Buffer species and general acid-base catalysis: why the same pH in two buffers is not the same condition for a peptide

Stability data for peptides in solution are usually reported against pH, and pH is treated as the principal formulation variable. It is an important one, but it describes only the activity of hydrogen ions. The species used to hold that pH are not chemically inert spectators. Many of the degradation reactions that limit peptide solution stability are susceptible to general acid-base catalysis, in which the buffer components themselves donate or accept protons in the rate-determining step. Two solutions at an identical measured pH, prepared with different buffers, can therefore degrade at measurably different rates.

This post reviews the distinction between specific and general catalysis, the reactions in which buffer species have been characterized as influential, the practical differences among commonly used buffers, and the reasons why buffer selection data do not transfer cleanly from one peptide to another.

Specific versus general catalysis

In specific acid or base catalysis, the observed rate depends only on the concentration of hydronium or hydroxide ions. Changing the buffer species at fixed pH has no effect, and the rate does not change with total buffer concentration. In general acid-base catalysis, the buffer acid and its conjugate base each contribute a rate term, so the observed first-order rate constant takes a form such as k_obs = k_0 + k_H[H+] + k_OH[OH-] + k_HA[HA] + k_A[A-]. The last two terms are the buffer contribution.

The practical diagnostic is straightforward. If the apparent first-order rate constant is measured at constant pH and constant ionic strength while total buffer concentration is varied, a plot of k_obs against buffer concentration gives a straight line. The slope reflects catalytic activity of the buffer, and the intercept, extrapolated to zero buffer concentration, represents the contribution of the solvent, hydronium, and hydroxide alone. A flat line indicates that buffer catalysis is negligible for that reaction under those conditions. This experiment is inexpensive relative to a full stability program and is a standard element of preformulation work.

Because the buffer terms scale with concentration, the effect is not fixed. A 5 mM buffer may contribute little, while a 100 mM buffer of the same species may dominate the rate. Reports that compare buffers without stating their concentrations and ionic strength are difficult to interpret for this reason.

Reactions in which buffer species have been implicated

Several well-characterized peptide degradation pathways involve proton transfer steps that buffers can mediate. Asparagine deamidation proceeds through a cyclic succinimide intermediate, and studies of model peptides have reported that the rate of succinimide formation at neutral pH is increased by phosphate and other polyprotic anions, consistent with bifunctional catalysis. Phosphate has an acid and a base site positioned to assist simultaneous proton removal and donation, which is a feature that simple monocarboxylate buffers lack. Aspartate isomerization through a similar cyclic imide has likewise been characterized as sensitive to buffer composition.

Hydrolysis of labile peptide bonds, including the acid-sensitive Asp-Pro linkage, can show buffer dependence, particularly in the mildly acidic range where citrate and acetate are commonly used. Diketopiperazine formation, in which an N-terminal dipeptide cyclizes and is excised, depends on the free amine being unprotonated, so it is influenced by the buffer’s effect on local acidity as well as by any direct catalysis. Oxidation is a somewhat different case: the buffer may not catalyze the chemistry directly, but it can modulate it through metal binding, as discussed below.

It should be noted that buffer influence is far from uniform across these reactions. Some are strongly catalyzed, some weakly, and some show no detectable dependence. Generalizing from one pathway to another is a common error.

Characteristics of commonly used buffers

Each frequently used buffer carries a set of properties beyond its pKa.

Phosphate. Its second pKa near 7.2 makes it a natural choice at neutral pH, and it is well tolerated in many analytical settings. Its polyprotic character makes it a documented catalyst in cyclic imide chemistry, as noted above. Phosphate also has a pronounced pH shift on freezing: when a sodium phosphate solution is cooled, the disodium salt can crystallize preferentially, and the pH of the freeze-concentrated fraction has been reported to fall by several units. This is relevant to any peptide solution that will be frozen, and it complements the cryoconcentration effects described elsewhere for freeze-thaw cycles. Phosphate can also form insoluble salts with some divalent cations.

Acetate. With a pKa near 4.76, acetate suits mildly acidic solutions. It is a monocarboxylate and so lacks the bifunctional geometry of phosphate or citrate. It is volatile, which makes it compatible with mass spectrometry and lyophilization, but this also means it can be partly lost during freeze-drying, shifting the pH of the reconstituted solution relative to the pre-lyophilization one. Counterion exchange to acetate is also a common step in peptide manufacture, so acetate may already be present in the material as a salt form.

Citrate. Citrate is triprotic, with pKa values near 3.1, 4.8, and 6.4, covering a broad acidic range. It is a multidentate chelator of metal ions, which can be a benefit where trace metals promote oxidation of methionine, histidine, or cysteine residues, and a liability where a metal is part of the system under study. Citrate has also been characterized as a catalyst in several reactions, and its carboxylate groups can interact with basic peptides. Citrate-containing solutions are known to be somewhat more prone to microbial growth than phosphate-containing ones if not otherwise protected, although this is a handling observation rather than a stability one.

Histidine, Tris, and others. Histidine, with an imidazole pKa near 6.0, is widely used in protein formulations and has some buffering capacity near neutral pH. Tris has a strongly temperature-dependent pKa, with a change of roughly 0.03 pH units per degree Celsius, so a solution prepared at room temperature will read differently at refrigerator temperature. Tris also bears a primary amine that can react with certain electrophiles. Each of these brings its own interactions.

Why buffer data transfer poorly between peptides

The central difficulty is that buffer catalysis is a property of the peptide-buffer pair, not of the buffer alone. Sequence determines which reactions are possible, and local environment around a susceptible residue determines how accessible it is to a catalytic anion. A peptide with an Asn-Gly sequence may show a marked phosphate effect, while one lacking any labile site may show none. Net charge is an additional factor: an anionic buffer can concentrate near a positively charged peptide through electrostatic attraction, raising the effective local catalyst concentration above what the bulk value would suggest. Ionic strength, which changes with buffer choice and concentration, can influence rates through this same mechanism.

Interactions between buffers and other formulation components add further complexity. Buffers can alter solubility and aggregation behavior, they can change the response of a stability-indicating chromatographic method if they absorb at the detection wavelength or alter retention, and they can contribute to osmolality, which matters for any solution intended to be compared across studies. Residual counterions from synthesis, particularly trifluoroacetate, contribute to the effective ionic composition of a reconstituted solution and can partially offset the apparent role of the chosen buffer when the buffer concentration is low.

For laboratories characterizing a peptide in solution, a useful practice is to report buffer species, concentration, ionic strength, measured pH at the temperature of the study, and the headspace and container conditions. Without these, a stability result describes an incompletely defined system. A pH-rate profile obtained in a single buffer system also carries the buffer’s own contribution, and it is good practice to extrapolate to zero buffer concentration, or to confirm the profile in a second buffer, before attributing features of the profile to the peptide alone.

Summary

Holding pH constant does not hold the chemical environment constant. Buffer acids and bases participate in proton transfer reactions that underlie several peptide degradation pathways, with polyprotic species such as phosphate and citrate showing the most pronounced and best-documented effects in cyclic imide and related chemistry. Freeze-induced pH shifts, temperature-dependent pKa values, volatility, and metal chelation add further buffer-specific behavior. Because the magnitude of each effect depends on sequence, charge, concentration, and ionic strength, buffer comparisons are most informative when each variable is explicitly stated and when buffer concentration dependence has been examined directly rather than assumed.