What Is HPLC Testing? How Peptide Purity Is Measured and Where the Method Falls Short

High-performance liquid chromatography (HPLC) separates a peptide sample into its components and reports purity as the main peak's share of the total UV-detected peak area. It is the standard purity test on a research peptide's certificate of analysis, and for good reason: it is sensitive, quantitative and reproducible. It is also method-dependent and blind to several kinds of impurity, so reading an HPLC result well means knowing how the number was produced.

How reversed-phase HPLC separates peptides

Almost all peptide purity testing uses reversed-phase HPLC (RP-HPLC). The column is packed with silica particles carrying bonded C18 alkyl chains, a hydrophobic surface. The sample is loaded in a mostly aqueous mobile phase, and a gradient of rising acetonitrile then elutes each component in roughly the order of its hydrophobicity. The generic quality-control method for synthetic peptides is a C18 column, an acetonitrile gradient with 0.1% trifluoroacetic acid (TFA) in both mobile phases, and UV detection at about 215 nm ^[1].

Trifluoroacetic acid does two jobs. It keeps the mobile phase near pH 2, protonating basic side chains and the N-terminus, and its anion pairs with those positive charges, sharpening peaks. Typical concentrations are 0.05–0.1% (6.5–13 mM), but a systematic study found that 0.2–0.25% gave better resolution of peptides carrying several positive charges, especially at elevated column temperature ^[2]. Swapping TFA for more hydrophobic perfluorinated acids changes selectivity further, increasingly separating peptides by their number of charges ^[3].

Preparative liquid chromatography is the main way synthetic peptides are purified ^[4]. That has a consequence for analysis: the impurities left in a purified product tend to be the ones that were not resolved from the target during purification, and they can co-elute with it in the QC run too ^[1]. How purification fits into manufacture is covered in how peptides are made.

Reading a chromatogram

A chromatogram plots detector response (usually milli-absorbance units, mAU) against time. Its main features:

• The void or solvent front in the first minutes: unretained material, including salts and very polar species.
• The main peak at the target's retention time. It should be sharp and symmetrical; a shoulder or split top can indicate a co-eluting species.
• Earlier peaks, which are more polar than the target. Truncated or deletion sequences lacking a hydrophobic residue and oxidised methionine variants often appear here.
• Later peaks, which are more hydrophobic. Incompletely deprotected species still carrying a hydrophobic protecting group often appear here.
• The baseline, which drifts upward as acetonitrile rises. A blank run under the same gradient identifies system peaks that are not part of the sample.

Peak area, not height, is what gets integrated, and the integration table, listing every peak with its retention time and area, is the part of a report that can actually be checked.

How the purity percentage is calculated

Purity by area normalisation is:

Purity (%) = area of main peak ÷ sum of all integrated peak areas × 100

The table below is an illustrative calculation with invented areas, not data from any batch:

This sample would be reported as 98.1% pure. Three assumptions sit behind that figure:

1. Equal response. Area percent treats every species as absorbing equally per unit mass. At 215 nm that is a reasonable approximation for closely related peptides, which share most of their peptide bonds, but less so for small fragments or species with very different aromatic content.
2. Everything was detected. Anything that does not absorb at the chosen wavelength is invisible. Quantitative NMR found undeclared mannitol at 20% and 43% w/w in two commercially sourced custom peptides, a UV-transparent constituent that LC methods with UV detection cannot see ^[5].
3. Everything was separated. A species hidden under the main peak is counted as product.

Area percent is a relative purity. Absolute content, the mass of target peptide per mass of powder, needs an assay against a characterised reference standard, quantitative NMR or amino acid analysis; in a USP collaborative study on oxytocin, the HPLC assay against a common bulk standard gave the lowest between-laboratory variability of the three ^[6].

Why the same sample can give different results

HPLC purity belongs to a method, not just to a batch.

• Co-elution. Closely related impurities not resolved during preparative purification can also co-elute in the QC run. A two-dimensional method that adds a second reversed-phase separation at basic pH was developed precisely to expose them, and was demonstrated on therapeutic peptides including exenatide, octreotide and oxytocin ^[1].
• Ion-pairing chemistry. Health Canada analysts found a major impurity in two commercial synthetic parathyroid hormone (1–84) products that TFA-based methods could not resolve; a pentanesulfonate ion-pairing system at pH 5.6 separated it ^[7].
• Isomers. Racemisation during synthesis creates D-amino acid variants. Because D and L forms differ only in stereochemical configuration, separating and quantifying them is analytically demanding ^[8], and most generic purity methods are not designed to do it.
• Integration settings. Peak-detection thresholds decide whether small peaks are counted at all; a high threshold silently raises the reported purity.

For a worked example of which impurities each technique would catch in one sequence, the BPC-157 monograph tabulates the expected masses of its likely deletion, aspartimide and isoaspartate variants.

HPLC with mass spectrometry

Coupling the separation to a mass spectrometer identifies what each peak is. There is a trade-off: TFA, the best additive for UV peak shape, causes significant signal suppression in electrospray ionisation, so LC–MS methods often use other additives; one group showed that adding 2 mM glycine to TFA mobile phases raised peptide MS signal-to-noise by more than tenfold on average without degrading the separation ^[9]. A different additive means different selectivity, so peak patterns from an LC–MS run and a TFA–UV purity run do not always map one to one.

LC–high-resolution MS can characterise impurities in depth. For the antimicrobial peptide Cbf-14, one process-related impurity and 32 degradation products were identified, including hydrolysis fragments, isomers and oxidised forms, none of which had been declared in custom-synthesised material ^[10]. A COA rarely goes that far; how to read the MS identity section it does contain is covered in understanding certificates of analysis.

For the bench

• Read the method, not only the number. Column, gradient, additive and wavelength should be on the report. A figure without them cannot be compared with anything.
• Ask for the chromatogram and integration table. These are the verifiable parts of a certificate of analysis.
• Treat 1% as method noise. A 98.6% and a 99.2% result from different laboratories may describe the same material; peptide purity explained covers when the last percent matters.
• Use HPLC for stability checks. Running a stock solution before and after storage, or assay medium before and after incubation, shows degradation directly. Peptides with methionine, such as Semax, or labile Asp and Asn motifs warrant this more than robust sequences; deamidation and oxidation products typically appear as new peaks near the main one.
• Keep the counter-ion in mind. Purification in TFA means many peptides are supplied as trifluoroacetate salts, which do not affect area percent but do affect weighed mass; see TFA vs acetate peptide salts and the molarity calculator.