Regulatory Mechanisms of Mitochondrial 2-Cys Peroxiredoxins 1 Calcium and Magnesium Ions Modulate the Oligomeric State and Function of Mitochondrial 2-Cys Peroxiredoxins in Leishmania Parasites *

Leishmania parasites have evolved a number of strategies to cope with the harsh environmental changes during mammalian infection. One of these mechanisms involves the functional gain that allowed mitochondrial 2-Cys peroxiredoxins to act as molecular chaperones when forming decamers. This function was demonstrated to be critical for the parasite infectivity in mammals and its activation was considered to be controlled exclusively by the enzyme redox state under physiological conditions. Herein, we revealed that magnesium and calcium ions play a major role in modulating the ability of these enzymes to act as molecular chaperones, surpassing the redox effect. These ions are directly involved in the mitochondrial metabolism and now also integrate a novel mechanism to stabilize the decameric form of 2-Cys peroxiredoxins in Leishmania mitochondrion. Moreover, we demonstrated that a constitutively dimeric Prx1m mutant impairs Leishmania's survival under heat stress, supporting the central role of chaperone function of Prx1m for Leishmania parasites during the transition from insect to mammalian hosts. The Leishmania parasites, causative agents of human and canine leishmaniasis, are exposed to different growth conditions during their life cycle due to the migration from the insect to the mammalian host environment. Among the differences are the elevation of temperature, exposure to oxidants produced by the macrophages, pH acidification and lower availability of oxygen and nutrients (1). In this context, parasite survival as well as the establishment of a successful intracellular infection rely on the development of adaptive mechanisms to hostile conditions. In Leishmania infantum, for instance, one such mechanisms involves a mitochondrial 2-Cys http://www.jbc.org/cgi/doi/10.1074/jbc.M116.762039 The latest version is at JBC Papers in Press. Published on March 14, 2017 as Manuscript M116.762039 Copyright 2017 by The American Society for Biochemistry and Molecular Biology, Inc. by gest on M arch 6, 2017 hp://w w w .jb.org/ D ow nladed from Regulatory Mechanisms of Mitochondrial 2-Cys Peroxiredoxins 2 peroxiredoxin ( LiPrx1m) (2), also named tryparedoxin peroxidase, which allows the parasite to cope with heat stress during the transition from the insect (25 °C) to the mammalian host (37 oC) (2,3). Prx1 subfamily members are peroxide-scavenging enzymes that display a 2Cys catalytic mechanism and can assume distinct oligomeric states (dimers, decamers and higher-order oligomers) (4). A dual function of peroxidase and molecular chaperone has been reported for several Prx1 enzymes and seems to be modulated by changes in quaternary structure (5-7). Factors such as pH (8-10), ionic strength (8,11,12), protein concentration (3,12,13) and protein redox state (14,15) can affect the dimerdecamer equilibrium of Prx1 members, but how these factors modulate the peroxidase and chaperone activities is still poorly understood. Overoxidation of peroxidasic cysteine, in particular, has been demonstrated to shut down the peroxidase function and to enhance the chaperone activity by stabilizing oligomers larger than decamers (5,6). However, this functional switch appears to be relevant for only some members of Prx1 subfamily (16). In Leishmania, the mitochondrial Prx1 can act as molecular chaperone and peroxidase, but only its role as chaperone is crucial for the parasite infectivity in mammals (2,3). Nonetheless, in vitro studies showed that the peroxidase catalytic cycle could modulate the chaperone reservoir of LiPrx1m, favoring chaperone-active decamers when reactive cysteine is reduced (C p-SH) and chaperoneinactive dimers when C p is oxidized into CpSSCR (3). On the other hand, the enzyme from Leishmania mitochondrion (2) seems to be resistant to the over-oxidation mechanism that inactivates the peroxidase function, stabilizes high-order oligomers, and enhances the chaperone activity of fungal (5) and mammalian (6) 2-Cys Prxs. Recently, we have demonstrated that pH variations also affect the dimer-decamer equilibrium of Leishmania braziliensis Prx1m (LbPrx1m), a close orthologue of the L. infantum enzyme, indicating that the chaperone function of these proteins might not be exclusively modulated by their redox state (9). A pH shift from 8.0 to 7.0 – commonly observed in mitochondria of nutrient-deprived cells (17) – is sufficient to stabilize oxidized decamers of LbPrx1m (9). Furthermore, it is unclear whether other factors from the mitochondrial environment affect the dimerdecamer interconversion of Leishmania Prx1m and its dual function. For 2-Cys Prxs from distant-related species, it has been demonstrated that, at least in vitro, such equilibrium can be affected by ionic strength variations (11,18) and post-translational modifications (19,20). Herein, we demonstrated that the divalent cations Ca 2+ and Mg, which are important co-factors of mitochondrial enzymes involved in cellular respiration (21-23), activate the chaperone function of oxidized LbPrx1m and enhance that of the reduced enzyme via a novel mechanism of decamer stabilization. Using an LbPrx1m mutant unable to decamerize, we showed that decamer formation is crucial for both chaperone and peroxidase activities of LbPrx1m as well as for the protective role of this protein against heat stress in the parasite context. Together, our findings unveil an exclusive and redundant system in Leishmania that uses Ca 2+ and Mg, besides pH and redox mechanisms, to maintain most of Prx1m in the decameric form and, hence, support parasite survival and infectivity in the mammalian host.

The Leishmania parasites, causative agents of human and canine leishmaniasis, are exposed to different growth conditions during their life cycle because of the migration from the insect to the mammalian host environment. Among the differences are the elevation of temperature, exposure to oxidants produced by the macrophages, pH acidification, and lower availability of oxygen and nutrients (1). In this context, parasite survival as well as the establishment of a successful intracellular infection relies on the development of adaptive mechanisms for hostile conditions. In Leishmania infantum, for instance, one such mechanism involves mitochondrial 2-Cys peroxiredoxin (LiPrx1m) 4 (2), also named tryparedoxin peroxidase, which allows the parasite to cope with heat stress during the transition from the insect (25°C) to the mammalian host (37°C) (2,3).
Prx1 subfamily members are peroxide-scavenging enzymes that display a 2-Cys catalytic mechanism and can assume distinctoligomericstates(dimers,decamers,andhigher-orderoligomers) (4). A dual function of peroxidase and molecular chaperone has been reported for several Prx1 enzymes and seems to be modulated by changes in quaternary structure (5)(6)(7). Factors such as the pH (8 -10), ionic strength (8,11,12), protein concentration (3,12,13), and protein redox state (14,15) can affect the dimer-decamer equilibrium of Prx1 members, but how these factors modulate the peroxidase and chaperone activities is still poorly understood. Overoxidation of peroxidasic cysteine, in particular, has been demonstrated to shut down the peroxidase function and to enhance the chaperone activity by stabilizing oligomers larger than decamers (5,6). However, this functional switch appears to be relevant for only some members of Prx1 subfamily (16). ). This work was also supported by grants from Project "NORTE-07-0124-FEDER-000002" (co-funded by the Norte Portugal Regional Operational Programme (NORTE 2020) under the Quadro de Referência Estratégico Nacional through the European Regional Development Fund (ERDF) and by the the Portuguese Foundation for Science and Technology) and from Project "NORTE-01-0145-FEDER-000012" (supported by NORTE 2020 under the PORTUGAL 2020 Partnership Agreement through ERDF) (to A. M. T.). The authors declare that they have no conflicts of interest with the contents of this article. 1  In Leishmania, the mitochondrial Prx1 can act as molecular chaperone and as a peroxidase, but only its role as chaperone is crucial for parasite infectivity in mammals (2,3). Nonetheless, in vitro studies show that the peroxidase catalytic cycle modulates the chaperone reservoir of LiPrx1m, favoring chaperoneactive decamers when reactive cysteine is reduced (C p -SH) and chaperone-inactive dimers when C p is oxidized into C p S-SC r (3). On the other hand, the enzyme from Leishmania mitochondrion (2) seems to be resistant to the over-oxidation mechanism that inactivates the peroxidase function, stabilizes highorder oligomers, and enhances the chaperone activity of fungal (5) and mammalian (6) 2-Cys Prx.
Recently, we have demonstrated that pH variations also affect the dimer-decamer equilibrium of Leishmania braziliensis Prx1m (LbPrx1m), a close orthologue of the L. infantum enzyme, indicating that the chaperone function of these proteins might not be exclusively modulated by their redox state (9). A pH shift from 8.0 to 7.0, commonly observed in the mitochondria of nutrient-deprived cells (17), is sufficient to stabilize oxidized decamers of LbPrx1m (9). Furthermore, it is unclear whether other factors from the mitochondrial environment affect the dimer-decamer interconversion of Leishmania Prx1m and its dual function. For 2-Cys Prx from distantly related species, it has been demonstrated that, at least in vitro, such equilibrium can be affected by ionic strength variations (11,18) and post-translational modifications (19,20).
Herein, we have demonstrated that the divalent cations Ca 2ϩ and Mg 2ϩ , which are important co-factors of mitochondrial enzymes involved in cellular respiration (21)(22)(23), activate the chaperone function of oxidized LbPrx1m and enhance that of the reduced enzyme via a novel mechanism of decamer stabilization. Using an LbPrx1m mutant unable to decamerize, we showed that decamer formation is crucial for both the chaperone and peroxidase activities of LbPrx1m as well as for the protective role of this protein against heat stress in the parasite context. Together, our findings unveil an exclusive and redundant system in Leishmania that uses Ca 2ϩ and Mg 2ϩ , in addition to pH and redox mechanisms, to maintain most of Prx1m in the decameric form and hence support parasite survival and infectivity in the mammalian host.

Ca 2؉ /Mg 2؉ ions induce LbPrx1m decamerization
Analytical size-exclusion chromatography (aSEC) and smallangle X-ray scattering (SAXS) experiments at physiological pH revealed that oxidized (disulfide-bonded) LbPrx1m assumes a dimeric structure in the presence of chelating agents, whereas it assembles into decamers upon the addition of Ca 2ϩ or Mg 2ϩ (Fig. 1, a-c). Post-treatment of Ca 2ϩ decamers with EDTA induced their disassembly into dimers, showing that the cation effect is a reversible process (Table 1). Upon comparison of several divalent cations, only Mg 2ϩ and Ca 2ϩ showed similar effects on LbPrx1m oligomerization (Table 2), indicating a specific role for these ions in stabilizing LbPrx1m decamers. The hypothesis that the Ca 2ϩ /Mg 2ϩ effect would be due to ionic strength variations in the medium was discarded, as LbPrx1m presented the same SEC elution profile in the presence or absence of 150 mM NaCl (data not shown). As the effects of Mg 2ϩ and Ca 2ϩ were indistinguishable from each other, and because LbPrx1m eluted as dimers when incubated with EGTA ( Fig. 1a), which chelates Ca 2ϩ with a higher selectivity (Ͼ10 5 ) (24) than Mg 2ϩ , some of the in vitro and in silico assays described here were performed with Ca 2ϩ only.  To study the effect of increasing concentrations of Ca 2ϩ and Mg 2ϩ in the dimer-decamer equilibrium of LbPrx1m, we monitored changes in the oligomerization state of oxidized samples titrated with CaCl 2 or MgCl 2 by measuring the anisotropy of intrinsic protein fluorescence (Fig. 1d). Based on these data, we estimated a K1 ⁄ 2 near 3 mM for both ions, indicating a low-affinity system. During these assays, we noticed that the cation effect was dependent on a critical protein concentration (ϳ80 M) below which oxidized LbPrx1m dimers became less sensitive to the presence of cation (data not shown). However, after the formation of cation-stabilized decamers, the protein dilution to levels below the critical concentration did not induce decamer disassembly, indicating that cation binding to LbPrx1m involves the formation of "transitional" decamers followed by the binding of Ca 2ϩ or Mg 2ϩ to yield stable cation-decamer complexes.
As a first approach to evaluating the physiological relevance of the Ca 2ϩ /Mg 2ϩ effect in the quaternary structure of LbPrx1m, we performed aSEC assays under two conditions: first by simulating mitochondrial basal concentrations of free Mg 2ϩ and Ca 2ϩ ions (25,26) and then by mimicking a Ca 2ϩ increase to levels already reported for the L. braziliensis mitochondrion (27). Basal concentrations of Mg 2ϩ /Ca 2ϩ were sufficient to maintain most of the reduced enzyme in the decameric form, indicating that physiological levels of Mg 2ϩ /Ca 2ϩ stabilize reduced decamers (Fig. 1, e and f). We next exposed the cation-stabilized, reduced decamers to a low concentration of H 2 O 2 and evaluated the aSEC profile of the oxidized (S-Sbonded) enzyme. In basal concentrations of Mg 2ϩ /Ca 2ϩ , most of the decamers dissociate into dimers upon C p oxidation/resolution (Fig. 1, e and f), which correlates with the low affinity of these cations to the oxidized enzyme (Fig. 1d). However, under the condition simulating a calcium overload, almost half of the population remained decameric, indicating that supraphysiological Ca 2ϩ concentration already reported for Leishmania mitochondria can increase the level of oxidized decamers (Fig. 1e).

Prx1m decamer stabilization by Ca 2؉ /Mg 2؉ is redox-independent and a unique feature of the mitochondrial Prx1 from Leishmania parasites
To further investigate how the redox state affects the oligomerization of LbPrx1m, the hydrodynamic behavior of oxidized and reduced proteins was assessed under chelating conditions or in the presence of Ca 2ϩ (Fig. 2). The aSEC results indicate that when the cation is absent the dimer-decamer equilibrium becomes more responsive to the protein redox state; the oxidized enzyme remains dimeric, regardless the protein concentration (Fig. 2a), whereas the reduced enzyme gets into a dimer-decamer equilibrium that is shifted to the decamer by increasing protein concentrations (Fig. 2b). In contrast, 25 mM CaCl 2 stabilizes a major population of oxidized and reduced decamers even when they are diluted to low protein concentrations (Fig. 2, c and d), supporting the conclusion that the Ca 2ϩ effect surpasses the redox state in stabilizing LbPrx1m decamers.
Unlike LbPrx1m, the cytoplasmic 2-Cys Prx from the same pathogen, as well as from two evolutionarily distant organisms, were not as dependent as LbPrx1m on the decamer-stabilizing effect of Ca 2ϩ , eluting mainly as decamers regardless the presence of this ion or EDTA (Fig. 3). These findings indicate that the high sensitivity of dimer-decamer equilibrium to Ca 2ϩ / Mg 2ϩ ions is a unique feature of LbPrx1m and possibly of other mitochondrial orthologues from Leishmania spp, according to the comparative structural and sequence analyses described below.

Structural basis for the cation-dependent mechanism of decamer stabilization
Despite extensive efforts, we were unable to crystallize LbPrx1m in complex with Ca 2ϩ . Crystals of LbPrx1m decamers prestabilized with CaCl 2 were observed only under acidic conditions (pH 4.4), known to enhance decamer stability in a cation-independent manner (9) and to decrease the Ca 2ϩ -binding affinity to proteins (28,29). Thus, to determine the Ca 2ϩ / Mg 2ϩ -binding site in LbPrx1m, we analyzed the crystal structure of the pH-stabilized decamer (PDB accession no. 4KB3 (9)) and used in in silico approaches combined with site-directed mutagenesis to validate the predicted site.
Like other members of the AhpC/Prx1 subfamily (30), the LbPrx1m decamer is formed when five dimers bind to each other via the A-type interface (9). Because the formation of this interface depends on conformational changes in region I (residues 75-79) preceding the C p -loop (residues 80 -84) (9), we hypothesized that the decamer-stabilizing effect of Ca 2ϩ /Mg 2ϩ might be related to their binding to the A-type interface and the concomitant stabilization of region I in a conformation that favors the decameric assembly.
To test this hypothesis, we searched in the A-type interface of LbPrx1m decamer (9) for negatively charged cavities in which a positive ion such as Ca 2ϩ could bind and maintain the closed Samples were pretreated with CaCl 2 (I) and then incubated for 10 min with EDTA (II) to illustrate the reversibility of the decamerization process. Note that the hydrodynamic radius (R H ) of the protein decreases upon the addition of EDTA in samples pretreated with CaCl 2 .

Regulatory mechanisms of mitochondrial 2-Cys peroxiredoxins
conformation of region I required for decamer stabilization (9). As expected, we found a site in which Ca 2ϩ could be coordinated by residues from both interfacing subunits, assuming a distorted trigonal bipyramidal geometry according to molecu-lar dynamics simulations (Fig. 4a). This site includes the main chain of Asp-76, a residue from region I, and is duplicated at each A-type interface, implying a stoichiometry of 10 cations/decamer. Numbers above the graphs represent the molecular mass (kDa) of standard proteins used for column calibration. For this assay, the His tag was removed, using TEV protease to show that the untagged protein behaves similar to the His-tagged samples upon CaCl 2 and EDTA treatments (see Fig. 1a). Note that the Ca 2ϩ -stabilized decamers, preformed at 94 M protein, did not dissociate upon protein dilution even in the absence of DTT. Samples not treated with DTT are air-oxidized (S-S-bonded). mAU, milliabsorbance units. We next evaluated the Ca 2ϩ effect on the aSEC profile of mutants lacking one of the side chains predicted to coordinate this cation. These side chains belong to the residues Asp-108, Ser-109, and Ser-112, located in region II (residues 107-120) of the A-type interface (Fig. 4). The mutants D108A and S109A eluted essentially as dimers in the presence of Ca 2ϩ at pH 7.5, whereas mutant S112A showed a concentration-dependent behavior, eluting as a decamer at 86 M and as a dimer at 13 M (Fig. 4, c and d). These results indicate that Asp-108 and Ser-109 are essential for the Ca 2ϩ -dependent stabilization of the LbPrx1m decamer, whereas Ser-112 plays a facultative role in this mechanism.
Intriguingly, the cation-binding site identified in LbPrx1m is highly conserved in some 2-Cys Prx, where the decamers remain stable without Ca 2ϩ /Mg 2ϩ ions, such as the cytoplasmic peroxiredoxin from Trypanosoma cruzi (TcPrx1a) in which the decamer is crystallized in the presence of EDTA (31) and human Prx2 (HsPrx2) (Fig. 4b). This finding prompted us to search for other structural elements that would possibly account for the Ca 2ϩ /Mg 2ϩ effect on LbPrx1m.
A comparison of the A-type interfaces of the LbPrx1m, TcPrx1a, and HsPrx2 crystallographic decamers (Fig. 5) showed that residues Cys-107 and Met-139 are exclusive of LbPrx1m and expose Asp-108 to the electrostatic attraction of Lys-137 ( Fig. 5a). However, in TcPrx1a, the Asp-108 counterpart (Asp-79) is shielded from the influence of such a lysine by a residue that is bulkier than Cys-107 (Met-78), which seems to favor the interaction between the Asp-79 side chain and a water molecule that occupies the predicted cation-binding site (Fig. 5b). The same water-mediated link occurs in HsPrx2, which lacks the corresponding Lys-137 (Fig. 5c). Thus, we hypothesized that the attraction of the Asp-108 side chain by Lys-137 prevents Asp-108 from coordinating a water molecule, but not a Ca 2ϩ or Mg 2ϩ ion, within the Ser/Asp cluster, destabilizing the A-type interface when Ca 2ϩ or Mg 2ϩ is absent and the medium is alkaline (9). To test this hypothesis, we mutated Cys-107 (region II) to a methionine and Met-139 (region III) to a lysine, thus mimicking the TcPrx1a Asp-79 (equivalent to Asp-108) microenvironment in which Met-78 blocks the access of the lysine to the aspartate (Fig. 5b). As envisioned, this double mutation rendered LbPrx1m decamer stabilization Ca 2ϩ -independent (Fig. 4c), demonstrating that changes in regions II and III of the A-type interface gave rise to the Ca 2ϩ /Mg 2ϩ effect on LbPrx1m oligomerization (Fig. 5a).

Ca 2؉ /Mg 2؉ ions enhance the peroxidase activity of LbPrx1m
The dimer-decamer switch seems to play a role during the peroxidase catalytic cycle of AhpC/Prx1 subfamily members by . Shown as sticks are the residues involved in the cation-dependent mechanism of decamer stabilization: Asp-76, region I; Cys-107, Asp-108, Ser-109, and Ser-112, region II; Lys-137 and Met-139, region III. b, sequence alignment of A-type interface regions from LbPrx1m and two cation-independent 2-Cys Prx, highlighting the residues predicted to play a role in cation binding (red boxes) and those predicted to determine the cation dependence of decamer stabilization (green boxes). c, aSEC chromatograms of air-oxidized LbPrx1m D108A, S109A, S112A, and C107M/M139K mutants at 86 M (500 l) in Tris buffer (pH 7.5) containing 25 mM CaCl 2 (red lines) or 5 mM EDTA (green line). The single mutant C107M was unstable in solution, and thus it was not included in our analyses. mAU, milliabsorbance units. d, aSEC chromatogram of air-oxidized LbPrx1m WT and S112A mutant at 13 M (200 l). Note that WT protein eluted as decamer, whereas the S112A mutant eluted mainly as dimer under the same conditions. aSEC assays were carried out in Tris buffer (pH 7.5) containing 25 mM CaCl 2 .

Regulatory mechanisms of mitochondrial 2-Cys peroxiredoxins
influencing enzymatic efficiency (15,32). We thus used the trypanothione-dependent enzyme cascade from Leishmania to evaluate in vitro how the decamer-stabilizing effect of Ca 2ϩ / Mg 2ϩ influences the capability of LbPrx1m to reduce hydrogen peroxide.
The chelation of metal ions drastically decreased the peroxidase activity of purified LbPrx1m (Fig. 6a). In samples pretreated with EDTA and then supplemented with Ca 2ϩ or Mg 2ϩ , the catalytic activity was recovered, showing that these divalent cations play a role in the peroxidase function of LbPrx1m (Fig.  6a). The increment of Ca 2ϩ in purified LbPrx1m did not alter its catalytic activity (Fig. 6a), an effect that can be ascribed to divalent cations from Escherichia coli that remained bound to LbPrx1m decamers during the purification process, as indicated by aSEC analysis of untreated samples (results not shown). The observed effects of EDTA and Ca 2ϩ cannot be attributed to interference from the other molecules of the trypanothione cascade, because these additives did not alter the peroxidase activity of cytoplasmic Prx1 from L. infantum (LiPrx1a) (Fig. 6b). Importantly, the observation that the Ca 2ϩ /Mg 2ϩ ions did not affect the peroxidase activity of LiPrx1a correlates with the fact that Ca 2ϩ /Mg 2ϩ ions are not required to stabilize the decameric structure of its ortholog in L. braziliensis (Fig. 3), corroborating our view of the spe- To demonstrate that the effect of Ca 2ϩ on peroxidase activity is due to its binding at the A-type interface and consequent decamer stabilization, we compared the peroxidase activity of Ca 2ϩ -binding site mutants with that of the wild-type (WT) protein or the mutant H113A (Fig. 6, a-c), which preserves the dimeric structure but is unable to decamerize in response to pH (9), redox, and cat-ion stimuli (Fig. 7a). Mutants D108A and S109A displayed only residual activity compared with the WT enzyme in the presence of Ca 2ϩ (Fig. 6c). As expected, the mutation S112A was not as efficient as the D108A and S109A substitutions in decreasing the peroxidase activity of LbPrx1m, which correlates with the facultative role of Ser-112 in Ca 2ϩ binding. The effects of the D108A and S109A mutations on peroxidase activity were equivalent to that caused by the decamer-disrupting H113A substitution, further . Ca 2؉ and Mg 2؉ ions enhance both peroxidase and chaperone activities of LbPrx1m. a, relative peroxidase activity of LbPrx1m left untreated, pretreated with 5 mM EDTA only, pretreated with 5 mM EDTA followed by the addition of 25 mM CaCl 2 or MgCl 2 , and pretreated with 20 mM CaCl 2 or MgCl 2 alone. b, relative peroxidase activity of LiPrx1a prepared as described for LbPrx1m. The absolute peroxidase activity values (mol NADPH⅐min Ϫ1 ⅐mg Ϫ1 ) were 0.18 Ϯ 0.04 (for untreated LbPrx1m) and 5.12 Ϯ 1.19 (for untreated LiPrx1a). c, comparison between the relative peroxidase activity of WT LbPrx1m and mutants pretreated with 5 mM EDTA followed by the addition of 25 mM CaCl 2 . d, relative chaperone activity of air-oxidized LbPrx1m left untreated and pretreated with 5 mM EDTA only or 5 mM EDTA followed by the addition of 25 mM CaCl 2 or MgCl 2 . The same treatments were carried out with the protein reduced using 2 mM DTT. e and f, comparison between the relative chaperone activity of WT LbPrx1m and muteins pretreated with 5 mM EDTA followed by the addition of 25 mM CaCl 2 without DTT (air-oxidized) (e) or with 2 mM DTT (reduced) (f). *, p Ͻ 0.1; **, p Ͻ 0.05; ***, p Ͻ 0.01. All experiments were performed in triplicate.

Regulatory mechanisms of mitochondrial 2-Cys peroxiredoxins
supporting that Ca 2ϩ /Mg 2ϩ ions enhance LbPrx1m peroxidase activity by stabilizing its decameric form.

Reduced and oxidized cation-stabilized decamers possess the required conformation to perform chaperone function
Our finding that Ca 2ϩ /Mg 2ϩ ions stabilize oxidized decamers of LbPrx1m led us to investigate whether these decamers were able to suppress luciferase thermal aggregation similarly to the reduced decamers (3). Under non-reducing conditions, metal chelation inhibited the chaperone activity of LbPrx1m, indicating a role for Ca 2ϩ /Mg 2ϩ in activating the chaperone function of oxidized LbPrx1m (Fig. 6d). When cation-free samples were supplemented with CaCl 2 or MgCl 2 , the chaperone activity was recovered, reaching levels significantly higher than that of the untreated samples in the case of Ca 2ϩ (Fig. 6d). Although LbPrx1m reduction also rescued the chaperone activity of cation-free LbPrx1m, the average activity of the reduced decamers was further stimulated by CaCl 2 supplementation (Fig. 6d), which is in agreement with aSEC data showing the better performance of Ca 2ϩ than protein reduction in stabilizing decamers (Fig. 2).
To demonstrate that protection against luciferase aggregation depends on the formation of decamers and to evaluate the effect of Ca 2ϩ /Mg 2ϩ -binding site mutations on chaperone function, we measured the chaperone activity of LbPrx1m mutants H113A, D108A, S109A, and S112A. The H113A substitution, which prevents the decameric assembly of dimers, completely abolished the chaperone activity of oxidized LbPrx1m and drastically decreased that of the reduced protein, demonstrating that decamer formation is a prerequisite for chaperone activity (Fig. 6, e and f). Unexpectedly, of the four Ca 2ϩ -binding site mutations, only D108A significantly decreased the chaperone activity under both non-reducing and reducing conditions (Fig. 6, e and f). However, the mutant D108A was more active than H113A, which prompted us to investigate whether the high temperature of the chaperone assay favors the formation of D108A decamers. In agreement with the chaperone activity data, mutant D108A entered into a dimer-decamer equilibrium at 42°C (Fig. 7, b and e), whereas mutant H113A remained in the dimeric state at this temperature (Fig. 7, a and d). Under the same conditions, the WT protein was decameric, indicating that decamer formation is necessary and sufficient to trigger the chaperone function of LbPrx1m (Fig. 7, c and f).
Based on studies of reduced decamers, it has been proposed that the chaperone function of Leishmania Prx1m is activated by an increase in temperature, which induces conformational rearrangements that expose hydrophobic regions (3). As our data showed that oxidized decamers also display chaperone activity, we investigated whether their behavior upon temperature increase supports the model in which thermo-induced conformational changes activate the chaperone function of Leishmania Prx1m. For this purpose, circular dichroism (CD) spectra of WT LbPrx1m were collected under reducing and non-reducing conditions at 25 and 42°C in the presence of Ca 2ϩ . Interestingly, at 25°C, the CD spectra of the reduced and oxidized decamers were virtually identical, showing that the conformational differences between their respective fully folded and locally unfolded active sites were undetectable by this technique (Fig. 8, a and b). Upon a temperature increase from 25 to 42°C, the CD spectrum of the reduced decamers showed a decrease in the absolute values at 193 and 222 nm and presented a shift of the minimum at 208 nm toward smaller wavelengths, indicating the loss of ␣-helical structures (Fig. 8a). In agreement with this result, differential scanning fluorimetry (DSF) data showed the exposure of hydrophobic surfaces in reduced decamers heated up to 42°C (Fig. 8c). However, no significant changes in CD spectra or exposure of hydrophobic regions in DSF analyses were observed for the oxidized decamers in the same temperature range (Fig. 8, b and c). This enhanced thermostability of the oxidized species correlates with the presence of the disulfide bond linking C p , located in the loop connecting strand ␤3 to helix ␣2, and C r , placed at the C-terminal extension downstream the helix ␣6. Because this is the only disulfide bond present in oxidized LbPrx1m (9), the loss of structure indicated by CD and DSF analyses of reduced decam-ers likely reflects, among other events, the unfolding of helices ␣2 and/or ␣6, which is favored when C p and C r are reduced.
Considering that the helices ␣2 and ␣6 are located at the external surface of the decamer, i.e. far from the luciferasebinding site (3), their unfolding might not interfere with the chaperone function of reduced decamers (Fig. 9). In support of this hypothesis, the Ca 2ϩ -stabilized reduced species preserved the decameric structure at 42°C (data not shown) and were as effective as the oxidized samples in suppressing the thermal aggregation of luciferase in vitro (Fig. 6d). Together, these analyses suggest that the thermo-induced conformational changes observed for the reduced decamer might not be required to activate the chaperone function of Leishmania Prx1m.
For comparison purposes, we also performed CD and DSF analyses for the WT LbPrx1m in presence of EDTA, a condition that favors the dimeric state, and the mutant H113A, which is fully dimeric in solution regardless of its redox state (Fig. 8, d-i). (c and f) as well as mutant H113A (i) treated as described for the CD experiments. In these assays, WT LbPrx1m data in the presence of Ca 2ϩ reflect the CD and DSF profiles of decamers, whereas those of WT LbPrx1m in the presence of EDTA or those of the mutant H113A represent the behavior of the dimers. The calculated melting temperatures are color-coded according to the respective curves. In c, note that the dashed line indicates the temperature of the chaperone assay in which the reduced decamers expose hydrophobic patches, whereas the oxidized decamers remain almost completely folded.
These experiments confirmed that the mutant H113A is properly folded (Fig. 8, g and h) and revealed that the C p S-SC r disulfide bond has a higher impact in thermostability than variations in the oligomeric state, leading to an increase of at least 12°C in the melting temperature of both the dimers and the decamers (Fig. 8, c, f, and i).

Preventing Prx1m decamer formation impairs leishmanial survival at 37°C
Previous work has shown that L. infantum parasites devoid of LiPrx1m are thermosensitive when exposed to 37°C, a phenotype that is partially reverted upon reintroduction of the enzyme (2). To investigate the phenotypic implications of Prx1m mutants unable to decamerize or with a lesser tendency to form Ca 2ϩ /Mg 2ϩ -stabilized decamers, L. infantum knockout parasites for LiPrx1m (prx1m Ϫ ) were transfected with WT LbPrx1m and the corresponding H113A and D108A muteins. Using indirect immunofluorescence (Fig. 10a), Western blotting (Fig. 10b), and PCR (data not shown) analyses, it was verified that knock-out parasites lacked the expression of LiPrx1m, whereas the transfected parasites expressed the WT protein and mutants into the mitochondrion.
As expected, at 25°C, all parasites had the same growth rates (Fig. 10c). However, at 37°C, the H113A dimers were incapable of rescuing the thermosensitive phenotype of Prx1m Ϫ promastigotes, demonstrating that the decameric state of Prx1m is crucial for Leishmania survival at temperatures akin to those encountered in the mammalian host (Fig. 10, d and e).
When exposed to 37°C, knock-out parasites expressing the D108A mutein, which is less prone to form Ca 2ϩ /Mg 2ϩ -stabilized decamers, presented a behavior similar to those complemented with the WT protein (Fig. 10, d and e). This result is likely explained by the fact that the chaperone activity of this mutant retains a residual response to Ca 2ϩ (Fig. 10f). Furthermore, mutant D108A is susceptible to the influence of medium acidification, which stabilizes a subpopulation of D108A as decamers in vitro (Fig. 7b) and probably counterbalances the lower responsiveness of D108A decamers to Ca 2ϩ /Mg 2ϩ in vivo. Interestingly, the chaperone activity of the mutant D108A in vitro is unresponsive to DTT treatment even under cationfree conditions (Fig. 10f), indicating that decamer-stabilizing factors other than protein reduction support the chaperone function of this mutein in vivo.

Ca 2؉ /Mg 2؉ ions stabilize decamers and activate the dual function of mitochondrial 2-Cys Prx from Leishmania parasites
In this work, we have revealed that Ca 2ϩ and Mg 2ϩ ions affect the quaternary structure and the dual function of mitochondrial 2-Cys peroxiredoxins from Leishmania parasites. Our data show that these divalent cations stabilize LbPrx1m decamers and thereby stimulate peroxidase and chaperone activities. The mechanism involves the binding of Ca 2ϩ or Mg 2ϩ at transitional A-type interfaces, stabilizing dimer-dimer interactions. Although the K1 ⁄ 2 estimated for Ca 2ϩ and Mg 2ϩ binding was near 3 mM for the oxidized enzyme, our analyses suggest that this affinity can be enhanced by protein reduction, which stabilizes transitional decamers and favors the formation of cation-bound decamers. Supporting this hypothesis, air-oxidized samples treated with EDTA and then incubated with 1 mM Ca 2ϩ behaved essentially as dimers (Fig. 1d); however, when these EDTA-treated samples were reduced with DTT, incubated with similar amounts of cation, and then reoxidized, about 50% of the enzyme remained in the decameric form (Fig.  1, e and f). In other words, the amount of cation-bound oxidized decamers increases when these complexes are formed under reducing conditions prior to enzyme oxidation. In summary, our data indicate that physiological concentrations of free Ca 2ϩ /Mg 2ϩ stabilize reduced decamers, stimulating the peroxidase function and contributing to the maintenance of a basal pool of chaperone-active LbPrx1m at the alkaline environment of mitochondria (pH 7.5-8.0) (33). When mitochondrial Ca 2ϩ uptake is stimulated (23), the level of oxidized decamers in- Figure 9. The disulfide bond between C p and C r probably suppresses the unfolding of helices ␣2 and or ␣6 at 42°C. a, schematic representation of LbPrx1m crystallographic decamer highlighting helices ␣2 (pink) and ␣6 (violet). b, magnified view of the boxed region in a, showing the disulfide bond between C p (gray, carbon atoms) and C r (green, carbon atoms). The C-terminal extension harboring C r is shown in green.
creases, contributing to an enlarged chaperone reservoir of Prx1m in Leishmania.

The closed conformation of region I is required to stabilize the fully folded conformation of the reduced C p -loop
According to data from the literature, the decameric structure of 2-Cys Prx contributes to stabilizing the fully folded conformation of the active site, allowing the optimal orientation of substrate and the activation of the catalytic C p to reduce the peroxide oxygen (15,32). Our data support the conception that for the C p -loop to adopt the fully folded conformation, the adjacent region I might be stabilized in a closed conformation via interactions involving residue Asp-76 (Fig. 11). The disruption of such interactions by point mutations severely decreases the peroxidase activity of LbPrx1m (Fig. 6c). Although the link between the Asp-76 and His-113 side chains is intramolecular and direct, those involving the Asp-76 main chain require decamer assembly and are mediated by Ca 2ϩ or Mg 2ϩ ions in the case of LbPrx1m (Fig. 11). Our data revealed that Mg 2ϩ or Ca 2ϩ ions increase the peroxidase activity by binding at the A-type interface of LbPrx1m, thus holding the Asp-76 main chain with the assistance of Asp-108, Ser-109, and optionally Ser-112 side chains. Noticeably, mutations at positions 108 and 109 displayed a higher impact on peroxidase activity in comparison with EDTA treatment, suggesting that, in the absence of divalent cations, a solvent molecule could link this Ser/Asp cluster allowing a suboptimal activity. Indeed, in several members of the AhpC/Prx1 subfamily, from bacteria to mammals, a water molecule plays the role of linking the highly conserved Ser/Asp cluster at the A-type interface (Fig. 12). By converting LbPrx1m into a cation-independent 2-Cys Prx, we have provided strong evidence that the microenvironment of Asp-108 selects LbPrx1m as a Ca 2ϩ /Mg 2ϩ -sensitive Prx1. According to our model, LbPrx1m requires Ca 2ϩ /Mg 2ϩ to surpass the electrostatic attraction caused by Lys-137 on Asp-108 and to restore the solvent-mediated link that holds the main chain of Asp-76 and contributes to maintain the fully folded conformation of the C p -loop, which is mandatory for substrate binding and catalysis. Comparative sequence analysis suggests that the high Ca 2ϩ /Mg 2ϩ sensitivity observed for LbPRx1m extends exclusively to mitochondrial 2-Cys Prx from the Leishmania species (Fig. 12).

Ca 2؉ /Mg 2؉ ions compose a redundant molecular system that supports the chaperone function in vivo
Our studies have demonstrated the similar efficiency of oxidized and reduced cation-stabilized decamers in preventing luciferase aggregation under heat stress conditions. This finding suggests that the Leishmania reservoir of Prx1m chaperones is not only formed by reduced proteins, as envisaged previously (3), but can also comprise oxidized Prx1m (S-Sbonded). As the S-S-bonded species represent about half of the Prx1m population present in the parasite (2), their capability to Figure 11. Stabilization of region I in the closed conformation is required to maintain the fully folded conformation of the C p -loop upon C p reduction. Shown is a 3D alignment of the loop-helix active-site motif (region I ϩ C p -loop) of LbPrx1m-oxidized dimer (shades of violet; PDB accession no. 4KCE), LbPrx1m-oxidized decamer (shades of orange; PDB accession no. 4KB3) after molecular dynamics simulations with Ca 2ϩ bound (green sphere) at the A-type interface (gray surface), and reduced decamer TcPrx1a (shades of pink; PDB accession no. 4LLR). Note that region I adopts an open conformation in dimers, but it assumes a closed conformation in the oxidized and reduced decamers. According to our data, the stabilization of region I in the closed conformation requires at least an H-bond between Asp-76 and His-113 as well as Ca 2ϩ (green sphere), Mg 2ϩ , or (less efficiently) the water molecule at the A-type interface. The establishment of such interactions is crucial to maintain the fully folded (FF) form of the reduced C p -loop. In the LbPrx1m dimer, the high entropy of the region I, trapped in an open conformation in chain B but disordered in chain A, might propagate to the adjacent C p -loop, hampering the stabilization of fully folded conformation required for substrate binding and catalysis. LU, locally unfolded. Figure 12. Stabilization of region I in the closed conformation is mediated by Ca 2؉ /Mg 2؉ in mitochondrial Prx1 from Leishmania species and by a water molecule in other AhpC/Prx1 subfamily members. a, sequence alignment of regions I, II, and III from the A-type interface as well as the C ploop from mitochondrial (Prx1m) and cytoplasmic (Prx1a) 2-Cys Prx from trypanosomatids with known crystallographic decamers. Note that residues involved in cation binding in LbPrx1m are highly conserved in the AhpC/Prx1 subfamily (red boxes). However, only the mitochondrial enzymes from Leishmania (black outlined area) conserve residues Cys-107, Lys-137, and Met-139 (green boxes) that render decamer stabilization of LbPrx1m highly sensitive to Ca 2ϩ and Mg 2ϩ ions (green circle). In other 2-Cys Prx, Lys-137 is replaced by neutral polar or hydrophobic residues or kept away from Asp-108 by residues bulkier than Cys-107. In most of the analyzed structures of AhpC/Prx1 subfamily members (blue circles), a water molecule mediates the link between the Asp-76 main chain and Asp-108, Ser-109, and eventually Ser-112 side chains. form chaperone-active decamers may be of great relevance for Leishmania virulence, considering that the chaperone function of Prx1m is crucial for the parasite survival in the mammalian host (3). Moreover, the resistance of Leishmania Prx1m to C p overoxidation (2) and the lack of the sulfiredoxin protein in these parasites further supports the importance of an alternative mechanism to modulate the chaperone activity of Prx1m in Leishmania species.
When LbPrx1m dimers fail to respond to the three components of the decamer-stabilizing system (pH, redox state, and Ca 2ϩ /Mg 2ϩ ), they lose their capacity to rescue the temperature-sensitive phenotype of prx1m Ϫ promastigotes, as indicated by our studies with mutant H113A. Based on our results, we suggest that the residue His-113 is not directly involved in cation binding but is necessary for the formation of a transitional A-type interface that is then stabilized by cation binding, His-113 protonation (9), or less effectively, by C p reduction. The crystal structure of oxidized LbPrx1m dimers shows that region I preceding the C p -loop is highly flexible and can transit between an open and a closed conformation, which favors dimers and decamers, respectively (9). The prerequisite to form transitional A-type interfaces likely involves the H-bond between His-113 and Asp-76 when region I accesses the closed conformation. Because this interaction is insufficient to lock region I in the closed conformation, decamer stabilization requires a second stimulus. According to our model (Fig. 13), in basal concentrations of enzyme, Mg 2ϩ , and Ca 2ϩ , most of Prx1m enzymes are decameric when reduced and dimeric when oxidized. During heat shock, which stimulates Ca 2ϩ uptake in Leishmania promastigotes (34), higher levels of Ca 2ϩ at the mitochondrion (27) feed the chaperone reservoir with oxidized decamers. Besides stimulating mitochondrial Ca 2ϩ uptake, cytosolic Ca 2ϩ elevations can also lead to a mitochondrial pH decrease of about 0.2 pH units (35), providing an extra stimulus to enlarge the reservoir of Prx1m decamers (9). At pH 7, for example, we detected in vitro a cooperative effect between pH and Ca 2ϩ in stabilizing oxidized decamers (data not shown).
Together, our studies show that basal concentrations of Mg 2ϩ /Ca 2ϩ ions support the dual function of mitochondrial Prx1 from Leishmania and reveal a molecular mechanism that may help explain why calcium uptake is crucial for Leshimania thermotolerance and differentiation in the mammalian host (34). Furthermore, we have demonstrated that the decameric structure, independently of its redox state, is both necessary and sufficient for the protective effect of Prx1m against heat During the peroxidase cycle, Leishmania Prx1m transits between the dimeric (chaperone-inactive) and the decameric (chaperone-active) forms. This oligomeric shift is regulated by the conformational changes of two regions from the loop-helix active-site motif: the C p -loop (LU N FF), which is redox-sensitive, and the region I (open N closed), which is Ca 2ϩ /Mg 2ϩ /pH-sensitive. When dimers are reduced (SH), they tend to form weak decamers in which free Mg 2ϩ and Ca 2ϩ ions bind to enhance their peroxidase and chaperone functions. In basal conditions, after the peroxidation (I) and resolution (II) steps, oxidized (S-S) decamers tend to release Mg 2ϩ and Ca 2ϩ ions and dissociate into dimers, a process that involves an intermediate state, here named the transitional A-type interface. However, under heat-stress conditions, two stimuli can boost the chaperone reservoir of Prx1m, mainly by stabilizing the oxidized decamers: red arrows, small pH decreases; green arrows, Ca 2ϩ overload. Because our data could not discriminate the preferable substrate for tryparedoxin (Prx1m dimers or decamers), we labeled the C p reduction step with a question mark. The same is applicable for the conformational state proposed for the active site in reduced dimers and reduced weak decamers. For purposes of clarity, the padlocks illustrating the conformational state of region I are shown only for a pair of dimers at the decamers. LU, locally unfolded; FF, fully folded. stress in Leishmania, a vital attribute for the establishment of a successful infection in the mammalian host (2,3). This finding implies that the search for compounds that prevent Prx1m decamerization represents the best strategy for inhibiting the crucial chaperone function of this attractive therapeutic target (36). Zhao et al. (37) already have identified chaperone inhibitors for the human Prx I, demonstrating the feasibility of such an approach.

Molecular cloning and site-directed mutagenesis
The LbPrx1m gene (RefSeq accession no. XM_001562186.1) was cloned into a pET28a-His-TEV vector as described previously (38). Human gene PRX2 (RefSeq accession no. NM_005809.5) was amplified by PCR and cloned into the pET28a vector between the NdeI and SalI restriction sites. The TSA1 gene from Saccharomyces cerevisiae was cloned into the pET15b vector as described previously (39). The pET28a construct containing the DNA sequence of the LbPrx1a gene (RefSeq accession no. XM_001563506.1) between the NdeI and SalI restriction sites was purchased from GenScript (Piscataway, NJ). All LbPrx1m mutants were produced using the QuikChange TM site-directed mutagenesis kit (Stratagene).

Protein expression and purification
The protein LbPrx1m and corresponding muteins were expressed and purified as described previously (9). LbPrx1a and HsPrx2 were produced in E. coli BL21(DE3)⌬SlyD cells containing the plasmid pRARE2, whereas TSA1 was produced in BL21(DE3) cells. After the cell culture reached A 600 nm ϳ 0.6 in LB medium, protein expression was induced with 0.5 mM isopropyl 1-thio-␤-D-galactopyranoside at 30°C for 4 h at 200 rpm (HsPrx2 and LbPrx1a) or with 1 mM isopropyl 1-thio-␤-D-galactopyranoside at 37°C for 3 h at 200 rpm (TSA1). Protein extraction and affinity purification followed that described for LbPrx1m (9). All purified proteins were dialyzed against the buffer used in the analytical size-exclusion chromatography step and concentrated using Amicon Ultra devices (Millipore). The oxidized (S-S-bonded) state of the purified proteins was confirmed by SDS-PAGE analysis under non-reducing conditions, whereas the reduced state of proteins treated with DTT was confirmed by quantification of free thiol groups using 5,5Јdithiobis(nitrobenzoic acid) as described previously (9).

Analytical size-exclusion chromatography
A total of 2 ml of LbPrx1m at 130 M, LbPrx1a at 43 M, yeast TSA1 at 43 M, and HsPrx2 at 130 M were loaded onto a HiLoad 16/600 Superdex 200 column (GE Healthcare) preequilibrated with 25 mM Tris-HCl (pH 7.5) containing 25 mM CaCl 2 , 25 mM MgCl 2 , 5 mM EDTA, or 5 mM EGTA. aSEC experiments for the LbPrx1m mutants D108A, S109A, and S112A at 86 M (500 l input) were carried out using the same column pre-equilibrated with buffer T-Ca (25 mM Tris-HCl, 150 mM NaCl, and 25 mM CaCl 2 , pH 7.5). The same assay was performed using the mutant C107M/M139K at 86 M (500 l input) in buffer T-EDTA (25 mM Tris-HCl, 150 mM NaCl, and 5 mM EDTA, pH 7.5). Additionally, WT LbPrx1m was cleaved with TEV protease to remove the His tag, preincubated at 94 M with or without 10 mM DTT in buffer T-Ca or T-EDTA, and divided into three samples at different protein concentrations (94, 23, and 9 M) that were loaded (250 l) onto a Superdex 200 10/300 GL column (GE Healthcare) pre-equilibrated with the sample buffer.
aSEC assays mimicking in vivo conditions (3, 25-27, 40, 41) were performed with TEV-cleaved LbPrx1m at 100 M (200 l input) in buffer T-EDTA plus the following additives: 5.7 mM MgCl 2 (free Mg 2ϩ ϭ 0.7 mM) and 200 nM CaCl 2 (condition I); and the same as described in condition I but with 90 M CaCl 2 (condition II). These assays were carried out in a Superdex 200 10/300 GL column (GE Healthcare). Samples were pretreated and eluted in the presence of 2 mM DTT or pretreated with 2 mM DTT, re-oxidized, and eluted in buffer without DTT. To obtain re-oxidized samples, DTT was removed as described by Morais et al. (9), and the concentrated protein was incubated with H 2 O 2 in a 1:1 (protein:H 2 O 2 ) molar ratio prior injection. The redox state of the samples was analyzed by non-reducing SDS-PAGE.
aSEC assays of LbPrx1m WT and S112A mutant at 13 M (200 l input) were performed in buffer T-Ca using a Superdex 200 10/300 GL column (GE Healthcare). Additional aSEC assays comparing WT LbPrx1m and H113A and D108A mutants at 48 M (250 l input) were performed at pH 4.0, as described in Morais et al. (9), in buffer T-EDTA plus 2 mM DTT or in buffer T-Ca. For comparative purposes, the molar concentrations estimated for all analyzed proteins refer to monomers. Columns were calibrated using the gel filtration calibration kits LMW and HMW (GE Healthcare).

Dynamic light scattering (DLS)
DLS measurements were performed on a Dynapro molecular sizing instrument at 25 or 42°C. Protein samples at 100 M were centrifuged previously for 20 min at 20,000 ϫ g. Data were collected at intervals of 10 s with at least 100 acquisitions. The diffusion coefficient (D) was determined from the analysis of measured time-dependent fluctuations in the scattering intensity and used to calculate the hydrodynamic radius (R H ) of the protein according to the Stokes-Einstein equation. Data analysis was performed using the software Dynamics V6.3.40.

Small angle X-ray scattering
SAXS data were collected at the D02A/SAXS2 beamline (Brazilian Synchrotron Light Laboratory, Campinas, Brazil). The radiation wavelength was set to 1.48 Å and a 165-mm MarCCD detector was used to record the scattering patterns. The sample-to-detector distance was set to 1534.5 mm to give a scattering vector range from 0.25 to 2.5 nm Ϫ1 . Protein samples at 108 M were prepared in 25 mM Tris-HCl (pH 7.5) with 5 mM EDTA or 25 mM CaCl 2 . Frames with an exposure time of 600 s were recorded, and buffer baselines were collected under identical conditions. Background scattering was subtracted from the protein-scattering pattern, which was then normalized and corrected. Experimental data fitting and evaluation of the pairdistance distribution function P(r) were performed using the program GNOM (42). The low-resolution envelopes were determined using ab initio modeling as implemented in the program DAMMIN (43). An averaged model was generated using the package DAMAVER (44) The low-resolution model and the crystal structure were superimposed using the program SUPCOMB (45).

Fluorescence anisotropy measurements
Fluorescence anisotropy data were collected in a PC-1 fluorimeter (ISS Instruments) coupled to a thermal bath at 25°C using an excitation wavelength of 280 nm (46). Samples of LbPrx1m at 80 M were preincubated in buffer at pH 7.5 containing 25 mM Tris-HCl, 150 mM NaCl, and 5 mM EDTA and increasing amounts of CaCl 2 or MgCl 2 . For data acquisition, samples were diluted in buffer consisting of 25 mM Tris-HCl and 150 mM NaCl (pH 7.5) to a final protein concentration of 2 M. Free cation concentrations were defined by subtracting the added Ca(Mg)Cl 2 concentration from the EDTA concentration of each sample. The mean data of three independent experiments were fitted to a nonlinear regression to estimate the constant K1 ⁄2 using GraphPad Prism v.6.0.

Molecular dynamics simulations
The most favorable geometric coordination of Ca 2ϩ by the residues Ser-109 (chain A) , Ser-112 (chain A) , Asp-76 (chain J) and Asp-108 (chain J) of the LbPrx1m decamer (PDB accession no. 4KB3) was evaluated using molecular dynamics simulations. The system was submitted to an explicit solvent simulation with a water density of 1 g cm Ϫ3 and neutralized using a 0.9% NaCl solvent (mass fraction) at 298 K. The protonation states for ionizable groups were set according to pH 7.0 using an empirical equation derived from experimental data that considers electrostatic potential, hydrogen bonds, and accessible surface area (47). The simulation was carried out for 10 ns using the YAMBER3 force field (48), which includes the cation parameters, on the program YASARA. The calcium coordination sphere for every 25-ps snapshot was analyzed using a customized script implemented in FindGeo (49).

Chaperone activity assays
To investigate the chaperone activity of reduced and oxidized LbPrx1m (WT or mutants), 100 nM luciferase (Promega) was incubated in 40 mM HEPES (pH 7.5) at 42°C with a molar ratio of 1:10 (Luciferase:LbPrx1m). LbPrx1m was pretreated with 20 mM CaCl 2 or MgCl 2 , 5 mM EDTA, or 5 mM EDTA plus 25 mM CaCl 2 or MgCl 2 . The reduced samples were incubated with 2 mM DTT (final concentration). The reactions were kept at 42°C, and luciferase aggregation was monitored in a Fluoromax-4 spectrofluorometer (Horiba) for 900 s using a wavelength of 360 nm for excitation and emission. Relative activities were calculated as mean values considering the activity of untreated samples as reference for those treated with additives (Ca 2ϩ , EDTA, and DTT) or the activity of WT protein as a reference for the muteins assayed in the same condition, unless stated otherwise. To exclude the effect that some additives have on luciferase aggregation, the relative activities were calculated according to the formula (N T Ϫ T)/(N R Ϫ R), where N T , T, N R , and R refer to light-scattering values recorded at 900 s of N T , the negative control reaction of the test condition (luciferase ϩ additives); T is the test condition (N T ϩ WT or mutant LbPrx1m); N R is the negative control reaction of the reference condition (luciferase ϩ additives); and R is the reference condition (N R ϩ WT LbPrx1m). All assays were performed with TEV-cleaved LbPrx1m in triplicate.

Circular dichroism
CD measurements were acquired at 25 or 42°C on a JASCO J-815 spectropolarimeter equipped with a Peltier temperature controller (Jasco Analytical Instruments). TEV-cleaved LbPrx1m samples (WT and H113A mutant) at 80 M were pretreated with 5 mM EDTA plus (or not) 25 mM CaCl 2 and diluted to a final concentration of 2 M in 10 mM sodium phosphate (pH 7.5) with or without 2 mM DTT. Far-UV CD spectra were recorded between 190 and 260 nm at a speed of 50 nm/min with a total of 16 accumulations. The CD data were buffer-subtracted and normalized to molar residual ellipticity allowing the comparison between different treatments.

Differential scanning fluorimetry
DSF assays were performed in triplicate using a real-time PCR machine 7300 (Applied Biosystems). Samples of TEVcleaved LbPrx1m (WT and H113A mutant) were pretreated as described above and diluted to a final concentration of 2 M in buffer consisting of 20 mM HEPES (pH 7.5), 150 mM NaCl, with or without 2 mM DTT, and containing 5ϫ SYPRO Orange fluorescent dye (Invitrogen-Molecular Probes). The 96-well plates were heated from 25 to 95°C, increasing 1°C/cycle, and the fluorescence emission was measured at 580 nm. The DSF melting curves were analyzed using GraphPad Prism software version 6.0.

Generation of L. infantum transfectants
To construct the pSSU-PHLEO-infantum-LbPrx1m plasmids, a DNA fragment corresponding to the mitochondrial targeting sequence of Prx1m was PCR-amplified with PfuTurbo from genomic DNA of Leishmania amazonensis with primers Regulatory mechanisms of mitochondrial 2-Cys peroxiredoxins P1 (5Ј-cgcggatccATGCTCCGTCGTCTTGCTA-3Ј) and P2 (5Ј-tgctctagagctagcaggcctGACAGTCGCCGTACGGTA-3Ј) and cloned into the BamHI and XbaI sites of pSSU-PHLEO-infantum-LiPrx1m vector (2). Clamp sequences are indicated in lower case, and restriction sites are in italics. The resulting plasmid was subsequently digested with StuI and NheI and ligated to the rest of the LbPrx1m ORF (either the WT or mutated versions of the gene) obtained by PCR amplification with Pfu-Turbo and primers P3 (5Ј-GATCCTGCGCCGCAGTTT-3Ј) and P4 (5Ј-ctagctagcTCACATATTCTTCTCAAAAAATT-3Ј) from the plasmids pET28a-His-TEV-LbPrx1m WT or mutants. The accuracy of all constructs was verified by DNA sequencing at GATC Biotech (Konstanz, Germany). Prior to transfection of L. infantum, the pSSU-PHLEO-infantum-LbPrx1m constructs were linearized by digestion with NdeI and PmeI and purified from agarose gels.

Transfection of L. infantum and isolation of mutants
Transfections were carried out on L. infantum promastigotes (MHOM MA67ITMAP263) missing both Prx1m alleles (i.e. Prx1m null mutants or Prx1m Ϫ ) produced previously (2). Parasites were grown to the logarithmic phase and electroporated at 450 V and 350 -400 F with 5 g of DNA as described elsewhere (51). Transfectants were allowed to recover in culture medium without selective drug for 24 h prior to being plated onto agar plates containing 17.5 g ml Ϫ1 bleomycin (Sigma-Aldrich). Upon 1 to 2 weeks of growth on agar, colonies were picked up, transferred to liquid medium, and analyzed by PCR, Western blotting, and indirect immunofluorescence to confirm LbPrx1m expression in the transfectants according to previously described procedures (2,52).

Thermotolerance assays
L. infantum promastigotes, synchronized previously by three to four daily changes of culture medium, were seeded at 10 6 cells ml Ϫ1 in 24-well plates containing RPMI 1640-GlutaMAX medium supplemented with 10% inactivated fetal bovine serum, 50 units ml Ϫ1 penicillin, 50 mg ml Ϫ1 streptomycin (all from Gibco), and 20 mM HEPES sodium salt (pH 7.4) (Sigma). Parasites were allowed to grow for 4 days at either 25 or 37°C. Every 24 h, cell densities were determined with a Neubauer counting chamber for growth curve determination. Two independent clones were analyzed for each transfectant.