Ebsulfur Is a Benzisothiazolone Cytocidal Inhibitor Targeting the Trypanothione Reductase of Trypanosoma brucei *

Background: The trypanosoma-specific trypanothione system is essential for the redox metabolism in the parasite. Results: The benzisothiazolone compound ebsulfur is a cytocidal agent that disrupts parasitic trypanothione system and redox balance. Conclusion: Ebsulfur and its analogues irreversibly inhibit TryR activity hampering ROS detoxification and leading to parasite death. Significance: The antiparasitic mechanism of benzisothiazolone helps the development of new strategies against trypanosomiasis. Trypanosoma brucei is the causing agent of African trypanosomiasis. These parasites possess a unique thiol redox system required for DNA synthesis and defense against oxidative stress. It includes trypanothione and trypanothione reductase (TryR) instead of the thioredoxin and glutaredoxin systems of mammalian hosts. Here, we show that the benzisothiazolone compound ebsulfur (EbS), a sulfur analogue of ebselen, is a potent inhibitor of T. brucei growth with a favorable selectivity index over mammalian cells. EbS inhibited the TryR activity and decreased non-protein thiol levels in cultured parasites. The inhibition of TryR by EbS was irreversible and NADPH-dependent. EbS formed a complex with TryR and caused oxidation and inactivation of the enzyme. EbS was more toxic for T. brucei than for Trypanosoma cruzi, probably due to lower levels of TryR and trypanothione in T. brucei. Furthermore, inhibition of TryR produced high intracellular reactive oxygen species. Hydrogen peroxide, known to be constitutively high in T. brucei, enhanced the EbS inhibition of TryR. The elevation of reactive oxygen species production in parasites caused by EbS induced a programmed cell death. Soluble EbS analogues were synthesized and cured T. brucei brucei infection in mice when used together with nifurtimox. Altogether, EbS and EbS analogues disrupt the trypanothione system, hampering the defense against oxidative stress. Thus, EbS is a promising lead for development of drugs against African trypanosomiasis.

Infection with subspecies of the protozoan parasite Trypanosoma brucei causes African trypanosomiasis, a disease that affects both humans and animals. The human disease has two stages. In the first stage, parasites are found in the blood and lymph from infected individuals. In the second stage, parasites cross the blood-brain barrier causing severe neurological symptoms (1)(2)(3)(4). The disease is invariably fatal if untreated.
Drugs used for treatment of the early stage of the disease do not penetrate the blood-brain barrier and are ineffective for the second stage treatment. Arsenic compounds such as melarsoprol used for treatment of the second stage of disease have severe side effects and are lethal in ϳ10% of cases. DL-␣-Difluoromethylornithine is used for treating the West African form of human African trypanosomiasis, caused by Trypanosoma brucei gambiense. However, this drug is given by intravenous injections, is expensive, and is not effective against Trypanosoma brucei rhodesiense that causes the East African form of human African trypanosomiasis. Thus, there is an urgent need for development of new alternative anti-trypanosomal compounds.
Redox active thiol groups in proteins and low molecular mass compounds like GSH play key roles in enzymatic reactions involved in DNA synthesis and in the control of the intracellular redox state. Trypanosomatids differ from other eukaryotes and prokaryotes in their specific thiol redox metabolism (5). These parasites possess a small thiol molecule named trypanothione (T(SH) 2 ), 4 a conjugate of the sulfur-containing tri-peptide GSH and the polyamine spermidine (6). Trypanothione reductase (TryR), an NADPH-dependent disulfide oxidoreduc-tase, reduces trypanothione disulfide to dihydrotrypanothione T(SH) 2 . The dithiol form of T(SH) 2 maintains an intracellular reducing environment in these parasites, substituting for GSH and glutathione reductase and the thioredoxin system (thioredoxin (Trx) and thioredoxin reductase (TrxR)) found in the mammalian host (7). Many proteins of the parasite-specific trypanothione metabolism enzymes are critical for the defense against oxidative stress (8,9). Moreover, TryR and T(SH) 2 delivers electrons to ribonucleotide reductase that is essential for DNA synthesis (10,11). Altogether, the absence of the trypanothione system in mammals, the lack of a functional redundancy, and the sensitivity of trypanosomes against oxidative stress render the components of this metabolism system attractive drug target molecules (10).
Ebselen (EbSe) is an anti-oxidative and anti-inflammatory seleno-organic compound (23). EbSe was shown to be a substrate for both mammalian thioredoxin and thioredoxin reductase and catalyzed H 2 O 2 reduction (24,25). However, high concentrations of EbSe competed with disulfide substrates for reduction by Trx and therefore acted as an inhibitor of protein disulfide reduction by the Trx system (24). In contrast to broad interest focused on the glutathione peroxidase-like activity of EbSe as antioxidant, little attention has been paid to its selective activity toward certain pathogenic bacteria (26). Recently, we have discovered that ebselen is an effective competitive inhibitor of bacterial TrxR, which is a highly promising strategy to develop antibiotics against pathogenic bacteria lacking a glutathione-glutaredoxin system (27).
Given the unique and essential features of thiol metabolism in T. brucei for DNA synthesis and defense against oxidative stress and the proposed mode of action of EbSe on bacteria, we studied the effect of EbSe and EbSe analogues on the viability of trypanosomes. We found that the benzisothiazolone ebsulfur (EbS), an EbSe analogue, but not EbSe is a very potent and highly selective trypanocidal agent of parasites. EbS irreversibly bound and oxidized TryR, resulting in an enhancement of the intracellular ROS level, leading to parasite death. EbS analogues also were able to cure T. brucei infection in vivo when combined with nifurtimox.

MATERIALS AND METHODS
Reagents and Enzymes-NADPH, 5,5Ј-dithiobis-2-nitrobenzoic acid (DTNB), and DMSO were from Sigma. Ebselen, ebsulfur (2-phenyl-1,2-benzisothiazol-3(2H)-one), and their analogues were freshly dissolved in DMSO prior to use. Recombinant Trypanosoma cruzi TryR was kindly provided by Prof. R. Luise Krauth-Siegel (Heidelberg University, Germany), and T(SH) 2 was kindly provided by Dr. Marcelo Comini (Institut Pasteur de Montevideo, Uruguay). The gene encoding Trypanosoma brucei brucei trypanothione reductase was cloned using the previously developed exonuclease I-dependent restriction-free cloning method into pNIC28-BSA4. The protein was recombinantly expressed by autoinduction in 2A medium and was subjected to immobilized metal affinity chromatography purification, tobacco etch virus protease cleavage, and subtractive immobilized metal affinity chromatography as described (28), resulting in a protein with only one extra serine added to the N terminus. To ensure maximal cofactor incorporation in the protein, bacterial lysis was done in the presence of 0.2 mM FAD (28,29).
For 1 H NMR analysis (CDCl 3 and CD 3 OD), 500 and 300 MHz spectrometers were used. 13 C spectra were recorded at 125 and 75 MHz. NMR chemical shifts are reported in ppm referenced to the solvent peak of the deuterated solvent used (CDCl 3 ; 7.26 ppm for 1 H and 77.16 ppm for 13 C, respectively. CD 3 OD; 3.31 ppm for 1 H spectra. For DMSO-d 6 , solvent signal reference at 39.52 was used for 13 C NMR spectra). MS-EI data were obtained using a Finnigan Mat DPC/GCQ system and direct inlet of a dichloromethane solution of compound. MS-ESI data were obtained using a Waters micromass ZQ system, and compounds were dissolved in methanol prior to analysis. Infrared spectra were recorded on a PerkinElmer Life Sciences Spectrum 100 FT-IR spectrometer, and melting points were obtained using a Stuart Scientific melting point apparatus. Dichloromethane used in reactions was dried with aluminum oxide and stored under nitrogen gas before use.
Inhibition of TryR by EbS and Its Analogs in Vitro-Recombinant TryR (20 nM) was incubated with the indicated concentrations of EbS or its analogs in TE buffer containing 200 M NADPH for different times. Then the enzymatic activity was measured with the T(SH) 2 -coupled DTNB reduction-based assay.
Determination of Parasite Viability-To measure drug sensitivity, 6 ϫ 10 3 /100 l of parasites were cultured in 96 flatbottomed well culture plates with serial drug dilutions for 72 h at 37°C. Cultures (100 l) were further incubated for 2 h with 10 l of WST-1 reagent (Roche Diagnostics) to monitor parasite viability. Viability was measured by the conversion of WST-1 reagent to formazan and recorded by a multiwell scanning spectrophotometer at an excitation wavelength of 450 nm. A direct correlation between WST-1 reagent oxidation and parasite numbers was confirmed.
Determination of Viability of Luciferase-conjugated Parasites-The cell viabilities of luciferase-conjugated T. brucei were measured by detection of the luciferase activity. Serial dilutions of ebsulfur were incubated with 200 l of T. brucei brucei (Rluc) 1.25 ϫ 10 6 /ml in 96-well plates, and parasite luminescence was measured relative to untreated control between 4 and 6 h after drug incubation. Luminescence was measured 5 min after the addition of 10 M of the substrate coelenterazine (Nanolight). Luminescence was quantified in a Microbeta Jet luminometer (PerkinElmer Life Sciences) in triplicate samples at each time point and drug dilution.
Preparation of Parasite Lysates-To analyze the TryR activity in the parasites, the cells were harvested by centrifugation, washed with PBS, then resuspended in PBS containing protease inhibitor mixture, and lysed by sonication followed by centrifugation to prepare the lysates.
Detection of Parasite Cell Death-Exposed phosphatidylserine was detected on the outer membrane of cells using annexin-V kit (BD Biosciences) following the instructions supplied by the manufacturer as a measure of programmed cell death. Fluorescence was measured using FACS analysis as described before (37). To analyze disruption of the plasma membrane after ebsulfur treatment, nuclei were stained with propidium iodide (5 g/ml) in the absence of a cell permeant and analyzed by flow cytometry.
ROS Measurement-To measure intracellular ROS, the cell-permeable nonfluorescent probe 2Ј,7Ј-dichlorofluorescin diacetate (DCF-DA) (Sigma) was used. The procedure was described earlier (38). In brief, T. brucei brucei were treated with various concentrations of EbS, washed with PBS, then suspended in fresh medium containing 5 M of H 2 DCF-DA for 30 min, and incubated at 37°C, 5% CO 2 . After the incubation, parasites were spun and resuspended in PBS, and the ROS produced was quantified by flow cytometry.
Determination of Non-protein Thiol Amount-To measure the amount of non-protein thiols in parasites, the parasite cells were washed with PBS and then treated with 5% TCA to precipitate the proteins. After removing the pellet, the pH of the supernatant was adjusted to 7.4 by 1 N NaOH and 1 M Tris-HCl, pH 8.0. Subsequently, 1 mM DTNB was added to the solution, and the free thiols were determined by the absorbance at 412 nm.
Determination of Trypanothione, GSH, and TryR in the Cell Lysate-The amount of trypanothione in cell lysates was determined by a DTNB reduction activity assay. The experiment was performed with a 96-well plate in the solution containing 50 mM Tris-HCl, pH 7.5, 200 M NADPH, 1 mM EDTA, 1 mM DTNB in the presence of 20 nM TryR. A sample without TryR was used as the control. The absorbance at 412 nm was measured for 5 min with a VERSA microplate reader, and the slope was used to represent DTNB reduction activity. A standard curve with T(SH) 2 ranging from 0.25 to 4.0 M was prepared at the same time. Thus, the DTNB reduction activity was converted to the amount of trypanothione. To measure GSH levels, 20 nM TryR was replaced by 50 nM glutathione reductase, using GSH standards in the range of 0.25-4.0 M. To measure the TryR amount, 10 M T(SH) 2 was used instead of 20 nM TryR, with a TryR standard ranging from 5 to 50 nM. Protein amount was determined by the Bio-Rad DC assay.
Mass Spectrometry-TryR (P, 50 nM) was incubated with EbS (L, 50 M) in the presence of 100 M NADPH for 30 min. Then the protein complex was concentrated with a Microsep TM centrifugal device (30-kDa MWCO, Pall Life Science, Ann Arbor, MI), filtered, and subjected to liquid chromatography-mass spectrometry (LC-MS) analysis. LC-MS analysis was performed using an Agilent 1200 Series LC System coupled to an Agilent 6510 quadrupole-time-of-flight (Q-TOF) mass spectrometer. A Phenomenex Aeris Widepore C4 column (2.1 ϫ 100 mm, 3.6 m) was used for the separation. The mobile phase buffer A consisted of 5% acetonitrile, 0.1% propionic acid, and 0.02% trifluoroacetic acid; buffer B consisted of 95% acetonitrile, 0.1% propionic acid, and 0.02% trifluoroacetic acid. Gradient conditions were 0 -80% buffer B during 20 min. The mobile phase flow rate was 0.2 ml/min. A 10 M sample solution (10 l) was injected into the LC-MS system. The detection by LC-MS was accomplished using electrospray ionization in the positive ion mode, with the scanning range from 500 to 3000 m/z, drying gas flow 11 liters/min, nebulizer pressure 45 p.s.i., dry gas temperature 320°C, ion spray voltage 3600 V, and fragmentor voltage 100 V. Data collection was performed using Agilent MassHunter WorkStation software. TryR Activity Assay-The experiment was performed in TE buffer (50 mM Tris-HCl, pH 7.5, 1 mM EDTA), containing 200 M NADPH, 10 M T(SH) 2 in the presence of 1 mM DTNB. The absorbance at 412 nm was followed with a VERSA microplate reader (Molecular Devices, Sunnyvale, CA). TryR activity was represented by the slope of absorbance change during the initial 2 min.
Animal Experiments-All animal experiments were conducted following protocols that received institutional approval and authorization by the local animal ethical committee. Mice were kept with food and water ad libitum under specific pathogen-free conditions. C57BL6J (B6) male mice, 8 -10 weeks old, were used to assess the efficacy of ebsulfur and analogues and combination treatment with nifurtimox. Compounds were dissolved in DMSO, and an equivalent dose to be administered was further diluted in cyclodextrin. Mice were infected with 2000 T. brucei brucei (AnTat1.1E) intraperitoneally and treated with the drugs as indicated. Weight loss and parasitemia were recorded every other day, and mice without detectable parasitemia for more than 60 days after treatment were considered cured.

RESULTS
Tryanocidal Activity of EbSe and EbS-First, we investigated the inhibitory effect of EbSe and EbS on the growth of blood stream forms of T. brucei brucei. EbS was more than 50-fold more toxic for T. brucei brucei than EbSe. The IC 50 was only 92 nM after 72 h of incubation (Table 1). EbS was also toxic for T. brucei rhodesiense and T. brucei gambiense, the human pathogenic subspecies of T. brucei. EbS was 100-fold more toxic than EbSe for the human pathogenic subspecies T. brucei gambiense, confirming the specificity of EbS for T. brucei (Table 1). Instead, EbS was less toxic for other trypanosomatides such as T. cruzi epimastigotes or the promastigote forms of Leishmania tropica. The IC 50 of EbS for the mammalian L929 cells was 55 and 250 times higher compared with that for T. brucei rhodesiense and T. brucei gambiense, respectively (Table 1).
By measuring ATP-independent luciferase activity in transgenic T. brucei brucei, we found that EbS caused a time-dependent trypanocidal effect (Fig. 1A and Table 1). Using this assay, the IC 50 of EbS was shown to be 1.4 and 0.91 M at 4 and 6 h, respectively. Several trypanocidal drugs have been shown to activate a programmed cell death of T. brucei (39). One of the earliest indications of programmed cell death is the transloca-  SEPTEMBER 20, 2013 • VOLUME 288 • NUMBER 38 tion of phosphatidylserine from the inner to the outer leaflet of the plasma membrane. Parasites treated with 2 M EbS showed translocation of phosphatidylserine as indicated by the binding of annexin V as early as 1 h after treatment (Fig. 1B). Parasite death as measured by cell permeability to DAPI was recorded 1 h after treatment with 2 M EbS (Fig. 1C), indicating that EbS-induced programmed cell death of T. brucei brucei was followed by a secondary necrosis.

Inhibition of Trypanothione Reductase by Ebsulfur
Ebsulfur Is an Irreversible Inhibitor of TryR-Because TryR is a main regulator of trypanosomal thiol metabolism (13), we measured TryR activity in parasites treated with EbS. TryR activity as measured with a T(SH) 2 -coupled DTNB reductionbased method was decreased after incubation of T. brucei with EbS (Fig. 1D). The amount of reduced non-protein thiols in parasites treated with EbS was also decreased upon the treatments with above 2 M of EbS (Fig. 1E). Thus, the thiol metabolism in T. brucei was hampered by EbS. We then investigated the effects of EbS on TryR in cell lysates from T. brucei and T. cruzi. The TryR activity in cell lysates from T. brucei was inhibited by the lower concentration of EbS than that in lysates from T. cruzi (Fig. 1F), consistent with the observation that T. brucei was more susceptible than T. cruzi to EbS (Table 1). To further understand the interaction mechanism of TryR with EbS, we compared the inhibition of EbS on the activity of recombinant T. cruzi and T. brucei TryR enzymes. EbS inhibited the activity of these two enzymes with similar efficiency (Fig. 1G). Both enzymes share more than 80% identity in amino acid sequence and are highly similar in structure, kinetics, and sensitivity to different inhibitors, suggesting that these variants could be "interchangeable" (40). Further studies were performed with T. cruzi TryR. EbS inhibited TryR in a time-and concentration-dependent manner ( Fig. 2A). The relationship between loss of TryR activity and time fit the Kitz-Wilson analysis pattern of irreversible enzyme inactivation (Fig. 2B) (41), and k inact (rate constant for the inactivation) was 0.38 min Ϫ1 and K i was 10.4 M. In line with this, the activity of TryR was not recovered after EbS was removed by filtration through Sephadex G-25, even in the presence of 1 mM DTT, indicating that EbS is an irreversible inhibitor of the enzyme. The inhibition of TryR by EbS is NADPH-dependent because EbS did not inhibit TryR activity in the absence of NADPH (Fig. 2C).
Because TryR reduces trypanothione disulfide to T(SH) 2 to regulate the parasite thiol metabolism, we further examined the effects of T(SH) 2 on the inhibition of TryR by EbS. T(SH) 2 protected TryR in a concentration-dependent manner (Fig. 2D). To study why the TryR activity in T. brucei was more susceptible to EbS inhibition than that in T. cruzi, we measured T(SH) 2 , GSH, and TryR contents in the two parasites using a DTNBcoupled reduction-based method. The T(SH) 2 , GSH, and TryR levels in T. brucei were all lower than those in T. cruzi (Table 2). In particular, T(SH) 2 contents in blood stream T. brucei were 10 times lower compared with those in T. cruzi, indicating that T. brucei has a frailer thiol metabolism system, making it more susceptible to the thiol-targeting drug.
To prove whether the enzyme inhibition involved the interaction between active site thiols in TryR and EbS, the oxidized forms of an EbS analog, a fluorinated EbS (EbS-F), and the cor-responding sulfone derivative (EbS(O) 2 -F), the latter compound unable to participate in redox reactions via thiol interactions, were tested. The EbS-F inhibited TryR and the growth of parasite with similar efficiency as EbS, whereas EbS(O) 2 -F showed neither inhibitory effects on the TryR activity nor cytotoxic effects for T. brucei brucei (Fig. 2, E and F).
Complex of TryR with EbS and EbSe-Furthermore, we used ESI-MS to confirm that EbS forms a complex with TryR accounting for the inhibition of enzyme activity. TryR (50 nM) was incubated with 100 M NADPH and 50 M EbS, and the mixture was subjected to ESI-MS. The recombinant TryRs incubated or not with EbSe were used as controls. The pure TryR yielded a peak at 53,609 Da (Fig. 3A). When TryR was incubated with EbS, a shift in the mass was observed resulting in a dominant peak corresponding to TryR in complex with one molecule of EbS (Fig. 3, B and D). Instead, upon incubation of TryR with EbSe, we detected a dominant peak of the enzyme forming a complex with five molecules of EbSe (Fig. 3C). A low abundant peak corresponding to TryR in complex with seven EbSe molecules was also found, suggesting that all the cysteines in TryR can react with EbSe. Interestingly, upon treatment with EbS, the complex was found to be in oxidized forms with more oxygen additions (Fig. 3D).
Parasite Cytotoxicity by EbS Is Mediated by ROS-One key function of the trypanothione system was to remove toxic ROS generated either by parasite metabolism, stress, or inflammatory responses, maintaining the cellular redox milieu. Treatment with EbS resulted in an increased accumulation of ROS in  living parasites (Fig. 4A). When T. brucei brucei were incubated with both EbS and glucose oxidase, an ROS generator mimicking an oxidative extracellular environment, a synergy in intracellular ROS accumulation compared with the treatments with either EbS or glucose oxidase was observed (Fig. 4, A-C). The addition of pharmaceutical glutathione, an anti-oxidant mixture containing GSH and vitamin C, diminished EbS-induced ROS accumulation in parasites (Fig. 4D) and decreased the tox- icity of EbS (Fig. 4E), indicating that parasite death caused by EbS is mediated by ROS. TryR is a flavoprotein enzyme that may act as an NADPH oxidase, catalyzing the reduction of O 2 to generate the superoxide anion. We found that the incubation with EbS enhanced H 2 O 2 levels generated by NADPH and TryR (Fig. 5A), and further addition of T(SH) 2 diminished ROS levels (Fig. 5A), providing a possible mechanistic explanation for the protection of T(SH) 2 against the inactivation of TryR by EbS. Because EbSe has been shown to mimic glutathione peroxidase activity (24), we studied whether EbS could influence the trypanothione system in a similar manner. When EbS was incubated with the whole trypanothione system, including NADPH, TryR, and T(SH) 2 together with hydrogen peroxide, EbS acted as a poor peroxidase mimic, increasing the consumption of NADPH at a low concentration (Fig. 5B). This is consistent with the observation that H 2 O 2 was not detected when EbS was incubated with the whole trypanothione system (Fig. 5A). Instead, at higher EbS concentrations, the activity of the trypanothione system to remove H 2 O 2 was blocked in a time-dependent manner as measured by NADPH consumption (Fig. 5B). To determine either T(SH) 2 or TryR was affected by EbS when the whole trypanothione system was incubated with EbS and H 2 O 2 , after the incubation we added additional TryR or T(SH) 2 and continued the peroxidase activity assay. The addition of TryR salvaged the H 2 O 2 peroxidase activity of this system, and the peroxidase activity was inhibited by EbS when T(SH) 2 was added, which demonstrated that TryR but not T(SH) 2 was inactivated when the whole trypanothione system was incubated with EbS and H 2 O 2 (Fig. 5C). We also investigated the effects of H 2 O 2 on the inactivation of TryR by EbS (Fig. 5D). The whole trypanothione system containing high concentrations of TryR was incubated with EbS in the presence or absence of H 2 O 2 overnight, and then small molecules were removed by a MicroSpin G-25 desalting column, and the TryR activity was measured by a T(SH) 2 -coupled DTNB reduction assay. H 2 O 2 enhanced the inactivation of TryR by EbS (Fig. 5D), which is in agreement with the results that glucose oxidase enhanced the EbS-induced H 2 O 2 accumulation (Fig. 4, A-C).

Treatment of T. brucei brucei-infected Mice with EbS-
The effect of administration of EbS in the outcome of murine infection with T. brucei brucei was then studied. EbS (30 mg/kg) was administered intraperitoneally daily starting at the day of infection for 3 days. Treated mice showed reduced parasitemia when compared with controls at 5 days after infection. Parasitemia in mice treated or not with a comparable dose of EbSe was similar (Fig. 6A).
To achieve a complete elimination of parasites, six EbS analogues and 11 selenium-containing ebselen analogues were synthesized and tested for their differential toxicity against T. brucei when compared with mammalian cell lines (Table 3). Although most of the selenium analogues showed a minor growth inhibition for T. brucei, a "bis-selenium compound, compound 13, was highly toxic against parasites with a suitable selective index over mammalian cells (Table 3); however, this compound also inhibited the thioredoxin reductase activity of mammalian cells and thus was not further studied. A similar analogue, a "bis-sulfur compound," compound 16, also showed a strong trypanocidal effect. Parasitemia was not detectable in mice treated with the bis-EbS analogue 16 for at least 6 days, but a 100% of parasite clearance was not achieved (Fig. 6B).
The ability of one of the more soluble analogues, EbS19, to cure T. brucei brucei-infected mice was also tested. A fraction of mice was cured when treated with 5 doses of 20 mg/kg EbS19 and 50 mg/kg nifurtimox, as measured by nondetectable parasitemia for more than 60 days after treatment, whereas all mice treated with either nifurtimox or EbS19 showed parasites in the circulation (Table 4).
Nifurtimox is a nitrofuran used for treatment of diseases caused by trypanosomes, including Chagas disease and sleeping sickness (42). Nifurtimox is a pro-drug and must be processed by nitroreductases to display toxicity. Increased toxicity was displayed by co-incubation of parasites with EbS19 and nifurtimox as compared with the incubation with each of these drugs (Fig. 6C). However, incubation of T. brucei with nifurtimox did not result in ROS elevation, probably indicating that it is metabolized by a type I O 2 -independent nitro-reductase (43) that does not generate superoxide anions (Fig. 6D), suggesting that the additive trypanotoxic effect is not due to an increased sensitivity of EbS-treated parasites to ROS generated after nifurtimox addition.

DISCUSSION
Because trypanosomes lack glutathione reductase and TrxR, TryR is the only enzyme connecting NADPH-and thiol-based redox systems in these parasites (10). TryR is an essential enzyme in the parasite (8), and its absence in mammalian cells makes it an attractive drug target. However, the TryR active site is large and relatively solvent-exposed (45), which may make the discovery of low molecular weight highly potent and competitive inhibitors problematic, even when some help in design can be obtained from structural information from inhibitors bound into the active site (19 -21). Moreover, genetic deletion of TryR indicated that Ͼ90% loss in activity of TryR is required for trypanosomal death (8), suggesting that if competitive a very potent inhibitor is required.
In this study, we show that EbS is an NADPH-, concentration-, time-dependent, and irreversible inhibitor of the T. bru- cei TryR. This may represent a useful direction for further development, because irreversible inhibitors are less susceptible to the effects of substrate accumulation than competitive inhibitors.
Recently, EbSe, EbS, and other EbS analogues were identified in a high throughput screen for inhibitors of T. brucei hexokinase, the first enzyme in glycolysis and another validated target for therapeutic development (46). EbSe was highly efficient to inhibit T. brucei brucei hexokinase, with at least a 40-fold lower IC 50 value for T. brucei hexokinase than EbS or other analogues.
However, and confirming our data, EbS and one of EbS analogues, but not EbSe, were highly toxic for T. brucei, suggesting that the T. brucei hexokinase is unlikely to be the only target for EbS in the living parasite and that subtle differences in the molecular composition or the pharmacology of these inhibitors result in remarkable changes in their activities. We also observed that the presence of bovine serum albumin can protect the enzyme against the inhibitory effects of EbSe much more than EbS. The mass spectral results also showed that EbSe can bind with all the Cys residues in the TryR, although EbS only reacts with one Cys in TryR (Fig. 3). This indicates that the replacement of selenium by sulfur in EbS changed the reaction mechanism, decreased the compound reactivity, but increased its selectivity.
EbS shows the trypanothione peroxidase activity, indicating that the reaction of EbS with trypanothione is reversible and forms a redox cycle. EbS may react with active site thiols of TryR. Both bacterial TrxR and TryR are flavoproteins, containing FAD and NADPH binding domains. However, TryR contains an extra interface domain. Moreover, bacterial TrxR has the CXXC active site motif instead of the CXXXXC active site motif in TryR. The bacterial TrxR active site is located in the NADPH binding domain, whereas the TryR active site is in the FAD binding domain. These differences in structure may explain that EbS causes an irreversible inhibition in TryR, which is different from the inhibition of bacterial TrxR by EbSe (27). The oxidized amino acid residues in the TryR-EbS complex that cause the inactivation of the enzyme may be the active site Cys or some other amino acid residues like tryptophan, and this needs further investigation.
It has been demonstrated that the "trypanothione peroxidase" activity is achieved by two enzymes working in concert: (i) tryparedoxin, a thioredoxin-related protein that is specifically reduced by T(SH) 2 ; and (ii) either a tryparedoxin peroxidase that belongs to the thiol peroxidase family of the peroxiredoxins, which is reduced by tryparedoxin (47), or a glutathione peroxidase (48). The tryparedoxin peroxidase is essential for survival of the blood stream form of the parasite. We demonstrate here that EbS inhibits TryR reducing activity in which T(SH) 2 acts as a substrate and blocks the electron transfer to tryparedoxin peroxidase. Moreover, we show that EbS induces an NADPH oxidase activity on TryR (49) that probably results in further inactivation of the enzyme. This means that TryR, which physiologically is responsible for maintaining high levels of thiols in the cell, is converted into an oxidase, producing reactive oxygen species. Thus EbS might raise the intracellular oxidative stress by several synergistic effects as follows: reducing equivalents of NADPH depletion, disulfide reduction inhibition, or increased ROS production. A model of inhibition of the trypanothione system by EbS is depicted in Fig. 7.
Very interestingly, two types of major antioxidant enzymes, catalase and selenium-dependent glutathione peroxidases, are missing in trypanosoma (10). Catalase and selenium-dependent GPx are very efficient enzymes in defense against oxidative stress with a reaction rate at 10 7 -10 8 M Ϫ1 s Ϫ1 . The absence of these enzymes in trypanosomes makes their antioxidant defenses dependent on the trypanothione system. Thus, it is surprising that the H 2 O 2 level in T. brucei brucei, which can get  to 70 M, is much higher than that in mammalian cells (50). An enhanced inhibition of TryR by EbS, due to high H 2 O 2 levels, might explain why T. brucei brucei is more sensitive to EbS than mammalian cells. We show that treatment of T. brucei brucei with EbS inhibited TryR activity in parasites and increased the intracellular accumulation of ROS. Mimicking environmental or inflammation-generated ROS, glucose oxidase acted in synergy with EbS in ROS accumulation. Thus, although off-targets cannot be ruled out, TryR is a main target of EbS in T. brucei. ROS accumulation resulted in parasite death because parasites survived when the ROS scavenger pharmaceutical glutathione was added into the culture medium. In line with this, TryRpartially defective Leishmania could not regenerate T(SH) 2 from T(S) 2 (51) and showed markedly decreased capacity to survive intracellularly within macrophages, when these produce ROS (52), and T. brucei with 10% of the wild type level of TryR showed hypersensitivity toward H 2 O 2 (8). Treatment with EbS reduced parasitemia in T. brucei-infected mice. Administration of subcurative doses of nifurtimox together with EbS cured the infection of mice with T. brucei brucei, at concentrations at which none of the drugs given alone was curative. Different metabolic pathways are probably targeted by these drugs, and in line with this, an additive effect of nifurtimox and EbS was observed in vitro (Fig. 6C), although nifurtimox did not increase ROS accumulation in parasites.
Altogether, these results reveal the importance of further development of a novel group of potential drugs that inhibit parasite TryR and thereby T. brucei oxidative defenses in an irreversible manner. Moreover, EbS is a substrate for Trx (27) and thus might simultaneously protect against infection-induced mammalian cell oxidative damage. EbS shows no similarity to current treatments for African sleeping sickness, and one EbS analogue has been reported to show oral bioavailability, low mutagenic potential, and CNS permeability (44).