How do you elute from Protein A without acid?

There is no single replacement for acid. Protein A binds the Fc at the CH2–CH3 interface through a largely hydrophobic contact with a few histidines in it, and low pH works because protonating those imidazole side chains is the cheapest way to break it. Every alternative below breaks the same contact some other way — by weakening the ligand so it lets go earlier, by competing for the contact surface, by changing what the surrounding solvent does, or by removing the dependence on chemistry altogether. Each one buys you something and costs you something else, and only three of the nine genuinely remove the acid.

In practice the answer is usually one of the first three: a mild-elution ligand releasing at pH 4.3–5.0, an amino-acid buffer system that carries the pool back towards neutral, or simply less time in the acid — which a convective format such as a MonoCore™ Protein A capsule, running at 5–10 monolith volumes per minute, shortens from minutes to seconds. Light-controlled release at neutral pH, where the ligand is switched by visible light instead of buffer chemistry (Solaris® Protein A, in Beta), is the last route here rather than the first.

The nine routes at a glance

Read as three groups. Three routes remove the acid: engineered calcium-dependent ligands, detergent elution, and light-controlled release. Five reduce it: mild-elution ligands, amino-acid buffers, arginine, glycols, and mixed-mode capture. One changes what binds in the first place: a different, subdomain-specific ligand. There is also an option that is not an elution at all — shortening how long the acid lasts — and for a large share of real problems it is the one that pays first, so it has its own section near the end.

Which failure mode do you actually have?

Before choosing a route, find out what actually fails. Two different problems hide behind "my antibody does not like the elution", and they respond to different things.

Aggregation at low pH is a cooperative unfolding process with a steep pH threshold and a kinetics measured in minutes. It responds strongly to a milder pH and strongly to less time. Acid-catalysed cleavage of a linker or conjugate is different: it follows roughly first-order kinetics in acid concentration and time — an acylhydrazone linker has a half-life of days at pH 7 and hours at pH 5 — so halving the contact time halves the cleavage and rarely more. Shortening helps proportionally; it does not rescue a chemistry that the process will expose to pH 3.6 again during virus inactivation.

There is a third case that is not about the elution at all. If the aggregate appears after the low-pH hold rather than in the eluate, the capture step is not your problem — see why low-pH elution fails on sensitive modalities. In a typical Fc process the elution takes minutes and the hold runs 30 to 60 minutes at pH 3.6 or below, so the hold usually dominates the total acid exposure — unless the eluate sits waiting before it is neutralised, which is common enough to be worth checking. Measuring aggregate at both points costs one analytical run and tells you which step to spend money on.

Can a mild-elution ligand replace acid elution?

Engineered Protein A variants that release the antibody around pH 4.3–5.0 rather than 3.0–3.5 are available from several suppliers and are the least disruptive option there is: same column format, same protocol shape, a different ligand. The mechanism is the mirror image of ordinary acid elution — additional histidines are built into the ligand itself, so the contact breaks at a pH the molecule still tolerates. Published evaluations across IgG subclasses and Fc-fusion proteins generally find comparable recovery with measurably less aggregate, and some of these ligands additionally remove a substantial share of the aggregate present in the load during the elution itself.

The limit is in the name: this is milder acid, not the absence of acid. A molecule whose weak point relaxes at pH 5 gains nothing. And the step after has to be checked — a virus-inactivation hold is typically validated at pH 3.6 or below, so a pool eluting at pH 4.8 has to be acidified separately. The acid does not disappear, it moves.

Can a buffer bring the pool back to neutral?

A group of published approaches replaces plain glycine-HCl with amino-acid systems — leucine, glycine and serine perform best in the published screens — whose buffering behaviour carries the collected pool up towards neutrality as it elutes, in some reports to around pH 7. The antibody still passes through an acidic elution front, but it does not then sit in acid while the rest of the peak collects, which is precisely the time-dependent part of the damage.

You pay in elution strength and peak shape: reported elution volumes are on the order of 2.4 column volumes with yields around 87–88 %, so the pool is more dilute than a hard acid elution would give. Additives have to be cleared and to be compatible with your assays. The attraction is that it needs no new hardware and no new ligand — it is a buffer development exercise, which is the cheapest kind there is.

How much does arginine raise the elution pH?

Arginine sits at 0.5–2 M in a mildly acidic eluate. It works by interacting with the aromatic and hydrophobic patches on both partners, masking the contact rather than titrating it, and it suppresses aggregation at the same time. That combination lets you raise the elution pH by roughly half a unit and still recover the product — but only that far: published data show a clear gain at pH 4.3, and at pH 5.0 even 2 M arginine leaves most of the antibody on the column. Arginine is a way to make an acid elution kinder, not a way to reach neutrality.

At process scale the arithmetic changes. Molar arginine is expensive, the solutions are viscous, and it has to be cleared and measured. It is a good answer for a clinical-scale process with a genuinely fragile product, and a poor one where cost of goods matters more than a point of monomer.

Can glycols shift the elution pH upwards?

Propylene glycol, ethylene glycol and similar polarity modifiers weaken a largely hydrophobic contact by changing what the solvent does around it, and they are usually combined with moderate salt rather than used alone. This is the documented route by which several subdomain-specific ligands elute a full unit higher than their nominal range. It rarely does the job by itself, which is why it is often invisible in method descriptions — but it is a genuine lever, and it stacks with mild-elution ligands, with arginine and with subdomain-specific ligands.

The costs are viscosity, pumping behaviour and one more component to clear and to keep out of your analytics. For a molecule that is close to tolerable at pH 4.5 and not quite there, it is frequently the cheapest way to find the missing half unit.

Can salt or chelation elute Protein A at neutral pH?

Two related ideas, and they are not equally practical. Eluting a conventional Protein A ligand with salt alone needs chaotropic conditions — on the order of 3–4 M magnesium chloride or 2 M thiocyanate. Moderate sodium chloride does not elute; at 0.5–1 M it is a wash buffer, and if anything it strengthens the hydrophobic part of the contact. Any claim of gentle salt elution at low ionic strength belongs to an engineered ligand, not to standard Protein A.

The engineered version is the interesting one: ligands whose Fc binding depends on a bound calcium ion, so the antibody is released by chelation at or near neutral pH, with affinity-grade selectivity and no acid at all. The published caveat matters — IgG2 and IgG4 elute at neutral pH, but IgG1, the most common therapeutic subclass, generally still needs around pH 5.5–6 or a substantial chelator concentration. Where it does work it works well: in one reported case a mild elution removed aggregation that amounted to about a third of the eluted antibody under a standard pH 3 elution. The downstream cost is real either way — a high-salt or chelator-containing pool is a poor load for cation exchange, which typically wants low conductivity, and the chelator is a clearance and analytics question rather than a conductivity one.

Can a detergent elute an antibody near neutral pH?

Alkyl amine-oxide detergents release bound antibody at pH 5.5–6.5, and recent work on a convective Protein A membrane reports recovery rising from under a third at pH 5.5 to near-quantitative with detergent present. Mechanistically the detergent competes for the hydrophobic contact rather than titrating the histidines, which is why the effect depends on chain length and on the specific ligand.

The obstacle is not the chemistry but the detergent, and here there is an argument the field tends to miss. Detergent virus inactivation is an established, regulatorily accepted process step in its own right. A detergent that both elutes the product and inactivates enveloped virus could remove both acid exposures — the elution and the hold — in exchange for a clearance burden you would then be carrying for a reason. That is a stronger proposition than "avoid acid, add a validation problem", and it is the version worth evaluating.

Does a different ligand solve the problem?

Changing the ligand changes the problem. Protein G is not an answer here — it generally needs a lower elution pH, around 2.5–2.8, and it answers a different question, covered in Protein G capture for murine IgG. More relevant are ligands engineered for a specific Fc subdomain or variable-region family, including camelid-derived ligands, which typically release their target around pH 4.0–4.5 — and above pH 5 only with additives such as magnesium chloride or propylene glycol, which is the polarity modifier doing the extra work rather than the ligand.

The trade is selectivity and revalidation. A different ligand means a different host-cell-protein profile and a different leaching signature, and the comparison has to be made on your feed rather than on a datasheet. Where the requirement is genuinely unusual, a ligand can also be immobilised on a blank matrix — that is what custom affinity means in practice.

Is mixed-mode capture an acid-free alternative?

Hydrophobic charge-induction chromatography binds through a hydrophobic interaction and releases the product when the ligand picks up charge as the pH drops. The common ligand has a pKa near 4.8, so elution runs at roughly pH 4.0–5.5. That is worth stating plainly: this is not an acid-free route, and the pH window is similar to the mild-elution ligands of Route 1.

What it does offer is a much lower medium cost, independence from subclass and species, and the ability to load directly without conditioning the feed — real reasons to choose it, not merely a compromise. What you give up is the largest single purification factor in the train: an affinity capture step delivers roughly 98–99 % purity and around three log of host-cell-protein reduction in one operation, and mixed-mode does not come close. Whether that is acceptable depends on what the rest of your train can absorb.

Can light replace the elution buffer?

The last route removes the chemical trigger entirely. Solaris® Protein A, built on the PureLight® ligand platform, carries a photoswitchable group inside the ligand. Under red light (630 nm) the ligand sits in its binding conformation and the antibody is captured at native pH and ionic strength; a blue pulse (480 nm) isomerises the switch, the affinity collapses, and the product leaves in the buffer already flowing. There is no acid, no elevated salt and no additive — the elution buffer is the running buffer.

Three honest boundaries. Solaris® Protein A is in Beta, with access by application, and the demonstrated scope is IgG on a Protein A ligand. In a process that still runs a low-pH virus-inactivation hold, removing the acid from the capture step removes one of two acid exposures, not both. And the physical constraint of any photo-controlled chromatography is that the light has to reach the whole bed uniformly — which is what makes a flat convective body the plausible format and a deep packed column the implausible one, and which means such a device scales by area rather than by bed depth. What light-controlled capture does and does not solve is set out in more detail on the PureLight® page.

Two families of light-controlled purification exist, and they are easy to confuse. In the published academic route the photoswitch sits on the target protein as a genetically encoded tag or a light-switchable binding domain, and release is triggered with UV-A. In the Solaris® approach the photoswitch sits in the ligand and switches with visible light, so the product carries nothing and is never engineered for the method. The first is the more general research tool; the second is the one that applies to a molecule you are not allowed to modify.

Can you keep the acid and just shorten it?

None of the nine routes is free, and for a large share of real problems the cheapest fix is not an alternative chemistry at all. Aggregation at low pH is kinetic: the same antibody in the same buffer aggregates far more in twenty minutes than in two. Three levers cost nothing — pre-load the neutralisation buffer so each fraction is neutralised as it lands rather than at the end, collect in smaller fractions, and verify that the pooled pH ends between 6.5 and 7.5 rather than assuming it.

The capture format belongs in that list too, with one condition attached. Packed-bed capture commonly runs at a residence time of a few minutes, while a MonoCore™ Protein A capsule runs at 5–10 monolith volumes per minute, which puts the column part of the elution in seconds. But the column is only part of the exposure: the product then waits in the collection vessel, still acidic, until it is neutralised — and that wait is the same whatever the column format. A shorter residence time pays off when the neutralisation is pre-loaded or in-line, and hardly at all when the pool stands. The buffer side of the same decision is in Protein A elution buffer: glycine, citrate or acetate.

Two approaches deliberately left off the list: competitive elution with a soluble Fc fragment or Protein A mimetic peptide, which works and is priced out of any real process, and raising the temperature, which trades one denaturation risk for another.

Which route should you try first?

For a full-length IgG that aggregates: pre-load the neutralisation and shorten the acid contact, then a mild-elution ligand. Between them they solve most cases, and neither adds a unit operation.

For a bispecific or fragile fusion protein: arginine at development scale to find out how much pH headroom the molecule has, then decide between a mild-elution ligand and an amino-acid buffer system for the process, with glycols available if you end up half a unit short.

For acid-labile conjugate chemistry: shortening helps proportionally but rarely enough, and the virus-inactivation hold will undo what the elution saved. Go to the genuinely acid-free routes — engineered calcium-dependent ligands, detergent elution, light-controlled release — and expect the downstream consequences to be the real work.

And in all three cases the same cheap experiment comes first: aggregate measured in the eluate and again after the hold, before any money is spent on a new medium.

Frequently asked questions

How can I avoid low-pH elution in Protein A chromatography?

There are nine practical routes: mild-elution ligands releasing at pH 4.3–5.0, amino-acid or pH-shift buffer systems, arginine, glycols and polarity modifiers, salt-mediated or calcium-dependent elution, detergent-induced elution, subdomain-specific alternative ligands, mixed-mode capture instead of affinity, and light-controlled release, where the ligand is switched by visible light instead of buffer chemistry (Solaris® Protein A, in Beta). Only three of them — calcium-dependent ligands, detergents and light — genuinely remove the acid; the rest make it milder or shorter. The most common answer in practice is a mild-elution ligand combined with faster neutralisation, because neither changes the shape of the process. Which one fits depends on whether the problem is time-dependent aggregation — where shortening the acid contact helps, and a convective format such as a MonoCore™ Protein A capsule at 5–10 monolith volumes per minute puts the column part of the elution in seconds — or genuine acid lability, where only a non-acidic route works.

What are the alternatives to acid elution for antibodies?

Truly acid-free elution is possible in three ways: chelation from an engineered calcium-dependent ligand, an alkyl amine-oxide detergent at pH 5.5–6.5, and light-controlled release at neutral pH. Each requires something you do not already have — a matched ligand, a detergent clearance argument, or a technology still in Beta. Mild-elution ligands, amino-acid buffers, arginine and glycols do not remove the acid; they reduce it, which is enough for most molecules.

What does "mild elution" mean, and where does it stop?

Mild elution describes Protein A ligands engineered to release the antibody at roughly pH 4.3–5.0 instead of pH 3.0–3.5, usually by building additional histidines into the ligand. It is a real and well-documented improvement for aggregation-prone molecules. It stops at two boundaries: molecules whose weak point already relaxes around pH 5, and the virus-inactivation hold, which is typically validated at pH 3.6 or below — so a mild-elution pool has to be acidified separately, and the acid moves rather than disappears.

My bispecific aggregates during Protein A elution — what should I change first?

Find out when it aggregates before changing anything. Take a sample directly from the eluate and another after the neutralisation and hold, and compare by size exclusion. If the aggregate is already in the eluate, the elution is the problem: pre-load the neutralisation, then try arginine or a mild-elution ligand. If it appears afterwards, the hold is the problem and a milder elution will not help. On an asymmetric format the heterodimer interface is usually the first thing to relax, which is why half a pH unit can matter disproportionately.

If I stop eluting at low pH, do I still need the low-pH virus-inactivation hold?

In a process that relies on low pH for viral clearance, yes — and this is the objection that decides most of these routes in a pharmaceutical setting. The hold is a separate, independently validated step, typically at pH 3.6 or below for 30 to 60 minutes, and a gentler elution does not touch it. Removing the acid from the capture step removes one of two acid exposures, and for a product that cannot survive either, the inactivation has to be solved on its own terms — orthogonal clearance such as detergent inactivation or nanofiltration, argued on its own viral clearance data. The one route that could address both at once is detergent elution, because detergent virus inactivation is itself an accepted method.

Does a shorter residence time reduce acid-induced aggregation?

It reduces the column part of the exposure: a MonoCore™ Protein A capsule at 5–10 monolith volumes per minute elutes in seconds where a packed bed at a few minutes residence time takes correspondingly longer. Whether that translates into less aggregate depends on what happens next — if the eluate then stands in a vessel before neutralisation, the saving is largely lost, because the pool is acidic for as long as it waits. For acid-catalysed cleavage of a linker the benefit is proportional to the time saved and usually too small to matter on its own.

Can salt elute an antibody from Protein A instead of acid?

Not from a conventional Protein A ligand at moderate concentrations — 0.5–1 M sodium chloride is a wash condition, not an elution condition, and it tends to strengthen the hydrophobic part of the contact. Salt elution of standard Protein A requires chaotropic conditions such as 3–4 M magnesium chloride. Gentle salt or chelation elution near neutral pH belongs to purpose-engineered ligands, and even there IgG1 often still needs around pH 5.5–6.

Is light-controlled elution available today?

Solaris® Protein A is in Beta, with access by application. The demonstrated scope is IgG capture on a Protein A ligand. It is a real option to evaluate where the acid elution is the step limiting your molecule, and not yet an option for a validated manufacturing process.