Most Protein A problems announce themselves as one of eight symptoms, and each has a most likely cause that can be confirmed in a single check before anything is changed. Work through the table, confirm the cause, then fix one variable at a time — changing buffer, flow and load together is how a bad run becomes an unexplained run.
Symptom table
| Symptom | Most likely cause | Check first | Fix |
|---|---|---|---|
| Low yield, product in the flow-through | Load exceeded capacity, or residence time too short for the medium | Read the UV trace of the flow-through and the wash — product there settles it; confirm by re-running that fraction over a fresh column if the trace is ambiguous | Reduce the load, or lengthen the residence time; verify with a breakthrough curve |
| Low yield, product not in the flow-through | Incomplete elution — buffer pH too high or elution too short | Measure the actual pH of the elution buffer; strip afterwards and look for a late peak | Correct the pH, extend elution to 5 CV, then strip at pH 2.5 |
| Aggregate after elution | Too long at low pH | Compare aggregate at the elution condition with aggregate after neutralisation | Neutralise in the collection tube, smaller fractions, consider a milder buffer at a higher pH |
| Host-cell protein in the eluate | Non-specifically bound impurity co-eluting | Does a longer plain wash change anything? If not, it is bound, not free | Add an intermediate wash, 3–5 CV: 0.25–1 M NaCl in phosphate at pH 7.4 — the effect saturates near 250 mM — or 0.5–1 M arginine, which works on the HCPs salt does not shift |
| Back pressure rising over cycles | Fouling at the inlet, or a feed that was not clarified enough | Pressure with buffer only, on a clean column, at the working flow rate | Improve clarification and filtration; clean earlier. On a packed bed the caustic step is commonly run in reverse flow because the inlet fouls first; on any other format, follow its datasheet |
| Capacity falling over cycles | Ligand loss from cleaning, or reversible fouling | Strip and clean, then repeat the breakthrough curve — recovered capacity means fouling | If fouling: adjust the cleaning regime. If ligand loss: it is permanent, plan the retirement |
| Carryover between runs | Incomplete regeneration | Run a blank cycle and look at the elution trace | Strip every cycle at the specified volume; if it persists, shorten the cleaning interval rather than the strip |
| Leaked ligand in the eluate | Ligand hydrolysis, ageing medium or harsh cleaning | Quantify with a ligand-specific assay across the cycle history | Soften the cleaning regime; leaching that rises with cycle count signals the end of service life |
The three that are usually misdiagnosed
Yield is low, so the column must be worn out
Usually it is not, but the first move is not the elution either — it is finding out where the product went. Re-read the UV trace of the flow-through and the wash: product there means the load exceeded the capacity at that residence time. Product nowhere means it is still bound, and then the elution is the suspect: measure the actual pH of the buffer, because a nominal pH 3.0 that reads 3.4 will not release everything, and look at the strip afterwards, where a large peak is the signature of an elution that was too mild. Measure capacity properly with a breakthrough run rather than inferring it from a yield number — what the figure means, and what it needs alongside it.
Aggregate is a product problem
It is usually a process problem. The same molecule in the same buffer aggregates far more when it sits at low pH for twenty minutes than for two. Fix the collection first — pre-loaded neutralisation, small fractions — and only then reach for a different buffer. Which buffer to reach for is covered in Protein A elution buffer.
Pressure is a pump problem
Sometimes. Confirm by running buffer through the system without the column: if the pressure is normal, the column is fouling, and the cause is upstream. Clarification and filtration decide how fast a capture step blocks; the mechanisms are described in why chromatography media foul.
Prevention checklist
- Clarify the feed, and match the filter train to this feed rather than the last project's — a packed bed needs the 0.22 µm step right before the column far more than an open-channel capsule does.
- Measure the pH of every buffer you prepare, including the elution buffer, at the temperature you will run it.
- Pre-load the collection tubes with neutralisation buffer before every run.
- Strip after every cycle at the volume the method specifies; clean in place on a fixed interval rather than on suspicion. If carryover persists, shorten the cleaning interval — stripping harder or longer costs ligand without fixing it.
- Record back pressure at a fixed flow rate once per cycle — it is the earliest warning you get.
- Re-measure binding capacity at fixed intervals against a baseline you recorded when the column was new.
- Load to 70–80 % of the capacity you measured yourself, and log cumulative cleaning cycles rather than runs — that is the number that predicts the end.
Running the same checks on a monolith capsule
The symptoms are the same; two of the checks change. A MonoCore™ capsule has no packed bed, so a rising back pressure is not bed compression — it is material on the inlet face, and the 15 µm channels make that slower to happen. Which cleaning direction a device tolerates is a question for its datasheet: on a packed bed, reverse-flow cleaning is the vendor default rather than an exception.
The second difference is the residence time. At about 8 column volumes per minute — 15 mL/min on the 1.8 mL format, 45 mL/min on the 5.4 mL — the product spends seconds in the medium rather than minutes, so "the residence time was too short" has a different meaning than on a packed bed. Check it with the residence time calculator before you conclude that the load was too high. The cleaning conditions for the affinity capsules come with the datasheet.
Frequently asked questions
Why is my Protein A yield low?
Two different failures look identical on the chromatogram. If the product is in the flow-through, the column did not hold it: the load was too high or the residence time too short. If it is not in the flow-through, it did not come off: check the elution pH and extend the elution, then strip at pH 2.5 and look for a late peak.
Why does the back pressure rise from run to run?
Because something is depositing on the medium. Confirm it is the column and not the system by running buffer without it. Then look upstream — insufficient clarification is the usual cause — and clean earlier rather than harder.
How do I remove host-cell protein that co-elutes?
With an intermediate wash before elution: typically sodium chloride in 20 mM phosphate at pH 7.4, or 0.5–1 M arginine, for 3–5 column volumes. A longer plain wash does not help, because the impurity is bound rather than merely present. Note that the salt effect saturates — around 250 mM in published work — so if a salt wash has not solved it, more salt will not either; arginine is the next step.
My column lost capacity. Is it dead?
Not necessarily. Strip and clean it, then repeat the breakthrough curve. Capacity that comes back was fouling; capacity that does not is ligand loss, which is permanent. Comparing against a baseline you recorded when the column was new is what makes the answer unambiguous.
Is ligand leaching a sign of a worn column?
Rising leaching over the cycle history is one of the clearest end-of-life signals, and it matters beyond the column because leached ligand has to be cleared downstream. Quantify it with a Protein A ELISA on an acid-dissociated sample — leached ligand travels bound to the Fc and is otherwise under-reported — and track it against cumulative cleaning cycles rather than against runs. What drives it, and how it fits into a lifetime study, is in Protein A column regeneration, CIP and service life. For a MonoCore™ affinity capsule, the cleaning conditions that belong in that study are in the datasheet.
Still an open question?
Tell us the molecule, the feed volume and the system you run it on, and we will say whether a MonoCore™ capsule is the right starting point — or whether it is not.