Chromatography polishing
Separation by hydrodynamic radius on a porous matrix, with no binding: molecules too large for the pores elute at the void volume and smaller ones are progressively retained. The reference method for aggregate, and a buffer exchange in the same run.
Our ruleSimulations prepare you for the bench but don't replace hands-on practice.
Size exclusion (SEC)
Separation by hydrodynamic radius on a porous matrix, with no binding: molecules too large for the pores elute at the void volume and smaller ones are progressively retained. The reference method for aggregate, and a buffer exchange in the same run.
How it differs from the standard method
- Load 0.5–2% of the column volume for preparative separation, up to 5% if resolution allows; this is the hard limit on throughput.
- The matrix must bracket your protein: a 10–600 kDa resin resolves a 34 kDa monomer from its dimer, a 5–5,000 kDa resin barely does.
- Always run in at least ~150 mM salt, or proteins interact ionically with the matrix and retention stops meaning size.
- The pool is always more dilute than the load, so plan a concentration step after it.
- Retention volume alone does not identify a species: calibrate the column, or use in-line light scattering (SEC-MALS).
1What capture left behind
An affinity column is very good at one thing: taking the protein that carries the tag. It cannot tell a correctly folded monomer from a dimer carrying the same tag, a proteolytic fragment that still has the tag, or a host-cell protein that happens to stick to the ligand.
2Three properties, three columns
Size exclusion
Separates by hydrodynamic radius with no binding at all: big molecules cannot enter the pores and leave first. The only step that shows you aggregate directly, and the only one that dilutes everything it touches.
Ion exchange
Binds by net charge, which depends on pH relative to the protein's isoelectric point. Above the pI the protein is negative and binds an anion exchanger; below it, positive and binds a cation exchanger. Elutes concentrated on a salt gradient.
Hydrophobic interaction
Loads at high salt, which drives exposed hydrophobic patches onto the resin, and elutes as the salt falls. The opposite sense to ion exchange, which is why it follows an ammonium sulfate precipitation naturally.
Mixed-mode and others
Ceramic hydroxyapatite, multimodal resins and heparin columns combine charge with something else, and are often what clears the last aggregate when ion exchange cannot.
3Size exclusion: resolution paid for in volume
SEC never binds anything, so recovery is high and nothing is denatured, but there is no concentration step to hide behind. The whole separation happens in the volume between V₀ (where excluded molecules come out) and Vt (the total column volume), so it is a narrow window, and the sample has to be injected into a small fraction of it: 0.5 to 2% of the column volume for a preparative run, 5% at most.
4Ion exchange: pH is the control knob
Net charge is a function of pH, so moving the pH a unit can turn a binding protein into one that passes straight through. That is a tool, not a hazard: picking a pH where your protein binds and a major contaminant does not (or the reverse, running the column in flow-through mode where only the contaminants bind) is often the cleanest separation available.
5Resolution against loading
Every polishing step trades resolution for throughput, and the usual way a good prep goes wrong is overloading. Peaks broaden well before a column is nominally full, because the dynamic capacity under flow is always lower than the static figure in the catalogue, and a broadened peak means the contaminant that was resolved at 5 mg is co-eluting at 50.
6Reading the trace
A single symmetric peak is not evidence of purity, only of one unresolved population. A leading shoulder on a SEC trace is almost always aggregate; a trailing shoulder is a clipped form or a late-eluting contaminant.
Key terms
Practise these as flashcards →Purification steps after capture, aimed at aggregate, product-related variants, host-cell protein, DNA and endotoxin rather than at bulk purification.
The partition coefficient in SEC, (Ve − V0)/(Vt − V0); linear in log molecular weight over the matrix's fractionation range.
Void volume (where excluded molecules elute) and total column volume (where small molecules elute); everything separates in between.
The pH at which a protein has no net charge: where it binds neither ion exchanger and is least soluble.
Protein a column holds at a given flow rate before breakthrough; always below the static capacity quoted on the datasheet.
Peak separation divided by the sum of the peak widths; around 1.5 is where two peaks can be cut apart cleanly.
Running a column at conditions where the contaminants bind and the product does not: fast, high capacity, and often overlooked.
Protein from the expression host that survives capture; measured by immunoassay or mass spectrometry, not by a Coomassie gel.
When to use it
- After an affinity capture, when the eluate still shows aggregate on analytical SEC or extra bands on a gel.
- When the application cannot tolerate aggregate: crystallography, biophysics, cell assays, anything injected into an animal.
- To separate a protein from a close variant (a clipped form, a deamidated or oxidised species, a charge variant) that affinity cannot see.
- To exchange buffer and remove imidazole, salt or chaotrope at the same time as purifying (size exclusion does both).
- To concentrate a dilute eluate: ion exchange binds from a large volume and elutes in a small one.
Compared with alternatives
The capture step that comes first; polishing exists to clean up what affinity cannot distinguish.
The analytical version of size exclusion, with light scattering so the peaks are assigned by mass rather than by retention.
A cheaper way to change buffer if that is all you need, with no separation and no dilution control.
A quick check for aggregate in a pool without running a column at all.
An ammonium sulfate cut is the classic step before hydrophobic interaction, since both work in high salt.
The same logic at process scale, where capacity, buffer volumes and cycle time decide the step order.
Sub-techniques
Choosing one here switches the whole page, including the protocol and the lab, to it.
Separation by hydrodynamic radius on a porous matrix, with no binding: molecules too large for the pores elute at the void volume and smaller ones are progressively retained. The reference method for aggregate, and a buffer exchange in the same run.
- Load 0.5–2% of the column volume for preparative separation, up to 5% if resolution allows; this is the hard limit on throughput.
- The matrix must bracket your protein: a 10–600 kDa resin resolves a 34 kDa monomer from its dimer, a 5–5,000 kDa resin barely does.
- Always run in at least ~150 mM salt, or proteins interact ionically with the matrix and retention stops meaning size.
- The pool is always more dilute than the load, so plan a concentration step after it.
- Retention volume alone does not identify a species: calibrate the column, or use in-line light scattering (SEC-MALS).