Thioredoxin reductase is essential for the survival of Plasmodium falciparum erythrocytic stages.

The human malaria parasite Plasmodium falciparum poses an increasing threat to human health in the tropical regions of the world, and the validation and assessment of possible drug targets is required for the development of new antimalarials. It has been shown that the erythrocytic stages of the parasites, which are responsible for the pathology of the disease in humans, are under enhanced oxidative stress and are particularly vulnerable to exogenous challenges by reactive oxygen species. Therefore it is postulated that the disruption of the antioxidant and/or redox systems of the parasite is a feasible way to interfere with their development during erythrocytic schizogony. In order to test this suggestion thioredoxin reductase (TrxR), an enzyme heavily involved in maintenance of redox homeostasis and antioxidant defense, was knocked out in P. falciparum. It was impossible to generate parasites with a disrupted trxR gene suggesting that TrxR is essential for P. falciparum erythrocytic stages. Technical problems were excluded by transfecting a 3' replacement construct, which recombined correctly and transfectants did not show any phenotypic alterations. In order to prove that the trxR knockout was responsible for the lethal phenotype of the null mutants, a co-transfection with both the knockout construct and a construct containing the trxR coding region under the control of the calmodulin promoter was conducted. Despite the disruption of the trxR gene, parasites were viable. In a Southern blot analysis a complicated restriction pattern was obtained, but it was shown by pulse field gel electrophoresis and field inverse gel electrophoreses that only the trxR gene locus on chromosome 9 was targeted by the constructs. It was found that the co-transfected constructs form concatemeric structures prior to integration into the trxR gene locus, which is further supported by plasmid rescue followed by restriction analyses of the plasmids. Northern and Western blot analyses proved that the co-transfectants highly overexpress TrxR from the introduced gene. Our results demonstrate that TrxR is essential for the survival of the erythrocytic stages of P. falciparum.

Aerobic organisms are exposed to reactive oxygen species (ROS), 1 which are generated as by-products of their metabolism throughout their lives. ROS are combated by the organism with antioxidative systems, which include enzymes such as superoxide dismutase, catalase, and peroxidases as well as non-enzymatic antioxidants like glutathione and vitamins E and C (1). These antioxidants prevent damage of nucleic acids, proteins, and membrane integrity and are crucial for the survival in an aerobic environment (1,2). Parasites have not only to cope with the oxidative stress generated by their own metabolic reactions but also have to fight ROS generated by the host immune system.
Plasmodium falciparum, the causative agent of malaria tropica, is one of the major threats to human health in the tropical regions of the world. The parasites have developed resistance to most commonly used drugs, and it is of great interest to identify, validate, and assess new potential targets suitable for the design of novel chemotherapeutics. The fact that Plasmodium-infected erythrocytes are under enhanced oxidative stress together with their susceptibility to exogenous oxidative challenge is a reason to believe that the antioxidant systems of the parasite are important for the survival of the parasite during their erythrocytic life stages (3)(4)(5).
It was shown that P. falciparum possesses two functional redox systems, which involve the low molecular weight thiol glutathione and the 12-kDa protein thioredoxin (Trx) (6 -11). Both are part of enzymatic redox cascades, which transfer electrons from NADPH to acceptor molecules such as hydrogen peroxide and alkyl hydroperoxides but also proteins like ribonucleotide reductase and a number of transcription factors (12)(13)(14)(15)(16)(17)(18)(19). Thus these systems are heavily involved in the redox regulation of the cell and are thought to contribute greatly to the antioxidant capacity of the cell.
Thioredoxin reductase (TrxR) is a flavo oxidoreductase, which exists in two different types in nature. The low molecular weight type occurs in bacteria, plants, and fungi, whereas mammalian cells, the fruit fly Drosophila melanogaster, the nematode Caenorhabditis elegans, and the malaria parasite P. falciparum contain the high molecular weight type of reductase (20 -28). Plasmodium TrxR has been investigated intensively and by employing mutagenesis and kinetic studies showed that it differs significantly from its mammalian counterpart (10,21,29,30). Both proteins belong to the family of high molecular weight TrxR, but due to different active site residues they are distinct in their substrate and inhibitor profiles that may be exploitable for the design of antimalarials (21). To establish the significance of the protein for the survival of P. falciparum, we have genetically modified blood stage forms of parasites and have demonstrated that TrxR is indeed essential for their survival.

EXPERIMENTAL PROCEDURES
Materials-[␣-32 P]ATP (3000 Ci/mmol) and the ECL Western blot detection system were purchased from Amersham Biosciences. Albumax II and RPMI 1640 were from Invitrogen. Pyrimethamine was purchased from Sigma. WR 99120 was a kind gift of Dr. Jacobus, Jacobus Pharmaceuticals.
Generation of TrxR Transfection Constructs-The TrxR knockout construct was generated by PCR using Pfu polymerase with P. falciparum 3D7 genomic DNA as a template and the specific oligonucleotides 5Ј-GGAAGATCTGTTTTATTGTTTGATTATGTAAAGCC-3Ј and 5Ј-CC-GCTCGAGTTATCCACATGCACCATATTCAATAGG-3Ј to obtain a 1028-bp fragment. The fragment starts at position 196 of the trxR open reading frame and ends at position 1224 where an artificial stop codon was introduced into the oligonucleotide (Fig. 1A). The sense oligonucleotide contains a BglII restriction site (boldface), and the antisense oligonucleotide contains an XhoI restriction site (boldface) in order to allow directional cloning into the previously cut transfection plasmid pHH1 (31) (see Fig. 1A) resulting in construct pHH1TrxR-ko. The 3Ј replacement control fragment was amplified from genomic DNA with the oligonucleotides 5Ј-GCGCGAGATCTCATATACCAGATGATGTT-GAAGG-3Ј and 5Ј-GCGCGCTCGAGTTATCCACATTTTCCACCCCCA-C-3Ј also containing the restriction sites required for cloning into pHH1. The fragment represents 1037 bp, starts at position 589 of the coding region of trxR and includes the genuine stop-codon at its 3Ј-end. The resulting construct is called pHH1TrxR-3ЈR (Fig. 1B).
In order to be able to clone the txrR coding region into the overexpression plasmid pHC1, an XhoI site was mutated according to Gilberger et al. (10) using the oligonucleotides sense 5Ј-GCTGCAGC-ACATGGTGCCCGGGTTTTATTGTTTGATTATG-3Ј, antisense 5Ј-CATAATCAAACAATAAAACCCGGGCACCATGTGCTGCAGC-3Ј, Pfu polymerase, and the expression plasmid pJC40-TrxR (10) as a template in a PCR-based protocol. The PCR fragment was cloned into the TOPO T/A vector (Invitrogen), and the mutation was verified by nucleotide sequencing. Subsequently, trxR was amplified from the TOPO clone with the specific oligonucleotides sense 5Ј-GCGCGCTCGAGAAAAA-CATGTGTAAAGATAAAAACG-3Ј and antisense 5Ј-GCGCGCTC-GAGTTATCCACATTTTCCACCCCCAC-3Ј. Both oligonucleotides possess an XhoI site (boldface) to allow subcloning of the PCR fragment into the P. falciparum overexpression plasmid pHC1 (32) resulting in plasmid pHC1TrxR-Ex (Fig. 1C). The nucleotide sequences of all constructs were verified by automated sequencing on an ABI 377 (Applied Biosystems Inc.). Plasmid DNA was prepared using the Qiagen Maxi Prep kit (Qiagen).
Parasites and Transfection of Erythrocytic Stages-P. falciparum D10 was cultivated according to Trager and Jensen (33) in RPMI 1640 medium containing 5% human serum O ϩ and 0.05% Albumax II in human erythrocytes blood group O ϩ under a reduced oxygen atmosphere. Prior to transfection parasites were synchronized by sorbitol lysis (34). Transfection was carried out as described previously (35)(36)(37) with 100 g of circular DNA of either pHH1TrxR-ko, pHH1TrxR-3ЈR, or both pHH1TrxR-ko and pHC1TrxR-Ex. Transfected parasites were cultured in 90-mm Petri dishes for 48 h without drug pressure before the medium was supplemented with 0.2 M pyrimethamine for 2 days followed by selection with 0.1 M pyrimethamine until 5% parasitemia was reached. The first parasites were observed after 21-32 days of selection. In order to select for homologous recombination of the constructs and loss of episomal DNA, transfectants were cultivated for 3 weeks without drug pressure followed by 2 weeks with pyrimethamine pressure. This selection cycle was repeated for several rounds, and at different stages parasite DNA was isolated and subjected to Southern blotting and PCR analyses to determine whether episomal DNA was present and whether integration into the trxR gene locus had occurred. After 3 rounds of selection, parasites were cloned by limiting dilution (38) and used for genotypic and phenotypic analyses. In the case of the co-transfection of knockout and overexpression constructs, clones were generated using selection with pyrimethamine (0.1 M) and WR 99120 (5 nM).
DNA and RNA were isolated from wild type parasites and transfectants as described elsewhere (39,40). 15 g of genomic DNA were digested with NotI and DraIII, separated on a 0.8% agarose gel, and transferred to positively charged nylon membranes (Roche Molecular Biochemicals). Subsequently the membranes were hybridized with either the trxR knockout fragment or the 3Ј-UTR of PbDT specific for pHH1TrxR-ko in order to identify the genotype of the respective parasite clone. Northern blots were performed according to Kyes et al. (40) and probed with the coding region of trxR.
In order to analyze the genotype of co-transfected parasite clones in addition to Southern blot analyses, a pulse field gel electrophoresis FIG. 1. Schematic presentation of the constructs used for transfection. The transfection plasmids pHH1TrxR-ko (A), pHH1TrxR-3ЈR (B), and pHC1TrxR-Ex (C) and their arrangement after a homologous recombination event into the trxR gene locus are shown. BglII, DraIII, NotI, and XhoI restriction sites are indicated. A, pHH1TrxR-ko contains a 1024-bp fragment of trxR starting 196 bp downstream of the ATG initiation codon and carries an artificially introduced stop codon at its 3Ј-end. The knockout fragment possesses the 3Ј region of P. berghei dihydrofolate reductase thymidylate synthase. After homologous recombination, the endogenous 5-kb NotI/ DraIII fragment should disappear upon Southern blot analysis, and new bands at 1.85 and 9.95 kb should appear when using the trxR-ko region as a probe. B, pHH1TrxR-3ЈR encompasses a 1037-bp fragment of the 3Ј region of trxR starting at position 1224 of trxR and possesses the genuine stop codon at its 3Ј-end. Upon homologous recombination into the trxR locus, the endogenous 5-kb restriction fragment obtained with NotI and DraIII should disappear, and new bands at 2.29 and 9.5 kb should appear. The pHH1 plasmids were initially constructed on a pGEM-3Z (Promega) vector backbone and contain the human DHFR as selection marker under control of the P. falciparum calmodulin promoter and the histidine-rich protein 2 (HRP2) 3Ј region. C, the TrxR overexpression plasmid carries the entire coding region of trxR under control of the calmodulin promoter and the heat shock protein 86 (HSP86) 3Ј region. pHC1 also has the pGEM-3Z vector backbone but contains the Toxoplasma gondii DHFR-TS as selection marker under control of the Plasmodium chabaudi DHFR-TS promoter and histidinerich protein 2 3Ј region. Arrows indicate the location of primers used in the PCR analysis of D10TrxR-ko and D10TrxR-3ЈR parasites shown in Fig. 5. The abbreviations used are as follows: DHFR-TS, dihydrofolate reductase thymidylate synthase; trxR, thioredoxin reductase; trxR-ko, thioredoxin reductase knockout fragment; trxR-3Ј, thioredoxin reductase 3Ј-end; HSP6-3Ј, P. falciparum heat shock protein 86 3Ј-UTR.
(PFGE) was performed. The chromosomes were separated by using contour clamped homogenous field apparatus at 4.2 V cm Ϫ1 and a pulse time of 225 s for 72 h (41). The blots were hybridized with three different probes (trxR knockout fragment, pGEM vector backbone, and PbDT-3Ј region).
Furthermore, a field inverse gel electrophoreses (FIGE) was carried out. Genomic DNA of co-transfectants was digested with ClaI, a restriction enzyme that cuts rarely in the P. falciparum genome, and the fragments were separated on a 1% agarose gel in a FIGE (at 200 V and a pulse time of 0.8 -1.5 s for 16 h). The DNA was blotted onto positively charged nylon membrane and hybridized with the trxR-coding region and PbDT-3Ј region according to standard methods (42).
Plasmid Rescue from Co-transfected Parasites-DNA from co-transfected parasites was isolated as described above and subsequently used to transform Escherichia coli DH5␣. Bacterial clones were grown overnight in Luria-Bertani medium containing 50 g/ml ampicillin and plasmid DNA was isolated according to Sambrook et al. (42). DNA was restricted with BglII and XhoI in single digests and double digests and subsequently separated on a 1% agarose gel.
Western Blot Analysis-Proteins of wild type and transfected parasite clones were isolated after saponin lysis (43). For Western blot analyses 5-l aliquots of parasite pellets were resuspended in 10 l of 2ϫ SDS sample buffer containing 5% 2-mercaptoethanol, separated on a 7.5% SDS-PAGE, and subsequently blotted onto nitrocellulose according to Sambrook et al. (42). The blots were blocked in 3% bovine serum albumin for 1 h and subsequently incubated with the primary rabbit anti-TrxR antiserum (1:1000) for 1 h at room temperature. As a loading control the blots were probed with antibodies raised against 1-Cys peroxiredoxin of P. falciparum (1:2000) (14). Afterward the blots were washed twice in Tris-buffered saline containing 0.1% Tween 20 (TBS-Tween), and the secondary anti-rabbit horseradish peroxidase-conjugated antibody was applied (1:10,000) for 1 h before the blot was washed in TBS-Tween and developed using the ECL detection system and exposed to BioMax MR-1 films (Eastman Kodak).

RESULTS
Transfection of P. falciparum-In order to elucidate the significance of TrxR in P. falciparum, it was attempted to generate trxR null mutants transfecting them with pHH1TrxR-ko (Fig. 1A). However, it proved difficult to obtain viable parasites, and there are a number of possible explanations as follows: (i) problems with the transfection technique, (ii) occurrence of an antisense effect (44,45), or (iii) trxR knockout has a lethal effect.
In order to exclude that the failure to obtain stable trxR null mutants is due to technical problems, a control construct was transfected that replaces the 3Ј region of the gene upon homologous recombination without disrupting it (see Fig. 1B) (46). By using this construct, parasites grew normally, and it was shown by Southern blot analyses that recombination events into the trxR locus already started to occur during the 1st cycle of pyrimethamine selection (Fig. 2). The transfectants were cloned, and Southern blot analyses were performed. Fig. 2 shows that clone 3 contains pHH1TrxR-3ЈR integrated in the trxR locus. Importantly, the transfectants did not show any phenotypic changes, and their growth was comparable with wild type P. falciparum D10. These results demonstrate that the trxR gene locus is targeted by the constructs transfected and supports the suggestion that the difficulty in obtaining trxR null mutants was due to the disruption of the trxR gene.
To confirm that we could target the trxR gene and disrupt its expression, we made transfection constructs that would disrupt the gene and complement lack of expression with an introduced trxR transgene (46). Parasites were concomitantly transfected with the knockout construct pHH1TrxR-ko and the overexpression construct pHC1TrxR-Ex (Fig. 1C), and the transfectants survived several selection cycles. In order to verify that the parasites contained both constructs, a plasmid rescue experiment was performed. 10 independent bacterial clones were analyzed. Interestingly, the restriction pattern obtained show that the knockout and overexpression plasmids have recombined after transfection. Furthermore, only two different vari-ants of clones could be isolated, implicating that "hot spots" exist where the plasmids prefer to recombine (Fig. 3) (47, 48).
These parasites were cloned to determine the nature of the plasmids co-transfected. The clones 7D and 11E were chosen for further analyses because they were shown to carry both plasmid constructs in some form. To prove the assumption that the clones no longer possess episomal DNA but have integrated the recombined plasmids into the trxR locus, they were analyzed by PCR, Southern blotting, PFGE, and FIGE.
Southern blot analyses of both clones showed a complicated hybridization pattern when probed with the trxR-ko fragment as shown in Fig. 4A. The new band at 1.85 kb, however, strongly indicates that the endogenous trxR gene has been disrupted by pHH1TrxR-ko (see Fig. 1A). Disruption of the endogenous trxR by pHH1TrxR-ko was verified by PCR analysis using primers that specifically bind upstream of the trxR knockout fragment and within the PbDT-3Ј region of pHH1TrxR-ko knockout construct (Fig. 5). In addition numerous other bands are found in the Southern blot, and we analyzed this genotype further by probing it with PbDT-3Ј. The 1.85-kb band diagnostic for the disruption of the endogenous trxR gene was also obtained with this probe, and a second band at 3.2 kb appears on the blot, indicating that at least two copies of pHH1TrxR-ko are present in the transfectants. The fact that DraIII only cuts pHC1TrxR-Ex once and NotI only cuts pHH1TrxR-ko once and there are no endogenous DraIII or NotI sites near the trxR gene locus, the appearance of several bands in the Southern blot indicates that they originate from recombined episomes and not from recombination events into other gene loci of the Plasmodium genome.
This interpretation was confirmed by PFGE, using chromosomes from wild type Plasmodium and the transfectants (Fig.  6). Hybridization of the blot with the trxR-ko fragment resulted in a single signal on chromosome 9. In clones 7D and 11E the bands have shifted slightly and also the intensities of the signal appear to be higher. The shift in size of the chromosome may be attributable to the integration of a concatamer of pHH1TrxR-ko and pHC1TrxR-Ex which leads to quite a considerable increase of size but also may have structural implications on the chromosome. It has been suggested by OЈDonnel et al. (48) that the formation of concatamers (namely trimers) during replication of the parasites is essential for correct segregation and maintenance of the episomes during schizogony, and therefore our results are not surprising. The formation of Genomic DNA was digested with NotI/DraIII. Lane 1, after transfectants grow up, they contain an episomal copy (6.8 kb) of pHH1TrxR-3ЈR and the 5-kb endogenous trxR fragment. Lane 2, in the 1st cycle a correct integration of pHH1TrxR-3ЈR into the trxR locus has occurred (appearance of two new fragments of 2.29 and 9.5 kb, respectively); the parasites still contain episomes (6.8 kb) and the endogenous gene (5 kb). such concatameric structures was established by a FIGE with subsequent blotting. As shown in Fig. 7, after ClaI restriction of P. falciparum D10 genomic DNA, a band of 30 kb was detectable using the trxR probe. This band shifts by at least 20 kb to a size ranging between 50 and 90 kb in clone 11E indicating that several copies of the transfection constructs were integrated into the trxR locus. Another important finding of this experiment is that the pHH1TrxR-ko-specific probe PbDT-3Ј also hybridized to the same band, further supporting our suggestion that the endogenous txrR gene of these clones is disrupted. Presumably, the survival of the parasites is ensured by overexpression of TrxR from the co-transfected plasmid. These findings strongly suggest that TrxR is essential for the survival of P. falciparum during the erythrocytic development and validate this protein as a target for chemotherapy of malaria.
Northern and Western Blot Analyses-Total RNA of wild type D10 and the transfectants (TrxR-3ЈR, Co-7D, and Co-11E) was separated on a 1.5% agarose gel containing 5 mM guanidine thiocyanate, transferred to nylon membranes, and subsequently hybridized with a trxR probe. Fig. 8 shows that the message of trxR (3.2 kb) in wild type and control clone TrxR-3ЈR is present in equal amounts, whereas in clones 11E and 7D the message was greatly increased. These results are consistent with the fact that several copies of pHC1TrxR-Ex are present in both clones and that transcription of the recombinant gene is driven by the strong calmodulin promoter (32).
To determine whether the level of expression of TrxR was also increased in the Co-7D and Co-11E parasites, we used Western blot analyses of parasite lysates. These were separated on SDS-PAGE, blotted onto a nitrocellulose membrane, and incubated with antibodies raised against TrxR. Consistent with the elevated TrxR mRNA observed in the Northern blot analyses, the wild type and the TrxR-3ЈR clone clearly show a FIG. 3. Plasmid rescue of co-transfectants. Re-isolated episomes from co-transfected P. falciparum parasites were digested with XhoI, BglII, and XhoI/BglI. The pattern of the double digest of clones 1 and 2 shows that both plasmids (pHH1TrxR-ko and pHC1TrxR-Ex) are present and have formed two different megaplasmids (chimera of both plasmids). In clone 1 the appearance of a 1-and a 5.8-kb band upon double digest shows that the pHH1TrxR-ko plasmid is present and that the 1-kb trxR knockout fragment is cut out with XhoI/BglII. The additional three bands (4.8, 2.4, and 1.6 kb) are attributable to pHC1TrxR-Ex where the trxR full-length gene is cut out with XhoI (see Fig. 1, A and C, for location of XhoI and BglII restriction sites). Clone 2 shows a different restriction pattern, which is due to different recombination events of pHH1TrxR-ko and pHC1TrxR-Ex compared with clone 1. Nevertheless, the 1-kb trxR knockout fragment and the 1.6-kb trxR full-length overexpression fragment are present. M, 1-kb ladder (MBI Fermentas).

FIG. 4. Southern blot analyses of co-transfected parasite clones Co-7D and Co-11E hybridized with four different probes: trxR-ko fragment (A) and PbDT-3-UTR (B).
Genomic DNA was digested with NotI (N), DraI (D), and NotI/DraI (N/D). In both clones the appearance of the 1.85-kb NotI/DraIII fragments (arrows), which hybridize with trxR and PbDT-3Ј-UTR-specific probes, indicates that the endogenous trxR-ko has been disrupted by pHH1TrxR-ko. The other bands are attributable to the formation of a megaplasmid (chimera), which contains both pHH1TrxR-ko and pHC1TrxR-Ex. For the exact location of the probes see Fig. 1, A and C. The abbreviations used are as follows: Co-7D and Co-11E, parasite clones 7D and 11E that were co-transfected with both plasmids, pHH1TrxR-ko containing the knockout fragment of trxR and pHC1TrxR-Ex overexpressing trxR; trxR, thioredoxin reductase.
FIG. 5. PCR analyses of co-transfectants. Clones Co-7D and Co-11E were used as template for the PCR-based analysis of the disruption of the endogenous trxR gene. Co-7D and Co-11E, sense primer binds upstream of the TrxR knockout fragment; the antisense primer is specific for pHH1 (see arrows in Fig. 1A for exact location of primers). ϩ, positive control PCR using primers specific for pHH1TrxR-ko; Ϫ, negative control without DNA.
FIG. 6. Pulse field gel electrophoresis of co-transfected parasite clones 7D (Co-7D) and 11E (Co-11E) in comparison with D10. The blot was hybridized with three different probes: trxR-ko, pGEM vector backbone, and PbDT-3Ј. All three probes hybridize to the same chromosome in the co-transfected clones 7D and 11E, which indicates that integration of the megaplasmid has occurred into chromosome 9. The hybridization of PbDT-3Ј, which is specific for pHH1TrxR-ko, with chromosome 9 proves that the knockout construct is present in the chromosome, and therefore homologous recombination must have occurred. To assign the chromosomes, the Yeast Chromosome PFG Marker and the data from Thompson et al. (53) were used. The abbreviations used are as follows: chr., chromosome; trxR, thioredoxin reductase; pGEM, pGEM vector backbone. weak band at 60 kDa, whereas a strong overexpression was found in clones 7D and 11E (Fig. 9). As a loading control antibodies raised against the 1-Cys peroxiredoxin from P. falciparum were used (14). Our results demonstrate that the integration of pHH1TrxR-3ЈR into the trxR gene locus has no deleterious effect on the expression of the protein. This result was not surprising, because we had already shown that the transfectants do not show any detectable phenotypic differences. Furthermore, using 1-Cys peroxiredoxin as a loading control in the Western blot experiment implies that the expression of proteins involved in the antioxidative defense mechanism and possibly interacting directly with the Trx system was not affected either. Most importantly the results presented here show that co-transfection of knockout and overexpression constructs is a possibility to complement loss of function of a lethal knockout in P. falciparum.

DISCUSSION
The thioredoxin redox system is involved in a variety of important cellular functions, which include the maintenance of the intracellular reducing environment, the reduction of ribonucleotides and hence DNA synthesis, and the regulation of transcription of certain genes by interaction with transcription factors (15)(16)(17)(18)(19)49). Furthermore, it has been shown that TrxR/ Trx are components of an efficient antioxidant system in the cell (7,12,13). Despite these important roles, a functional thioredoxin redox system is dispensable in bacteria and yeast, and it was shown that there appears to be a high degree of redundancy for almost all proteins of the thioredoxin redox cascade, as long as there is a second system present in the cell that can compensate for the loss of function of the Trx system (50). Knockout studies of the Trx system in mice and D. melanogaster impair embryonic and larval development, and it is presumed that the reduced capacity to adequately protect cells from cytotoxic damage is the cause for the deleterious effect (51,52).
Based on the finding that trxR null mutants were not viable, we postulate that in P. falciparum blood stage forms the Trx system is not dispensable and cannot be replaced by, for instance, the glutathione redox system despite their overlapping functions. Thus the level of redundancy between both systems appears to be not as profound as in yeast or bacterial systems and is more similar to the situation in higher eukaryotes.
In order to prove that the trxR locus was indeed targeted, a control experiment was performed that leads to the replacement of the 3Ј region of the trxR gene upon homologous recombination of the construct pHH1TrxR-3ЈR. Parasites obtained after integration of the control construct did not show any phenotypic alterations when compared with the wild type strain. To rescue trxR null mutants, a co-transfection of both pHH1-TrxR-ko and pHC1-TrxR was conducted. This was mainly done because currently there are only limited numbers of selectable markers available for the transfection of P. falciparum. Both constructs possess DHFR as selectable markers and confer resistance to pyrimethamine; however, the human DHFR also confers resistance to WR 99120 (31,32,36,37). Therefore, parasite clones were selected by using both drugs, and it was shown that indeed both plasmids reside in the trxR gene locus of clones 7D and 11E. These results indicate that the lethal effect of the trxR knockout can be compensated by overexpressing the gene from a recombinant source.
The genotype of the co-transfected parasites was investigated in more detail, and it was shown that the co-transfected plasmids form complexes via homologous recombination prior to integration into the trxR gene locus. The plasmid rescue as well as the subsequent Southern blot, PFGE, and FIGE experiments strongly support this hypothesis. It has been shown previously that recombination events lead to the formation of concatameric complexes (unstable replicating forms) between episomal copies of the constructs transfected into P. falciparum (47,48). It is presumed that transfected plasmids need to form unstable replicating forms to be maintained within the parasite, because they are essential for correct segregation to daughter cells during the erythrocytic cycle of the parasites FIG. 7. FIGE of co-transfected parasite clone 11E (Co-11E) in comparison with D10. DNA was digested with ClaI and separated by FIGE. The blot was hybridized with two different probes (trxR-ko fragment and PbDT-3Ј). The hybridizing fragment of Co-11E is at least 20 kb larger than the fragment of D10. This demonstrates that at least three copies of plasmids are integrated with one of them being the knockout construct as the probe PbDT-3Ј, which is specific for pHH1TrxR-ko, hybridizes with this fragment. Sizes were estimated using Yeast Chromosome PFG Marker and Lambda DNA-HindIII Digest Standards (New England Biolabs). The abbreviations used are as follows: trxR, thioredoxin reductase; pGEM, pGEM vector backbone.

FIG. 8. Northern blot analysis of transfectants.
Northern blot analysis of co-transfectants 7D (Co-7D) and 11E (Co-11E) in comparison with wild type D10 and trxR-3Ј parasites. The expression levels are increased severalfold in co-transfectants, which proves that the construct pHC1TrxR-Ex leads to overexpression of trxR mRNA in the co-transfected parasites and complements the lethal trxR knockout. rRNA was used as a loading control. The size of the mRNA was estimated using RNA Markers 0.28 -6.58 (Promega). TrxR-3Ј, thioredoxin reductase 3Ј-replacement parasite clone 3. FIG. 9. Western blot analysis. The Western blot analysis shows the higher expression of TrxR (60 kDa) in the co-transfectants (Co7D and Co11E) in comparison with D10 and TrxR-3ЈR parasites. The loading control represents 1-Cys peroxiredoxin (24 kDa) from P. falciparum and shows that similar amounts of protein were loaded in the different lanes of the gel. Furthermore, it shows that no alterations in the expression of this protein upon transfection of the parasites occur. To estimate the size of the protein, Kaleidoskop Prestained Standards (Bio-Rad) were used. ␣Px1, antibodies raised against 1-Cys peroxiredoxin; ␣TrxR, antibodies raised against thioredoxin reductase; TrxR-3Ј, thioredoxin reductase 3Ј-replacement parasite. (48). PFGE shows that both constructs were present on chromosome 9, and FIGE shows that at least three copies of the constructs were integrated into the trxR gene locus. The hypothesis that the concomitant transfection of the overexpression construct led to a compensation of trxR loss of function is shown by the normal growth of clones 7D and 11E.
Northern blot analyses confirmed that trxR was expressed at high levels, which is also reflected by the elevated protein levels found in those clones. The amplified expression on mRNA and protein level in comparison with wild type or control parasites is mainly attributable to the calmodulin promoter that drives the expression of the recombinantly introduced trxR gene from pHC1 (32,46).
In conclusion the results presented here show that a cotransfection of knockout and overexpression constructs is a suitable method to complement a lethal disruption of an essential gene in P. falciparum. By using this method our study validates P. falciparum thioredoxin reductase as a potential target for the development of new antimalarials and highlights the importance of functional antioxidant and redox systems for the survival of the erythrocytic stages of P. falciparum.