Size-exclusion chromatography concepts

Size-exclusion chromatography separates by how deeply a molecule can partition into pores. You choose it when the decision is aggregate versus monomer, a fragment versus a full chain, or a buffer exchange that must not rely on binding. It is a weak way to fish one protein out of a lysate. The surrounding menu of modes is in chromatography methods in life-science labs. Affinity capture, which often comes first, is treated in affinity and ion-exchange protein purification.

Partitioning, not binding

In a well-chosen buffer the resin is a porous gel the protein does not stick to. IUPAC chromatography is still the right name: the molecule distributes between interstitial mobile phase and the stagnant liquid inside pores. A species larger than the pores never enters them. It elutes at the void volume, the liquid outside the beads. A species small enough to enter every pore explores the included volume and elutes later. Proteins in between have a partial access and elute between those two limits. There is no useful retention beyond the included volume. Anything later than the salt peak is interacting with the resin, and the run has stopped being size exclusion.

Resolution between aggregate and monomer is the usual protein question. Aggregates are larger hydrodynamically and lead. The monomer follows. Fragments and free dye or salt trail toward the included volume. Whether those zones separate depends on the pore size distribution, the bed height, the linear velocity, and how much volume you loaded. A short, wide bed is a desalting tool. A long bed is where aggregate resolution has a chance.

Sequence mass from UniProt is a planning hint. Hydrodynamic size in your buffer, with your tag and your oligomeric state, is the quantity the column sees. A mass-spectrometry identification of what was in the tube is a later claim, in the spirit of the reporting HUPO exists to improve. The early peak is not automatically "aggregate" until an orthogonal check says so.

The load is a small slice of the bed

Sample volume is part of the peak width. You are not sharpening the band on a gradient. You are asking a narrow plug to stay narrower than the separation between two sizes. Typical analytical loads are a few percent of bed volume. Some resin notes are tighter than that. Preparative loads can be larger when you only need to move salt away from protein, and they destroy aggregate resolution if you treat a desalting volume as a polishing volume. Follow the resin note for the column in front of you. A 24 millilitre bed and a 0.5 millilitre injection are in the few-percent range. A 5 millilitre injection on that bed is a different experiment.

Measure the injection with a device you trust. ISO 8655-1 is the general standard for piston-operated volumetric apparatus. A sloppy 2 percent load that was actually 6 percent will look like a column that "cannot resolve aggregates" when the column never had the chance.

Concentration has a separate ceiling. Very high protein can increase viscosity and create viscous fingering: the peak leans and trails for a hydrodynamic reason. Dilute into the elution buffer rather than stacking more mass into the same microlitres. The resin note's mass load is the limit, not a target to hit.

Hardware and buffer class

You need a packed bed of known column volume, a pump or a gravity head that holds a steady linear velocity inside the resin's pressure rating, a detector if you are tracing, and a collector. HPLC-format size-exclusion columns use finer particles and higher pressure. Soft prep resins do not. Do not move an HPLC flow onto a soft bed because the peak looked good on the scout.

The buffer should suppress electrostatic binding. Many protein methods add a moderate salt for that reason, at a concentration the resin note suggests, often on the order of a hundred millimolar rather than a multi-molar extreme. pH should keep the protein soluble and inside the resin's window. Reducing agents, detergents, and organic solvents are allowed only when the insert says the bed tolerates them. A buffer that differs wildly from the sample buffer will still exchange, but the first runs can show distorted peaks while the bed finishes equilibrating. Equilibrate with several column volumes until conductivity and pH match the buffer you think you applied.

Filter the sample. Clarify it. Size exclusion has no binding step that ignores debris. Particles sit on the frit and raise pressure, which on a soft gel quickly becomes a compressed bed.

The run, and the branches

Equilibrate. Inject a volume inside the resin's fraction of bed volume. Run at the recommended linear velocity until at least one included volume has passed, so salt and small molecules have cleared. Mark the void with a large standard if you need to teach the column to a new user, and mark the included volume with a small inert marker the resin note accepts. Your protein should fall between them if it is a soluble monomer inside the fractionation range.

If aggregate and monomer form one peak, the pore range may be wrong, the bed may be too short, the load may be too large, or the flow may be too fast for pore equilibration. Change one of those and repeat. Do not interpret a single 280 nanometre hump as "mostly monomer". If the protein elutes after the salt, it is sticking. Raise the ionic strength into the range the insert allows, or accept that this resin is behaving as an ion exchanger and move the separation to a mode designed for charge.

If the void peak grows across a day of injections, the sample may be aggregating in the vial, or the column may be contaminated and shedding. Run a blank injection. A blank that still shows the void peak points at the column or the buffer, not at the next sample. A blank that is clean points at the sample sitting in the autosampler.

Collect fractions across the peak rather than one tube under the apex. Assay the leading edge, the apex, and the trailing edge by activity, by a gel, or by absorbance only as a locating tool. Pool the tubes the assay supports. Absorbance at 280 nanometres locates aromatic protein. It does not score aggregate content.

Failure modes

A tail into the included volume can be a fragment, a protease nick that happened on the column, or interaction. Hold an aliquot of the load and compare it on a gel with the late fractions. If the load was already nicked, the column did its job. If the load was intact and the late tubes are fragments, look at time, temperature, and proteases, not at the pore size first.

Channeling from a dried or poorly packed bed makes everything elute early and broad. A standard pair that used to resolve and now overlaps is a bed problem until proved otherwise. Repack or replace. Do not keep "subtracting a blank" from a channeling column.

High-pressure size-exclusion hardware is still size exclusion. Moving a method from a soft gravity column onto an HPLC column changes particle size, pressure, and sometimes the pore. Re-establish void and included volume on the new bed. A retention time copied from the open column will not land.

Research-use limits

The trace supports a size-based cut for research. It does not prove a pharmaceutical oligomeric specification, and it does not make a diagnostic claim. Aggregates can be immunogenic or inactive in a research assay. Your assay decides whether the leading edge is a problem. Biosafety of the protein sample is an institutional decision. The WHO laboratory biosafety manual is a public reference for how laboratories frame that decision, not a permit for this run.

What to ask for

State the two sizes you must resolve, the sample volume and concentration, the buffer composition, the acceptable pressure, and analytical versus preparative intent. Ask for fractionation range, particle size, column volume, and a recommended linear flow. Hardware classes are listed in the scientific instruments catalogue. Send the specification with the quote request. If size exclusion is the polish after a tagged capture, the method question can sit beside the custom protein expression and purification reference. That page is a prompt for the discussion. ---