Size-Exclusion Chromatography for Peptides: Oligomers, Aggregates and Non-Ideal Elution

What SEC separates

Size-exclusion chromatography (SEC) pumps a sample through a column of porous particles. In ideal SEC nothing binds: a molecule simply explores the pore volume it can fit into. Species too large to enter the pores elute at the void volume (V₀); very small molecules that reach all of the pore volume elute near the total permeation volume (Vt). Everything else elutes in between, largest first. For peptides that means aggregates and oligomers elute before the monomer, which is why SEC is the usual first look at aggregation.

Unlike reversed-phase HPLC (reading an HPLC chromatogram), SEC ignores hydrophobicity, so species of similar size co-elute.

Kav and calibration arithmetic

Elution position is usually expressed as the partition coefficient Kav = (Ve − V₀) ÷ (Vt − V₀), where Ve is the elution volume. Kav runs from 0 (fully excluded) to 1 (fully included).

Worked example (illustrative numbers): V₀ = 5.0 mL and Vt = 15.0 mL. A monomer peak at 11.0 mL gives Kav = 6.0 ÷ 10.0 = 0.60; an earlier peak at 9.8 mL gives 4.8 ÷ 10.0 = 0.48. Suppose standards run under the same conditions fit log₁₀(MW) = 5.2 − 2.5 × Kav. The monomer then reads 10^(5.2 − 1.50) = 10^3.70 ≈ 5,000 Da and the earlier peak 10^(5.2 − 1.20) = 10^4.0 = 10,000 Da, consistent with a dimer. The peptide mass calculator gives the expected monomer molecular weight for that comparison.

Two cautions. SEC responds to hydrodynamic size and shape, not mass, so an extended peptide can read heavier than a compact one of the same mass. And a calibration is only valid when the sample elutes by size alone, which is the subject of the next section.

Ideal and non-ideal elution

Peptides often do not behave ideally on SEC packings. Mant, Parker and Hodges made a set of synthetic peptide standards, Ac-(G-L-G-A-K-G-A-G-V-G)n-amide with n = 1 to 5, to monitor both ideal elution and non-ideal behaviour caused by hydrophobic or ionic interactions with the column material.^[1] Adding salts or organic solvents overcame those interactions, and the authors concluded that when the conformation of a sample in a given mobile phase is uncertain and pure size separation is needed, SEC should be run under highly denaturing conditions.

Ionic interactions depend on charge, so the pI and net charge calculator is a useful check: a strongly basic peptide at neutral pH is a candidate for binding to residual negative charges on the packing, the same chemistry described in peptide adsorption to labware.

Choosing the column and mobile phase

A 2026 benchmark evaluated eleven modern SEC columns, including prototypes designed to minimise non-specific interactions, for peptide analysis.^[2] Columns with pores of 100 Å or larger performed best, and one prototype reached a reduced plate height close to 1, compared with the typical range of 2 to 3. The most denaturing mobile phase tested, 30% acetonitrile with 0.1% trifluoroacetic acid, gave the best performance for linear, macrocyclic, disulfide-constrained and lipid-conjugated peptides by suppressing hydrophobic and ionic interactions. For hydrophobic lipidated peptides such as liraglutide and semaglutide, the authors proposed a phosphate mobile phase with 20% isopropanol instead, to limit denaturation and potentially allow non-covalent aggregates to be characterised.

Additives can also protect aggregates. Working with proteins, one group found that arginine in the mobile phase reduced the loss of aggregates to the column, giving a reliable aggregate estimate even on a new column with a small load; sodium chloride, by contrast, increased protein loss, so the effect was not simply ionic strength.^[3]

When the mobile phase changes the answer

Insulin shows how much conditions matter. The United States and European Pharmacopoeias use an acidic-pH SEC method for insulin aggregates. A comparison found that neutral-pH SEC and capillary gel electrophoresis gave similar aggregate contents, while the acidic method showed polymeric complexes not seen by either neutral method; arginine and acetonitrile both affected the SEC profiles during method development.^[4]

For synthetic human parathyroid hormone fragment 1-34, non-covalent aggregates dissociated through hydrophobic and electrostatic interactions with the packing until salts and organic solvent were added to the mobile phase; under the optimised conditions the disaggregation was completely suppressed.^[5] The mobile phase is part of the result and belongs in the report.

Aggregates that fall apart on the column

Even with interactions suppressed, a weak aggregate may not survive the run. For an acylated GLP-1 analogue, free-solution capillary electrophoresis was compared with SEC; SEC showed a significantly slower apparent rate of aggregation because the aggregate was disrupted during separation, possibly through on-column dilution and interactions with the packing.^[6] CE has no stationary phase, as described in capillary electrophoresis for peptides.

Worked example of on-column dilution (illustrative): 20 µL of a 1.0 mg/mL solution is loaded and the peak elutes in about 0.5 mL. The peptide is diluted about 25-fold, to roughly 0.04 mg/mL, by the time it reaches the detector, enough to shift a concentration-dependent equilibrium toward monomer. The dilution calculator handles the same arithmetic.

Stable oligomers and their history

Some oligomers are robust. Using SEC, one group isolated stable low-molecular-weight oligomers of GLP-1 and its C-terminal amide, containing two to five chains, with a highly disordered structure.^[7] Under the conditions studied they resisted fibrillation and dissociation over time, temperature and agitation; their non-covalent character was established by LC-MS and SDS-PAGE. They formed off the pathway to amyloid fibrils, which the thioflavin T assay follows.

Oligomer state can also carry over from earlier steps. Liraglutide forms oligomers that can be stabilised by pH and organic solvents, and as a synthetic peptide it kept its oligomerization state in dried lyophilized powder, a memory effect from synthesis and purification that influenced its tendency to aggregate and form fibrils.^[8] How a sample was dissolved therefore matters; see peptide solubility and aggregation.

Area percent and recovery

SEC results are usually reported as area percent of high-molecular-weight species, dimer and monomer. Worked example (illustrative): peaks of 120 (aggregate), 300 (dimer) and 9,580 (monomer) total 10,000, so high-molecular-weight species plus dimer are 420 ÷ 10,000 = 4.2%.

That value assumes everything loaded came off the column. A simple check compares the total SEC area with a reference load of freshly dissolved material. If the stressed sample gives a total of 9,000 against a reference of 10,000, recovery is 90%, and up to 10% of the material, possibly aggregate retained on the column, is missing from the area percent. Report recovery alongside the distribution.

Record checklist

Record enough that another lab could reproduce the size distribution.

• Column. Packing chemistry, pore size, dimensions, age or number of runs, V₀ and Vt markers.
• Mobile phase. Buffer, pH, salt, organic solvent and additives with concentrations, flow rate, temperature.
• Sample. Solvent, concentration, time since dissolution, loaded volume and amount.
• Results. Elution volumes or Kav, area percent per species, total area and recovery, calibration standards if MW is estimated.
• Cross-check. Which orthogonal method (CE, thioflavin T, circular dichroism, MS) was compared and whether it agreed.

Orthogonal methods are described in circular dichroism and the guides above. This guide describes general laboratory methods and published literature only. It is not a statement about the composition of any particular material. Terms are defined in the glossary; questions can be sent to support.