Glyoxalase I is part of the glyoxalase system present in the cytosol of cells. The glyoxalase system catalyses the conversion of reactive, acyclic a-oxoaldehydes to the corresponding a-hydroxyacids. Glyoxalase I catalyses the isomerisation of the hemithioacetal, formed spontaneously from a-oxoaldehyde and GSH, to S-2-hydroxyacylglutathione derivatives RCOCH(OH)-SG ® RCH(OH)CO-SG, and in so doing decreases the steady concentrations of physiological a-oxoaldehydes and associated glycation reactions. Physiological substrates of glyoxalase I are: methylglyoxal, glyoxal and other acyclic a-oxoaldehydes. Human glyoxalase I is a dimeric zinc metalloenzyme of molecular mass 42 kDa. Glyoxalase I from Escherichia coli was a nickel Ni2+ metalloenzyme. The crystal structures of human and E. coli glyoxalase I have been determined to 1.7 and 1.5 Å resolution. The zinc Zn2+ site is two, structurally equivalent residues from each domain – Gln-33A, Glu-99A, His-126B, Glu-172B and two water molecules. The Ni2+ binding site is His-5A, Glu-56A, His-74B, Glu-122B and two water molecules. The catalytic reaction involves base-catalysed shielded‑proton transfer from C1 to C2 of the hemithioacetal to form an ene‑diol intermediate and rapid ketonisation to the thioester product. R- and S- enantioners of the hemithioacetal are bound in the active site, displacing the water molecules in the metal ion primary coordination shell. It has been proposed that Glu-172 is the catalytic base for the S-substrate enantiomer and Glu-99 the catalytic base for the R-substrate enantiomer; Glu-172 then reprotonates the ene-diol stereospecifically to form the R-2-hydroxyacylglutathione product. By analogy with human enzyme, Glu-56 and Glu-122 may be the bases involved in the catalytic mechanism of E. coli glyoxalase I. The suppression of a-oxoaldehyde-mediated glycation by glyoxalase I is particularly important in diabetes and uraemia where a-oxoaldehyde concentrations are increased. Decreased glyoxalase I activity in situ in the ageing process and oxidative stress increases glycation and tissue damage. Inhibition of glyoxalase I pharmacologically with specific inhibitors leads to the accumulation of a-oxoaldehydes to cytotoxic levels; cell permeable glyoxalase I inhibitors are anti-tumour and anti-malarial agents. Glyoxalase I has a critical role in the prevention of methylglyoxal, glyoxal and other a-oxoaldehyde-mediated glycation reactions in vivo.
The glyoxalase system
Glyoxalase I (EC 18.104.22.168) is part of the glyoxalase system present in the cytosol of all cells. The glyoxalase system catalyses the conversion of reactive, acyclic a-oxoaldehydes to the corresponding a-hydroxyacids. It is comprised of two enzymes, glyoxalase I and glyoxalase II (EC 22.214.171.124) and a catalytic amount of glutathione GSH. Glyoxalase I catalyses the isomerisation of the hemithioacetal, formed spontaneously from a-oxoaldehyde RCOCHO and GSH, to S-2-hydroxyacylglutathione derivatives RCH(OH)CO-SG.
RCOCHO + GSH ⇌ RCOCH(OH)-SG ® RCH(OH)CO-SG
For the methylglyoxal-glutathione hemithioacetal and human glyoxalase I, the KM is 71-130 µM and the kcat is 7-11 x 104 min-1. Glyoxalase II catalyses the conversion of S-2-hydroxyacylglutathione derivatives to a-hydroxyacids and reforms GSH consumed in the glyoxalase I-catalysed reaction step.
Figure 1. The glyoxalase system
The major physiological substrate for glyoxalase I is methylglyoxal and this accumulates markedly when glyoxalase I is inhibited in situ by cell permeable glyoxalase I inhibitors and by depletion of GSH [1-3]. Methylglyoxal is formed mainly by the degradation of triosephosphates, and also by the metabolism of ketone bodies, threonine degradation and the fragmentation of glycated proteins. Other substrates are glyoxal – formed by lipid peroxidation and the fragmentation of glycated proteins, hydroypyruvaldehyde HOCH2COCHO and 4,5-doxovalerate H-COCOCH2CH2CO2H [1;4]. Glyoxalase I activity prevents the accumulation of these reactive a-oxoaldehydes and thereby suppresses a-oxoaldehyde-mediated glycation reactions . It is, therefore, a key enzyme of the anti-glycation defence.
Molecular properties of glyoxalase I
Glyoxalase I activity is present in all human tissues. Specific activities of fetal tissues are ca. 3 times higher than corresponding adult tissues. There is ca. 0.2 μg of glyoxalase I per gram of protein in human tissues and blood cells. Human glyoxalase I is a dimer, expressed at a diallelic genetic locus GLO which encodes for two similar subunits in heterozygotes; the three allozymes are designated GLO 1-1, GLO 1-2 and GLO 2-2. All allozymes have molecular mass of 46 kDa (gel filtration) or 42 kDa (sequence) and isoelectric point pI values of 4.8-5.1; however, they have distinctive charge densities and/or molecular shapes and are resolved by ion exchange chromatography and non-denaturing gel electrophoresis. Each subunit contains one zinc ion, Zn2+ . Surprisingly, glyoxalase I from Escherichia coli is a nickel Ni2+ metalloenzyme .
Sixty-one glyoxalase I sequences have been reported: human, murine, yeast (Saccharomyces cerevisae and Schizosaccharomyces pombe), plant (Arabidopsis thaliana, Avicennia marina, Brassica oleracea, Brassica juncea, Cicer arietinum, Glycine max, Oryza sativa, Sporobolus stapfianus, Solanum lycopersicu, and Triticum aestivum), insect (Drosophila melanogaster, Anopheles gambiae), protozoal (Plasmodium falciparum and Plasmodium yoelii), fungi (Paracoccidioides brasiliensis) and 42 strains of bacterial enzyme (Escherichia coli, Pseudomonas putida and others). The human, bacterial and plant glyoxalase I enzymes are dimeric. The yeast enzymes of Saccharomyces cerevisae and Schizosaccharomyces pombe are monomers of 32 and 37 kDa, respectively, with two copies of a segment equivalent to the monomer of the human protein. The sequence identity of human glyoxalase I with the bacterial enzyme (Pseudomonas putida) is 55% and with the yeast enzyme between residues 1-182 and 183-326 (Saccharomyces cerevisiae) is 47%, suggesting glyoxalase I of different origins may have arisen by divergent evolution from a common ancestor.
Structure and catalytic mechanism of glyoxalase I
The translation product of human glyoxalase I contains 184 amino acids. The N-terminal Met is removed in post-translational processing and the N-terminal Ala blocked by an as-yet, unknown modification. There are at least 4 possible phosphorylation sites. The structure of human glyoxalase I enzyme in complex with S-benzyl-glutathione was determined to 2.2 Å resolution  – Figure 2a. Each monomer consists of two, structurally equivalent domains. The active site is situated in the dimer interface, with the inhibitor and essential Zn2+ ion interacting with side chains from both subunits. The zinc binding site is two structurally equivalent residues from each domain – Gln-33A, Glu-99A, His-126B, Glu-172B and two water molecules in octahedral coordination. The competitive inhibitors S-p-bromobenzylglutathione and S-(p-nitrobenzoxycarbonyl)glutathione were bound in the second coordination shell of Zn2+ in extended conformation, whereas the transition state analogue inhibitor S-[N-hydroxy-N-(p-iodophenyl)carbamoyl]glutathione was bound directly to the Zn2+ ion by its carbamoyl oxygens in a cis conformation. The carboxylate group of the g-Glu residue forms a salt bridge link to the guanidino group of Arg-37A and the amide group Asn-103A; Arg-122B was within hydrogen bonding distance of the g-Glu residue. Gly was not bound [7;8].
Glyoxalase I of Escherichia coli is composed of two identical 135 amino acid subunits. The crystal structure of the Ni2+-bound enzyme has been determined to 1.5 Å resolution. Each subunit consists of two domains, residues 3-60 and 72-126, that are linked by an intervening 12 residue segment. Each domain consists of a babbb motif. The Ni2+ ion is coordinated by His-5A, Glu-56A, His-74B, Glu-122B and two water molecules in an octahedral coordination – Figure 2b.
Figure 2 Structure and catalytic mechanism of glyoxalase I. a. and b., Solid ribbon representations of the crystal structures of human and Escherichia coli glyoxalase I [7,11]. Subunits are shown as light and dark grey; metal ion coordinating amino acid residues are shown in stick format. Active sites are located in the central, vertical cleft of each protein.
The mechanism proposed for the glyoxalase I reaction involves base-catalysed shielded‑proton transfer from C1 to C2 of the hemithioacetal, bound in the active site, to form an ene‑diol intermediate and rapid ketonisation to the thioester product - Figure 3. Both R- and S-forms of the hemithioacetal are bound in the active site of glyoxalase I and are therein deprotonated; the subsequent reprotonation of the putative ene-diol intermediate occurs stereospecifically to form the R-2-hydroxyacylglutathione derivative. It has been proposed that Glu-172 is the catalytic base for the S-substrate enantiomer and Glu-99 the catalytic base for the R-substrate enantiomer. Both reaction mechanisms form and cis-ene-diol intermediate coordinated directly to the Zn2+ ion: this is deprotonated to a cis-ene-diolate by Glu-172 which then reprotonates C2 stereospecifically to form the R-2-hydroxyacylglutathione product  – Figure 1c. S-Glycolylglutathione, S-D-lactoylglutathione and S-L-glyceroylglutathione are formed from glyoxal, methylglyoxal and hydroypyruvaldehyde by glyoxalase I and hydrolysed to glycolate, D-lactate and L-glycerate by glyoxalase II, respectively . By analogy with human enzyme, Glu-56 and Glu-122 may be the bases involved in the catalytic mechanism .
Figure 3. Catalytic mechanism of glyoxalase I. Catalytic mechanism of human glyoxalase I for the isomerisation of the R-hemithioacetal – from 
Genetics and polymorphism
There are 3 phenotypes of human glyoxalase I, GLO 1‑1, GLO 1-2 and GLO 2‑2, representing the homozygous and heterozygous expression of a the diallelic gene, GLO1 and GLO2 at an autosomal locus; GLO2 allele is the ancestral allele with the GLO1 allele arising by mutation. GLO alleles are inherited in a simple co-dominant manner, with characteristic phenotypic expression present in all tissues. The GLO locus is on chromosome 6, between the centromere and HLA‑DR. Population genetics of GLO1 allele frequency show that the GLO1 allele frequency is highest in native tribes in Alaska, and decreases geographically South and East to Europe and South America, through Africa, the Middle East and India, to the very low GLO1 allele frequencies of the Far East and Oceania . The allele expression products differ in amino acid sequence only at position 111: in subunit GloI-A there is an alanine residue, and in subunit GloI-E there is a glutamic acid residue . The gene promoter contains insulin response element (IRE) and metal response element (MRE) . There was a phenotypic disturbance of GLO in type 1 diabetic patients with and without chronic microvascular complications (retinopathy and neuropathy). Diabetic patients without complications had a significant increase in GLO 1-1 homozygote .
The role of glyoxalase I in the anti-glycation defence
The formation of methylglyoxal is an intrinsic feature of the Embden‑Meyerhof pathway; a corollary to the presence of triose phosphate intermediates in glycolysis. It is minimised by maintaining low concentrations of triosephosphates at steady state and capping of the active site in triosephosphate isomerase. Consequently, the formation of methylglyoxal accounts for only ca. 0.1 ‑ 0.4% of glucotriose flux. Methylglyoxal has high reactivity in glycation reactions in vivo, forming advanced glycation endproducts (AGEs) of protein, nucleotides and probably also basic phospholipids  . Similar adducts are formed by glyoxal and probably by other glyoxalase I substrates too. The glyoxalase system is an efficient enzymatic detoxification system suppressing the formation of methylglyoxal- and glyoxal derived AGEs. Recent quantitative analysis of methylglyoxal-derived glycation adducts of cellular and extracellular proteins indicates that 0.1 – 2% of total cellular arginine is modified by methylglyoxal – the highest estimate in human lens proteins of elderly subjects with low glyoxalase I activity [18;19]. A patient with endstage renal disease and unusually low glyoxalase I activity also had high levels of advanced glycation endproducts and a susceptibility to recurrent events of macrovascular disease when conventional risk factors were controlled . Convincing experimental evidence that glyoxalase I suppressed the formation of AGEs came from studies of endothelial cells in normoglycaemic and hyperglycaemic culture. Hyperglycaemia induced increases in the concentrations of methylglyoxal, D-lactate and cellular protein AGEs. Overexpression of glyoxalase I prevented totally the increase in methylglyoxal and cellular protein AGEs, and increased the concentration of D-lactate. This indicated that glyoxalase I had a critical role in suppressing the formation of protein AGEs  – Figure 4, a.-d.
Figure 2 Prevention of cellular glycation by glyoxalase I. Glyoxalase I and methylglyoxal metabolism in GM7373 endothelial cells overexpressing glyoxalase I (GLO1) and sham-transfected controls (Neo). a. Glyoxalase I activity; ***, P<0.001, with respect to the sham transfected control. b., c. and d., cellular concentrations of methylglyoxal, AGEs and D-lactate of incubations in normoglycaemia (5 mM glucose, ℵ) and hyperglycaemia (30 mM glucose, ) ** and ***, P<0.01 and P<0.001, with respect to the normoglycaemic control. From .
Glyoxalase I is a GSH-dependent enzyme. Under physiological conditions in situ, the rate of fragmentation of hemithioacetal to GSH and methylglyoxal is of the order of 103 times faster than the rate of isomerisation by glyoxalase I. Therefore, there is a rapid pre-equilibrium of GSH and methylglyoxal with hemithioacetal and the activity of glyoxalase I in situ is proportional to the cellular concentration of GSH. Experimental depletion of GSH, by oxidative or non-oxidative mechanisms (exposure to cytotoxic levels of hydrogen peroxide or the glutathione transferase substrate 1-chloro-2,4-dinitrobenzene CDNB), induced marked accumulation of methylglyoxal, a much smaller increase in glyoxal, and induced cytotoxicity. Scavenging of a-oxoaldehydes by aminoguanidine prevented the accumulation of a-oxoaldehydes and cytotoxicity. This indicates that methylglyoxal accumulates markedly in oxidative stress and may mediate oxidant-induced cytotoxicity  – Figure 5, a. and b.
The cytotoxicity associated with the accumulation of methylglyoxal and other glyoxalase I substrates should be avoided under normal physiological states. Hence, the function of glyoxalase I is the detoxification of a-oxoaldehydes as part of the enzymatic defence against glycation. In certain disease states – such as cancer and microbial infections, we may wish to induce cytotoxicity of tumour cells and microbial organisms pharmacologically. A cell permeable glyoxalase I inhibitor achieves this.
S-p-Bromobenzylglutathione (SpBrBzGSH) is a potent inhibitor of human glyoxalase I (Ki = 83 nM). Diesterification of SpBrBzGSH stabilises this GSH conjugate to extracellular degradation by g-glutamyl transpeptidase and makes it cell permeable. Inside cells, SpBrBzGSH diesters are de-esterified and glyoxalase I is inhibited. S-p-Bromobenzylglutathione cyclopentyl diester (SpBrBzGSHCp2) was a potent antitumour agent in vitro and had antitumour activity in vivo . More recent studies have shown overexpression of glyoxalase I associated with multidrug resistance in cancer chemotherapy where potent antitumour activity was achieved with SpBrBzGSHCp2 – Figure 5, c. and discussed elsewhere in this issue . S-p-Bromobenzylglutathione ethyl diester (SpBrBzGSHEt2) had potent anti-malarial activity against the red blood cell stage of Plasmodium falciparum. This stage of the malarial parasite has only anaerobic glycolysis and an associated high flux of methylglyoxal formation. It is, therefore, particularly sensitive to pharmacological inhibition of glyoxalase I with associated accumulation of methylglyoxal to cytotoxicity levels  - Figure 5, d.
Figure 5 Inhibition of glyoxalase I and cytotoxicity. a. Accumulation of methylglyoxal in P388D1 macrophages in vitro. Key: (X-X), control; (○-○) + hydrogen peroxide, and (□-□) + CDNB. b. Prevention of cytotoxicity to P388D1 macrophages in vitro induced by hydrogen peroxide and CDNB with aminoguanidine. Key: (ℵ), control; () + aminoguanidine; **, P<0.01 with respect to the control. Data from . c. Inhibition of human leukaemia 60 cell growth by SpBrBzGSHCp2 in vitro; the median inhibitory growth concentration GC50 value was 4.23 ± 0.01 mM . d. Inhibition of malarial parasite Plasmodium falciparum growth in human red blood cells in vitro by SpBrBzGSHEt2 in vitro; the median inhibitory growth concentration GC50 value was 4.77 ± 0.12 mM .
There is established and further emerging evidence that the physiological substrates of glyoxalase I are potent glycating agents of proteins, nucleotides and basic phospholipids, and these glycation processes contribute to cell and tissue damage in vivo. This is particularly marked in diabetes and uraemia where a-oxoaldehyde concentrations are increased [22;23]. By the efficient metabolism of a-oxoaldehydes, glyoxalase I provides an enzymatic defence against a-oxoaldehyde-mediated glycation. Decrease of glyoxalase I activity in situ by the ageing process and oxidative stress increases glycation and tissue damage. The glyoxalase I defence against glycation appears to be a factor associated with risk of developing vascular complications of diabetes and uraemia. Surprisingly, glyoxalase I activity is also linked with multidrug resistance in cancer chemotherapy, suggesting a role of methylglyoxal-induced apoptosis in the mechanism of action of some antitumour drugs.
1 Thornalley, P. J. (1993) Molecular Aspects of Medicine 14, 287-371
2 Abordo, E. A., Minhas, H. S., and Thornalley, P. J. (1999) Biochem.Pharmacol. 58, 641-648
3 Thornalley, P. J., Edwards, L. G., Kang, Y., Wyatt, C., Davies, N., Ladan, M. J., and Double, J. (1996) Biochem.Pharmacol. 51, 1365-1372
4 Thornalley, P. J. (1998) Chem.-Biol.Interact. 111-112, 137-151
5 Shinohara, M., Thornalley, P. J., Giardino, I., Beisswenger, P. J., Thorpe, S. R., Onorato, J., and Brownlee, M. (1998) J.Clin.Invest. 101, 1142-1147
6 Clugston, S. L., Barnard, J. F. J., Kinach, R., Miedema, D., Ruman, R., Daub, E., and Honek, J. F. (1998) Biochemistry 37, 8754-8763
7 Cameron, A. D., Olin, B., Ridderstrom, M., Mannervik, B., and Jones, T. A. (1997) EMBO J. 16, 3386-3395
8 Cameron, A. D., Ridderstrom, M., Olin, B., Kavarana, M., Creighton, D. J., and Mannervik, B. (1999) Biochemistry 38, 13480
9 Himo, F. and Siegbahm, P. E. M. (2001) J.Am.Chem.Soc. 123, 10280-10289
10 Clelland, J. D. and Thornalley, P. J. (1991) J.Chem.Soc.Perkin.Trans.I 3009-3015
11 He, M. M., Clugston, S. L., Honek, J. F., and Mathews, B. W. (2000) Biochemistry 39, 8719-8727
12 Thornalley, P. J. (1991) Heredity 67, 139-142
13 Kim, N.-S., Sekine, S., Kiuchi, N., and Kato, S. (1995) J.Biochem. 117, 359-361
14 Ranganathan, S., Ciaccio, P. J., Walsh, E. S., and Tew, K. D. (1999) Gene 240, 149-155
15 McCann, V. J., Davis, R. E., Welborn, T. A., Constable, J., and Beale, D. J. (1981) Aust.N.Z.J.Med. 11, 380-382
16 Ahmed, N., Argirov, O. K., Minhas, H. S., Cordeiro, C. A., and Thornalley, P. J. (2002) Biochem.J. 364, 1-14
17 Thornalley, P. J. (2003) Biochem.Soc.Trans., in press (Protecting the genome)
18 Thornalley, P. J., Battah, S., Ahmed, N., Karachalias, N., Agalou, S., and Dawnay, A. (2003) Biochem.J., in press
19 Ahmed, N., Thornalley, P. J., Dawczynski, J., Franke, S., Strobel, J., Stein, G., and Haik JR, G. M. (2003) Invest.Ophthalmol.Vis.Sci., in press
20 Miyata, T., van Ypersele de Strihou, C., Imasawa, T., Yoshino, A., Ueda, Y., Ogura, H., Kominami, K., Onogi, H., Inagi, R., Nangaku, M., and Kurokawa, K. (2001) Kidney Internat. 60, 2351-2359
21 Thornalley, P. J., Strath, M., and Wilson, R. J. M. (1994) Biochem.Pharmacol. 268, 14189-14825
22 McLellan, A. C., Thornalley, P. J., Benn, J., and Sonksen, P. H. (1994) Clin.Sci. 87, 21-29
23 Thornalley, P. J. (1999) Clin.Lab. 45, 263-273