A Trypanothione-Dependent Glyoxalase I with a Prokaryotic Ancestry in Leishmania Major

A trypanothione-dependent glyoxalase I with a prokaryotic ancestry in Leishmania major

Tim J. Vickers, Neil Greig, and Alan H. Fairlamb*

Division of Biological Chemistry and Molecular Microbiology, Wellcome Trust Biocentre, School of Life Sciences, University of Dundee, DD1 5EH Dundee, Scotland

Edited by Anthony Cerami, The Kenneth S. Warren Institute, Kitchawan, NY, and approved July 29, 2004 (received for review April 26, 2004)

Glyoxalase I forms part of the glyoxalase pathway that detoxifies to hydrolyze the S-D-lactoylglutathione product of GLO1 to reactive aldehydes such as methylglyoxal, using the spontaneously D-lactate and glutathione (1). GLO2 is not as well characterized formed glutathione hemithioacetal as substrate. All known eu- as GLO1 and is not seen as a target for chemotherapy, because karyotic enzymes contain zinc as their metal cofactor, whereas the the reaction catalyzed by GLO1 is essentially irreversible under Escherichia coli glyoxalase I contains nickel. Database mining and physiological conditions; thus, inhibition of GLO2 would only sequence analysis identified putative glyoxalase I genes in the cause the accumulation or excretion of the nontoxic S-D- eukaryotic human parasites Leishmania major, Leishmania infan- lactoylglutathione (10). Interestingly, GLO1 inhibitors are se- tum, and Trypanosoma cruzi, with highest similarity to the cya- lectively toxic to proliferating cells, possibly as a result of an nobacterial enzymes. Characterization of recombinant L. major inhibition of DNA synthesis produced by the accumulation of glyoxalase I showed it to be unique among the eukaryotic enzymes methylglyoxal (11, 12). GLO1 inhibitors have also been reported in sharing the dependence of the E. coli enzyme on nickel. The to have antimalarial (13) and antitrypanosomal (14) activities. parasite enzyme showed little activity with glutathione hemithio- New antiparasitic drugs are required for the treatment of acetal substrates but was 200-fold more active with hemithioac- tropical diseases caused by the pathogenic trypanosomes and etals formed from the unique trypanosomatid thiol trypanothione. Leishmania. These protozoa cause Chagas’ disease, sleeping L. major glyoxalase I also was insensitive to glutathione derivatives sickness, and the various forms of leishmaniasis, which cause that are potent inhibitors of all other characterized glyoxalase I serious illness and death, particularly in many tropical regions. enzymes. This substrate specificity is distinct from that of the The currently available treatments, on the whole, are poor and human enzyme and is reflected in the modification in the L. major suffer from serious side effects, limited efficacy, and both natural sequence of a region of the human protein that interacts with the and acquired resistance (15). New drugs and drug targets are glycyl-carboxyl moiety of glutathione, a group that is conjugated therefore urgently required (16). Such a target may be offered to spermidine in trypanothione. This trypanothione-dependent by the glyoxalase system of the pathogenic kinetoplastids, as a glyoxalase I is therefore an attractive focus for additional biochem- consequence of these protozoa possessing an unusual thiol ical and genetic investigation as a possible target for rational drug metabolism. In these organisms, instead of glutathione, the 1 8 design. major low molecular mass thiol is trypanothione [N ,N - bis(glutathionyl)spermidine] (17). Previous investigations on methylglyoxal ͉ D-lactate ͉ glutathione ͉ nickel Leishmania donovani showed that intact cells quantitatively converted methylglyoxal to D-lactate, suggesting the presence of an active glyoxalase system (18). However, only very low GLO1 ll organisms require systems to shield them from chemical and GLO2 activities could be detected in extracts using gluta- Astress, such as the antioxidant enzymes that detoxify en- thione as a cofactor (19). These observations prompted us to dogenous oxidants and the enzymes that metabolize exogenous ␣ investigate whether trypanothione would substitute for glutathi- toxins. However, endogenous toxins such as the reactive -oxoal- one in the glyoxalase pathway. Here we report the cloning, dehyde methylglyoxal (2-oxopropanal) are also by-products of expression, and characterization of a Leishmania major GLO1 metabolism (1). The main source of methylglyoxal is the non- that is similar to prokaryotic glyoxalases but unique in its enzymatic fragmentation of triose phosphates, but it also is pronounced specificity for trypanothione hemithioacetal over formed during threonine catabolism and acetone oxidation (2). glutathione hemithioacetal as substrate. L. major GLO1 thus In some prokaryotes, dihydroxyacetone phosphate is enzymat- represents a member of a previously uncharacterized family of ically converted to methylglyoxal and inorganic phosphate by GLO1 enzymes. During our investigations into L. major GLO1, methylglyoxal synthase (EC 4.2.3.3) in response to phosphate an independent study reported that GLO2 from Trypanosoma starvation (2). Methylglyoxal reacts rapidly with both proteins brucei was highly selective for mono- and bis-S-D-lactoyltrypano- and nucleic acids and thus is both toxic and mutagenic (3, 4). thione (20). Taken together, these observations provide com- The glyoxalase system is a ubiquitous detoxification pathway pelling evidence for trypanosomatids possessing a unique that protects against the cellular damage caused by methylg- trypanothione-dependent glyoxalase system. lyoxal. This pathway comprises glyoxalase I (GLO1) (lactoylglutathione lyase, EC 4.4.1.5) and glyoxalase II (GLO2) (hydroxya- Materials and Methods cylglutathione hydrolase, EC 3.1.2.6), which act in concert to Materials. Saccharomyces cerevisiae GLO1 was purchased from convert the spontaneously formed hemithioacetal adduct be- Sigma. Methylglyoxal was prepared from methylglyoxaldimethy- tween glutathione (␥-L-glutamyl-L-cysteinylglycine) and methylglyoxal into D-lactate and free glutathione (1). Thus, glutathione acts as a cofactor in the overall reaction pathway. GLO1 is This paper was submitted directly (Track II) to the PNAS office. a metalloenzyme that isomerizes the glutathione hemithioacetal Freely available online through the PNAS open access option. to S-D-lactoylglutathione through proton transfer to a metal- Abbreviations: GLO1, glyoxalase I; GLO2, glyoxalase II; trypanothione, N1,N8-bis(glutathio- bound enediol intermediate (5). This metal cofactor is zinc in nyl)spermidine. eukaryotes but nickel in Escherichia coli (6, 7). The difference in Data deposition: The sequence reported in this paper (Leishmania major GLO1 gene) has cofactor dependence is reflected in differences between the been deposited in the GenBank database (accession no. AY604654). active sites of the human and E. coli enzymes, suggesting that the *To whom correspondence should be addressed. E-mail: [email protected]. latter may be a target for antimicrobial therapy (8, 9). GLO2 acts © 2004 by The National Academy of Sciences of the USA

13186–13191 ͉ PNAS ͉ September 7, 2004 ͉ vol. 101 ͉ no. 36 www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402918101 Downloaded by guest on September 30, 2021 lacetal and confirmed to be free of formaldehyde, as described pKa of the protein was measured by chromatofocusing with a in ref. 21. S-p-bromobenzylglutathione was synthesized as de- MonoPHR5͞20 column (Amersham Pharmacia) and a pH scribed in ref. 22. All other reagents were standard commercial 7.0–4.0 gradient of Polybuffer 74. products of high purity. Production and Assay of Thiols. Trypanothione disulphide (2.5 Determination of Dissociation Constants. Methylglyoxal stocks were ␮mol; Bachem) in 0.5 ml of 400 mM Hepes buffer, pH 7.4, was assayed by preparation of the disemicarbazone adduct (23) and reduced by the addition of a 0.5-ml suspension of Tris(2- ␮ with S. cerevisiae GLO1 (24), both giving identical values. The Kd carboxyethyl)phosphine agarose (containing 4.8 mol of reduc- values for the adducts between glutathione or trypanothione and tant; Pierce). After mixing (25°C for 30 min), the agarose was phenylglyoxal were determined as described in ref. 24. However, removed, HCl was added to 20 mM, and the acidified solution the extinction coefficient of the trypanothione-phenylglyoxal was stored at Ϫ20°C. Thiol concentrations were determined by adduct was determined by using excess phenylglyoxal rather than titration with 5,5Ј-dithio-bis(2-nitrobenzoic acid) (26) by using excess trypanothione, because the use of millimolar levels of an extinction coefficient of 14.14 mMϪ1⅐cmϪ1 at 412 nm for the trypanothione is impractically expensive. A dual-beam Shi- thionitrobenzoate anion (27). madzu UV-2401 PC spectrophotometer was used with a blank assay lacking thiol in the reference beam. The measured extinc- Enzyme Kinetics. Kinetic constants were determined at 27°Cin tion coefficients then were used to determine the concentration 0.5-ml assays containing 100 mM (Naϩ) phosphate, pH 7.0. To of adduct in a set of equilibrium mixtures in 100 mM (Naϩ) ensure full saturation with cofactor, GLO1 was diluted with ␮ ͞ phosphate, pH 7.0. The concentrations used (expressed as [SH assay buffer containing 20 M NiCl2 and 0.02% (wt vol) BSA. group]) were 100 ␮M phenylglyoxal plus 100 ␮M thiol, 100 ␮M The required concentration of hemithioacetal and 0.1 mM free phenylglyoxal plus 200 ␮M thiol, and 200 ␮M phenylglyoxal plus thiol were produced by varying the thiol and aldehyde concen- 100 ␮M thiol. trations, with the concentrations quoted of trypanothione being that of the sulfydryl group. Hemithioacetal concentrations were Growth and Lysis of Leishmania. The L. major Friedlin A1 clone calculated by using the published Kd values of 3 and 0.6 mM for was grown as described in ref. 25. Frozen cells were thawed on the methylglyoxal– and phenylglyoxal–glutathione equilibria, ice and resuspended in lysis buffer [75 mM (Naϩ) phosphate, pH respectively (24). Trypanothione hemithioacetal concentrations ͞ ͞ ͞ ͞ ␮ ⅐ Ϫ1 7.5 500 mM NaCl 1mMDTT 1 mM benzamidine 3 g ml were calculated by using the Kd value of 3 mM for the gluta- leupeptin͞250 ␮M 4-(2-aminoethyl)benzenesulphonyl fluo- thione–methylglyoxal equilibrium and the experimental value of ride͞1 ␮M pepstatin A]. Cells were lysed by sonication (four 30-s 0.5 mM for the phenylglyoxal–trypanothione equilibrium. In the Ͻ bursts), with cooling to 4°C between pulses. After centrifuga- case of glutathione hemithioacetals, Km values were too large for ϫ ͞ tion (100,000 g for 1 h), the extract was dialyzed against assay accurate measurement of kinetic constants, and thus kcat Km bufferplus1mMDTTfor2htoremove components of Ͻ10 values were determined from the slope of plots of rate versus kDa. enzyme concentration. Trypanothione-dependent GLO2 activity was determined at Cloning and Expression of L. major GLO1. A putative L. major GLO1 27°C in 0.5 ml of 100 mM Mops, (pH 7.2)͞150 ␮M trypano- gene (GenBank accession no. AY604654) was identified in the thione-SH͞200 ␮M methylglyoxal (20). Lactoyltrypanothione L. major genome database (www.genedb.org). This gene was was produced by adding 0.3 nmol of recombinant L. major amplified by PCR from genomic DNA by using the sense primer GLO1, and the reaction was monitored at 240 nm. Once the 5Ј-CATATGCCGTCTCGTCGTATGCTGCACACC-3Ј and reaction was complete, cell extract was added, and the rate of the antisense primer 5Ј-GGATCCGGACCTTACGCAGTGC- decrease in absorbance at 240 nm was measured. Glutathione- CCTGCTCCTTC-3Ј and fully sequenced. These primers added dependent GLO2 activity was assayed under identical conditions a5Ј NdeI site and a 3Ј BamHI site (underlined), allowing cloning using 150 ␮M glutathione and 0.3 nmol S. cerevisiae GLO1 in into the expression plasmid pET15b (Novagen). place of trypanothione and L. major GLO1. Methylglyoxal Transformed BL21(DE3)pLysS E. coli were grown in LB reductase and methylglyoxal dehydrogenase were assayed as BIOCHEMISTRY medium containing 100 ␮g⅐mlϪ1 carbenicillin and 12.5 ␮g⅐mlϪ1 described in ref. 19. chloramphenicol to an absorbance of 0.6 at 600 nm. Expression was induced for 4 h at 25°C with 1 mM isopropyl-␤-D- Apoenzyme Production, Metal Reconstitution, and Analysis. Apoen- galactopyranoside. Cells were lysed as described above in buffer zyme was produced by overnight dialysis at 4°C of 50 nmol of modified by the substitution of 10 mM 2-mercaptoethanol for GLO1 against 1 liter of 20 mM imidazole͞1 mM EDTA, pH 7.0, DTT. The supernatant after centrifugation (30,000 ϫ g for 30 containing5gofChelex resin and then for an additional 4 h min) was applied at 2 ml⅐minϪ1 to a nickel-charged 5-ml HiTrap against 2 liters of 20 mM imidazole, pH 7.0, containing5gof chelating Sepharose column (Amersham Pharmacia) in 50 mM Chelex resin. Aliquots (0.1 ml) were incubated on ice with a (Naϩ) phosphate͞200 mM NaCl, pH 7.5. GLO1 was eluted with 100-fold molar excess of metal chloride and then assayed. a linear gradient of 0–500 mM imidazole in the same buffer, and Chelex-treated ultrapure water and buffers were used in plastic the tag was removed by digestion for4hat25°C with thrombin containers at all stages in this procedure to reduce adventitious (10 units⅐mgϪ1; Pharmacia). After dialysis against 20 mM Bis- metal-ion contamination. Tris, pH 6.5, the protein was applied at 2 ml⅐minϪ1 to a 5-ml The zinc and nickel contents of a 0.1-␮mol sample of GLO1 HiTrap Q-Sepharose column (Amersham Pharmacia). GLO1 were determined by atomic absorption spectrophotometry. This was eluted with a linear gradient of 0–500 mM NaCl in the same sample was dialyzed overnight at 4°C against 2 liters of 10 mM buffer. Hepes, pH 7.0, containing5gofChelex resin. The metals then were measured by using a Unicam 939 AA spectrophotometer, Protein Characterization. Subunit masses were measured by ma- with conditions being according to manufacturer guidelines. trix-assisted laser desorption ionization mass spectrometry on a Voyager-DE STR (PerSeptive Biosystems, Framingham, MA). Results The native Mr of GLO1 was measured by using a Superdex 200 Dissociation Constants and Absorbance Coefficients. The dissocia- HR 10͞30 size-exclusion column (Amersham Pharmacia) in 50 tion constants for the glutathione and trypanothione hemithio- mM Hepes͞300 mM NaCl, pH 7.4. Unknowns and standards acetals of phenylglyoxal were determined spectrophotometri- were applied and eluted at 0.5 ml⅐minϪ1 in the same buffer. The cally by the change in absorbance at 280 nm after hemithioacetal

Vickers et al. PNAS ͉ September 7, 2004 ͉ vol. 101 ͉ no. 36 ͉ 13187 Downloaded by guest on September 30, 2021 Table 1. Methylglyoxal-catabolizing activities in L. major extract Trypanosomatid GLO1 Sequences. Putative GLO1 genes were Speciﬁc activity, identified in the trypanosomatid genome databases by BLAST Enzyme and cosubstrate nmol⅐minϪ1 per mg searches (L. major, LmjF35.3010; Leishmania infantum, LI0762b01.p1k; and Trypanosoma cruzi, Tc00.1047053510659.240 GLO1 Trypanothione 40 Ϯ 7 and Tc00.1047053510743.70). These genes are highly similar to each Glutathione Ͻ5 other, with the L. infantum and T. cruzi sequences showing 97% and GLO2 Trypanothione Ͻ2 72% identity, respectively, to the L. major sequence at the amino Glutathione Ͻ2 acid level. The two T. cruzi GLO1 sequences differ by only three Methylglyoxal reductase NADPH 4.0 Ϯ 0.4 silent nucleotide substitutions and may represent different alleles of NADH Ͻ2 the gene. Methylglyoxal dehydrogenase NADPH Ͻ2 Interestingly, the putative GLO1 sequences are most similar to NADH Ͻ2 bacterial glyoxalases, with, for instance, the L. major GLO1 being 47% and 51% identical at the amino acid level to the glyoxalases of the proteobacteria E. coli and the cyanobacteria formation. Because trypanothione contains two SH groups per Synechococcus sp. WH 8102, respectively. In contrast, L. major mole, data are presented per mole of sulfydryl group for GLO1 is only 33% identical to human GLO1, with the human consistency with data for glutathione. With glutathione in glyoxalase sequence containing an N-terminal extension and two ⌬␧ inserts that are lacking from both the bacterial and trypanoso- 20-fold excess, the 280 nm for the phenylglyoxal-glutathione adduct was estimated to be 0.76 Ϯ 0.02 mMϪ1⅐cmϪ1, which is in matid enzymes (Fig. 1). The trypanosomatid GLO1 sequences reasonable agreement with the value of 1.17 mMϪ1⅐cmϪ1 pre- are also grouped with proteobacterial and cyanobacterial GLO1 viously determined by this method (24). With phenylglyoxal in enzymes in phylogenetic analyses (data not shown). 20-fold excess, similar values of 0.83 Ϯ 0.1 mMϪ1⅐cmϪ1 for the glutathione-phenylglyoxal adduct and 1.04 Ϯ 0.04 mMϪ1⅐cmϪ1 Expression and Molecular Characterization of L. major GLO1. L. major for the trypanothione-phenylglyoxal adduct were obtained. GLO1 was expressed and purified to homogeneity by metal- affinity and ion-exchange chromatography. The mass of the These extinction coefficients then were used to determine Kd values. These values were 0.5 Ϯ 0.04 mM for the glutathione– protein was measured by mass spectrometry before and after Ϯ removal of the hexahistidine tag. Before cleavage, the mass was phenylglyoxal equilibrium and 0.5 0.03 mM for the trypano- Ϯ thione–phenylglyoxal equilibrium, close to the previously pub- 18,370 18 Da (predicted, 18,481 Da), with thrombin cleavage reducing this mass to 16,607 Ϯ 17 Da (predicted, 16,599 Da). No lished value of 0.6 Ϯ 0.05 mM for the glutathione–phenylglyoxal residual hexahistidine-tagged enzyme could be detected in the equilibrium (24). final preparation by mass spectrometry or SDS͞PAGE. Size- The values of ⌬␧ for the isomerization of the trypano- 240 nm exclusion chromatography against known standards showed the thione hemithioacetals of methylglyoxal and phenylglyoxal were protein to be a dimer in solution, with an M of 38,700 (predicted, determined by using L. major GLO1 to convert a known quantity r 33,200). The isoelectric point was also determined to be pH 4.9 of trypanothione to the corresponding 2-hydroxyacyl thiol ester. (predicted, pH 4.7). For the isomerization of the trypanothione–methylglyoxal and Ϯ phenylglyoxal hemithioacetals, these values were 3.0 0.3 and Metal Specificity of L. major GLO1. Ϯ Ϫ1⅐ Ϫ1 The cofactor specificity of L. 5.59 0.06 mM cm , respectively. Within experimental major GLO1 was investigated by reconstitution of apoenzyme error, these are the same as the values of 2.86 and 5.69 Ϫ1⅐ Ϫ1 with metal ions (Fig. 2). Of the metals tested, only manganese, mM cm found for the corresponding reactions with gluta- cobalt, and nickel reactivated the enzyme, with the largest thione (28). To determine whether hemithioacetals formed by recovery in activity being seen with nickel. No activity was seen either or both sulfydryl groups in trypanothione would act as with zinc, the cofactor of the human enzyme. After extensive substrates, reactions containing 1 mM of the trypanothione dialysis of the recombinant protein to remove loosely bound sulfydryl group were run to completion in the presence of 3 mM metals, 0.9 mol of nickel and 0.1 mol of zinc per mole of GLO1 methylglyoxal. The concentration of residual thiol then was dimer were detected by atomic absorption spectrophotometry. determined by using 5,5Ј-dithio-bis(2-nitrobenzoic acid). The Ϯ This stoichiometry suggests that one active site in the ho- addition of L. major GLO1 to these assays removed 99.6 0.2% modimeric protein does not tightly bind metal cofactor, an of the sulfydryl groups that were measurable before addition of observation also made with the E. coli enzyme (7). the enzyme, indicating that hemithioacetals formed by both sulfydryl groups of trypanothione can be isomerized by L. major Catalytic Activities of the L. major GLO1. With trypanothione hemi- GLO1. thioacetal as substrate, the L. major GLO1 is a highly efficient ͞ enzyme with kcat Km values approaching the diffusion limit Methylglyoxal-Catabolizing Activities in L. major. L. major extract (Table 2). These kinetic constants are comparable with pub- was assayed for enzymes capable of metabolizing methylglyoxal. lished values for the human enzyme using glutathione hemith- Only trypanothione-dependent GLO1 and NADPH-dependent ioacetal substrates (29). Significantly, the L. major enzyme shows methylglyoxal reductase activities could be detected, with the a marked preference for trypanothione hemithioacetal, with a ͞ GLO1 activity being 10-fold higher than the methylglyoxal 200-fold difference in kcat Km between this substrate and glu- reductase activity (Table 1). The specific activity of methylg- tathione hemithioacetal. The hemithioacetal formed with meth- lyoxal reductase in this extract agrees well with the previously ylglyoxal and the mono-thiol N1-glutathionylspermidine is also a Ϫ reported level of 4.8 Ϯ 0.7 nmol⅐min 1 per mg in L. donovani,as substrate for this enzyme (data not shown), suggesting that the does the lack of significant GLO1 activity with glutathione hemithioacetals formed by the two sulfydryl groups of trypano- hemithioacetal substrate (19). Because of the different initial thione could bind to the active site independently. This would be absorbances and wavelengths of the various assays, the maximum consistent with the observation that both of these groups are amount of protein that can be added varies. Hence, the assays are substrates for the L. major GLO1. This specificity of the L. major most sensitive for reductase͞dehydrogenases (limit of detection, GLO1 for glutathione amides such as trypanothione is also seen Ͻ0.4 nmol⅐minϪ1 per mg), less sensitive for GLO1 (limit of in the effect of glutathione derivatives on the activity of the detection, Ͻ5 nmol⅐minϪ1 per mg), and least sensitive for GLO2 enzyme. A range of S-alkyl and S-benzyl glutathione derivatives (limit of detection, Ͻ18 nmol⅐minϪ1 per mg). that have been characterized previously as GLO1 inhibitors were

13188 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402918101 Vickers et al. Downloaded by guest on September 30, 2021 Fig. 1. Alignment of trypanosomatid, human, and E. coli GLO1 sequences. The L. major, L. infantum, and T. cruzi GLO1 sequences are shown aligned with the human (Swiss-Prot accession no. P78375), E. coli (Swiss-Prot accession no. P77036), and Synechococcus sp. WH 8102 (TrEMBL accession no. Q7U3T2) enzymes. The alignment is shaded according to identity: white text on a black background, complete identity; white text on a dark-gray background, Ͼ75% identity; black text on a light-gray background, Ͼ50% identity. The positions of the active-site metal-binding residues in the human and E. coli enzymes are indicated under the sequences by asterisks.

tested against the L. major enzyme. The S-methyl, S-propyl, methylglyoxal synthases in a single unassembled sequence S-butyl, S-hexyl, and S-octyl glutathione derivatives were tested (LM32D2b09) in the L. major genome database, this sequence is at 1 mM, whereas S-p-nitrobenzyl and S-p-bromobenzylgluta- not found in other trypanosomatids. Indeed, this sequence seems to thione were tested at 250 ␮M. No inhibition was observed at be of Pseudomonas origin, and attempts to clone this gene from L. these concentrations (data not shown). major genomic DNA were unsuccessful. Previous studies on the metabolism of exogenous methylglyoxal Discussion by L. donovani showed that 98% was recovered as D-lactate, with Because methylglyoxal is produced from several different metabolic only 1.2% being recovered as lactaldehyde (19). These data are intermediates, almost all cells will be exposed to its toxic effects and entirely consistent with our observation of 10-fold higher trypano- thus require a methylglyoxal detoxification system. Methylglyoxal thione-dependent GLO1 than methylglyoxal reductase activity in synthase may produce an additional toxic challenge, but the exis- extracts and also indicate that a GLO2-like activity must be present tence of this activity in trypanosomatids is controversial (18, 19). in these cells. The relative specific activities of methylglyoxal-

Although we identified an ORF with high similarity to prokaryotic catabolizing enzymes in L. major therefore suggest that this parasite BIOCHEMISTRY has a highly active glyoxalase system that differs from other organisms in being trypanothione-dependent rather than glutathione-dependent (Fig. 3). Although the L. major genome contains a putative GLO2 (LMFLCHR12࿝92), which is 51% identical to T. brucei GLO2 at the amino acid level (20), we were unable to detect any GLO2 activity in cell extracts of L. major, which may be because of the insensitivity of the assay method or instability of the enzyme. T. brucei GLO2 readily loses activity because of loss of its metal cofactor (20), and thus L. major GLO2 may have lost activity during dialysis to remove trypanothione and glutathione cosubstrates. BLAST

Table 2. Kinetic parameters of the L. major GLO1

7 Hemithioacetal kcat͞Km ϫ 10 , kcat͞Km Ϫ1 Ϫ1 Ϫ1 substrate Km, ␮M kcat,s M ⅐s (relative)

Trypanothione Methylglyoxal 32 Ϯ 3 800 Ϯ 30 2.5 100 Phenylglyoxal 51 Ϯ 7 570 Ϯ 40 1.1 44 Fig. 2. Metal reconstitution of the L. major GLO1. Reactivation of the apo Glutathione form of the L. major GLO1 by a 100-fold excess of metal ions. Assays contained Methylglyoxal Ͼ1900 ND 0.013 0.5 0.1 mM trypanothione hemithioacetal and 0.1 mM free trypanothione-SH. All Phenylglyoxal Ͼ80 ND 0.008 0.3 assays were performed in triplicate, with the mean activity relative to the nickel-activated enzyme displayed. ND, not determined.

Vickers et al. PNAS ͉ September 7, 2004 ͉ vol. 101 ͉ no. 36 ͉ 13189 Downloaded by guest on September 30, 2021 Fig. 3. Metabolism of methylglyoxal in Leishmania spp. The major end product is D-lactate (98%) with a trace amount of L-lactaldehyde (2%) formed via methylglyoxal reductase (MEGR) (19). Enzymatic formation of methylglyoxal via methylglyoxal synthase is controversial (18).

searches of the T. cruzi genome identified homologues of GLO1 The high level of specificity of the L. major GLO1 for (Tc00.1047053510659.240 and Tc00.1047053510743.70) and GLO2 trypanothione hemithioacetals defines this enzyme as a (Tc00.1047053507603.230 and Tc00.1047053509429.290). The two trypanothione-dependent GLO1 (lactoyltrypanothione lyase, GLO2 homologues differ by three residues at the amino acid level, EC unassigned). This substrate requirement and the insensi- with both having 66% identity to T. brucei GLO2. There are no tivity of the L. major enzyme to glutathione-based inhibitors publications on glyoxalase activity in T. cruzi, but the presence of are highly significant, because this specificity contrasts with these genes suggests that the glyoxalase pathway may be present. that of the human enzyme. In human and S. cerevisiae GLO1, Unexpectedly, we were unable to identify a strong homologue of the glutathione glycine carboxylate makes a significant con- GLO1 in the almost completed T. brucei brucei genome. Although tribution to the binding of substrates and transition-state this result is consistent with an earlier report that a closely related analogues (36, 37). Binding occurs through an interaction of subspecies (T. brucei gambiense) does not produce D-lactate (30), the carboxylate with main-chain atoms of a flexible loop the presence of a functional GLO2 in T. brucei brucei (20) suggests formed from residues Pro-152–Gly-159 in the human enzyme that GLO1 might be present because the genome sequence for this (8). This flexible loop is also present in the E. coli structure organism is incomplete. Additional studies are required to resolve (Pro-102–Val-109) (9), but four of these eight residues are this question. absent from the corresponding region in the trypanosomatid The high degree of sequence similarity and the conservation sequences (Fig. 1). The lack of sequence conservation suggests of the predicted metal-binding residues between the trypanoso- that structural differences in this part of the L. major protein matid and the E. coli GLO1 enzymes suggest that the parasite may confer specificity for trypanothione hemithioacetals. An- enzymes all might be nickel-dependent. However, the cofactor other structural difference that may affect binding of hemi- requirement of a GLO1 enzyme cannot be predicted from thioacetal substrates is seen between the E. coli and human sequence alone, because the zinc-dependent Pseudomonas enzymes. The E. coli active site is significantly larger (9), and putida and S. cerevisiae enzymes have identical metal-binding it is possible that it therefore may allow binding of not only the residues as those of the E. coli enzyme (31, 32). Thus, the hemithioacetal with glutathione but also with glutathionyl unambiguous demonstration of the requirement of the L. major spermidine, a metabolite that represents Ͼ80% of glutathione- GLO1 for nickel, but not zinc, by metal analysis and reconsti- containing thiols in stationary-phase E. coli cells (38). tution is important in confirming the unique relationship of these The contrasting substrate specificities of the human and eukaryotic glyoxalases to the E. coli enzyme. Importantly, the trypanosomatid enzymes suggest that GLO1 may be an attractive similarity in sequence to cyanobacterial enzymes suggests that target for drug design. However, gene-knockout and inhibitor the presence of these genes in the trypanosomatids may be the studies are required to determine whether this pathway plays an result of gene transfer from an ancestral endosymbiont (33). It essential role in parasite growth, survival, or virulence. is interesting that a rice GLO1 that is of bacterial origin and belongs to a group of plant glyoxalases with strong sequence We thank Dean Wilson for performing metal analysis, Douglas Lamont similarity to the bacterial enzymes has been identified (34). for performing mass spectrometry, and John Richardson for assistance However, these plant enzymes are twice the size of the dimeric with organic synthesis. The use of the genome sequence data from the bacterial enzymes and therefore are probably monomeric GLO1 Wellcome Trust Sanger Institute Pathogen Sequencing Unit (www. enzymes more closely related to those of S. cerevisiae and genedb.org) is gratefully acknowledged. This work was funded by the Plasmodium sp. (35) than those of the trypanosomatids. Wellcome Trust.

13190 ͉ www.pnas.org͞cgi͞doi͞10.1073͞pnas.0402918101 Vickers et al. Downloaded by guest on September 30, 2021 1. Thornalley, P. J. (1996) Gen. Pharmacol. 27, 565–573. 21. Pourmotabbed, T. & Creighton, D. J. (1986) J. Biol. Chem. 261, 14240–14244. 2. Cooper, R. A. (1984) Annu. Rev. Microbiol. 38, 49–68. 22. Vince, R., Daluge, S. & Wadd, W. B. (1971) J. Med. Chem. 14, 402–404. 3. Lo, W. C. T., Westwood, M. E., McLellan, A. C., Selwood, T. & Thornalley, 23. Alexander, N. M. & Boyer, J. L. (1971) Anal. Biochem. 41, 29–38. P. J. (1994) J. Biol. Chem. 269, 32299–32305. 24. Vander Jagt, D. L., Han, L. P. & Lehman, C. H. (1972) Biochemistry 11, 4. Marinari, U. M., Ferro, M., Sciaba, L., Finollo, R., Bassi, A. M. & Brambilla, 3735–3740. G. (1984) Cell Biochem. Funct. 2, 243–248. 25. Ariyanayagam, M. R. & Fairlamb, A. H. (2001) Mol. Biochem. Parasitol. 115, 5. Davidson, G., Clugston, S. L., Honek, J. F. & Maroney, M. J. (2001) 189–198. Biochemistry 40, 4569–4582. 26. Beutler, E., Duron, O. & Kelly, B. M. (1963) J. Lab. Clin. Med. 61, 882–888. 6. Aronsson, A. C., Marmstal, E. & Mannervik, B. (1978) Biochem. Biophys. Res. 27. Collier, B. (1973) Anal. Biochem. 56, 310–311. Commun. 81, 1235–1240. 28. Vander Jagt, D. L., Daub, E., Krohn, J. A. & Han, L. P. (1975) Biochemistry 7. Clugston, S. L., Barnard, J. F., Kinach, R., Miedema, D., Ruman, R., Daub, E. 14, 3669–3675. & Honek, J. F. (1998) Biochemistry 37, 8754–8763. 29. Ridderstrom, M. & Mannervik, B. (1996) Biochem. J. 314, 463–467. 8. Cameron, A. D., Ridderstrom, M., Olin, B., Kavarana, M. J., Creighton, D. J. 30. Darling, T. N., Balber, A. E. & Blum, J. J. (1988) Mol. Biochem. Parasitol. 30, & Mannervik, B. (1999) Biochemistry 38, 13480–13490. 253–257. 9. He, M. M., Clugston, S. L., Honek, J. F. & Matthews, B. W. (2000) Biochemistry 31. Saint-Jean, A. P., Phillips, K. R., Creighton, D. J. & Stone, M. J. (1998) 39, 8719–8727. Biochemistry 37, 10345–10353. 10. Sellin, S. & Mannervik, B. (1983) J. Biol. Chem. 258, 8872–8875. 32. Frickel, E. M., Jemth, P., Widersten, M. & Mannervik, B. (2001) J. Biol. Chem. 11. Ayoub, F. M., Allen, R. E. & Thornalley, P. J. (1993) Leuk. Res. 17, 397–401. 276, 1845–1849. 12. Egyud, L. G. & Szent-Gyorgyi, A. (1968) Science 160, 1140. 13. Thornalley, P. J., Strath, M. & Wilson, R. J. (1994) Biochem. Pharmacol. 47, 33. Hannaert, V., Saavedra, E., Duffieux, F., Szikora, J. P., Rigden, D. J., Michels, 418–420. P. A. M. & Opperdoes, F. R. (2003) Proc. Natl. Acad. Sci. USA 100, 1067–1071. 14. D’Silva, C., Daunes, S., Rock, P., Yardley, V. & Croft, S. L. (2000) J. Med. 34. Usui, Y., Nakase, M., Hotta, H., Urisu, A., Aoki, N., Kitajima, K. & Matsuda, Chem. 43, 2072–2078. T. (2001) J. Biol. Chem. 276, 11376–11381. 15. Fairlamb, A. H. (2003) Trends Parasitol. 19, 488–494. 35. Iozef, R., Rahlfs, S., Chang, T., Schirmer, H. & Becker, K. (2003) FEBS Lett. 16. Croft, S. L. & Coombs, G. H. (2003) Trends Parasitol. 19, 502–508. 554, 284–288. 17. Fairlamb, A. H., Blackburn, P., Ulrich, P., Chait, B. T. & Cerami, A. (1985) 36. Hamilton, D. S. & Creighton, D. J. (1992) Biochim. Biophys. Acta 1159, Science 227, 1485–1487. 203–208. 18. Darling, T. N. & Blum, J. J. (1988) Mol. Biochem. Parasitol. 28, 121–127. 37. Murthy, N. S., Bakeris, T., Kavarana, M. J., Hamilton, D. S., Lan, Y. & 19. Ghoshal, K., Banerjee, A. B. & Ray, S. (1989) Mol. Biochem. Parasitol. 35, Creighton, D. J. (1994) J. Med. Chem. 37, 2161–2166. 21–29. 38. Smith, K., Borges, A., Ariyanayagam, M. R. & Fairlamb, A. H. (1995) Biochem. 20. Irsch, T. & Krauth-Siegel, R. L. (2004) J. Biol. Chem. 279, 22209–22217. J. 312, 465–469. BIOCHEMISTRY

Vickers et al. PNAS ͉ September 7, 2004 ͉ vol. 101 ͉ no. 36 ͉ 13191 Downloaded by guest on September 30, 2021