A certificate that reports 99.2% purity is also reporting 0.8% of something else, and that remainder is rarely a single substance. For a long synthetic peptide such as LY3437943 or LY3298176, both 39 residues with non-coded amino acids and a fatty diacid attached to a lysine, the remainder is a population: sequences one residue short, sequences still carrying a protecting group, the right sequence with one residue in the wrong hand, and the right sequence a few weeks older than it should be.
This guide names each class, says where it comes from, how it shows up on reversed-phase HPLC and on mass spectrometry, and whether a routine certificate detects it. The regulatory vocabulary comes from ICH Q6B, which separates product-related impurities (variants of the molecule itself) from process-related ones (reagents, solvents, resin leachables), and from ICH Q3A, which sets the thresholds at which an impurity has to be reported, identified and qualified.
Two families: made during synthesis, made afterwards
Synthesis-related impurities are fixed at the moment the peptide is cleaved from the resin and do not grow in the vial. Degradation-related impurities are present at a low level on the day of release and increase with time, temperature and water. The distinction matters because a certificate dated at release captures the first family completely and the second only as it stood that day.
D'Hondt and colleagues (2014) reviewed the related impurities of peptide medicines in J Pharm Biomed Anal and organised them along exactly this line; Vergote and colleagues (2009) set out how a specification for a peptide should handle each. Solid-phase peptide synthesis explained covers the chemistry that produces the first family, and peptide degradation pathways the second.
Deletion and truncation sequences
In stepwise solid-phase synthesis each amino acid is coupled to the growing chain in turn. A coupling that does not reach completion leaves some chains without that residue; if the next coupling then succeeds on those chains, the result is a deletion sequence, the target minus one internal residue. A chain that stops growing altogether, because its N-terminus was capped or became inaccessible, is a truncation: the C-terminal part of the sequence only. For a 39-residue molecule there are 38 couplings at which a deletion can occur, which is why length is the single largest driver of impurity burden.
On HPLC a deletion sequence usually elutes close to the main peak, because losing one residue from 39 changes hydrophobicity slightly; the most hydrophobic deletions can co-elute with the target. On mass spectrometry a deletion is unambiguous: the mass is the target minus the residue mass of the missing amino acid, and the deficit identifies which residue is missing where two have different masses. Truncations are shorter, elute earlier and are easy to resolve.
Incomplete deprotection
Side chains are protected during synthesis with groups such as tert-butyl, trityl and Pbf, which are removed in the final acid cleavage. A protecting group that survives cleavage leaves a peptide heavier than the target by the mass of that group. These are often more hydrophobic than the target and elute later on reversed-phase HPLC, and their masses are diagnostic: the increment matches a known protecting group. A related impurity arises when a scavenger or cleavage by-product attaches to a nucleophilic side chain such as tryptophan or methionine.
Racemisation: the right sequence in the wrong hand
Amino acids are L-configured, and the activation step that makes a coupling possible also makes the activated residue vulnerable to losing that configuration. The result is a diastereomer: the correct sequence with one D-residue. Cysteine, histidine and phenylalanine are the most susceptible, and Han, Albericio and Barany (1997) documented the extent of cysteine racemisation in stepwise solid-phase synthesis and the coupling conditions that minimise it.
A diastereomer has the identical mass, so mass spectrometry cannot see it. Reversed-phase HPLC sometimes resolves it as a shoulder on the main peak and sometimes does not. The method that sees it reliably is chiral amino acid analysis after hydrolysis, and almost no routine certificate includes that test.
Aspartimide, isomerisation and deamidation
Aspartate and asparagine share a route. The backbone nitrogen following the residue attacks the side-chain carbonyl to form a five-membered succinimide ring (an aspartimide). The ring then opens either way, giving a mixture of the original aspartate and iso-aspartate, in which the backbone now runs through the side chain. From asparagine the same ring forms with loss of ammonia, so the product is deamidated: aspartate or iso-aspartate, one dalton heavier than the parent.
During synthesis the base used for Fmoc removal drives aspartimide formation, and Mergler and colleagues (2003) described the problem and its dependence on the sequence around the aspartate. In solution the same chemistry runs slowly at neutral pH, which is why deamidation is the first thing to grow in a stored solution. Yang and Zubarev (2010) reviewed how mass spectrometry handles these products: deamidation is a one-dalton shift that high-resolution instruments resolve from the isotope pattern, while the aspartate to iso-aspartate isomer has the same mass and needs fragmentation or chromatography to detect. On HPLC the iso-aspartate form typically elutes slightly earlier than the parent and the aspartimide, 18 daltons lighter, slightly later.
Oxidation of methionine and tryptophan
Methionine oxidises to the sulfoxide (plus 16 daltons) and then the sulfone (plus 32); tryptophan oxidises to a family of products including kynurenine and hydroxytryptophan. Both routes need only dissolved oxygen, light or trace metals to proceed. Ji and colleagues (2009) mapped methionine, tryptophan and histidine oxidation in parathyroid hormone as a model and showed how each is detected and slowed. Methionine sulfoxide elutes earlier on reversed-phase HPLC and is one of the few impurities a routine purity method identifies well; tryptophan products are more varied and more often missed.
Aggregation, and two things that are not impurities
Aggregates are the target molecule associated with itself, non-covalently or through disulfide exchange. They are not seen by a reversed-phase purity method, which either dissociates them in the organic mobile phase or retains them on the column, so a certificate can report high purity for material that is partly aggregated. Size-exclusion chromatography or light scattering is needed. Zapadka and colleagues (2017) reviewed the physical factors that drive peptide aggregation, and long acylated sequences are among the most susceptible.
Two components of the powder are not impurities at all. A lyophilised peptide is a salt. The trifluoroacetate or acetate counterion and the residual water are real mass in the vial but are not related substances: they are not variants of the peptide, they do not appear on a chromatogram as peptide peaks, and purity by HPLC ignores them entirely. They are the reason a vial's net peptide content is lower than its gross mass, which is a separate line on a certificate from purity and is covered in what a certificate does not test for.
Each class at a glance
| Class | Origin | On reversed-phase HPLC | On mass spectrometry | Routine certificate sees it? |
|---|---|---|---|---|
| Deletion sequence | Incomplete coupling in synthesis | Near the main peak; can co-elute | Target minus one residue mass | Usually, unless co-eluting |
| Truncation | Chain stopped growing | Earlier-eluting | Mass of the C-terminal fragment | Yes |
| Incomplete deprotection | Protecting group survived cleavage | Later-eluting | Target plus the group's mass | Yes |
| Racemisation (D-residue) | Activation during coupling | Shoulder or hidden | Identical mass | Rarely; needs chiral analysis |
| Aspartimide / iso-Asp | Base during synthesis; neutral pH in solution | Adjacent peaks | Minus 18 (aspartimide); identical (iso-Asp) | Partly |
| Deamidation (Asn, Gln) | Hydrolysis in solution and slowly in the solid | Slightly earlier | Plus 1 dalton | Partly; needs resolution |
| Oxidation (Met, Trp) | Oxygen, light, trace metals | Earlier (Met sulfoxide) | Plus 16, plus 32 | Yes for Met; Trp often missed |
| Aggregation | Self-association, disulfide exchange | Not seen | Not seen as intact mass | No; needs SEC |
| Counterion and water | Purification and lyophilisation | No peptide peak | Not applicable | Not an impurity; separate content line |
Origin, appearance and routine detectability by class. HPLC and MS behaviour summarised from D'Hondt and colleagues (2014) and the specific papers cited in the text; 'routine certificate' means a reversed-phase HPLC purity method with a single intact-mass identity test.
Why the two catalogue examples carry more of this
LY3437943 and LY3298176 are 39 residues long, include the non-coded residue Aib, and carry a C20 fatty diacid on a lysine side chain through a linker, according to the catalogue records. Each of those features adds impurity classes: more couplings mean more deletion opportunities; non-coded residues couple more slowly and leave more incomplete chains; and the lipid attachment is a further synthetic step that can fail (unlipidated peptide), double up, or land on the wrong lysine, producing a positional isomer with the identical mass that an intact-mass identity test cannot distinguish from the target.
That is why HPLC purity for these two is only informative alongside the method conditions, and why an identity test that fragments the molecule earns its place on their certificates. The figure on the catalogue record, ≥99% by HPLC, is a specification; the certificate for a lot is where the actual number and the actual impurity profile live.