pH and Buffer Selection: Competing Dependencies, Buffer Properties and a Recording Procedure

There is no universal answer, and the honest one is sequence dependent: a mildly acidic buffer is the compromise that suits the largest number of sequences, but it is a default to be checked rather than a specification to be assumed. pH is the single most influential variable a laboratory controls in solution — it governs more degradation routes than any other controllable condition, and acts on them in opposing directions 12. Choosing a pH means deciding which loss route matters most here; choosing a buffer means accepting a second set of liabilities belonging to the buffer itself.

Why no single value is correct

The chemistry of each route is set out in the degradation pathways entry. What matters here is narrower: the direction each takes along the pH axis, and the fact that those directions conflict.

Deamidation of asparagine is base-catalysed — its rate-determining step involves deprotonation of a backbone nitrogen, so it accelerates above neutrality and slows as pH falls 3. Acid-catalysed cleavage at aspartyl residues does the reverse: the protonated side-chain carboxyl is the attacking species, so it becomes dominant below roughly pH 4 2. Disulfide scrambling proceeds through the thiolate and requires near-neutral or alkaline conditions. Solubility passes through a minimum at the isoelectric point, where net charge is zero and repulsion least — the condition that most favours association 4.

Four constraints on one variable. Three are covalent and point in different directions; the fourth is physical and can land anywhere — including inside the window the others recommend.

The mildly acidic window and its limits

For a large fraction of sequences a mildly acidic buffer is the best available compromise 12. It sits below the region where deamidation and scrambling accelerate, and above the region where aspartyl cleavage becomes dominant. It minimises no single route; it avoids being badly wrong about several at once.

The qualification matters more than the recommendation. The window is defensible for a sequence about which nothing specific is known, and stops being defensible the moment something specific is. A sequence with no asparagine, glutamine or free cysteine has little to gain from acid and a susceptible aspartyl bond to lose by it. A sequence whose pI falls inside the window has its worst solubility where its covalent stability is best, and must be moved away from the pI even at a covalent cost 4. A labile asparagine motif alongside a susceptible aspartyl bond leaves no good pH at all — only a documented decision about which loss to accept 1.

Estimating the isoelectric point

The pI is the pH at which the molecule carries no net charge. It enters twice: solubility is at its minimum there, the tendency to associate at its maximum 4. The objective is to hold the solution some distance away, on whichever side the covalent constraints prefer.

1. List every ionisable group: the N-terminal α-amino group, the C-terminal α-carboxyl group, and the side chains of aspartate, glutamate, cysteine, tyrosine, histidine, lysine and arginine.
2. Account for terminal modifications. An acetylated N-terminus and an amidated C-terminus each remove an ionisable group; a calculation run on the unmodified sequence is wrong.
3. Assign each group a standard tabulated pKa from a single published source. Do not mix tables — the sets differ and a mixed set is not reproducible.
4. Express net charge against pH — each basic group positive below its pKa, each acidic group negative above its own — and solve for the pH at which the sum is zero.
5. Sanity-check against composition. Lysine- and arginine-rich sequences return a basic pI, aspartate- and glutamate-rich ones an acidic pI; a contradiction is an arithmetic error.
6. Record the pKa set alongside the value. Different sets return different figures for one sequence.

A pI obtained this way is an estimate. It treats every ionisable group as though it sat in bulk solvent with a textbook pKa, when local environment — neighbouring charges, hydrogen bonding, partial burial, conformation — shifts individual values in both directions. Short unstructured peptides estimate more reliably than folded ones, but the figure is an approximation and observation overrides it. If material comes out of solution at a pH the calculation calls safe, the calculation is wrong.

Buffer systems and their specific liabilities

A buffer holds pH against acid or base acquired from dissolved carbon dioxide, container leaching and the peptide's own counter-ion, and does so only within about one unit either side of its pKa; outside that range it is a salt. Selection means matching the range to the target pH, then checking the buffer's own liabilities.

Phosphate carries two liabilities, easy to overlook because it is the default in so much laboratory work. It acts as a general base catalyst for deamidation, so the buffer holding a convenient near-physiological pH also accelerates the route that pH already favours, and the effect scales with concentration 2. And sodium phosphate shifts pH sharply on freezing: one salt crystallises before the other, so freeze-concentrated material ends up well away from its ambient pH. Do not hold material intended for frozen storage in sodium phosphate without a reason.

Tris has an unusually large temperature coefficient. A Tris buffer titrated on the bench is not at that pH at refrigerator or freezer temperature — it moves appreciably more alkaline as it cools — so a solution adjusted at ambient and then chilled has drifted toward the region where deamidation and scrambling accelerate. Titrate at the temperature of use, or record pH with the temperature of measurement.

Citrate chelates metal ions, and whether that is an advantage or an interference depends on what happens downstream. As an advantage it sequesters iron and copper leached from glassware, steel and ordinary-grade salts, suppressing metal-catalysed oxidation. As an interference it strips the metal from any assay requiring one, and at an unfavourable ratio it can turn pro-oxidant.

Ammonium bicarbonate is volatile, which is exactly why it is convenient immediately before lyophilisation: it leaves little residue, so the dried solid carries no salt load it never needed. The same volatility makes it a poor holding buffer — it loses carbon dioxide and drifts, and its range sits where base-catalysed routes run fastest.

Ionic strength

Ionic strength acts on both solubility and aggregation, and not in a single direction. At low ionic strength added salt screens electrostatic attraction between oppositely charged surfaces and raises solubility. Beyond a point, further additions compete for water of hydration, expose hydrophobic surface and drive the molecule out of solution. The crossover is specific to sequence and salt, and interacts with pH because the charge state being screened is itself pH-dependent 4.

Two practical points. Keep buffer concentration only as high as the job requires — the phosphate case shows a buffer component can be a reactant as well as a pH control, and that contribution scales with concentration 2. And record total ionic composition, not the buffer alone: added sodium chloride, the purification counter-ion and any salt in a stock diluent all contribute.

When the assay dictates the buffer

In practice the buffer is frequently not chosen on stability grounds at all. A binding assay, a chromatographic method, a mass-spectrometric workflow or a cell-based readout each imposes compatibility constraints, and those usually win — a stability-optimal buffer that suppresses the measurement is of no use. Phosphate-buffered saline near neutrality is used constantly for this reason, despite being, on stability grounds, a poor choice for a deamidation-prone sequence 12.

That is a legitimate outcome. The failure is not choosing the assay-compatible buffer; it is pretending the choice carried no cost. Bound the exposure and write it down: hold assay-buffer solutions for the shortest interval the work permits, keep the long-term stock in a better condition, and prepare the assay solution from that stock immediately before use. Where two datasets later disagree, a recorded buffer condition separates a traceable explanation from an unexplained one.

Selecting and recording buffer conditions

1. List the liabilities in the sequence: asparagine or glutamine and the residue following each, aspartyl motifs, free or paired cysteine, methionine and tryptophan, and both termini.
2. Estimate the pI by the procedure above and record the pKa set used.
3. Decide the direction of the compromise from those liabilities, not from habit. Deamidation and scrambling argue downward; a susceptible aspartyl bond argues against going far down.
4. Set a target pH inside the chosen region and at a stated distance from the estimated pI, then select a buffer whose range brackets that target near its middle rather than at the edge where capacity has already failed.
5. Check that buffer against its own liabilities: phosphate against deamidation and freezing, Tris against the storage temperature, citrate against downstream metal dependencies, volatile salts against holding.
6. Keep buffer concentration to the minimum that holds pH over the intended interval.
7. Measure the pH of the final solution — after peptide and counter-ion are dissolved — at the temperature of storage. Do not infer it from the recipe.
8. Record as one entry: buffer, concentration, measured pH, measurement temperature, added salt, purification counter-ion, container type and date.
9. Where a downstream method dictated the buffer, record that with the cost accepted and the holding interval permitted.
10. Re-measure pH on any solution held for an extended interval before drawing conclusions from it. Drift is silent.
11. On unexplained loss, retrieve the record first. Buffer, pH and ionic strength are the conditions most often unrecorded.

None of this yields a number transferable to the next sequence. It yields a documented condition for this one with the reasoning attached — the only form in which a buffer choice can be defended, revisited or corrected when a result fails to reproduce 1.