Role of Peroxidoxins in Leishmania chagasi Survival: Evidence of an Enzymatic Defense Against Nitrosative Stress

The mechanisms by which Leishmania parasites survive exposure to highly reactive oxygen (ROS) and nitrogen (RNS) species within phagosomes of macrophages are not well known. Recently it has been shown that RNS alone is sufficient and necessary to control L. donovani infection in mice (Ref. 17). No enzymatic defense against RNS has been discovered in Leishmania to date. We have previously isolated two peroxidoxins (LcPxn1 and LcPxn2) from L. chagasi and showed that recombinant LcPxn1 protein was capable of detoxifying hydrogen peroxide, hydroperoxide and hydroxyl radicals (Ref. 25). In further characterizing the physiological role of peroxidoxins in Leishmania survival, we show here that recombinant LcPxn1 protein can detoxify RNS in addition to ROS, whereas recombinant LcPxn2 protein can only detoxify hydrogen peroxide. LcPxn1 and LcPxn2 are localized to the cytoplasm and over-expression of LcPxn1 in L. chagasi parasites enhanced survival when exposed to exogenous ROS and RNS and also enhanced survival within U937 macrophage cells. Site-directed mutagenesis studies revealed that the conserved Cys52 residue is essential for detoxifying hydrogen peroxide, t-butyl hydroperoxide and hydroxyl radicals, whereas the conserved Cys173 residue is essential for detoxifying t-butyl hydroperoxide and peroxynitrite. This is the first report of an enzymatic defense against RNS in Leishmania. various RNS and ROS. A, activity was assessed by the ability of the LcPxn1 proteins to protect 50µM Pyrogallol Red from ONOO - -induced bleaching and to protect supercoiled pGEM-2 plasmid (s) from ONOO - attack into the slower migrating nicked band (n) (A-inset); B, activity towards protecting 5mM ABTS from • NO-induced oxidation into the highly absorbing green ABTS + complex. C, Activity expressed as nmol/min/µg recombinant protein towards 100µM H 2 O 2 and 100µM t-butyl hydroperoxide (D); E, activity towards protecting 0.8mM 2-deoxy-D-ribose from • OH-induced damage and the ability to protect supercoiled (s) pGEM-2 DNA from • OH-induced nicking (n) (E-inset). The average +/- SE of at least four independent experiments are shown. Insets were representative of three independent experiments.

Glutathione has recently been implicated in protecting L. major from • NO-induced cytotoxicity (29). To date, an enzymatic defense against RNS has not been identified in Leishmania.
Peroxidoxins (or peroxiredoxins) are highly conserved enzymes found in all kingdoms ranging from bacteria to humans. 2-Cys peroxidoxin proteins are characterized by two conserved cysteine residues corresponding to approximately positions 47 and 170 and exist in nature predominantly as head-to-tail dimers, although high molecular weight multimers have been reported (25,(30)(31)(32). Peroxidoxins were initially characterized as enzymes able to detoxify ROS, namely H 2 O 2 and alkyl hydroperoxides (33), with • OH recently being added to the substrate list (25). Peroxidoxins have also been implicated in detoxifying RNS in bacteria, yeast and human cells (34)(35)(36).
We have previously isolated two peroxidoxin genes from L. chagasi that are differentially regulated, where LcPxn1 RNA transcripts are highly abundant in the amastigote stage and LcPxn2 transcripts are highly abundant in the promastigote stage (25). Recombinant LcPxn1 protein was shown to detoxify H 2 O 2 , ROOH and • OH, but the mechanism of its action and the role that L. chagasi peroxidoxins play in detoxifying RNS and in parasite survival has not been characterized. In this paper, we demonstrate that recombinant LcPxn1 protein, but not LcPxn2, can detoxify RNS in addition to ROS and show that LcPxn1 protects L. chagasi parasites from ROS-and RNS-mediated toxicity in vitro and enhances survival within macrophages. by guest on March 24, 2020 http://www.jbc.org/ Downloaded from Furthermore, we have identified the key catalytic residues of LcPxn1 involved in detoxifying both ROS and RNS which differs from peroxidoxins isolated from other organisms.

EXPERIMENTAL PROCEDURES
Site-directed Mutagenesis-LcPxn1 mutants C52A, C173A and C52A/C173A were generated by site directed PCR mutagenesis as previously described (37). All peroxidoxin constructs were amplified using PCR, cloned into the pGEX-2T vector (Amersham Biosciences) and confirmed by sequencing. Transformed E. coli DH5α cells were grown shaking at 37°C in Luria-Burtani broth containing 100µg ml -1 ampicillin for 8 hours, after which 0.2mM isopropyl-1-thio-β-Dgalactoside (IPTG) was added to the culture and shaken overnight. GST fusion proteins were harvested by sonication and passed over a glutathione-agarose resin column as described by manufacturer (Amersham Biosciences). The fusion proteins were cleaved with Thrombin overnight at 24°C, further purified and protein purity (>95%) was verified on a SDS-PAGE gel.
Protein concentrations were determined using the BCA Protein assay kit (Pierce Chemical, Rockford, IL).
Peroxide Assays-Peroxide metabolism was measured as previously described (38). Briefly, the reaction mixture contained 50mM Tris-HCl (pH 8.0), 0.2mM dithioerythritol (DTE), 50µM H 2 O 2 or 50µM t-butyl hydroperoxide, and 0.125µg/ml of protein (pre-incubated with 0.2mM DTE for 30 mins at 37°C). The reaction was stopped with the addition of 1ml of trichloroacetic acid (10% w/v). 0.2ml of 10mM ferrous ammonium sulfate and 0.1ml of 2.5M potassium thiocyanate were added and the peroxide concentrations were determined spectrophotometrically by guest on March 24, 2020 http://www.jbc.org/ Downloaded from at 480nm using known amounts of peroxide (1-50µM) as a standard. All solutions were made fresh immediately before use.
Deoxyribose Degradation Assay for • OH Scavenging-The production of • OH and the • OHinduced damage of 2-deoxy-D-ribose were measured as previously described (39). A 50µl reaction mixture was set up to contain the following components to give the final concentrations as stated: 10mM potassium phosphate buffer (pH 7.4); 63mM NaCl, 0.8mM 2-deoxy-D-ribose; 0.2mM DTE; 0.125µg/µl protein (proteins were pre-incubated in 0.2mM DTE for 30 mins at 37°C). 21µM ferrous ammonium sulfate was added and the tubes were incubated at 37°C for 15 mins. 100µl of thiobarbituric acid (TBA) (1% w/v) and 100µl of trichloroacetic acid (TCA) (2.8% w/v) were then added to the mixture and boiled for 10 mins. Fluorescence was measured in a 96-well plate using a SpectraMax Gemini plate reader (Molecular Devices) with six reads per well (Excitation= 532nm, Emission= 553nm). All solutions were made fresh immediately before use.
• OH-induced DNA Nicking Assay-3µM FeCl 3 , 0.1mM EDTA and 10mM DTE were allowed to react for 10 mins at 37°C to generate • OH as previously described (40). 0.5mg/ml protein (preincubated with 0.2mM DTE for 30 mins at 37°C) was then added to the mixture and incubated at 37°C for 30 mins. 2µg of pGEM-2 plasmid (Promega) was then added to each tube and incubated at 37°C for 4 hours. The DNA was separated on a 1% agarose gel containing 0.2µg/ml ethidium bromide at 100V constant. All solutions were made fresh immediately before use.
Pyrogallol Red Bleaching Assay for ONOO -Scavenging-Reagent peroxynitrite was generated from acidified hydrogen peroxide and nitrite using the quenched-flow method (41) and passaged over MnO 2 column as previously described (42). The reaction assay was carried out as previously described (43). The reaction mixture contained 100mM phosphate buffer (pH 7.0), 1µM DTE, 50µM Pyrogallol Red (ε=2.4x10 4 mol -1 .liter.cm -1 ) and 20µM protein at 25°C. 20µM of reagent peroxynitrite was added to the reaction for 5 mins after which the absorbance at 542nm was measured. All solutions were made fresh immediately before use.
ONOO --induced DNA Nicking Assay-A reaction mixture containing 50mM sodium phosphate (pH 7.0), 10mM NaCl, 0.1mM diethylenetriaminepentaacetic acid (DTPA), 0.5µg intact pGEM-2 plasmid DNA and 20µM protein was prepared as previously described (44). 50µM reagent ONOOwas added to the reaction mixture and incubated at room temperature for 5 mins. The DNA was separated on a 1% agarose gel containing 0.2µg/ml ethidium bromide at 100V constant. All solutions were made fresh immediately before use.
• NO Detoxification Assay-• NO levels were measured as previously described (45). 100mM sodium nitroprusside and 0.125µg/ml protein were incubated in phosphate buffered saline (PBS) pH 7.4 for 5 mins. 2,2'-azinobis(3-ethylbenzthiazoline-6-sulfonic acid) (ABTS) was then added to the reaction mixture (5mM final concentration) and the • NO-induced oxidation of ABTS to ABTS + was measured by monitoring the change in absorbance at 420nm at room temperature.

Construction of Expression Vectors-The
Western blotting was performed using anti-GFP (1:2500) antibodies and detected using an

Macrophage Infection Assay-Infection of U937 cells from human origin (American Type Cell
Collection, Rockville, Md.) was carried out as previously described (47). U937 cells were seeded at a concentration of 2.5x10 5 cells/cm 2 in 8-chamber slides and differentiated into adherent macrophages by treatment with 7.5ng phorbol myristate acetate (Sigma) per ml of RPMI 1640 with 10% fetal calf serum, 2mM glutamine and 50µg/ml gentamicin (Invitrogen) (RPMI) under 5% CO 2 at 37ºC for 72 hours. Non-adherent cells were washed 3-5 times with warm RPMI media followed by incubation with L. chagasi parasites at a parasite to U937 cell ratio of 10:1 for 6 hours. Non-engulfed parasites were washed away 3-5 times with warm RPMI and incubated in fresh RPMI media. The level of infection in 200 U937 cells was determined at 12, 24, and 48 hours by optical microscopy following Diff Quick staining of cell preparations (48). Values are expressed as the average number of parasites per infected U937 cell.

Recombinant LcPxn1 Protein can detoxify RNS, whereas LcPxn2 cannot
Assays containing reagent ONOOor the • NO donor sodium nitroprusside (SNP) and target molecules susceptible to RNS-induced oxidation were used to test whether recombinant LcPxn1 protein can detoxify RNS. Pyrogallol Red (PR) has been previously shown to be bleached by ONOObut not by decomposed ONOO -, nitrite or nitrate (49). When added to the reaction mixture, recombinant LcPxn1 and the ONOOscavenger Trolox significantly (P<0.001) protected PR from bleaching compared to boiled LcPxn1, GST and BSA controls ( Figure 1A).
DNA has also been shown to be a target of ONOOwhich converts the supercoiled DNA into a slower migrating nicked DNA (50). Consistent with the results observed above, LcPxn1 protein and Trolox were able to protect supercoiled DNA from ONOO --induced nicking, whereas boiled LcPxn1, GST and BSA were unable to provide protection ( Figure 1A A colorimetric assay for measuring • NO produced from SNP was used to test the ability of recombinant LcPxn1 protein to protect ABTS from • NO-induced oxidation into the strongly absorbing green ABTS + complex (45). When added to the reaction mixture, LcPxn1 protein and the • NO scavenger PTIO significantly (P<0.0001) protected ABTS from oxidation compared to boiled LcPxn1 and GST controls ( Figure 1B). There was no significant difference in protection between LcPxn2 protein and boiled LcPxn2 or GST controls. As further controls, scavengers of does not appear to detoxify RNS.

Amino acid residues involved in RNS-and ROS-detoxifying activity
The two cysteine residues corresponding to Cys47 and Cys170 in all 2-Cys peroxidoxins are highly conserved among all organisms. The amino terminus Cys47 has been implicated as the catalytic residue in the detoxification of H 2 O 2 and alkyl hydroperoxides (51,52) and peroxynitrite (36). To identify the catalytic residues in Leishmania LcPxn1, we performed site-directed mutagenesis of the corresponding conserved Cys52 and Cys173 residues and constructed three recombinant LcPxn1 protein mutants containing Cys to Ala mutations (LcPxn1-C52A, LcPxn1-C173A and LcPxn1-C52A/C173A).
In studying the amino acid residues involved in detoxifying ROS, we found that LcPxn1-C52A protein failed to detoxify H 2 O 2 , t-butyl hydroperoxide (tBOOH) and • OH (P<0.001 in each case) compared to wildtype LcPxn1 protein ( Figure 1C-E respectively). In addition, LcPxn1-C52A protein failed to protect supercoiled DNA from • OH-induced nicking ( Figure 1E-inset). Since there was no significant ROS-detoxifying activity by LcPxn1-C52A compared with boiled LcPxn1 or GST, Cys52 appears to be essential for detoxifying ROS which is consistent with previous findings with other peroxidoxins. There was no significant difference observed in the ability of LcPxn1-C173A to detoxify H 2 O 2 ( Figure 1C) or • OH ( Figure 1E) compared to wildtype LcPxn1 and LcPxn1-C173A protected supercoiled DNA from • OH-induced nicking ( Figure 1E-inset). In contrast, LcPxn1-C173A did not demonstrate significant tBOOHdetoxifying activity compared to boiled LcPxn1 and GST ( Figure 1D). In studying the amino acid residues involved in detoxifying RNS, we found that LcPxn1-C52A protein exhibited similar levels of activity in detoxifying ONOOand protecting supercoiled DNA from ONOO --induced nicking compared to wildtype LcPxn1 protein ( Figure 1A), suggesting that Cys52 is not essential for detoxifying ONOO -. LcPxn1-C173A protein did not demonstrate a significant difference in detoxifying ONOOcompared to boiled LcPxn1 or GST (P<0.005) and could not protect DNA from nicking ( Figure 1A), suggesting that Cys173 is the catalytic cysteine residue and is essential for detoxifying ONOO -. Both LcPxn1-C52A/C173A and LcPxn2 proteins did not detoxify ONOOand could not prevent ONOO --induced nicking of DNA. Interestingly, wild type LcPxn1, LcPxn1-C52A, LcPxn1-C173A and LcPxn1-C52A/C173A proteins were all found to significantly (P<0.002) detoxify • NO compared to boiled LcPxn1, LcPxn2 and GST controls ( Figure 1B). Remarkably, LcPxn1-C52A/C173A protein provided significantly (P<0.0001) more protection from • NO-induced oxidation compared to wildtype LcPxn1 ( Figure 1B). Collectively, these results suggest that the Cys52 residue is essential for detoxifying H 2 O 2 , tBOOH and • OH, Cys173 is essential for detoxifying tBOOH and ONOOand neither Cys52 nor Cys173 are essential for detoxifying • NO (summarized in Table 1).  Figure   3A-E). Consistent with previous findings (5), we found that the control (pX) stationary phase parasites were significantly (P<0.01) more resistant to H 2 O 2 toxicity than early log phase parasites ( Figure 3A). We also found that the control (pX) stationary phase parasites were significantly more resistant to tBOOH (P<0.001) ( Figure 3B) and ONOO -(P<0.03) ( Figure 3D).

Over-expression of
In support of our findings that Cys52 is essential in detoxifying H 2 O 2 , tBOOH and • OH and that Cys173 is essential in detoxifying tBOOH and ONOO -, parasites over-expressing LcPxn1-C52A did not exhibit an enhanced survival upon exposure to H 2 O 2 , tBOOH or • OH but did exhibit enhanced survival upon exposure to ONOOcompared to pX control parasites ( Figure 3A-D).
LcPxn1-C173A exhibited an enhanced survival upon exposure to H 2 O 2 and • OH but not upon exposure to tBOOH or ONOO - (Figure 3A-D) which is consistent with our observations with recombinant LcPxn1-C173A protein (Figure 1). Contrary to our findings with recombinant LcPxn1 proteins, parasites over-expressing LcPxn1-C52A and LcPxn1-C173A did not exhibit enhanced survival upon exposure to • NO ( Figure 3E). Parasites over-expressing LcPxn1-C52A/C173A did not exhibit enhanced survival upon exposure to any of the ROS or RNS.

Cellular localization of LcPxn1 and LcPxn2 proteins
To further define the functions of LcPxn1 and LcPxn2 in parasite survival, we studied the cellular localization of these proteins within L. chagasi. The amino acid sequence of LcPxn1 does not appear to contain a typical organellar-targeting signal sequence which suggests that it may be localized to the cytoplasm. The last three amino acids at the carboxyl terminus of LcPxn2 are SKQ and conspicuously resembles the glycosomal targeting signal sequence SKL.

Although mutational analysis of the SKL glycosomal targeting signal in Trypanosoma brucei
showed that this signal is highly degenerate, mutation of the signal to SKQ was not sufficient to target proteins to the glycosome but rather remained cytosolic (53).
We created GFP-LcPxn1 and GFP-LcPxn2 fusion protein gene constructs and over-expressed them in L. chagasi parasites. Parasites were selected at 50µg/ml G418 and Western blot analysis of each parasite extract with anti-GFP revealed the presence of an ~48kDa fusion protein suggesting that both the GFP-LcPxn1 and GFP-LcPxn2 fusion proteins were intact ( Figure 4A).
No bands corresponding to the GFP protein alone (27kDa) were observed in either of the extracts isolated from parasites expressing the fusion proteins. Fluorescence microscopy showed that both the GFP-LcPxn1 and GFP-LcPxn2 fusion proteins are localized throughout the parasite including the flagella, similar to the fluorescence pattern observed with the control parasites expressing the GFP protein alone ( Figure 4B-D). The fluorescence patterns observed for both LcPxn1 and LcPxn2 are distinct from the pattern of glycosomal localization (data not shown).
These results suggest that both LcPxn1 and LcPxn2 are localized to the cytoplasm.

Over-expression of LcPxn1 protein in L. chagasi enhances intracellular survival within macrophages
During the initial stages of infection with a foreign pathogen, an oxidative burst occurs in human macrophages (9,18,54,55) including U937 cells (56)(57)(58) in which ROS is produced in response to phagocytosis. Human macrophages including U937 cells have also been shown to produce RNS once infection is established (18,59,60).
We have previously shown that the level of LcPxn1 mRNA expression increases significantly towards the amastigote phase compared to early log phase parasites (25). In order to gain insight into the role that LcPxn1 plays in intracellular survival within macrophages, we tested the ability

Previous reports have shown that both ROS and RNS contribute to the early control of
Leishmania infection and that RNS alone is necessary and sufficient to control Leishmania infection (11,17,18 (62,63). Peroxidoxins have also been shown to be able to protect biological targets such as DNA from attack by • OH (25,40) and furthermore we have shown here that LcPxn1 is cytoplasmic which significantly increases its chance of coming into contact with • OH (Figure 4) and whose over-expression can protect parasites from an exogenous source of • OH ( Figure 3C).
Numerous studies have identified alkyl hydroperoxides as substrates for peroxidoxins, implicating peroxidoxins as very important enzymes in reducing phospholipid hydroperoxides which can arise from oxidation and thereby protecting cells from plasma membrane damage.
Leishmania parasites lack typical glutathione peroxidases which are well-known protectors of lipid peroxides in eukaryotes. We previously demonstrated that recombinant LcPxn1 protein can detoxify alkyl hydroperoxides and have extended these studies to show that LcPxn1 can also protect L. chagasi from tBOOH ( Figures 1D and 3B). Interestingly, we found that both the conserved Cys52 and Cys173 residues are essential for detoxifying tBOOH which suggests an alternative mechanism for the detoxification of ROOH compared to the detoxification of H 2 O 2 and • OH. Recent studies have emphasized the importance of the microenvironment surrounding the active site residues of peroxidoxins (52,64). It is therefore possible that the Cys173 residue of LcPxn1 may be essential for coordinating the active site into a more favourable environment for donating a proton to the poor and much more bulky ROleaving group. The lack of an available proton donor could cause the sulfenic acid intermediate (R-SOH) that forms on the catalytic cysteine to be further oxidized into R-SOOH which has been found to lead to reduced activity (65). Remarkably, we could not detect alkyl hydroperoxidase activity with recombinant LcPxn2 protein which is 89% identical to LcPxn1 ( Figure 1D). There has been a report of a peroxidoxin from Leishmania major (Lmf30 TryP) that is highly homologous to LcPxn2 and is also incapable of significantly detoxifying alkyl hydroperoxides (26). Crystallographic studies with the peroxidoxin AhpC from S. typhimurim has revealed that structural conformations and the mobility of key residues present in loop structures encompassing the active site cysteines is very important for activity of the protein (64). The main difference between LcPxn1 and LcPxn2 is the presence of a nine amino acid extension at the carboxyl terminus of LcPxn2. It is possible that this extension alters the microenvironment surrounding the active site into one that is not favorable for activity or is more prone to inactivation as described above. Characterization of the crystal structures of LcPxn1 and LcPxn2 will provide more insight into this mechanism.
Bacterial and yeast peroxidoxins have been previously shown to protect cells from • NO and ONOO --mediated toxicity (34)(35)(36). Our findings also show that LcPxn1 can detoxify • NO and ONOOand protect Leishmania from RNS-mediated toxicity, however the mechanism by which this occurs in Leishmania differs from the proposed mechanism by which bacterial peroxidoxins detoxify ONOO -. The conserved amino terminus cysteine (Cys46) residue of the AhpC peroxidoxin from S. typhimurium was found to be essential for activity (36), whereas we demonstrate that the carboxyl terminus cysteine (Cys173) residue of LcPxn1 is essential for activity ( Figure 1A and 3D). This finding will be important for future drug design studies with  (34,35), no evidence has been presented of a recombinant peroxidoxin protein capable of detoxifying • NO directly. We found that overexpression of LcPxn1 protein in parasites also protected them from both ONOO --and • NOmediated toxicity, but we also found evidence that recombinant LcPxn1 protein can detoxify • NO ( Figure 1B). Interestingly, site-directed mutagenesis of LcPxn1 protein revealed that neither of the conserved cysteine residues (Cys52 or Cys173) were essential for detoxifying • NO and the mutation of both cysteines led to increased activity. This result suggests that LcPxn1 possesses a different mechanism for detoxifying • NO which we are currently investigating. Oddly, the overexpression of wildtype LcPxn1 but not the mutant proteins within the parasites provided significant protection to the parasites when exposed to • NO ( Figure 3E). It is possible that formation of heterogeneous multimers between mutant and wildtype LcPxn1 monomers within the parasites led to an inhibitory effect.
Of significant interest is our finding that over-expression of LcPxn1 in Leishmania parasites enhanced survival within macrophages ( Figure 5 infantum (27,66) have not been shown to possess the ability to detoxify RNS. It is possible that LcPxn2 is an evolutionary precursor to LcPxn1 which through evolution underwent a deletion in the carboxyl terminus. This deletion could have altered the microenvironment of the active site surrounding the conserved cysteine residues resulting in the acquisition of higher functions such as being able to detoxify phospholipid hydroperoxides, • OH, ONOOand • NO which ultimately led to a selective advantage for L. chagasi survival. It will be interesting to see whether the acquisition of the higher functions of LcPxn1 contribute to the pathogenecity of L. chagasi. Our findings provide a better understanding of the mechanisms that Leishmania utilize for intracellular survival and the role that peroxidoxins play in Leishmania survival. We are currently using homologous recombination and anti-sense technology and mouse infection