Key points
- Reversed-phase HPLC separates peptides by hydrophobicity: a non-polar stationary phase, a polar mobile phase, and a rising gradient of organic solvent that releases peptides in order of increasing hydrophobicity.
- The standard system is a C18 column with water and acetonitrile, both containing 0.1% trifluoroacetic acid, monitored at 214–220 nm where the peptide bond itself absorbs.
- Peptides elute over a surprisingly narrow window of organic solvent, which is why shallow gradients are used and why small changes in gradient slope change resolution so much.
- The hydrophobicity that governs retention is the same property the hydropathy plot describes, so GRAVY is a useful, rough predictor of elution order.
How reversed-phase separation works
In normal-phase chromatography the stationary phase is polar and the mobile phase is not. Reversed-phase inverts this: the stationary phase consists of silica particles whose surface has been derivatised with long alkyl chains, most often C18, and the mobile phase is water mixed with an organic solvent.
A peptide injected in mostly water adsorbs to the alkyl surface through its hydrophobic residues. As the proportion of organic solvent rises, the mobile phase becomes progressively better at solvating those residues, until at some point the peptide desorbs and travels to the detector. More hydrophobic peptides need more organic solvent and therefore elute later.
Peptide retention behaves differently from small-molecule retention in one important respect. Because a peptide binds through several residues at once, the transition from fully adsorbed to fully desorbed happens over a very narrow range of solvent composition, often less than 1% acetonitrile. Isocratic elution is therefore impractical: a peptide either sticks or runs through with the solvent front. Gradient elution is effectively mandatory, and gradients are kept shallow, typically 0.5–1% acetonitrile per minute, so that peptides separate rather than co-eluting.
Standard conditions
| Parameter | Typical choice | Notes |
|---|---|---|
| Column | C18, 4.6 × 150–250 mm analytical | C8 or C4 for very hydrophobic or larger peptides |
| Particle size | 3–5 µm (sub-2 µm for UHPLC) | Smaller particles give sharper peaks at higher back-pressure |
| Pore size | 300 Å for peptides above ~3 kDa; 100–120 Å for smaller | Molecules must enter the pores to see the surface |
| Mobile phase A | Water + 0.1% TFA | |
| Mobile phase B | Acetonitrile + 0.085–0.1% TFA | Slightly less TFA in B balances the baseline drift |
| Gradient | 5→65% B over 30–60 min | Shallower near the peptide's elution point for difficult separations |
| Detection | 214–220 nm | 280 nm only for peptides containing Trp or Tyr |
| Temperature | 25–40 °C, controlled | Higher temperature sharpens peaks and lowers back-pressure |
Detection at 214–220 nm works because the amide bond absorbs there, as explained in the guide to the peptide bond. It sees every peptide regardless of composition, which makes it the right choice for purity assessment. Acetonitrile is preferred over methanol partly because it has much lower absorbance in this region.
The role of TFA and ion pairing
Trifluoroacetic acid does two things. It lowers the pH to about 2, which protonates all carboxyl groups and suppresses interactions between negatively charged peptides and residual silanol groups on the silica surface. And its trifluoroacetate anion pairs with the positively charged groups of the peptide, masking their charge and making the peptide behave as a more hydrophobic molecule. The result is sharper peaks and better retention of basic peptides.
The choice of ion-pairing agent changes selectivity, which is a useful tool when two peptides will not separate:
| Additive | Effect | MS compatible |
|---|---|---|
| 0.1% TFA | Strong ion pairing, sharpest peaks, longest retention of basic peptides | Poorly; suppresses ESI signal |
| 0.1% formic acid | Weaker ion pairing, slightly broader peaks | Yes; the standard for LC-MS |
| 10 mM ammonium acetate or bicarbonate, pH 7–10 | Different selectivity entirely; needs a pH-stable column | Yes |
| 0.1% heptafluorobutyric acid | Stronger ion pairing than TFA; retains very hydrophilic peptides | No |
TFA is the reason purified synthetic peptides are usually obtained as TFA salts, which affects how much peptide a weighed sample actually contains. The consequences for concentration calculations are worked through in the molecular weight guide.
Predicting elution order
Retention correlates with the summed hydrophobicity of the residues, which is why GRAVY, the average Kyte–Doolittle value over the sequence, is a serviceable first guess. Peptides rich in Trp, Phe, Leu, Ile and Val elute late; those rich in Asp, Glu, Lys, Arg, Ser and Gly elute early.
Dedicated retention prediction models do better because they account for effects that a simple average misses:
- Position. Residues near the termini contribute less than those in the middle.
- Length. Longer peptides retain more strongly, but the effect saturates because only one face of the peptide contacts the surface.
- Conformation. A peptide that forms an amphipathic helix presents a continuous hydrophobic face to the column and retains far more strongly than its composition alone suggests. This is the same amphipathicity that a hydropathy plot cannot see.
In practice, elution order is used more often as a diagnostic than as a prediction. A single modification frequently shifts retention in a characteristic direction: oxidised methionine elutes earlier than the parent peptide, deletion sequences elute close to it, and a peptide with its disulfides formed usually elutes differently from the reduced form.
From analysis to purification
Analytical and preparative runs have different goals. The analytical run measures purity; the preparative run recovers material. Moving between them follows a straightforward scaling rule: keep the linear velocity, the gradient slope in column volumes, and the stationary phase the same, and scale the flow rate and the load with the square of the column diameter.
Typical preparative practice:
- Load. Roughly 1–10 mg of peptide per gram of packing for difficult separations, more when the impurities are well resolved. Overloading broadens and distorts peaks.
- Injection solvent. Dissolve the crude peptide in as little organic solvent as possible; injecting in a strong solvent causes the band to spread at the head of the column.
- Fraction collection. Collect narrow fractions across the main peak and analyse each one, rather than pooling on the basis of the UV trace alone. The shoulder of a peak is where the closest impurities sit.
- Recovery. Pooled fractions are lyophilised. Hydrophobic peptides can be lost to tubing and vial walls during this step, as described in the stability guide.
Two-step purification is common for difficult peptides: a first pass under TFA conditions, then a second at a different pH or with a different additive, so that the two steps separate on genuinely different bases.
What a purity figure means
When a supplier states 95% purity, the figure almost always means the percentage of the total area under the UV trace at 214–220 nm that belongs to the main peak. This is a reasonable measure, but it has limits worth knowing:
- It measures peptide bonds, not molecules. A short impurity contributes less area per mole than a long one, so the molar purity and the area purity are not the same number.
- Co-eluting impurities are invisible. Deletion sequences and isomers such as isoaspartate often sit under the main peak. A second method, normally mass spectrometry, is needed to see them.
- Non-peptide content is not counted. Water and counterions can account for 20–40% of the mass of the powder and do not appear in the chromatogram at all. Purity and net peptide content are different quantities.
A complete characterisation therefore pairs an HPLC trace with a mass spectrum, and for quantitative work with an independent concentration measurement such as amino acid analysis or UV absorbance at 280 nm.
Common problems
| Symptom | Likely cause | What to try |
|---|---|---|
| Peak tailing | Silanol interactions; column overload | Check that TFA is present and fresh; reduce the load |
| Broad or split peaks | Aggregation; proline cis/trans isomerism; injection in too strong a solvent | Raise the temperature; dilute the injection; change the injection solvent |
| No peak at all | Peptide too hydrophobic and still bound; adsorbed to the system; too hydrophilic and in the void volume | Wash with a high-organic step; check the void volume; try a C4 column |
| Retention time drifts between runs | Temperature variation; incomplete equilibration; evaporating acetonitrile | Control the column temperature; equilibrate for several column volumes; prepare fresh mobile phase |
| Rising baseline during the gradient | TFA absorbance differs between A and B | Use slightly less TFA in the organic phase |
| New peak appears in stored samples | Oxidation, deamidation or disulfide formation | Check the mass difference against the modification tables |
Other separation modes
Reversed-phase dominates, but it is not always the right choice:
- Ion-exchange chromatography separates by charge and is the natural complement. The working pH is chosen relative to the peptide's isoelectric point: below the pI the peptide is positive and binds a cation exchanger, above it an anion exchanger.
- Size-exclusion chromatography separates by hydrodynamic size and is useful for detecting aggregates and disulfide-linked dimers, which reversed-phase may not resolve.
- HILIC retains very polar peptides that reversed-phase cannot hold, and is widely used for glycopeptides.
For characterisation, combining orthogonal methods is far more informative than running the same separation twice.
Frequently asked questions
Why 214 nm rather than 280 nm?
Because 214 nm detects the peptide bond, which every peptide has, while 280 nm detects aromatic side chains, which many peptides lack. A peptide with no Trp or Tyr is essentially invisible at 280 nm.
Should I use C18, C8 or C4?
C18 is the default for peptides up to a few kDa. Shorter chains retain less strongly, so C8 and C4 are used for larger or very hydrophobic peptides that would otherwise need extreme organic concentrations to elute.
Can I use the same method for LC-MS?
Not directly. TFA suppresses electrospray ionisation badly. Switch to 0.1% formic acid, accepting somewhat broader peaks and shifted retention times, or use a very low TFA concentration with a post-column additive.
How do I remove TFA from the purified peptide?
By salt exchange: repeated lyophilisation from dilute hydrochloric acid gives the chloride salt, or from dilute acetic acid or ammonium acetate the acetate salt. Ion-exchange cartridges are also used. Complete removal is difficult, and a residual level is normal.
References
- Mant CT, Chen Y, Yan Z, et al. (2007) HPLC analysis and purification of peptides. Methods in Molecular Biology 386:3–55.
- Krokhin OV (2006) Sequence-specific retention calculator: algorithm for peptide retention prediction in ion-pair RP-HPLC. Analytical Chemistry 78:7785–7795.
- Snyder LR, Kirkland JJ, Dolan JW (2010) Introduction to Modern Liquid Chromatography, 3rd edition. Wiley.