Glutathione, L-γ-glutamyl-L-cysteinylglycine, is found in most cells in high concentration (0.1 to 10 mM). It has an unusual γ-peptide bond that prevents its hydrolysis by all but one peptidase and a thiol moiety that is key to its physiological and biochemical functions. Glutathione participates as a coenzyme in several reactions and is involved in maintaining thiol/disulfide balance. Additionally, it plays a key role in the protection against toxic compounds through nonenzymatic reactions and those catalyzed by reduced glutathione (GSH) S-transferases and GSH peroxidase. A product of the peroxidase reaction, glutathione disulfide (GSSG), is normally reduced by glutathione disulfide reductase using NADPH. Normally there is little cellular GSSG, and the ratio of GSH to GSSG is high. Glutathione has many cellular functions, and the study of these has been aided by genetic mutants of enzymes involved in GSH metabolism and by the use of inhibitors of GSH metabolism.1-5
Six enzymes form the γ-glutamyl cycle, which catalyzes the synthesis and degradation of GSH (Fig. 96-1). The γ-glutamyl cycle was identified by Meister and coworkers in 1970.6 Since then, a large body of experimental evidence supports its function in vivo.3,7 Although many tissues and cells have all of the γ-glutamyl cycle enzymes, some cells, such as the erythrocytes, which lack a nucleus, lack γ-glutamyl transpeptidase and 5-oxoprolinase. Glutathione is synthesized intracellularly and exported from most cells. Tissues have very different rates of GSH turnover. For example, rat kidney, liver, lymphocyte, lung, and skeletal muscle have GSH half-times of about 20 min, 60 min, 25 min, 45 min, and 16 h, respectively; the reasons for such differences in GSH turnover are not known. The γ-glutamyl cycle plays a role in amino acid transport, and it is likely that it has an important role in the salvaging of cysteine and GSH moieties and for maintaining cellular thiol and GSH status.
With the exception of some gram-positive bacteria,8 GSH is synthesized in all cells from glutamate, cysteine, and glycine by the sequential actions of γ-glutamylcysteine synthetase (GCS) and glutathione synthetase (GS). Both enzymes are cytosolic and use ATP, which is hydrolyzed to ADP and Pi.
γ-Glutamylcysteine Synthetase (GCS)
GCS (glutamate cysteine ligase; EC 6.3.22) catalyzes the first and rate-limiting step in the biosynthesis of GSH, the formation of γ-glutamylcysteine:
The enzyme has been purified or partially purified from many sources, from E. coli to human cells.3
Rat kidney is the source of the most highly purified and characterized GCS.9,10 The enzyme is active toward several l-glutamate and l-cysteine analogs. Because of spontaneous oxidation, which causes problems in in vitro assays, l-α-aminobutyrate is often used instead of cysteine as the acceptor amino acid. Rat GCS is a heterodimer of separately encoded subunits (molecular weight 73 kDa and 31 kDa).9,10 Human GCS is also a heterodimer and is highly homologous to rat GCS.11 The heavy subunit (73 kDa) has catalytic activity, but low affinity for glutamate; this catalytic subunit (28 kDa) is more susceptible to feedback regulation by GSH.9 The light subunit, which by itself is not catalytically active, increases the affinity (lowers the Km) of the heavy subunit for glutamate and decreases the inhibition by GSH (increases the Ki), and is thus regulatory. The light subunit has 6 cysteine residues and the heavy subunit has 14, one of which is needed for activity. The two subunits are associated by strong hydrophobic interactions. They can also interact through disulfide bond formation, which increases the affinity of the enzyme for glutamate.10 The disulfide form of the enzyme is likely to be present when GSH levels are low and there is oxidative stress. Thus, GCS activity is controlled by cellular needs for GSH. Also GCS is feedback inhibited, that is, non-allosterically, by GSH at physiological concentrations suggesting that in vivo the enzyme is not fully active but can respond rapidly to changing cellular needs for GSH.9,12 In addition to regulation of GCS activity by GSH and by disulfide formation between the heavy and light subunits, GCS activity may be inhibited by phosphorylation.13
The enzyme GCS is inhibited by a number of compounds, which are reactive with thiols.2 Such inhibitors are cystamine, l-2-amino-4-oxo-5-chloropentanoic acid, and l- and d-3-amino-1-chloro-2-pentanone, all of which interact with a thiol near the active site. d-γ-Methylene glutamate irreversibly inhibits the enzyme by addition to an enzyme thiol moiety. The enzyme is also inactivated by both l- and d- isomers of S-sulfocysteine and S-sulfohomocysteine, which bind tightly to the enzyme and are stabilized by a thiol near the active site. l-Glutamate protects against inactivation by all of these compounds, which suggests that there is an essential thiol at or near the active site.9,10,14
Methionine sulfoximine (MSO) inhibits both GCS and glutamine synthetase.15 Meister and colleagues showed that glutamine synthetase is inhibited by l-methionine-S-sulfoximine phosphate, a transition state inhibitor. Glutamine synthetase is selectively inhibited by α-ethyl MSO, whereas GCS is inhibited by longer amino acid sulfoximines, such as prothionine sulfoximine and buthionine sulfoximine (BSO). These studies suggest that like glutamine synthetase, GCS functions by forming an enzyme-bound γ-glutamyl phosphate intermediate.16,17
GCS activity may be regulated in several ways. The availability of substrates, especially of cysteine, may be limiting. GCS activity is also regulated by feedback inhibition by GSH and disulfide interaction of the subunits. The activity of the catalytic subunit has been reported to be decreased by phosphorylation (Sekhar, 1999; Sun et al, 38). The ratio of the regulatory to the catalytic subunit will affect enzyme catalytic efficiency. The amount of enzyme in tissues and its transcriptional upregulation are important for capacity to synthesize GSH. There has been much interest in transcriptional regulation of GCS subunits.
GCS is a phase II enzyme, a family of inducible enzymes involved in detoxification of electrophiles that includes, for example, GSH S-transferases and NAD(P)H: quinine reductase (NQO1) (Dinkova-Kostova et al, 2001). Phase II enzyme regulation is of importance for prevention of cancer and may offer targets for chemotherapy. Inducers of oxidative stress and compounds that deplete GSH levels upregulate both subunits of GCS, as well as other phase II enzymes (for example, see the following reviews and references therein: Dickinson et al, 2003; Griffith, Mulcahy, 1999; Kondo et al, 1999; Krzywanski et al, 2004; Wild, Mulcahy, 2000). The regulation of constitutive and inducible expression of GCS is complex and still evolving.
The human gene for the heavy or catalytic subunit of GCS, called GLCLC, has been localized to chromosome 6p12111 (Gipp, Mulcahy, 2000). The promoter for GCS has been characterized from rat, mouse, and humans (Galloway et al, 1999; Rahman et al, 1998; Mulcahy, Gipp, 23). The 5′ flanking region of the gene for the human catalytic subunit has multiple transcription start sites (two of which are major), a TATA box, a CAAT box, and GC-rich regions. Consensus sequences for activator protein-1 (AP-1) and AP-1-like binding sites, a putative AP-2 site, metal responsive element (MRE), and antioxidant response element (ARE)/electrophile response element (EpRE) (ARE/EpRE) have been identified (Mulcahy, Gipp, 23). In addition to AP-1/TRE [TPA (tetradecanoyl phobol-13-acetate) response element], other cis elements identified include NF-κB and ARE-like (ARE4) elements (Galloway et al, 1999; Kondo et al, 1999; Mulcahy et al, 1997; Rahman et al, 1998).
The gene for the light or regulatory subunit, called GLCLR, has been localized to chromosome 1p21112 (Rozet et al, 1998). The regulatory, or light, subunit of GCS, like the catalytic subunit, is inducible. The 5′ flanking region is reported (Galloway et al, 1997; Moinova et al, 24) not to have a TATA box or binding sequences for NF-κB, but rather to be rich in GC with several putative Sp-1, AP-1, AP-1-like, MRE, and ARE/EpRE sites. Either AP-1 or EpRE may be used for constitutive expression, while EpRE is important for the induction of the regulatory subunit of GCS. Recent studies show that one of the two ARE sites has a variant sequence (Erickson et al, 2002).
Trans-Activating Factors of GCS
Inhibitors of the activation of mitogen-activated protein kinase (MAPK) family members have been implicated in transcriptional upregulation of both subunits of GCS (Zipper, Mulcahy, 2002). These studies suggest that the binding of Nrf2 plus JunD to GCS EpREs is a downstream consequence of extracellular signal-regulated protein kinase (ERK) and p38 pathways. Additional studies (Miyahara et al, 2002) show that bovine serum albumin with Nε-(carboxymethyl) lysine adducts (CML-BSA) induces the catalytic subunit of GCS through AP-1 via the protein kinase C, Ras/Raf-1, and ERK1/2 pathway.
Both subunits of GCS have AP-1 and EpRE/ARE binding sites and are induced by stresses that induce other phase II enzymes. EpRE, which may have embedded or nearby AP-1 sites, controls the basal and inducible expression of the catalytic subunit (Wild et al, 1999; Wild et al, 2000). Only the inducible expression of the regulatory subunit is under the control of EpRE/ARE; the basal expression of the regulatory subunit is influenced by a nearby AP-1 binding site (Moinova et al, 1999; Zipper et al, 2000). The sequence of EpRE is similar to NF-E2 and T-MARE, and trans-acting factors that bind to those have been shown to bind to EpRE (Wild et al, 1999; Zipper, Mulcahy, 2000) and upregulate GCS expression.
Nuclear factor E2 p45-related factor 2 (Nrf2), cJunD, and small Maf (MafK, MafG, MafF) proteins form complexes that bind to the EpREs of the catalytic and regulatory subunit of GCS (Wild et al, 1999). Nrf2 (a member of the Cap ’n Collar subfamily of bZip transcription factors) plus c-Jun heterodimer is required for upregulation, but high concentrations of c-Jun downregulate the expression of the catalytic subunit of GCS (Jeyapaul, Jaiswal, 2000). The latter is suggested to be due to the formation of a c-Jun plus c-Fos complex that blocks the binding of the Nrf2 plus c-Jun complex to EpRE. Nrf2 forms heterodimers with bZIP proteins and binds to EpRE to regulate GCS expression. Nrf2 knockout mice [nrf2 (-/-)] have decreased mRNA for both catalytic and regulatory subunits of GCS (Chan, Kwong, 2000). The overexpression of Nrf2 cDNA in nrf2 (-/-) fibroblasts resulted in higher levels of the transcripts for the subunits and higher GSH levels. Nrf2 (-/-) mice are more sensitive to carcinogenesis and are not protected by inducers (Ramos-Gomez et al, 2001).
Nrf2 heterodimer binding to EpRE/ARE regulates the expression of GCS. Normally Nrf1 is located in the cytosol attached through its N-terminal to the actin-binding protein Kelch-like ECH-associated protein 1 (Keap1; a cysteine-rich protein and key for regulating the induction of phase II enzymes) (Dinkova-Kostova et al, 2002; Itoh et al, 1999). Inducers release Nrf2 from the complex with Keap1, and Nrf2 translocates to the nucleus to stimulate transcription of phase II enzymes (Itoh et al 1999).
Further studies on the transcription regulation of GCS will provide information on the role of GCS in carcinogenesis and prevention, chemotherapy, and diseases, as well as a deeper understanding of the cellular roles of GSH.
Polymorphisms have been reported in the promoter region of both the catalytic and regulatory subunits of GCS. A trinucleotide repeat has been reported in the 5′ untranslated region of the catalytic subunit of human GCS (Walsh et al, 1996). Studies of a panel of tumor cell lines between alleles and drug sensitivity are noted (Walsh et al, 2001). Ethnic diversity and 11 polymorphisms also were noted in the catalytic subunit (Willis et al, 2003). A polymorphism also has been reported in the 3′ untranslated region of the catalytic subunit of GCS (Buetler et al, 1999). A polymorphism in the catalytic subunit promoter (-129 C/T), where the -129T allele is associated with decreased (>50%) promoter response to hydrogen peroxide, decreased endothelium-dependent coronary artery dilation, and increased risk for myocardial infarction (Koide et al, 2003). The same group also reported that a polymorphism (-588 C/T) in the promoter region of the regulatory subunit is associated with decreased promoter activity in luciferase reporter gene assays, and this -588T allele is a risk factor for myocardial infarction (Nakamura et al, 2002). Polymorphisms in GCS genes may have important functional significance. For example, a single-nucleotide polymorphism may prevent induction of a GCS subunit in response to an electrophile or reactive oxygen species, thus altering susceptibility to carcinogens or increasing the risk of other diseases (Ramos-Gomez et al, 2001; Talalay et al, 2003).
Glutathione synthetase (GS) (OMIM 601002) is widely distributed in mammalian tissues and catalyzes the formation of GSH.30 The enzyme has been purified from a number of sources including E. coli, yeast pigeon liver, plants, rat kidney, human brain, and erythrocytes. Studies on yeast GS, using pulse labeling and chemical intermediate synthesis, showed that GS catalyzes the formation of an acyl phosphate intermediate, γ-glutamyl-α-aminobutyryl phosphate.30 Thus, GS, like GCS, proceeds via an acyl phosphate intermediate:
The most widely studied and first cloned mammalian GS is the rat kidney enzyme.31 This enzyme is a homodimer with subunits of 52 kDa. It has about 2 percent carbohydrate content,32 but it is not required for enzyme activity. The human enzyme33 has 88 percent amino acid identity and 98 percent amino acid similarity to the rat enzyme.
Recent studies on human GS have shown that mutations altering Cys-422 lead to decreased enzyme activity.34 Comparison of human and rat GS shows that there is a highly conserved domain of 13 amino acids that may be important in substrate binding.35 Furthermore, mutations altering some residues in this domain lead to dramatic loss of enzyme activity [Bhansali et al 2003]. The enzyme exhibits negative cooperativity for both γ-glutamylcysteine and γ-glutamyl-α-aminobutyrate, that is, the binding of the first molecule of γ-glutamyl substrate diminishes the affinity for the second molecule.35 Kinetic analysis revealed departure from linearity of the Lineweaver-Burk double reciprocal plot for the binding of -glutamyl substrate, indicating negative cooperative binding [Njalsson et al, 2001]. This may reflect changes in subunit interaction or altered kinetic mechanism. Additionally, a large conformational change has been reported to occur during the catalytic cycle of the yeast enzyme (38% identical; 59% similar to mammalian GS) [Gogos and Shapiro 2002].
GS deficient yeast are reported to grow slowly [Grant et al, 1997]. The levels of GS have been reported to be decreased in Downs syndrome [Gulesserian et al, 2001] and after surgical trauma [Luo et al, 1998].
Human GS has an ATP grasp binding fold and is the first mammalian member of the glutathione synthetase ATP-binding superfamily, whose members include D-ala:D-ala ligase, biotin carboxylase domain of carbamoyl phosphate synthetase, enzymes in the purine biosynthetic pathway, the N terminal domain of pyruvate phosphate dikinase, and the C terminal domain of bovine brain synapsin Ia [Murzin et al, 1995; Andreeva et al, 2004]. When conserved residues in ATP grasp binding fold are mutated there is substantial loss of enzyme activity [Dinescu et al, 2004].
There is only GS structural gene in the human genome and it maps to chromosome 20.q 11.2.36 Studies on drug resistant human tumor cells show that GS is not induced, even when GCS transcription is strongly induced,18 suggesting that there are different response elements in the promotors of GCS and GS, and that there is sufficient GS activity to form GSH. The murine GS gene encodes six transcripts each with different 5′ ends.37 Mouse GS is reported to be induced by the metal chelator 1,10-phenanthroline.38 Partial hepatectomy or treatment of cultured hepatocytes with thioacetamide, BSO, diethyl maleate and t-butylhydroquinone is reported to increase GS expression [Huang et al, 2000]. Deficiencies of folate, vitamin E and apolipoprotein E have recently been reported lead to increased transcription of GS and also of its activity in a mouse model [Tchantchou et al, 2004].
Four enzymes catalyze the degradative reactions of the γ-glutamyl cycle (Fig. 96-1). The activity of each of these enzymes varies in different tissues. For instance, erythrocytes lack γ-glutamyl transpeptidase (GT) and OP, and thus GSH turnover in erythrocytes depends on export of GSH to other tissues with a complete set of degrading enzymes.
GT (E.C. 184.108.40.206) is a membrane-bound enzyme whose active site faces the external side of a cell (ectoenzyme), and it is the only enzyme that cleaves the γ-glutamyl bond of GSH. The GT activity is found in many plants and microorganisms, as well as in mammals.1 In mammals, it is found in cells that have secretory or absorptive functions. The highest levels are found in kidney, pancreas, and intestine, but it is also found in spleen, liver, epididymis, lymphoid cells, choroid plexus, and ciliary body.1 A soluble form of GT is found in bile and plasma, and its levels in plasma are used clinically as a marker of liver function. The enzyme is localized to microvilli of epithelial cells in the kidney proximal tubules, and to the jejunum, bile duct, retinal pigment epithelium, bronchioles, thyroid follicles, pancreas acinar and ductile epithelial cells, the canalicular portions of hepatocytes, and seminal vesicles.39
γ-Glutamyl transpeptidase catalyzes the coupled hydrolysis of the γ-glutamyl bond of GSH (and other γ-glutamyl compounds) with the transfer of the γ-glutamyl moiety to acceptor molecules:
The acceptor is usually an amino acid, yielding a γ-glutamyl amino acid, but it may also be water or GSH. When the acceptor is water, GSH is hydrolyzed to glutamate and cysteinylglycine. Maleate and hippurate increase the hydrolysis activity of GT.39 Bile acids and conjugated bile acids also modulate the hydrolysis and γ-glutamyl transfer activity of GT.1 When GSH is the acceptor, γ-glutamyl GSH is formed.40 The enzyme has a γ-glutamyl intermediate and a ping-pong mechanism.1
Gene Location and Structure of GT
The human GT gene family is composed of at least seven different loci,42 several of which are located on the long arm of chromosome 22.43 Sakamuro and colleagues (1988) reported the primary structure of human GT based on studies of a cDNA. The enzyme consists of two highly glycosylated peptide chains, heavy and light, composed of 351 (Mr 38,336) and 189 (Mr 20,000) amino acids, respectively. Both are coded by a single gene; the two subunits of the mature enzyme are the products of processing of the single precursor peptide.41 The active site of GT is located in the light subunit of the mature enzyme. The enzyme is membrane-bound with its active site facing the external side of the cell. The heavy subunit has an N-terminal membrane anchor and is noncovalently linked to the light subunit.
The transcriptional organization of GT is complex. There are multiple regulatory sites that allow for differential expression, and different 5′ untranslated sequences that could create different transcript stability and translation efficiency (Chikhi et al, 1999).
In contrast, rodent GT is encoded by a single copy gene with multiple promotors, some of which are regulated by oxidative stress.46-48 Both the human and the rat GT mRNAs are differentially expressed in tissues and during development.47,49-54 One rat promotor has a negative regulatory region that is similar to the silencer element in the GSH S-transferase gene.55
The active site of GT faces the outside of the cell and has three active subsites. The γ-glutamyl donor site has a broad specificity using l- and d- γ-glutamyl donors.1,39 The cysteinylglycine site binds amino acids and dipeptides (in the l-configuration or glycine) and has two subsites. The best acceptor amino acids are the neutral amino acids, whereas branched chain amino acids are weak acceptors, and d-amino acids and l-proline are inactive.39 The glycine subsite prefers dipeptides with glycine in the C-terminal position.3
γ-Glutamyl transpeptidase is inhibited by l- or d-serine and borate, which forms a transition state inhibitor-complex with an enzyme hydroxyl group.56 The enzyme is also inhibited by a variety of compounds, such as γ-glutamyl derivatives and glutamine antagonists, and such inhibition is decreased by γ-glutamyl compounds, such as GSH and glutamine.39 Site-specific mutagenesis and inhibitor-binding studies suggest that several amino acid hydroxyl groups and charged amino acids are important for GT catalysis (Taniguchi et al, 45). These studies45,57-59 suggest that the active site of the enzyme is at an interface between the heavy and light subunits.
Mutations have yet to be identified in patients with GT deficiency; however, there are many possible locations for alternations.
γ-Glutamylcyclotransferase (GCT) (GCT; E.C. 220.127.116.11) catalyzes the hydrolysis of the γ-glutamyl bond of certain γ-glutamyl compounds to form free amino acid and 5-oxoproline, a cyclic form of glutamate:
GCT activity is present in many tissues, including kidney, liver, testes, spleen, brain, lung, heart, thymus, thyroid, skeletal muscle, skin, and adrenal gland.1 The GCT peptide is 27 kDa and contains 7 half cystines.1
Although GCT catalyzes the hydrolysis of l-γ-glutamyl-l-amino acids, it does not catalyze the hydrolysis of the γ-glutamyl bond of GSH or GSSG.1 It has a broad specificity toward l-γ-glutamyl-l-amino acids; many l-γ-glutamyl-l-amino acids are substrates except for the aromatic, branched chain, and proline l-γ-glutamyl derivatives. The enzyme is not active toward d-γ-glutamyl derivatives. The substrate specificity of GCT parallels that of γ-glutamyl transpeptidase. Studies using NMR on whole erythrocytes suggest that one product, 5-oxoproline, is an inhibitor of GCT (York et al, 1989)
A similar enzyme, γ-glutamylamine cyclotransferase, catalyzes the formation of 5-oxoproline and amine compound from γ-glutamylamines, such as Nε-(γ-glutamyl)lysine formed during the degradation of glutaminyl-lysine crosslinked proteins and peptides (Danson et al, 2002, Fink et al, 1983). The specificity of the enzymes does not overlap with GCT. Interestingly, Nε-(γ-L-glutamyl)-L-lysine levels are increased substantially in Alzheimer brains and in the cerebrospinal fluid of patients with Huntington’s disease (Cooper et al, 1999; Jeitner et al, 2001; Kim et al, 1999). One or both cyclotransferases may be important for the production of 5-oxoproline, which has been implicated in regulation of amino acid (Vina et al, 1989) and energy metabolism (see GS deficiency, below), including 5-oxoprolinuria without GS deficiency (see below).
Membrane-bound dipeptidase (MBD) (E.C. 18.104.22.168; previously E.C. 22.214.171.124) is the enzyme that hydrolyzes dipeptides, including S-substituted L-cysteinylglycine compounds, such as the oxidized GT product cystinyl-bis-glycine and the conversion of leukotriene D4 to E4 (Anderson et al, 1982). Because MBD also hydrolyzes certain β-lactam antibiotics, there has been much interest in it. MBD has been found in most tissues. It is a homodimer and is anchored to the membrane by a glycosylphosphatidylinositol anchor. The MWs are 42 kDa and 63 kDa for unglycosylated and glycosylated, respectively. Several residues have been identified as being important for dimerization or catalysis (Hooper et al, 1997; Keynan et al, 1997). The crystal structure of free and inhibitor-bound human MBD has been reported (Nitanai et al, 2002). Two N-linked glycosylations were identified, as were two zinc atoms (per monomer). It is an α/β barrel with structural architecture similar to other metallohydrolases. Renal MBD has been mapped to human chromosome 16 at q24 (Nakagawa et al, 1992). A deficiency has not yet been reported for this enzyme.
A mouse has at least four distinct membrane dipeptidase mRNA transcripts that result from the use of two promoters and two different poly(A) addition sites.60
5-Oxoprolinase (OP; E.C. 126.96.36.199) catalyzes ATP-dependent ring-opening of 5-oxo-l-proline (pyroglutamic acid; pyrrolidone carboxylate) to yield glutamate.1
5-Oxoproline can be formed by the action of GCT on γ-glutamyl compounds, enzymatic degradation of proteins with N-terminal 5-oxoprolyl residues, by the action of γ-glutamylamine cyclotransferase on γ-glutamylamines, such as those formed during the degradation of glutamine-lysine transglutaminase crossed-linked proteins or peptides, and by nonenzymatic formation from glutamine as well as from dietary sources, such as tomato juice and some processed foods, and intestinal microorganism flora.60
The enzyme is found in microorganisms, plants, and most mammalian tissues, except erythrocytes.1 OP is also found in porcine brain microvessels and brain microvessel endothelial cells, important for the blood brain barrier (Weber et al, 1999). The activity of OP in cells is low relative to most of the other enzymes of the γ-glutamyl cycle, but normally little 5-oxoproline is found in tissues or urine. The mammalian enzyme is not well studied, but the rat kidney enzyme is composed of two apparently identical 142-kDa subunits, each with several thiol moieties. Studies on the bacterial enzyme suggest that the enzyme forms a 5-oxoproline phosphate intermediate that couples the hydrolysis of the “internal peptide bond” of 5-oxoproline to the hydrolysis of ATP.63
The substrate specificity of OP is narrow.1 5-Oxo-D-proline is neither a substrate nor an inhibitor of OP. While ATP is the preferred nucleotide, OP will use dATP and other nucleoside triphosphates (NTPs) to a lesser extent. The enzyme requires both K+ and Mg2+ for activity. An interesting substrate is l-2-oxothiazolidine-4-carboxylate (OTC) (see below, Cysteine Delivery), which is opened to form intracellular cysteine. The best competitive inhibitor of OP is l-2-imidazolidone-4-carboxylate.
The level of OP in tumors is somewhat lower than match normal tissues; thus OTC may help protect normal vs tumor tissue during chemotherapy (Chen et al, 1998; Meister, 1983; Srivenugopal et al, 1997).
Experimental Deficiency of γ-Glutamyl Cycle Enzymes
A number of methods have been used to study the roles of the individual enzymes of the γ-glutamyl cycle, from cultured cells and tissue slices to rodent models.1,15 These studies have given insights into the functions of GSH and provided models for testing potential therapies for deficiency of GSH and GSH-related enzymes, as well as for chemoprevention.
Experimental γ-Glutamylcysteine Synthetase Deficiency
Several compounds have been used to decrease cellular GSH levels; however, many of these compounds (such as diethyl maleate) are nonspecific. As discussed above, GCS and glutamine synthetase are both inhibited by MSO. Administration of MSO to rodents leads to decreased tissue levels of GSH and glutamine. In addition, BSO, a specific inhibitor of GCS, is transported into many tissues, except across the blood-brain and blood-testis barriers. Glutathione levels decrease in tissues as GSH is used for cellular functions and exported in the absence of GSH synthesis.1 The levels of GSH decreases in rodent tissues after a few doses of BSO to 10 to 20 percent of control. A series of experiments showed that most of the remaining GSH is present in mitochondria. Mitochondria do not synthesize GSH, but take it up from the cytosol.64 After long-term administration of BSO to rodents, even mitochondrial GSH is depleted primarily because the electron transport system leaks free radicals, and, eventually, the mitochondria swell and lyse.2 Long-term BSO administration leads to damage of several organs, including lung, muscle, cerebral cortex, liver, jejunum, colon, and proximal renal tubule, but not heart and stomach; also plasma levels of triglycerides and cholesterol become elevated.2 Long-term BSO treatment provides a model of endogenous oxidative stress.65 Guinea pigs and newborn rodents, like humans, cannot synthesize vitamin C, and thus may be useful models for understanding human endogenous oxidative stress. Studies show that administration of vitamin C to rodents can prevent mitochondrial GSH deficiency caused by BSO treatment and that elevating cellular GSH can delay scurvy in vitamin C-deficient rodents.66-69 Thus, GSH “spares” vitamin C and vitamin C “spares” GSH.
Administration of BSO to rodents increases the toxicity of melphalan, BCNU (1.3-bis(2-chloroethyl)-1-nitrosourea), cisplatin, cadmium ions, mercury ions, and monocrotaline, as well as of radiation. Furthermore, BSO treatment blocks the activation of human T lymphocytes.4 Such studies emphasize the importance of GSH in protection mechanisms. In addition, BSO is being tested in clinical trials for cancer therapy70 (Bailey, 1998).
Human colon cancer cell lines transfected with a hammerhead ribozyme to the catalytic subunit of GCS have deceased GSH levels, decreased efflux of cisplatin-GSH adduct, and decreased expression of ABC transporters (Iida et al, 2001).
γ-glutamylcysteine synthetase knockout mice have been developed for both the heavy and light subunit (Dalton et al, 2000; Shi et al, 2000; Yang et al, 2002). The knockout of the heavy subunit shows that heterozygous mice have lower (20%) liver GSH and increased (30%) liver ascorbic acid, while the homozygous embryos fail to gastrulate and they die before day 8.5 of gestation (Shi et al, 2000). This mouse model has complete GSH deficiency, and the authors conclude that lethality is caused by apoptosis rather than by a reduction in cell proliferation. However, cell lines from homozygous mutant blastocysts grow indefinitely in GSH-free medium supplemented with N-acetylcysteine as a replacement of GSH. The authors conclude that GSH is needed for mammalian development but dispensable in cell culture, and the functions of GSH--not GSH itself--are essential for cell growth (Shi et al, 2000). The knockout mice of the light subunit are viable and fertile and have no overt phenotype (Yang et al, 2002). They do, however, have low GSH levels (9-16 percent of littermate controls) in several tissues, as well as GCS activity (catalytic subunit) with a twofold increase in Km for glutamate, in accord with in vitro studies. Fetal fibroblasts with a knockout of the light subunit were strikingly more sensitive to chemical oxidants such as H2O2. These knockout mice could serve as a model system for the study of chronic GSH deficiency and enhanced sensitivity to oxidative stress (Yang et al, 2002) even though these mouse models are unlike humans and are capable of synthesizing ascorbate.
Experimental Glutathione Synthetase Deficiency
No selective inhibitor of mammalian GS is available, and so far no knockout mouse model is available. Accordingly, there is no pharmacologic model of human GS deficiency. This is unfortunate because such a model would facilitate the testing of therapies for patients with GS deficiency.
Experimental γ-Glutamyl Transpeptidase Deficiency
Administration to mice of acceptor amino acids, such as glycylglycine, stimulate transpeptidation and lead to decreased cellular levels of GSH. When an inhibitor of GT is given prior to the amino acid, GSH levels do not decrease.15 Administration of effective inhibitors of GT to mice leads to glutathionuria and glutathionemia. After acivicin treatment, urinary thiols (GSH, γ-glutamylcysteine, cysteine) increased; these same thiols were found in the urine of a GT-deficient patient.71
γ-Glutamyl transpeptidase knockout mice have been made.72 Recent studies using these mice showed that GT deficiency leads to glutathionuria, glutathionemia, growth failure, cataracts, lethargy, shortened life span, and infertility73 (Chevez-Barrios et al, 2000; Held et al, 2003). A closer study of the reproductive phenotype of GT-deficient mice show that they are hypogonadal and infertile (Kumar et al, 2000). Administration of N-acetylcysteine to mutant mice for 2 weeks restored the fertility (Kumar et al, 2000).
Experimental γ-Glutamylcyclotransferase Deficiency
The best inhibitor of GCT is β-aminoglutaryl-l-α-aminobutyrate with a Ki of 0.6 mM, and it is competitive. When administered to mice, 5-oxoproline levels decrease substantially, suggesting that GCT activity is important in vivo for the formation of 5-oxoproline and also that the γ-glutamyl cycle functions in vivo.74
Experimental 5-Oxoprolinase Deficiency
Administration to mice of competitive inhibitors of OP, such as L-2-imidazolidone-4-carboxylate or D,L-3-methyl-5-oxoproline, showed that the metabolism of 5-oxoproline is significantly decreased and 5-oxoproline may be found in urine.1
Experimental Dipeptidase Deficiency
MBD1 knockout mice are viable and fertile (Habib et al, 1998). Cystinyl-bis-glycine in the urine is increased about threefold, and the mice lack β-lactamase activity in lung, kidney, small intestine, and heart. The conversion of leukotriene D4 (LTD4) to E4 (LTE4) was reduced. However, this activity has also been shown to be carried out by MBD2 (Habib et al, 2003). The findings in the MBD1 knockout mice suggest that there is an alternative cystinyl-bis-glycine hydrolyzing activity and also an alternative pathway for conversion of LTD4 to LTE4 (Habib et al, 1998).
Elevation of GSH Levels—Potential Therapies
Deficiency of GSH results from inborn errors in the GSH synthetic pathway. Other diseases have also been associated with GSH deficiency, such as HIV infection,75-79 hepatitis C infection,78 diabetes,79 ulcerative colitis,80 idiopathic pulmonary fibrosis,81 and adult respiratory distress syndrome.82 In normal aging, GSH levels in blood also decrease (Julius et al, 1994). Various toxic compounds can be rendered less harmful by reaction with GSH; such reactions can be nonenzymatic or via the GSH S-transferase family of enzymes. Superoxide dismutase and reactions involving GSH detoxify reactive oxygen species. Glutathione can react through glutathione peroxidase to detoxify hydrogen peroxide and organic peroxides. Reactive oxygen species have been associated with various diseases involving ischemia-reperfusion, namely stroke and myocardial infarction, cancer, AIDS, arthritis, atherosclerosis, adult respiratory distress syndrome, and Parkinson disease.83,84 It is not clear for any of these pathologic conditions whether the synthesis of GSH is affected, whether the metabolism of GSH is increased, or whether both are affected. Increasing cellular GSH levels may be beneficial for treating GSH deficiency caused by reactive oxygen species and by deficiency of GSH-related enzymes.
Administration of GSH is only effective as a cysteine delivery compound. GSH is not readily taken up by most cells; it is broken down and its component amino acids used to resynthesize GSH.15 Administration of cysteine supplies what is usually the limiting substrate for GSH synthesis; however, it is readily oxidized to cystine, which has limited solubility and is reported to be toxic.85 N-Acetyl cysteine (NAC) is commonly used to treat patients who are exposed to paracetamol in overdose. It is transported into cells and deacetylated to form cysteine for GSH synthesis. Two reports of treatment of GS-deficient patients with NAC suggested that NAC has little or no effect.86 NAC may serve as a reducing agent and liberate GSH from GSSG and other disulfide forms thereby supplying cysteine for GSH synthesis. Patients with GS deficiency, however, accumulate γ-glutamylcysteine and to some extent also cysteine, at least in fibroblasts (Ristoff et al, 2002). Since cysteine is known to be neurotoxic in excessive amounts (Janáky et al 2000; Olney et al, 1972), treatment with NAC should not be recommended in patients with GS deficiency as this might increase the intracellular cysteine levels even more. 2-Oxothiazolidine-4-carboxylate (OTC; procysteine) is a substrate for OP and is opened to form intracellular cysteine.5,87 It is well transported and increases cellular GSH levels in many cells;85 however, both GCS and GS activities are required. Administration of OTC has been reported to raise cellular GSH levels in HIV patients.88-90 An early clinical trial using OTC in patients infected with HIV showed that whole-blood GSH increased after OTC (Barditch-Crovo et al, 1998). An in vitro study showed that OTC exhibited anti-HIV effects, and even at a high dose, there was little cytotoxicity (Oiry et al, 1999). Administration of OTC has been shown to improve flow-mediated dilation in patients with coronary artery disease (Vita et al, 1998). In rats, OTC promotes growth and protects against ischemia.91,92
OTC has been used to increase cysteine and GSH levels in many systems, from cultured cells to whole animal and human studies. It protects liver (Iimuro et al, 2000; Ozguc et al, 2003), lung (Levy et al, 1998), pancreas (Luthen et al, 1997), and lymphocytes (Gross et al, 1997). It also promotes spinal cord function after trauma (Kamencic et al, 2001).
In GT-deficient mice, OTC administration partially restores hepatic GSH levels and improves their growth and appearance (Held et al, 2003); it also increases GSH levels in thymocytes and diminishes dexamethasone-induced apoptosis (Will et al, 2002).
The cellular level of γ-glutamylcysteine is low and normally it is the rate-limiting substrate for GS. Administration of this dipeptide can increase GSH in cells that can transport γ-glutamyl amino acids, such as those of the kidney.93 Administration of γ-glutamylcysteine bypasses GCS, the feedback inhibited and rate-limiting step of GSH synthesis. The use of γ-glutamylcysteine for GSH biosynthesis is part of the salvage or alternative pathway of GSH synthesis. In this pathway, GSH is exported, reacts via GT with cystine to form γ-glutamylcystine, which is transported and reduced to γ-glutamylcysteine and used directly by GS for GSH synthesis; the extent to which this pathway functions in vivo is not known.
γ-Glutamylcysteine has been shown to increase brain GSH levels after intraventricular injection.94 γ-Glutamylcysteine mono- and diesters have been used to transport γ-glutamylcysteine into cells for the synthesis of GSH.95,96 These compounds are protective against several cytotoxic agents (Drake et al, 2002; Hobbs et al, 1998; Naito et al, 2000; Okamoto et al, 1999).
There has been one report that γ-glutamylcysteine ester is selectively cytotoxic against HIV-1 infected H-9 T lymphocytes (Kubota et al, 1999).
Because GSH is not readily taken up by cells, GSH mono- and diesters were developed.5,97-99 These esters are more hydrophobic and are transported into cells in many tissues, except brain, and are hydrolyzed, in an unknown manner, to yield free GSH.100,101 The other product of the reaction is the corresponding alcohol, usually ethanol. Treatment of rodents and cell lines with GSH monoester protects against the toxicity listed above for experimental GSH deficiency produced by BSO treatment. It is also effective in the endogenous oxidative stress model involving newborn rodents (see above, “Experimental GCS Deficiency”). Thus, administration of GSH monoester to rodents prevents cataract formation and mitochondrial damage, and protects against the effects of vitamin C deficiency. Esters of GSH have potential in the therapy of patients with GSH deficiency.
GSH mono- and diesters have been used in various studies to modulate GSH levels and for protective functions (e.g., against oxidative stress, toxic compounds, aging); such studies are being added rapidly (Chen et al, 2000; Gao et al, 2002; Grattagliano et al, 1999; Leeuwenburgh et al, 1998; Lindsay et al, 1998; Madesh et al, 1998; Ohinata et al, 2000; Sakamaki et al, 1999; Singhal et al, 2000; Yamaguchi et al, 1998).
Vitamin C and GSH can spare each other in rodent models (see above, “Experimental GCS Deficiency”). Short-term treatment of GS-deficient patients with vitamin C has been reported to increase lymphocyte GSH levels.86 It is not known whether long-term vitamin C treatment is beneficial, but because vitamin C and GSH are not able to carry out the same functions, it is likely that vitamin C treatment is only partially effective. Treatment with other antioxidants, such as vitamin E, may have effects similar to treatment with vitamin C, but further studies are needed. A long-term follow up study of 28 patients with GS deficiency indicates that early supplementation with vitamins C and E may improve the long-term clinical outcome by protecting the CNS from damage (Ristoff et al, 2001).