The cytosolic hyperoxidation-sensitive and -robust Leishmania peroxiredoxins cPRX1 and cPRX2 are both dispensable for parasite infectivity

Abstract

Typical two-cysteine peroxiredoxins (2-Cys-PRXs) are H₂O₂-metabolizing enzymes whose activity relies on two cysteine residues. Protists of the family Trypanosomatidae invariably express one cytosolic 2-Cys-PRX (cPRX1). However, the Leishmaniinae sub-family features an additional isoform (cPRX2), almost identical to cPRX1, except for the lack of an elongated C-terminus with a Tyr-Phe (YF) motif. Previously, cytosolic PRXs were considered vital components of the trypanosomatid antioxidant machinery. Here, we shed new light on the properties, functions and relevance of cPRXs from the human pathogen Leishmania infantum. We show first that LicPRX1 is sensitive to inactivation by hyperoxidation, mirroring other YF-containing PRXs participating in redox signaling. Using genetic fusion constructs with roGFP2, we establish that LicPRX1 and LicPRX2 can act as sensors for H₂O₂ and oxidize protein thiols with implications for signal transduction. Third, we show that while disrupting the LicPRX-encoding genes increases susceptibility of L. infantum promastigotes to external H₂O₂in vitro, both enzymes are dispensable for the parasites to endure the macrophage respiratory burst, differentiate into amastigotes and initiate in vivo infections. This study introduces a novel perspective on the functions of trypanosomatid cPRXs, exposing their dual roles as both peroxidases and redox sensors. Furthermore, the discovery that Leishmania can adapt to the absence of both enzymes has significant implications for our understanding of Leishmania infections and their treatment. Importantly, it questions the conventional notion that the oxidative response of macrophages during phagocytosis is a major barrier to infection and the suitability of cPRXs as drug targets for leishmaniasis.

Highlights

• •Leishmania feature hyperoxidation-sensitive and -robust cytosolic peroxiredoxins.
• •Both peroxiredoxins can sense and transfer H₂O₂ signals to target protein thiols.
• •Leishmania peroxiredoxins are dispensable for parasite survival and infectivity.
• •H₂O₂-hypersensitive cPRXs knockouts resist the macrophage oxidative attack.

1. Introduction

Peroxiredoxins (PRXs) are a family of enzymes that, owing to their remarkable reactivity with H₂O₂ and high abundance in cells, act as first line sensors and recipients for this oxidizing agent [1,2]. Their peroxidase activity depends on a conserved peroxidatic cysteine residue (C_P), which directly reduces H₂O₂ and becomes oxidized to cysteine sulfenic acid (C_P–SOH), and a second ‘resolving’ cysteine (C_R) that can undergo a condensation reaction with C_P-SOH to form a disulfide bond [3]. Typical 2-Cys peroxiredoxins form head-to-tail dimers, where the C_P of each subunit can form a disulfide bond with the C_R of the adjacent subunit for a total of two disulfide bonds [4,5]. The resulting intermolecular C_P-C_R disulfide bonds can be reduced by an oxidoreductase of the thioredoxin family.

Beyond their involvement in thioredoxin-dependent H₂O₂ reduction, 2-Cys-PRXs can engage in alternative redox mechanisms with various implications for cell physiology. One such mechanism is oxidative inactivation, which occurs by hyperoxidation of C_P-SOH to sulfinic (C_P–SO₂H) or sulfonic (C_P–SO₃H) acid, above certain H₂O₂ threshold concentrations [6]. This shutdown mechanism frees H₂O₂ [5] and thioredoxin-like enzymes [7] from their dominant PRX interaction partners, allowing them to participate instead in other physiologically relevant phenomena [6,7]. Another important mechanism is the participation of PRXs in redox relays for redox signaling [8,9]. In this case, PRXs sense H₂O₂ signals, and relay oxidizing equivalents, stored in the form of C_P-SOH or C_P-C_R, to redox-regulated proteins. Both hyperoxidation and thiol-condensation mechanisms place PRXs as key proteins in redox signaling.

Two-Cys PRXs were the first peroxidases identified in trypanosomatids, a protist family that comprises the human parasites Leishmania, Trypanosoma brucei and Trypanosoma cruzi. Present in the cytosol, mitochondria and, in the case of Leishmania, in glycosomes, the discovery of these enzymes in 1997 [10] provided a long-sought explanation as to how trypanosomatids can efficiently remove peroxides in the absence of catalases and classic selenocysteine-containing glutathione peroxidases. The subsequent biochemical characterization of trypanosomatid PRXs established that these enzymes had an exquisite preference for using tryparedoxin (TXN), a unique thioredoxin-like enzyme restricted to these organisms, as a reducing agent. For this reason, trypanosomatid PRXs were labeled as “tryparedoxin peroxidases” (TXNPx). Upstream of the TXN enzymes lies trypanothione (TS₂), a small-molecule dithiol exclusive to the trypanosomatids, and its NADPH-dependent reducing enzyme trypanothione reductase (TR) [10]. The distinctive TR-TS₂-TXN redox cascade is absolutely central for redox balance in trypanosomatids, as demonstrated by the essentiality of each of the components [11,12]. It is now known that, in addition to PRXs, trypanosomatids can express other peroxidases, non-selenium glutathione peroxidases (nsGPXA and nsGPXB), also dependent on the trypanothione system [13,14], and heme peroxidases (cytochrome c peroxidase [15], pseudoperoxidase [16] and a third uncharacterized enzyme), the latter absent from T. brucei. None of the trypanosomatids expresses 1-Cys PRXs.

Trypanosomatid PRXs have always been viewed as major components of the antioxidant machinery of parasites, with hardly any studies exploring the possibility of these enzymes displaying alternative functions. One exception to this is the mitochondrial 2-Cys PRX of Leishmania infantum, whose genetic silencing and editing uncovered its crucial peroxidase-independent function as a molecular chaperone [17,18]. Previous functional studies on cytosolic PRXs (cPRXs) focused on phenotypic analyses of parasites ectopically overexpressing these enzymes. These studies emphasized the potential of cPRXs to protect parasites from oxidative challenges, as well as to boost infectivity in the context of in vitro and in vivo infection models [[19], [20], [21], [22], [23], [24]]. Along the same line, cPRXs were found to be upregulated in highly virulent [25,26] and highly metastatic [27] Leishmania strains, as well as in parasites resistant to microbicidal drugs [21,24,[28], [29], [30], [31], [32]].

The apparent correlation between cPRX upregulation and increased parasite virulence and drug resistance has led to the general acceptance of these enzymes as critical determinants of trypanosomatid survival and virulence [21,24,26,33]. Indirect evidence supporting this conclusion includes the essentiality of the TR, TS₂, and cytosolic TXN components that support cPRX activity [34]. Importantly however, robust reverse genetic analyses to prove the essentiality of cPRXs are lacking. The only study performed to date that go in this direction is an RNA interference-based trial carried out in T. brucei in which down regulation of cPRX was found to impair the growth and tolerance of bloodstream parasites to H₂O₂ [35]. The recently reported elimination of cPRXs in Leishmania mexicana [36], could not have been achieved because the strategy followed did not target all cPRX open reading frames annotated in the L. mexicana genome [37].

In summary, the conclusion that trypanosomatid cPRXs are essential components of the antioxidative machinery, with a function restricted to peroxide detoxification, has not been rigorously experimentally demonstrated. Recent developments, for example in the direction of genetic validation tools in trypanosomatids call for new and more robust proofs of cPRXs essentiality. With this in mind, the present manuscript revisits the functional relevance of cPRXs in the human pathogen L. infantum [19,38,39]. While biochemical properties of L. infantum cPRXs support their potential participation in redox signaling, homologous recombination and CRISPR-Cas9 gene knockout strategies uncover that cPRXs are dispensable for L. infantum promastigotes and their differentiation into infective amastigotes.

2. Material and methods

2.1. Animals and ethics

Mice were purchased from the i3S animal facility and handled in strict accordance with good animal practice as defined by national authorities (directive 113/2013 from 7th August) and European legislation (directive 2010/63/EU, revising directive 86/609/EEC). Animal procedures were approved by the Local Animal Ethics Committee of i3S, licensed by Direção Geral de Alimentação e Veterinária (DGAV), Govt. of Portugal.

2.2. Oligonucleotides and DNA sequencing

Primers (Table S1) were purchased from Sigma-Aldrich. The accuracy of DNA constructs was verified by sequencing at Macrogen Europe (The Netherlands) and Eurofins Genomics (Germany).

2.3. Parasite culture and transfection

Leishmania infantum promastigostes (MHOM MA67ITMAP263) were cultured as described before [17]. For transfection, 5×10⁷ promastigotes in the logarithmic phase of growth (2 days) were suspended in Tb-BSF buffer [40] in the presence of 2–10 μg DNA, electroporated using the program X-001 of the Amaxa Nucleofector^TM (Lonza), and selected according to Sousa et al. [41].

2.4. Indirect immunofluorescence analysis (IFAT) of LicPRX1 and LicPRX2

For detection of c-myc-LicPRX1 and c-myc-LicPRX2, the coding sequences of each gene were PCR-amplified with primers P8/P9 and P8/P10, and cloned into the EcoRV/XhoI sites of pTEX-NEO^R-c-myc [42] to guide the expression of N-terminally c-myc tagged cPRX chimeras. Promastigotes transfected with the resulting plasmids were processed for IFAT as described before [17], using anti-c-myc (Santa Cruz Biotechnology; 1:100 dilution) and Alexa Fluor 488 anti-mouse IgG (Molecular Probes; 1:2000 dilution). IFAT diagnosis of cPRX knockouts was carried with an antiserum raised against purified recombinant LicPRX2 (Davids Biotechnologie GmbH; 1:1000 dilution), and Alexa Fluor 488 anti-rabbit IgG (Molecular Probes; 1:2000 dilution). Slides were examined with an AxioImager Z1 microscope and photographed with an Axiocam MR 3.0 using the Axiovision 4.6 software (all from Carl Zeiss).

2.5. Phylogenetic analysis

The evolutionary history of Trypanosomatidae PRXs was inferred in MEGA X [43], based on 22 nucleotide sequences (Table S2). The bootstrap consensus tree inferred from 500 replicates was assumed to represent the evolutionary history of the taxa analyzed. Branches corresponding to partitions reproduced in less than 50% bootstrap replicates were collapsed. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test are shown next to the branches. The evolutionary distances were computed using the Maximum Composite Likelihood method and are in the units of the number of base substitutions per site.

2.6. In vitro peroxidase assays

6His.LicPRX1, 6His.LicPRX2, 6His.LiTXN1, and LiTR.6His were produced as recombinant proteins in Escherichia coli according to established protocols [39]. N-terminal histidine tags of purified 6His.LicPRX1 and 6His.LicPRX2 were cleaved by thrombin, as described earlier [17]. Proteins were quantified by the bicinchoninic acid protein assay (Pierce), using bovine serum albumin as standard. Peroxidase activity assays were performed in cuvettes containing 300 μl-reaction mixtures of 100 μM NADPH (Sigma), 0.5 U ml⁻¹ LiTR.6His, 20 μM trypanothione (Bachem), 4 μM 6His.LiTXN1, and either 0.75 μM LicPRX1 or 0.1 μM LicPRX2, in 50 mM Tris-HCl, 1 mM EDTA, pH 8.0 [39]. Negative control reactions contained no cPRX enzyme. After 10 min pre-incubation, 50–125 μM H₂O₂ was added to the cuvettes and the NADPH consumption was monitored at 340 nm in a Shimadzu UV-2401 PC spectrophotometer (Shimadzu Corporation), equipped with temperature control (25 °C).

2.7. Fluorescence measurement of roGFP2-LicPRX1/2 sensors in S. cerevisiae

Codon-optimized LicPRX1 and LicPRX2 (GenScript Biotech, Netherlands) were cloned into the EcoRI/HindIII restriction sites of p416TEF-roGFP2 [44] to yield p416TEF-roGFP2-LicPRX1 and p416TEF-roGFP2-LicPRX2. The remaining mutein-carrying plasmids were generated from wild type constructs following standard site-directed mutagenesis protocols. Constructs were transferred to BY4742 Δtsa1Δtsa2 S. cerevisiae strain [45]. Yeast genetically encoding roGFP2-LicPRX1/2 probes were employed in roGFP2 measurements, as described before [46]. Briefly, yeast cultures were grown to late exponential phase (OD_600nm = 3–4), suspended in buffer 100 mM MES-Tris pH 6.0 to a final concentration of 7.5 OD_600nm units ml⁻¹. Aliquots of 180 μl of the yeast suspension were transferred to a flat-bottomed 96-well imaging plate (BD Falcon 353219), challenged with either 20 mM diamide (fully oxidized control), 100 mM DTT (fully reduced control) or H₂O₂ (0–1000 μM), and monitored for 90 min at 30 °C in CLARIOstar fluorescence plate reader (BMG Labtech). Oxidized and reduced roGFP2 were excited at 400 nm and 480 nm, respectively, and emission of both redox states measured at 520 nm. At each time point, the degree of roGFP2 oxidation (OxD) was calculated as before [46].

2.8. DNA constructs for homologous recombination

The construct to target LicPRX1 was built from a pTEX-HYG^R-based plasmid previously used to disrupt LimTXNPx [17], by sequential cloning of the LicPRX1 5’UTR (primers P11/P12) and 3’UTR (primers P13/P14) into BamHI/EcoRV and KpnI sites, respectively. The DNA cassette to target LicPRX2 was assembled from the pGL-HYG^R-(5′UTRLicPRX1)-LiDRP-(3′UTRLigPRX) [47], by sequentially cloning the LicPRX2 5’UTR (primers P15/P16) and the LicPRX2 3’UTR (primers P17/P18) into HindIII/XmaI and BamHI/BglII restriction sites, respectively. Prior to transfection, integration cassettes were separated from plasmid backbones by digestion with BamHI (LicPRX1 cassette) or HindIII/BglII (LicPRX2 cassette).

2.9. Genetic silencing of LicPRX1 and LicPRX2 by CRISPR-Cas9

Promastigotes of L. infantum were engineered to i) express humanized Streptococcus pyogenes Cas9 nuclease gene and T7 RNA polymerase from the pT007-Cas9-T7-Tub plasmid [48], and ii) to integrate LiDRP in the ribosomal locus from the pSSU-PHLEO^R-L. infantum-LiDRP construct. The latter was built by excising the LiDRP ORF and its flanking 5’UTR and 3’UTR from the LicPRX2 disruption construct (see above) by sequential treatment with HindIII, Klenow and SpeI, and subsequent cloning into the XhoI/Klenow and SpeI sites of pSSU-PHLEO^R-L. infantum-mTXNPx [17]; this construct was linearized with NdeI/PmeI-digestion prior to transfection of parasites. The resulting Cas9/DRP transgenics were transformed with a mixture of synthetic sgRNAs [amplified with primer P19 (sgRNA scaffold) and a target-specific primer (P20 or P21)], and BSD^R and PURO^R donor fragments (amplified from pTBlast or pTPuro plasmids [48], with primers P22/P23).

2.10. Complementation plasmid

The pXG-NEO-LicPRX1 plasmid was assembled by cloning the PCR-amplified (primers P24/P25), and SmaI-, BglII-digested LicPRX1 ORF into the SmaI and BamHI sites of pXG-NEO [49] and used to rescue mutant phenotypes.

2.11. Western blotting

Proteins from promastigote extracts were separated by SDS-PAGE and transferred to nitrocellulose membranes that were subsequently decorated with polyclonal antibodies raised against purified recombinant LicPRX2 [39] in rabbits (Davids Biotechnologie GmbH), or LiTRYS as loading control [41] (both at 1:1000 dilution). The secondary antibody was an anti-rabbit IgG HRP conjugate (Invitrogen; 1:10000 dilution).

2.12. Determination of IC₅₀ values against H₂O₂ and SIN-1

Promastigotes were seeded at 3.75×10⁶ cells ml⁻¹ in 96-well plates containing increasing concentrations (in duplicate) of either H₂O₂ (3–800 μM; Sigma) or 3-morpholinosydnonimine hydrochloride (SIN-1, 32.5–4000 μM; Santa Cruz Biotechnology, Inc-EUR). Upon 24 h incubation, 125 μM resazurin (Sigma) was added to cultures, and parasite viability measured in a Synergy 2 (BioTek) microplate reader at λ_exc = 540 nm and λ_em = 620 nm. The effect of the vehicle (DMSO) was considered for the determination of cell viability (only for SIN-1). Half-maximal inhibitory concentration (IC₅₀) values were determined in GraphPad Prism v8.0 and using the nonlinear fit log (inhibitor) vs. response (variable slope, four parameters).

2.13. Infection of bone marrow derived macrophages (BMDM)

BMDM, differentiated from monocytes of C57BL/6 mice following previously described protocols [50], were co-incubated with stationary phase promastigotes freshly recovered from spleens of infected mice at a macrophage:parasite ratio of 1:10. Three hours later, non-phagocytosed promastigotes were washed away and the time of the experiment set to zero. At defined points post infection, cultures were processed for fluorescence microscopy and followed up for determination of percentage of infection and number of parasites per infected macrophages, as described elsewhere [51].

2.14. In vivo infection of murines

Stationary phase promastigotes freshly recovered from spleens of infected mice were inoculated intravenously in 6–8 weeks old C57BL/6 mice (2×10⁷ parasites/mouse). At defined days post infection, mice were sacrificed and their livers and spleens processed for determination of the number of parasites per gram of organ (parasite burden) [17]. The detection limit of this method is 500 parasites per gram of organ (i.e. 2.7 log units).

2.15. Statistical analysis

Experimental groups were compared by the one-way analysis of variance (ANOVA), followed by pairwise multiple comparisons using Tukey’s post hoc test, or two-tailed unpaired Student’s t-test, using GraphPad Prism v8.0.

3. Results

3.1. Leishmania harbors two cPRX isoforms, cPRX1 and cPRX2

The genomes of Leishmania encode two cytosolic 2-Cys PRX enzymes: cPRX1 and cPRX2 [38,39]. Both proteins are assigned to the Prx1/AhpC subgroup of PRXs [52] and preserve amino acid residues critically involved in peroxidase activity (namely, the redox-active C_P and C_R residues), as well as in activation of the C_P (Pro and Thr/Ser residues located contiguous to C_P and a distant Arg) [2,53,54], and in interaction with the TXN reducing partner [55,56] (Fig. 1A). Leishmania cPRX1 and cPRX2 are highly similar (e.g., 86.5% identity between the Leishmania infantum enzymes), the most remarkable exception being their C-termini. In cPRX1, this part is extended by a 9-residue alpha-helix harboring a Tyr-Phe (YF) motif (Fig. 1A), which in other typical 2-Cys PRXs confers susceptibility to hyperoxidation [57]. Another noteworthy sequence divergence is found at position 49, which in cPRX2 (as in a low percent of 2-Cys PRXs [54]) is occupied by a Ser, and in cPRX1 by a Thr residue. An identical Thr-to-Ser amino acid substitution artificially imposed on the L. donovani cPRX1 enzyme was reported to potentiate its peroxidase activity [55]. Finally, cPRX1 and cPRX2 peptide sequences lack canonical organelle targeting signals in agreement with their localization to the cytosol of Leishmania (Fig. 1B) [19,22].

Fig. 1.

The Leishmania repertoire of cytosolic PRXs is composed of two potentially distinct enzymes, cPRX1 and cPRX2. (A) Sequence alignment of the predicted amino acid sequence of L. infantum cPRX1 (LicPRX1) and cPRX2 (LicPRX2). Strict identity across all sequences (including the active site C_P and C_R residues) is shown shaded in black; equivalent residues are typed in bold; red triangles represent residues implicated in activation of C_P [2,4,54,55]; red circles highlight residues putatively interacting with TXN [55,56]; the orange bar signals the LicPRX1 C-terminal extension; orange boxes signal the GGLG and YF motifs. (B) Representative fluorescence micrographs of L. infantum promastigotes ectopically expressing c-myc-LicPRX1 or c-myc-LicPRX2, showing the merged channels of the anti-c-myc antibody (green) and DAPI (blue). Scale bar: 2.5 μm. (C) Phylogenetic analysis of cPRX genes of trypanosomatids. Neighbor-Joining tree with bootstrapped values (500 replicates) of cPRX-coding sequences across five subfamilies of the family Trypanosomatidae. The percentage of replicate trees in which the associated taxa clustered together in the bootstrap test are shown next to the branches. Orange and blue circles signal YF-containing cPRX1 and YF-devoid cPRX2 isozymes, respectively. The asterisk and dotted lines in LpcPRX2 indicate that, due to incomplete sequence information of genome databases, we cannot ascertain the presence of this molecule in Leptomonas pyrrhocoris. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

To gain insight into the distribution of cPRX1 and cPRX2 isozymes across Trypanosomatidae, we carried out genome mining of 11 species (some of which encompass two strains), covering five (out of six) subfamilies. This delivered 22 cPRX-coding sequences, which we subjected to in silico phylogenetic analysis, using the free-living eubodonid Bodo saltans as an out-group. The resulting topology tree (Fig. 1C) shows that cPRX1 isoforms are conserved along the Trypanosomatidae lineage, whereas cPRX2 is found exclusively in the Leishmaniinae sub-family. Retention of cPRX2-coding genes is a generalized feature of all Leishmania species with complete sequence annotation of the cPRX loci (L. aethiopica, L. braziliensis, L. donovani, L. enriettii, L. mexicana, L. panamensis, and L. tarentolae). Interestingly, Leishmaniinae are also unique in harboring a glycosomal PRX (gPRX) isoform [47], whose coding sequence falls within the same genetic locus as cPRX1 and cPRX2. Based on the highly repetitive character of cPRX loci, e.g., multi-copy genes and stretches of identical untranslated regions, it is conceivable that the co-evolutionary acquisition of cPRX2 and gPRX genes in Leishmaniinae resulted from gene duplication events that took place after the divergence of this phylogenetic arm.

3.2. LicPRX1, but not LicPRX2, is prone to inactivation by hyperoxidation

Based on its primary structure, LicPRX1 is predicted to be susceptible to hyperoxidation. In this enzyme, the C-terminal YF motif together with the conserved Gly-Gly-Leu-Gly (GGLG) domain (Fig. 1A)] [57], is expected to impose a structural hindrance that delays C_P-SOH from reaching and condensing with the C_R residue of the adjacent PRX subunit. This delay, known as a “kinetic pause”, increases the chance of C_P-SOH to react with additional hydroperoxide molecules before it can form a disulfide bond, thereby becoming hyperoxidized. However, the presence of the YF motif does not per se prove that PRXs are hyperoxidation prone [58]. Therefore, to test LicPRX1 sensitivity to oxidative inactivation, we resorted to in vitro enzymatic assays, in which we indirectly assessed peroxidase activity by monitoring the kinetic of NADPH consumption in coupled enzymatic assays containing the NADPH-LiTR-TS₂-LiTXN1-LicPRX1-H₂O₂ redox cascade (Fig. 2A). In parallel, we also assayed LicPRX2. The results of these assays support the primary structure-based prediction that LicPRX1, but not LicPRX2, undergoes hyperoxidation, as indicated by the shape of NADPH consumption curves obtained for each enzyme. In the case of LicPRX1, NADPH oxidation kinetics resulted in curved lines, reflecting the progressive loss of enzyme activity (Fig. 2B, left panel). On the contrary, for LicPRX2, NADPH consumption kinetics were linear for all H₂O₂ concentrations tested (50–125 μM), consistent with a stable enzymatic activity along the time course of the reaction (Fig. 2B, right panel). The curving effect denoted for LicPRX1 is more evident at higher concentrations of H₂O₂ and reflects the time-dependent accumulation of inactive PRX forms, likely resulting from irreversible hyperoxidation of the C_P residue [6]. Under these assay conditions, we could estimate a C_hyp1% of 44 μM for LicPRX1, meaning that at 44 μM H₂O₂, one in 100 PRX molecules is inactivated per catalytic cycle [59]. Compared with other peroxiredoxins, for example human PRX3, PRX1 and PRX2, which feature C_hyp1% of 127, 50 and 5 μM respectively [60], LicPRX1 exhibits an intermediate susceptibility to hyperoxidation. The kinetic constants, determined using the Dalziel algorithm [61], support the argument that cPRX1 is prone to hyperoxidation. Specifically, LicPRX1 exhibits a non-classical pattern with coefficient φ _1,2 = 69.2 s μM, while LicPRX2 follows a classical ping-pong pattern with a null coefficient φ_1,2 (Fig. S1 and Table 1).

Fig. 2.

LicPRX1, but not LicPRX2, is prone to hyperoxidation in vitro. (A) The PRX-dependent pathway for H₂O₂ reduction in trypanosomatids. TR, trypanothione reductase; TS₂, trypanothione; TXN, tryparedoxin; red, reduced; ox, oxidized. (B) Assessment of LicPRX1/2 peroxidase activity in in vitro test systems containing 100 μM NADPH, 0.5 U ml⁻¹LiTR.6His, 20 μM TS₂, 4 μM 6His.LiTXN1, and either 0.75 μM LicPRX1 (left panel) or 0.1 μM LicPRX2 (right panel). Reactions were started with bolus addition of H₂O₂ at different concentrations (50–125 μM), and NAPDH consumption monitored spectrophotometrically at OD_{340 nm} along time. Dashed lines [(−) cPRX] refer to the negative control reaction (i.e., without LicPRX1/2), carried out in the presence of 125 μM H₂O₂. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Table 1.

Steady-state kinetic constants of LicPRX1 and LicPRX2 peroxidase activity.

Enzyme	Φ0 (s)	Φ1 (s.μM)	Φ2 (s.μM)	Φ1,2 (s μM)	Kcat (s−1)	KMLiTXN1 (μM)	KM H2O2 (μM)
LicPRX1	0.22	∼0	15.2	69.2	4.5	39.0	∼0
LicPRX2	0.26	∼0	4.38	0	3.8	16.6	∼0

The observation that LicPRX1 is hyperoxidation-sensitive raises questions about its capacity to deal with acute oxidative challenges and might rather support a role in redox signal transduction events in Leishmania. On the contrary, LicPRX2, by resisting oxidative inactivation, could be better suited for classic antioxidative defense, i.e. H₂O₂ reduction.

3.3. LicPRX1 and LicPRX2 relay oxidizing potential to protein thiols employing different mechanisms

To gain insight into the potential role of trypanosomatid cPRXs as H₂O₂ sensors in disulfide-dependent redox signaling cascades [62], we constructed genetic fusions between redox-sensitive green fluorescent protein (roGFP2) and either LicPRX1 or LicPRX2. These redox probes allow the non-invasive and real-time monitoring of roGFP2 oxidation by H₂O₂-scavenger proteins to which it is fused [46].

We started by cloning codon-optimized roGFP2, roGFP2-LicPRX1, or roGFP2-LicPRX2 (as well as cPRX variants in which C_P and/or C_R were mutated to Ser) into the p416TEF yeast expression vector. The constructs were then transformed into the Saccharomyces cerevisiae strain BY4742 Δtsa1Δtsa2 strain, which has previously been developed and used for the monitoring of PRX activity in the context of roGFP2 fusion constructs [45]. The resulting transformants were subsequently used for monitoring the dynamic change in the degree of roGFP2 oxidation (OxD _roGFP2) in response to treatment with exogenous H₂O₂ at concentrations from 50 to 1000 μM.

These assays confirmed that both LicPRX1 and LicPRX2 are capable of sensing and transferring H₂O₂ signals to roGFP2. This is shown by the higher and more sustained OxD_roGFP2 of the fused probes relative to the unfused roGFP2 control (Fig. 3A, roGFP2-LicPRX1/2 vs. roGFP2). The OxD_roGFP2 is also significantly higher in active roGFP2-LicPRX1/2 fusion constructs, compared to inactive constructs in which sensing of H₂O₂ by the PRX is blocked either by a C_P-to-Ser mutation (Fig. 3A, roGFP2-LicPRX1/2ΔC_P) or by elimination of both catalytic cysteine residues (Fig. 3A, roGFP2-LicPRX2ΔC_PΔC_R).

Fig. 3.

LicPRX1 and LicPRX2 relay oxidative potential from H₂O₂to protein thiols, albeit via distinct mechanisms. (A) Time-course measurements of the ratiometric degree of oxidation (OxD) for BY4742 Δtsa1Δtsa2 yeast cells with genetically encoded roGFP2-LicPRX1/2 constructs, after different initial pulses of H₂O₂ (0–1000 μM). Data represent means and standard deviations of three independent biological replicates. (B) Proposed pathways for H₂O₂-triggered redox reactions of roGFP2-LicPRX1 and roGFP2-LicPRX2 probes. The LicPRX1 moiety is depicted in orange, LicPRX2 in blue, and roGFP2 in dark (reduced) and light (oxidized) grey. In steps common to both roGPF2-LicPRX1 and -LicPRX2 probes, the PRX moiety is shown in violet. In LicPRX1/2, the sulfhydryl groups of peroxidatic and resolving cysteines, are depicted as S_PH and S_RH, respectively. Curved, red arrows represent the attack of the electron donor. Also depicted are the excitation and emission wavelengths of reduced (left edge) and oxidized roGFP2 probe (right edge). From left to right, the scheme reads as follows. First, S_PH reacts with H₂O₂ to form a sulfenic acid (S_POH). In LicPRX2, S_POH readily condenses directly the S_RH in the PRX dimer (lower branch). The resulting S_P-S_R disulfide is then transferred to a target protein (in this case roGFP2; in physiological conditions, usually, a thioredoxin-type enzyme) via thiol-disulfide exchange. In LicPRX1, the kinetic pause (dotted lines) imposed by the YF C-terminal motif delays the reaction between S_POH and S_RH, thus freeing S_POH to react with a second and third H₂O₂ molecules, and to become hyperoxidized to sulfinic (S_PO₂H) and sulfonic (S_PO₃H) acid forms, respectively (upper branch). In yeast, this sulfinic acid can be retroreduced by sulfiredoxin (SRX), following a reaction with low catalytic rate and turnover (dotted lines) [78]. Alternatively, LicPRX1 kinetic pause can prompt S_POH to react with a thiol on a neighboring protein, in this case roGFP2 (middle branch). Finally, LicPRX1/2-roGFP2 mixed disulfides rearrange to form the roGFP2 disulfide, which is recycled to its reduced state by yeast glutaredoxin (GRX) enzymes [45]. (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Notably, LicPRX1 and LicPRX2 display different response kinetics, as indicated by the different shapes of the OxD_roGFP2 curves recorded for each roGFP2 fusion probe. For roGFP2-LicPRX2, OxD_roGFP2 increases sharply after peroxide addition, peaking to transient plateaus, which after ∼1000 s start to decline (Fig. 3A, roGFP2-LicPRX2). In comparison, the peaks of roGFP2-LicPRX1 oxidation are lower (e.g., maximum OxD_{roGFP2-LicPRX1} = 0.23 ± 0.02 versus OxD_{roGFP2-LicPRX2} = 0.31 ± 0.03, at 1000 μM H₂O₂), and decline much faster (Fig. 3A, roGFP2-LicPRX1). This faster decline might be an indication for the progressive accumulation of hyperoxidized LicPRX1, in contrast to LicPRX2, along the time course of the reactions. A cumulative (rather than instantaneous) inactivation of the roGFP2-LicPRX1 probe conforms with our in vitro observation that LicPRX2 is resistant and that LicPRX1 is only mildly sensitive to oxidative inactivation. In contrast, highly hyperoxidation sensitive roGFP2-PRX fusions usually respond by decreasing maximal OxD values despite increasing H₂O₂ concentration [44,45].

Finally, these assays suggest that LicPRX1 and LicPRX2 might employ distinct mechanisms to transfer oxidative equivalents from H₂O₂ to target protein thiols. According to the mechanistic model in Fig. 3B, the kinetic pause imposed by the C-terminal YF motif of LicPRX1 not only favors hyperoxidation of C_P-SOH but could also favor C_P-SOH condensation with Cys residues of other protein thiols [63], in this case roGFP2. A different redox relay mechanism might take place in the LicPRX2-roGFP2 probe, wherein C_P-SOH, free of structural constraints, readily condenses with C_R. The redox relay to roGFP2 then presumably occurs via a thiol-disulfide exchange reaction with the C_P-C_R mixed disulfide of LicPRX2 being transferred to roGFP2 via a transient LicPRX2/roGFP2 mixed disulfide intermediate. Consistent with this model, removing the C_R residue from LicPRX2 negatively affects the redox relay activity of this enzyme (Fig. 3A, roGFP2-LicPRX2ΔC_R), making the dynamics of roGFP2-LicPRX2 oxidation more similar to those of roGFP2-LicPRX1ΔC_R.

Overall, these data show that LicPRX1 and LicPRX2 have the potential to participate in signal transduction pathways by sensing H₂O₂ and relaying its oxidizing equivalents to protein thiols. Importantly, the data also suggest that the kinetic pause imposed by the C-terminal YF motif of LicPRX1 could impact on the mechanism by which this enzyme relays H₂O₂ signals, i.e., via C_P-SOH. This might not be the case of the YF-devoid LicPRX2 enzyme, whose redox relay presumably occurs by means of the C_P-C_R mixed disulfide.

3.4. L. infantum tolerates the selective deletion of either LicPRX1 or LicPRX2

To gain insight into the functional relevance of LicPRX1 and LicPRX2 for L. infantum parasites, we devised a homologous recombination (HR)-based strategy to delete the encoding genes. LicPRX1-and LicPRX2-coding genes share the same diploid locus at chromosome 15 (Fig. 4A). This locus comprises two additional open reading frames (ORFs): i) LigPRX, encoding the glycosomal PRX enzyme, which is dispensable for L. infantum survival and infectivity [47], and ii) LiDRP, a double copy gene that intercalates LicPRX1, LicPRX2 and LigPRX, and that encodes a putative developmentally regulated protein (LiDRP) of unknown function.

Fig. 4.

Generation and phenotypic analysis of Δcprx1 and Δcprx2 L. infantum promastigotes. (A) Schematic representation of the diploid cPRX1/cPRX2/gPRX locus at chromosome 15 of L. infantum, showing LicPRX1 (tritrypdb ID: LINF_150019000), LicPRX2 (LINF_150018800), and LigPRX (LINF_150018600) coding sequences, as well as the intercalating double copy LiDRP gene (LINF_150018700, LINF_150018900). Generation of Δcprx1 and Δcprx2 mutants followed two steps of homologous recombination (HR1 and HR2). HR1 consisted in the generation of a heterozygous PRX knockout line (SKO) by replacement of one of the cPRX1/cPRX2/gPRX alleles with a LiDRP/BSD^R replacement cassette, as described before [47]. In HR2, SKO promastigotes were independently transfected with LiDRP/HYG^R disruption constructs aimed at selectively targeting LicPRX1 and LicPRX2 to yield Δcprx1 and Δcprx2 mutants. Location of primers (P1–P6) employed in PCR analysis of the resulting transformants (in B) is also depicted. (B) PCR, WB and IFAT diagnoses of Δcprx1 and Δcprx2, as well as WT and SKO controls. The anti-LicPRX2 antibody indiscriminately recognizes LicPRX1, LicPRX2, and LigPRX. (C) Proliferation of Δcprx1 and Δcprx2, and of control WT and SKO L. infantum promastigotes. Parasites were seeded at 5×10⁵ cells ml⁻¹ (day 0) in complete RPMI medium and counted daily throughout six days of culture. Results represent means and standard deviation of three independent growth curves. (D) Relative normalized IC₅₀ values of H₂O₂ (left panel) and SIN-1 (right panel) estimated for WT, SKO, Δcprx1 and Δcprx2 promastigotes, 24 h after exposure to the compounds. Columns represent means and standard deviations of three (H₂O₂) and four (SIN-1) independent assays. For each assay, data were normalized relative to WT to minimize to inter-experimental variation. Statistically significant differences (two-way ANOVA, Tukey’s): *, Δcprx1 are different from WT (p < 0.001), SKO (p < 0.005), and Δcprx2 (p < 0.05); **, Δcprx2 are different from WT (p < 0.0001), SKO (p < 0.001), and Δcprx1 (p < 0.05); #, Δcprx1 and Δcprx2 are different from WT (p < 0.05). (E) Infection parameters of murine bone marrow derived macrophages (MO) parasitized with WT, SKO, Δcprx1 and Δcprx2 at different time points after infection. Percentage of infected macrophages (left panel) and the average number of amastigotes per infected macrophage (right panel) are shown. Values represent mean and standard deviations of three independent experiments, each performed in triplicate. (F) Parasite burden (PB) in livers (left panel) and spleens (right panel) of C57BL/6 mice, 17 days after intravenous injection with WT, SKO, Δcprx1 and Δcprx2. PB was determined by limiting dilution as the number of parasites per gram of organ and is represented in log₁₀. The detection limit of this technique is 2.7 log units. Each circle represents one mouse, and horizontal and vertical lines represent averages and standard deviations, respectively. Statistically significant differences (two-way ANOVA, Tukey’s): *, Δcprx1 are different from WT (p < 0.005), and SKO (p < 0.005); **, Δcprx2 are different from WT (p < 0.05); #, Δcprx1 are different from SKO (p < 0.05); ##, Δcprx2 are different from WT (p < 0.001), and SKO (p < 0.0001).

As a starting point to individually target LicPRX1 and LicPRX2, we made use of L. infantum cprx1^+/− cprx2^+/− gprx ^± heterozygous mutants (single knockouts, SKO), previously generated in our lab by replacement of one of the cPRX1/cPRX2/gPRX alleles with a disruption cassette carrying the BSD^R ORF and one copy of LiDRP (Fig. 4A) [47]. To specifically eliminate LicPRX1 and LicPRX2, we next transformed SKO promastigotes with DNA constructs harboring the HYG^R gene flanked by ∼750 nt sequences homologous to the unique intergenic regions of each cPRX ORF, as well as one copy of LiDRP to ensure the preservation of this gene upon HR events. Following this approach, we isolated HYG-resistant parasites which, as confirmed by PCR (Fig. 4B), are homozygous knockouts for either LicPRX1 (Δcprx1) or LicPRX2 (Δcprx2). Western blot and indirect immunofluorescence analyses of the mutants, using an anti-LicPRX2 antiserum that indiscriminately recognizes LicPRX1, LicPRX2 and LigPRX, confirmed a drop in the total amount of these enzymes in Δcprx1 and Δcprx2 promastigotes relative to their parental SKO strain. This decrease was more evident in Δcprx1 than Δcprx2, suggesting that the cytosolic concentration of LicPRX1 in cultured promastigotes is higher than the concentration of LicPRX2.

Phenotypic analysis of Δcprx1 and Δcprx2 promastigotes revealed that these mutants are morphologically normal and replicate as control wild type (WT) and SKO parasites (Fig. 4C). When assessed for their ability to cope with oxidative challenges, Δcprx1 and Δcprx2 were found to be more susceptible to exogenously added H₂O₂ relative to the control strains (Fig. 4D, left panel). Noticeably, Δcprx2 promastigotes are significantly more sensitive to H₂O₂ than Δcprx1 (p < 0.05), suggesting that LicPRX2, albeit apparently less abundant, contributes more to parasite antioxidant protection at high H₂O₂ concentrations than LicPRX1, as expected from LicPRX2 resistance to oxidative inactivation (Fig. 2B). Apart from reducing hydroperoxides, PRXs can also display peroxynitrite (ONOO⁻) reductase activity [64]. Aligned with this, we observed that selective depletion of either LicPRX1 or LicPRX2 impairs promastigote resistance to the indirect ONOO⁻ donor 3-morpholinosydnonimine (SIN-1, [65]) (Fig. 4D, right panel).

Next, we investigated the capacity of Δcprx1 and Δcprx2 strains to infect macrophages. Resorting to in vitro infections of macrophages derived from the bone marrow of C57BL/6 mice, we found that removal of LicPRX1 or LicPRX2 did not affect the capacity of parasites to penetrate macrophages (Fig. 4E), despite the oxidative burst that is triggered by these host cells at initial stages of infection (Fig. S2). On the following days post infection, we witnessed a sharp decrease in the percentage of infected macrophages for all parasite lines, with no obvious advantage of WT and SKO relative to Δcprx1 and Δcprx2 strains (Fig. 4E).

Finally, we assessed the impact of LicPRX1 or LicPRX2 removal on L. infantum virulence in an in vivo murine model of infection. These trials revealed that both Δcprx1 and Δcprx2 mutants are capable of infecting and thriving in mammalian hosts, albeit with some disadvantages relative to WT and SKO controls (Fig. 4F). This is evidenced by the lower parasite burdens yielded by Δcprx1 and Δcprx2 strains in livers and spleens of mice, 17 days after intravenous inoculation with 2x10⁷ parasites.

From the data above, we conclude that selective removal of either LicPRX1 or LicPRX2 yields fully viable parasites that are nevertheless less resistant to exogenous peroxide challenges. Despite displaying impaired virulence, Δcprx1 and Δcprx2 strains still develop as infective amastigotes in murine hosts.

3.5. The triple PRX locus on chromosome 15 is dispensable for parasite infectivity

Next, we investigated whether L. infantum tolerates the concomitant removal of both LicPRX1 and LicPRX2. To this end, we planned the generation of Δcprx1Δcprx2Δgprx mutants, i.e., parasites lacking both cPRX1/cPRX2/gPRX alleles. This strategy included co-disruption of the LigPRX ORF, which sits in the same chromosomal locus as LicPRX1 and LicPRX2 and encodes a glycosomal enzyme 99% identical to LicPRX1 (Fig. S3) that is dispensable for L. infantum survival and infectivity [47]. We started by attempting elimination of the intact cPRX1/cPRX2/gPRX allele in the SKO transformants (Fig. 4A) following a HR strategy. However, after several failed attempts, we decided to adopt the more robust CRISPR-Cas9-based strategy [48].

For the CRISPR-driven disruption of cPRX1/cPRX2/gPRX alleles, we started by genetically modifying L. infantum wild-type parasites to introduce i) a Cas9-expressing plasmid and ii) the LiDRP ORF in the ribosomal locus (Fig. 5A). The resulting “Cas9/DRP” parasites were subsequently transformed with sgRNA sequences matching unique sequences at the 5’UTR of LicPRX1 and at the 3’UTR of LigPRX, plus two donor cassettes containing BSD^R or PURO^R genes flanked by sequences of 30 nucleotides identical to the targeted locus, for homology-directed repair of the nicks produced by Cas9. Following transfection, and selection with BSD and PURO, we isolated several L. infantum clones that tested negative for LicPRX1, LicPRX2, and LigPRX ORFs. Fig. 5B shows PCR, WB and IFAT analyses of three of these clones (c1-c3), confirming the successful generation of a Δcprx1Δcprx2Δgprx mutant line.

Fig. 5.

Generation and phenotypic analysis of Δcprx1Δcprx2Δgprx L. infantum promastigotes. (A) Schematic representation of the diploid cPRX1/cPRX2/gPRX locus at chromosome 15 of L. infantum, showing LicPRX1, LicPRX2, and LigPRX coding sequences, as well as the intercalating double copy LiDRP gene. For CRISPR-Cas9 targeted deletion of both cPRX1/cPRX2/gPRX alleles, parasites were subjected to two rounds of transfection. The first round (pre-CRISPR) consisted in genetically modifying WT to accommodate a Cas9-expressing plasmid and LiDRP integrated in the ribosomal locus. On the second round (CRISPR), the resulting “Cas9/DRP” transgenics were transfected with sgRNA plus two BSD^R- and PURO^R-donor cassettes for homology-directed repair of the nicks produced by Cas9. The resulting Δcprx1Δcprx2Δgprx mutants retain the Cas9/DRP genetic background. The scheme also depicts the location of primers employed in PCR analysis of transformants (in B). (B) PCR, WB and IFAT diagnoses of three independent Δcprx1Δcprx2Δgprx clones (c1-c3) and control Cas9/DRP promastigotes. The anti-LicPRX2 antibody indiscriminately recognizes LicPRX1, LicPRX2, and LigPRX. (C) Proliferation of Δcprx1Δcprx2Δgprx, and of control Cas9/DRP L. infantum promastigotes. Parasites were seeded at 10⁶ cells ml⁻¹ (day 0) in complete RPMI medium and counted daily throughout six days of culture. Results refer to one representative experiment. (D) Relative normalized IC₅₀ values of H₂O₂ (left panel) and SIN-1 (right panel) estimated for Δcprx1Δcprx2Δgprx and Cas9/DRP promastigotes, 24 h after exposure to the compounds. Results represent means and standard deviations of three (H₂O₂) and four (SIN-1) independent assays. For each assay, data were normalized relative to Cas9/DRP promastigotes to minimize to inter-experimental variation. Statistically significant differences (two-way ANOVA, Tukey’s): *, c1 and c2 are different from Cas9/DRP (p < 0.05); #, c1 and c3 are different from Cas9/DRP (p < 0.005); ##, c2 are different from Cas9/DRP (p < 0.0005). (E) Infection parameters of murine bone marrow derived macrophages (MO) parasitized with Cas9/DRP and Δcprx1Δcprx2Δgprx (pool of clones c1, c2 and c3) at different time points after infection. Percentage of infected macrophages (left panel) and the average number of amastigotes per infected macrophage (right panel) are shown. Values represent mean and standard deviations of three independent experiments, each performed in triplicate. (F, G) Parasite burden (PB) in livers (left panels) and spleens (right panels) of C57BL/6 mice, (F) 7 days after intravenous injection with Cas9/DRP and Δcprx1Δcprx2Δgprx (pool clones c1, c2 and c3), and (G) 21 days upon intravenous inoculation with Δcprx1Δcprx2Δgprx (pool clones c1, c2 and c3) in grey and Δcprx1Δcprx2Δgprx/+cPRX1 strains in red. PB was determined by limiting dilution as the number of parasites per gram of organ and is represented log₁₀. The detection limit of this technique is 2.7 log units. Each circle represents one mouse and horizontal and vertical lines represent averages and standard deviations, respectively. Statistically significant differences (two-way ANOVA, Tukey’s): ***, Δcprx1Δcprx2Δgprx are different fromΔcprx1Δcprx2Δgprx/+cPRX1 (p < 0.0001). (For interpretation of the references to colour in this figure legend, the reader is referred to the Web version of this article.)

Δcprx1Δcprx2Δgprx promastigotes are morphological normal. Their growth is slightly delayed in comparison to strain Cas9/DRP, albeit they manage to reach cell densities similar to those of the control line (Fig. 5C). As expected from the complete lack of cPRXs, Δcprx1Δcprx2Δgprx promastigotes tend to be very sensitive to external H₂O₂ and the indirect ONOO⁻ donor SIN-1 (Fig. 5D). One exception to this is clone c3, for which no significantly different IC₅₀ value was estimated to H₂O₂ relative to control parasites (Fig. 5D).

To assess LicPRX1/2 essentiality in the amastigote stage of L. infantum, we carried out in vitro macrophage and in vivo mouse infections. In the in vitro assays, Δcprx1Δcprx2Δgprx strains behaved as the control Cas9/DRP parasites (Fig. 5E and S4). Likewise, in the in vivo infection setting, we could not detect any disadvantage of Δcprx1Δcprx2Δgprx parasites to infect and thrive in mice within the first 7 days post-infection (Fig. 5F), further indicating that none of the genes in the triple LiPRX locus is crucial for promastigotes to differentiate into amastigotes and, subsequently, infect the organs. However, lack of these enzymes impairs L. infantum proliferation (Fig. 5G), a phenotype rescued in Δcprx1Δcprx2Δgprx mutants complemented with cPRX1 (Δcprx1Δcprx2Δgprx/+cPRX1, Fig. S5). As a side note, we call attention to the fact that Cas9/DRP, Δcprx1Δcprx2Δgprx and Δcprx1Δcprx2Δgprx/+cPRX1 parasites are overall less virulent than WT, SKO, Δcprx1 and Δcprx2 parasites (compare Fig. 4, Fig. 5F and G), which might be tentatively attributed to the relocation of LiDRP to the ribosomal locus or to long-term Cas9 off-target effects. Nonetheless, from this set of experiments, we conclude that both cPRXs are dispensable for survival of L. infantum promastigotes. Notably, albeit advantageous in the context of murine infections, none of the three LiPRXs in the triple locus is individually or in combination required for parasites to infect and establish themselves as amastigotes in vitro and in vivo.

4. Discussion

This report provides novel insights into the properties, functions and relevance of Leishmania cytosolic 2-Cys-PRXs. First, it puts cPRXs under the spotlight of redox-signaling phenomena in these protist parasites. Second, it overthrows the pre-conceived dogma that cPRXs are crucial for Leishmania survival and infectivity. Third, it challenges the general oxidative stress theory in which oxidative attack of macrophages is a major leishmanicidal weapon. Finally, it questions the importance of cPRXs as drug targets.

The ability of some peroxiredoxins to alternate between hyperoxidized (inactive) and reduced/oxidized (active) states is regarded as a mechanism by which these enzymes fine-tune redox-signaling pathways [9,57]. In this study, resorting to in vitro and in yeast assays, we show that Leishmania cPRX1-type enzymes, contrary to cPRX2, are, in principle prone to inactivation by H₂O₂ (Fig. 2, Fig. 3 and 3A). This phenomenon is also evidenced upon exposing L. infantum promastigotes to H₂O₂ (Fig. S6). However, it is very unlikely that cPRX1 hyperoxidation is physiologically relevant for redox-signaling phenomena in Leishmania. First, LicPRX1 is relatively resilient to hyperoxidation when compared to other enzymes [44,60]. It features a C_hyp1% of 44 μM in vitro, whilst the roGFP2-LicPRX1 oxidation kinetics in response to H₂O₂ challenge are only consistent with mild hyperoxidation in yeast. Under physiological conditions, where peroxide concentrations rarely exceed 0.7 μM [66], it is plausible that most LicPRX1 molecules resist hyperoxidation. Second, Leishmania are unable to recycle hyperoxidized cPRXs. In the OFF/ON cycles that accompany PRX oscillation between hyperoxidized and reduced/oxidized states, sulfiredoxins (SRXs) are the key recyclers of inactivated PRXs [67]. SRX activity is what renders hyperoxidation a dynamic process by which cells swiftly control redox-signal transduction, instead of a dead-end, futile reaction [58,68]. In Leishmania, SRXs are thought to be absent. Not only are canonical SRX-coding genes missing from the genomes of Leishmania (and from other trypanosomatids) [58], no other enzymes seem to take over SRX functions in these parasites. On a preliminary assay to survey SRX activity in L. infantum promastigotes, we found that these parasites are unable to reduce hyperoxidized LicPRX1 within the 300 min (5 h) that follow an H₂O₂ challenge (Fig. S7). Together, the relative resilience of LicPRX1 to oxidative inactivation coupled with the incapacity of Leishmania to actively recycle hyperoxidized PRXs, question the significance of cPRX hyperoxidation as a regulatory-switch for redox signaling events. It is nonetheless plausible that LicPRX1 specialization in redox signaling proceeds via an enzymatic mechanism that does not involve C_P inactivation but rather its interaction with other protein thiols.

The C-terminal YF domain governs the oxidative inactivation of PRXs by delaying the reaction between C_P-SOH and the C_R residue of the adjacent PRX subunit (Fig. 3B and S8) [57]. This kinetic pause offers C_P-SOH the chance to react with a second or even third molecule of hydroperoxide and to become hyperoxidized. Importantly, as suggested by Rizza et al. [63] and shown in our roGFP2-based assays (Fig. 3), it also provides a unique mechanism for redox signaling by these PRX isoforms. In hyperoxidation-sensitive enzymes, the extended lifetime of C_P-SOH allows it to react with other protein thiols and, in this way, to relay oxidative signals via an intermediate mixed disulfide without the involvement of C_R. Conversely, in hyperoxidation-robust PRXs, the C_P-SOH tends to be more likely to react with the C_R residue of the neighboring PRX subunit. In this case, the oxidation of interacting partners could preferentially proceed via the mixed C_P-C_R disulfide, albeit under direct competition with the highly efficient C_P-C_R reducing agents, i.e., thioredoxin-type enzymes. Importantly, these dissimilar mechanisms of redox relay by hyperoxidation-prone and -resistant PRXs can translate into physiological differences. This is evidenced for two human PRX isoforms featuring different sensitivities to hyperoxidation [63]: the sensitive enzyme, but not its robust homologue, forms disulfide-linked conjugates with a series of transducers, such as the transcription factor STAT3 [69]. Whether or not cPRX1 can also act as a sensor and redox relay of H₂O₂-derived signals in Leishmania is a topic that deserves further investigation in the future.

This manuscript documents the paradigm-shifting discovery that Leishmania is tolerant of the simultaneous removal of gPRX and both cPRXs. The early assumption that trypanosomatid cPRXs are generally essential enzymes was based on two sets of indirect evidence. One was the essentiality of the molecules supporting the peroxidase activity of cPRXs, namely TR, TS₂, and TXN [11]. Here, by showing that cPRXs are non-essential to L. infantum, we circumstantially demonstrate that the vital role of the TR-TS₂-TXN cascade stems from its well-known ability to reduce a myriad of acceptors and/or pathways beyond cPRX enzymes [11]. The other line of evidence used to support the essentiality of Leishmania cPRXs was the fact that RNAi-induced down-regulation of the single cPRX1 enzyme in phylogenetically related T. brucei parasites resulted in a strongly reduced growth rate [35]. Interestingly though, when directly assessing cPRX function in L. infantum, we found the opposite, i.e., that these parasites tolerate the genetic excision of their full cPRX repertoire. Perhaps these conflicting T. brucei- and Leishmania-cPRX silencing outcomes reflect genuine species-specific differences in the requirement for these peroxidases. Alternatively, they might merely reflect technical specificities of the gene silencing tools employed in each case. In this latter regard, it is conceivable that stable L. infantum cPRX knockouts feature genetic compensatory phenomena that in transient T. brucei cPRX knockdowns would not have enough time to occur and/or manifest. In the future, systematic omic-based analyses of the Δcprx1Δcprx2Δgprx strain might help elucidate which mutations, if any, allow these parasites to cope with the absence of cPRXs. On a small-scale trial, we could exclude upregulation of the non-selenium glutathione peroxidase-like enzymes nsGPXA, the most obvious candidate substitutes of cPRX antioxidant function, as well as of mitochondrial peroxiredoxin and of ascorbate peroxidase [11,15,70], as compensation events in Δcprx1Δcprx2Δgprx mutants (Figs. S9 and S10). Whichever cPRX-bypass mechanisms, if any, are active in the Δcprx1Δcprx2Δgprx strain, they do not fully compensate for the loss of the oxidative and nitrosative shields otherwise granted by cPRXs. Indeed, knockouts are 2–3 times more sensitive to H₂O₂ and ONOO⁻ than control parasites (Fig. 5D). The finding that cPRXs are dispensable for L. infantum survival, albeit groundbreaking to the trypanosomatid community, is not surprising in the peroxiredoxin research field, where it actually echoes the trend of other non-essential cytosolic peroxiredoxins [[71], [72], [73], [74], [75]].

This study questions the general oxidative stress theory according to which the oxidative burst of macrophages is a major weapon against invading Leishmania pathogens. It reports that a strain of L. infantum (Δcprx1Δcprx2Δgprx), with compromised capacity to cope with external H₂O₂ challenges (IC₅₀ values estimated as 5 ± 3 μM and 47 ± 2 μM for Δcprx1Δcprx2Δgprx and WT promastigotes, respectively; Fig. 5D), does not feature any obvious defect to infect macrophages in vitro (Fig. 5E) and to establish an early infection in mice (Fig. 5F and 7 days p.i.). Since Leishmania promastigotes are enveloped in a lipophosphoglycan (LPG) coat, which provides an effective barrier against the external oxidative insults [76], it is plausible that the amount of H₂O₂ generated during the oxidative burst may not be harmful for Leishmania. Perhaps the peak of H₂O₂ generated during phagocytosis is even beneficial for parasites, for instances, by contributing to promastigote differentiation into amastigotes [77].

In conclusion, by reassessing biochemical and functional traits of Leishmania cPRXs, this report brings novel insights into the current state-of-knowledge on trypanosomatid 2-Cys-PRXs. First, by directly comparing LicPRX1 and LicPRX2, this study contributes to understand the diversity of cPRXs in trypanosomatids. Second, by defining the hyperoxidation-profiles and redox relay activities of Leishmania cPRXs, it positions these enzymes at the crossroad of signal transduction pathways, beyond mere antioxidant protection. Finally, by showing that L. infantum tolerates elimination of the cytosolic pool of PRXs, it overthrows the long-lasting assumption that these molecules are absolutely essential for parasite survival, raising serious doubts on the success of drugs targeting cPRXs [36].

CRediT authorship contribution statement

Helena Castro: Writing – review & editing, Writing – original draft, Supervision, Project administration, Investigation, Funding acquisition, Formal analysis, Conceptualization. Maria Inês Rocha: Investigation. Margarida Duarte: Writing – review & editing, Investigation, Formal analysis. Jordi Vilurbina: Investigation. Ana Georgina Gomes-Alves: Investigation. Teresa Leao: Investigation. Filipa Dias: Investigation. Bruce Morgan: Writing – review & editing, Supervision, Resources, Funding acquisition, Formal analysis, Conceptualization. Marcel Deponte: Writing – review & editing, Supervision, Funding acquisition, Formal analysis, Conceptualization. Ana Maria Tomás: Writing – review & editing, Visualization, Supervision, Project administration, Methodology, Investigation, Funding acquisition, Formal analysis, Conceptualization.

Declaration of competing interest

None

Appendix A. Supplementary data

The following is the Supplementary data to this article:

Data availability

No data was used for the research described in the article.