Capillary Electrophoresis for Peptides: Mobility, Purity, Charge Variants and Counter-Ions

Why a second separation principle

Most peptide purity values come from reversed-phase HPLC, which separates by hydrophobicity; see reading an HPLC chromatogram. An impurity with nearly the same hydrophobicity as the main peptide can hide under the main peak. Capillary electrophoresis separates by a different property, charge relative to size, which makes it an orthogonal check.

The difference can be large. For the basic peptide aviptadil, capillary electrophoresis on a coated capillary coupled to a mass spectrometer revealed 15 detectable impurities, while a nano reversed-phase HPLC-MS method detected two, and none of the impurities seen by one method appeared in the other.^[1] The CE-MS detection limit was about 1 fmol, roughly 2,500 times lower than the 3 pmol found for the LC method, and the authors concluded that the silica column and flow rate influenced what the HPLC method could see.

How capillary zone electrophoresis works

In capillary zone electrophoresis (CZE), a fused-silica capillary, typically 25 to 75 µm wide, is filled with a background electrolyte (buffer). A small sample plug is loaded at the inlet and a high voltage is applied. Each ion moves at its own electrophoretic mobility, while the whole solution also moves by electroosmotic flow (EOF), driven by the charged capillary wall. A UV detector near the outlet, often at 200 to 214 nm where the peptide bond absorbs, records an electropherogram.

Worked example (illustrative numbers): total capillary length 50 cm, 40 cm to the detector, 20 kV applied. A neutral EOF marker arrives at 4.0 min (240 s) and a peptide at 5.0 min (300 s). Apparent mobility is (detector length × total length) ÷ (voltage × time). For the peptide: (40 × 50) ÷ (20,000 × 300) = 3.33 × 10⁻⁴ cm²/V·s. For the EOF marker: 2,000 ÷ (20,000 × 240) = 4.17 × 10⁻⁴ cm²/V·s. The peptide’s effective mobility is the difference, about −0.83 × 10⁻⁴ cm²/V·s: it is an anion moving against the flow and is carried past the detector only because EOF is faster. Reporting effective rather than apparent mobility removes day-to-day EOF drift.

Charge and size predict migration

Peptide migration in CZE is unusually predictable. In a data set of 125 peptides of 2 to 14 residues measured at pH 2.5, the Offord model, mobility proportional to Q/M²ᐟ³ (charge over mass to the two-thirds power), gave the best correlation of the empirical models tested; adding steric and bulk descriptors and a neural network improved predictions, especially for peptides containing arginine, histidine and lysine.^[2] A later model built from more than 4,000 tryptic peptides confirmed charge and molecular mass as the two major factors, reached a correlation of R² ≈ 0.995, and found that Asp or Glu near the N-terminus lowers the N-terminal amine charge, by about 0.27 units for N-terminal Asp.^[3]

Net charge depends on pH and on each group’s pKa, so the pI and net charge calculator is a useful first step when choosing a buffer. At low pH, carboxyl groups are mostly protonated and peptides are cations; well above the isoelectric point, they are anions.

Charge variants: deamidation as an example

Deamidation turns asparagine into aspartate or isoaspartate, a mass change of only +0.984 Da but, at a pH where the new carboxyl is ionized, a change of one full charge. Worked example: for a 1,500 Da peptide, M²ᐟ³ ≈ 131. A net charge of −2 gives Q/M²ᐟ³ = 0.0153; after deamidation, −3 gives 0.0229, about 50% higher, while the mass changes by 0.07%. At pH 2.5 the same pair differs far less, because the new carboxyl is largely uncharged.

Published insulin work shows this in practice. Comparing three recombinant insulins, a CZE method using a basic electrolyte in bare fused silica separated forms differing by 0.984 Da and even same-mass forms with slightly different shapes, with UV detection limits of 0.6 to 0.93 mg/L; deamidation was minimal at −20 to −80 °C and accelerated above +5 °C.^[4] Electrospray MS detection was also used, with detection limits of 2.17 to 4.37 mg/L. The chemistry is covered in peptide deamidation.

Purity by CZE and corrected peak areas

For the synthetic LHRH analogue lecirelin, CZE in an acidic electrolyte (100 mM phosphoric acid, 50 mM Tris, pH 2.25) in bare fused silica with detection at 206 nm gave purity of 60.1 to 80.9% for crude material and 96.4% or higher for HPLC-purified batches, in agreement with capillary isotachophoresis.^[5]

One CE-specific step matters for area percent. Zones pass the detector at different speeds, so a slower zone gives a larger area for the same amount. Areas are therefore divided by migration time before normalizing. Worked example: main peak area 9,800 at 5.0 min and an impurity of 200 at 6.0 min. Uncorrected purity is 9,800 ÷ 10,000 = 98.0%. Corrected, 9,800 ÷ 5.0 = 1,960 and 200 ÷ 6.0 = 33.3, so purity is 1,960 ÷ 1,993.3 = 98.3%. Stating whether areas were corrected makes CE and HPLC values comparable.

Hydrophobic peptides and organic modifiers

Purely aqueous buffers do not suit every peptide. Analysing GnRH, CRF and analogues, Miller and Rivier found that hydrophobic amino acids and peptides often migrated as broad, tailing zones or even precipitated during CZE in aqueous buffers.^[6] Adding acetonitrile, methanol or isopropanol to sodium phosphate or triethylammonium phosphate buffers improved behaviour and allowed nine native GnRH decapeptides to be separated. Several had identical formal charges yet still separated, pointing to a mixed mechanism that also involves structure and interactions with the capillary.

Wall adsorption and capillary coatings

Bare silica is negatively charged above about pH 3, and basic peptides can stick to it, which broadens peaks and shifts migration times; the same chemistry is described in peptide adsorption to labware. A CE-MS protocol states that coatings are necessary to prevent peptide adsorption and achieve good intermediate precision, comparing a neutral coating with two cationic coatings that differ in the electroosmotic flow they induce.^[7] The method tolerated samples with up to 500 mmol/L sodium chloride, and the authors flag possible cross-contamination when switching between coatings on the same instrument.

Counter-ions measured by electrophoresis

Small anions migrate quickly, so CE also measures the counter-ions that make up part of a peptide’s weighed mass. A validated method with an internal standard and indirect UV detection determined acetic acid and trifluoroacetic acid together in synthetic peptide samples, with migration times of about 2 to 8 minutes.^[8] In the lecirelin study, isotachophoresis found counter-ion contents ranging from zero to about 9% trifluoromethanesulfonate, 3% TFA and 11% acetate by mass.^[5]

How counter-ions and water feed into net content is covered in TFA vs acetate counter-ions and water content and Karl Fischer titration.

Record checklist

Record enough that another lab could reproduce the electropherogram and the numbers taken from it.

• Capillary. Inner diameter, total and effective length, bare or coated (coating type), conditioning between runs.
• Electrolyte. Buffer composition, pH, organic modifier and percentage, temperature.
• Run. Voltage and polarity, sample loading mode (pressure or voltage) and amount, detection wavelength, EOF marker if used.
• Results. Migration times, effective mobilities, whether peak areas were corrected for migration time, and replicate spread.
• Cross-check. Which orthogonal method (HPLC, MS) was compared and how the impurity profiles matched.

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