Glutathione 1500mg

Glutathione is the tripeptide gamma-L-glutamyl-L-cysteinyl-glycine, formula C10H17N3O6S, average mass 307.32 g/mol, CAS 70-18-8, PubChem CID 124886. It is the most abundant low-molecular-weight thiol in mammalian cells, present at roughly 1 to 10 mM depending on tissue, with the highest concentrations in liver. Plasma carries far less: basal plasma glutathione in healthy volunteers measured 6.2 micromol/L in the Bern pharmacology group's study, about a thousandfold below the intracellular figure. That gradient, not the absolute number, is the fact most descriptions of this molecule leave out.

The compound was first noticed in 1888 by de Rey-Pailhade, who found a substance in yeast that reacted with elemental sulfur to liberate hydrogen sulfide and called it philothion. Frederick Gowland Hopkins isolated it from animal tissue in 1921 and named it glutathione, but described it as a dipeptide of glutamate and cysteine. The glycine residue was only established at the end of that decade, and the structure was settled definitively when Harington and Mead achieved a chemical synthesis in 1935 that reproduced the natural material, including the unusual linkage between the first two residues.

That linkage is the whole story of the molecule. The bond between glutamate and cysteine is formed from the gamma-carboxyl group of glutamate rather than the alpha-carboxyl used in every ribosomally translated protein. It is an isopeptide bond, and ordinary aminopeptidases do not recognise it. Only one mammalian enzyme, gamma-glutamyl transpeptidase, can remove the glutamyl residue, and that enzyme sits on the outer face of the plasma membrane rather than inside the cell. Glutathione is therefore chemically protected wherever it does its work and degradable only after it has been exported, which is why intracellular concentrations reach millimolar and why the question of what happens to an orally ingested dose is genuinely difficult to answer.

Glutathione has no receptor. It is not a signalling peptide in the sense that the rest of a peptide catalogue is, and looking for a binding site is the wrong model. Its activity comes entirely from the sulfhydryl group on the cysteine residue, which is a good nucleophile and a good one-electron and two-electron donor at physiological pH. Everything the molecule does follows from that thiol: reduction of peroxides, conjugation to electrophiles, reversible modification of protein cysteines, and coordination of metals.

When glutathione reduces something, two molecules are consumed and one molecule of glutathione disulfide is produced, linking the two cysteine residues through a disulfide bridge. Oxidised glutathione is a distinct chemical species, C20H32N6O12S2, 612.6 g/mol, and is usually written GSSG. Because the oxidation is a two-for-one reaction, the Nernst expression for the half-cell contains the square of the reduced species: the reduction potential tracks the logarithm of [GSH]squared over [GSSG]. This is the reason the ratio matters more than the concentration, and it is the reason a small rise in GSSG shifts the cellular redox potential much further than an equivalent fall in GSH.

As a cofactor, glutathione feeds two enzyme families. The glutathione peroxidases use it to reduce hydrogen peroxide and organic hydroperoxides to water and the corresponding alcohols. Mills identified the first of them in 1957 as the erythrocyte activity that protected haemoglobin from oxidative breakdown, and Rotruck and colleagues showed in 1973 that the enzyme is a selenoprotein, explaining a large part of the nutritional biology of selenium in a single paper. The glutathione S-transferases use it differently, catalysing attack of the thiolate on electrophilic carbon, nitrogen and sulfur centres to produce glutathione conjugates that are exported and processed to mercapturic acids for excretion.

The disulfide produced by all of this is not discarded. Glutathione reductase reduces GSSG back to two molecules of GSH using NADPH supplied largely by the pentose phosphate pathway, so the cell runs a closed redox cycle in which glutathione is the working fluid and NADPH is the power source. A parallel set of reactions, catalysed by glutaredoxins, adds and removes glutathione from protein cysteine residues. Protein S-glutathionylation of this kind is a reversible post-translational modification that acts as a redox switch on enzymes, ion channels and transcription factors, which is how a small thiol ends up participating in signalling without ever binding a receptor.

Synthesis is a two-step, ATP-dependent cytosolic process. Glutamate-cysteine ligase joins glutamate and cysteine through the gamma-carboxyl, and glutathione synthetase adds glycine. The ligase is the rate-limiting enzyme and is a heterodimer of a 73 kDa catalytic subunit, GCLC, and a 31 kDa modifier subunit, GCLM, which is enzymatically inactive but lowers the Km for glutamate and raises the Ki for feedback inhibition. Reported values are a Km of about 1.8 mM for glutamate, 0.1 to 0.3 mM for cysteine, and non-allosteric competitive feedback inhibition by glutathione with a Ki near 2.3 mM. Intracellular glutamate sits roughly tenfold above its Km while cysteine hovers near its own, which is the quantitative reason cysteine supply, not enzyme abundance, usually sets the synthetic rate.

In a protein, residue one connects to residue two through the alpha-carboxyl of the first amino acid. In glutathione it does not. Glutamate carries a second carboxyl on its side chain, and it is that gamma-carboxyl that forms the amide bond to the cysteine amino group. The resulting isopeptide bond has the same chemistry as any amide, but it does not present the geometry that aminopeptidases and most endopeptidases evolved to cleave.

The consequence is protection. A linear tripeptide with a conventional backbone would be degraded within minutes of appearing in cytosol. Glutathione instead accumulates to concentrations of 1 to 10 mM, higher than any other small thiol in the cell and higher than most metabolites of any class. Hepatic cytosolic glutathione still turns over with a half-life of only 2 to 3 hours, but that turnover is dominated by export and consumption, not by proteolysis of the peptide backbone.

One enzyme can undo the linkage. Gamma-glutamyl transpeptidase transfers the gamma-glutamyl moiety to an acceptor amino acid, with cystine the best acceptor, leaving cysteinylglycine which a dipeptidase then splits. Critically, GGT is an ectoenzyme: its active site faces the extracellular space. Glutathione must be exported before it can be broken down, and the products are then taken back up as amino acids and reassembled inside the cell. Orlowski and Meister described this loop in a 1970 paper in the Proceedings of the National Academy of Sciences and named it the gamma-glutamyl cycle, proposing that its real function was amino acid transport rather than antioxidant defence.

The cleanest test of that proposal came from knocking the transpeptidase out. GGT-deficient mice generated at Baylor in 1996 appeared normal at birth, then grew to roughly half wild-type weight by six weeks, failed to reach sexual maturity, developed cataracts and a grey coat cast, and mostly died between 10 and 18 weeks. Urinary glutathione was elevated 2,500-fold and plasma glutathione sixfold, while tissue glutathione fell in eye, liver and pancreas and plasma cysteine dropped to about a fifth of normal. Feeding those mice N-acetylcysteine restored growth and partially restored coat colour. The animals were not dying of oxidative stress; they were starving for cysteine because the only enzyme that can recover it from exported glutathione was missing.

Oxidation of glutathione is a two-molecule event. Two GSH give up one electron each and join through a disulfide to make one GSSG. Writing the Nernst equation for that half-cell puts the reduced species in the numerator squared, so the reduction potential is proportional to the logarithm of [GSH]squared over [GSSG]. A cell that loses ten percent of its glutathione to oxidation does not shift its redox potential by ten percent of anything; the doubling of GSSG in the denominator dominates. This is why any measurement that reports total glutathione, or reduced glutathione alone, throws away the information that actually matters.

In a cell keeping up with its own oxidant load, more than 98 percent of the glutathione pool is reduced. Schafer and Buettner assembled the published values in 2001 and found that the half-cell reduction potential of the couple maps onto cell state with surprising consistency: about -240 mV in proliferating cells, about -200 mV as cells differentiate, and about -170 mV in cells committing to apoptosis. The potential is not a passive readout of damage. It moves first, and the phenotype follows.

The couple is not uniform across the cell. Cytosol and mitochondria are held strongly reducing, while the endoplasmic reticulum is deliberately kept oxidising because disulfide bond formation in nascent secreted proteins requires it. Hwang, Sinskey and Lodish measured this directly in 1992 with a glycosylatable trapped peptide probe and found the secretory pathway running a reduced-to-oxidised glutathione ratio of 1:1 to 3:1 while the overall cellular ratio ran 30:1 to 100:1. A single whole-cell number averages compartments the cell is at considerable metabolic expense keeping different.

The practical problem is that measuring the ratio is harder than it sounds. Thiols oxidise during sample handling, and the artefact runs in the direction that inflates GSSG. Rossi and colleagues showed in 2006 that when blood samples were pretreated with N-ethylmaleimide to block free thiols immediately on collection, apparent GSSG concentrations were up to threefold different from untreated handling of the same material, and glutathionylated protein measurements shifted by around 50 percent under oxidant challenge. Any published GSH/GSSG figure that does not state its thiol-blocking step should be read with that in mind.

The peroxidase family was the first glutathione-dependent enzyme system to be characterised. Mills reported in the Journal of Biological Chemistry in 1957 that erythrocytes contained an activity that protected haemoglobin from oxidative breakdown and that the activity required glutathione. Sixteen years later Rotruck and colleagues incubated haemolysates from selenium-deficient rats with ascorbate or hydrogen peroxide and found that added glutathione no longer protected the haemoglobin, because those cells were effectively devoid of peroxidase activity. Purified enzyme carried most of the selenium-75 label from erythrocytes labelled in vivo. Glutathione peroxidase is a selenoprotein, and the catalytic residue is selenocysteine.

Eight human glutathione peroxidases are now recognised. GPX1 is the cytosolic workhorse acting on hydrogen peroxide, GPX3 is secreted into plasma, and GPX4 is the outlier: it is a monomer, it accepts complex lipid hydroperoxides while they are still esterified in membranes, and it is the only one that can act on phospholipid hydroperoxides in situ. That specificity turned out to matter enormously. In 2014 Yang and colleagues at Columbia looked for a common mediator behind 12 chemically unrelated ferroptosis-inducing compounds and found GPX4. One class of inducers depleted glutathione and thereby switched GPX4 off indirectly; a second class inhibited the protein directly. Manipulating GPX4 levels changed the potency of all 12 and of none of 11 control lethal compounds. Sensitivity profiling across 177 cancer cell lines flagged diffuse large B cell lymphoma and renal cell carcinoma as unusually dependent.

The transferases work by a different chemistry. Rather than donating electrons, glutathione S-transferases lower the pKa of the bound glutathione thiol to generate a thiolate, then position it to attack an electrophilic carbon, nitrogen or sulfur. The cytosolic classes alpha, mu, pi, theta and omega between them handle epoxides, quinones, alkyl halides and reactive metabolites of drugs and combustion products. The resulting conjugate is too polar to diffuse out, so multidrug resistance-associated protein transporters export it, and the gamma-glutamyl and glycine residues are then stripped and the cysteine acetylated to yield the mercapturic acid excreted in urine.

Two of these genes are commonly deleted. Homozygous null genotypes at GSTM1 and GSTT1 are found in a substantial fraction of the population, and they are among the most-studied polymorphisms in molecular epidemiology precisely because they remove an entire conjugation capacity rather than tuning it. That variation is one reason glutathione biology resists population-level generalisation.

Nothing about glutathione synthesis is limited by glutamate or glycine. Both sit well above the Km of their respective enzymes in a normal cell. Cysteine does not. Reported Km values for glutamate-cysteine ligase are around 1.8 mM for glutamate against an intracellular pool roughly tenfold higher, and 0.1 to 0.3 mM for cysteine against an intracellular pool that sits close to that figure. Whenever cysteine availability falls, the ligase slows in direct proportion, and the intermediate gamma-glutamylcysteine stays at vanishingly low concentration because the second enzyme is fast and is not feedback inhibited.

The ligase itself is regulated at three levels. Glutathione competitively inhibits the catalytic subunit with a Ki near 2.3 mM, which is inside the physiological range, so the pathway senses its own product. The modifier subunit GCLM, catalytically inert on its own, lowers the Km for glutamate and raises the Ki for that feedback, so the holoenzyme is substantially more productive than GCLC alone. And transcription of both subunits runs through Nrf2, which is held in the cytosol by Keap1 and degraded until oxidative or electrophilic stress releases it to bind antioxidant response elements in the promoters.

Genetics has bracketed how much of this is dispensable. Homozygous knockout of the mouse Gclc gene is lethal before gestational day 13. Heterozygotes are viable and fertile, show a gene-dose reduction in both GCLC protein and ligase activity, and yet carry only about 20 percent less glutathione, with a roughly 30 percent compensatory rise in ascorbate. At the other end of the pathway, inherited glutathione synthetase deficiency produces 5-oxoprolinuria: gamma-glutamylcysteine accumulates, is cyclised by gamma-glutamyl cyclotransferase to 5-oxoproline, and spills into urine as a metabolic acidosis that can be picked up on newborn screening.

This is the biochemistry behind the most common question about the molecule. N-acetylcysteine is not glutathione and does not act as glutathione; it is a cysteine delivery vehicle that is deacetylated intracellularly and feeds the rate-limiting substrate directly. Buthionine sulfoximine, the standard experimental tool for depleting glutathione, works from the opposite direction by inhibiting the ligase transition state. Between them they define the two levers the pathway actually responds to, and neither of them is exogenous glutathione.

Roughly 10 to 15 percent of cellular glutathione sits in mitochondria, with 80 to 85 percent in cytosol and a small remainder in the endoplasmic reticulum. Because the mitochondrial matrix is a much smaller volume, that fraction represents a matrix concentration comparable to or higher than the cytosolic one. It is also the pool that matters most for survival, since the respiratory chain is where most endogenous superoxide originates.

Meredith and Reed separated the two pools in isolated rat hepatocytes in 1982 using graded digitonin permeabilisation, releasing the cytosolic pool alongside lactate dehydrogenase and the mitochondrial pool alongside citrate synthase. The turnover figures they obtained were startling: a half-life of 30 plus or minus 3 hours for the mitochondrial pool against 2 plus or minus 0.1 hours for the cytosolic one, and incorporation of labelled methionine or cysteine into cytosolic glutathione running fifteen times faster than into the mitochondrial pool after correction for turnover. Depleting cytosolic glutathione with diethyl maleate to 40 percent of control, or to 10 percent with a glutathione reductase inhibitor added, left the mitochondrial pool untouched for an hour.

Griffith and Meister showed in 1985 that the reason for the separation is that mitochondria do not make glutathione at all. The matrix contains neither synthetic enzyme. Everything in there arrived from cytosol, through a process they described as slow net transport superimposed on faster exchange. The identity of the carrier stayed unresolved for another 36 years.

In 2021 the Birsoy laboratory at Rockefeller combined organellar proteomics with metabolomics and identified SLC25A39, an inner membrane carrier of previously unknown function, as the transporter. Loss of SLC25A39 cut mitochondrial glutathione without changing whole-cell glutathione at all, which is exactly the signature a dedicated importer should produce. Cells lacking both SLC25A39 and its paralogue SLC25A40 showed defects in the activity and stability of iron-sulfur cluster proteins, and mitochondrial import proved necessary for red blood cell development in mice. Expressing a bifunctional bacterial glutathione synthetic enzyme inside the matrix rescued the metabolic and proliferative defects, confirming that the phenotype was about matrix glutathione rather than about the transporter doing something else. A parallel CRISPR screen published in 2022 reached the same protein by an independent route and connected it to iron homeostasis and oxidative phosphorylation. Glutathione availability also turns out to control SLC25A39 protein abundance, so the import step is itself redox-regulated.

The reason this compartment matters clinically is the acetaminophen model. The reactive metabolite NAPQI consumes hepatic glutathione, and hepatocyte necrosis tracks the loss of the mitochondrial pool rather than the cytosolic one, because it is mitochondrial protein adduction and the resulting oxidant burst that commits the cell. Raising plasma glutathione does not obviously reach that pool, since anything circulating must first be broken down at the cell surface, resynthesised in cytosol and then imported through SLC25A39. It is a three-step barrier, and it is the reason a rise in blood glutathione is a weaker result than it sounds.

This is the most searched question about glutathione and the one where confident answers are least warranted. The classical negative result is from Bern in 1992: seven healthy volunteers received a single oral dose of 0.15 mmol per kilogram, about 3 g of glutathione, and plasma glutathione, cysteine and glutamate were followed for 270 minutes. None of the three rose significantly. Basal plasma glutathione was 6.2 micromol/L, cysteine 8.3 micromol/L and glutamate 54 micromol/L. The authors attributed the result to hydrolysis by intestinal and hepatic gamma-glutamyl transferase and concluded that systemic availability of oral glutathione in man is negligible.

A second negative came from a randomised placebo-controlled trial of 40 adults given oral glutathione twice daily for four weeks. Urinary F2-isoprostanes and 8-hydroxy-2-deoxyguanosine were unchanged, and so were erythrocyte reduced glutathione, oxidised glutathione and their ratio. Thirty-nine participants completed per protocol. On its own terms that study found nothing.

The positive result that changed the conversation was a six-month randomised, double-blinded, placebo-controlled trial in 54 non-smoking adults at Penn State, comparing two oral doses of glutathione, 250 mg and 1,000 mg taken daily. Blood glutathione rose at one, three and six months at both doses. At six months the high-dose group showed increases of 30 to 35 percent in erythrocytes, plasma and lymphocytes and 260 percent in exfoliated buccal cells, with the low-dose group up 17 percent in whole blood and 29 percent in erythrocytes. The whole-blood oxidised-to-reduced ratio fell. Natural killer cell cytotoxicity more than doubled against placebo at three months. Levels returned to baseline after a one-month washout, which is the internal consistency check that makes the result hard to dismiss.

A one-month pilot of a liposomal preparation in 12 healthy adults reported increases of 40 percent in whole blood, 25 percent in erythrocytes, 28 percent in plasma and 100 percent in peripheral blood mononuclear cells by two weeks, with a 35 percent fall in plasma 8-isoprostane. It also reported no difference between the two dose groups, was funded by the manufacturer of the liposomal product tested, had no placebo arm and enrolled twelve people. Those are the caveats the study itself lists.

The unresolved core is mechanistic rather than statistical. Nobody has shown, in humans, whether an oral dose raises body stores because intact tripeptide is absorbed and distributed, or because it is hydrolysed at the brush border into glutamate, cysteine and glycine which are then absorbed and used to synthesise glutathione de novo. Both routes predict a rise in blood glutathione after six months. Only the first would justify the way the molecule is usually described. Settling it needs a stable-isotope tracer study in humans that follows the intact molecule rather than the pool it might have fed, and that study has not been published. Anyone who tells you the question is closed in either direction is ahead of the data.

The delivery problem is why clinical work has kept trying other routes, with mixed results. A Phase IIb randomised placebo-controlled study in 45 people with Hoehn and Yahr stage 1 to 3 Parkinson's disease compared intranasal glutathione at 100 mg and 200 mg three times daily against saline for three months. The high-dose group improved on total Unified Parkinson's Disease Rating Scale by 4.6 points and on the motor subscore by 2.2 points against its own baseline, but neither active arm beat placebo, and the placebo response was larger than in previous Parkinson's studies. One high-dose participant developed cardiomyopathy. Glutathione is also a licensed pharmaceutical in Japan, marketed under the name Tathion, while it carries no equivalent marketing authorisation in North America, where it circulates as a supplement or through compounding.

Three mechanisms are proposed for the depigmenting effect. Glutathione inhibits tyrosinase, the copper enzyme that commits tyrosine to melanin synthesis. It shifts the branch point in melanogenesis away from dark eumelanin toward lighter pheomelanin, since the pheomelanin route requires a thiol. And it scavenges the reactive oxygen species that stimulate melanocytes in the first place. There is a mechanistic catch that vendor material almost never mentions: unmodified glutathione crosses melanosomal membranes poorly, and esterified derivatives such as the monoethyl ester penetrate substantially better in cell work, which suggests the free tripeptide is not the ideal molecule for this particular target.

The human evidence is four small trials. A randomised, double-blind, placebo-controlled study in 60 Thai medical students aged 19 to 22 gave oral glutathione for four weeks in two divided doses and found melanin index reduced at all six measured sites, statistically significant at two, the right side of the face and the sun-exposed left forearm. An open-label single-arm pilot in 30 Filipino women aged 22 to 42 with Fitzpatrick types IV and V used a buccal lozenge for eight weeks and reported significant reductions at all sites with 90 percent reporting moderate lightening on global assessment. A randomised double-blind split-face study in 30 Filipino women applied 2 percent oxidised glutathione lotion twice daily for 10 weeks and found a significant fall in melanin index plus improvements in moisture and texture. A three-arm randomised trial in Bangkok gave 250 mg of reduced or oxidised glutathione orally for 12 weeks and found melanin index and ultraviolet spots trending below placebo, with reduced wrinkles at some sites and increased elasticity.

The intravenous route is where the picture turns. The one published controlled study, in 50 Pakistani women aged 25 to 47 given 1,200 mg intravenously twice weekly for six weeks against saline, found minimal efficacy on a visual hyperpigmentation scale, with improvement rates of 37.5 percent versus 18.7 percent at end of treatment falling to 6.2 percent versus zero at six months. Adverse effects were not minor: liver dysfunction in 32 percent of treated participants, eight patients in total, and one case of anaphylactic shock. Dropout was 36 percent. Reviewers have described the design and the statistical reporting as inadequate.

Regulators reached the obvious conclusion. The Philippine Food and Drug Administration issued a public advisory in May 2011 against intravenous glutathione for skin lightening, citing Stevens-Johnson syndrome and toxic epidermal necrolysis, severe abdominal pain, thyroid and renal dysfunction, and life-threatening complications including air embolism and sepsis arising from improper administration or counterfeit product. The Philippine Dermatological Society issued a corresponding position statement. The 2018 review that gathered this literature titled itself around the question of whether skin lightening by glutathione is a regnant myth or an evidence-based verity, and concluded that the intravenous case rests on a single study with a dubious design, while the oral and topical effects are modest, reversible on discontinuation and of unknown duration.

One further route-specific finding is worth recording because it is counterintuitive. Nebulised glutathione, tested as an airway antioxidant, provoked bronchoconstriction in patients with mild asthma. In a randomised double-blind crossover study of eight patients, 600 mg of nebulised glutathione produced a 19 percent fall in FEV1 and a 61 percent rise in total pulmonary resistance against negligible change on saline, with cough in four and breathlessness in three. Salbutamol pretreatment blocked it, and the response correlated with metabisulfite sensitivity rather than methacholine sensitivity, implicating sulfite generated from the thiol. A molecule that is protective inside a cell is not automatically benign delivered onto an epithelium.

The free thiol that makes glutathione useful also makes it fragile outside a cell. In neutral aqueous solution the sulfhydryl autoxidises to the disulfide, and the reaction is catalysed by trace transition metals, particularly copper and iron, that are present in ordinary water and glassware at concentrations sufficient to matter. Oxygen exposure, alkaline pH and warmth all accelerate it. The lyophilized solid is comparatively stable because there is no solvent to carry the metal ions and no mobility for the bimolecular step, which is why the compound is supplied as a powder rather than a solution.

Two variables control the rate, and neither is usually stated. The reactive species is the thiolate, not the thiol, so the oxidation rate is pH dependent: the cysteine sulfhydryl of glutathione has a pKa of roughly 8.8 to 9.2 depending on the method, which means the molecule is markedly more stable in acidic solution and degrades quickly as pH approaches neutral and above. And the reaction is metal-catalysed, so a chelator suppresses it while trace copper from glassware or unpurified water accelerates it. A solution held at pH 7.4 in a metal-contaminated buffer and left open to air is running the worst case on both counts.

This has a direct analytical consequence that trips up published numbers as well as bench work. Because sample preparation itself oxidises glutathione, measured GSSG is systematically inflated unless free thiols are blocked at the moment of collection. Rossi and colleagues quantified the problem in human blood and in cannulated rats, comparing HPLC measurement with and without N-ethylmaleimide pretreatment: with artefacts prevented, GSSG rose up to threefold over basal on oxidant challenge and glutathionylated protein by around 50 percent, while protein carbonyls and malondialdehyde in the same samples barely moved. The redox forms of glutathione are more sensitive markers of oxidative stress than the classical lipid and protein damage markers, but only when the preanalytical step is controlled.

Characterisation of the material itself relies on a small standard set. Reversed-phase HPLC with ultraviolet detection near 210 nm resolves the tripeptide and separates it from the disulfide, which elutes distinctly and is the usual impurity of interest. Ellman's reagent, 5,5'-dithiobis(2-nitrobenzoic acid), gives a direct colorimetric readout of free thiol at 412 nm and is the simplest check that a lot has not oxidised. Mass spectrometry confirms identity at m/z 308.09 for the protonated monoisotopic species and 613.16 for protonated GSSG. Nuclear magnetic resonance distinguishes the gamma linkage from the alpha isomer unambiguously, because the glutamate side-chain protons shift on amidation in a pattern the alpha-linked isomer cannot reproduce.

Two practical points follow for anyone comparing certificates of analysis. A purity figure by HPLC area percent says nothing about oxidation state unless GSSG is separately quantified, since both species are peptide-related and both absorb. And a total glutathione assay that includes a reducing step before measurement will report the disulfide as though it were reduced glutathione, so the two numbers are not comparable and should never be read as though they were.

One production fact is worth correcting because it is stated backwards so often. Bulk glutathione is not made by solid-phase peptide synthesis. The commercial routes are fermentation, typically with Saccharomyces cerevisiae or Candida utilis strains selected for high intracellular accumulation, and enzymatic synthesis using glutamate-cysteine ligase and glutathione synthetase or a bifunctional bacterial enzyme that performs both steps. Standard Fmoc chemistry builds alpha-linked backbones; the gamma linkage that defines this molecule is exactly what that chemistry does not produce without side-chain protection strategies that the biological routes make unnecessary.