What Is Glutathione?

Published August 2026 · Patriot Labs Research Team · ~11 min read

A molecule with no real controversy

Glutathione is the most abundant low-molecular-weight thiol in most animal cells, present intracellularly in the low millimolar range while circulating extracellular concentrations are orders of magnitude lower. It is a tripeptide — glutamate, cysteine, glycine, in that order — and it has been studied continuously for more than a century. A search of Europe PMC on 10 August 2026 for the word “glutathione” in the title alone returned a hit count of 35,582 records.

That number is unusual for this catalogue. Most of the peptides discussed in these guides have literatures measured in dozens or low hundreds of papers, much of it unreplicated. Glutathione has the opposite problem: the core biochemistry is so thoroughly established that there is nothing to hedge about. Alton Meister and Mary Anderson's 1983 Annual Review of Biochemistry article, titled simply “Glutathione,” was already a synthesis of decades of work.¹

So this guide inverts the usual structure. It is confident about mechanism, and careful instead about the much narrower question of what happens when the intact tripeptide is administered rather than synthesised. Those are two different literatures, and conflating them is the most common error in writing about this molecule. If the vocabulary below is unfamiliar, the introduction to what peptides are covers amide bonds, residues and nomenclature.

The gamma bond, and why it changes everything

Write out the full name and the oddity is right there: glutathione is gamma-L-glutamyl-L-cysteinylglycine. In an ordinary peptide, every amide bond runs from the alpha-carboxyl of one residue to the alpha-amino group of the next. That is the geometry a ribosome produces, and it is the geometry every guide on this site otherwise describes.

Glutathione's second bond, between cysteine and glycine, is a normal alpha linkage. Its first bond is not. The amide between glutamate and cysteine is formed from the glutamate side-chain carboxyl — the gamma carboxyl, three carbons out from the alpha carbon — onto the alpha-amino group of cysteine. The glutamate residue is effectively hung off the chain sideways. Three consequences follow, and they explain most of what is distinctive about the molecule.

It cannot be made by a ribosome. Translation only builds alpha linkages. Glutathione is therefore not a gene product; it is assembled enzymatically, one bond at a time, by two dedicated ligases.

It is resistant to ordinary peptidases. Aminopeptidases and most endopeptidases recognise the standard alpha-amide backbone and its flanking geometry. The gamma linkage does not present that geometry, so the bond is invisible to them. This is the single most interesting structural fact about glutathione: a three-residue peptide would normally be degraded almost immediately in any biological fluid, and glutathione is not. Compare that with the hydrolytic vulnerability described in the guide to how peptides degrade, where backbone cleavage is one of the dominant failure routes for conventional sequences.

One enzyme can cleave it, and it sits outside the cell. Gamma-glutamyl transpeptidase, an ectoenzyme anchored on the outer face of the plasma membrane in tissues such as kidney, biliary epithelium and intestinal brush border, hydrolyses or transfers the gamma-glutamyl group. Extracellular glutathione is therefore taken apart at the cell surface rather than imported intact. This asymmetry — built inside, dismantled outside — is the mechanistic core of the bioavailability discussion below.

GSH and GSSG: a buffer, not a scavenger

All of glutathione's redox chemistry runs through one atom: the sulfur of the cysteine thiol. In reduced glutathione, written GSH, that sulfur carries a hydrogen and is available to donate an electron, to attack an electrophile, or to exchange with another sulfur.

When two GSH molecules are oxidised, their thiols join to form a disulfide bridge, producing one molecule of glutathione disulfide, written GSSG. The reaction is reversible, and the cell spends energy keeping it pushed to the left:

2 GSH ⇌ GSSG + 2 H⁺ + 2 e⁻

In a healthy cytosol the pool sits overwhelmingly in the GSH form, with GSSG a small minority species. The ratio between them, taken together with the absolute size of the pool, defines a reduction potential — and it is that potential, rather than any single molecule's activity, that constitutes the cell's principal redox buffer. Freya Schafer and Garry Buettner's much-cited 2001 paper in Free Radical Biology and Medicine is titled, precisely, “Redox environment of the cell as viewed through the redox state of the glutathione disulfide/glutathione couple.”²

Two points are worth making explicit because popular writing usually gets them wrong. First, glutathione is not primarily a radical scavenger that gets consumed; it is a buffer that gets recycled. Describing it as an antioxidant that “neutralises free radicals” is not exactly false, but it misses the architecture: the pool is a reservoir of reducing equivalents held at a defined potential, and enzymes draw on it and return it.

Second, because the oxidised form contains two glutathione units and the reduced form one, the couple is not a simple one-to-one ratio in thermodynamic terms. Halving the total pool while keeping the GSH:GSSG ratio constant still shifts the potential. Total concentration and ratio both matter, which is one reason single-number readouts of “glutathione status” are hard to interpret.

Compartments differ, too. The cytosol and mitochondrial matrix are held strongly reducing; the endoplasmic reticulum lumen is deliberately more oxidising, because disulfide bond formation in secreted and membrane proteins requires it. There is no single cellular redox state — only compartment-specific ones.

The enzymes that spend it

Glutathione does very little on its own at physiological rates. Almost all of its work is enzyme-catalysed, and the enzyme families involved were reviewed comprehensively by Marcel Deponte in 2013 under the title “Glutathione catalysis and the reaction mechanisms of glutathione-dependent enzymes.”³ The table below sets out the main ones.

The cycle is worth stating in one sentence, because it is the thing people most often garble. A peroxidase or a transferase consumes GSH and produces GSSG or a conjugate; glutathione reductase pulls electrons off NADPH to convert GSSG back to two GSH; the pool turns over continuously and the cell pays for it in NADPH rather than in newly synthesised glutathione. Net synthesis is only needed to replace what is lost as conjugates and efflux.

Where glutathione actually comes from

Because the gamma bond is not something translation can produce, glutathione is built by two dedicated cytosolic ligases, each consuming one molecule of ATP. Shelly Lu's 2013 review in Biochimica et Biophysica Acta, titled “Glutathione synthesis,” is the standard entry point,⁶ alongside the 2009 overview by Forman, Zhang and Rinna.⁷

The practical consequence is that cysteine supply, not glutamate or glycine supply, is usually what constrains the rate. Cysteine is the least abundant of the three, it is itself readily oxidised, and it is the residue that carries the functional thiol. This is why so much of the experimental literature on raising cellular glutathione works on cysteine delivery rather than on glutathione itself.

Two further structural facts matter. Synthesis is cytosolic; mitochondria, which hold a substantial and functionally distinct pool, do not make their own and rely on carrier-mediated import. And step 1 is feedback-inhibited by the product, so a cell with a full pool throttles its own synthesis.

Put those together and the reason “raising glutathione” is not as simple as supplying glutathione becomes a matter of pathway architecture rather than opinion. The pool is intracellular, built under negative feedback from amino acid precursors, and the intact tripeptide is dismantled at cell surfaces rather than imported. Nothing in that arrangement is set up to accept a delivery of finished product.

The bioavailability problem

This is where the confident tone of the preceding sections has to stop. The biochemistry above is not contested. What happens when intact glutathione is administered orally is contested, and the literature on it is far smaller than the literature on the molecule.

The scale difference is easy to quantify. Against 35,582 Europe PMC records with “glutathione” in the title, a search combining the exact phrase “oral glutathione” with “supplementation” returned a hit count of 85 on the same day. Two orders of magnitude separate the biochemistry from the administration question.

The question was put directly and early. A 1992 paper in the European Journal of Clinical Pharmacology by Witschi, Reddy, Stofer and Lauterburg is titled “The systemic availability of oral glutathione,” and it remains the reference most often reached for.⁸ Its bibliographic record was verified for this guide, but neither its abstract nor its full text could be retrieved, so nothing is asserted here about its findings beyond the fact that it asked the question.

More recent work states the obstacle plainly. A 2026 study in the British Journal of Nutrition comparing liposomal and plain glutathione in cellular and human models opens its abstract with the position that “oral glutathione supplementation is limited by poor bioavailability due to gastrointestinal degradation and low membrane permeability.”⁹ That is a fair summary of the mechanistic expectation set out in the previous section: a gamma-glutamyl bond that survives peptidases still meets a brush border dense with the one enzyme that cleaves it, and a tripeptide with two ionisable carboxyls and a free amine is not a molecule that crosses lipid membranes readily.

The response in the applied literature has been formulation chemistry rather than new biology. A 2026 randomised crossover pilot in Antioxidants compared three oral presentations of glutathione — standard, liposomal and micellar — in 14 healthy adult volunteers, using baseline-corrected pharmacokinetic parameters and a targeted panel of glutathione-derived metabolites.¹⁰ The relevant finding is structural rather than promotional: systemic exposure differed substantially between presentations of the same molecule, which is the signature of a delivery-limited compound. The authors describe the work as a pilot with a modest sample size, conducted on a proprietary commercial formulation.

A 2026 review in Inflammopharmacology is titled “A review of the potential benefits and limitations of oral glutathione supplementation”¹¹ — limitations still in the title of a review thirty-four years after Witschi, which is itself informative about how settled the question is.

None of this bears on in-vitro work, where glutathione is added directly to a defined medium and the absorption problem does not arise. It bears entirely on the interpretive leap from a biochemical role that is beyond dispute to a claim about what supplying the molecule accomplishes.

Glutathione, NADPH and NAD+

Glutathione reductase does not run on nothing. Every GSSG it reduces costs one NADPH, which means the glutathione pool's capacity to stay reduced is downstream of the cell's NADPH supply — largely from the oxidative branch of the pentose phosphate pathway, with contributions from malic enzyme and the NADP-dependent isocitrate dehydrogenases.

That places glutathione in the same conceptual family as the nicotinamide cofactors covered in the guide to what NAD+ is. Three points make the parallel worth drawing.

Both are pools defined by a ratio. The absolute amount of NAD, NADP or glutathione in a cell says much less than the proportion in each oxidation state. A large, fully oxidised glutathione pool is not a functional one, and the same logic applies to NAD+/NADH.

The two nicotinamide pools do different jobs. NAD+/NADH is held oxidised and serves catabolism, ferrying electrons to the respiratory chain. NADP+/NADPH is held reduced and serves reductive biosynthesis and antioxidant defence, including glutathione reductase. They are interconverted by NAD kinase, so total NAD availability sits upstream of NADPH capacity, but they are not interchangeable currencies.

Both are recycled, not consumed. Neither cofactor is used up stoichiometrically in normal operation. Framing either as something that gets “depleted” and needs “topping up” misdescribes the system — which is one reason both recur in the longevity-adjacent literature surveyed in the anti-aging peptides overview. NAD+ and glutathione are catalogued separately here, but researchers working on either pool frequently work on both.

Bench handling: a free thiol is a liability

The same chemistry that makes glutathione useful makes it fragile in a vial. A free thiol is an easily oxidised functional group, and it does not need an enzyme to react.

Reduced glutathione in aqueous solution oxidises to the disulfide on standing in air. The rate rises with pH, because the reactive species is the thiolate anion and deprotonation increases with alkalinity, and it rises sharply with trace transition metals — copper and iron contamination in buffers or glassware catalyses thiol autoxidation efficiently. Dissolved oxygen and headspace volume are therefore real experimental variables, not fussiness. Related chemistry for other sequences is covered in the guide to how peptides degrade, where thiol oxidation and disulfide scrambling are treated alongside deamidation and hydrolysis.

This has direct implications for measurement. An assay that reduces the sample before quantification reports total glutathione, GSH plus twice GSSG; an assay that measures free thiol directly reports GSH only. The two numbers answer different questions, and a sample that oxidised during handling looks identical to one that was genuinely oxidised in the system under study. Reported GSH:GSSG ratios are notoriously sensitive to how quickly and how coldly the sample was processed. For solid material the usual constraints for lyophilised research compounds apply: cold, dark, dry, minimal freeze–thaw cycling of stock, and identity confirmed against a certificate of analysis rather than assumed.

What the evidence does not establish

Being confident about mechanism makes it more important, not less, to be precise about where the confidence stops.

Established biochemistry is not evidence for supplementation. That glutathione is essential to peroxidase and transferase catalysis says nothing about the consequences of administering the molecule. The first is enzymology; the second is pharmacokinetics. A well-characterised endogenous role is one of the most common sources of overclaiming in this field.

The human pharmacokinetic literature is small and heterogeneous. Eighty-five Europe PMC hits for oral glutathione supplementation covers everything from dermatology to hepatology to sports science, across different formulations, endpoints and analytical methods. The 2026 crossover study cited above enrolled 14 volunteers and describes itself as a pilot. Small crossover pilots on proprietary formulations, frequently with commercial involvement, are hypothesis-generating.

Blood glutathione is a poor proxy for tissue glutathione. The pool that matters is intracellular and compartment-specific. Whole-blood or plasma measurements are dominated by erythrocytes and by ex-vivo oxidation artefacts, and there is no general reason to assume they track hepatic, neuronal or mitochondrial pools.

Low glutathione in a disease state does not establish direction. Depressed glutathione is reported across many pathologies. Whether that is cause, consequence or correlate is not answered by the observation, and correcting a marker is not the same as correcting a mechanism. Nor are in-vitro results transferable: cell-culture work adds glutathione or its precursors at experimenter-defined concentrations in a defined medium, and those concentrations are not evidence about any other system.

“More antioxidant” is not self-evidently better. Reactive oxygen species are signalling molecules as well as damaging ones. A 2026 review in Antioxidants on antioxidants and exercise frames the trade-off explicitly, noting that exercise-derived ROS “are required for mitochondrial and hypertrophic adaptations, creating a practical trade-off.”¹² Shifting a redox buffer is an intervention in a signalling system, with the ambiguity that implies.

Nothing above should be read as a claim about any Patriot Labs product. Glutathione is catalogued here as a research chemical, and the references cited describe published research rather than product performance.

Frequently asked questions

Is glutathione a peptide? Chemically, yes — it is a tripeptide of glutamate, cysteine and glycine. Structurally it is an outlier, because one of its two amide bonds is a gamma linkage from the glutamate side chain rather than the standard alpha linkage. It cannot be produced by a ribosome and is not cleaved by ordinary peptidases.

What does the GSH/GSSG ratio actually measure? GSH is the reduced form with a free cysteine thiol; GSSG is two GSH joined through a disulfide. The proportion of the pool in each form, combined with the pool's absolute size, sets a reduction potential. That potential is what people mean by the cell's redox state, and it differs between compartments.

Why does the cycle need NADPH rather than NADH? Glutathione reductase is specific for NADPH. The two nicotinamide pools are held at different redox states for different purposes: NAD+/NADH oxidised for catabolism, NADP+/NADPH reduced for biosynthesis and antioxidant defence. Regenerating a reduced buffer draws on the reduced pool.

Is oral glutathione bioavailability actually settled? No. The mechanistic expectation is that the intact tripeptide is degraded at cell surfaces by gamma-glutamyl transpeptidase and crosses membranes poorly, and recent papers state that expectation directly. The human pharmacokinetic literature is small, uses varied proprietary formulations, and mostly reports pilot-scale crossover designs.

Does research on glutathione overlap with NAD+ research? Frequently. Glutathione reductase consumes NADPH, and the nicotinamide pools are interconverted by NAD kinase, so the two systems are coupled through cellular reducing power. The NAD+ guide covers that pool in its own right.

Is it approved for human use? No. Glutathione supplied by Patriot Labs is sold strictly for in-vitro research and laboratory use only. It is not for human or veterinary consumption, and nothing in this guide describes how to use it.

All Patriot Labs products are sold strictly for in-vitro research and laboratory use only. Not for human or veterinary consumption. This guide is educational and describes peptide chemistry and published research in general terms; it is not medical advice, does not describe how to use any product, and the references cited do not constitute a product claim.