Peptide Folding and Tertiary Structure

Why a peptide's shape matters more than its sequence

Sequence is the alphabet. Structure is the sentence. You can hand a ribosome the exact same string of amino acids twice and, depending on what happens in the microseconds after synthesis, end up with two different three-dimensional objects with different affinities, different stabilities, and completely different fates in the body.

That's the short version of why peptide folding keeps pharmacologists busy.

I've been knee-deep in folding thermodynamics for the better part of a decade, and I'll admit I got something wrong for years: I thought of tertiary structure as mostly a protein concern , something you worry about with enzymes and antibodies, not with the 10–50 residue molecules we call peptides. Smaller peptides, I assumed, were too short to fold meaningfully. That framing is mostly wrong, and the past few years of structural data have made that very clear.

The four levels , and where peptides sit

A quick orientation. Protein and peptide structure is described at four levels.

Primary structure is the linear sequence of amino acids linked by peptide bonds , the covalent backbone. It carries all the information needed for folding, but it isn't the fold itself.

Secondary structure is local geometry: alpha-helices and beta-sheets formed by hydrogen bonds between backbone atoms. Many bioactive peptides adopt a partial helix or a beta-turn. These aren't decorative. A helix presents its side chains in a specific spatial pattern; a receptor that evolved to recognise that helix won't bind a disordered coil of the same sequence.

Tertiary structure is the overall three-dimensional fold; how the secondary elements pack together in space, stabilised by side-chain interactions: disulfide bridges, salt bridges, hydrophobic packing, and more hydrogen bonds. For most proteins this is the "final answer." For peptides the picture is more complicated, because short chains exist on a spectrum from fully disordered (intrinsically disordered peptides, or IDPs) to rigidly folded.

Quaternary structure, assemblies of multiple chains, is largely a protein story, though some therapeutic peptides do self-associate, which matters enormously for formulation and bioavailability.

Folding forces: what actually drives a peptide into shape

The textbook answer is "the hydrophobic effect," and that's fair as far as it goes. In an aqueous environment, nonpolar side chains are thermodynamically penalised by contact with water. They cluster inward, and the chain arranges itself to minimise that penalty. The resulting compaction is tertiary structure.

But that's one driver among several. Disulfide bonds between cysteine residues are the most familiar covalent stabiliser; insulin's three disulfide bridges are probably the most cited example in the pharmacology literature, and they're what give the hormone its remarkable stability relative to its size. Electrostatic interactions between charged residues can either stabilise or destabilise a fold depending on pH, which is one reason peptide stability is so pH-sensitive. Van der Waals packing in the hydrophobic core provides incremental stabilisation that adds up. And backbone hydrogen bonds, the same ones defining secondary structure, contribute to the overall free energy of the folded state.

Worth noting: none of these forces is large. That's why peptide folding is reversible and why small changes in temperature, pH, or ionic strength can unfold a peptide. Proteins are more forgiving because more contacts accumulate over longer chains. For a 20-residue peptide, a single point mutation can destabilise the entire fold.

The folding problem: sequence to structure isn't trivial

Here's the thing that kept theoreticians awake for decades. Cyrus Levinthal pointed out in 1969; what became known as Levinthal's paradox, that if a protein explored all possible backbone conformations randomly, it would take longer than the age of the universe to find its native fold. Yet proteins fold in milliseconds. Something guides the search.

The resolution, broadly, is that folding doesn't explore random conformations; it follows a funnel-shaped energy landscape. Early folding events (collapse of the hydrophobic core, formation of key secondary structure elements) constrain the search dramatically. The peptide rolls downhill on this landscape, occasionally trapping in local minima, misfolded states; but generally finding the global minimum, the native fold.

For small peptides the landscape is shallower and the native state less clearly defined. Many bioactive peptides are "floppy" in solution and only adopt a defined structure upon binding a receptor. This is called coupled folding and binding, and it has important pharmacological consequences. The entropic cost of ordering the peptide is paid during binding, which reduces affinity relative to a pre-folded ligand of the same sequence. Medicinal chemists exploit this: conformational constraint via cyclisation, N-methylation, or the introduction of non-natural amino acids pre-organises the peptide and often improves binding affinity and metabolic stability simultaneously.

Tertiary structure and receptor recognition

Receptors are not just sequence readers. A GLP-1 receptor agonist, for instance, engages its receptor through a defined helical segment. The helix positions key residues, particularly the N-terminal region, in precise geometry relative to the receptor binding cleft. Disrupt the helix by replacing a helix-stabilising residue with one that breaks it, and binding drops dramatically even if every "pharmacophore" residue is nominally present. The three-dimensional arrangement is the pharmacophore, not just the residue identity.

This is also why pharmacokinetic behaviour and structure are inseparable. A peptide's tertiary shape determines its hydrodynamic radius; how large it appears to renal filtration machinery. It determines surface exposure of proteolytic cleavage sites. A compact, disulfide-bridged tertiary fold physically occludes peptidases that would otherwise degrade the chain in plasma. The fold isn't decoration; it's defence.

Misfolding and aggregation: the dark side of the funnel

Energy landscapes have local minima as well as global ones. A peptide (or protein) can settle into a misfolded state. For short synthetic peptides this might mean aggregation, chains associating into amorphous precipitates or, more worryingly, into ordered beta-sheet-rich fibrils. Amyloid fibrils are the extreme case: stable, ordered aggregates with characteristic cross-beta architecture that the body cannot easily clear.

For pharmaceutical scientists this is a formulation headache. Aggregated peptide is not just inactive, it can be immunogenic. Regulatory agencies including the TGA expect manufacturers to demonstrate that peptide-based products are free of aggregates and that the fold is the intended one, which is part of why physicochemical characterisation requirements for biologics and peptide therapeutics are so demanding.

Honestly, I find the aggregation literature almost as gripping as a good true-crime episode. The conditions under which perfectly well-behaved peptides suddenly decide to misfold and stack into fibrils have a chaotic, slightly ominous quality. Agitation, trace metal contamination, a fraction of a pH unit; any of them can tip the balance.

Tools for studying peptide tertiary structure

Nuclear magnetic resonance (NMR) spectroscopy is the workhorse for peptides in solution. Unlike X-ray crystallography, NMR captures the ensemble of conformations a peptide adopts in a physiologically relevant solvent, which is more informative for intrinsically disordered peptides. Circular dichroism (CD) spectroscopy gives faster, lower-resolution secondary structure estimates, useful for screening how formulation conditions affect fold. Cryo-electron microscopy has transformed structural biology for large assemblies but isn't quite the tool for isolated short peptides. Computational methods, particularly AlphaFold2 (published in Nature in 2021 by Jumper et al.), have changed the landscape for predicting structure from sequence, though their accuracy for short and disordered peptides remains an active area of research.

A point I'd argue is underappreciated: computational structure prediction is not a substitute for experimental structural data when the molecule is intended for human use. The fold in silico may not match the fold in formulation buffer, let alone in plasma.

Why this matters beyond the lab

Understanding tertiary structure is foundational to understanding the entire pipeline, from synthesis through to whatever regulatory pathway a peptide might eventually enter. In Australia, peptide-based therapeutics face the same rigorous TGA scrutiny as any biological medicine. The Special Access Scheme and Authorised Prescriber framework govern the limited circumstances under which unapproved peptides can reach patients. Critically, many peptides being marketed in Australia for performance or body composition purposes are unapproved therapeutic goods, and supplying them without TGA approval is unlawful regardless of claims about their structural or biological properties. Structural sophistication doesn't confer regulatory legitimacy.

The biochemistry here is genuinely fascinating on its own terms. A short chain of amino acids folding into a precise three-dimensional object in fractions of a second, guided by thermodynamics alone, then presenting exactly the right surface to a receptor; there's something almost improbable about how reliably it works. And when it doesn't work, when the fold goes wrong, the consequences reverberate all the way through to stability, safety, and efficacy.

That's the level at which peptide science actually operates. Sequence is just the starting point.

Sources

• Protein Structure, NCBI Bookshelf (Berg, Tymoczko & Stryer, Biochemistry)
• Highly accurate protein structure prediction with AlphaFold, Nature (Jumper et al., 2021)
• Biologicals Regulatory Framework; Therapeutic Goods Administration (TGA)
• Peptide therapeutics: current status and future directions, NCBI / Drug Discovery Today

, Priya Nandakumar, Pharmacology writer, peptides & PK