Recombinant protein purification
Enzymes, growth factors, vaccine antigens, fusion proteins — the work is the same shape every time: separate one protein from everything the host made alongside it. Which chemistry does that, and where the feed decides the route.
Most recombinant protein work has the same shape. Something expresses your protein, and it expresses a great deal else besides. The purification has to separate the two, usually in two or three chromatography steps.
What changes between projects is not the principle but the feed — and the feed is decided by the expression system long before anyone picks a column.
The expression system decides the first problem
| System | What you get | What that means for the first step |
|---|---|---|
| E. coli, soluble | Lysate: host protein, DNA, membrane fragments, endotoxin | Clarification matters as much as the column. Endotoxin removal usually rides on an anion exchange step |
| E. coli, inclusion bodies | Insoluble protein requiring refolding | Chromatography starts after refolding; the feed carries refolding buffer and misfolded species |
| Yeast (Pichia, S. cerevisiae) | Secreted protein in a dense medium | Often the cleanest start — but the medium is viscous and proteases are active |
| CHO or HEK, secreted | Supernatant with host-cell protein and DNA | Similar to an antibody harvest without the Protein A shortcut |
| Cell-free | Reaction mix with the expression machinery in it | Small volumes, high value, and a background unlike any cell-based system |
What these have in common: the target is a protein among many proteins, and the separation runs on charge, hydrophobicity or a specific ligand. How far the feed has to be clarified before any of that begins is covered in feed preparation.
The chemistries, and what each one is for
| Chemistry | Role | Typical mode |
|---|---|---|
| Anion exchange (Q) | Capture for acidic proteins; DNA, endotoxin and acidic host protein in flow-through | Either, depending on the target's pI |
| Cation exchange (S) | Capture for basic proteins; polishing where charge variants have to be resolved | Usually bind and elute |
| Hydrophobic interaction (HIC) | Polishing, and the step that removes aggregate and misfolded species | Bind and elute, or flow-through |
| Custom affinity | Where a specific ligand exists for your target and the selectivity is worth it | Bind and elute |
A typical route puts ion exchange first, chosen by whether the target sits above or below the pI of the bulk contaminants, then HIC as an orthogonal second step. Two mechanisms that do not correlate remove more than two steps of the same kind.
Where the format matters
Two situations come up often enough in protein work to be worth naming.
A large or awkward target. Large proteins, multimeric complexes and heavily glycosylated species diffuse slowly, and a process that waits on diffusion runs at the pace of the slowest molecule in it. Convective transport does not depend on diffusion to the same degree — which is why the format is worth testing early for these targets rather than after two chemistries have been screened. The mechanism is set out in convective vs diffusive mass transport. MonoCore™ ships with 15 µm channels.
A difficult feed. Lysates and dense yeast media carry material that blocks narrow flow paths. The wider channel is margin against that, which is why it is the format question rather than the chemistry question — the chemistry would be the same on any support.
Where neither applies — a small, soluble protein in a clean supernatant, and time is not the constraint — a packed bed is a well-understood and widely supported choice, and we would say so.
Process development
Most recombinant protein work happens at a scale where method development, not manufacture, is the cost. Screening pH and salt across a few conditions means many short runs, and the time per run is what limits how many conditions get tested in a week.
That is the argument for a convective format at this stage: runs at flow rates a packed bed cannot reach within its pressure limit, so a screen that took two days takes one. How that arithmetic works is in rapid cycling chromatography, and the run planner lets you put your own method in rather than take ours.
- Which chromatography steps does a recombinant protein need?
- Two or three, working on different properties. Where no affinity step applies, the separation runs on charge and hydrophobicity. Ion exchange first — anion or cation depending on the target's pI relative to the bulk contaminants — then hydrophobic interaction as a second mechanism that does not correlate with the first. A third step is added where aggregate or a closely related species survives the first two.
- Which chromatography step comes first for a recombinant protein?
- Usually ion exchange, because it is robust, well understood and tolerant of the volumes a fermentation produces. Which of the two depends on the pI of your target relative to the host background: anion exchange where the target is acidic, cation exchange where it is basic. The choice is worth testing rather than calculating, since pI predicts behaviour imperfectly.
- How is protein from E. coli different from protein from CHO?
- The feed, not the chromatography. An E. coli lysate carries host protein, DNA, membrane fragments and endotoxin, so clarification is a real process step and endotoxin clearance has to be demonstrated. A CHO supernatant is closer to an antibody harvest — cleaner, but without the Protein A shortcut that makes antibody processes simple.
- What removes aggregate from a recombinant protein preparation?
- Hydrophobic interaction chromatography, most often. Aggregate exposes hydrophobic surface that the monomer keeps folded away, which makes HIC the more selective tool for it than ion exchange. Size exclusion also separates aggregate but does not scale well, which is why it stays a polishing step at small volumes.
- Which MonoCore™ chemistry should I start with?
- Anion or cation exchange, chosen by your target's pI, with HIC as the orthogonal second step. If your protein is large, part of a complex, or your feed is a lysate rather than a supernatant, the open channel format is worth testing early — those are the cases where it changes the result rather than only the timing.