How Peptides Degrade

Published August 2026 · Patriot Labs Research Team · ~11 min read

Degradation is not one process

“Degraded” is a single word covering at least half a dozen unrelated reactions. A peptide can lose integrity because a water molecule cut its backbone, because a side-chain amide fell off, because a sulfur atom picked up an oxygen, or because the molecule never changed chemically at all and simply stuck to its neighbours. Those are different events with different triggers, different rates and different analytical signatures.

The useful split is chemical versus physical. Chemical degradation makes covalent changes: bonds break, atoms are added or removed, and the product is a genuinely different molecule. Physical degradation leaves the covalent structure alone and changes the state — folded to partially unfolded, monomer to dimer to visible particle. Reviews of protein pharmaceutical stability treat these as coupled rather than independent, because a partially unfolded molecule exposes side chains that were previously buried and therefore chemically protected.¹

Almost everything that follows depends on three variables: how much water is present, what temperature the sample is held at, and what else is in the vial.

Hydrolysis: water cutting the backbone

The amide bond that links one residue to the next is unstable with respect to hydrolysis. It survives because the reaction is slow, not because it is unfavourable. Put a peptide in water and, given enough time and thermal energy, the chain will be cut. This is why a reconstituted vial has a working life measured in weeks while a dry one is measured in years.

Hydrolysis is acid- and base-catalysed at the extremes and slowest somewhere in the middle. Some positions are far more labile than the average. Aspartate is the notorious one. Where Asp sits next to proline, acid conditions promote formation of a cyclic imide that resolves by cleaving the chain, which is why Asp-Pro is treated as a designed-in weak point in a sequence.² Asp-Gly behaves similarly through the same cyclic intermediate.

The kinetics were mapped carefully in a model hexapeptide, Val-Tyr-Pro-Asp-Gly-Ala. Between roughly pH 0.3 and 3.0 the dominant event was specific acid-catalysed hydrolysis of the Asp-Gly amide bond alongside cyclic imide formation; between pH 4 and 5, cleavage and isomerisation ran in parallel; above pH 6 the chain stopped breaking and isomerisation took over, producing only isoaspartyl product. The ionised form of the aspartyl side chain was substantially more reactive than the neutral form.³ That single study is a good mental model for the whole class: the pathway a peptide takes is not fixed, it is selected by pH.

In dry powder, hydrolysis is nearly absent. Not because the bond became stronger but because the reactant — water — is no longer there in a mobile, reactive form. This is the entire logic behind the storage hierarchy set out in how to store research peptides: dry beats cold, and dry-and-cold beats both.

Deamidation: the charge change nobody sees

Asparagine and glutamine carry a neutral amide on the side chain. Deamidation converts that amide to a carboxylic acid, turning Asn into aspartate and Gln into glutamate. Under acidic conditions this can happen by direct hydrolysis, but at neutral and alkaline pH the dominant route is intramolecular: the backbone nitrogen of the following residue attacks the side-chain carbonyl, expelling ammonia and forming a cyclic succinimide, which then opens by hydrolysis at either of two carbonyls to give a mixture of normal aspartate and isoaspartate.⁴

Two consequences fall straight out of the mechanism. First, Asn deamidates faster than Gln, because Asn forms a five-membered succinimide ring while Gln must form a strained six-membered one.⁴ Second, the reaction is exquisitely sequence-dependent, because the attacking nitrogen belongs to the next residue and steric bulk there slows everything down. Glycine, having no side chain at all, offers no hindrance — which is why Asn-Gly is the textbook hot spot. Serine, threonine and aspartate following Asn also accelerate the reaction substantially.⁴

The scale of the sequence effect is not subtle. A mass-spectrometric survey measured deamidation rates across 306 asparaginyl sequences, and showed that folded structure slows the reaction by restricting the backbone geometry the cyclisation requires.⁵ Broader stability reviews describe deamidation and the related aspartate isomerisation as governed jointly by sequence, pH, temperature and local structure.¹

The analytical implication is worth dwelling on. Losing an amide and gaining a carboxylate changes the nominal mass by only about one dalton — NH₂ leaves, OH arrives — which on a molecule of several thousand daltons is easy to miss at low resolution. What it changes dramatically is charge: a neutral side chain becomes a negative one at working pH. A deamidated peptide is therefore often near-invisible by mass and obvious by charge-based separation. Isoaspartate is worse again: it is isobaric with aspartate and inserts an extra methylene into the backbone, altering shape without altering mass at all.

Oxidation: oxygen, light and trace metals

Oxidation attacks side chains rather than the backbone, and it is selective. Methionine and cysteine are the most reactive, with histidine, tryptophan and tyrosine close behind.⁴ Methionine oxidises to the sulfoxide and, further, to the sulfone; the second step is effectively irreversible. Cysteine oxidation forms disulfides, is accelerated at higher pH, and is at least chemically reversible with reducing agents such as DTT or TCEP.²

Three triggers matter, and they compound.

Dissolved oxygen. Ordinary aerated water carries enough O₂ to support slow oxidation indefinitely. Formulation practice responds by purging headspace with nitrogen, argon or helium and minimising oxygen contact during processing.⁴ A vial that has been repeatedly opened has been repeatedly re-aerated.

Light. Aromatic residues absorb in the near-UV and can photoionise. Irradiation of tryptophan generates radical species and singlet oxygen, which then attack neighbouring residues that would otherwise be unreactive.⁴ Amber glass and opaque secondary packaging are not decoration.

Trace transition metals. Iron and copper at trace levels redox cycle and generate reactive oxygen species, producing site-specific oxidation near the metal binding site rather than uniform damage.⁴ Metal can enter from glass, stoppers, water or raw materials, which is why metal burden is specified at all. Chelators are the standard countermeasure and are genuinely double-edged: the Fe(II)–EDTA complex reacts rapidly with hydrogen peroxide to make hydroxyl radicals, so a chelator can accelerate the reaction it was added to suppress, and triethylenetetramine has been reported as more effective than EDTA against copper-mediated oxidation.⁴ Excipients can be the source too: polysorbate surfactants generate peroxide on storage.⁴

The copper-peptide complexes are the instructive special case, because there the metal is deliberate and structural rather than contaminating. In GHK-Cu and AHK-Cu the Cu(II) ion is coordinated by the peptide itself, and that coordination is the point of the molecule. It does not exempt them from redox chemistry — a bound copper is still a redox-active centre, and complex stability depends on pH and on what else in solution can compete for the metal — but it does mean reasoning about them by analogy to a metal-free peptide is a mistake in both directions.

Aggregation: the physical route

Aggregation is different in kind. No bond in the backbone breaks; the molecules simply associate with each other instead of with solvent. The result runs a continuum from soluble dimers, through subvisible particles, to visible haze and precipitate. Some aggregates are held together only by non-covalent hydrophobic contact; others acquire covalent crosslinks such as dityrosine or intermolecular disulfides once they have been in contact long enough.⁴

The standard mechanistic account starts with a partially unfolded intermediate. A fully folded molecule keeps its hydrophobic surface buried; the dangerous population is in between — enough structure to present a complementary sticky face, enough flexibility to find a partner. Aggregation propensity is therefore controlled by two things at once: conformational stability, how strongly the molecule resists unfolding, and colloidal stability, how strongly molecules repel each other in solution. Temperature, pH, ionic strength, cosolutes, preservatives and surfactants all act on one or both.⁶

Aggregation is also self-accelerating, which makes it qualitatively different from hydrolysis. Hydrolysis proceeds at a rate set by conditions and does not care how much has already happened. Aggregation is nucleated: forming the first stable multimer is the hard, slow step, and once nuclei exist they template further addition. A sample can sit apparently stable through a long lag phase and then turn over quickly, and existing particulates can seed the process.

The interface is the real culprit. Peptides and proteins are amphipathic, and the air-liquid interface is a place where a molecule can bury its hydrophobic surface in air instead of paying the entropic cost of keeping it in water. Adsorption is thermodynamically favourable and essentially immediate. Once adsorbed, the molecule spreads and partially unfolds, exposing regions that were internal; an interferometric study of monoclonal antibody films at the air-water interface observed layer thickness and amide signal reaching equilibrium over roughly three hours while surface pressure continued to rise, consistent with continued reorganisation, further unfolding and formation of intermolecular beta sheet within the film.⁷

This is precisely why shaking is worse than swirling, and the distinction is mechanistic rather than folkloric. Gentle swirling moves bulk liquid without creating much new surface. Shaking generates bubbles and repeatedly stretches, compresses and ruptures the interfacial film, and each rupture delivers a patch of denatured, aggregation-competent material into the bulk, where it can nucleate. The stress is not shear in the liquid so much as the repeated creation and destruction of interface. The same logic explains why surfactants protect: polysorbate 20 stabilised the antibody by occupying the interface and preventing adsorption, not by interacting with it in bulk.⁷ Something has to be at the interface; if it is not the surfactant, it will be the peptide. The corollary shows up in how to reconstitute peptides, where diluent is directed down the vial wall rather than jetted onto the cake.

Freeze-thaw damage

Freezing is not simply a slower version of cold. The transition itself is a stress, distinct from the storage state that follows.

As ice forms, it excludes solutes. Everything dissolved — peptide, buffer salts, excipients — is pushed ahead of the advancing ice front into a shrinking volume of unfrozen liquid. Solute redistribution and cryoconcentration of this kind are documented drivers of freeze-induced aggregation, because the peptide spends the transition at a far higher effective concentration and ionic strength than the nominal formulation specifies.⁸

Buffer components do not freeze out together. In a phosphate system the individual phosphate salts have different solubilities, so one crystallises preferentially and the pH of the remaining liquid drifts as freezing proceeds — an effect that correlates with aggregation and that is worst where buffer concentration is too low to hold the pH.⁸ A formulation nominally at neutral pH can pass through conditions well away from it during the seconds or minutes of the transition.

Alongside these, the ice-water interface is itself a surface that molecules adsorb to and unfold at, in the same way as an air-liquid interface, and cold denaturation weakens the hydrophobic interactions that hold a structure together, exposing buried residues including free thiols that can then take part in disulfide chemistry.⁸

All of this is why cycling is the problem rather than cold. Continuous cold storage means passing the transition once. Ten cycles means passing it ten times, and reported aggregation increases severalfold with additional cycles, with the duration of the thaw among the most influential variables.⁸ The practical consequence — aliquot once, cycle nothing — falls out of the chemistry, and is covered from the logistics side in cold-chain peptide storage.

Why lyophilization solves most of this

Look back at the list. Hydrolysis needs water as a reactant. Deamidation resolves its succinimide by hydrolysis. Oxidation needs dissolved oxygen and mobile metal ions. Aggregation needs molecules free to diffuse and interfaces to unfold against. Every one of those requirements is a property of the liquid state.

Freeze-drying removes the water by sublimation from the frozen state and leaves an amorphous, glassy solid. In that glass, molecular mobility is orders of magnitude lower than in solution; reactants cannot find each other, and the ones that require water do not have it. This is why lyophilized peptide is the standard shipping and storage form, and why the guidance for solid peptide is storage at −20 °C or −80 °C with the material kept dry rather than in solution.² The process itself is described in how peptides are made.

The cake is the visible record of whether that worked. A well-formed cake is a porous solid that fills the vial footprint, holds its shape and height, has a matte rather than glassy surface, and dissolves quickly and completely — the porosity is what gives the solvent access. Freeze-drying literature calls this an “elegant” cake and treats appearance as a quality attribute, while cautioning that not every deviation is meaningful and that acceptance criteria should be risk-based rather than cosmetic.⁹

A collapsed cake — shrunken, dense, glossy, pooled at the bottom of the vial — means the product temperature rose above the collapse temperature of the frozen matrix during drying, so the concentrated phase softened and flowed before the structure was fixed. The concern is not appearance. Collapse restricts vapour escape, so residual moisture tends to be higher; higher residual moisture plasticises the glass and restores exactly the molecular mobility that drying was meant to remove. Reconstitution is also typically slower, because a dense collapsed mass presents far less surface to the solvent. Discoloration, particularly yellowing, is a separate signal and can indicate chemical degradation such as oxidation or tryptophan modification.⁹

Reconstitution restarts the clock deliberately. Once diluent goes in, the whole liquid-phase menu is available again — see what is bacteriostatic water. Benzyl alcohol addresses microbial growth; it does nothing about hydrolysis or oxidation.

What degradation looks like in practice

Some degradation is visible. Aggregation past the subvisible range produces haze, opalescence, floating particulates, fibrous strands or a film on the glass. A cake that has slumped, shrunk away from the wall, discoloured or turned glassy is reporting on the drying process or on subsequent moisture ingress. Material that will not go into solution cleanly — leaving gel, flecks or a residue after prolonged gentle mixing — is the same signal, since a genuinely intact amorphous cake dissolves readily.

Most degradation is invisible. That is the honest and important half of this section. Hydrolysis produces shorter peptides that are usually as soluble as the parent. Deamidation changes charge and about one dalton of mass. Methionine oxidation adds a single oxygen atom. Isomerisation changes nothing detectable by eye at all. None of these cloud a solution, none of them change its colour, and none of them can be ruled out by looking at a vial.

Visual inspection is therefore a one-way test. Visible change is strong evidence that something is wrong; absence of visible change is very weak evidence that anything is right. A perfectly clear, colourless solution can be substantially degraded. Separating those cases requires analysis rather than observation — chromatographic separation to resolve species differing in hydrophobicity or charge, and mass spectrometry to identify them. What a purity figure does and does not cover is discussed in understanding peptide COA testing; note also that a COA characterises material at the time of testing, not material since stored, shipped, opened and reconstituted.

What the evidence does not establish

The mechanisms above are well characterised. Their quantitative application to any specific research peptide is much less so, and the gap is wider than most summaries admit.

Most of the primary data comes from proteins, not short peptides. The stability literature is dominated by therapeutic proteins and monoclonal antibodies, because that is where the analytical budgets are.^1,6 Those molecules have tertiary structure that shields buried residues and slows deamidation and oxidation. A short synthetic peptide is comparatively unstructured, so the mechanisms transfer but the rates generally do not, and usually not in the conservative direction.

Kinetics from model peptides are sequence-specific. The pH-dependent partition between cleavage and isomerisation was measured in one hexapeptide.³ Deamidation rate constants were measured in pentapeptide libraries.⁵ These are excellent for demonstrating mechanism and for ranking motifs. They are not a shelf-life calculator for an unrelated sequence, and predicting an absolute rate for a given peptide from sequence alone remains unreliable.

Published shelf-life figures for research peptides are largely not experimental. Numbers circulated as “stable for X days refrigerated” usually derive from analogy, from manufacturer convention, or from accelerated studies on a different molecule — not from stability-indicating assays on that specific product, in that specific diluent and container. Where such data has not been published, the honest position is that the figure is an estimate.

Accelerated stability testing extrapolates imperfectly. Raising temperature to compress a study assumes the same reaction dominates at both temperatures. That fails when pathways have different activation energies or when a phase change intervenes, which is common in amorphous solids near their glass transition.

Chemical change and functional change are not the same measurement. A purity or intact-mass result quantifies molecules, not activity. Whether a given oxidation or deamidation event alters behaviour in an assay depends on where in the sequence it occurred, and for most research peptides that has not been mapped residue by residue.

None of this is a reason to disregard the chemistry. It is a reason to treat storage conditions as risk management rather than as a formula with a guaranteed answer.

Frequently asked questions

Why do peptides degrade so much faster in solution than as a dry powder? Because water is a reactant in the two dominant chemical routes — backbone hydrolysis and the hydrolytic opening of the succinimide in deamidation — and it is also the medium that gives the molecule the mobility to unfold, diffuse and collide. Lyophilization removes both the reagent and the mobility at once.

Which residues are the usual weak points? Asparagine and glutamine deamidate, with Asn-Gly the classic fast motif and Asn-Ser, Asn-Thr and Asn-Asp also elevated. Aspartate isomerises and, next to proline or glycine, promotes chain cleavage. Methionine and cysteine oxidise readily; tryptophan, histidine and tyrosine are also susceptible, particularly under light or trace metals.

Does a clear solution mean the peptide is intact? No. Cloudiness and particulates indicate aggregation, but hydrolysis, deamidation and oxidation all produce soluble products that leave a solution looking unchanged. A substantially degraded sample can be perfectly clear. This asymmetry is the reason analytical testing exists.

Why is repeated freeze-thaw worse than continuous cold? Because the damage is concentrated in the transition, not the storage. Each cycle recreates the advancing ice front that cryoconcentrates solutes, the differential crystallisation of buffer salts that shifts local pH, and the fresh ice-water interface. Reported aggregation rises with cycle count.⁸

If oxidation needs trace metals, why do some peptides contain a metal deliberately? In copper complexes such as GHK-Cu and AHK-Cu the copper is a coordinated structural component rather than an adventitious contaminant, and the complex is the molecule of interest. Contaminating metals are uncontrolled, unbound and free to redox cycle against whatever is nearby; coordinated metal is defined and constrained. The two situations are chemically distinct even though the element is the same.

Is any of this approved for human use? No. All Patriot Labs products are sold strictly for in-vitro research and laboratory use only. They are not approved for human or veterinary consumption, and nothing on this page describes use in a person.

All Patriot Labs products are sold strictly for in-vitro research and laboratory use only. Not for human or veterinary consumption. This guide is educational and describes peptide chemistry and published research in general terms; it is not medical advice, does not describe how to use any product, and the references cited do not constitute a product claim.