What Interactions Occur in the Quaternary Structure?

Quaternary structure arises when two or more protein chains, called subunits, assemble into a functional complex, and the forces holding them together are overwhelmingly non-covalent: hydrogen bonds, hydrophobic contacts, electrostatic attractions, and van der Waals forces. These same interactions that shape a single protein’s fold also operate between subunits, but with important differences in strength, composition, and biological consequence. In some cases, covalent bonds and coordinated metal ions join the mix, and the whole system can be tuned by the cell’s chemical environment in ways that directly control what a protein does.

The Non-Covalent Forces That Dominate Subunit Interfaces

Most of the glue holding a multi-subunit protein together is non-covalent. If you could zoom in on the surface where two subunits meet, you would see a patchwork of complementary shapes and charges. Hydrophobic side chains from each subunit cluster together, burying themselves away from water in much the same way that oil droplets merge. Flanking these hydrophobic cores are hydrogen bonds between polar residues and, at the edges, salt bridges between positively and negatively charged amino acids. Van der Waals contacts fill in the remaining gaps, contributing modest but cumulative attraction wherever atoms from opposite subunits sit close together.

Hydrogen bonds at subunit interfaces behave differently from those buried inside a single chain. Because the two surfaces were not originally designed as a single continuous structure, many polar residues at the interface end up forming weaker hydrogen bonds with each other than they would with water. To compensate, water molecules frequently act as bridges, linking donor-acceptor pairs across the interface that are not perfectly aligned. These bridging water molecules mediate contacts that would otherwise be impossible, connecting pairs that face the wrong direction or that both carry the same charge polarity.¹ The result is a hydrogen-bonding network at the interface that is more flexible, more water-dependent, and less energetically optimal than the network inside a single subunit. This looseness is not a defect. It allows many protein complexes to assemble and disassemble as needed.

The overall character of an interface depends on whether the complex is permanent or fleeting. Permanent complexes, the kind that stay together for the life of the protein, tend to have larger contact surfaces rich in hydrophobic residues. Transient complexes, which form and break apart in response to signals, have smaller interfaces with more polar residues.² The distinction matters: permanent interfaces behave more like the hydrophobic core of a single protein, while transient interfaces resemble the protein’s solvent-exposed surface, making them easier to pry apart when the cell needs to.

How Interface Size Relates to Binding Strength

There is a direct relationship between the amount of surface area buried when two subunits come together and how tightly they bind. Larger interfaces generally mean tighter binding. But the correlation is rough. For any given amount of buried surface, the actual binding strength can vary across four orders of magnitude.³ Two complexes can bury the same number of square angstroms and yet differ by a factor of ten thousand in how easily they fall apart.

What explains the scatter? The chemical makeup of the interface matters, but not in the simple way you might expect. Across large datasets, there is no obvious single chemical property that predicts binding strength once you account for interface size. The geometry matters enormously: how well the two surfaces fit together, how many water molecules get displaced, and how the remaining bridging waters are positioned all contribute. One useful concept is surface energy density, the amount of binding energy per unit of buried area. For smaller interfaces (roughly 500 to 2,000 square angstroms), packing more surface area in actually improves efficiency. Beyond about 2,000 square angstroms, the energy per unit area levels off, as if larger interfaces start running into diminishing returns.³

Metal Ions and Covalent Bonds

Not all quaternary interactions are non-covalent. Disulfide bonds, formed between cysteine residues on different subunits, can lock chains together with a covalent link. Antibodies, for instance, use interchain disulfide bonds to hold their heavy and light chains in place. These bonds provide a baseline of stability that non-covalent forces alone might not achieve, especially in harsh extracellular environments.

Metal ions play a subtler but equally important role. The insulin hexamer, the storage form of insulin in the pancreas, is held together by two zinc ions. Each zinc coordinates with histidine residues from three different insulin subunits, pulling the subunits into the correct geometry. Without zinc, the hexamer falls apart into dimers and monomers. The binding is sequential: the first zinc brings three subunits together, and the second stabilizes the remaining three, with accompanying changes in protonation and water displacement.⁴

The yeast enzyme arginase illustrates how different metals can serve different structural roles within the same protein. In arginase, a tightly bound zinc ion is essential for maintaining the trimer’s shape. Remove the metals by chelation and the trimer breaks down into monomers, entering a reversible equilibrium. Add the zinc back and the trimer locks in place with no detectable dissociation. A more loosely associated manganese ion, by contrast, provides catalytic activity but contributes less to structural integrity.⁵ The same metal can therefore serve an architectural or a catalytic purpose depending on its binding site and affinity.

Allostery and Cooperativity

One of the most consequential features of quaternary structure is that interactions between subunits enable allostery, where binding of a molecule at one site on the complex changes the behavior of a distant site on another subunit. Hemoglobin is the textbook example: when the first oxygen molecule binds to one of hemoglobin’s four subunits, it shifts the subunit interfaces in a way that makes it progressively easier for the remaining subunits to pick up oxygen. This cooperativity, where the binding of one ligand promotes the binding of the next, depends entirely on changes in the contacts between subunits along the way.⁶

The physical picture of how allosteric signals travel through a complex has become clearer in recent years. Rather than a single domino chain of residues, allosteric communication uses networks of rigid-body motions. Groups of residues that move together, sometimes an entire domain, shift relative to their neighbors. When those shifting groups form closed loops of connected motions, the mechanical coupling is strong enough that moving one part of the complex forces another part to respond.⁷

Tryptophan synthase offers a detailed molecular example. This enzyme has two types of subunits (alpha and beta) that catalyze consecutive reactions, connected by a hydrophobic tunnel through which the intermediate product travels. The beta subunit’s active site is tightly coupled to a set of hinge residues near the interface between the two beta subunits. Those hinge residues simultaneously control the opening and closing of the tunnel and coordinate the anti-correlated motion of the two beta chains. From there, motions propagate through interfacial helices to the alpha subunit’s catalytic residues, linking the two active sites into a single coordinated machine.⁸ Without quaternary interactions, these enzymes would function as isolated units, slower and less coordinated.

Temporary Assemblies and Substrate Channeling

Not all quaternary complexes are permanent fixtures. Cells build and dismantle multi-enzyme complexes in response to metabolic demand. These transient assemblies, sometimes called metabolons, bring enzymes that catalyze sequential steps of a pathway into close physical proximity so that the product of one reaction can be handed directly to the next enzyme without diffusing away into the cytoplasm. The process is called substrate channeling, and it can dramatically improve the speed and efficiency of a pathway.

A striking example comes from the citric acid cycle. Structural analysis of the complex formed by malate dehydrogenase, citrate synthase, and aconitase shows that when these enzymes associate, polar residues become concentrated on the surface while nonpolar residues are buried more deeply than usual. Positively charged residues line a groove connecting the active sites of malate dehydrogenase and citrate synthase, creating an electrostatic track that guides the negatively charged intermediate oxaloacetate from one active site to the next.⁹ The quaternary arrangement is not just about proximity; it physically sculpts the path the substrate must take.

A similar story is emerging for the purinosome, a transient multi-enzyme complex that assembles in the cytoplasm to carry out the ten-step synthesis of purine nucleotides. Simulations and kinetic analysis of this complex support cluster channeling, where the spatial organization of the enzymes promotes efficient handoff of intermediates, as the dominant mechanism for moving substrates through the pathway.¹⁰ These metabolons can form and dissolve within minutes, regulated by signals that the cell needs more or less of a particular end product.

How the Environment Tunes Quaternary Structure

Because the forces holding subunits together are mostly non-covalent, they are sensitive to the surrounding chemical conditions. Changes in pH, salt concentration, temperature, and the availability of specific ions can all shift the equilibrium between assembled and disassembled states.

Bovine alpha-crystallin, a protein that helps maintain the transparency of the eye lens, illustrates this vividly. Its quaternary structure disassembles in a multi-step process as pH rises above 8 or ionic strength drops, first losing some subunits to form a smaller but still spherical particle, then reversibly dissociating further above pH 9. Raising the temperature at neutral pH causes a parallel process of slow dissociation and swelling. Adding calcium ions stabilizes the structure so substantially that no dissociation occurs even above pH 10.¹¹

The ability of pH to regulate assembly can also be engineered. Researchers have converted avidin, a protein that normally holds its four subunits together with extreme stability, into a pH-sensitive version by replacing specific interface residues with histidine. Histidine changes its charge near physiological pH, so introducing histidines at the subunit interface made the assembly fall apart or come together depending on whether the environment was slightly acidic or neutral.¹² Experiments like these confirm that the balance of quaternary interactions is fine enough that swapping just a few residues at the interface can make or break an entire complex.

When Quaternary Interactions Go Wrong

If the forces between subunits weaken or shift in the wrong direction, the consequences can be severe. Transthyretin, a protein that normally circulates in the blood as a stable tetramer carrying thyroid hormones, is a case study in how small changes at the quaternary level cause disease. The V122I variant, where a single amino acid is swapped, destabilizes the tetramer’s quaternary structure while leaving each subunit’s individual fold mostly intact. The loosened tetramer is more likely to dissociate into monomers, which can then misfold and aggregate into amyloid fibrils, leading to familial amyloid cardiomyopathy, the most common hereditary transthyretin amyloid disease.¹³

Another transthyretin variant, S112I, is even more dramatically destabilized. At physiological pH, this mutant exists not as a tetramer but as a dimer, which self-assembles into spherical aggregates and kills human neuroblastoma cells in culture.¹⁴ The therapeutic strategy for these diseases, exemplified by the drug tafamidis, is to stabilize the tetramer’s quaternary structure, essentially reinforcing the subunit interfaces so the complex does not fall apart. It is a direct application of everything discussed above: strengthening non-covalent interactions at the interface prevents the cascade of dissociation, misfolding, and aggregation.

Disordered Regions at Subunit Interfaces

Not every protein region involved in quaternary assembly starts out with a defined shape. Some subunit interfaces include stretches that are intrinsically disordered, meaning they lack a stable fold until they encounter their binding partner. When contact is made, the disordered region folds and binds simultaneously, in what is sometimes described as a fold-and-bind mechanism. The binding gains energy both from the new contacts formed and from displacing water molecules that were sitting in thermodynamically unfavorable positions on the partner’s surface.¹⁵

Disordered regions are especially important in proteins that serve as hubs, interacting with many different partners. The nuclear pore protein Nup153, for example, has a long disordered tail containing multiple binding motifs for a transport protein called Kap. Thermodynamic analysis shows a complex arrangement: one high-affinity site for the first binding event, followed by a series of lower-affinity sites that bind additional copies with negative cooperativity, meaning each additional binding event is slightly less favorable than the last.¹⁶ This kind of tunable, graded response would be difficult to achieve with rigid, prefolded interfaces.

Post-Translational Modifications as Interface Switches

Cells have a powerful way to modulate quaternary interactions after a protein has already been made: post-translational modifications. Adding a phosphate group, a methyl group, an acetyl group, or a sugar to an amino acid at or near a subunit interface changes the local charge, size, and hydrogen-bonding capacity of that residue, which can strengthen or weaken the interaction.

Phosphorylation is the best-studied example. Researchers have designed peptide interfaces where adding a phosphate to a specific serine residue changes the binding strength in either direction. In some cases, phosphorylation tightened binding by up to eleven-fold. In others, it weakened binding by up to 180-fold.¹⁷ Both the rate at which subunits come together and the rate at which they fall apart were affected, demonstrating that phosphorylation does not just flip a binary switch but reshapes the entire kinetic landscape of the interaction. In living cells, kinases and phosphatases that add and remove phosphate groups thereby act as real-time controllers of which complexes exist and for how long.

Chaperones and Guided Assembly

Large or complex quaternary structures sometimes cannot assemble on their own. The subunits may be prone to misfolding or aggregation if they are left unassisted. Cells solve this problem with molecular chaperones, proteins whose job is to guide other proteins toward their correct fold and correct partners.

The assembly of G protein complexes, which relay signals from cell-surface receptors, is a well-characterized example. The beta subunit of a G protein has a complex propeller-like fold that requires the assistance of the cytosolic chaperonin CCT. Once CCT helps the beta subunit achieve its proper shape, a co-chaperone called PhLP1 releases it from CCT and facilitates its pairing with the gamma subunit.¹⁸ Without this chaperone relay, the beta subunit would likely aggregate before finding its partner, and the downstream signaling pathway would never form. This illustrates that the interactions in quaternary structure are not always spontaneous; the cell invests energy and dedicated machinery to ensure they happen correctly.

Phase Separation and Condensate Formation

A frontier in quaternary structure research is the realization that some assemblies do not form discrete, stoichiometric complexes at all. Instead, multivalent interactions among proteins drive liquid-liquid phase separation, where proteins concentrate into droplet-like condensates that coexist with the surrounding dilute solution. These condensates are not disordered blobs; their formation depends on specific quaternary interactions.

In the inner ear, a group of proteins associated with Usher syndrome (a cause of combined deafness and blindness) assemble into condensates through a combination of interactions: dimerization domains, PDZ-binding motifs, and other multivalent contacts. When the proteins WHRN and USH2A-CT are mixed, they form spherical, liquid-like droplets enriched in both proteins, behavior that neither protein shows on its own.¹⁹ This kind of quaternary assembly depends on having multiple weak interaction sites per molecule, so that each protein can simultaneously contact several partners, building up an extended network rather than a neat complex of defined size.

A related example comes from spider silk. The aciniform spidroin protein uses both hydrophobic and electrostatic interactions in its disordered linker and terminal domains to drive a hierarchical assembly process through phase separation.²⁰ These condensates represent a different paradigm from the classical picture of a hemoglobin tetramer or an enzyme dimer: the same fundamental forces are at work, but the architecture is open-ended and fluid.

How Scientists Study Quaternary Interactions

Characterizing the interactions that hold subunits together requires tools that can see both the shapes and the forces involved. X-ray crystallography and cryo-electron microscopy provide atomic or near-atomic snapshots of complexes, revealing which residues contact each other and how the surfaces fit together. Cryo-EM has been particularly revealing for complexes that resist crystallization. For the bacteriophytochrome from Deinococcus radiodurans, cryo-EM showed that the dimer interface was not limited to the C-terminal region, as long assumed, but also included contacts from the N-terminal light-sensing domain, fundamentally revising the structural model.²¹

Native mass spectrometry has emerged as a complementary approach. By gently transferring intact protein complexes into the gas phase, native MS can measure the mass of the whole assembly, determine how many copies of each subunit are present, and even probe which subunits sit next to each other by inducing partial dissociation.²² Because the technique preserves non-covalent interactions, it captures assemblies that might fall apart under harsher analytical conditions.

Designing Quaternary Interactions from Scratch

Understanding what holds subunits together has opened the door to designing entirely new protein assemblies that do not exist in nature. Computational tools now allow researchers to sculpt protein surfaces so that specific subunit interfaces form with predictable strength and geometry. One approach, learned surface fingerprints, captures the physical and chemical features that determine whether two protein surfaces will recognize each other, enabling the design of novel interactions from first principles.²³

AI-driven tools like ProteinMPNN have been experimentally validated on designing proteins with novel folds and large symmetrical assemblies.²⁴ These designed assemblies range from simple dimers to elaborate cages with dozens of subunits arranged in precise geometric patterns. The fact that researchers can now build quaternary structures to order, choosing which interactions to include and how strong to make them, confirms that the principles governing natural assemblies are well enough understood to be applied as engineering rules. It also suggests a future where custom protein machines, built from designed quaternary interactions, serve as drug delivery vehicles, biosensors, and synthetic catalysts.