Functional Characterization of Peroxiredoxins from the Human Protozoan Parasite Giardia intestinalis

The microaerophilic protozoan parasite Giardia intestinalis, causative of one of the most common human intestinal diseases worldwide, infects the mucosa of the proximal small intestine, where it has to cope with O2 and nitric oxide (NO). Elucidating the antioxidant defense system of this pathogen lacking catalase and other conventional antioxidant enzymes is thus important to unveil novel potential drug targets. Enzymes metabolizing O2, NO and superoxide anion (O2 −•) have been recently reported for Giardia, but it is yet unknown how the parasite copes with H2O2 and peroxynitrite (ONOO−). Giardia encodes two yet uncharacterized 2-cys peroxiredoxins (Prxs), GiPrx1a and GiPrx1b. Peroxiredoxins are peroxidases implicated in virulence and drug resistance in several parasitic protozoa, able to protect from nitroxidative stress and repair oxidatively damaged molecules. GiPrx1a and a truncated form of GiPrx1b (deltaGiPrx1b) were expressed in Escherichia coli, purified and functionally characterized. Both Prxs effectively metabolize H2O2 and alkyl-hydroperoxides (cumyl- and tert-butyl-hydroperoxide) in the presence of NADPH and E. coli thioredoxin reductase/thioredoxin as the reducing system. Stopped-flow experiments show that both proteins in the reduced state react with ONOO− rapidly (k = 4×105 M−1 s−1 and 2×105 M−1 s−1 at 4°C, for GiPrx1a and deltaGiPrx1b, respectively). Consistent with a protective role against oxidative stress, expression of GiPrx1a (but not deltaGiPrx1b) is induced in parasitic cells exposed to air O2 for 24 h. Based on these results, GiPrx1a and deltaGiPrx1b are suggested to play an important role in the antioxidant defense of Giardia, possibly contributing to pathogenesis.


Introduction
Giardia intestinalis is the amitochondriate protist causing giardiasis, one of the most common human intestinal diseases worldwide, responsible for 280 million symptomatic infections per year [1,2,3,4]. This early divergent parasite has a relatively simple life cycle, alternating between two forms: the oro-fecally transmitted cyst and the vegetative trophozoite, which causes the disease by proliferating attached to the mucosa of the proximal small intestine. In this location this microaerophilic pathogen lacking most of the conventional antioxidant enzymes (including catalase, superoxide dismutase (SOD) and glutathione peroxidase [5,6]) has to cope with both O 2 and nitric oxide (NO). O 2 tension in the proximal small intestine indeed not only is higher than in distal tracts of the gut [7,8,9], but it also fluctuates over time, peaking at every meal in order to meet metabolic demand. Moreover, the fine microcirculatory vascular network perfusing the intestinal submucosa reportedly contributes to formation of a steep O 2 gradient such that O 2 tension declines from up to 80-100 mm Hg at the submucosa to near anoxia at the luminal midpoint [10]. Living attached to the intestinal mucosa, it is therefore likely that Giardia trophozoites are physiologically exposed to fairly high O 2 levels, as well as to the NO released by the NO-synthases in intestinal epithelial cells or derived from reduction of dietary nitrate/nitrite (see [10,11] and references therein). Elucidating the Giardia antioxidant defense system that enables parasite survival to oxidative and nitrosative stress conditions is thus important, particularly in the perspective of unveiling novel potential pharmacological targets.
Defense systems against O 2 , NO and superoxide anion (O 2 2N ) have been recently identified in Giardia as: a flavodiiron protein that, like a previously characterized NADH oxidase [12], is able to convert O 2 to H 2 O [13,14]; an inducible flavohemoglobin able to aerobically metabolize NO to nitrate [15,16] and, more recently, a superoxide reductase reducing O 2 2N to hydrogen peroxide (H 2 O 2 ) [17]. Identification of the H 2 O 2 -producing superoxide reductase is puzzling, because Giardia lacks catalase and no H 2 O 2 -metabolizing enzymes have been characterized in the parasite to date; a NADH peroxidase activity has been reported in membrane extracts of the parasite [5], but the enzyme responsible for this activity is as-yet unidentified. On the other hand, toxicity of H 2 O 2 against Giardia trophozoites is well documented [18,19], being associated with depletion of cellular thiols, inactivation of O 2 consumption, loss of membrane potential and cell motility [20], prompt degradation of the flavodiiron protein [14] and induction of a peculiar programmed cell death [21].
Peroxiredoxins (Prxs) [22,23] are ubiquitous cysteine-dependent peroxidases found from bacteria and archaea to mammals, able to reduce H 2 O 2 to H 2 O. Based on their catalytic mechanism and the number of cysteine (Cys) residues participating in catalysis, these enzymes are named '1-Cys', '2-Cys' or 'atypical 2-Cys' Prx. In homodimeric 2-Cys Prxs, one Cys-SH residue ('peroxidatic Cys', C p ) is specifically oxidized by H 2 O 2 to sulfenic acid (Cys-SOH), which in turn reacts with another Cys-SH ('resolving Cys', C r ) on the adjacent monomer to produce a disulfide bond. This bond is then eventually reduced by a thiol-based protein, such as thioredoxin (Trx), to restore the initial fully reduced state of the enzyme. Besides detoxifying H 2 O 2 , Prxs also play a role in oxidative damage repair due to their alkyl hydroperoxide-reducing activity [24], and are involved in the detoxification of peroxynitrite (ONOO 2 ) [25]. This is a harmful species generated by the reaction of nitric oxide (NO) with O 2 2N at diffusion-controlled rates, causing oxidation and/or nitration of many biomolecules, including proteins, nucleic acids, lipids and thiols [26]. Given its cytotoxicity, ONOO 2 is a key effector produced by the host immune system to counteract microbial infections. Prxs are multifunctional proteins, playing a role not only in the defense from nitroxidative damage, but also in signal transduction [27,28] and protein folding [29], as well as in inflammation, tissue repair and tumor progression in higher Eukaryotes [30]. Interestingly, in several parasitic protozoa Prxs have been shown to be implicated in virulence and drug resistance [31].
Crystallographic studies revealed that all Prxs exhibit a similar topology, with a central 5-stranded b sheet, 5 a-helices and a 2stranded b hairpin (see [22] and references therein). The two cysteine residues essential for the catalytic activity (C p and C r ), are often found in a conserved Val-Cys-Pro motif: C p at the Nterminal part of the a 2 helix, at the bottom of a pocket surrounded by the three additional conserved residues (Pro, Thr and Arg), and C r at the N-terminal region. Depending on the redox state, typical 2-Cys Prxs alternate between two quaternary structures, a homodimer in the oxidized state and larger oligomeric forms, typically (do)decamers, in the reduced state. The crystal structure of two typical 2-Cys Prxs from protozoan parasites have been solved, the mitochondrial Prx from Plasmodium (P.) falciparum in the oxidized state [32] and tryparedoxin peroxidase from Trypanosoma (T.) cruzi [33].
The genome of G. intestinalis (http://giardiadb.org/giardiadb/) encodes two typical 2-Cys Prxs homologues belonging to the Prx1 subfamily (according to the nomenclature in [31]): GiPrx1a (ORFs 16076 and 14521) and GiPrx1b (ORF 15383). The enzymes are yet uncharacterized, but in a recent transcriptomic investigation [34] the expression of GiPrx1a has been shown to be stimulated in Giardia trophozoites upon interaction with rat intestinal epithelial cells, pointing to a role of the protein in pathogenesis. Here, we report a detailed characterization of both Prxs from Giardia, focusing on their ability to metabolize H 2 O 2 , alkyl-hydroperoxides and ONOO 2 , and their expression profile in parasitic cells in response to O 2 exposure.

Cloning, expression and purification of GiPrx1a and GiPrx1b
The GL50803_16076 gene (coding for GiPrx1a) was amplified from G. intestinalis genomic DNA (150 ng) by PCR using the Taq DNA Polymerase High Fidelity and the primers 59 -GAGATGA ATTCATATGCCCGTC -39 and 59 -GATTTGAAAGC TTCCCTCCTG -39 with restriction sites for Nde1 and HindIII, respectively. Similarly, the GL50803_15383 was amplified using Vent polymerase (New England Biolabs) and the primers 59 -CT GCATGCAGCATATGACAACC -39 and 59 -GTTAATGG-GAGCTCTTCACTTTG -39 with restriction sites for Nde1 and SacI, respectively. The GL50803_15383 gene (coding for GiPrx1b) was amplified without the portion encoding the protein N-terminal domain (a putative 46 aa-long signal peptide), thus resulting in a truncated version of the protein (here denoted as 'deltaGiPrx1b'), because attempts to obtain the full length protein in a stable soluble form were unsuccessful. The amplified genes were NcoI and XhoI cloned into the expression vector pET28a(+) (GEN-EART GmbH, Regensburg, Germany).
Transformed E. coli BL21-Gold (DE3) cells were grown aerobically at 37uC in LB medium supplemented with 30 mg mL 21 kanamycin. At OD 600 nm ,0.6 , expression of the His-tagged proteins was induced with 0.4 mM isopropyl-b-D-1thiogalactopiranoside, and the cells were further grown overnight at 20uC. The cells (typically ,6 g from 2 L of culture) were harvested by centrifugation (20 min at 5000 g), resuspended in 70 ml of 50 mM Tris + 500 mM NaCl + 1 mM DTT, and lysed by sonication. After centrifugation (30 min at 14000 g), the supernatant was loaded onto a His-Trap affinity column (Amersham). The recombinant His-tagged protein (GiPrx1a or deltaGiPrx1b) was then eluted with 400 mM imidazole, which was then removed by gel filtration chromatography.

Authors Summary
Giardia intestinalis causes one of the most common human intestinal diseases worldwide, called giardiasis. This microorganism infects the small intestine where it has to cope with O 2 , nitric oxide (NO) and related reactive species that are toxic for Giardia as it lacks most of the conventional antioxidant enzymes. Understanding how this pathogen survives oxidative stress is thus important because it may help to identify novel drug targets to combat giardiasis. Some enzymes playing a role in the antioxidant defense of Giardia have been recently reported, but it is yet unknown how the parasite copes with two well-known oxidants, hydrogen peroxide (H 2 O 2 ) and peroxynitrite (ONOO 2 ). In this study, the Authors show that Giardia expresses two enzymes (called peroxiredoxins), yet uncharacterized, that are able not only to degrade both H 2 O 2 and ONOO 2 , but also to repair damaged molecules (called hydroperoxides) that accumulate in the cell under oxidative stress conditions. These results are totally unprecedented because no enzymes with these types of functions have been reported for Giardia to date. If these two enzymes will prove to be essential for Giardia virulence in future studies, a new way will be paved towards the discovery of novel drugs to treat giardiasis.
Concentration of the isolated proteins was determined using the bicinchoninic acid assay and their purity assessed by sodium dodecyl sulphate polyacrylamide gel electrophoresis (SDS-PAGE).

Reaction of GiPrxs with H 2 O 2 and alkyl-hydroperoxides
The reaction of GiPrx1a and deltaGiPrx1b with H 2 O 2 , t-butylOOH or CumOOH was investigated at 25uC, using a Jasco V-650 spectrophotometer. The peroxydatic activity of the two proteins (2-4 mM each) was measured in 50 mM HEPES + 1 mM EDTA + 100 mM NaCl pH = 7.2, following the oxidation of NADPH (100-200 mM) at 340 nm in the presence of E. coli TrxR (0.4-12 U mL 21 ) and E. coli Trx (10-80 mM), acting as artificial electron donors for the Prxs, and 100 mM H 2 O 2 (or t-butylOOH or CumOOH) as the final electron accepting substrate. The rates were corrected for the rate of NADPH oxidation measured in the presence of all reactants prior to the addition of Prxs. The concentration of NADPH and H 2 O 2 was determined photometrically using the extinction coefficients e 340 nm = 6.22610 3 M 21 cm 21 and e 240 nm = 43.6 M 21 cm 21 [35], respectively.

Reaction of GiPrxs with peroxynitrite
The reaction of reduced GiPrx1a and deltaGiPrx1b with peroxynitrite (ONOO 2 , e 302 nm = 1.67610 3 M 21 cm 21 [36]) was investigated by time-resolved spectroscopy, using a thermostated stopped-flow instrument (DX.17MV, Applied Photophysics, Leatherhead, UK) equipped with a 1-cm path length observation chamber. Experiments were carried out according to the 'initial rate approach' described in [37]. Briefly, each of the two Prxs in 100 mM phosphate buffer pH = 7.0+0.2 mM diethylenetriamine pentaacetic acid was reduced by 2 h-incubation with 10 mM DTT at room-temperature. Prior to the experiment, DTT was removed by concentration/dilution cycles and each protein at increasing concentration (from 10 to 150 mM) was anaerobically mixed at 4uC with a solution of 40 mM ONOO 2 in 10 mM NaOH. According to [37], initial rates of ONOO 2 decomposition were obtained from the absorption decrease measured at 310 nm, using e = 1600 M 21 cm 21 .

Cultures of G. intestinalis trophozoites
Trophozoites of the G. intestinalis strain WB clone C6 (ATCC 50803) were cultured axenically at 37uC in modified Diamond's TYI-S-33 medium as previously described [38]. The medium was supplemented with 10% heat-inactivated bovine serum (Invitrogen) and 0.05% bovine bile (Sigma). Cultures were inoculated in 25 cm 2 -flasks, filled to 90% of their total volume in order to attain low-O 2 tension conditions. Trophozoites were transferred every 48 h into fresh medium when cells became confluent. For preparation of the inocula, trophozoites were detached by chilling the cultures on ice for 30 min.

Immunoblotting assays
For immunoblotting assays, Giardia trophozoites were plated in sterile 6-well plates at a density of 1610 6 cells mL 21 in a volume of 3 mL medium/well. Incubation was performed at 37uC allowing the plates to equilibrate for 24 h either under anaerobic conditions (Anaerocult A minisystem, Merck) or with air (atmospheric O 2 level), in the presence or absence of 120 U mL 21 catalase.
After incubation, trophozoites were detached on ice for 30 minutes, collected by centrifugation and lysed (lysis buffer C3228, Sigma). After total protein content determination by the bicinchoninic acid assay, cell extracts (20 mg protein/lane) were subjected to SDS-PAGE and proteins blotted onto a polyvinylidene difluoride (PVDF) membrane (Immobilon pSQ, Merck).
Blots were then incubated with rabbit polyclonal antibodies raised against GiPrx1a or deltaGiPrx1b (Davids Biotechnologie GmbH), followed by incubation with alkaline peroxidase-conjugated secondary antibodies (NA934, GE Healthcare) and detection by enhanced chemiluminescence (ECL kit RPN2132, GE Healthcare). In these assays, using either the anti-GiPrx1a or the anti-deltaGiPrx1b antibodies was irrelevant, because in dot-blot experiments each of the two antibodies showed cross-reactivity with both GiPrxs.

Real time qPCR
To quantify gene expression by real-time quantitative polymerase chain reaction (qPCR), based on genomic information (http:// giardiadb.org/giardiadb/), primers specific for the G. intestinalis ORFs coding for GiPrx1a (GL50803_16076), GiPrx1b (GL50803_15383) and the housekeeping ribosomal small subunit protein S26 (GL50803_17364) were designed with the software Primer3 (v. 0.4.0, http://frodo.wi.mit.edu/primer3/) ( Table 1) and purchased from Primm (Milan, Italy). Due to the very high nucleotide sequence identity between the two genes, primers designed for the 16076 gene are expected to target also the 14521 gene ( Figure S1).
As for immunoblotting assays, trophozoites were plated in sterile 6-well plates at a density of 1610 6 cells mL 21 in a volume of 3 mL medium/well, and incubated up to 24 h under air, with or without 120 U mL 21 catalase, or under anaerobic conditions (Anaerocult A minisystem, Merck). After incubation, trophozoites were detached on ice for 30 minutes, washed in sterile phosphate buffered saline (PBS), collected by centrifugation and lysed with the lysis buffer from the High Pure RNA Isolation Kit from Roche. A DNaseI digestion step was included to remove possible residual genomic DNA. Quality of the extracted RNA was assessed by 1% agarose gel electrophoresis and from the A 260 nm / A 280 nm absorbance ratio. According to the manufacturer's instructions, 5 mg total RNA were used for synthesizing the first cDNA strand, using the Thermo Scientific kit #K1641, with both random and oligo(dT) 18 primers. qPCR assays were carried out with a Mx3000P Q-PCR System instrument (Agilent Technologies), using 2 mL cDNA as template, the Thermo Scientific Maxima SYBR Green qPCR Master Mix (26) (# K0252), and the primers at a final concentration of 300 nM.
DNA amplification was carried out by running 40 cycles, each cycle including a denaturation (95uC, 15 s), an annealing (55uC, 30 s) and an extension (72uC, 30 s) step. Melting curve analysis was performed at the end of each run. No DNA was amplified if no reverse transcription was carried out, thus confirming the lack of contaminant genomic DNA in the samples. Each experimental condition was assayed in triplicate in at least three independent experiments. Data were normalized to the mRNA levels of the ribosomal small subunit protein S26 (ORF 17364).

Data analysis
Multiple sequence alignments were obtained using Clustal Omega [39,40]. Data from time-resolved spectroscopy were analysed using the software MATLAB (MathWorks, South Natick, MA, USA). Densitometric analysis of blotted membranes was carried out with the software Image J (http://imagej.nih.gov/ij). Statistical significance of the data was determined using the Student t-test in Microsoft Excel; all P values correspond to twosided sample t-test assuming unequal variances. Graphs were generating using the software Origin7. The reported error bars represent the standard error of the mean (SEM).

Results
In the genome of G. intestinalis assemblage A (GL50803), there are three genes annotated as putatively coding for 2-Cys Prxs ( Figure S1): the almost identical 16076 and 14521 genes (coding for GiPrx1a), and the 15383 gene (coding for GiPrx1b). Due to the very high (,99%) nucleotide sequence identity between the 16076 and 14521 genes, only the proteins encoded by the 16076 (GiPrx1a) and 15383 (GiPrx1b) genes were considered for this study. Amino acid sequence analysis ( Figure S2) shows that both GiPrx1a and GiPrx1b share significant similarities with Prxs from other protozoan parasites and, as expected, retain the two catalytically relevant, redox active cysteines (namely, Cys58 and Cys174 in GiPrx1a and Cys95 and Cys219 in GiPrx1b). At variance with GiPrx1a, GiPrx1b exhibits at its N-terminus 46 residues that are recognized by the software SignalP-4.1 [41] as a signal peptide, with a cleavage site between positions 15 and 16 (Figures S2); this suggests that GiPrx1b may have a different intracellular localization as compared to GiPrx1a, or it may even represent a secretory protein. Attempts to obtain the recombinant full-length GiPrx1b protein in a stable soluble form were unsuccessful. The protein was therefore herein characterized as a truncated form (deltaGiPrx1b) devoid of the N-terminal signal peptide.
The recombinant His-tagged proteins GiPrx1a and delta-GiPrx1b were purified to homogeneity by affinity chromatography with a typical yield of .20 mg protein per g of cells. As determined by SDS-PAGE (see a representative gel in Figure S3), both proteins were purified as single polypeptide chains with a molecular mass ,23 KDa, consistent with the values calculated from amino acid sequences (22,540 and 22,690 Da for GiPrx1a and deltaGiPrx1b, respectively).

H 2 O 2 -and alkylhydroperoxide-reductase activity of GiPrxs
The ability of the isolated recombinant GiPrx1a and delta-GiPrx1b proteins to reduce H 2 O 2 or alkyl-hydroperoxides, such as CumOOH and tButylOOH, was tested spectrophotometrically at 25uC by measuring NADPH oxidation at 340 nm, upon addition of the enzymes to a solution containing E. coli TrxR and Trx to mediate electron transfer to the Prxs. A representative assay with H 2 O 2 as the final electron accepting substrate is reported in Fig. 1A, which shows a clear peroxidase catalytic activity of the Prxs. The activity, determined from the initial rate of NADPH oxidation, is underestimated due to the limited efficiency of the E. coli chimeric reducing system in providing electrons to Giardia Prxs. This has been confirmed in experiments in which the concentration of either TrxR (Fig. 1B) or Trx (Fig. 1C) from E. coli was systematically increased. Keeping [Trx] constant at 10 mM and increasing the concentration of TrxR, the H 2 O 2 -reductase activity of both GiPrx1a and deltaGiPrx1b progressively increased saturating at .15 U mL 21 TrxR (Fig. 1B), while at high TrxR concentration (12 or 24 U mL 21 ) the apparent turnover number of GiPrx1a increased almost proportionally with the concentration of Trx up to 4 s 21 (Fig. 1C). In these assays, where H 2 O 2 was used at a maximal concentration of 100 mM, a progressive but slight inactivation of the recombinant Prxs was observed during the reaction, likely due to protein cysteine(s) hyperoxidation [42]. From these data, we conclude that both GiPrx1a and deltaGiPrx1b can effectively reduce H 2 O 2 with similar rates. V max , however, could not be determined because the concentrations of E. coli TrxR and Trx used in the assays did not prove to be saturating for GiPrxs.
Importantly, by keeping the concentration of NADPH, TrxR and Trx constant and replacing H 2 O 2 with CumOOH or tButylOOH, we could show that, under the experimental conditions tested, both GiPrx1a and deltaGiPrx1b reduce alkylhydroperoxides at least as efficiently as they metabolize H 2 O 2 (Fig. 2).

Reaction of GiPrxs with peroxynitrite
The reaction of DTT-reduced GiPrx1a and deltaGiPrx1b with peroxynitrite (ONOO 2 ) was investigated by stopped-flow spectroscopy, following the experimental protocol described in [37]. The experimental temperature was set at 4uC so to slow down the reaction and measure the initial rate of the reaction with higher accuracy. Upon rapidly mixing under anaerobic conditions ONOO 2 with either of the two Prxs in the reduced state, over the first 100 ms a fast consumption of ONOO 2 was detected as a rapid absorption decrease at 310 nm; a representative experiment is reported in Fig. 3. The observed fast ONOO 2 consumption is protein-mediated and dependent on the redox state of the Prxs; consistently, the reaction does not take place over the same time window (100 ms) in the absence of the proteins (thick lines in Fig. 3) or following their oxidation by an excess of H 2 O 2 (not shown).
To estimate the second-order rate constant for the reaction of ONOO 2 with reduced GiPrx1a and deltaGiPrx1b, the kinetics of ONOO 2 consumption was investigated at increasing concentrations of the two proteins. As expected, faster ONOO 2 consumption was observed at higher protein concentrations (Fig. 3). In full agreement with previous data [37], the initial rate of the reaction was found to be proportional to the protein concentration (Fig. 4) and linear regression of the data allowed us to estimate the secondorder rate constants k,4610 5 M 21 s 21 and ,2610 5 M 21 s 21 for GiPrx1a and deltaGiPrx1b, respectively. ONOO 2 decomposition is thus much faster in the presence of either of the two proteins than in their absence; for instance, the initial consumption rate of 20 mM ONOO 2 increases by 70-or 35-fold in the presence of 20 mM reduced GiPrx1a or deltaGiPrx1b, respectively. As internal control, we investigated the kinetics of the reaction of ONOO 2 with free cysteine and found that, at pH = 7 and 4uC, the reaction proceeds with a second order rate constant k,1610 3 M 21 s 21 ( Figure S4). This rate constant, consistent with the one (k,5.9610 3 M 21 s 21 ) previously measured at 37uC and pH = 7.4 by Radi et al. [43], is more than 100-fold smaller than the rate constant measured for the reaction of ONOO 2 with GiPrx1a or deltaGiPrx1b under otherwise identical experimental conditions.

Effect of O 2 on GiPrxs expression
Expression of the two Prxs in Giardia trophozoites was investigated both by immunoblotting and real time qPCR. Immunoblotting assays, however, did not allow us to discriminate between GiPrx1b and GiPrx1a (produced by either the 16076 or the almost identical 14521 gene), because polyclonal antibodies raised against either of two proteins cross-reacted with both Prxs (not shown). As shown in Fig. 5, Prxs can be immunodetected in trophozoites grown under standard anaerobic conditions and, interestingly, their expression is overall increased (,1.6 fold) when parasitic cells are exposed for 24 h to air levels of O 2 . Notably, this effect is partly reverted by addition of catalase in the medium, pointing to a role of H 2 O 2 in modulation of GiPrxs expression.
In order to discriminate between the two Prxs, transcription of the genes encoding GiPrx1a (16076 and 14521) or GiPrx1b (15383) was individually analyzed by real time qPCR in parasitic cells, as a function of the incubation time (from 1 to 24 h) with air. As shown in Fig. 6, after normalization to the mRNA level of the housekeeping ribosomal small subunit protein S26, regardless of the presence or absence of O 2 , at any incubation time the mRNA of GiPrx1a was found to be more abundant (at least 4 fold) than that one of GiPrx1b. After 1 h-exposure to air O 2 , both Prxs showed slightly increased (,2 fold) mRNA levels. However, at longer exposure times, the transcription profile of the two proteins was different: the mRNA level of GiPrx1b constantly decreased, whereas at t $6 h the transcription of GiPrx1a was stimulated, and after 24 h-exposure to air O 2 the protein mRNA level was ,3 fold higher than measured under anaerobic conditions. Interestingly, and in agreement with the immunoblotting analysis, in the presence of catalase scavenging H 2 O 2 , lower expression levels were detected for both GiPrx1a and GiPrx1b in parasitic cells exposed to air for 24 h (Fig. 7). In the light of the results obtained by real time qPCR, we conclude that GiPrx1a is the Prx primarily detected by immunoblotting.

Discussion
In several parasitic protozoa, Prxs have been shown to be implicated in virulence and drug resistance [31]. For instance, Prx1a from T. brucei has been validated as a drug target [44], and in T. cruzi the expression of two Prxs (TcruPrx1a and TcruPrx1m) was found to be up-regulated during infection, correlating with parasitic virulence [45]. Along the same line, in the invasive form of Entamoeba (E.) histolytica, the Prx expression level proved to be much higher than in the closely related E. dispar species, incapable of invasion [46]; even more revealing, the level of Prxs in nonvirulent E. histolytica strains was shown to be lower than in the  virulent ones [47], although in either strain the transcriptional levels did not change upon H 2 O 2 or NO stress [48]. A Prx-deleted mutant of the mouse parasite P. berghei showed a reduced number of gametocytes [49] and oocysts [50], and more recently in P. falciparum PfTPx-1 was found to have a hypertermal-protective function, relevant for survival of the parasite in the human body after repeated incidences of fever [51]. Over and above these functions, overexpression of Prxs in parasitic protozoa has been also shown to enhance resistance to some drugs, namely to antimony in Leishmania (L.) donovani [52], to benznidazole in T. cruzi [53,54] and to metronidazole in both E. histolytica [55] and Trichomonas vaginalis [56]. Finally, some Prxs appear to be also promising antigens for the development of new vaccines (see [31] and references therein), as exemplified by the Prx1 from L. donovani patented for such application (U.S. Patent 7795406). For all these reasons, Prxs are considered potential targets for the development of new antiparasitic treatments.
Despite the great body of information currently available on Prxs from parasitic protozoa [31], this is the first study focusing on the functional characterization of the two Prxs (GiPrx1a and GiPrx1b) identified in G. intestinalis. As a key observation, it is found that these proteins are both able to metabolize not only H 2 O 2 , but also the harmful ONOO 2 and alkyl-hydroperoxide model compounds. These results are unprecedented for G. intestinalis because, to our knowledge, no enzymes from this parasite able to catalyze these physiologically relevant reactions have been characterized to date.
Prxs typically belong to an electron transfer chain that includes a NADPH-dependent thioredoxin-reductase (TrxR) and thioredoxin (Trx) as electron carrier. A TrxR has been previously identified in Giardia, purified and characterized [57]. However, attempts to measure the activity of this enzyme in crude extracts of   the parasite in the presence of the putative Giardia Trx protein (encoded by the GL50803_3910 gene), recombinantly produced in E. coli, or its homologues from E. histolytica or T. vaginalis, were unsuccessful [58]. In the absence of the physiological redox partner of Giardia Prxs, the ability of GiPrx1a and deltaGiPrx1b to turnover with H 2 O 2 and alkyl-hydroperoxides was tested here following NADPH oxidation in the presence of the E. coli TrxR and Trx proteins ( Figs. 1 and 2). While allowing measurements, this non-physiological chimeric reducing system even at the highest concentrations tested proved to rate-limit substrate consumption by the two Prxs (see Figs. 1B and 1C), thus preventing V max measurements. Nevertheless, under identical experimental conditions the two Prxs exhibited the same apparent peroxidatic activity, regardless of the oxidizing substrate used in the assay (H 2 O 2 , CumOOH or tButylOOH, Fig. 2). Over time, a progressive slight inactivation of the two Prxs was observed as inferred from the non-linear time course of NADPH consumption measured after the addition of the enzymes to the reaction mixture (Fig. 1A). Such a slow activity decline did not revert upon readdition of NADPH (not shown); in analogy with other Prxs, it was therefore interpreted as a progressive accumulation of the inactive form of the enzyme with hyperoxidized cysteine(s) [42].
Notably, both GiPrxs are also highly reactive towards ONOO 2 (Figs. 3 and 4). In alkaline solutions, ONOO 2 is rather stable; otherwise, it decomposes rapidly (t 1/2 = 1 s at pH 7.4, 37uC) upon protonation to peroxynitrous acid (ONOOH) (pKa = 6.8) [43]. In our assays, GiPrx1a and deltaGiPrx1b were found to catalyze the consumption of ONOO 2 with second-order rate constants k,4610 5 M 21 s 21 and k,2610 5 M 21 s 21 , respectively (Fig. 4). These values are within the range of those published for Prxs from different microbial sources [37], taking into account that the published values were obtained at higher temperature (25uC or 37uC), and assuming the rates to double for every 10uC degrees increase in temperature. The occurrence of enzymes promptly degrading ONOO 2 in Giardia is not only consistent with a previous study [59], in which ONOO 2 was reported to kill parasitic cells only at high concentrations (apparent IC 50 ,3 mM), but also likely relevant for parasite survival in vivo. Giardia trophozoites are indeed known to express enzymes, such as DTdiaphorase, that by reaction with O 2 generate O 2 2N [60], a potential source of ONOO 2 in the fairly NO-rich environment of the mucosa of the proximal small intestine. Moreover, Giardia is known to utilize arginine as energy source [61] and secrete an arginine-consuming enzyme, arginine deiminase, upon interaction with intestinal epithelial cells [62]. The reduced availability of arginine, however, can establish favorable conditions for the production of ONOO 2 [63] by the NO-synthases, via the combined release of ROS and NO. In this regard, the combined action of arginine deiminase and Prxs represents a safe strategy for the parasite to counteract NO stress. In this context, it is worth mentioning that we are not aware of reports providing direct evidence for the production of ONOO 2 in the human small intestine under physiological conditions. However, a basal level of nitrotyrosine (used as an indirect marker of ONOO 2 ) has been reported in the small intestine of animal models [64,65].
The effect of air O 2 on the expression of the two GiPrxs in Giardia trophozoites has been investigated here both by immunoblotting and qPCR. As shown in Fig. 5, GiPrxs are overall already expressed to sufficiently high levels to be immunodetected in cells grown under standard anaerobic conditions, which may be consistent with Prxs being constitutively expressed to act as a first line defense against oxidative stress. Nonetheless, exposure of parasitic cells to air for 24 h caused a ,50% increase in Prxs content, further suggesting a defense role against O 2 toxicity. Consistently, qPCR experiments showed that in terms of mRNA levels GiPrx1a is more abundant than GiPrx1b and its transcription is further stimulated following cell exposure to air O 2 (Fig. 6). Notably, as proved by addition of exogenous catalase, in both type of experiments (immunoblotting and qPCR) H 2 O 2 appears to be responsible for the O 2 -mediated up-regulation of expression (Figs. 5 and 7). All together these data suggest an involvement of GiPrxs both in the early and in the late phase of the response to oxidative stress, in agreement with the ability of these enzymes to detoxify nitroxidative stressors and repair oxidatively damaged molecules.
In conclusion, GiPrxs are the first enzymatic defense system against peroxides, alkyl-hydroperoxides and ONOO 2 having been characterized in Giardia as yet. Owing to their ability to protect from nitroxidative stress, these enzymes are likely involved in parasite survival in vivo, possibly playing a role in pathogenesis. No direct information supporting such a role is currently available, but in this regard it is interesting that recently GiPrx1a has been found to be up-regulated upon interaction of Giardia trophozoites with intestinal epithelial cells [34]. Future work should aim at testing whether GiPrxs are implicated in Giardia virulence, thus representing potential drug targets.  brucei (Q71SQ4), Leishmania (L.) donovani (Q9BP39), T. cruzi (O79469). The two conserved cysteine residues in the active site are indicated in yellow. Grey blocks represent conserved residues. B) Pairwise comparison in terms of % identity of the predicted amino acid sequences. C) Schematic drawing of GiPrx1a and GiPrx1b, with the two active site cysteines (C p and C r ) and the GiPrx1b signal peptide highlighted in black and red, respectively. (DOC)