Freeze-thaw cycles and peptide integrity: what happens at the molecular level

Freeze-thaw cycling is one of the most commonly cited handling variables in peptide stability literature, and one of the least precisely understood in practice. The standard guidance — aliquot your stock, avoid repeated freeze-thaw — is repeated across supplier documentation and protocol notes without much explanation of what the cycling actually does to the molecule. The mechanisms are worth understanding, because they determine which peptides are sensitive to cycling and which are relatively robust, and because they explain why the damage is not linear in the number of cycles.

This post covers what happens physically and chemically during a freeze-thaw cycle, which degradation pathways it accelerates, and how aliquoting practice maps onto those mechanisms.

What freezing actually does to a solution

The intuitive model of freezing — the solution gets cold and becomes a solid — is misleading for stability purposes. What happens in a dilute aqueous peptide solution is closer to a fractional crystallization.

As the temperature drops below the freezing point, pure water crystallizes out first. Solutes are excluded from the growing ice lattice and become concentrated in the shrinking volume of liquid that remains between ice crystals. This residual liquid phase is called the freeze-concentrate, and the process is cryoconcentration. By the time the sample is fully solidified, the peptide, buffer salts, and any excipients have passed through a phase in which their local concentration was many times higher than the nominal concentration of the starting solution — in some systems by one to two orders of magnitude.

Two consequences follow. First, concentration-dependent degradation pathways are accelerated during the transition, not during storage at the final temperature. Aggregation is bimolecular or higher order, so a transient tenfold concentration increase disproportionately increases aggregation rate. Second, buffer components crystallize at different rates and different temperatures, which means the pH of the freeze-concentrate can drift substantially from the pH of the starting solution.

The phosphate buffer problem

The pH shift during freezing is well characterized in sodium phosphate systems and is the standard illustration of the effect. Disodium hydrogen phosphate is less soluble than sodium dihydrogen phosphate and crystallizes out preferentially as the solution freezes. This selectively removes the basic component of the buffer pair from the liquid phase, and the freeze-concentrate becomes acidic. Studies have measured drops of two to three pH units in sodium phosphate systems during freezing.

For a peptide with pH-sensitive residues, this matters. Aspartic acid residues undergo acid-catalyzed isomerization to isoaspartate via a succinimide intermediate, and the rate is strongly pH-dependent. Asparagine and glutamine deamidation has a pH-dependent profile as well, with distinct acid- and base-catalyzed regimes. A peptide that is stable at pH 7.4 in solution may be transiting through pH 4-5 during every freeze, for a period determined by how quickly the sample passes through the phase transition.

Potassium phosphate systems shift in the opposite direction, and citrate and histidine systems shift less. This is one of the reasons formulation buffer selection is not interchangeable across peptides, and one of the reasons a certificate of analysis that specifies the diluent is more informative than one that does not.

Interfacial adsorption at the ice front

The second mechanism is mechanical rather than chemical. Freezing creates an enormous amount of new ice-water interface, and peptides — like proteins generally — are surface-active. They adsorb to interfaces, and adsorption at an interface tends to partially unfold the molecule, exposing hydrophobic residues that were internally oriented in solution.

For small linear peptides with little tertiary structure, this is a minor effect. For larger peptides and those with defined secondary structure, partial unfolding at the ice interface is a nucleation event for aggregation: the exposed hydrophobic surfaces of two adsorbed molecules can associate, and the resulting dimer is a template for further association. The aggregate does not necessarily redissolve on thawing.

This is why the damage from freeze-thaw cycling is not simply proportional to time spent frozen. A sample held at −80 °C for a year experiences one freezing event; a sample cycled weekly for three months experiences twelve. The second sample has been through twelve rounds of interface generation and cryoconcentration despite spending less total time frozen.

Why lyophilized material behaves differently

Lyophilized peptide is in a fundamentally different physical state, and the freeze-thaw framing does not apply to it in the same way. In a properly lyophilized cake, water content is typically reduced to low single-digit percentages, and the peptide is immobilized in an amorphous solid matrix below its glass transition temperature. Molecular mobility is minimal, and the bimolecular collisions required for aggregation are largely prevented by the physical constraint of the matrix.

The practical implication is that the sensitive window is the reconstituted state, not the lyophilized state. Removing a lyophilized vial from a freezer and returning it does not cycle the peptide through cryoconcentration, because there is no bulk aqueous phase to concentrate. What it does risk is condensation: a cold vial brought into ambient humidity will collect water on and potentially in the cake if the seal is compromised or if the vial is opened before equilibrating to room temperature. Moisture uptake lowers the glass transition temperature of the matrix, and if it drops toward storage temperature, mobility increases and the protective effect of the amorphous solid is lost.

This is the reasoning behind the standard instruction to allow a lyophilized vial to reach room temperature before opening it. The concern is not thermal shock to the peptide but water condensing onto the cake.

What aliquoting is actually solving for

Aliquoting reconstituted stock into single-use volumes is the standard mitigation, and the mechanisms above explain what it is and is not doing.

It eliminates repeat cycling of the bulk stock, which is the main benefit — each aliquot experiences one freeze and one thaw rather than one per experiment. It does not eliminate the first cycle, which every aliquot still undergoes, so a peptide that is genuinely freeze-labile is not rescued by aliquoting alone; for those, refrigerated storage of a working stock with a short assigned shelf life is sometimes preferred over freezing at all.

Aliquot volume involves a trade-off that is often overlooked. Smaller aliquots mean less waste and fewer cycles, but they also mean a higher surface-area-to-volume ratio, which increases the proportion of peptide exposed to container-wall adsorption — a separate loss mechanism that matters most at low concentrations. For dilute solutions, this is a real effect, and it is part of why low-binding polypropylene is specified over standard tubes and why carrier proteins or surfactants appear in some formulation protocols.

Freezing rate also matters, and in a direction that is not always intuitive. Faster freezing produces smaller ice crystals and more total interfacial area, which is worse for interface-driven aggregation, but it shortens the time spent in the cryoconcentrated intermediate state, which is better for concentration-driven and pH-driven pathways. Which effect dominates is peptide-specific, and this is one of the variables that formulation development work is designed to resolve empirically rather than predict.

Synthesis

The shorthand instruction to avoid repeated freeze-thaw compresses three distinct mechanisms: cryoconcentration accelerating concentration-dependent degradation, buffer-component crystallization shifting pH away from the nominal formulation value, and ice-interface adsorption nucleating aggregation. They have different sensitivities across different peptides, which is why blanket stability guidance transfers poorly between compounds and why supplier stability data is most useful when it specifies the conditions tested rather than reporting a single shelf-life figure.

The practical corollary is that the state of the material determines which precautions apply. Lyophilized material stored dry and sealed is largely insulated from these pathways and is limited instead by moisture ingress and matrix mobility. Reconstituted material in aqueous buffer is where cycling exposure accumulates, and where aliquoting practice, buffer selection, and container chemistry do their work. Stability figures quoted for one state should not be assumed to describe the other.