Depletion of Plasmodium berghei Plasmoredoxin Reveals a Non-Essential Role for Life Cycle Progression of the Malaria Parasite

Proliferation of the pathogenic Plasmodium asexual blood stages in host erythrocytes requires an exquisite capacity to protect the malaria parasite against oxidative stress. This function is achieved by a complex antioxidant defence system composed of redox-active proteins and low MW antioxidants. Here, we disrupted the P. berghei plasmoredoxin gene that encodes a parasite-specific 22 kDa member of the thioredoxin superfamily. The successful generation of plasmoredoxin knockout mutants in the rodent model malaria parasite and phenotypic analysis during life cycle progression revealed a non-vital role in vivo. Our findings suggest that plasmoredoxin fulfils a specialized and dispensable role for Plasmodium and highlights the need for target validation to inform drug development strategies.


Introduction
Plasmodium, the etiologic agent of malaria, is a unicellular facultative intracellular parasite of the phylum Apicomplexa. A hallmark of the malaria parasite is its remarkable capacity to specifically invade and replicate inside red blood cells. This intraerythrocytic proliferation phase ultimately leads to the disease known as malaria. Due to the high metabolic rates of the rapidly growing and multiplying parasite, large quantities of toxic redoxactive by-products are generated. Additional reactive oxygen and nitrogen species are generated by immune effector cells of the host in response to parasite infection and during hemoglobin degradation in the food vacuole of the parasite. Therefore, inside erythrocytes, the ability of Plasmodium to defend itself against oxidative damage is of vital importance for parasite survival [1,2], and it appears to be highly effective in this respect [3]. Plasmodium employs multiple biochemical pathways that mediate antioxidant defense and redox-regulation and play a central role in pathogenesis [4][5][6][7][8][9].
In most eukaryotic organisms, redox-active enzymes, such as catalase, superoxide dismutase, and peroxidases as well as an enzymatic cascade that generates reduced electron donors, i.e. glutathione (GSH) and thioredoxin (Trx), sustain the cellular redox homeostasis [10]. This redox network is split into two major arms, the GSH and the Trx system, that serve complementary functions in antioxidant defense and DNA synthesis. Interestingly, the malarial parasite Plasmodium lacks two central antioxidant enzymes: (i) catalase that typically detoxifies hydrogen peroxide and (ii) a classical glutathione peroxidase, a selenoenyzme that reduces lipid hydroperoxides to their alcohols [9]. This apparent deficiency further underscores the central importance of the thioredoxin system in the parasite. In good agreement, Trx reductase, which transfers electrons from NADPH to Trx, appears to perform vital functions for asexual development of the malaria parasite in vitro and is considered an attractive target for antimalarial drug development [11].
Intriguingly, malaria parasites possess-in addition to the classical thioredoxins-a Plasmodium-specific member of the thioredoxin superfamily termed plasmoredoxin (Plrx) (P. falciparum GenBank AAF87222) [12]. Plrx is a 22 kDa dithiol protein with the unique active site sequence WCKYC. Plrx is not reduced by thioredoxin reductase but can react with glutaredoxin and glutathione. Although a non-enzymatic reaction between reduced Trx and glutathione disulfide (GSSG) has been described in insects, which lack glutathione reductase [13], the physiologic electron donor for P. falciparum plasmoredoxin still remains to be identified. As described for certain thioredoxins and glutaredoxins, Plrx has been shown to serve as electron donor for ribonucleotide reductase. Furthermore, the protein is capable of reducing disulfide bonds in general and in particular P. falciparum thioredoxin peroxidase 1, the major cytosolic peroxiredoxin of the parasite [12,14].
In this study, we addressed the cellular role of Plrx in the rodent malaria model parasite Plasmodium berghei. Targeted gene disruption permits drug target validation or elucidation of the in vivo role of the gene product. The corresponding predicted experimental outcomes are refractoriness to gene targeting, which would correlate with a vital role in asexual blood stage development, or a detectable phenotype during life cycle progression, respectively. Because of the potential to design tailor-made inhibitors of the Plasmodium redox network as innovative antimalarial drugs [4,5], target validation of individual redox-active enzymes paves the way for future drug discovery approaches. Our findings that loss of Plrx function does not affect parasite development under normal growth conditions exemplifies the important role of reverse genetics in guiding drug development against malaria.

Generation of Plrx(-) parasites
To test whether plasmoredoxin is important for asexual replication of P. berghei, we first targeted the PbPlrx gene with an integration vector that disrupts the gene locus via a single cross over event. After a single transfection we successfully integrated the disruption plasmid (data not shown). To confirm our unexpected finding that PbPlrx can be disrupted we constructed a replacement vector containing the PbPlrx 59 and 39 UTRs that flank the positive selectable marker, which confers resistance to the antifolate pyrimethamine (Fig. 1A). Upon a double cross over event this vector is predicted to disrupt the entire PbPlrx open reading frame (ORF). After continuous selection with oral pyrimethamine a resistant population was obtained and genotyped (data not shown). This parental parasite population contained the correct integration mixed with WT parasites and was used for cloning of four independent parasite Plrx-deficient lines, named Plrx(-)Rep. Replacement-specific PCR analysis verified the correct replacement event after homologous recombination (Fig. 1B). To confirm the absence of Plrx transcripts in Plrx(-) parasites, RT-PCR and subsequent cDNA synthesis was performed with poly(A)+ RNA from mixed blood stages (Fig. 1C). In good agreement with the previous expression profiling of PfPlrx [12], we detected the PbPlrx transcript in asexual blood stage parasites of the WT. As predicted, no Plrx transcripts were detected in the knockout parasite lines. Moreover, Western blot analysis of Plrx(-) blood stages with a PbPlrx-specific anti-peptide antiserum confirmed complete absence of the protein in Plrx(-) parasites (Fig. 1D). Together, the successful generation of Plrx-deficient parasites demonstrates that this gene is not essential for proliferation of the asexual blood stages.
Plrx is dispensable for in vivo and in vitro growth of asexual blood stages To test whether Plrx serves an auxiliary role during blood stage growth we first performed an in vivo growth assay by intravenous injection of 1,000 asexual parasites, followed by parasitemia counts every 12 hours post infection ( Fig. 2A). Notably, proliferation of Plrx(-) parasites was indistinguishable from WT parasites indicating that Plrx is dispensable for the parasite under normal growth conditions. We next wanted to determine whether loss of Plrx function affects parasite growth under redox stress conditions. To this end, we performed an in vitro culture assay over 24 hours to determine the IC 50 values in WT and Plrx(-) parasites upon exposure to antimalarial drugs, some of which act also as elicitors of redox stress [15]. The compounds tested included 4-aminoquinolines, i.e. chloroquine and amodiaquine, the first synthetic antimalarial agent methylene blue, mefloquine, and the central compound of artemisin-combination therapies (ACT) recommended as first-line antimalarial treatment, artemisinin. In order to minimize stage-specific effects we used non-synchronized asexual blood stages [16]. Notably, the IC 50 values of the Plrx(-) parasite line did not differ significantly from WT parasites, excluding, at least ex vivo, a central function of Plrx in anti-redox stress defense (Table 1).
To elucidate whether this dispensability can be observed in vivo we applied a growth assay as outlined above under enhanced oxidative stress conditions. We selected methylene blue (MB) as this antimalarial was shown to challenge the parasites intracellular reducing milieu through the generation of pro-oxidant H 2 O 2 [17]. The WT and Plrx(-) infected mice were treated orally with 50 mg/ kg body weight methylene blue as previous dose finding experiments revealed rapid elimination of WT parasites with 100 mg/kg body weight MB (data not shown). Even with enhanced oxidative stress in vivo Plrx-deficient parasites grow indistinguishable from WT parasites (Fig. 2B), excluding a central role of P. berghei Plrx in antioxidant defense.

Depletion of Plrx induces only weak changes of transcript levels of selected redox proteins
Since our in vitro and in vivo data exclude an essential function of Plrx to maintain the parasite's redox equilibrium, we extended our analysis of the Plrx-deficient strain to expression profiling of selected redox proteins. This analysis was expected to further reveal the modulation of intracellular redox networks. We studied the effects of the Plrx-deletion on mRNA levels of genes related to cellular redox metabolism with a focus on the cytosolic components because Plrx is a cytosolic member of the antioxidant network [12]. Gene transcript levels were measured by quantitative real-time RT-PCR and the effect on a target gene is reported as differences in comparison to a WT control population (Fig. 3). Transcript levels of mRNA for the two major sustainers of redox homeostasis, thioredoxin reductase (TrxR) and glutathione reductase (GR) increased only slightly. Plrx was previously shown to directly interact with these two systems [12], thereby potentially acting as an additional antioxidant defence line of Plasmodium. The prediction was that deletion of Plrx may be accompanied by a compensatory upregulation of functional paralogues that balance the reducing capacity of Plrx. Such a function can most likely be fulfilled by thioredoxin (Trx) and/or glutathione (GSH). While this assumption is supported by the weak increase of Trx mRNA levels, GSH cannot be tested directly because it is only a tripeptide.
Another member of the thioredoxin superfamily, glutaredoxin (Grx) did not change significantly. Moreover, Plrx-deficient parasites show a slight decrease in mRNA levels of thioredoxin peroxidase 1 (TPx1), the major cytosolic peroxiredoxin of the parasite, and ribonucleotide reductase (RiboR). Collectively, these data show that depletion of PbPlrx caused only weak alterations in gene expression of selected members of the cytosolic redox network compared to wild type parasites.  Plrx is not essential for completion of the Plasmodium life cycle.
Since we could exclude a discernible function of Plrx in blood stage development we extended our phenotypic analysis to the entire Plasmodium life cycle ( Table 2). Plrx(-) parasites did not differ from WT parasites in sexual development, which is a prerequisite for transmission to mosquitoes (data not shown). Dissection of infected mosquitoes showed similar numbers of oocysts, midgutassociated and salivary gland-associated sporozoites in WT and Plrx deficient parasites (Table 2). Therefore, Plrx is also dispensable for sporogony and sporozoite maturation. When mature salivary gland sporozoites were tested for infectivity to the mammalian host in vivo and in vitro again no phenotypic differences between the two parasite lines could be observed. Hepatocytes infected with Plrx(-) sporozoites were indistinguishable from WT infected cells and produced high numbers of mature liver stage parasites (Table 2). When tested in vivo by intravenous injection or natural mosquito bite the recipient animals became patent after similar prepatent periods compared to WT sporozoite inoculation (Table 2). Together these data exclude a vital role for Plrx in Plasmodium life cycle progression under standard conditions.

Discussion
We initiated this study to test the potential of plasmoredoxin (Plrx), a Plasmodium specific member of the thioredoxin superfamily [12], as a novel antimalarial drug target. Using classical reverse genetics we could demonstrate that Plrx is dispensable for Plasmodium development inside its host cells. This finding rejects future drug discovery efforts that aim at specifically targeting Plrx, most likely even in combination with existing antimalarial drugs. Successful generation of Plrx(-) mutants permitted a detailed observation of the in vivo function of Plrx during life cycle progression of the malaria parasite. Again, no vital role at any stage of the parasite life cycle was revealed. Therefore, specific targeting of plasmoredoxin is not suitable either for transmissionblocking or causal-prophylactic malaria intervention strategies.
The redox-active proteins thioredoxin and glutaredoxin are founding members of the thioredoxin superfamily. Additional members include tryparedoxin of Trypanosomes, the protein disulfide isomerase and a few bacterial disulfide bondforming proteins. This group of proteins carries out oxidation and reduction reactions based on the chemistry of the catalytic cysteine residues. The unifying features of all family members are (i) the typical active site motif C-X-X-C and (ii) similarity in the overall tertiary protein structure, the so-called thioredoxin fold, despite low overall amino acid sequence similarity [18]. Intriguingly, homozygous deletions of the mouse cytoplasmic Trx1 or the mitochondrial isoform Trx2 resulted in early embryonic lethality indicating key roles of the thioredoxin system in the development of multicellular organisms [19,20]. In analogy, the recent characterization of the Plasmodiumspecific dithiol:disulfide oxidoreductase Plrx raised the attractive possibility that the protein performs an important function for the intracellular life style of the malaria parasite.
Thus far, five thioredoxin related proteins have been identified in Plasmodium falciparum, in addition to plasmoredoxin [21]. Cytosolic Trx1 is the major substrate of thioredoxin reductase. It reduces thioredoxin-dependent peroxidases and is also capable of reacting with peroxides, dehydroascorbate, lipoic acid, and lipoamide directly. Trx2 (and TPx2) were shown to be mitochondrial [22]. Trx2 also displays general disulfide reducing activity and serves as electron donor for thioredoxin peroxidises and GSSG. Trx3, which also carries a targeting sequence, has been shown to be redox active and reducible by PfTrxR. In addition, two thioredoxin-like proteins, Tlp1, which might  Percentage of mosquitoes that contain oocysts at day 11-13 post-feeding. b Midgut-associated sporozoites per infected mosquito at day 12-14 post-feeding. represent a small dynein subunit, and Tlp2 have been reported in Plasmodium falciparum and might exhibit partially overlapping functions with classical thioredoxins.
An explanation for the non-vital role of Plrx in malarial parasites might thus be redundancy in the function of the multiple members of the thioredoxin superfamily. There is precedence from other systems. For instance, the bacterial cytoplasm typically contains two thioredoxins, thioredoxin 1 (trxA) and thioredoxin 2 (trxC), as well as the glutaredoxin system comprising glutaredoxins 1 to 3 (grxA-C) [23,24]. Out of these proteins three (trxA, trxC, and grxA) are able to reduce ribonucleotide reductase. Importantly, null mutants lacking any of these genes are viable [23,25], whereas the corresponding E. coli triple knockout is not [26]. The combined vital role of these complementary bacterial redox proteins is best explained by their overlapping functions [27]. Therefore, in E. coli the presence of any one of these oxidoreductases is sufficient for survival [26]. Another example is found in Saccharomyces cerevisiae that encodes two cytoplasmic glutaredoxins (Grx1/2) [24,28,29] and two thioredoxins (Trx1 and Trx2) [30]. In analogy to bacteria, only a quadruple yeast mutant is non-viable, and a single glutaredoxin or thioredoxin is able to restore viability [31]. However, strains deleted for glutaredoxins display altered sensitivity when exposed to various reactive oxygen species [28,29]. In addition, the Trx double knockout displays a profound cell cycle defect, resulting in a prolonged S and a shortened G1 phase [30]. This effect again is presumably based on the inefficient ribonucleotide reduction, which results in insufficient supply of deoxyribonucleotides, a rate-limiting step in DNA synthesis. Due to advanced reverse genetic tools in yeast and E. coli, multiple knockouts of the GSH, glutaredoxin and thioredoxin pathways could be generated [26,[28][29][30] that established the overlapping nature of the two systems and provided crucial insight into the cellular responses to oxidative and reductive stress.
Currently, these genetic tools are not available for P. berghei. It will however be interesting to test in the future which redox-active proteins functionally overlap with Plrx in the malaria parasite. Plasmoredoxin may still fulfill specific and important functions in the cell in concert with other members of the thioredoxin superfamiliy that are not readily revealed under normal in vivo growth conditions. Under more pronounced stress conditions and in certain parasite stages these specific functions of plasmoredoxin might become evident. However, because of the non-vital role and the likelihood of major functional redundancy in Plasmodium redox defence and ribonucleotide reduction, Plrx can largely be excluded as a valid target for antimalarial intervention strategies. We propose that target validation by reverse genetics in the rodent malaria model parasites is an indispensable requirement for preclinical drug development.

Experimental animals
Sprague-Dawley rats, NMRI mice and C57bl/6 mice were obtained from Charles Rivers Laboratories. All animal experiments were conducted in accordance with European regulations and approved by the state authorities (Regierungsprä sidium Karlsruhe).

Immunoblotting
Three rats were infected with WT and Plrx(-) blood stage parasites, respectively, and cytosolic parasite proteins were extracted using 2M Urea buffer. For immunoblot analysis, proteins were separated on 15% polyacrylamide gels and transferred to nitrocellulose membranes (Bio-Rad) by electroblotting. Plrx (21.5 kDA) was detected by incubation of membranes with a polyclonal anti-P. berghei Plrx antiserum (dilution 1:1,500). This antiserum was obtained by immunization of rabbits with synthetic peptides of the N-terminal region of Plrx (Plrx-1: CYYKNNELKKIDSSYFQDKY and Plrx-2: CKVDKALEHSTQNEAPSK) (Eurogentec, Seraing, Belgium). Bound antibodies were detected using peroxidase coupled antirabbit and anti-mouse antibodies, to detect PbPlrx and actin, respectively. Immunostained proteins were visualized with enhanced chemoluminescence detection (Pierce). Recombinant P. berghei Plrx was used as a positive control (obtained according to [12]). The anti-Dictyostelium discoideum actin antiserum, which cross-reacts with apicomplexan actins, was kindly provided by Dr. Markus Meissner (Heidelberg).

Determination of IC 50 values
NMRI mice were infected with P. berghei WT and Plrx(-) parasites. Collection of infected erythrocytes was done by cardiac puncture. Infected erythrocytes were cultured in an isotopic drug sensitivity assay based on incorporation of radioactive [3H] hypoxanthine (protocol kindly provided by Sergio Wittlin, Basel, Switzerland). Briefly, serial dilutions of various drugs were prepared in 96-well microtiter plates (Nunc). 100 ml infected red blood cells, resulting in 5% hematocrit and 1-5% parasitemia, were added per well and incubated in hypoxanthine-free medium at 37uC in a gas mixture consisting of 94% N 2 , 3% O 2 and 3% CO 2 . After 16 h incubation, 0.5 mCi of [3H] hypoxanthine (Amersham Pharmacia) in 50 ml medium was added and plates were incubated for an additional 8 h. Parasites were harvested onto glass fiber filters (Perkin-Elmer, Rodgau-Jügesheim, Germany), washed and dried. Radioactivity was counted using a b-counter (Matrix 9600; Packard). The results were recorded as counts per minute (cpm) per well at each drug concentration and growth inhibition was expressed as percent 3H incorporation compared with untreated controls. Fifty percent inhibitory concentrations (IC 50 ) were calculated from plotted data. All experiments were done in duplicate from at least three independent mouse infections each.
Expression profiles of Plasmodium WT and Plrx(-) parasites using quantitative real-time RT-PCR For qRT-PCR analyses poly(A) + RNA was isolated using oligo(dT) columns (Macherey-Nagel, Düren, Germany) from mixed blood stage of Plasmodium berghei WT and Plrx(-) lines. After a DNase treatment, aliquots of 800 ng of each sample were reversely transcribed to cDNA using an Abgene cDNA synthesis kit and oligo(dT) primers (Abgene, Hamburg, Germany). The SYBR Green Jumpstart Taq Ready Mix (Sigma, Steinheim, Germany) was utilized for quantitative real-time PCR on Rotor-Gene 3000 Real-Time PCR system (Corbett Research, Sydney, Australia) using the primers listed in Table 3. Specificity of the amplification products was confirmed by melting curve analysis; a no-template control was added in every run. The Rotor-gene 6.0 software was used to retrieve the real time PCR results, to analyse the data and to determine cycle threshold values. Relative expression of target genes were calculated based on efficiency (determined with dilution series of each gene and the Rotor-Gene software) and Ct-values for genes of Plrx-deficient parasites versus control WT-parasites, and expressed in comparison to a reference gene. The relative amount of mRNA was determined according to the Pfaffl equation [34], using P. berghei seryl-tRNA synthetase as the reference gene. In our experiments, each RT-PCR run was carried out in triplets of one mRNA pool or in doublets of two different mRNA pools and the whole series was reproduced in an independent experiment.
Plasmodium life cycle and phenotypic analysis of Plrx(-) parasites Anopheles stephensi mosquitoes were kept at 21uC, 80% humidity and daily feeding on 10% sucrose. Asynchronous blood-stages of P. berghei NK65 (WT) and Plrx(-) were maintained in NMRI mice and checked for gametocyte formation and exflagellation of microgametes prior to mosquito feeding. For mosquito infection, A. stephensi mosquitoes were allowed to bloodfeed on anesthetized mice for 15 minutes. Dissection of mosquitoes was conducted at days 10, 14 and 17 in order to determine infectivity and sporozoite numbers in midguts and salivary glands, respectively. Gliding motility was assessed by deposition of sporozoites onto precoated glass coverslips and visualisation by indirect immunofluorescence using a primary antibody against the P. berghei circumsporozoite protein (CSP) [35] followed by detection with an Alexa Fluor 488-conjugated antimouse antibody. To analyse liver stage development sporozoites were deposited onto a semi-confluent monolayer of hepatoma cells (HuH7) and incubated for 2 h, followed by washing and incubation in cell culture medium. Liver stages were detected after 48 h with a primary antibody directed against the P. berghei heat shock protein 70 (HSP70) [36], followed by an Alexa Fluor 488-conjugated antimouse antibody. To analyse sporozoite infectivity in vivo, Sprague-Dawley rats were infected intravenously with 10,000 WT or Plrx(-) sporozoites, respectively. Parasitemia was followed by daily examination of Giemsa-stained blood smears. The occurrence of a single parasite marked the first day of patency.