Glutathione (GSH): The γ-Glutamyl Tripeptide and Its Redox Chemistry

Glutathione is written γ-Glu-Cys-Gly, with formula C10H17N3O6S and a mass of 307.3 g/mol. Two of its three bonds are ordinary: the cysteine-to-glycine linkage is a standard α-peptide bond. The first one is not. Glutamate is joined to cysteine through the γ-carboxyl of its side chain rather than through the α-carboxyl that peptide bonds normally use.

That single structural quirk has large consequences. Aminopeptidases and most proteases recognise the geometry of α-peptide bonds; the γ-linkage does not present that geometry, so glutathione is not degraded by the general proteolytic machinery that would otherwise dismantle a free tripeptide in short order. A cell can therefore maintain glutathione at millimolar concentrations in its cytosol without it being consumed as a nitrogen source.

Breaking the γ-bond requires a specialised enzyme. That enzyme is γ-glutamyl transpeptidase, which sits on the extracellular face of the plasma membrane in tissues including kidney, liver and intestine, and which initiates the γ-glutamyl cycle by transferring the glutamyl group to an acceptor and releasing cysteinylglycine for further cleavage. The location of that enzyme - outside the cell, concentrated in gut and liver - turns out to be central to the bioavailability question discussed below.

Synthesis runs in two ATP-dependent steps. Glutamate-cysteine ligase joins glutamate and cysteine to form γ-glutamylcysteine, and glutathione synthetase adds glycine. The first step is rate-limiting and is feedback-inhibited by glutathione itself, but in practice the constraint on synthesis is usually the supply of cysteine, the scarcest of the three amino acids.

It is a mistake to think of glutathione as a substance that gets used up. The functional entity is a couple. The reduced form, GSH, carries a free thiol on its cysteine residue; when it donates a reducing equivalent, two GSH molecules join through a disulphide bond to form one molecule of glutathione disulphide, GSSG. Glutathione reductase then reduces GSSG back to two GSH, drawing electrons from NADPH.

This makes glutathione the connection point between two systems. The pentose phosphate pathway generates NADPH; glutathione reductase spends it; glutathione peroxidases and other thiol-dependent systems consume the resulting GSH. The relevant measurement is therefore not the amount of glutathione present but the ratio of GSH to GSSG, which in a healthy cytosol is heavily weighted toward the reduced form.

Those ratios are also compartment-specific, which is easy to miss. Cytosolic, mitochondrial and endoplasmic reticulum glutathione pools are maintained at different redox potentials, and the ER is deliberately more oxidising because disulphide bond formation in nascent proteins requires it. Mitochondria have no glutathione synthesis capacity of their own and import GSH from the cytosol. A whole-cell or whole-blood glutathione number averages across pools that are doing different jobs.

A practical corollary follows for handling the reduced compound as a reagent: GSH air-oxidises to GSSG readily in aqueous solution. Measurements of glutathione redox status are notoriously sensitive to sample handling, because oxidation that occurs after collection is indistinguishable from oxidation that occurred in the cell.

Glutathione is a cosubstrate for two large, mechanistically distinct enzyme families, and conflating them produces most of the imprecision in popular accounts of what glutathione does.

The glutathione peroxidases reduce hydrogen peroxide and organic hydroperoxides to water and the corresponding alcohols, oxidising GSH to GSSG in the process. Brigelius-Flohé and Maiorino's review is the source to read on this family, and its main lesson is specificity: the isoforms differ in tissue distribution, subcellular location and substrate preference, with GPX4 uniquely able to act on lipid hydroperoxides within membranes. Most human glutathione peroxidases carry selenocysteine in the active site, which is why selenium status feeds directly into this capacity.

The glutathione S-transferases do something different. They catalyse conjugation, attaching the GSH thiol to electrophilic centres on xenobiotics, reactive metabolites and endogenous electrophiles. This is the committed step of phase II metabolism: the conjugate is more water-soluble than the parent electrophile and is exported and processed through the mercapturic acid pathway to a N-acetylcysteine conjugate for excretion. Here glutathione is not reducing anything - it is being consumed stoichiometrically as a nucleophilic tag, and the resulting loss must be replaced by synthesis rather than by reduction of GSSG.

The distinction matters for interpreting glutathione depletion. Oxidative consumption shifts the GSH/GSSG ratio and is in principle recoverable by reductase activity. Conjugative consumption removes glutathione from the system entirely and requires new synthesis, which is why heavy electrophile exposure depletes glutathione in a way that a purely oxidative challenge does not.

This is the most contested question in the glutathione literature, and the honest summary is that it is unresolved rather than settled in either direction.

The case against oral availability is mechanistic and has direct human support. γ-Glutamyl transpeptidase is abundantly expressed on intestinal and hepatic surfaces, and it is the enzyme evolved to dismantle exactly this molecule. Anything crossing that barrier faces first-pass exposure to it. The 1992 study by Witschi and colleagues tested the prediction in humans with a large single oral load and found no meaningful rise in circulating glutathione, a result frequently cited as showing that oral glutathione is degraded to its constituent amino acids before reaching the systemic circulation.

The case for it comes from trials with different designs. The 2015 randomised controlled trial by Richie and colleagues supplemented over months and measured glutathione in blood compartments rather than tracking a single acute load, reporting increased body stores. These two results are not straightforwardly contradictory, because they measure different things: an acute plasma excursion and a chronic accumulation in stores are separable outcomes, and it is possible for the second to occur through salvaged constituent amino acids even if the intact tripeptide never survives absorption.

The comparative crossover work by Schmitt and colleagues probes that ambiguity from another angle by placing oral glutathione alongside N-acetylcysteine, which supplies cysteine to endogenous synthesis rather than delivering the tripeptide, and alongside a sublingual route that avoids first-pass metabolism. Studies of this kind are where the question will actually be settled, since they distinguish between 'glutathione cannot be absorbed' and 'glutathione status can be raised, but not by delivering intact glutathione to the gut'.

For present purposes the accurate statement is narrow: the intracellular biochemistry of glutathione is thoroughly characterised, whereas the effect of exogenous glutathione on human tissue glutathione status is a live disagreement in the primary literature with credible studies on both sides.

Beyond bioavailability, a measurement problem runs underneath much of this literature. Because GSH oxidises so readily once a sample leaves its cellular environment, reported GSH/GSSG ratios depend heavily on collection and processing methods, and cross-study comparison is correspondingly hazardous. Differences between studies sometimes reflect assay handling rather than biology.

There is also a persistent inferential gap between glutathione status and clinical outcome. Depleted glutathione is reported in association with a long list of human conditions, but association does not establish direction: a condition that generates electrophiles or peroxides will deplete glutathione, so low glutathione may be a consequence rather than a cause. Interventional human evidence that raising glutathione status changes disease outcomes is much thinner than the observational literature on glutathione levels.

Where glutathione is unambiguously well characterised is as biochemistry and as a reagent. Its structure, synthesis, enzymology and redox behaviour are established in detail, which is what makes the reduced tripeptide a standard material in enzyme assays, redox chemistry and analytical work. The uncertainty in this field concerns systemic supplementation, not the underlying chemistry.