Characterizing a peptide: HPLC, LC-MS, amino acid analysis, and NMR

A finished peptide is a solid powder in a vial. The label says it is a particular sequence at a particular mass. To know whether the label is true, four independent questions have to be answered:

1. Identity — is this molecule actually the declared sequence?
2. Purity — what fraction of the peptide species in the vial is the target, versus deletions, truncations, oxidations, and dimers?
3. Content (assay) — how many milligrams of active peptide are in the vial, separate from water, salt, and counter-ion?
4. Structural integrity — for sequences with disulfides, cyclic constraints, or modification-prone residues, is the structure actually correct?

These are four different measurements. A "99% pure" label that does not specify content is incomplete; a content number without an identity confirmation is meaningless.

Analytical reverse-phase HPLC is the workhorse for purity. A small aliquot of the dissolved peptide is injected onto a C18 column with a controlled gradient (typically 5% to 60% acetonitrile in 0.1% TFA over 20 to 40 minutes) and detected by UV absorbance at 214 nm, where the peptide bond absorbs strongly.

The chromatogram is integrated and the area of the main peak is divided by the total integrated area, giving HPLC purity as a percent. Modern systems can quantify down to ~0.05% impurity (the limit of quantification, LOQ) with ~0.01% reporting precision.

What HPLC purity does not tell you:

• Anything about UV-invisible impurities — residual solvents, water, salts, counter-ion.
• Anything about co-eluting impurities (single-residue deletions of Gly/Ala, D-isomers).
• Anything about absolute content — a 99% pure peptide can still be a fraction of the labeled mass once water and counter-ion are subtracted.

For these reasons, HPLC purity is necessary but not sufficient.

LC-MS pairs an HPLC separation with a mass spectrometer that measures the mass-to-charge ratio of every eluting species. For peptides, electrospray ionization (ESI) produces a series of multiply charged ions: a 2,000 Da peptide commonly appears as M+2H²⁺ at m/z 1001, M+3H³⁺ at m/z 668, and so on. Software deconvolutes that ladder back into the underlying neutral molecular mass.

The deconvoluted mass is compared to the theoretical monoisotopic mass calculated from the declared amino acid sequence. A match to within ±0.1 Da on a high-resolution instrument is the standard identity confirmation. Modern Orbitrap and Q-TOF instruments routinely deliver mass accuracy below 5 ppm.

LC-MS also identifies impurities. The deletion of a glycine residue produces a satellite peak 57.02 Da below the target; oxidation of methionine produces a +16 Da satellite; aspartimide produces a −18 Da satellite. The mass-difference fingerprint is often more diagnostic than the chromatographic separation itself.

The dry mass of a peptide vial is not the same as the active peptide mass. Lyophilized peptide powder typically contains 5% to 15% water, plus counter-ion (TFA, acetate, HCl) at 5% to 20% of total mass depending on the composition. To determine what is actually in the vial, you have to measure the peptide independently of dry weight.

Amino acid analysis (AAA) does that. The peptide is hydrolyzed in 6 N HCl at 110°C for 24 hours, breaking every amide bond and releasing free amino acids. The hydrolysate is then quantified against calibrated amino acid standards, typically by ion-exchange chromatography with ninhydrin detection or by reverse-phase HPLC with pre-column derivatization (OPA, AQC, or PITC).

The measured molar amounts of each amino acid are compared to the expected stoichiometry from the declared sequence. The result is the true peptide content of the vial — the gold standard for assigning value to a peptide reference standard, and what the European Pharmacopoeia and USP accept for active content of peptide actives.

For routine release testing of a well-characterized peptide, HPLC + LC-MS + content (by AAA or quantitative HPLC) is sufficient. For investigational or reference-standard work, additional methods come in:

• 1D and 2D NMR are used to confirm cyclic-disulfide topology, distinguish iso-aspartate from aspartate, detect aspartimide intermediates, and verify residue stereochemistry. NMR requires milligram quantities and is slow but uniquely structural.
• Circular dichroism (CD) reports on secondary structure (alpha-helix, beta-sheet, random coil) and is used for stability comparison and conformational fingerprinting.
• Edman degradation or LC-MS/MS sequencing reads the actual amino acid sequence one residue at a time and is the orthogonal confirmation that the declared sequence is what was actually synthesized — particularly important when an LC-MS mass match is consistent with multiple permuted sequences.
• Endotoxin testing (LAL) is a bacterial-contamination assay required for any peptide intended for parenteral use. It is not a structural method but it is part of any complete release package for an injectable.
• Size-exclusion HPLC (SEC-HPLC) and host-cell protein (HCP) ELISA are mandatory once you cross from synthetic peptides into recombinant proteins like somatropin — SEC resolves covalent dimers and non-covalent aggregates that RP-HPLC dissociates, and HCP ELISA quantifies residual host-cell antigen the UV detector cannot see. See the HGH testing methodology guide for the full recombinant-protein assay set.

A complete Certificate of Analysis for a research- or pharmaceutical-grade peptide pulls all of these threads into one document. At minimum it should report:

• HPLC purity (% area at 214 nm), with the chromatogram available on request.
• LC-MS identity (observed monoisotopic mass vs. theoretical, with the deconvoluted spectrum).
• Quantitative content (mg/vial of active peptide, by AAA or by HPLC against a calibrated reference).
• Counter-ion identity and content (% TFA or acetate by ion chromatography).
• Water content (Karl Fischer titration).
• Endotoxin level (EU/mg) for any injectable-grade material.
• Visual appearance, solubility, and storage conditions.

A vendor COA without all of these fields is not necessarily wrong — but it is incomplete, and the gaps are exactly where independent verification adds value.