What the nitrogen said

A vial labelled 5 mg contains, in the ordinary case, rather less than 5 mg of peptide. This is not fraud and it is not usually even a discrepancy. A lyophilised peptide is supplied as a salt — most commonly a trifluoroacetate or an acetate — and the counter-ion contributes mass. So does residual water, which a hygroscopic powder acquires readily, and so does residual solvent that lyophilisation did not fully remove. A vial can be 99% pure by chromatography and 82% peptide by mass, and both statements can be true of the same powder at the same time.

The number that resolves this is peptide content: the fraction of the vial’s contents that is actually peptide. It is determined most commonly by elemental nitrogen analysis, sometimes by quantitative amino-acid analysis, occasionally by chromatography against a certified reference standard. All three cost meaningfully more than a purity run, and none of the twenty companies in the Journal’s dossier programme reports content as a standard release test.

Nine vials

Between March and July we submitted nine vials from six suppliers for nitrogen determination. Every vial was labelled with a nominal mass. Every supplier had provided a purity figure above 97%. We asked one question: how much peptide is in there?

The median was 93% of label. Two vials were below 85%. The vial with the highest stated purity in the set — 99.4% — returned 95% content, and the vial with the lowest stated purity returned 98%. There is no relationship between the two numbers in this dataset, and there is no reason there should be: they measure different things.

Why this matters more than purity

A 2% impurity burden in a research peptide is, in most realistic scenarios, a smaller problem than a 16% content shortfall. The impurities in a well-made peptide preparation are overwhelmingly closely related species — deletion sequences, deamidated forms, oxidised methionine — present at fractions of a per cent each. A content shortfall, by contrast, propagates directly into every calculation the user performs.

Work the arithmetic. A vial labelled 5 mg, reconstituted with 2 mL of bacteriostatic water, is assumed to hold 2.5 mg per mL. On a 100-unit insulin syringe, 20 units is 0.2 mL and therefore, on that assumption, 0.5 mg. If the vial actually contains 4.2 mg — the worst case in our set — then the same 20 units delivers 0.42 mg, sixteen per cent less than intended. Nothing about the appearance of the solution, the label, or the purity certificate would reveal it.

What the suppliers said

We put each result to the supplier concerned. Six of the six responded, which is a better rate than we expected. Two accepted the finding without qualification and one of those has since begun reporting content on its certificates. Three argued, reasonably, that content is not a specification they publish and that the nominal mass on the label refers to the mass of material filled rather than the mass of peptide — which is true, is standard practice in the research-chemical trade, and is not what any buyer assumes.

The sixth disputed the method, arguing that nitrogen determination overstates non-peptide nitrogen in preparations containing certain excipients. That is a legitimate objection in general and did not apply to the vial in question, which contained no excipient. We said so; the company did not respond further.

The counter-ion question

Trifluoroacetic acid is used ubiquitously in peptide purification, and TFA salts of basic peptides can carry a substantial counter-ion fraction — for a peptide with several basic residues, ten to twenty per cent of the total mass is not unusual. Acetate salts carry less. Some manufacturers perform a salt exchange specifically to reduce residual TFA, and a certificate that reports the salt form and the counter-ion content is telling you something genuinely useful about the mass in the vial.¹

Two of the twenty suppliers we track report the salt form. None reports counter-ion content as standard. One will report it on request, at additional cost, on a four-week turnaround.

What it costs to know

This is the part of the story that explains the rest of it. A purity run on a generic method costs a private buyer a low three-figure sum. A nitrogen determination costs roughly two and a half to three times as much. Quantitative amino-acid analysis costs more again. A full characterisation package — identity by mass spectrometry, purity by two orthogonal gradients, content, water, counter-ion and residual solvent — was quoted to us by one supplier at approximately 2.4 times the catalogue price of the vial itself.

That figure is the market price of knowing what you have, and the fact that almost nobody pays it is the whole explanation for why the trade competes on purity. Purity is what you get for a hundred euros. Content is what you would need.

The Journal’s position is not that every buyer should commission a full characterisation package. It is that the label mass of a research vial is an upper bound rather than a value, that the median shortfall in our small sample was seven per cent and the worst was nineteen, and that anybody performing arithmetic with a nominal mass should know which of those they are doing.