universe, implying that proteins must follow directed folding pathways rather than sampling every conformation.

1995 • Energy Landscape (Funnel) Theory

Ken Dill, José Onuchic, and Peter Wolynes proposed the folding funnel model, depicting the free-energy landscape as a rough funnel that biases the polypeptide toward the native state without requiring a single defined pathway.

2020 • AlphaFold2

DeepMind's AlphaFold2 achieved near-experimental accuracy in protein structure prediction, marking a computational milestone and enabling the prediction of hundreds of millions of protein structures.

The central question that unifies this history remains deceptively simple: given an unfolded polypeptide chain in aqueous solution, what physical forces drive it to adopt one particular shape, how stable is that shape, and what happens when those forces are overwhelmed? Answering these questions requires a synthesis of thermodynamics, noncovalent interactions, solvent effects, and kinetic reasoning—the very topics explored in the sections that follow.

SECTION 2

Core Principles of Protein Folding & Stability

Protein folding is governed by the interplay of several noncovalent forces and the thermodynamic requirement that the native state occupies the global free-energy minimum of the polypeptide–solvent system. Although individual noncovalent interactions are weak—typically 4–30 kJ mol⁻¹ each—hundreds or thousands of them act cooperatively to stabilize a folded protein by a net margin that is itself surprisingly small, often only 20–65 kJ mol⁻¹. This marginal stability is not a defect; it is a feature that allows proteins to undergo the conformational changes essential for function. The following core principles frame the entire field.

Anfinsen's Thermodynamic Hypothesis

The native structure of a protein is determined solely by its amino acid sequence and corresponds to the state of lowest Gibbs free energy under physiological conditions. No external template is required.

The Hydrophobic Effect

Nonpolar side chains are driven into the protein interior to minimize their contact with water, releasing ordered water molecules and increasing solvent entropy—the dominant driving force for folding.

Conformational Entropy Opposition

Folding restricts the polypeptide to a narrow region of conformational space, producing a large unfavorable decrease in chain entropy (−TΔS > 0) that must be overcome by favorable enthalpy and solvent entropy changes.

Marginal Net Stability

The net stabilization free energy ΔG°(folding) is typically only −20 to −65 kJ mol⁻¹—the small difference between enormous opposing contributions—making proteins sensitive to perturbation.

Cooperative Folding

Most small single-domain proteins fold in an all-or-none (two-state) transition: partially folded intermediates are thermodynamically unstable, and the protein is essentially either fully folded or fully unfolded.

✦

KEY TAKEAWAY

Think of protein folding like a ball rolling down a hilly landscape into a valley. The valley floor is the native state—the lowest energy point. The ball doesn't need to explore every hill and dale; gravity (analogous to the hydrophobic effect and hydrogen bonding) funnels it downward. But the valley is only slightly lower than the surrounding plains: a modest perturbation—heat, a denaturant, an extreme pH—can push the ball back out of the valley. This marginal stability is what allows proteins to be both stable enough to function and flexible enough to be regulated.

SECTION 3

The Folding Funnel Energy Landscape

The most powerful conceptual tool for understanding protein folding is the energy landscape or folding funnel diagram. Instead of a single defined pathway, the unfolded polypeptide can begin from any of an astronomically large number of conformations at the top of the funnel. As it forms favorable intramolecular contacts, the chain descends the energy surface toward the native state at the funnel's bottom. The width of the funnel at any height represents the conformational entropy (number of accessible conformations), while the depth represents the free energy. Roughness on the funnel surface corresponds to kinetic traps—partially folded intermediates or misfolded states that temporarily slow the folding process.

The folding funnel illustrates how the conformational entropy (horizontal width) decreases as the polypeptide descends toward the native state at the funnel bottom. Kinetic traps (local minima) on the funnel surface represent misfolded intermediates that can temporarily stall the folding process.

In the diagram above, the red dots at the rim represent the vast ensemble of unfolded conformations—each with high free energy and maximal conformational entropy. As favorable contacts form (hydrogen bonds, hydrophobic packing, van der Waals interactions), the chain moves down the funnel, progressively restricting its conformational freedom while lowering its free energy. The amber-colored kinetic traps illustrate an important reality: folding is not always smooth. Some intermediates may be kinetically stabilized by partially correct contacts that must be broken and reformed before the chain can proceed to the thermodynamically favored native state. Molecular chaperones in vivo often assist by preventing aggregation of these partially folded intermediates, effectively smoothing the funnel surface.

SECTION 4

Thermodynamic Framework of Folding

The thermodynamics of protein folding can be described quantitatively using the Gibbs free energy framework. For the two-state folding equilibrium U ⇌ N (unfolded ⇌ native), the stability of the native state is defined by the free energy change of folding, ΔG°_fold. A negative value indicates that the native state is thermodynamically favored. Because the overall free energy change is a balance between enormous opposing enthalpic and entropic contributions, understanding each term is critical for predicting how perturbations—temperature, pH, denaturants—shift the equilibrium.

GIBBS FREE ENERGY OF FOLDING

ΔG°fold = ΔH°fold − TΔS°fold

Where ΔG°fold is the standard free energy of folding (kJ mol⁻¹), ΔH°fold is the enthalpy change (reflects noncovalent interactions: hydrogen bonds, van der Waals, hydrophobic desolvation), T is absolute temperature (K), and ΔS°fold is the entropy change (combines decreased chain entropy and increased solvent entropy upon burial of hydrophobic residues).

FOLDING EQUILIBRIUM CONSTANT

Keq = [N] / [U] = e^(−ΔG°fold / RT)

Where Keq is the equilibrium constant for folding, [N] and [U] are concentrations of native and unfolded states, R is the gas constant (8.314 J mol⁻¹ K⁻¹), and T is the absolute temperature. When ΔG°fold = −40 kJ mol⁻¹ at 298 K, Keq ≈ 1.1 × 10⁷, meaning only about 1 in 10 million molecules is unfolded at equilibrium.

LINEAR EXTRAPOLATION MODEL (LEM) FOR DENATURANT

ΔG°fold = ΔG°H₂O − m × [denaturant]

Where ΔG°H₂O is the free energy of folding in pure water (maximum stability), m is the m-value (kJ mol⁻¹ M⁻¹), proportional to the change in solvent-accessible surface area upon unfolding, and [denaturant] is the molar concentration of urea or guanidinium chloride. The midpoint concentration where ΔG°fold = 0 is Cm = ΔG°H₂O / m.

MELTING TEMPERATURE (Tm)

Tm = ΔH°fold / ΔS°fold

At the melting temperature Tm, ΔG°fold = 0 and the protein is 50% folded, 50% unfolded. Above Tm the entropic penalty of folding dominates and the protein unfolds. Below Tm the enthalpy of noncovalent interactions dominates and the protein remains folded.

ℹ

Why Is ΔG°fold So Small?

Consider that the total favorable enthalpy from forming hundreds of hydrogen bonds and van der Waals contacts upon folding might be −2000 kJ mol⁻¹. But this is almost exactly offset by the loss of conformational entropy (≈ +1600 kJ mol⁻¹ at 298 K) and the cost of desolvating polar groups (≈ +350 kJ mol⁻¹). The net ΔG°fold of −40 kJ mol⁻¹ is the small residual—like the profit margin of a business with enormous revenues and nearly equally enormous expenses. This is why single-point mutations can shift ΔΔG by 5–15 kJ mol⁻¹ and dramatically alter stability.

SECTION 5

Stabilizing Forces and Modes of Denaturation

Several classes of noncovalent interactions contribute to protein stability, and each mode of denaturation disrupts a different subset of these forces. Understanding which forces are perturbed by each denaturing condition is essential for experimental protein biochemistry and for designing stable proteins. The table below summarizes the major stabilizing forces and their approximate energetic contributions, while the diagram that follows illustrates how different denaturants act on a folded protein.

Stabilizing Force	Typical Strength (kJ mol⁻¹)	Role in Folding	Disrupted By
Hydrophobic effect	~4–12 per residue buried	Primary driving force; burial of nonpolar side chains increases solvent entropy	Urea, GdnHCl, detergents, high temperature
Hydrogen bonds	~8–30 per bond	Stabilize secondary structure (α-helices, β-sheets) and tertiary contacts	Urea, GdnHCl, extreme pH
Van der Waals interactions	~2–4 per contact	Close packing in the hydrophobic core; collectively substantial	High temperature, high pressure
Electrostatic / ion pairs	~15–40 per pair	Salt bridges on protein surface; more important in thermophilic proteins	Extreme pH, high ionic strength
Disulfide bonds (covalent)	~170 (covalent)	Reduce conformational entropy of unfolded state; especially important for extracellular proteins	Reducing agents (β-mercaptoethanol, DTT)

Major forces stabilizing the native protein fold and conditions that disrupt them

The native (folded) protein sits at the center, while five major modes of denaturation radiate outward. Heat overcomes the hydrophobic effect and weakens hydrogen bonds. Chemical denaturants (urea, GdnHCl) favorably solvate the peptide backbone and nonpolar side chains. Extreme pH alters protonation states, disrupting salt bridges and introducing charge–charge repulsion. Reducing agents cleave disulfide bonds, and detergents compete for hydrophobic interactions.

It is important to distinguish between reversible and irreversible denaturation. In Anfinsen's classic experiment, ribonuclease A was denatured with 8 M urea and β-mercaptoethanol, and upon removal of these agents the enzyme refolded with full recovery of catalytic activity—a hallmark of reversible denaturation. However, many proteins, when heated above their Tm, aggregate through intermolecular hydrophobic contacts between exposed nonpolar surfaces, a process that is effectively irreversible under normal conditions. The distinction is not merely academic: pharmaceutical proteins must be formulated to avoid irreversible aggregation, and laboratory protocols for protein purification depend on knowing whether a protein can be reversibly unfolded and refolded.

SECTION 6

Worked Example: Analyzing Protein Stability

Consider a small single-domain protein that unfolds in a two-state (U ⇌ N) transition. A chemical denaturation experiment with urea yields the following data: at 0 M urea and 25 °C, the fraction of unfolded protein is measured as f_U = 2.0 × 10⁻⁴. The midpoint of the unfolding transition (C_m) is observed at 4.8 M urea. Calculate ΔG°_H₂O, the m-value, and the fraction unfolded at 3.0 M urea.

Chemical Denaturation Analysis of a Two-State Protein

1
Step 1 — Determine Keq at 0 M Urea

For a two-state system, f_U + f_N = 1, so f_N = 1 − 2.0 × 10⁻⁴ = 0.9998. The equilibrium constant for unfolding is K_unfold = f_U / f_N = 2.0 × 10⁻⁴ / 0.9998 ≈ 2.0 × 10⁻⁴.

K_unfold ≈ 2.0 × 10⁻⁴
2
Step 2 — Calculate ΔG°H₂O

Using ΔG° = −RT ln K_unfold with R = 8.314 × 10⁻³ kJ mol⁻¹ K⁻¹ and T = 298 K: ΔG°_H₂O = −(8.314 × 10⁻³)(298) ln(2.0 × 10⁻⁴) = −(2.478)(−8.517) = +21.1 kJ mol⁻¹. Note: here ΔG° is for unfolding, so a positive value means the native state is favored—the free energy of folding is −21.1 kJ mol⁻¹.

ΔG°_H₂O (unfolding) = +21.1 kJ mol⁻¹
3
Step 3 — Determine the m-Value

From the linear extrapolation model, at the transition midpoint C_m = 4.8 M, ΔG° = 0, so: 0 = ΔG°_H₂O − m × C_m. Solving: m = ΔG°_H₂O / C_m = 21.1 / 4.8 = 4.40 kJ mol⁻¹ M⁻¹.

m = 4.40 kJ mol⁻¹ M⁻¹
4
Step 4 — Calculate ΔG° at 3.0 M Urea

ΔG° = ΔG°_H₂O − m × [urea] = 21.1 − 4.40 × 3.0 = 21.1 − 13.2 = +7.9 kJ mol⁻¹. The native state is still favored, but substantially less so than in pure water.

ΔG°(3.0 M urea) = +7.9 kJ mol⁻¹
5
Step 5 — Find the Fraction Unfolded at 3.0 M Urea

K_unfold = e^{(−ΔG°/RT)} = e^{(−7.9/2.478)} = e^−3.19 = 0.041. Then f_U = K / (1 + K) = 0.041 / 1.041 = 0.039, or about 3.9% unfolded.

f_U ≈ 3.9% at 3.0 M urea

SECTION 7

Molecular Chaperones, Misfolding, and Disease

Although Anfinsen's thermodynamic hypothesis holds true in vitro for many small proteins, the crowded intracellular environment—where macromolecular concentrations reach 300–400 mg mL⁻¹—presents additional challenges. Newly synthesized polypeptides risk aggregation: exposed hydrophobic surfaces on partially folded chains can associate intermolecularly rather than intramolecularly, leading to non-functional aggregates or toxic amyloid fibrils. Cells employ a sophisticated network of molecular chaperones to mitigate this risk.

Chaperone System	Mechanism	Example / Context
Hsp70 / DnaK	Binds exposed hydrophobic segments on nascent chains; prevents premature folding and aggregation via ATP-dependent binding–release cycles	Co-translational folding at the ribosome; heat-shock response
Hsp60 / GroEL–GroES	Encapsulates partially folded proteins in an internal cavity (Anfinsen cage), providing a protected environment for folding free of aggregation risk	Post-translational folding of ~10–15% of E. coli cytoplasmic proteins
Hsp90	Assists late-stage folding and conformational maturation; stabilizes metastable client proteins in signaling pathways	Steroid hormone receptors, kinases; cancer drug target
Protein disulfide isomerase (PDI)	Catalyzes formation and reshuffling of disulfide bonds in the ER to reach the thermodynamically correct pattern	Secretory and membrane proteins with multiple disulfides
Peptidyl-prolyl isomerase	Accelerates cis–trans isomerization of Xaa–Pro peptide bonds, which can be rate-limiting in folding	Immunosuppressant targets (cyclophilin, FKBP)

Major molecular chaperone and folding-catalyst systems

✦

KEY TAKEAWAY

Molecular chaperones do not provide folding information—the sequence still dictates the native structure. Instead, they function like soundproof practice rooms for a musician: the musician (polypeptide) already knows the piece (folding code), but the practice room (chaperone cavity) provides an isolated environment free from the noise and distractions (aggregation with neighboring chains) of the concert hall (crowded cytoplasm). When the player emerges, the performance (native fold) is the same one the sequence encoded all along.

When quality-control systems fail, misfolded proteins can accumulate as insoluble aggregates with cross-β amyloid architecture. These deposits are hallmarks of devastating protein-misfolding diseases: amyloid-β plaques in Alzheimer's disease, α-synuclein Lewy bodies in Parkinson's disease, polyglutamine aggregates in Huntington's disease, and prion protein (PrP^Sc) fibrils in transmissible spongiform encephalopathies. Understanding the energetics of folding versus misfolding is therefore of profound medical importance, driving research into small-molecule stabilizers, chaperone-induction therapies, and immunotherapies targeting aggregated species.

SECTION 8

Connections to Advanced Protein Science

The thermodynamic and kinetic principles of protein folding established at the introductory level connect directly to several active areas of advanced research. The table below maps foundational concepts to their more sophisticated counterparts encountered in graduate-level biochemistry, biophysics, and computational biology.

Foundational Concept	Advanced Extension	Significance
Two-state (U ⇌ N) equilibrium	Multi-state folding with on-pathway and off-pathway intermediates; φ-value analysis of transition states	Reveals residue-level structure of the transition-state ensemble
Folding funnel model	All-atom molecular dynamics simulations of folding; Markov state models; coarse-grained structure-based (Gō) models	Enables computational prediction of folding kinetics and pathways
ΔG° from chemical denaturation (LEM)	Differential scanning calorimetry (DSC); hydrogen-deuterium exchange mass spectrometry (HDX-MS); single-molecule force spectroscopy	Provides site-resolved and kinetic stability measurements beyond global ΔG°
Molecular chaperone assistance	Proteostasis network; unfolded protein response (UPR); ER-associated degradation (ERAD); autophagy of aggregates	Integrates folding quality control with cellular signaling and disease
Sequence determines structure	Machine-learning structure prediction (AlphaFold2, RoseTTAFold); de novo protein design (Rosetta, ProteinMPNN)	Protein engineering for therapeutics, catalysts, and biomaterials

From foundations to frontier: bridging undergraduate and graduate-level protein science

An especially exciting frontier is the concept of intrinsically disordered proteins (IDPs), which challenge Anfinsen's hypothesis by remaining unfolded—or existing as dynamic ensembles of conformations—under physiological conditions. IDPs are prevalent in signaling and transcriptional regulation, where disorder confers functional advantages such as the ability to bind multiple partners, undergo coupled folding-and-binding, and be rapidly degraded for tight regulatory control. Their study has expanded the protein folding field from asking 'how do proteins fold?' to asking 'when is it advantageous for a protein not to fold?'

SECTION 9

Practice Problems

Question 1: CONCEPTUAL

Explain why the hydrophobic effect is considered the dominant driving force for protein folding, even though hydrogen bonds and van der Waals interactions also contribute favorably. In your answer, distinguish between the enthalpic and entropic contributions of the hydrophobic effect.

Show answer

The hydrophobic effect is dominant because it provides the largest net free-energy contribution to folding. When nonpolar side chains are exposed to water in the unfolded state, water molecules form ordered 'cages' (clathrates) around them, which is entropically unfavorable. Upon folding, these nonpolar residues are buried in the protein interior, releasing the ordered water molecules into bulk solvent, thereby producing a large favorable increase in solvent entropy (positive ΔS). Although the enthalpic component of the hydrophobic effect is more nuanced (it can be slightly favorable or unfavorable depending on temperature), the entropic gain from water release at physiological temperatures is substantial, often contributing 4–12 kJ mol⁻¹ per nonpolar residue buried. Hydrogen bonds and van der Waals contacts contribute individually to stability but are partially offset by the loss of equivalent interactions with solvent in the unfolded state, making their net contribution per residue smaller than that of the hydrophobic effect.

Question 2: BASIC CALCULATION

A two-state protein has ΔG°(unfolding) = +30 kJ mol⁻¹ at 25 °C (298 K). Calculate the equilibrium constant for unfolding (K_unfold) and the fraction of protein that is unfolded at equilibrium. (R = 8.314 × 10⁻³ kJ mol⁻¹ K⁻¹)

Show answer

Using ΔG° = −RT ln K, we rearrange to K =

e^{−ΔG°/RT}

. Substituting: K =

e^{−30 / (8.314 × 10⁻³ × 298)}

e^{−30/2.478}

e^{−12.11}

≈ 5.5 × 10⁻⁶. For the fraction unfolded:

f_U

= K / (1 + K) = 5.5 × 10⁻⁶ / (1 + 5.5 × 10⁻⁶) ≈ 5.5 × 10⁻⁶, or approximately 0.00055%. This means only about 5–6 molecules out of every million are unfolded at equilibrium—the protein is extremely stable under these conditions. Note that even modest changes in ΔG° produce large changes in K due to the exponential relationship.

Question 3: INTERMEDIATE

A urea denaturation experiment on lysozyme gives ΔG°_H₂O = +38 kJ mol⁻¹ (for unfolding) and a midpoint C_m = 5.0 M urea. (a) Calculate the m-value. (b) At what urea concentration will 20% of the protein be unfolded at 25 °C?

Show answer

(a) From m = ΔG°_H₂O /

C_m

= 38 / 5.0 = 7.6 kJ mol⁻¹ M⁻¹. (b) If

f_U

= 0.20, then

f_N

= 0.80 and K_unfold = 0.20/0.80 = 0.25. The free energy at this concentration is ΔG° = −RT ln K = −(2.478) ln(0.25) = −(2.478)(−1.386) = +3.44 kJ mol⁻¹. Using the LEM: ΔG° = ΔG°_H₂O − m[urea] → 3.44 = 38 − 7.6[urea] → [urea] = (38 − 3.44)/7.6 = 34.56/7.6 = 4.55 M. So approximately 4.55 M urea is required to unfold 20% of the protein. Note that this is close to but below the midpoint (5.0 M), consistent with the cooperative nature of the transition: most of the change in

f_U

occurs over a narrow concentration range near

C_m

Question 4: APPLIED

A pharmaceutical company is developing a therapeutic antibody that must remain stable during shipping at 4 °C and during storage at 25 °C. The antibody has T_m = 68 °C, ΔH°_unfold = 420 kJ mol⁻¹, and ΔS°_unfold = 1.23 kJ mol⁻¹ K⁻¹. Estimate ΔG°_unfold at 25 °C (assuming ΔH and ΔS are temperature-independent) and comment on whether the antibody is sufficiently stable. What formulation strategy might improve shelf stability?

Show answer

At 25 °C (298 K): ΔG°_unfold = ΔH° − TΔS° = 420 − 298(1.23) = 420 − 366.5 = +53.5 kJ mol⁻¹. This is a substantial positive free energy of unfolding, indicating high thermodynamic stability at 25 °C (K_unfold ≈ 4 × 10⁻¹⁰). Verification at

T_m

: ΔG° = 420 − 341(1.23) = 420 − 419.4 ≈ 0 kJ mol⁻¹, confirming

T_m

= 68 °C. Although the antibody is thermodynamically stable, practical concerns include aggregation of the small fraction of transiently unfolded molecules (kinetic instability) and chemical degradation (deamidation, oxidation). Formulation strategies to improve shelf stability include: (1) adding excipients such as sucrose or trehalose that preferentially exclude from the protein surface, increasing ΔG°_unfold; (2) optimizing pH to maximize net charge and colloidal stability; (3) controlling ionic strength; (4) lyophilization for long-term storage. Note that our assumption of temperature-independent ΔH and ΔS is a simplification; in reality, the large heat capacity change upon unfolding (ΔCp) makes both quantities temperature-dependent.

Question 5: CRITICAL THINKING

Intrinsically disordered proteins (IDPs) lack a stable tertiary structure under physiological conditions, yet they account for ~30% of the eukaryotic proteome and perform essential functions. How does the existence of IDPs challenge or refine Anfinsen's thermodynamic hypothesis? Propose a modified thermodynamic framework that accommodates both structured and intrinsically disordered proteins.

Show answer

Anfinsen's thermodynamic hypothesis states that the native state corresponds to the global free-energy minimum. IDPs challenge this by demonstrating that for some sequences, the free-energy minimum is not a single compact structure but rather a broad, shallow basin encompassing a dynamic ensemble of interconverting conformations. The key refinement is that the 'native state' should be redefined as the thermodynamically favored state—which for IDPs is the disordered ensemble itself. In thermodynamic terms, IDP sequences have amino acid compositions enriched in charged and polar residues and depleted in hydrophobic residues, so the hydrophobic driving force for compaction is insufficient to overcome the conformational entropy favoring disorder: ΔG°_fold ≈ |ΔH°_fold| − T|ΔS°_chain| > 0 because |ΔH°_fold| is small relative to T|ΔS°_chain|. A modified framework might define a conformational free-energy landscape where the global minimum can be either a narrow, deep well (structured protein) or a broad, shallow basin (IDP), depending on sequence composition. Many IDPs undergo coupled folding-and-binding, where interaction with a partner protein provides the additional enthalpic contacts needed to make ΔG°_fold < 0 locally—folding is induced by the binding event. This expanded view preserves the thermodynamic logic of Anfinsen's hypothesis while acknowledging that the free-energy landscape, not a single structure, is determined by the amino acid sequence.

SUMMARY

Protein Folding, Stability, and Denaturation — Summary

Summary

Protein folding is the process by which a linear polypeptide chain adopts its native three-dimensional structure, driven primarily by the hydrophobic effect and reinforced by hydrogen bonds, van der Waals interactions, and electrostatic contacts. Anfinsen's thermodynamic hypothesis establishes that the amino acid sequence alone encodes the native fold, which corresponds to the global Gibbs free-energy minimum. The folding funnel model resolves Levinthal's paradox by showing that proteins navigate a biased energy landscape rather than searching all conformations randomly.

Protein stability is marginal (typically −20 to −65 kJ mol⁻¹), making proteins sensitive to denaturation by heat, chemical denaturants (urea, GdnHCl), extreme pH, reducing agents, and detergents. Quantitatively, stability is described by ΔG° = ΔH° − TΔS° and probed experimentally using the linear extrapolation model or thermal melting curves. In vivo, molecular chaperones (Hsp70, Hsp60/GroEL, Hsp90) prevent aggregation and assist folding without altering the final structure. Failure of the protein homeostasis network leads to misfolding diseases such as Alzheimer's, Parkinson's, and prion diseases—underscoring the profound biomedical significance of understanding protein folding, stability, and denaturation.