Globular Proteins

Drop a spoonful of egg white into water and it dissolves into a clear, faintly viscous solution. Thousands of ovalbumin molecules, each folded into a compact particle roughly 3 nm across, tumble freely among water molecules. Heat that same solution and the clear liquid turns opaque white. The folded particles unravel, and their formerly buried oily interiors flip outward. The exposed hydrophobic patches stick together into an insoluble tangle. This everyday kitchen transformation is a direct demonstration of what makes a protein "globular" in the first place. It also shows why that architecture is fragile in a very specific, informative way. Globular proteins are the workhorses of the soluble proteome. They include enzymes, transport carriers, immune receptors, signaling hubs, and chaperones. All of them share one architectural strategy. They fold the polypeptide chain into a roughly spherical or ellipsoidal shape. Nonpolar side chains are sequestered in a tightly packed core. Polar, charged side chains are presented to solvent on the surface. This is the physical logic that lets a protein be soluble in cytoplasm or blood plasma. At the same time, it creates buried, precisely shaped pockets — active sites, binding clefts, allosteric interfaces — where chemistry actually happens. The contrast with fibrous proteins (keratin, collagen, fibroin) is instructive. Those extend indefinitely along one axis and repeat a simple secondary-structure motif. They serve mechanical rather than catalytic or transport roles. Globular proteins instead mix helices, sheets, loops, and turns into an irregular, information-rich three-dimensional object. Its sequence-specified shape is its function. The big picture connects directly to ideas about Tertiary Structure and the Fold. It also connects to the classification schemes used to catalog folds in Protein Domains and Structural Classification. Globular proteins are the physical objects that those frameworks describe. What a systematic treatment of globular architecture adds is the energetics. It explains why burial of nonpolar surface area is thermodynamically favorable. It explains why a marginal stability of only 20–60 kJ/mol separates a working enzyme from a random coil. And it explains why that marginal stability is not a design flaw but a functional necessity. It allows conformational breathing, allosteric motion, and controlled turnover. This article works through the physical chemistry of compact folding. Topics include the hydrophobic effect as an entropic phenomenon. It covers the packing density and radius-of-gyration scaling that define "globular." It also covers the thermodynamic bookkeeping of folding free energy. Finally, it introduces the structural vocabulary — core, surface, domains, quaternary assembly — that lets you read a fold and predict its behavior. It closes by setting up the transition to a specific, exhaustively studied globular protein — myoglobin — whose oxygen-binding chemistry is the natural next stop.

Key takeaways

Globular proteins fold into compact, roughly spherical shapes (radius of gyration scaling as $R_g \propto N^{0.4}$ ) that bury nonpolar side chains and expose polar ones to water.
Folding is driven mainly by the hydrophobic effect — an entropic phenomenon caused by ordered water around nonpolar surfaces being released upon burial.
Net folding stability ( $\Delta G_{fold}$ ) is a small, marginal difference (20–60 kJ/mol) between large favorable and unfavorable terms, making native structure only barely more stable than the unfolded state.
Core packing density approaches that of organic crystals (~0.74), which is why buried mutations that swap side-chain volume are often destabilizing even without changing polarity.
Globular folds organize into independently folding domains, and multiple polypeptide chains often assemble into quaternary structures stabilized by the same hydrophobic-burial logic.
Fibrous proteins repeat one secondary-structure motif for mechanical function; globular proteins mix motifs into irregular, compact folds built for chemistry.
Key idea: solubility and catalytic specificity both come from the same act — burying nonpolar surface area to release ordered water, while leaving a precisely shaped, chemically active surface exposed.

Where this lives today

Globular protein architecture is the direct target of modern rational drug design: nearly every small-molecule drug that acts on an enzyme or receptor binds inside a pocket created by the folded globular tertiary structure, and computational tools (AlphaFold, Rosetta) now predict these compact folds from sequence with near-experimental accuracy, accelerating structure-based drug discovery. Protein engineering and industrial biocatalysis exploit exactly the packing and hydrophobic-burial logic developed here: directed evolution campaigns routinely optimize buried-core mutations to raise $\Delta G_{fold}$ for industrial enzymes that must survive high-temperature or high-solvent conditions in biofuel and detergent manufacturing. In medicine, diseases of globular misfolding — from sickle-cell hemoglobin's surface-patch aggregation to the amyloid-forming globular-to-fibrillar transitions in Alzheimer's-associated amyloid-β and transthyretin amyloidosis — are understood mechanistically as failures of the same core-burial/surface-exposure logic that ordinarily keeps globular proteins soluble and monomeric. In synthetic biology, de novo protein design (the Baker lab and others) now routinely builds novel globular folds from scratch by directly encoding the packing rules described here — specifying a hydrophobic-core sequence pattern first, then filling in a polar surface, and verifying the intended fold appears exactly as predicted. And the immediate conceptual next step from this architecture — how a specific globular fold creates a functional ligand-binding pocket — is realized concretely in Myoglobin Structure and Oxygen Binding, where the same buried hydrophobic pocket principle creates a heme-binding site tuned for reversible, controlled oxygen capture.

Ask the tutor: How does the same hydrophobic-burial logic explain the specific oxygen-binding pocket geometry of myoglobin's heme site?

The physical picture — thorough prose

A globular protein is best pictured not as a rigid marble but as a densely packed oil droplet with a polar skin, held together by thousands of weak, individually reversible interactions that collectively resist thermal disruption. Three structural facts define the "globular" state and distinguish it from the extended, repetitive architecture of fibrous proteins.

First, compactness. An unfolded polypeptide chain behaves statistically like a random coil, and its radius of gyration scales with the number of residues $N$ as $R_g \propto N^{0.6}$ (the Flory exponent for a self-avoiding chain in a good solvent). A folded globular protein instead scales as

R_g \approx 2.24\, N^{0.392}\ \text{Å}

a much weaker dependence on chain length, reflecting three-dimensional packing rather than one-dimensional extension. A 150-residue globular domain typically has $R_g \approx 14$ –16 Å and an overall diameter of 30–40 Å — genuinely small, comparable to the diameter of the ribosome's mRNA channel or a few turns of B-DNA.

Second, core burial and packing density. Nonpolar side chains (Leu, Ile, Val, Phe, Met, Trp, Ala) preferentially occupy the interior, while charged and polar side chains (Lys, Arg, Glu, Asp, Gln, Asn, Ser, Thr) dominate the surface, with occasional buried polar residues almost always neutralized by an intramolecular hydrogen bond. The interior is not merely "hydrophobic" — it is packed to a density close to that of a crystalline organic solid, with a packing coefficient (fraction of core volume actually occupied by atoms, versus voids)

C_{pack} \approx 0.72\text{–}0.76

This is nearly the packing density of close-packed spheres (0.74) and far higher than the packing density of a liquid alkane (~0.6). This tight packing is why buried cavity-creating mutations (e.g., Leu → Ala at a core position) are so often destabilizing: removing volume without filling it leaves a void, and voids cost enthalpy by losing van der Waals contacts.

Third, mixed secondary structure organized into domains. Unlike keratin's near-uniform α-helical coiled coil or silk fibroin's β-sheet stacks, a typical globular protein interleaves α-helices, β-strands, turns, and loops in an irregular topology dictated by the specific sequence. This topology is usually decomposable into one or more independently-folding structural domains, each roughly 50–200 residues, each capable in principle of folding on its own. Multidomain globular proteins arrange these units like beads, sometimes rigidly packed against each other (forming a single functional unit, as in most enzymes) and sometimes connected by flexible linkers (permitting large-scale domain motions, as in kinases or multi-domain transcription factors).

Figure

Cross-section of a globular protein: buried hydrophobic core vs. polar solvent-exposed surface

Water plays an active, not merely passive, role in this architecture. Around an isolated nonpolar solute, water molecules cannot form their full complement of hydrogen bonds in random orientations; they instead adopt a partially ordered "clathrate-like" cage to preserve hydrogen bonding, which lowers the entropy of that shell of water relative to bulk. When two nonpolar surfaces come together — as when a protein folds and buries Leu against Ile — the ordered water released from the now-merged, now-smaller surface returns to the bulk, and its entropy rises. This is the hydrophobic effect, and it is fundamentally an entropic driving force from the water's perspective, even though it manifests as a favorable (negative) free-energy change for burying nonpolar surface area:

\Delta G_{transfer} \approx \gamma \, \Delta A_{nonpolar}

with an empirical proportionality constant $\gamma \approx 25$ cal/(mol·Å²) (about 100 J/(mol·Å²)) for transferring nonpolar surface area from water to a nonpolar environment (see worked calculation below).

Figure

Ordered water around exposed hydrophobic surface vs. released water upon burial

The core ideas — governing relationships

The net stability of a folded globular protein is a difference of large numbers, and every term matters:

\Delta G_{fold} = \Delta H_{fold} - T\Delta S_{fold}

where:

$\Delta H_{fold}$ includes favorable van der Waals packing and hydrogen bonding in the folded core, opposed somewhat by the enthalpy of breaking protein–water hydrogen bonds
$\Delta S_{fold}$ has a large unfavorable (negative) chain-conformational-entropy term (the backbone loses freedom on folding) opposed by a favorable solvent entropy term (the hydrophobic effect)
$T$ is absolute temperature, and $\Delta G_{fold}$ is defined as $G_{folded} - G_{unfolded}$ , so a stable protein has $\Delta G_{fold} < 0$

The scale of the individual terms dwarfs the net result. For a typical 150-residue protein, the unfavorable conformational entropy of folding is roughly

-T\Delta S_{conf} \approx +(20\text{ to }30)\,RT \text{ per residue-scale estimate, summing to hundreds of kJ/mol}

while favorable van der Waals and hydrophobic terms are comparably large and opposite in sign, leaving a net $\Delta G_{fold}$ of only

\Delta G_{fold} \approx -20 \text{ to } -60\ \text{kJ/mol}

— comparable in magnitude to just a handful of hydrogen bonds or a few van der Waals contacts. This is the origin of "marginal stability": globular proteins sit only barely on the folded side of the folding/unfolding equilibrium, which is exactly what allows conformational flexibility, ligand-induced fit, allosteric switching, and regulated proteolytic turnover — a protein locked into concrete-like rigidity could do none of these. The folding equilibrium constant and free energy connect through the standard relation

\Delta G_{fold} = -RT\ln K_{fold}, \qquad K_{fold} = \frac{[N]}{[U]}

where:

$[N]$ , $[U]$ are the equilibrium concentrations of native and unfolded states
$R = 8.314\ \text{J mol}^{-1}\text{K}^{-1}$ , $T$ in kelvin

Reading this: a $\Delta G_{fold}$ of $-30$ kJ/mol at 298 K gives $K_{fold} = e^{30000/(8.314\times298)} \approx e^{12.1} \approx 1.8\times10^5$ , meaning roughly 1 unfolded molecule per 180,000 folded ones at equilibrium — apparently overwhelming stability, yet a modest destabilizing mutation of only +10 kJ/mol (easily achieved by one poorly packed core substitution) drops that ratio by a factor of $e^{10000/2478}\approx 55$ , illustrating how sensitive globular architecture is to local packing defects.

Derivation AHydrophobic driving force from buried surface area

Goal: estimate the favorable free-energy contribution from burying nonpolar surface area when a chain folds, and show it is large enough to plausibly drive folding despite the unfavorable conformational entropy cost.

Step 1 — Define the empirical transfer free energy. Experiments transferring small nonpolar molecules (e.g., alkanes, benzene) from water into a nonpolar solvent give a free energy of transfer per unit exposed nonpolar area,

\Delta G_{transfer} = -\gamma\, \Delta A_{np}

where $\gamma \approx 25\ \text{cal mol}^{-1}\text{Å}^{-2} \approx 105\ \text{J mol}^{-1}\text{Å}^{-2}$ , and $\Delta A_{np}$ is the nonpolar surface area removed from water contact.

Step 2 — Estimate the nonpolar surface area buried on folding. For a 150-residue globular protein, roughly 55–60% of the total accessible surface area of the fully unfolded chain is nonpolar, and folding buries on the order of 60–70% of that nonpolar area (some nonpolar surface remains exposed even in the folded state, e.g. on surface loops). A typical unfolded 150-residue chain has total accessible surface area

A_{unfolded} \approx 1.48\times 10^2 \times N^{0.75}\ \text{Å}^2\ (\text{empirical scaling}),\qquad N=150

A_{unfolded} \approx 148\times150^{0.75}\approx 148\times 41.6 \approx 6.2\times10^{3}\ \text{Å}^2

Taking 57% nonpolar: $A_{np,unfolded}\approx 0.57\times 6200\approx 3.5\times10^{3}\ \text{Å}^2$ . If folding buries 65% of that:

\Delta A_{np}\approx 0.65\times 3500\approx 2.3\times10^{3}\ \text{Å}^2

Step 3 — Compute the hydrophobic free-energy contribution.

\Delta G_{hydrophobic}\approx -\gamma\,\Delta A_{np} = -(105\ \text{J mol}^{-1}\text{Å}^{-2})(2.3\times10^{3}\ \text{Å}^2)

\Delta G_{hydrophobic}\approx -2.4\times10^{5}\ \text{J/mol} = -240\ \text{kJ/mol}

Step 4 — Compare to the net observed stability. This hydrophobic term alone (−240 kJ/mol) is roughly 5–10 times larger in magnitude than the net observed $\Delta G_{fold}$ of a typical protein (−20 to −60 kJ/mol). The remainder must be an unfavorable conformational-entropy cost of comparable size, roughly +180 to +220 kJ/mol, confirming that folding stability is indeed a small difference between two large, opposing terms — exactly the marginal-stability picture asserted above. This also shows why a single point mutation exposing or burying even a modest patch of nonpolar area (say, 100 Å²) can shift $\Delta G_{fold}$ by $\sim 10$ kJ/mol, comparable to the entire native margin of stability.

Go deeper · Derivation B · 1 minRadius of gyration scaling: globular vs. random-coil chains

Derivation BRadius of gyration scaling: globular vs. random-coil chains

Goal: derive and contrast the scaling exponents that make a folded chain "compact" (globular) versus an unfolded chain "extended" (random coil), and use real numbers to show the two states are structurally distinguishable well before any spectroscopic probe is applied.

Step 1 — Random coil scaling. For a self-avoiding polymer chain in a good solvent (a reasonable model for a chemically denatured polypeptide in water), the radius of gyration follows the Flory scaling law

R_g^{unfolded} = R_0\, N^{\nu}, \qquad \nu \approx 0.598 \text{ (empirically } \approx 0.6\text{)}

with $R_0 \approx 1.9$ – $2.0$ Å for polypeptides (fit from small-angle X-ray scattering, SAXS, of chemically denatured proteins).

Step 2 — Globular (native) scaling. For folded globular proteins, empirical SAXS/crystallographic surveys give

R_g^{native} \approx 2.24\, N^{0.392}\ \text{Å}

Notice the exponent is close to 1/3, the value expected for a uniform-density three-dimensional sphere ( $R_g \propto V^{1/3} \propto N^{1/3}$ if mass is proportional to $N$ and density is constant) — a compact object grows in radius much more slowly with added residues than a chain does, since added residues fill in volume rather than extending length.

Step 3 — Evaluate both at N = 150 residues. Random coil (denatured):

R_g^{unfolded} = 1.9\times 150^{0.598}

150^{0.598} = e^{0.598\ln 150} = e^{0.598\times5.011} = e^{2.997}\approx 20.0

R_g^{unfolded}\approx 1.9\times20.0\approx 38\ \text{Å}

Native globular fold:

R_g^{native} = 2.24\times150^{0.392}

150^{0.392}=e^{0.392\times5.011}=e^{1.964}\approx7.13

R_g^{native}\approx2.24\times7.13\approx16.0\ \text{Å}

Step 4 — Interpret the contrast. The native state ( $R_g\approx16$ Å) is roughly 2.4-fold more compact by radius, or about $(38/16)^3\approx 13$ -fold more compact by volume, than the unfolded chain at identical chain length. This is precisely the observable SAXS/gel-filtration signature used to detect folding transitions in vitro: a two-state folding/unfolding titration (by urea or temperature) shows $R_g$ jump discontinuously between these two scaling curves, with no stable intermediate radius for most small single-domain proteins — direct structural evidence for the classic two-state model of protein folding thermodynamics.

Limiting case check. As $N\to1$ (a single residue), both expressions should converge toward a small, comparable size since a monomer cannot be meaningfully "coiled" or "globular" — indeed $R_g^{unfolded}(1)\approx1.9$ Å and $R_g^{native}(1)\approx2.24$ Å, both physically sensible amino-acid-scale distances, confirming the empirical prefactors are consistent at the boundary of applicability, even though the power-law scaling itself is only meaningful for $N \gtrsim 30$ –50 (below which finite chain-end and hydration effects dominate).

Worked examples

Example 1 (foundational) — Estimating the equilibrium unfolded fraction from a modest destabilizing mutation. A small globular protein has native-state stability $\Delta G_{fold} = -28.0$ kJ/mol at 298 K. A buried Ile → Val mutation (loss of one methyl group's worth of packing) destabilizes the fold by $\Delta\Delta G = +6.5$ kJ/mol (i.e., the new $\Delta G_{fold} = -21.5$ kJ/mol). Find the fractional population of unfolded protein at equilibrium before and after the mutation. Before mutation:

K_{fold} = e^{-\Delta G_{fold}/RT} = e^{28000/(8.314\times298)}

= e^{28000/2477.6} = e^{11.30}\approx 8.06\times10^{4}

f_U = \frac{1}{1+K_{fold}} = \frac{1}{1+8.06\times10^{4}}\approx1.24\times10^{-5}

After mutation ( $\Delta G_{fold}=-21.5$ kJ/mol):

K_{fold}' = e^{21500/2477.6}=e^{8.679}\approx5.87\times10^{3}

f_U' = \frac{1}{1+5.87\times10^{3}}\approx1.70\times10^{-4}

The unfolded fraction rises by a factor of roughly 13.7, from about 1 molecule in 80,000 to about 1 in 5,900 — a dramatic population shift from a single buried side-chain truncation, illustrating exactly why core-packing mutations are so often used experimentally (and clinically implicated in disease) as stability probes.

Example 2 (applied) — Comparing surface charge distribution to explain solubility differences. Hemoglobin (a globular, oxygen-carrying tetramer) and sickle hemoglobin (HbS) differ by a single surface substitution: Glu6 → Val on the β-chain, at a position on the outside of the folded subunit. Wild-type hemoglobin, with Glu6 exposed, presents a negatively charged patch to solvent, contributing to high solubility (~340 mg/mL achievable in red cell cytoplasm). The mutation replaces this charged, hydrogen-bonding side chain with a small nonpolar isopropyl group. Because the substitution is on the surface, it does not perturb core packing density ( $C_{pack}$ stays ≈0.74) or global fold stability much — HbS folds essentially normally and binds oxygen almost normally. The consequence is instead intermolecular: the new exposed hydrophobic Val6 patch on deoxygenated HbS fits into a complementary hydrophobic pocket on a neighboring deoxy-HbS tetramer (formed near residues Phe85 and Leu88 on the β-chain of the adjacent molecule), nucleating a fiber. This is a direct illustration of the article's central logic in reverse: globular solubility depends on keeping nonpolar patches buried or absent from the surface, and a single surface mutation that creates an exposed nonpolar "sticky patch" can convert a soluble globular protein into a polymer-forming one — the molecular basis of sickle cell disease. The take-home distinction: core mutations perturb $\Delta G_{fold}$ (stability); surface mutations perturb solubility and intermolecular association without necessarily perturbing $\Delta G_{fold}$ at all.

Example 3 (exam/competition level) — Packing density and cavity-creation free energy. A core Leu residue (side-chain volume $\approx167$ Å³) is mutated to Ala (side-chain volume $\approx 88$ Å³) in the hydrophobic core of a globular enzyme, and the crystal structure shows the surrounding residues do not rearrange to fill the gap — a genuine internal cavity of volume

\Delta V_{cavity} = 167 - 88 = 79\ \text{Å}^3

is created. Empirically, cavity formation in protein cores costs free energy at a rate of about $\Delta G_{cavity} \approx 24\ \text{cal/mol per Å}^3$ of unfilled volume (from high-pressure protein-folding studies and systematic core-mutation surveys), i.e. $\approx 100$ J/(mol·ų). Estimate the destabilization, and use it to back-calculate the new folding equilibrium constant at 310 K (body temperature) given a wild-type $\Delta G_{fold} = -35.0$ kJ/mol. Step 1 — cavity destabilization:

\Delta\Delta G_{cavity} = 100\ \text{J mol}^{-1}\text{Å}^{-3}\times79\ \text{Å}^3 = 7.9\times10^{3}\ \text{J/mol} = 7.90\ \text{kJ/mol}

Step 2 — new stability:

\Delta G_{fold}^{mut} = -35.0+7.90 = -27.1\ \text{kJ/mol}

Step 3 — equilibrium constants at $T=310$ K ( $RT = 8.314\times310 = 2577.3$ J/mol):

K_{WT} = e^{35000/2577.3}=e^{13.58}\approx7.9\times10^{5}

K_{mut} = e^{27100/2577.3}=e^{10.51}\approx3.66\times10^{4}

Step 4 — ratio of native populations:

\frac{K_{WT}}{K_{mut}} = \frac{7.9\times10^{5}}{3.66\times10^{4}}\approx21.6

So the mutant is roughly 22-fold less "locked" into the folded state at physiological temperature — enough, in many real enzymes, to noticeably lower the apparent melting temperature $T_m$ and to sensitize the protein to intracellular proteolytic quality-control systems, even though the mutation touches neither the active site nor the surface. This exemplifies why folding stability and function are not orthogonal: a completely "silent" core mutation, far from any catalytic residue, can still compromise enzymatic activity in vivo purely through decreased population of the folded, active state.

Limiting cases, scope & common traps

Regime / comparison	Behavior	Why
Small globular domain, $N\lesssim 100$	Often two-state folding, no stable intermediates	Cooperative core too small to support partially-folded states with independent stability
Large multidomain protein	Domains can fold/unfold independently; partial unfolding populates real intermediates	Each domain has its own hydrophobic core and $\Delta G_{fold}$
High salt / high temperature	Hydrophobic effect can strengthen then weaken (non-monotonic with T)	Water's ordering entropy around nonpolar solutes itself depends on T; heat-driven denaturation is enthalpically, not purely hydrophobically, driven at high T
Membrane (integral) proteins	"Inside-out" globular logic: hydrophobic residues face the lipid, polar residues face the aqueous pore/loops	Solvent is nonpolar lipid, not water — burial logic inverts
Fibrous proteins (keratin, collagen)	No defined hydrophobic core; stability from repetitive H-bonding/covalent crosslinks	Function is mechanical, not catalytic; extended, not compact
Very small peptides ( $N<30$ )	Radius-of-gyration scaling laws break down	Insufficient residues for genuine 3D packing statistics; end effects dominate

Common trapMisconception: "hydrophobic" means the core repels water like grease repels a countertop

The favorable free energy of burial is not primarily about nonpolar groups "disliking" water in an enthalpic sense — direct van der Waals interactions between nonpolar solutes and water are actually mildly favorable enthalpically. The driving force is almost entirely an entropy effect: water molecules near an exposed nonpolar surface must adopt a more ordered hydrogen-bonded cage to avoid "wasting" hydrogen bonds, and this ordering is entropically costly. Burial releases that ordered water back into the bulk, raising solvent entropy — even though it costs the polypeptide chain conformational entropy. Confusing this with a simple "oil and water don't mix" enthalpic repulsion is the single most common conceptual error in describing protein folding.

The chemist behind it

Christian B. Anfinsen's classic experiments on bovine pancreatic ribonuclease A, culminating in his 1972 Nobel Prize, established the thermodynamic hypothesis at the heart of this entire article: that the native, globular, folded structure of a protein is determined entirely by its amino acid sequence, representing the lowest accessible free-energy state under physiological conditions — not a structure imposed or "assembled" by cellular machinery. By fully denaturing and reducing ribonuclease A with urea and β-mercaptoethanol (destroying all four native disulfide bonds and abolishing catalytic activity), then slowly removing the denaturant and allowing air reoxidation, Anfinsen showed that the enzyme spontaneously refolded into its correct, fully active, native globular conformation — the disulfides re-forming in the right pairing pattern out of the many statistically possible incorrect pairings. This was direct proof that the compact globular fold is a thermodynamically encoded, self-organizing outcome of sequence, not a template-guided assembly process, and it set the entire subsequent research program of protein folding — energetics, kinetics, and later chaperone biology — on a rigorous physical footing.

Origins · 1 minChristian B. Anfinsen and how this was discovered

Christian B. Anfinsen

1916–1995 · American biochemist, NIH; refolding experiments on ribonuclease A established the thermodynamic hypothesis of protein folding, shared the 1972 Nobel Prize in Chemistry

History

- 1957 — Anfinsen begins systematic denaturation/renaturation studies on bovine pancreatic ribonuclease A.

- 1961 — Anfinsen and coworkers demonstrate full reactivation of reduced, denatured RNase A upon reoxidation, establishing spontaneous refolding.

- 1963 — Kauzmann formalizes the hydrophobic-effect model of protein stability, linking folding thermodynamics explicitly to water structure.

- 1972 — Anfinsen awarded the Nobel Prize in Chemistry for the thermodynamic hypothesis of protein folding.

- 1973 onward — Systematic packing-density and cavity-mutation studies (Richards, Matthews and others) quantify core packing coefficients and cavity destabilization energetics used throughout modern protein engineering.

Why don't globular proteins just maximize hydrophobic burial to become maximally stable?

Maximizing burial would mean maximizing compactness and rigidity, but functional proteins need controlled flexibility — active sites that open and close, allosteric surfaces that shift conformation upon ligand binding, and susceptibility to regulated proteolysis for turnover. Marginal stability (a $\Delta G_{fold}$ of only 20–60 kJ/mol) keeps the protein poised near the folding/unfolding boundary, allowing these functional motions without requiring large energy inputs. Additionally, evolutionary selection pressure typically only demands "stable enough" under physiological conditions, not maximal stability; mutations that increase stability beyond what's needed offer no further fitness benefit and may even be selected against if they reduce catalytic flexibility.

How is a globular protein's compactness actually measured experimentally?

The most direct method is small-angle X-ray or neutron scattering (SAXS/SANS) in solution, which yields the radius of gyration $R_g$ directly from the low-angle scattering profile via the Guinier approximation. Analytical ultracentrifugation and size-exclusion chromatography give complementary hydrodynamic size estimates. X-ray crystallography and cryo-EM give atomic-resolution structures from which packing density, buried surface area, and core composition can be calculated directly by computational surface-area algorithms. Comparing native-state $R_g$ to the random-coil scaling law ( $R_g \propto N^{0.6}$ ) for the same chain under denaturing conditions is the standard way to quantify "how globular" a given fold is.

Are all soluble proteins globular, and are all globular proteins soluble?

Almost all natively-folded, monomeric soluble proteins are globular in the sense described here, since the burial of nonpolar surface is essentially required for solubility in water. However, exceptions exist: intrinsically disordered proteins (IDPs) are soluble without a well-defined compact fold, achieving solubility instead through high net charge and low overall hydrophobicity along an extended, dynamic chain. Conversely, a globular protein can lose solubility despite a compact fold if surface mutations create sticky patches (as in sickle hemoglobin) or if partial unfolding exposes core hydrophobic residues (as in many amyloid diseases), showing that "globular" and "soluble" are correlated but not logically identical.

How does the hydrophobic effect differ from a simple attraction between nonpolar molecules?

A simple attractive force (like van der Waals dispersion) would predict roughly temperature-independent or weakly temperature-dependent behavior. The hydrophobic effect instead shows a distinctive, non-monotonic temperature dependence: the free energy of transferring a nonpolar solute from water to oil is most favorable near room temperature and becomes less favorable (or even unfavorable) at both very low and very high temperatures, a signature of its origin in water's hydrogen-bonding entropy rather than a direct solute-solute attraction. This is why heat denaturation of proteins is possible at all — at sufficiently high temperature, the entropic penalty of ordered hydration water shrinks (because bulk water itself becomes more disordered), weakening the hydrophobic driving force for folding relative to the ever-present unfavorable conformational entropy of the chain.

Why do fibrous proteins like collagen not fold into a globular shape?

Fibrous proteins have repetitive sequences (e.g., collagen's Gly-X-Y repeat) that are structurally and energetically optimized for a single, extended, repeating secondary-structure element — the collagen triple helix — rather than for burying a diverse, sequence-specific set of hydrophobic side chains into a compact tertiary fold. Their function is mechanical load-bearing (tensile strength in collagen, shear resistance in keratin), which favors long fibers with regular crosslinking over compact, chemically versatile pockets. Globular folding requires an irregular sequence with strategically placed nonpolar residues that can only pack correctly in one specific compact arrangement — precisely the sequence complexity that fibrous structural proteins forgo in favor of simple repetition.

What role does the protein surface play beyond just providing solubility?

The globular surface is far from a passive solubility wrapper — it is the primary site of molecular recognition. Surface loops and exposed side chains form antibody epitopes, protein-protein interaction interfaces, allosteric regulatory sites, and the rims of active-site clefts that guide substrates into the buried catalytic pocket. Surface electrostatics (the distribution of Lys, Arg, Glu, Asp) steer diffusion-limited association with charged partners (as in electrostatic steering toward a complementary enzyme or substrate), and surface post-translational modification sites (glycosylation, phosphorylation) are almost always located on exposed loops precisely because they must remain accessible to modifying enzymes.

Does every domain within a multidomain globular protein have its own independent hydrophobic core?

In most well-characterized multidomain proteins, yes — each folding domain (as classified structurally in schemes like Protein Domains and Structural Classification) typically has its own self-contained hydrophobic core and its own characteristic $\Delta G_{fold}$ , which is why isolated domains, expressed and purified alone, often fold correctly and remain stable even without the rest of the protein. Some domains, however, share a single continuous hydrophobic core across a domain interface (domain-swapped or tightly interdigitated architectures), in which case the two domains cannot be meaningfully separated thermodynamically and must be treated as one folding unit despite being classified as separate structural domains by sequence or topology criteria.