A certificate of analysis can report 99.4% purity and an observed mass matching theory to two decimal places, and still tell you nothing about whether the vial will quietly distort your next experiment. Purity and identity are chemical questions. Endotoxin is a biological one — and the instruments that answer the first two are, for practical purposes, blind to it.
Most discussion of research-peptide quality runs along two axes: is it the right molecule (mass spec), and how much of the sample is that molecule (HPLC). There is a third axis that rarely makes it onto the paperwork at all, and it is the one that most often explains an experiment that refuses to reproduce.
Why purity assays cannot see endotoxin
The reason is not that endotoxin is exotic. It is a sensitivity mismatch.
Reversed-phase HPLC with UV detection resolves impurities down to roughly a tenth of a percent of sample mass. Below that, an impurity simply lives in the baseline. Mass spectrometry, in standard peptide workflows, is looking for a target mass — and endotoxin is not one molecule but a heterogeneous family of lipopolysaccharides that does not present as a clean confirmatory peak.
Endotoxin, meanwhile, is biologically active at picogram-per-millilitre concentrations. That is orders of magnitude below the floor at which a purity assay resolves anything at all.
So a peptide can be genuinely 99% pure by HPLC, genuinely mass-confirmed, and still carry an endotoxin load that is pharmacologically meaningless by mass but biologically loud. Purity and endotoxin are independent variables. A high number on one says nothing about the other. This is the same lesson as HPLC versus mass spec — different questions need different instruments — extended one axis further.
What endotoxin actually is
Endotoxin is lipopolysaccharide (LPS), a structural component of the outer membrane of Gram-negative bacteria. It is not a secreted toxin; it is part of the bacterial envelope, shed during growth and released in quantity when the cell dies.
Structurally it has three regions: a variable O-antigen polysaccharide, a core oligosaccharide, and lipid A — the membrane anchor that carries essentially all of the biological activity. Lipid A is recognised in mammals by the TLR4/MD-2 receptor complex, which triggers innate immune signalling and inflammatory cytokine release.
Two properties make it a persistent contaminant. It is heat-stable, surviving standard autoclaving (121 °C) that reliably kills the bacteria that produced it. And it is small enough to pass through 0.22 µm sterilising filters that retain intact cells.
It is also ubiquitous. Water systems, glassware, raw materials, packaging components, and handling are all documented entry points — the contaminant does not require a contamination event in the dramatic sense.
Sterile is not the same as endotoxin-free
This is the single most consequential distinction in the topic, and three separate tests sit behind it:
- Sterility testing asks whether viable organisms are present.
- Bioburden testing counts the viable organisms present.
- Endotoxin testing measures a molecule left behind by organisms that may be long dead.
Because LPS is heat-stable and filter-permeable, a preparation can pass sterility, report zero bioburden, and still carry a substantial endotoxin load. Killing the bacteria does not remove the endotoxin. Removing it is a separate problem: depyrogenation of glassware requires dry heat far above autoclave temperatures (on the order of 250 °C), and removing LPS from a solution requires dedicated chromatographic or affinity approaches.
A vendor saying "sterile filtered" is answering a different question from the one endotoxin testing answers. Both can be true; neither substitutes for the other.
How endotoxin is measured: the LAL test
The standard assay exploits an accident of comparative immunology. In the 1960s, Frederik Bang and Jack Levin showed that the blood of the horseshoe crab (Limulus polyphemus) clots on contact with Gram-negative bacteria. The clotting is driven by an enzyme cascade in the animal's amebocytes, and that cascade is exquisitely sensitive to lipid A.
Isolated, this became Limulus Amebocyte Lysate (LAL), codified in the United States as USP General Chapter <85>, Bacterial Endotoxins Test, in three formats:
- Gel-clot — the reaction either forms a solid clot or it does not, read against a defined sensitivity. Qualitative to semi-quantitative, and under
<85>it remains the referee method when results conflict. - Turbidimetric — instruments track the cloudiness developing as the cascade proceeds; the rate is proportional to endotoxin concentration.
- Chromogenic — the cascade cleaves a synthetic substrate that releases a coloured product, with intensity read quantitatively.
Results are reported in endotoxin units (EU). It is worth being precise about what an EU is: it is a unit of biological activity, defined against a Reference Standard Endotoxin — not a unit of mass. Conversions to nanograms circulate widely, but the factor depends on the LPS preparation and the standard used, which is exactly why the pharmacopeia standardises the activity rather than the weight.
One complication matters specifically for peptides. Some compounds interfere with the cascade, enhancing or inhibiting it and skewing the reading. This is why a valid test includes a positive product control — a spike of known endotoxin into the sample — to demonstrate the result is trustworthy in that particular matrix. The interference is not hypothetical for this catalog: LL-37 is a cathelicidin whose established biology includes binding and neutralising LPS, which is precisely the sort of activity that confounds an LAL reading.
The recombinant shift: rFC and USP <86>
The LAL supply chain depends on bleeding wild horseshoe crabs, which has drawn both conservation pressure and supply-resilience concern for years. The technical answer was to clone the cascade.
Recombinant Factor C (rFC) reproduces the first enzyme in the pathway; recombinant cascade reagents (rCR) reconstruct more of it. Neither requires an animal.
The regulatory arc is recent and worth stating precisely. The European Pharmacopoeia moved first, adopting a recombinant Factor C chapter (2.6.32) in 2021. In the United States, rFC sat for years as an alternative method — usable, but placing the validation burden on whoever wanted to use it. That changed when USP published General Chapter <86>, Bacterial Endotoxins Test Using Recombinant Reagents, released for early adoption in November 2024 with an official date of 1 May 2025.
<86> gives recombinant reagents their own compendial standing alongside <85> rather than relegating them to alternative-method status. It permits rather than mandates: LAL remains entirely valid, and method suitability still has to be demonstrated. But as of 2025, "we use a recombinant method" is a compendial position in the US, not a deviation requiring justification.
For anyone reading testing documentation, the practical upshot is that a modern endotoxin result may cite either chapter, and neither is a red flag.
Where the synthesis route changes the risk
Not every research compound carries the same endotoxin exposure, and the reason is manufacturing route.
Most peptides in a research catalog are made by solid-phase peptide synthesis — built chemically, residue by residue, on a resin. There is no bacterium anywhere in the process, so endotoxin can only enter through water, reagents, glassware, packaging, or handling. Real risk, but incidental risk.
Recombinantly expressed products are a different situation. When a sequence is expressed in E. coli, the production organism is a Gram-negative bacterium — the endotoxin source is the manufacturing platform itself, and its removal is an active purification objective rather than a matter of housekeeping. HGH 191AA, a recombinant somatropin, is the clearest example on this shelf of a product whose expression system makes endotoxin control central rather than peripheral.
This is not an argument that one route is cleaner than the other. It is an argument that the question is more load-bearing for some products than others, and that knowing how a compound was made tells you how hard to look.
How it appears on a COA — and what silence means
Where endotoxin is tested, it is normally reported as EU/mg (units of activity per milligram of peptide), with the method and often the chapter cited.
Two things are worth understanding about that number.
First, it is completely independent of the purity figure. It is not derived from the chromatogram, it does not trade off against net peptide content, and it cannot be inferred from either. It is a separate assay producing a separate result, which is why it belongs in the full COA picture rather than as a footnote to purity.
Second, the thresholds that matter are application-dependent, and the research literature has been getting more specific. A 2025 paper in mAbs on establishing endotoxin limits for in-vitro immunogenicity assessment reported that levels above 0.1 EU/mg were sufficient to drive T-cell proliferation and CD14+ myeloid expansion — that is, to generate immune signals that look like a result but are the contaminant talking. Work with immune or primary cells is unusually exposed to this; other applications are less so.
And when the line is simply absent — which, on research-peptide documentation, is the common case — the honest reading is "not tested," not "not present." Endotoxin testing is an additional assay with additional cost. Its absence from a COA is a gap in the record, not evidence of a clean result. That distinction is the whole of the quality question: a certificate documents what was measured, and everything unmeasured is simply unknown.
FAQ
Does a high HPLC purity number imply low endotoxin? No. They are independent measurements of different things. Endotoxin is bioactive far below HPLC's detection floor, so a purity figure carries no information about it.
Does sterile filtration remove endotoxin? No. LPS passes through 0.22 µm sterilising filters and survives autoclaving. Sterility and endotoxin are separate specifications.
Is LAL being phased out?
Not phased out. USP <86> (official 1 May 2025) gave recombinant reagents compendial standing alongside the LAL methods of <85>. Both are valid; <86> permits recombinant methods rather than requiring them.
Why do some peptides give unreliable LAL results? Certain compounds interfere with the enzymatic cascade, either enhancing or inhibiting it. Peptides with LPS-binding activity are an obvious case. A positive product control is what demonstrates the assay works in that specific sample.
Endotoxin sits alongside stability and counterion identity in a category of quality attributes that the headline purity number was never designed to capture. Reading documentation well means knowing which questions a given assay answers — and which it never asked. More compound-level detail is in the research library.
This article is educational and for the laboratory research community. Trulogic Labs products are sold for laboratory and research use only and are not for human consumption.