Disulfide scrambling: why two peptides with identical mass can be different molecules

Most degradation pathways in peptide chemistry change something measurable. Oxidation adds mass. Hydrolysis produces fragments. Deamidation shifts charge and, marginally, mass. Disulfide scrambling changes none of these. A peptide with four cysteines that has rearranged from its intended pairing to an alternative pairing has the same sequence, the same molecular formula, the same nominal and exact mass, and frequently a similar enough hydrophobic surface that it elutes close to the correctly folded form. It is a different molecule with the same analytical signature under most routine testing.

This makes disulfide connectivity a distinctive problem in peptide characterization. It is not a purity question in the ordinary sense — the scrambled species is not an impurity introduced during synthesis so much as a structural isomer of the target. And it is not an identity question in the way mass spectrometry usually resolves identity, because mass confirmation returns the expected value for every isomer in the set. For cysteine-containing sequences, connectivity has to be established by methods chosen specifically for it, or it is simply assumed.

The combinatorics of pairing

A peptide with two cysteines has one possible intramolecular disulfide. A peptide with four has three possible pairings, only one of which is typically the intended arrangement. Six cysteines give fifteen. Eight give one hundred and five. The number of distinct pairings for 2n cysteines grows as the double factorial (2n−1)!!, and it climbs quickly enough that for the larger disulfide-rich sequences the intended structure is one option among many dozens.

Intermolecular pairing expands the space further. Two molecules can link through a cysteine from each, producing covalent dimers that carry roughly double the mass and are, in principle, detectable by mass measurement — though only if the analysis is looking for them and the chromatographic method resolves them. Higher oligomers follow the same logic. In practice, the intramolecular isomers are the harder analytical problem precisely because they are mass-silent.

Whether the correct isomer dominates is a thermodynamic and kinetic question. For sequences whose native fold is a deep energy minimum, oxidative folding under appropriate conditions converges on the intended connectivity with reasonable efficiency, because the correctly folded state is both accessible and stable. For sequences where the alternatives are close in energy, or where the folding pathway has kinetic traps, the product is a mixture, and the composition of that mixture depends on conditions rather than on sequence alone.

Thiol–disulfide exchange is reversible by design

The reaction that forms a disulfide is the same reaction that scrambles it. A thiolate anion attacks a disulfide bond, displacing one of the two sulfurs as a new thiolate and forming a new disulfide with the other. The process is a nucleophilic substitution at sulfur, and it is freely reversible. Nothing about a formed disulfide is chemically committed; it is a bond that can be exchanged whenever a nucleophilic thiolate is present.

Two consequences follow. First, scrambling requires a free thiol, or something that can generate one. A fully oxidized peptide with no unpaired cysteine and no reducing species in the environment is relatively resistant to rearrangement. A preparation containing even a small fraction of reduced or partially reduced material carries its own catalyst: those free thiols can initiate exchange throughout the population, so a minor reduced impurity has an effect disproportionate to its abundance.

Second, because the attacking species is the thiolate rather than the neutral thiol, the rate is strongly pH-dependent. Cysteine thiol groups have pKa values typically in the range of roughly 8 to 9, though local electrostatic environment shifts this considerably in structured molecules. Below that range the equilibrium favors the protonated, unreactive form. Above it, thiolate concentration rises and exchange accelerates. This is the basis for the common observation that mildly acidic conditions suppress scrambling while neutral-to-alkaline conditions promote it — the same directional dependence seen in deamidation, but for entirely different mechanistic reasons.

Trace metals compound the problem from a different direction. Copper and iron in solution catalyze the oxidation of free thiols to disulfides, which sounds helpful but is not: uncontrolled metal-catalyzed oxidation forms whatever pairing is kinetically accessible rather than whatever pairing is correct, and it can also drive intermolecular linkage. This is one of several places where trace metal content, usually treated as a separate analytical concern, connects directly to structural outcome.

What routine testing does and does not resolve

Reversed-phase HPLC can separate disulfide isomers, and often does. Different connectivity produces different three-dimensional structure, different solvent-exposed surface, and therefore different retention. But the separation is not guaranteed, and it is not predictable in advance. Isomers that happen to present similar hydrophobic faces co-elute. A purity chromatogram showing a single sharp peak is consistent with a homogeneous correctly folded product and equally consistent with two isomers that the method does not resolve.

Mass spectrometry in its routine identity-confirmation role is essentially blind here. Intramolecular isomers are isobaric; the measured mass matches the target regardless of pairing. What mass spectrometry does resolve is oxidation state — a fully oxidized species differs from the fully reduced form by two daltons per disulfide, so the presence of reduced or partially reduced material is detectable when the analysis is examined for it. That is a useful check, and it addresses the population most likely to seed scrambling, but it does not speak to connectivity among oxidized species.

Establishing connectivity requires a method built for the purpose. The standard approach is proteolytic digestion under conditions chosen to suppress exchange — typically low pH, where thiolate concentration is minimal — followed by mass analysis of the resulting fragments. If the digestion cleaves between the cysteines, disulfide-linked peptides appear as fragment pairs whose combined mass identifies which residues are joined. Partial reduction with controlled alkylation, differential labeling strategies, and NMR-based approaches for smaller molecules serve related roles. The common feature is that all of them are deliberate, method-specific, and absent from a routine CoA unless someone asked for them.

The practical reading of a certificate of analysis follows from this. For a peptide with no cysteine, none of this applies. For a peptide with two cysteines, a single disulfide is the only intramolecular option, and the analytical question reduces to whether the material is oxidized, reduced, or mixed — which mass data can answer. For a peptide with four or more cysteines, purity and mass together do not establish structure, and a CoA that reports only those two parameters is silent on the property most likely to distinguish correctly folded material from a plausible-looking alternative.

Conditions that preserve connectivity

The handling implications track the mechanism rather than adding anything separate. Exchange requires thiolate; thiolate concentration rises with pH; therefore conditions in the mildly acidic range suppress rearrangement relative to neutral or alkaline conditions. Reducing agents obviously accelerate scrambling and should not share an environment with material whose connectivity matters. Trace metals promote uncontrolled oxidation, so chelation or attention to water and buffer quality is more consequential for disulfide-containing sequences than for others.

The solid state is markedly more protective than solution, for the same reason it is protective against most degradation chemistry: molecular mobility is restricted and the reactants cannot readily find one another. Lyophilized disulfide-rich material is comparatively stable; the same material in aqueous solution at neutral pH and ambient temperature is in the regime where exchange proceeds. Freeze-thaw cycling deserves specific mention here, since the transient concentration and pH shifts that accompany freezing create conditions locally that the bulk solution does not exhibit.

Oxygen exposure sits on the other side of the same balance. Air oxidation of free thiols is slow but not negligible, and in a partially reduced preparation it converts free thiols into disulfides without regard to which pairing was intended. Headspace management is therefore not only a general good practice but a structurally specific one for these sequences.

What connectivity means for characterization

Disulfide scrambling occupies an unusual position among peptide degradation routes. It does not reduce purity as chromatography defines purity, does not change mass as identity testing measures mass, and does not produce fragments or high-molecular-weight species that routine assays are designed to catch. It produces a molecule that answers every standard question correctly while differing in the arrangement that determines its three-dimensional structure.

For research work where structural fidelity matters — which is most work involving disulfide-rich sequences, since the disulfide framework is generally what holds the fold together — this argues for treating connectivity as a distinct characterization parameter rather than an implication of purity and mass. It also argues for reading a CoA with attention to what is absent. A high purity figure and a confirmed mass are real information about a cysteine-containing peptide. They are not information about how its cysteines are paired.