Bacterial endotoxin is a heat-stable component of the outer membrane of Gram-negative bacteria, and it is one of the few contaminants that a purity figure will never tell you about. A synthetic peptide can be 99% pure by HPLC, correctly identified by mass spectrometry, and still carry enough endotoxin to change what a cell-based assay reports. Because endotoxin survives autoclaving and passes through a 0.22 µm filter, "sterile" and "low endotoxin" are separate claims established by separate tests. Peptide endotoxin testing is the analysis that closes that gap. This article covers what endotoxin is, how it is measured, what the units mean, and how to read — or notice the absence of — an endotoxin line on a certificate of analysis. All products discussed are for research use only.
What Endotoxin Actually Is
Endotoxin is lipopolysaccharide (LPS): a glycolipid built from a hydrophobic lipid A anchor, a non-repeating core oligosaccharide, and a distal O-antigen polysaccharide. It is a structural component of the outer membrane of Gram-negative organisms such as Escherichia coli and Pseudomonas species, released when cells divide or lyse.
The lipid A portion is chemically robust. It is not a protein, so it does not denature, and it is not destroyed by the temperatures and hold times used in routine autoclaving. It is also a small, soluble molecule relative to a bacterial cell, so sterile filtration through a 0.22 µm membrane removes the organism while leaving dissolved LPS in the filtrate. This is the single most consequential fact in the subject: a preparation can be genuinely sterile — no viable organisms — and still contain biologically active endotoxin from organisms that were present earlier in the process, in a water system, or on glassware.
"Endotoxin-free" is therefore not something sterility implies. It is an analytical determination, and like any analytical determination it has a method, a unit, and a detection limit attached to it.
Why Endotoxin Matters in Peptide Research
In a research setting the concern with endotoxin is experimental validity, not product safety. LPS is recognised by a receptor system comprising LPS-binding protein, CD14, and the TLR4–MD-2 complex, and engagement of that system in responsive cells drives NF-κB activation and the transcription of pro-inflammatory cytokines such as IL-6, TNF-α, and IL-1β.
For anyone running a cell-based assay, that is a problem of attribution. Cytokine output, NF-κB reporter signal, proliferation, differentiation, and apoptosis readouts are all commonly used endpoints in peptide work, and all of them can move in response to contaminating LPS at concentrations well below anything visible in a purity chromatogram. The observed effect then belongs to the contaminant rather than to the peptide. In animal models, low-level endotoxin can produce a systemic inflammatory background that obscures the response actually attributable to the compound under study.
The practical consequence is reproducibility. Two batches of the same sequence at the same stated purity can behave differently in the same assay if their endotoxin burden differs, and without endotoxin data there is nothing in the paperwork that would explain the discrepancy. Peptide endotoxin testing exists to remove that hidden variable.
How Endotoxin Is Measured
The established method is the Limulus Amebocyte Lysate (LAL) assay, which uses the clotting cascade of horseshoe crab (Limulus polyphemus) amebocytes. LPS activates a protease cascade beginning with Factor C, and the three commercial formats differ in how that activation is read out:
Gel-clot. The oldest format and a limit test rather than a continuous measurement. The sample is run against a lysate of known labelled sensitivity, and the endpoint is the last dilution at which a solid clot still forms. Results are semi-quantitative — reported against a threshold or as a doubling-dilution endpoint — not as a precise concentration.
Turbidimetric. Reads the increase in optical density as coagulin precipitates. The onset time or rate of turbidity change is compared to a standard curve, giving a quantitative result across a wider range than gel-clot.
Chromogenic. Substitutes a synthetic peptide substrate that releases a chromophore when cleaved by the activated clotting enzyme. Absorbance is proportional to endotoxin concentration, and the format suits plate-based, higher-throughput work.
Recombinant Factor C (rFC). A non-animal alternative that uses recombinantly expressed Factor C and a fluorogenic substrate, isolating the first step of the cascade rather than the whole clotting pathway. It removes the dependence on horseshoe crab lysate and, because it does not involve Factor G, is not subject to the β-glucan cross-reactivity that can produce false positives in LAL formats.
Whatever the platform, the result is expressed in endotoxin units — EU/mg of peptide, or EU/mL of a solution at a stated concentration. These are not interchangeable. A figure quoted without its unit, or an EU/mL figure quoted without the concentration it was measured at, cannot be converted or compared to anything.
Two limitations are worth knowing before trusting any single number. Interference occurs when sample components enhance or inhibit the cascade, which is why samples are usually diluted and checked with a spiked control. Low endotoxin recovery (LER) is a distinct, time-dependent masking effect, associated particularly with matrices combining a chelator and a surfactant, in which endotoxin known to be present is progressively under-recovered by the assay. Masked endotoxin is not inert: it can escape detection in Factor C-based assays while remaining detectable by cell-based readouts, which is precisely the failure mode that matters to a researcher.
Reading Endotoxin Data on a COA
A usable endotoxin entry on a certificate of analysis states three things — value, unit, and method:
Endotoxin: <0.05 EU/mg (kinetic chromogenic LAL)
The "<" is doing real work in that line. It means the assay did not detect endotoxin above its own sensitivity limit for that sample and dilution. It does not mean zero, and a lower stated limit sometimes reflects a more sensitive method rather than a cleaner batch.
Three things to check. First, that a method is named — gel-clot, turbidimetric, chromogenic, or rFC — because a bare number is not interpretable. Second, that the unit is stated, and if it is EU/mL, that the concentration tested is stated with it. Third, that the result is batch-specific rather than a generic specification reproduced on every document.
The most common gap is simpler than any of these: no endotoxin line at all. An absent entry is not a low result and should not be read as one. It most often means the test was not performed on that batch. If endotoxin is relevant to the intended work, its absence from the paperwork is itself the finding, and the question to ask is whether batch-specific data exists rather than what the number probably is.
Limitations and Common Misreadings
Reading "<0.05 EU/mg" as "zero endotoxin" is the most frequent error; it is a detection limit, not an absence. The second is comparing figures in different units, or comparing an EU/mL result to an EU/mg result without the concentration needed to relate them. The third is assuming a single result transfers across matrices — recovery in a simple buffer says little about recovery in serum-supplemented medium, which is why matrix-matched spike controls exist. And because LER develops over time, a result generated at release does not automatically describe the same material after extended storage in the same formulation.
Conclusion
Endotoxin and purity answer different questions. Purity methods describe what fraction of the material is the intended sequence; peptide endotoxin testing describes a contaminant that those methods are blind to by design. A complete picture of a research-grade peptide requires both, reported per batch with methods and units stated. Where endotoxin data is missing, the honest position is that the value is unknown rather than low.
Frequently Asked Questions
Is a sterile-filtered peptide also endotoxin-free?
No. Filtration at 0.22 µm removes bacterial cells but not dissolved lipopolysaccharide, and autoclaving does not destroy it. Sterility and endotoxin content are established by separate tests, and only the latter produces an endotoxin number.
What is the difference between EU/mg and EU/mL?
EU/mg normalises endotoxin to the mass of peptide and is what a lyophilised solid is usually reported against. EU/mL describes a solution and is only meaningful alongside the concentration at which it was measured. The two cannot be compared without that concentration.
Why do LAL and rFC results sometimes disagree?
The assays read different points in the same cascade. LAL formats depend on the full clotting pathway and can respond to β-glucans through Factor G, producing false positives; rFC isolates Factor C and does not. Matrix interference and low endotoxin recovery affect the formats differently as well, so an orthogonal method is a reasonable check when a result is unexpected.
All products are for research use only — not for human or animal consumption.