What Bonds Are in Tertiary Protein Structure?

Tertiary protein structure is held together by a mix of noncovalent interactions and, in some cases, covalent bonds between amino acid side chains. The major players are hydrophobic interactions, hydrogen bonds, ionic interactions (salt bridges), van der Waals forces, and disulfide bonds. No single type dominates on its own in every protein; instead, the three-dimensional shape emerges from thousands of these interactions working in concert, with the hydrophobic effect generally considered the largest single driving force behind folding.

Hydrophobic Interactions Are the Main Driver

When a protein folds in water, amino acids with oily, water-repelling side chains cluster together in the protein’s interior, away from the surrounding water. This clustering is called the hydrophobic effect, and researchers widely regard it as the single most important force shaping tertiary structure.¹ The effect is not really an “attraction” between greasy side chains. It is driven by water: water molecules near nonpolar surfaces lose rotational freedom and become more ordered, which is thermodynamically unfavorable. By burying those surfaces inside the protein, the surrounding water regains its disorder, and the overall system reaches a lower energy state.

Molecular simulations bear this out. When researchers compared the behavior of water molecules around folded and unfolded protein states, the folded (native) state showed the least water entropy loss, confirming that folding minimizes the ordering penalty water pays around exposed hydrophobic surfaces.² The upshot is that the hydrophobic effect acts as a kind of general compressive force, pushing the chain into a compact globule. The finer details of shape then depend on the other interaction types.

Hydrogen Bonds Between Side Chains

Hydrogen bonds form when a hydrogen atom bonded to nitrogen or oxygen gets shared with a nearby electronegative atom. In secondary structure (alpha helices and beta sheets), hydrogen bonds between backbone atoms are the defining feature. In tertiary structure, the relevant hydrogen bonds are typically between amino acid side chains, or between a side chain and a backbone atom in a distant part of the chain. These bonds are individually weak compared to covalent bonds, but there are a lot of them, and they impose precise geometric constraints that guide exactly where loops, helices, and sheets sit relative to one another.

A well-studied example comes from the lambda repressor protein. In that protein, the side chain of one aspartate residue forms both a salt bridge with a nearby arginine and a separate hydrogen bond with a serine located much further along the chain. That single serine-aspartate hydrogen bond stabilizes the tertiary fold by roughly 1.5 kcal/mol, a meaningful contribution at the scale of protein stability.³ What makes this interesting is that buried polar residues forming such hydrogen bonds tend to be highly conserved across protein families over evolutionary time, suggesting they serve as hidden structural supports that evolution is reluctant to alter.⁴

Salt Bridges and Ionic Interactions

Salt bridges form between amino acids that carry opposite charges at physiological pH, such as a positively charged lysine or arginine paired with a negatively charged aspartate or glutamate. They are essentially a combination of an electrostatic attraction and a hydrogen bond. Salt bridges are common structural motifs in both secondary and tertiary structure, and they are generally associated with increased stability.⁵

That said, salt bridges are not always stabilizing. The energy balance involves a tug-of-war: the favorable electrostatic attraction between the charged groups competes against the energy cost of burying those charged groups away from water (desolvation). In a large analysis, about 78% of salt bridges in a dataset of proteins turned out to be stabilizing, contributing on average around −3 kcal/mol net energy, while the remaining 22% actually destabilized the protein because the desolvation penalty outweighed the electrostatic gain.⁶ Researchers have also found a strong relationship between where salt-bridging residues appear in protein sequences and the tendency to form salt bridges, suggesting these interactions play a role not just in stabilizing the final fold but in guiding the folding process itself.⁷

Disulfide Bonds

Disulfide bonds stand apart from the other interactions because they are covalent. They form when two cysteine residues, which can be far apart in the amino acid sequence, are brought close enough by folding for their sulfur atoms to link. The result is a strong, permanent cross-link that pins two parts of the chain together. These bonds are fundamental structural elements in many proteins and peptides, particularly those that function outside the cell or as toxins.⁸

Disulfide bonds are especially important for secreted proteins and those exposed to harsh extracellular environments where the gentler noncovalent forces might not be enough. Structurally, they provide extra stability to the native fold and also reduce solvent access to the bond itself, which prevents unwanted chemical exchange reactions.⁹ Inside the cell, the chemical environment is too reducing for most disulfide bonds to form, which is why intracellular proteins rely almost entirely on noncovalent interactions for their tertiary structure.

Van der Waals Forces

Van der Waals forces are the weakest individual interactions in a protein, but they are everywhere. They arise from fleeting fluctuations in electron distribution that create temporary dipoles between atoms in close contact. Each van der Waals interaction contributes a tiny amount of energy, but when the interior of a protein is tightly packed, the sheer number of atom-atom contacts makes the collective contribution substantial.

Research has shown that van der Waals interactions favor the compact intramolecular contacts that characterize folded proteins, partly because the structural geometry of a polypeptide chain is simply a better fit for itself than for surrounding water molecules.¹⁰ In fact, the energy derived from optimized van der Waals packing in a protein’s hydrophobic core has been estimated to be comparable in magnitude to the hydrophobic effect itself, which makes the tightness of internal packing a major evolutionary constraint on protein cores.¹¹ This is why mutations that create cavities in the core, even without changing the overall fold, tend to destabilize a protein significantly.

Aromatic Stacking and Cation-Pi Interactions

Amino acids with aromatic rings (phenylalanine, tyrosine, tryptophan) can interact with each other through pi-stacking, where the electron clouds of their ring systems overlap favorably. They can also form cation-pi interactions with positively charged side chains like lysine and arginine, where the cation is stabilized by the negatively charged face of the aromatic ring.

The strength of these interactions in real proteins has been debated. In one study using a model peptide in water, only the arginine-phenylalanine pairing provided significant stabilization, and even then the researchers concluded that hydrophobic packing was more likely responsible than a pure cation-pi effect. The stabilizing contribution of cation-pi interactions in that context was much smaller than computational models had predicted.¹² However, in membrane environments the story changes. When researchers tested combinations of aromatic and cationic residues in transmembrane helices, they found that cation-pi interactions, particularly those involving lysine paired with tryptophan, tyrosine, or phenylalanine, significantly enhanced helix-helix association, sometimes forming complexes four times stronger than a well-known high-affinity dimer.¹³ The takeaway is that these weaker interactions are context-dependent: modest in water-soluble proteins, potentially quite significant in membranes where competition from water is reduced.

Metal Ion Coordination

Some proteins use metal ions, most commonly zinc, as structural anchors. The metal ion sits at the junction of several amino acid side chains (typically cysteines and histidines), and the coordination bonds to the metal hold distant parts of the chain in fixed positions relative to each other. Zinc finger proteins are the classic example: a zinc ion coordinated by four residues creates a compact, stable domain that would not fold properly without the metal.

Structural studies have shown that zinc can tie an alpha helix in place with a single coordination link, but to firmly constrain the helix’s orientation, two attachment points are needed.¹⁴ In the LIM domain protein family, each domain binds two zinc ions in tetrahedral coordination sites, one using four sulfur atoms from cysteines and one using three sulfurs plus a nitrogen from a histidine. Without metal binding, these proteins appear substantially unfolded, confirming that the zinc coordination is not merely decorative but essential for establishing the tertiary fold.¹⁵

Why the Overall Balance Is So Delicate

A common surprise for newcomers to protein biochemistry is how marginal the net stability of a folded protein actually is. Since at least the late 1950s, researchers have recognized that the folding free energy of a typical protein represents a small difference between two large opposing factors, each contributing roughly 5 kilojoules per residue: the hydrophobic effect pulling the protein together and the loss of chain flexibility (configurational entropy) resisting compaction.¹⁶ The net stability of an entire folded protein is often just 20 to 60 kJ/mol total, which is equivalent to the strength of a few hydrogen bonds. This is why a single mutation in the protein interior can sometimes be enough to destabilize the entire structure.

The cell accounts for this fragility by deploying molecular chaperones, families of helper proteins that assist newly made proteins in reaching their correct fold, protect partially assembled complexes from heat damage, and prevent aggregation.¹⁷ Chaperones do not dictate the final structure; they create a protected environment that gives the protein’s own sequence-encoded interactions the best chance to find the correct arrangement.

When These Bonds Are Disrupted

Denaturing agents like urea and guanidinium chloride are widely used in the laboratory to unfold proteins. Despite being structurally similar molecules, they work through different mechanisms. Urea denatures proteins by forming hydrogen bonds directly with the peptide backbone, competing with the intramolecular hydrogen bonds that hold the structure together. Guanidinium, surprisingly, does not appear to work through hydrogen bonding at all, even though it seems like it should. Its denaturing mechanism remains a separate, less well-understood process.¹⁸ Heat, extreme pH, and organic solvents also disrupt tertiary structure by shifting the balance of the noncovalent forces described above. Covalent disulfide bonds, being much stronger, generally survive conditions that destroy noncovalent structure, which is one reason reducing agents are often added alongside denaturants to achieve complete unfolding.

Misfolding and Disease

When the tertiary interactions go wrong, partially folded or misfolded intermediates can expose hydrophobic surfaces that are normally buried. These sticky surfaces tend to attract each other, causing the protein molecules to clump into aggregates. Environmental stress, mutations, oxidative damage, and errors in post-translational processing can all trigger misfolding.¹⁹ In the worst cases, misfolded intermediates self-assemble into amyloid fibrils, highly ordered aggregates rich in beta-sheet structure that accumulate as toxic deposits. Amyloid deposits in the brain and nervous system are linked to Alzheimer’s disease, Parkinson’s disease, and prion diseases.

Transthyretin, a blood protein, provides a particularly well-characterized example. Under mildly acidic conditions, it dissociates from its normal four-subunit assembly into monomers that retain substantial beta-sheet content but adopt a non-native tertiary structure. This altered monomer is the amyloidogenic intermediate that goes on to self-associate into amyloid fibrils through a series of intermediate assemblies.²⁰ The key insight is that it is not total unfolding that leads to disease; it is partial unfolding that exposes surfaces normally hidden by correct tertiary interactions.

Engineering Tertiary Bonds

Because disulfide bonds are covalent, they are the most straightforward tertiary interaction to engineer deliberately. Researchers routinely try to stabilize proteins by inserting new cysteine pairs at positions that should be close in the folded structure. This approach does not always work, though, because a cross-link stabilizes both the folded and the unfolded state. Whether you get a net gain in stability depends on which state benefits more. In one systematic study using a small trypsin inhibitor, one engineered disulfide raised the melting temperature by 17°C while another, different pairing actually decreased overall stability, because it preferentially stabilized the unfolded form.²¹

More recent work has pushed beyond natural cysteine chemistry. One group used noncanonical amino acids to create “extended” disulfide bonds, cross-links that span a slightly longer distance than traditional cysteine-cysteine bridges. They identified a mutant of the enzyme beta-lactamase that was stabilized by about 9°C through a single such extended bond, demonstrating that expanding the chemical toolkit beyond the standard 20 amino acids opens up stabilization strategies that conventional mutagenesis cannot achieve.²² Similarly, cysteine-engineered versions of a fungal uricase showed enhanced structural rigidity in molecular dynamics simulations without compromising the protein’s overall shape or flexibility.²³ These engineering successes underscore that understanding which bonds maintain tertiary structure is not just an academic exercise; it has practical applications in designing more stable therapeutic proteins and industrial enzymes.

Post-Translational Modifications Can Reshape the Fold

After a protein is synthesized, the cell can chemically modify its side chains through phosphorylation, glycosylation, acetylation, methylation, and many other reactions. These post-translational modifications change the charge, size, or polarity of specific residues, which in turn shifts the balance of tertiary interactions.²⁴ A phosphate group added to a serine, for instance, introduces a bulky negative charge that can create new salt bridges, disrupt existing hydrogen bonds, or repel neighboring negative residues.

Interestingly, structural comparisons suggest that most modifications do not dramatically overhaul the entire protein. Only about 7% of glycosylated proteins and 13% of phosphorylated proteins show global structural shifts larger than 2 angstroms. Phosphorylation actually appears to stabilize tertiary structure on average, reducing global conformational variability by around 25%.²⁵ The modifications are more like fine-tuning knobs than on-off switches, adjusting local and sometimes distant parts of the structure through subtle allosteric effects rather than wholesale refolding.

Allosteric Communication Through Tertiary Contacts

Tertiary interactions do more than hold a static shape. They also form the physical pathways through which information travels within a protein. When a molecule binds at one site on a protein and changes the protein’s behavior at a distant site, that is allostery, and it works because the network of tertiary contacts transmits conformational shifts across the structure like a chain of falling dominoes.

Computational analysis of allosteric proteins has revealed that communication between a substrate-binding site and a distant regulatory site relies on networks of both tertiary and quaternary (subunit-subunit) motions. In a study of 15 allosteric protein systems, many of the individual communication pathways between substrate and effector sites were “interdependent,” meaning they could not be explained by tertiary contacts alone or by subunit rearrangements alone but required both working together.²⁶ This means the noncovalent bonds that define tertiary structure are not just passive glue. They are an active communication infrastructure, and changes to even a single interaction in a critical pathway can alter the protein’s functional response to signals.