Monothiol Glutaredoxin-1 Is an Essential Iron-Sulfur Protein in the Mitochondrion of African Trypanosomes*

African trypanosomes encode three monothiol glutaredoxins (1-C-Grx). 1-C-Grx1 occurs exclusively in the mitochondrion, and 1-C-Grx2 and -3 are predicted to be mitochondrial and cytosolic proteins, respectively. All three 1-C-Grx are expressed in both the mammalian bloodstream and the insect procyclic form of Trypanosoma brucei, with the highest levels found in stationary phase and starving parasites. In the rudimentary mitochondrion of bloodstream cells, 1-C-Grx1 reaches concentrations above 200 μm/subunit. Recombinant T. brucei 1-C-Grx1 exists as a noncovalent homodimer, whereas 1-C-Grx2 and 1-C-Grx3 are monomeric proteins. In vitro, dimeric 1-C-Grx1 coordinated an H2O2-sensitive [2Fe-2S] cluster that required GSH as an additional ligand. Both bloodstream and procyclic trypanosomes were refractory to down-regulation of 1-C-Grx1 expression by RNA interference. In procyclic parasites, the 1-c-grx1 alleles could only be deleted if an ectopic copy of the gene was expressed. A 5–10-fold overexpression of 1-C-Grx1 in both parasite forms did not yield a growth phenotype under optimal culture conditions. However, exposure of these cells to the iron chelator deferoxamine or H2O2, but not to iron or menadione, impaired cell growth. Treatment of wild-type bloodstream parasites with deferoxamine and H2O2 caused a 2-fold down- and up-regulation of 1-C-Grx1, respectively. The results point to an essential role of the mitochondrial 1-C-Grx1 in the iron metabolism of these parasites.

and raises the question of the function(s) of the proteins in trypanosomatids.
Here we report on the structural and cell biological characterization of the three 1-C-Grx in the different life and growth stages of African trypanosomes. We show that the concentration of these proteins is highest in parasites that are grown to the stationary/starvation phase. For the mitochondrial 1-C-Grx1, we provide strong evidence that the homodimeric protein can coordinate an iron-sulfur center using GSH as ligand, fulfilling an essential role in the iron and redox homeostasis of the parasite.

EXPERIMENTAL PROCEDURES
Reagents-All enzymes employed in the molecular biology experiments were purchased from MBI Fermentas. Kits for the purification of genomic DNA, gel extraction, and mini-and midi-preps were from Qiagen or Macherey-Nagel. Oligonucleotides were synthesized by MWG Biotech, Operon, or Metabion. E. coli strains DH5-␣ and Novablue served as host cells in all genetic procedures. Glutathione disulfide, menadione, and 5,5Ј-dithiobis-(2-nitrobenzoic acid), used for thiol determinations (15), were purchased from Sigma. DNase, complete mini protease inhibitor mixture, and recombinant Aspergillus niger glucose oxidase were from Roche Applied Science and hydrogen peroxide and isopropyl-␤-D-thiogalactopyranoside from Merck. Antibiotics were purchased from Sigma and Invitrogen. Polyclonal guinea pig antibodies against T. brucei His 6 -1-C-Grx1, untagged 1-C-Grx2, and 1-C-Grx3 were produced by Eurogentec. The antisera against T. cruzi trypanothione reductase (TR) and T. brucei 1-C-Grx1 were purified (13). Mouse monoclonal anti-c-Myc (clone 9E10) antibodies were purchased from Roche Applied Science. Anti-guinea pig and antirabbit IgG secondary antibodies conjugated with horseradish peroxidase were obtained from Santa Cruz. Alexa Fluor488labeled goat anti-guinea pig IgG and anti-mouse IgG were purchased from Molecular Probes TM (Invitrogen). His-tagged 1-C-Grx1 was prepared essentially as described (13). Before induction, the cultures were kept at 4°C for 30 min, 200 M isopropyl-␤-D-thiogalactopyranoside, was added and the cells were incubated at 18°C overnight. The pH of all buffers was 7.0.
Cloning, Expression, and Purification of His-tagged 1-C-Grx2 and 1-C-Grx3-The coding regions of 1-c-grx2 and 1-c-grx3 were amplified from T. brucei (strain 427) (16) genomic DNA with the primers given in the supplemental table. The amplicon for 1-c-grx2 was blunt end-cloned using the Perfectly Blunt cloning kit (Novagen), which yielded plasmid pETBlue1-1-c-grx2. This construct served as template to amplify 1-c-grx2 without the first 21 residues. The PCR product was BamHI/ PstI-digested and ligated into the vector pQE30 (Qiagen) yielding pQE30-1-c-grx2(Ϫ)MTS. The PCR product for 1-c-grx3 was ligated into the pET19b vector (Novagen). All expression constructs were completely sequenced in both directions. Expression of pQE30-1-c-grx2(Ϫ)MTS and pET19b-1-c-grx3 yielded recombinant His 6 -1-C-Grx2 and His 10 -1-C-Grx3, respectively. The plasmids were overexpressed in E. coli BL21-DE3 cells grown in 2ϫ YT medium and Tuner-DE3 cells cultivated in Terrific Broth medium with 1% (w/v) glucose, respectively. The cultures containing 100 g/ml carbenicillin were incubated at 37°C and 180 rpm. At an A 600 of 0.6, the temperature was lowered to 4°C for 30 min, 0.2-1 mM isopropyl-␤-Dthiogalactopyranoside was added, and the cells were grown at 18°C overnight. Cells from 1 to 2 liters of culture were resuspended in 25-50 ml of 300 mM NaCl, 50 mM NaH 2 PO 4 , pH 8.0, and disintegrated by sonification. The supernatant was applied onto 10 ml of nickel-nitrilotriacetic acid Superflow resin (Qiagen). After washing with 10 mM imidazol, the His-tagged proteins were eluted with a linear gradient of 10 -250 mM imidazol. Fractions containing pure recombinant proteins were pooled, and the protein concentration was determined.
Gel Chromatography and Mass Spectrometric Analysis of 1-C-Grx1 Treated with GSSG or Hydrogen Peroxide-In a total volume of 5 ml of PBS buffer, pH 7.4, 50 M tag-free 1-C-Grx1 (expressed from a modified pET9d vector (New England Biolabs) containing a tobacco etch virus (TEV) protease-cleavable thioredoxin-His 6 fusion vector (kindly provided by Gunther Stier, EMBL, Heidelberg, Germany)) was incubated with 500 M GSSG overnight at 4°C and with 500 M H 2 O 2 for 2 h at room temperature, respectively. The latter reaction was stopped by adding 0.1 unit of bovine liver catalase (Fluka). The GSSG-treated sample was subjected to two PD-10 columns (GE Healthcare) to remove the low molecular mass components and stored at 4°C for 3 days. Both 1-C-Grx1 samples contained 0.3 free thiol/protein subunit as shown by Ellman's reaction (15). The protein solutions were concentrated in Amicon 10-kDa concentrators (Millipore) to 3.6 and 4.0 mg/ml for the subsequent size exclusion chromatography. A Superose 12 HR 10/30 column (GE Healthcare) was pre-equilibrated in 100 mM NaCl, 50 mM NaH 2 PO 4 , pH 8.0. 200-l samples of the oxidized 1-C-Grx1 species as well as of the untreated fully reduced protein, corresponding to about 0.8 mg of protein, were chromatographed at a flow rate of 0.2 ml/min at 8°C. The low molecular weight gel filtration calibration kit (GE Healthcare) served for standardization. The elution fractions were analyzed by SDS-PAGE under nonreducing and reducing conditions. The peak fractions were subjected to ESI-MS on an API-QSTAR TM Pulsar instrument (Applied Biosystems) with a high pressure liquid chromatography system (Agilent) on-line-coupled to the ESI-QTOF (quadrupole time-of-flight) instrument. 50 l of a 10 M protein solution was incubated with 20 mM iodoacetamide and injected into the mass spectrometer essentially as described previously (12).
Reconstitution Assay and [Fe-S] Cluster Stability-In vitro reconstitution of 1-C-Grx1 was performed as described (17). 100 -200 M 1-C-Grx1 was incubated under an argon atmosphere at room temperature with 0.01 molar equivalents of E. coli cysteine desulfurase IscS, 2.5-5 equivalents of cysteine, 2-4 equivalents of Fe(NH 4 ) 2 (SO 4 ) 2 , 1 mM GSH, 5 mM DTT, and 10 M pyridoxal phosphate in 50 mM sodium phosphate buffer, pH 8.0, containing 200 mM NaCl. After 2-3.5 h, the protein was separated from the reaction components on Sephadex G25 columns (NAP-5 or PD-10, GE Healthcare). The reconstitution mixture was subjected to gel chromatography on a Superdex 200 10/300 GL column (GE Healthcare) equilibrated in 50 mM sodium phosphate buffer, pH 8.0, containing 200 mM NaCl, and 2 mM GSH. The molecular mass of the complex was determined by a standard curve provided by the supplier's manual and by standard proteins. Absorbance at 280 and 420 nm was recorded. UV-visible spectra and the kinetics of cluster disassembly, measuring the absorbance at 420 nm in the presence or absence of 1 mM GSH, as well as of 0, 0.01, 0.1, 0.36, and 1 mM H 2 O 2 , were recorded on a Shimadzu UV-2100 spectrophotometer. Kinetic measurements were performed in 50 mM sodium phosphate buffer, pH 8.0, containing 200 mM NaCl at 25°C.
Determination of Iron and Acid-labile Sulfide-Protein bound iron was quantified as described by Fish (18), and acidlabile sulfide was quantified according to Broderick et al. (19). Three determinations were carried out on samples from five independent reconstitution assays.
Cell Lines, Culture Conditions, and Transfection-The T. brucei cell line 449 (referred to here as wild type, encoding one copy of the Tet repressor protein, Phleo R ) (20) and cell line 514 -1313 (encoding the T7-RNA Pol and two copies of the Tet repressor protein, Phleo R and Neo R , respectively) (21) employed in this work derive from strain 427 (16). Bloodstream cells were grown in HMI-9 medium (22) at 37°C in a humidified atmosphere with 5% CO 2 and procyclic T. brucei in MEM-Pros medium (Biochrom) containing 7.5 g/ml hemin at 27°C. Both culture media were supplemented with 10% (v/v) fetal calf serum (Biochrom, unless otherwise stated), 50 units/ml penicillin, and 50 g/ml streptomycin. Depending on the resistance gene harbored by the trypanosomes, the medium of bloodstream and procyclic cells contained 0.2 and 0.5 g/ml phleomycin, respectively, and/or 2.5 and 15 g/ml G-418, respectively.
3 ϫ 10 7 Trypanosomes were transfected with 10 -100 g of the DNA constructs according to standard techniques (23,24) in an ECM 630 electroporator (BTX). Single clones of bloodstream and procyclic cells were obtained in 24-wells culture plates by splitting the cultures immediately after transfection and by limiting dilution 24 h after transfection, respectively. Selection of stable transfectants was initiated about 16 h after electroporation by adding the respective antibiotics.
Generation and Growth Phenotype Analysis of 1-C-Grx1overexpressing T. brucei Cells-T. brucei bloodstream (cell line 449) and procyclic (cell line 514-1313) were transfected with pHD1700-1-c-grx1-c-myc 2 and selected with 10 and 50 g/ml hygromycin, respectively. Ectopic expression was induced with tetracycline (0.01, 0.1, 1, 10, 100, and 1000 ng/ml) and evaluated by Western blotting as described below. Bloodstream parasites were inoculated at 1 ϫ 10 5 cells/ml in 10 ml of fresh medium with (Tet-induced) or without (control) 1 g/ml tetracycline. The cells were counted daily, and the cultures were diluted to the initial cell density with fresh medium with or without tetracycline. In the case of procyclic parasites, the starting cell density was 5 ϫ 10 5 cells/ml. The medium was not replaced during cultivation, and the cell density was assessed at different time points. Sustained expression of 1-C-Grx1-c-Myc 2 was achieved by adding 1 g/ml tetracycline every 48 h.
Generation and Analysis of 1-c-grx1 Knock-out T. brucei Cell Lines-Bloodstream and procyclic T. brucei cells were transfected with the BLA-KO cassette and grown in the presence of 5 and 10 g/ml blasticidin, respectively. In procyclic cells, replacement of the second 1-c-grx1 allele was attempted in two ways. A loss of heterozygosity was induced by stepwise increasing the concentration of blasticidin (10, 70, 150 and 300 g/ml). The cells were seeded at 5 ϫ 10 5 cells/ml, cultivated for 4 to 7 days in the presence of the selecting drug, and diluted back to the initial cell density before increasing the concentration of the antibiotic. Alternatively, 1 ϫ 10 6 cells/ml were grown for 1 week in the presence of 300 g/ml blasticidin, the culture was diluted 1:2, and cultivation was continued. In the second approach, BLA-resistant cells were transfected with the PAC-KO cassette and grown in the presence of puromycin (2 g/ml) and blasticidin (10 g/ml).
To generate 1-c-grx1 conditional KOs of procyclic T. brucei, cell line 514-1313 Tb1-c-grx1 Ti HYG (clone B5) displaying a tightly regulated expression of the ectopic copy of 1-c-grx1 was first transfected with the BLA-KO cassette. Cell cloning in a 24-well microculture plate was performed as described above in the presence of 10 g/ml blasticidin. Loss of the second allele was achieved by transfecting a BLA-resistant cell line with the PAC-KO cassette in the presence of 1 g/ml tetracycline (replaced every 48 -72 h), 10 g/ml blasticidin, and 2 g/ml puromycin. The successful replacement of the 1-c-grx1 gene was verified by PCR and Western blot analysis (described below). Genomic DNA was subjected to PCR using oligonucleotides complementary to the open reading frames upstream and downstream of the 1-c-grx1 gene in combination with 1-c-grx1 specific primers (see supplemental table). Genomic DNA from wild-type cells served as negative control. PCR products were analyzed on a 1% agarose gel. Procyclic cells with a single allele replaced and wild-type parasites (control) were grown continuously, and the cell density was determined daily. The growth of procyclic trypanosomes with a conditional-1-c-grx1-KO genotype was followed in medium with Tet-free certified fetal calf serum (Cambrex). Steady expression of 1-c-grx1c-myc 2 in the control cultures was achieved by adding fresh tetracycline (1 g/ml) every 48 h. On day 0, 5 ϫ 10 5 cells/ml were diluted in 10 ml of medium and incubated at 27°C. For a 6and 15-days experiment, every second and fifth day, respectively, the cells were counted in a Neubauer chamber, samples were withdrawn for further analysis, and the cultures were diluted to the initial cell density.
Immunofluorescence-2.5 ϫ 10 6 Bloodstream and procyclic parasites from logarithmic phase cultures, 24 -48 h after induction of 1-C-Grx1-c-Myc 2 expression, were harvested by centrifugation at 2000 ϫ g for 10 min at 27°C and incubated for 15 min at 27 and 37°C in 10 ml of fresh medium containing 0.2 and 0.75 M MitoTrackerRed CMXRos (Molecular Probes), respectively. Subsequently, the cells were washed with PBS, suspended in 10 ml of fresh medium, cultivated for 20 min, and again washed. After fixation in 1 ml PBS, 4% (w/v) paraformaldehyde for 18 min at room temperature, the cells were pelleted at 5000 rpm for 1 min and washed three times with 500 l of PBS, resuspended in 800 l of PBS, and allowed to settle on a 4-well culture slide (200 l/well; BD Falcon TM ) by overnight incubation at 4°C. The cells were permeabilized with 0.2% (v/v) Triton X-100 in PBS for 20 min, washed twice with PBS, and blocked for 20 min with 0.5% (w/v) gelatin in PBS. After washing with PBS, the slides were incubated for 1 h with purified guinea pig anti-1-C-Grx1 or mouse monoclonal anti-c-Myc (1:400 in PBS), washed again with PBS, and then incubated for 1 h with the secondary antibodies (Alexa Fluor 488-labeled goat anti-guinea pig IgG and anti-mouse IgG, dilution 1:250). The parasite DNA was stained with 500 ng/ml (500 l/well) 4Ј,6-diamidino-2-phenylindole dihydrochloride. The cells were examined under a Leica DMRXA fluorescence microscope at ϫ1000 magnification, and images were recorded with a digital charge-coupled device camera (Hamamatsu).
Experiments Involving Iron Homeostasis and Oxidative Stress-Constant expression of 1-C-Grx1-c-Myc 2 in bloodstream T. brucei was achieved by adding 1 g/ml tetracycline 24 h prior to and throughout the experiments. The initial cell density was adjusted to 1 ϫ 10 6 and 5 ϫ 10 5 cells/ml for the assays involving iron and oxidative challenges, respectively. In all experiments, untreated Tet-induced and noninduced cells served as controls. Upon addition of the different stressors, the parasites were incubated under optimal culture conditions, and viable cells were counted in a Neubauer chamber after 16 -24 h.
Data Analysis-If not stated otherwise, the experimental data presented here refer to the analysis of at least two different clones in at least three independent experiments. A two-tailed Student's t test was applied for statistical analysis.

Subunit Composition of the T. brucei Monothiol Glutaredoxins-
The genes encoding T. brucei 1-C-Grx1, 1-C-Grx2, and 1-C-Grx3 were cloned and overexpressed in E. coli both as N-terminally His-tagged proteins and as tag-free proteins. Gel chromatography of the His-tagged proteins revealed 1-C-Grx1 dimers (13), but 1-C-Grx2 and 1-C-Grx3 occurred in mixtures of monomeric and oligomeric forms (not shown). To evaluate whether the oligomerization was an artifact caused by the N-terminal histidine stretch, the tag-free proteins were prepared and subjected to gel chromatography. In the case of 1-C-Grx1 the tag-free protein, containing two free thiol groups per monomer, also eluted with an apparent molecular mass of 31-33 kDa, corresponding to a dimer (theoretical mass of the subunit 16.1 kDa; Fig. 1). The noncovalent nature of the dimer was verified by SDS-PAGE under nonreducing conditions, which revealed only the monomer (Fig. 1C). Because 1-C-Grx1 is susceptible to specific thiolation of Cys 184 and can form an intramolecular disulfide bridge when treated with GSSG (12), we considered whether the oligomeric state of the protein was dependent on its redox state. 1-C-Grx1 was treated with GSSG or H 2 O 2 after which removal of the low molecular mass components resulted in samples with 0.3 free thiol group/monomer instead of the 2 free cysteines present in the untreated protein.
Treatment of the protein with either GSSG or H 2 O 2 led to the appearance of two major peaks with apparent masses of about 32,000 and 48,000 or 44,000 Da as well as high molecular mass species (Fig. 1, D and E). SDS-PAGE under nonreducing conditions revealed that all high molecular mass fractions represent covalent polymers. In contrast, both the 32,000-and the 44/48,000-Da fractions contained nearly exclusively monomeric protein, independent of the presence or absence of DTT (not shown). To reveal the molecular nature of the protein species, the samples were subjected to ESI-MS analysis ( Table 1). The fraction eluting from the gel chromatography column with a mass of 32,000 Da (Fig. 1D, elution fraction F3) showed for the GSSG-treated 1-C-Grx1 a main peak (16,362 Da) that corresponded to the reduced protein carboxamidomethylated at both cysteines. In the case of the H 2 O 2 -treated 1-C-Grx1, a second prominent peak at 16,337 Da corresponded to 1-C-Grx1 with two oxygen atoms and one carboxamidomethyl group bound. This species most probably represents 1-C-Grx1, which was oxidized to the sulfinic acid state at Cys 104 and alkylated at Cys 181 (12). The fractions of the GSSG-or H 2 O 2 -treated 1-C-Grx1 species that eluted with an apparent mass of 48,000 and 44,000 Da (Fig.  1D, elution fraction F1), respectively, contained exclusively or nearly exclusively the protein with intramolecular disulfide bridge (16,246 Da). A monomeric peak was never observed for the native protein; therefore, we can rule out the possibility that the 44 -48-kDa peaks correspond to a trimeric species (see also next paragraph). Instead, the formation of the intramolecular disulfide in 1-C-Grx1 likely induced a conformational change in the protein that altered its elution profile toward a species with apparent higher molecular masses.
Tag-free, reduced 1-C-Grx2 and 1-C-Grx3 eluted from the Superose 12 HR (Fig. 1A) as well as a Hi-load 26/60 Superdex 75 column (not shown) in a single peak with an apparent mass of 18 -20 and 28 -32 kDa, respectively (Fig. 1B). These masses are higher than the theoretical masses of 12.6 and 24.5 kDa for the subunits but are clearly below the mass for the respective dimers, which may indicate an overall nonglobular structure. Such unusual behavior upon gel chromatography has been described for other glutaredoxins. A plant protein was reported to run with apparent masses of 34,000 and 17,000 Da for the dimer and monomer in comparison with the theoretical masses of 25,000 and 12,500 Da, respectively (26).
Identification of the [Fe-S] Cluster in 1-C-Grx1-His 6 -1-Cys-Grx1 was purified as a slightly colored protein. In addition to the absorption maximum around 280 nm, the protein displayed peaks at 320 and 420 nm ( Fig. 2A). These spectral properties resembled those described before for glutaredoxins coordinating [2Fe-2S] clusters (10,11,25,26) and became more evident after in vitro reconstitution of the cluster ( Fig. 2A). Assembly (not shown) and stabilization of the Fe-S cluster in 1-C-Grx1 proved to require GSH as cofactor (Fig. 2B). Hydrogen peroxide triggered the disassembly of the cluster from 1-C-Grx1 in a concentration-dependent manner (Fig. 2C). By colorimetric methods we determined that 0.78 Ϯ 0.29 mol of iron and 1.15 Ϯ 0.22 mol of acid-labile sulfides were bound per mol of monomeric protein, which is consistent with the presence of one [2Fe-2S] center/protein dimer. The molecular mass of the com-  (--) is included for comparison. Elution fractions (F1, 12.5-13 ml; F2, 13-13.5 ml; and F3, 13.5-14 ml) that were subjected to mass spectrometry (see Table 1) are framed with a gray background. E, molecular mass determination for the major peaks of GSSG (‚)-and H 2 O 2 -treated (ƒ) as well as untreated reduced (E) 1-C-Grx1 with the standard proteins depicted as black dots. plex was determined by subjecting a reconstitution mixture to gel filtration and recording the elution profile at 280 and 420 nm. Except for aggregates (Ͼ450 kDa) and very small fragments (Ͻ6.5 kDa), only one protein with absorbance at 420 nm eluted from the column (Fig. 2D). The protein had an apparent molecular mass of about 34 kDa (Fig. 2E), and a UV-visible spectrum confirmed the presence of a [2Fe-2S] cluster (not shown).
Expression Pattern and Cellular Concentration of the 1-C-Grx in T. brucei-In both bloodstream and procyclic trypanosomes, the concentration of 1-C-Grx1 was highest in stationary and starving parasites (Fig. 3). The significance of this finding is underlined by the fact that the concentration of the cytosolic trypanothione reductase remained constant or even decreased in the final stage. The concentration of 1-C-Grx2 in logarithmic bloodstream cells was almost below the detection limit, but the protein became detectable in the stationary phase. 1-C-Grx3 proved to be a rather abundant protein with the highest concentration also in stationary parasites.
Different amounts of the recombinant tag-free 1-C-Grx were subjected to Western blot analysis and served as standards to estimate the cellular concentration of the three proteins in the parasites. With a cell volume of 58 femtoliters for bloodstream trypanosomes (27), the cellular concentrations of 1-C-Grx1, 1-C-Grx2, and 1-C-Grx3 ranged from 5 to 30 M, up to 0.5 M, and 1 to 6 M, respectively (Fig. 3B). In procyclic parasites (cell volume of 96 femtoliters), 4 the cellular concentrations of 1-C-Grx1, 1-C-Grx2, and 1-C-Grx3 were calculated to be 2 to 6 M, up to 0.5 M, and 0.5 to 3 M, respectively. Taking into account the exclusive mitochondrial localization of 1-C-Grx1 (13) and the published relative volume of the mitochondrion in bloodstream and procyclic T. brucei, which represents 2.3% (28) and 25% (27), respectively, of the total cell volume, the concentration of 1-C-Grx1 in the organelle ranges from 0.2 to 1.3 mM and 8 to 24 M, respectively, depending on the growth stage.
Expression of 1-C-Grx1 Could Not be Down-regulated by RNA Interference-Any attempt to deplete the mRNA of 1-c-grx1 in bloodstream or procyclic T. brucei by RNA interference using different vector/host cell combinations failed. We obtained viable transfectants that were resistant to the selection marker, but none of the cell lines showed significant depletion of 1-C-Grx1 as observed by immunoblot when the formation of specific double-stranded RNA was induced (not shown). PCR on genomic DNA from different clones verified the integrity of the RNA interference cassette, and sequencing of the amplicon confirmed the absence of mutations in the RNA interference control elements, namely the T7 promoter and the tetracycline operon (not shown). The failure to deplete the 1-C-Grx1 mRNA was probably caused by epigenetic mechanisms or by selection of cell lines with a defect in the RNA interference machinery (29), as it has been observed for other essential genes of T. brucei (24, 30 -33).
Gene Deletion of 1-c-grx1 in Procyclic T. brucei-We next attempted to replace the 1-c-grx1 alleles by a classical double-KO approach (Fig. 4A). In the case of bloodstream T. brucei, we have not been able thus far to isolate any viable cell line lacking a 1-c-grx1 allele. With procyclic trypanosomes, 1-c-grx1 single knock-out clones were obtained. In most of these clones, replacement of one allele with the blasticidin resistance gene correlated with a significantly lower level of 1-C-Grx1 (Fig. 4B). This depletion of 1-C-Grx1 had no effect on the growth rate (not shown). To delete the second allele, two strategies were followed. In the first approach, the blasticidin concentration was increased in steps up to 300 g/ml (30-fold higher than the concentration commonly used) to induce a loss of heterozygosity. This resulted in resistant cell lines all of which had retained a copy of 1-c-grx1 and expressed the protein at levels of about 50% when compared with wild-type cells (not shown). If the blasticidin concentration was increased directly from 5 to 300 g/ml, no viable clones were obtained. The respective experiment was carried out in parallel with single-KO cells for the glyoxalase II gene, which encodes an enzyme that is not essential for procyclic cells. 5 In this case, cells with a double-KO genotype were obtained. Because in both approaches the genetic background of the strain and the maximum concentration of the selection marker were identical, it is likely that the stepwise addition allowed genomic rearrangements to retain a copy of 1-c-grx1 and overcome drug toxicity by gene duplica- 1-C-Grx1 was incubated with 500 M GSSG or H 2 O 2 , subjected to gel chromatography (see Fig. 1D), and analyzed by ESI-MS as described under "Experimental Procedures." Prior to ESI-MS analysis, elution fractions F1 (12.5-13 ml), F2 (13-13.5 ml), and F3 (13.5-14 ml) were treated with iodoacetamide. Carboxamidomethylation (CM), or glutathionylation (GS), or over-oxidation to a sulfinic acid (2 oxygen molecules, (2 O) of free thiols (SH) caused an increase of the protein mass by 57, 305, and 32 Da, respectively. S-S, intramolecular disulfide. The main masses observed in each fraction and the respective modifications are given in bold letters.

Treatment
Mass ( tion or increased transcription of the resistance gene. This was not the case when a high concentration of the selecting agent was added immediately. The second approach was based on conventional gene replacement by another resistance marker. Substitution of the second allele with a puromycin cassette resulted in procyclic cells that were resistant to both blasticidin and puromycin (BLA R and PAC R ) but still harbored a copy of the 1-c-grx1 gene (Fig. 4C) and expressed the protein at levels similar to single-KO or wild-type cells (Fig. 4D). The refractoriness of T. brucei to 1-c-grx1 gene silencing or deletion is a strong indication that the protein is indispensable for the parasite.
Conditional Knock-out of 1-c-grx1 in Procyclic T. brucei-Finally, we tried to generate conditional KO cell lines containing a tetracycline-inducible copy of 1-c-grx1-c-myc 2 . To assure a tightly regulated expression of the ectopic gene, bloodstream and procyclic parasites harboring two copies of the Tet repressor protein (cell line 514-1313) (21) were chosen as host cells. The cells were transfected with the 1-c-grx1-c-myc 2 construct (Fig. 5A), which integrated in the untranscribed spacer of the ribosomal RNA locus. Western blot analyses of protein extracts from transformed parasites grown in the absence of tetracycline displayed a faint band with the expected molecular mass of 1-C-Grx1-c-Myc 2 indicating a minor leakiness of the expression system. After 24 h of growth in the presence of tetracycline, the cells showed a 5-10fold higher concentration of 1-C-Grx1-c-Myc 2 when compared with the authentic protein (Fig. 5B). The correct subcellular localization of 1-C-Grx1-c-Myc 2 was verified by immunofluorescence microscopy. Procyclic and bloodstream T. brucei overexpressing 1-C-Grx1-c-Myc 2 revealed an identical staining pattern for anti- 1-C-Grx1 were reconstituted and purified by gel filtration on Sephadex G25, and the cluster stability was recorded at 420 nm in the presence (black line) and absence (gray line) of 1 mM GSH. C, holo-His 6 -1-C-Grx1, with an initial absorbance at 420 nm of 0.35, was incubated with different concentrations of H 2 O 2 (from 0 to 1 mM as indicated), and cluster disassembly was followed as absorbance decrease at 420 nm until all holoprotein turned into apoprotein (ϳ20 h). Only the first hour is depicted. D, gel filtration on a Superdex 200 column revealed only one prominent peak with absorbance at 420 nm in the molecular mass range that would correspond to His 6 -1-C-Grx1 monomers, dimers, or tetramers. E, the molecular mass of the [Fe-S]-1-Cys-Grx1 complex (gray square) was determined using a standard chromatogram provided by the suppliers manual (black circles: ferritin (440 kDa), albumin (67 kDa), ␤-lactoglobulin (35 kDa), ribonuclease A (13.7 kDa), cytochrome c (6.5 kDa), and vitamin B12 (1.4 kDa)) and proven by separation of albumin (67 kDa), ovalbumin (43 kDa), chymotrypsinogen (25 kDa), and ribonuclease A (13.7 kDa) (gray circles). c-Myc and the mitochondrial marker (Fig. 5C). The signal obtained with the anti-1-C-Grx1 serum also overlapped with that of the mitochondrion. In addition, a weak and diffuse nuclear staining was observed. Because this was not the case with the anti-c-Myc monoclonal antibody, it was most likely due to a cross-reaction with other proteins (e.g. see in Fig. 5B the unspecific band detected in cell extracts). Taken together, the bloodstream and procyclic cell lines showed a Tet-responsive expression and proper compartmentalization of the protein expressed from the ectopic gene copy. For bloodstream trypanosomes, deletion of a single 1-c-grx1 allele was again not successful, not even when 1 g/ml tetracycline was added freshly to the medium every 48 h to guarantee a steady level of recombinant 1-C-Grx1-c-Myc 2 . In this respect, we cannot rule out that insufficient and/or unstable PAC and BLA mRNA was produced from the 1-C-Grx1 locus, as proposed for the unsuccessful gene replacement of another essential mitochondrial protein (32). Procyclic cell lines with a tetracycline-inducible (Ti) conditional knock-out genotype  (1)(2)(3)(4)(5) is given as percentage of the wild-type control (wt) and was normalized using the TR signal. C, PCR genotyping of cell lines resistant to both selection markers (BLA R /PAC R ). PCR was performed as described under "Experimental Procedures" using genomic DNA isolated from wild-type cells (negative control) and five resistant clones (clones 7, 5, 10, 2, and 0). The primer combinations should not yield PCR products if both 1-c-grx1 alleles have been successfully replaced (see Fig. 6A). All BLA R -PAC R double-resistant clones showed the conservation of 1-c-grx1. The minor PCR products amplified with the primer pairs f/D (ϳ750 bp) and U/r (ϳ520 bp) are nonspecific amplicons; the first one was also obtained from genomic DNA of wild-type cells. M, molecular size standards (GeneRuler TM 1-kb DNA ladder; MBI Fermentas). D, representative Western blot of two double-resistant clones (clones 0 and 10). In agreement with the PCR genotyping, 1-C-Grx1 was detected in all clones analyzed. The content of 1-C-Grx1 is given as percentage of the wild-type cell line after normalization against TR. (1-c-grx1-c-myc 2 Ti ⌬1-c-grx1::BLA/1-c-grx1::PAC) were obtained starting from a single-KO cell line resistant to blasticidin. The cells were transfected with the 1-c-grx1myc 2 construct and then in the presence of tetracycline with the PAC cassette and finally were cultivated in the presence of tetracycline, blasticidin, and puromycin (Fig. 6A). In these conditional KO clones, depletion of 1-C-Grx1-c-Myc 2 was subsequently achieved by growing the cells in a medium sup-plementedwithTet-freefetalcalfserum.Becauseofthephasedependent expression of 1-C-Grx1, the growth behavior and the degree of 1-C-Grx1-c-Myc 2 depletion were evaluated in logarithmically growing parasites (Fig. 6, B and C) as well as in cells harvested in the stationary phase (Fig. 6, D and E). The growth behavior of cells depleted of 1-C-Grx1-c-Myc 2 and of the corresponding overexpressing controls was identical (Fig. 6B), even if the concentration of 1-C-Grx1-c-Myc 2 was only about one-fifth of the authentic 1-C-Grx1 in logphase wild-type parasites (Fig. 6C, clone cKO3 and wt). Also upon long-term cultivation, the growth behavior of the conditional KO cell lines was independent of tetracycline ( Fig.  6D). Under this experimental setting, 1-C-Grx1-c-Myc 2 could be lowered to 29 Ϯ 7% (n ϭ 5) of the level of the authentic protein in wild-type cells (Fig. 6E), which corresponds to a minimum concentration of 1-C-Grx1-c-Myc 2 in the mitochondrion of cKO procyclic cells of about 3 M. 15 Days after tetracycline withdrawal, 1-C-Grx1-c-Myc 2 expression started to recover (Fig. 6E). Moreover, in the absence of tetracycline, all conditional KO clones displayed  Fig. 4, A and C, and the supplemental table). In the clones, PCR fragments of 1150 and 1213 bp (primer pairs U/r and f/D, respectively) were not obtained, proving that the 1-c-grx1 alleles had been successfully replaced. PCR with the primer pair f/r amplified a fragment of 540 bp of the 1-c-grx1 gene, which in the conditional KO cell lines corresponds to the ectopically integrated 1-c-grx1c-myc 2 . M, molecular size standards (GeneRuler TM 1-kb DNA ladder, MBI Fermentas). B, the conditional KO clones that had been continuously grown in the presence of tetracycline were washed three times with tetracycline-free medium and seeded at a cell density of 5 ϫ 10 5 cells/ml in fresh medium containing 1 g/ml tetracycline (cKO-1, --f--; cKO-3, --F--) or without tetracycline (cKO-1, --Ⅺ--; cKO-3, --E--). Every 48 h, cell viability was assessed by light microscopy, samples were withdrawn for Western blot analysis, and the cultures were diluted to 5 ϫ 10 5 cells/ml with fresh medium. C, Western blot analysis of cKO-1 and cKO-3 grown in the absence (Tet Ϫ) or presence (Tet ϩ) of tetracycline. TR served as loading control. wt and re represent extracts from wild-type parasites and 5 ng of recombinant His 6 -1-C-Grx1, respectively. D, long term cultivation of conditional KO cell lines (n ϭ 5) in the presence (---F---) or absence (---E---) of 1 g/ml tetracycline. Every 5 days, the cells were counted, samples were withdrawn for Western blot analysis, and the cultures were diluted to the initial cell density of 5 ϫ 10 5 cells/ml. E, the levels of 1-C-Grx1-c-Myc 2 and 1-C-Grx3 were followed by Western blot analysis with anti 1-C-Grx1 and 1-C-Grx3 antibodies in the clones cKO-17 and cKO-20 grown in the absence of tetracycline. Extracts from cells harvested at day 0 show the level of 1-C-Grx1-c-Myc 2 prior to withdrawal of Tet. The numbers below the blots give the amount of 1-C-Grx1-c-Myc 2 and 1-C-Grx3, respectively, relative to the level in wild-type trypanosomes normalized against the TR signal. The bands below the c-Myc 2 -tagged protein are probably due to cross-reactions of this batch of antiserum because they were also detected in extracts of wild-type cells (see Fig. 5B). The black and gray arrows mark the position of wild-type and 1-C-Grx1-c-Myc 2 , respectively. Trypanosoma brucei Monothiol Glutaredoxin-1 OCTOBER 10, 2008 • VOLUME 283 • NUMBER 41 1-C-Grx1-c-Myc 2 levels that were higher than those in noninduced wild-type cell lines harboring a Tet-inducible 1-c-grx1-c-myc 2 copy (compare for instance Fig. 6E, day 10, with Fig. 5B, PC 1A4 Tet Ϫ, ␣-Grx1). Probably the transcriptional control of the ectopic gene by the Tet repressor got lost at least partially. This is a common phenomenon reported for essential genes controlled with this system (24, 30 -33). To evaluate whether the intracellular concentration of 1-C-Grx1 correlates with that of the other monothiol glutaredoxins in T. brucei, extracts from the 1-C-Grx1-c-Myc 2 -depleted conditional KO cell lines were subjected to Western blot analyses with anti-1-C-Grx2 and anti-1-C-Grx3 sera. Under these conditions, expression of 1-C-Grx2 (not shown) resembled that in wild-type cells (Fig. 3B), whereas the 1-C-Grx3 level was slightly increased (190 Ϯ 48%, n ϭ 4) when the content of 1-C-Grx1-c-Myc 2 was lowest (Fig. 6E). The failure to obtain parasites devoid of 1-C-Grx1 strongly suggests that despite their homology and probably mitochondrial co-localization, 1-C-Grx2 cannot functionally replace 1-C-Grx1. The slight up-regulation of 1-C-Grx3 in cells with lowered 1-C-Grx1 content may be an indirect effect, or it may reflect a cross-talk between the cytosolic and mitochondrial monothiol glutaredoxins. Taken together, 1-C-Grx1 probably plays an essential and specific function in the mitochondrion of T. brucei.
Involvement of 1-C-Grx1 in the Iron and Redox Metabolism of T. brucei-The role of 1-C-Grx1 in the iron and redox homeostasis of the mammalian infective form was studied in wild-type and 1-C-Grx1 overexpressing parasites. As shown in Fig. 5, 1-C-Grx1-c-Myc 2 can be expressed at a 5-10-fold concentration of that of the authentic protein and is targeted to the mitochondrion. This overexpression did not affect the growth (Fig.  7A) or morphology (not shown) of the parasites when cultivated under optimal conditions. Parasites containing the 1-c-grx1-c-myc 2 gene were subsequently grown in the absence and presence of tetracycline and treated with Fe 3ϩ , the iron chelator deferoxamine (Dfx), shown to deplete cellular iron (34), hydrogen peroxide, and the redox cycler menadione (2-methyl-1,4naphthoquinone) (35), respectively. The addition of 100 M Fe 3ϩ did not significantly affect the growth of the parasites independent of the 1-C-Grx1 level. In contrast, trypanosomes overexpressing 1-C-Grx1 displayed significant growth retardation in the presence of 100 M Dfx (Fig. 7, B and C). The cells also showed enhanced sensitivity toward hydrogen peroxide, either given as a bolus or generated continuously by glucose oxidase. In procyclic cells overexpressing 1-C-Grx1-c-Myc 2 a comparable sensitivity toward hydrogen peroxide was observed (not shown). Although wild-type and 1-C-Grx1-overexpressing parasites were more sensitive toward treatment with 10 M menadione compared with 100 M hydrogen peroxide (not shown), overexpression of 1-C-Grx1 did not enhance the sensitivity of the parasites against the quinone, in contrast to the effect observed for hydrogen peroxide (Fig. 7B). This suggests that menadione has a different mode of action to exert its cytotoxicity in the parasites, which may involve depletion of the cellular low molecular mass thiols and inactivation of thiol proteins (Ref. 36 and references therein). We further studied the combined effect of Dfx and hydrogen peroxide (Fig. 7C). Independent of the overexpression of 1-C-Grx1-c-Myc 2 , a combination of oxidative stress (100 M H 2 O 2 or 1 milliunit/ml glucose oxidase) and iron depletion (25 or 100 M Dfx) had a more deleterious effect on parasite viability and proliferation than the respective single treatment (Fig. 7C). For instance, a highly cytotoxic effect was observed when the parasites were incubated for 18 h in the presence of 100 M Dfx and either 1 milliunit/ml glucose oxidase or 100 M H 2 O 2 . Parasites subjected to oxidative stress in the presence of lower concentrations of the iron chelator (25 or 5 M Dfx) were generally arrested. However, 1-C-Grx1-overexpressing cells treated with 100 M H 2 O 2 and 25 M Dfx or with 100 M Dfx and 1 milliunit/l glucose oxidase showed a significantly stronger growth impairment than the noninduced cultures. Taking together, high levels of 1-C-Grx1-c-Myc 2 appear to exacerbate the cytostatic or cytotoxic effect caused by oxidative stress and/or iron depletion in bloodstream T. brucei (Fig. 7, B and C).
In addition, we studied the impact of the different stressors on the 1-C-Grx1 levels in wild-type T. brucei (Fig. 7D). Exposure of bloodstream cells to 100 M hydrogen peroxide resulted in a 2-fold up-regulation of 1-C-Grx1. In contrast, in cells treated with 100 M Dfx, the 1-C-Grx1 content was lowered by 50%, whereas exogenous iron did not affect the level of the protein. This again suggests a physiological role of the mitochondrial 1-C-Grx1 in the iron and redox metabolism of African trypanosomes.

DISCUSSION
African trypanosomes possess three distinct 1-C-Grx. Whereas the parasite 1-C-Grx2 and 1-C-Grx3 like homologues from yeast, plasmodia, and bacteria are monomeric proteins (1,8,37), T. brucei 1-C-Grx1 forms noncovalent homodimers. As shown recently (12), treatment of the protein with GSSG leads to the specific thiolation of Cys 181 , which triggers the formation of an intramolecular disulfide bridge with Cys 104 (12). In a 1-Cys-D-peroxiredoxin, glutathionylation of the catalytic cysteine induces a switch from a noncovalent dimer to the monomeric form (38). As shown here, exposure of 1-C-Grx1 to GSSG or H 2 O 2 did not induce dissociation of the noncovalent dimer. Besides nonspecific covalent polymers, the protein appeared in Overexpression of 1-C-Grx1-c-Myc 2 was induced by adding tetracycline (1 g/ml) 24 h prior to and during the course of the experiment. The initial cell density was 5 ϫ 10 5 and 1 ϫ 10 6 cells/ml in the case of iron and oxidative challenges, respectively. 100 M Dfx, 100 M iron (Fe 3ϩ ), as well as H 2 O 2 , glucose oxidase, and menadione at the depicted concentrations were added to the cultures. Tet-induced and noninduced cells without treatment showed identical growth (none) and are set to 100%. Cell growth and viability were evaluated at 16 -24 h after stress induction. The relative cell density (%) refers to the growth of the Tet-induced cells to the respective noninduced cells. The values are the mean of three experiments that differed by less than 5%. C, combined effect of Dfx and hydrogen peroxide on bloodstream T. brucei overexpressing 1-C-Grx1-c-Myc 2 . 1 g/ml tetracycline was added 24 h prior to and during the course of the experiment. Noninduced parasites were included as control. The initial cell density was 5 ϫ 10 5 cells/ml and is depicted with a gray bar (Co, control) and a dashed line. The white and black bars give the mean cell densities of Tet-induced and noninduced parasites, respectively, after 18 h growth in the presence of 0, 5, 25, and 100 M deferoxamine, 100 M H 2 O 2 , and 1 milliunit/ml glucose oxidase. The data represent the mean of three independent experiments. Asterisks denote significant differences (n ϭ 6, p Ͻ 0.1, two-tailed t test). D, one representative experiment of three showing the expression of 1-C-Grx1 in wild-type bloodstream T. brucei subjected to different stress conditions. Parasites in the late log phase of growth (C 0 ) were harvested and diluted in fresh medium to a density of 5 ϫ 10 5 and 1 ϫ 10 6 cells/ml for the experiments involving iron and oxidative challenges, respectively. 100 M H 2 O 2 , 2 milliunits/ml glucose oxidase (GOD), 100 M FeCl 3 (Fe 3ϩ ), and 100 M Dfx, respectively, was added to the cultures. Nontreated cultures served as control (C 1 for oxidative stress treatment and C 2 for experiments involving iron homeostasis). After 20 h, viable cells were counted, and the extract of 4 ϫ 10 6 cells was subjected to Western blot analyses with 1-C-Grx1 and TR antibodies. The amount of 1-C-Grx1 in cells exposed to different stresses was estimated by densitometric analysis and is expressed relative to that in the corresponding untreated control. The 1-C-Grx1 content in the control cultures C 1 and C 2 is also compared with that in the inocula (C 0 ). two major peaks. The fraction eluting with an apparent molecular mass of 44/48 kDa contained 1-C-Grx1 with an intramolecular disulfide bridge. The peak at 32 kDa was a mixture of reduced, glutathionylated, and over-oxidized (sulfinic acid) protein. Thus, formation of the intramolecular disulfide bridge probably causes a marked conformational change. The oxidant-induced structural changes should affect or regulate the function(s) of 1-C-Grx1. For instance, formation of an intramolecular disulfide between Cys 181 and Cys 104 triggered by exposure to oxidants is a reversible state that could protect 1-C-Grx1 (especially Cys 104 ) against irreversible over-oxidation. Interestingly, the orthologous proteins from Leishmania major, Leishmania infantum, T. cruzi, and T. congolense contain only a single cysteine that corresponds Cys 104 in T. brucei 1-C-Grx1. Because an intramolecular disulfide cannot be formed, it remains to be investigated whether and/or how these proteins undergo a redox regulation.
Very recently, two new 1-C-Grx from yeast (named Grx6 and Grx7) were shown to dimerize noncovalently in a GSH-and GSSG-independent manner (11). The proteins appear to represent a novel subgroup of the 1-C-Grx family. In contrast to most 1-C-Grx, they share several structural and catalytic (e.g. transhydrogenase activity) features with dithiol glutaredoxins (11). Yeast Grx6 and Grx7 are much more distantly related to the T. brucei 1-C-Grx1 than are yeast Grx3-5, which indicates that the oligomeric structure of monothiol glutaredoxins is unpredictable.
Several dithiol (17,26,39) and, very recently, monothiol (10, 11) Grx proteins have been reported to complex a [2Fe-2S] cluster employing GSH as thiol cofactor. As shown here for the first time for a trypanosomatid protein, 1-C-Grx1 can coordinate one [2Fe-2S] center/protein dimer using GSH as the nonprotein ligand. For the glutaredoxins studied thus far, ISC assembly to the apoprotein leads to protein oligomerization, i.e. conversion of a monomer or dimer into the corresponding dimer and tetramer, respectively (10,11,25,26). Interestingly, 1-C-Grx1 appears to be the first example of a glutaredoxin-type protein in which the incorporation of the [2Fe-2S] cluster into the homodimeric apoprotein does not cause a change in the oligomeric structure. Our data did not allow us to decide whether binding of the cluster is accompanied by local or large conformational changes in the dimer. Picciocchi et al. (10) suggested that the ability to coordinate an ISC is an evolutionarily conserved feature of monomeric 1-C-Grx presenting a conserved CGFS motif. However, T. brucei 1-C-Grx1 and yeast Grx6 disprove this generalization. The proteins have a CAYS and CSYS active site motif, respectively, and exist as dimers in the apo-state. Instead, the structural motif common to all 1-C-Grx capable of binding an ISC seems to be the absence of a proline residue at the active site, which, in the case of dithiol Grx, was shown to be a key factor for cluster incorporation (17,26). In line with this statement, S. cerevisiae Grx7 with its CPYS motif does not coordinate an ISC (11).
Both the mitochondrial 1-C-Grx1 and the 1-C-Grx3 are abundant proteins in both life stages of T. brucei. In contrast, 1-C-Grx2, the second putative mitochondrial protein, was detectable only in stationary phase trypanosomes. The levels of all three monothiol glutaredoxins were highest in parasites har-vested in the stationary and starvation growth phases. This expression pattern is unusual considering the down-regulation of RNA and protein during stationary phase reported for several housekeeping enzymes in different trypanosomatids (40 -42) and as observed here for TR. Because in kinetoplastids gene expression is almost exclusively controlled at the post-transcriptional level (43), the high protein concentration in stationary parasites may be due to an increased half-life or translation of the corresponding transcript and/or protein. A comparable up-regulation in the stationary phase was reported for Grx4, the single monothiol glutaredoxin of E. coli (8), whereas the transcript levels of S. cerevisiae and Schizosaccharomyces pombe Grx3-5 achieved a maximum in the exponential growth phase and a minimum in stationary phase cells (1,44,45). Taking into account that 1-C-Grx1 can complex an ISC and its biological function is linked to the iron metabolism of trypanosomes, the strong enrichment in nondividing stationary parasites suggests that the protein may suit for the rapid de novo synthesis of iron-containing proteins when the cells resume growth.
Despite a variety of strategies, it was not possible to silence the expression of the authentic 1-C-Grx1 in bloodstream cells, and in procyclic T. brucei, the maximum achievable depletion was by 70%. In procyclic parasites, the chromosomal alleles could be replaced but only if an ectopic copy of 1-c-grx1 was expressed. Down-regulation of 1-C-Grx1-c-Myc 2 expression in these cells was not accompanied by an up-regulation of 1-C-Grx2, which also has an N-terminal sequence that is recognized by the mitochondrial import machinery of yeast (13). This clearly demonstrates that the two 1-C-Grx proteins lack functional redundancy and points to an indispensable function of 1-C-Grx1 in the organelle. 1-C-Grx from different phyla have been shown to efficiently substitute for the mitochondrial Grx5 in yeast and to participate in ISC biogenesis (5,9,14,46). Interestingly, T. brucei 1-C-Grx1, but not 1-C-Grx2, partially rescued the phenotype of grx5-deficient yeast cells when targeted to the mitochondria (13), suggesting that the parasite protein plays a similar role in ISC biogenesis. In T. brucei, two components of the ISC biosynthetic pathway have been characterized thus far. Both the cysteine desulfurase (TbiscS2) and a scaffold protein (TbiscU) are essential for procyclic cells (47), which adds value to the relevance of this metabolism for African trypanosomes.
Overexpression of an ectopic copy of 1-c-grx1 resulting in a 5-10-fold concentration compared with the authentic protein, did not affect the proliferation rate and morphology of bloodstream cells when grown under optimal conditions. In contrast, a 1.5-fold overexpression of the mitochondrial Grx5 in S. pombe impaired the growth of the fission yeast (44), although the underlying mechanism was not investigated. Interestingly, overexpression of 1-C-Grx1 in T. brucei resulted in an increased sensitivity of the parasite toward the iron chelator Dfx. The unphysiologically high concentration of 1-C-Grx1 may augment the depletion of the "free iron" pool due to an increased synthesis of ISC proteins and/or to iron sequestration in the form of the [Fe-S]-1-C-Grx1 complex. The 2-fold downregulation of 1-C-Grx1 in wild-type bloodstream parasites upon treatment with Dfx could thus be a compensatory mechanism to balance the cellular iron levels.
Overexpression of 1-C-Grx1 also sensitized trypanosomes against H 2 O 2 -mediated oxidative stress. One mechanism as to how hydrogen peroxide damages cells is via oxidation of ISCs with release of iron and initiation of Fenton chemistry (48,49). As shown in vitro, formation of the [Fe-S]-1-C-Grx1 complex requires the reduced protein, and the complex is rapidly destroyed by exposure to H 2 O 2 (estimated half-life, ϳ40 min for 100 M H 2 O 2 ). In addition, treatment of apo-1-C-Grx1 with H 2 O 2 or GSSG leads to formation of an intramolecular disulfide probably accompanied by strong conformational changes. Under conditions of oxidative stress, both reactions may occur in vivo, which would shift the equilibrium of holo-/apo-1-C-Grx1 to an inactive oxidized form of the protein with the concomitant impairment of ISCs biosynthesis. In the case of cells overexpressing 1-C-Grx1, this condition may be worsened due to a high (and potentially toxic) concentration of oxidized 1-C-Grx1-c-Myc 2 in the mitochondrion of proliferating trypanosomes. Although the underlying molecular mechanisms are not yet known, it is clear that overexpression of 1-C-Grx1 lowers the capacity of bloodstream cells to withstand an exogenous hydrogen peroxide challenge.
The detrimental effect of oxidative stress toward trypanosomes was amplified by iron depletion. This is not surprising when considering that H 2 O 2 destroys ISCs (48,49) and that the released iron is sequestered by Dfx and thus will not be available for the de novo assembly or repair of ISCs. Again, in parasites overexpressing 1-C-Grx1, this phenotype was augmented in accordance with a probable shift to inactive oxidized apoprotein.
Strikingly, exposure of wild-type trypanosomes to hydrogen peroxide induced a 2-fold increase of the 1-C-Grx1 levels. This can be interpreted as a physiological response to increase the biosynthesis and/or repair of ISCs that were oxidatively damaged. In contrast to T. brucei 1-C-Grx1, expression of Grx3-5 in S. cerevisiae is constitutive and unaffected by H 2 O 2 (1). In the fission yeast, oxidative stress due to heavy metals and reactive oxygen species down-regulates the expression of grx5 but induces expression of grx4 (44). Despite the overall evolutionarily conserved involvement of 1-C-Grx in ISC biogenesis, distinct stimuli probably regulate the levels of 1-C-Grx in the different organisms.
In conclusion, the three monothiol glutaredoxins occur in both life stages of African trypanosomes, with the highest levels found in the stationary phase, and display distinct cellular functions. Mitochondrial 1-C-Grx1 is an abundant protein that plays a crucial role in the iron and redox homeostasis of the parasites. Work is in progress to crystallize recombinant 1-C-Grx1 for structural analysis and to identify direct interaction partners of 1-C-Grx in these parasites.