Ask what is in a peptide preparation besides the peptide and the intuitive answer is "contamination", something foreign that got in. For synthetic peptide the more important answer is stranger: the main impurity is usually the target itself, slightly wrong.
Where they come from
Solid-phase synthesis builds a chain one residue at a time, and each coupling step succeeds with high but not perfect efficiency. When a step fails on a particular chain, that chain simply carries on without the residue, and is extended normally from there.
The arithmetic is unforgiving and is set out in how a peptide is actually made: at 99% efficiency per step, a thirty-step synthesis leaves only about 74% of chains complete. The missing quarter is not dirt. It is a population of chains that are the target minus one residue, in various positions.
Why they are the expensive part
An impurity that differs substantially from the target is straightforward to remove. It behaves differently in a purification column and separates cleanly.
A deletion sequence does not.
Contrast with an impurity that is obviously different
An unrelated impurity
- Different size, charge and hydrophobicity
- Elutes at a clearly different time, easy to separate
- Biologically inert in the assay; dilutes but does not confuse
A deletion sequence
- Differs by one residue out of fifteen or more
- Elutes very close to the target; may not resolve at all[1]
- May retain partial binding, behaves like a weak version of the compound
Missing one residue out of fifteen changes hydrophobicity by a small increment. On a reversed-phase column the two species elute at nearly the same time, and on a shallow gradient they may not resolve into separate peaks at all1. Removing them is the costly step in manufacturing a peptide, which is a large part of why purity above 98% commands the price it does.
The part that affects results
Here is where this stops being a manufacturing curiosity and becomes an experimental problem.
A truly inert impurity dilutes the sample. If 10% of the mass is inert, you have 10% less compound, and a correctly labelled peptide content figure would let you correct for it.
A deletion sequence may not be inert. It carries most of the target's structure, and depending on which residue is missing it can retain some fraction of binding. It is, functionally, a weak analogue of the compound mixed into the compound.
That does not shift a dose, response curve cleanly; it distorts it. The maximal effect may be depressed, the apparent potency shifted, and the result reproducible enough to look real. Nothing in the experiment announces that two related species are present.
Why one test cannot cover it
One can be fooled by co-elution; the other cannot
Sample contains target plus a deletion chain
Say 8% of chains are missing one glycine
Chromatography, shallow gradient
The two species may merge into a single peak, reported as one clean component
Mass spectrometry
A satellite peak appears exactly 57 Da below the target, glycine's residue mass
Together they cover each other
Which is the entire argument for requiring both on a certificate
Chromatography answers "how much of what the detector saw was the main peak", and if two species co-elute, the integrator counts them as one. The number is not dishonest; it is answering the question it was asked.
Mass spectrometry answers a different question, and does so in a way co-elution cannot defeat. A missing residue changes the molecular mass by that residue's exact mass, 57 Da for glycine, 71 for alanine, 128 for lysine. The logic is developed in confirming identity by mass, and it is why a satellite peak at a specific offset is such a clear signal.
This is the whole argument for requiring both analyses. They fail in different directions, and together they cover each other.
What to look for on a certificate
Three things make the difference between a purity figure you can interpret and one you cannot.
The method line. A percentage without the chromatographic conditions cannot be reproduced or compared. A shallow gradient and a steep one do not have equal power to resolve near-neighbours.
The mass result, with tolerance. Expected versus found, stated explicitly.
The chromatogram itself, where it is provided. A trace with a clean baseline and one symmetrical peak communicates more than a number does. A shoulder on the main peak is exactly what a partially resolved deletion sequence looks like.
And the broader habit, discussed in reading a certificate of analysis: what a document does not report was not measured. Degradation during storage produces related species too, oxidised and deamidated forms among them2, and a certificate describes the material on the day it was tested, not the day it reached you.
