Plasmodium parasites possess two endosymbiotic organelles: a mitochondrion and a relict plastid called the apicoplast. To accommodate the translational requirements of these organelles in addition to its cytosolic translation apparatus, the parasite must maintain a supply of charged tRNA molecules in each of these compartments. In the present study we investigate how the parasite manages these translational requirements for charged tRNACys with only a single gene for CysRS (cysteinyl-tRNA synthetase). We demonstrate that the single PfCysRS (Plasmodium falciparum CysRS) transcript is alternatively spliced, and, using a combination of endogenous and heterologous tagging experiments in both P. falciparum and Toxoplasma gondii, we show that CysRS isoforms traffic to the cytosol and apicoplast. PfCysRS can recognize and charge the eukaryotic tRNACys encoded by the Plasmodium nucleus as well as the bacterial-type tRNA encoded by the apicoplast genome, albeit with a preference for the eukaryotic type cytosolic tRNA. The results of the present study indicate that apicomplexan parasites have lost their original plastidic cysteinyl-tRNA synthetase, and have replaced it with a dual-targeted eukaryotic type CysRS that recognizes plastid and nuclear tRNACys. Inhibitors of the Plasmodium dual-targeted CysRS would potentially offer a therapy capable of the desirable immediate effects on parasite growth as well as the irreversibility of inhibitors that disrupt apicoplast inheritance.
aaRSs (aminoacyl-tRNA synthetases) are essential for protein translation and ensure the fidelity of translation from nucleic acid to amino acid sequence. These enzymes catalyse the correct attachment of an amino acid to its cognate tRNA, forming aminoacyl-tRNAs that can be used as substrates for codon/anticodon recognition and amino acid polymerization by the ribosome . Recognition of the correct amino acid and cognate tRNA are crucial to the function of aaRSs. This specificity is achieved via two means: fidelity in substrate recognition and selective editing of incorrectly synthesized aminoacyl-tRNAs . To preferentially select the correct amino acid, recognition can be achieved by shape or by induced fit . In some cases discrimination is aided through the use of metal ions, such as a zinc ion for cysteine residue recognition . Additionally, the binding of the proper tRNA can also increase the correct selection of the amino acid, as seen for glutaminyl-tRNA synthetase in Escherichia coli .
In eukaryotes, translation occurs in mitochondria, plastids and the cytosol. Translation in each of these compartments requires aaRSs. With few exceptions, organellar aaRSs are encoded in the nuclear genome, translated in the cytosol and post-translationally targeted to their destination. Despite the considerable evolutionary distance between the bacteria that gave rise to mitochondria (α-proteobacteria) and plastids (cyanobacteria), a specific tRNA found in these endosymbiotic organelles is sometimes sufficiently similar to the corresponding tRNA in the other organelle that a dual-targeted aaRS can recognize and charge the tRNA in both organelles . Indeed, many organellar aaRSs in Arabidopsis are coded by a single nuclear gene whose gene products are shared between mitochondria and plastids [7,8]. Differences between nuclear (eukaryotic)- and organelle (bacterial)-type tRNA molecules are more profound, and fewer cases exist for aaRSs being shared between endosymbiotic organelles and the cytosol. Instances of sharing include an alternatively spliced IleRS (isoleucyl-tRNA synthetase) that is shared between the cytosol and mitochondrion in Trypanosoma , a GlnRS (glutaminyl-tRNA synthetase) shared between the Saccharomyces cerevisiae cytosol and mitochondria , an alternatively-spliced LysRS (lysyl-tRNA synthetase) shared between human cytosol and mitochondria , and several Arabidopsis thaliana aaRSs shared between the cytosol and organelles [7,12].
The causative agent of severe malaria, Plasmodium falciparum, has three translationally active compartments. In addition to translation in the cytosol and mitochondrion, the parasite possesses a plastid obtained through secondary endosymbiosis , referred to as the apicoplast. Like plants, Plasmodium encodes too few aaRSs to serve each compartment with its own set of 20 unique enzymes [14,15]. However, unlike plant chloroplasts, the apicoplast resides in the endomembrane system , so mechanisms for dual targeting identified in primary chloroplasts cannot be directly shared or co-inherited. Whereas the transit peptides of primary chloroplasts and mitochondria are rather similar, apicoplast targeting is mediated by a very different sequence: a bipartite leader consisting of a signal peptide followed by a transit peptide [16,17]. Hence, apicoplast trafficking sequences are very unlikely to function as mitochondrial import elements, or to allow easy partitioning between apicoplast and cytosol.
A number of Plasmodium aaRSs that are dual-targeted to the apicoplast and the cytosol have now been identified [18,19], and a handful of similar enzymes have been identified in the related parasite Toxoplasma gondii . However, the means of dual targeting is unclear in each case. P. falciparum possesses only one CysRS (cysteinyl-tRNA synthetase) (PfCysRS) and we show in the present study, using a series of stable transgenic parasites, that this enzyme is dual-targeted to the cytosol and apicoplast through alternative splicing to produce isoforms with and without bipartite apicoplast-targeting presequences. Although alternative splicing has been observed before in Plasmodium, to our knowledge this is the first demonstration of differential localization through alternative splicing in Plasmodium species, and the first report of alternative splicing mediating dual targeting to a complex plastid of secondary endosymbiotic origin. No mitochondrial fraction is observed for this protein, which confirms that aminoacylation of tRNAs to be used in mitochondria must occur outside the organelle in Plasmodium, by a mechanism not yet understood.
As noted above, aminoacylation of organellar and nuclear tRNAs presents a potential challenge, as the tRNAs resident in these compartments are separated by more than a billion years of divergence. We show through phylogenetic analysis that the PfCysRS is of eukaryotic rather than organellar origin, and that it has displaced the bacterial version in the apicoplast organelle. Despite this ancient divergence, we show that the eukaryotic-type PfCysRS is able to rescue an E. coli cysRS mutant, demonstrating that the enzyme can recognize and can charge bacterial tRNACys. In addition, using a specific in vitro aminoacylation assay, we confirm that recombinant PfCysRS charges its cognate cytosolic tRNACys, while also displaying an ability to cross charge both apicoplast and E. coli tRNACys, although the activity with the apicoplast tRNA was lower relative to the cytosolic tRNACys.
aaRSs in both the cytosol and the apicoplast of Plasmodium falciparum have been pursued as targets for inhibitors that could potentially be exploited to design novel antimalarial drugs [18,21,22]. Demonstration that the PfCysRS is dual-targeted and can cross charge both the cytosolic and the apicoplast tRNACys suggests that inhibition of this enzyme might achieve more profound growth perturbations through disruption of both translation machineries in the parasite.
MATERIALS AND METHODS
Characterization of the CysRS locus
The sequence of the annotated PfCysRS (PF3D7_1015200.1) was obtained from PlasmoDB  (http://www.plasmodb.org). Putative orthologues were retrieved using BLAST searches  of the GenBank® non-redundant  and orthoMCL  databases. Bioinformatic analysis of the N-terminal targeting sequence was conducted using SignalP v4.0 , apicoplast transit peptide predictors PATS version 1.2.1  and PlasmoAP , as well as manual inspection of residues. Parsimony and neighbour-joining distance trees were constructed using the PHYLIP tools  implemented at the mobile webserver . Maximum-likelihood trees were constructed using the PhyML algorithm .
PfCysRS transcripts were characterized using RNA isolated from P. falciparum cultures as described previously  and converted into cDNA according to the Omniscript Reverse Transcription kit protocol (Qiagen). Oligonucleotide pairs corresponding to multiple possible splice junctions were designed for the locus, and final amplification of the apicoplast and cytosolic PfCysRS isoforms was conducted using the primer pairs 5′-TGGAATTTTCTTGTATTTATCAAATTC-3′/3′-AAAGCTA-ACATATGTTCTTGCATGTCC-5′ and 5′-ATGAAAACATAC-TGTTACTAATATTAC-3′/3′-AAAGCTAACATATGTTCTTGC-ATGTCC-5′ respectively. PCR-amplification of genomic DNA by the oligonucleotide pairs was performed for comparison. Sequencing was conducted at the Applied Diagnostic Unit, The University of Melbourne, Melbourne, Australia.
Construction of parasite transfection vectors
To determine the subcellular localization of the two PfCysRS mRNA isoforms, two C-terminal GFP fusions were created using a pGlux.1 plasmid as described previously . P. falciparum cDNA was amplified to create truncated N-terminal regions of the two mRNA gene isoforms and introduced into XhoI/KpnI restriction sites of pGlux.1 using the primers 5′-CCGctcgag-CGGATGAATCTTTTCCTTTTTATATGTATAGC-3′/5′-CGGg-gtaccAAAGCTAACATATGTTCTTGCATGTCC-3′ for the longer predicted apicoplast targeting transcript, and 5′-CCGctcg-agCGGATGGATAATGCTAATAAATTACC-3′/5′-CGGggtaccA-AAGCTAACATATGTTCTTGCATGTCC-3′ for the predicted cytosolic targeting transcript (letters in lower case indicate the restriction sites).
To generate a construct for epitope tagging of the endogenous PfCysRS locus, we used a modified version of the pHA3 plasmid as described previously . A targeting sequence to allow recombination into the PfCysRS locus was created by cloning the terminal 3′-end 760 bp of the coding region in front of a triple HA (haemagluttinin) (3×HA) tag into the BglII and PstI sites, using the primers 5′-GAagatctAAGTAGGTATAATACAAAGTGGGC-3′/5′-TGCActgcagCCACTATTTTTATCAACCCATTAAGCA-3′ (letters in lower case indicate the restriction sites).
Plasmodium culture and transfection
Plasmodium falciparum parasites (3D7 strain) were maintained in O+ human erythrocytes (Australian Red Cross Blood Services, Melbourne, Australia) using a modification  of the method established by Trager and Jensen . Transfections of ring-stage P. falciparum parasites with 100 μg of purified plasmid DNA were performed as described previously  and were grown in the presence of 5 nM WR99210 to select transfectants. For 3′ integration vectors, plasmids were cycled on and off WR99210 to obtain stable recombinant parasites.
Toxoplasma culture and transfection
T. gondii ∆KU80 tachyzoites were kindly supplied by Dr Giel van Dooren (Research School of Biology, Australian National University, Canberra, Australia), and were cultured on human foreskin fibroblasts according to the method of Striepen and Soldati . A 1.4 kb fragment of the TgCysRS (T. gondii CysRS) 3′ end was amplified by PCR from genomic DNA using the primers TgCysRS-LIC SP (5′-TACTTCCAATCCAATTTAGC-AACGAGCAAAGGAACTGAAGG-3′) and TgCysRS-LIC ASP (5′-TCCTCCACTTCCAATTTTAGCAGTGTGTCCTTGAGA-CGCGA-3′) and inserted into the pLIC-HA3/DHFR plasmid (a gift from Dr Michael White, University of South Florida, Tampa, FL, U.S.A.) as described previously . Parasites were transfected as described in , with selection in 1 μM pyrimethamine, using resistance conferred by a T. gondii-specific mutation of dihydrofolate reductase . Parasites were cloned by limiting dilution as described in , before being analysed by Western blotting and immunofluorescence assay.
Live-cell imaging of the PfCysRS-A1–153GFP and PfCysRS-B1–70GFP transfectants was performed using a Zeiss Axioplan2 epifluorescence microscope as described previously . To determine the subcellular trafficking of PfCysRS-tagged transfectants, immunofluorescence assays were performed as described in . After fixation and permeabilization, samples were incubated in 3% BSA containing rat anti-HA (1:200 dilution, Roche), rabbit anti-ACP (acyl carrier protein) (1:500 dilution ), mouse anti-GFP (1:200 dilution, Roche) and rabbit anti-CysRS (1:500 dilution) primary antibodies. Alexa Fluor® 488- and Alexa Fluor® 594-conjugated secondary antibodies (1:1000 dilution, Molecular Probes) diluted with 3% BSA were used. Samples were incubated in combinations of the antibodies before mounting. Cells were visualized by epifluorescence microscopy, by 3D reconstruction using a DeltaVision Elite imaging system, and by confocal microscopy using a Leica SP5 laser-scanning confocal microscope. Localization experiments using a rabbit anti-CysRS antibody were performed with an apicoplast protein, PfRRF1 (PF3D7_0208600), with a GFP tag used as a marker for the apicoplast. To specifically visualize labelling inside the apicoplast, parasite membranes were differentially solubilized and cytosolic contents extracted as described previously .
A peptide corresponding to the 14 C-terminal amino acids of the PfCysRS protein (AGDQKKGNQNEKRE) conjugated to KLH (keyhole limpet haemocyanin) at the N-terminus of the peptide, was used to immunize rabbits for production of affinity-purified antibodies using a commercial antibody service (GenScript Corporation).
Western blot analysis
Western blots on parasite-derived protein were prepared as described previously  using rat anti-HA (1:500 dilution, Roche), mouse anti-GFP (1:500 dilution, Roche) and rabbit anti-CysRS (1:500 dilution) antibodies.
Generation of E. coli constructs and complementation experiments
The PfCysRS gene, without the 75-amino-acid N-terminal apicoplast-targeting sequence, was synthesized as a codon-optimized sequence for protein expression in E. coli (Mr Gene). We investigated the aminoacylation activity of PfCysRS by complementation of a temperature-sensitive mutant cysRS E. coli strain, UQ818  (provided by The E. coli Genetic Stock Center, Yale University, New Haven, CT, U.S.A.). The recombinant codon-optimized PfCysRS was subcloned into the pTrc99A plasmid (Pharmacia Biotech) with the restriction enzymes XbaI and SalI. Transformed UQ818 cells with empty pTrc99A and pTrc99A-PfCysRS were streaked on to LB agar plates supplemented with ampicillin (100 μg/ml) and nalidixic acid (50 μg/ml). Transformants were incubated at the permissive temperature of 30°C, and the non-permissive temperature of 42°C, for 48 h. As a control, growth curves of DH5α cells transformed with pTrc99A and pTc99A-PfCysRS were also analysed. Each strain was grown overnight in liquid culture (at 30°C) and a 10−3 dilution was used to inoculate 20 ml of LB media supplemented with ampicillin (100 μg/ml) and nalidixic acid (50 μg/ml). Aliquots of 1 ml were taken every 30 min to measure the D600 using a SmartSpec 3000 instrument (Bio-Rad Laboratories).
Generation of a CysRS expression vector, bacterial expression and purification
The synthetic PfCysRS gene (see above) was designed to include the desired restriction sites and was subcloned into the final plasmid pET21a (Invitrogen). The pET21a-PfCysRS clones were transformed into E. coli Rosetta 2 (DE3) pLysS cells. Transformants were grown in 1 litre of LB-ampicillin medium to a cell density (D600) of 0.6. Protein expression was induced with 1 mM IPTG and incubation was carried out at 37°C for 3 h. Cells were harvested and lysed with lysozyme (Sigma–Aldrich) in the presence of EDTA-free protease inhibitor cocktail (Roche). The cleared lysate was passed through a nickel-affinity column and recombinant PfCysRS was purified according to the manufacturer's protocol (Novagen). The purified histidine-tagged PfCysRS was judged to be of 95% purity by SDS gel after Coomassie Brilliant Blue staining.
Preparation of tRNA transcripts
Transcripts of cytosolic and apicoplast Plasmodium tRNACys were prepared and purified as described previously . The P. falciparum tRNACys transcript was synthesized with a 5′ extension that encoded a ribozyme to self-cleave the 5′ extension . The cleavage reaction occurred during transcription and the processed tRNA lacking the 5′ extension was purified by extraction from a denaturing gel. [35S]Cysteine was obtained from PerkinElmer. The tRNA transcripts were heat-denatured at 85°C for 3 min and annealed at 37°C for 20 min before use. These tRNAs, after denaturation by heat and annealing in the presence of 10 mM Mg2+, exhibited a plateau level of aminoacylation of approximately 20%. Steady-state aminoacylation assays were monitored at 37°C in 20 mM Tris/HCl (pH 7.5), 20 mM KCl, 10 mM MgCl2, 25 mM DTT, 50 μM cysteine and 2 mM ATP with 10–50 nM CysRS (depending on the activity towards different tRNAs) and varying tRNA concentrations (0.06–8 μM) to maintain steady-state conditions.
Assays for aminoacylation of tRNA with cysteine
The aminoacylation reaction was initiated by mixing the CysRS enzyme with various concentrations of an annealed tRNA transcript and [35S]cysteine (specific activity of 5200–6600 d..p.m/pmol). The reaction mixture was aliquoted, quenched with the CAM solution [0.24 M iodoacetic acid and 0.1 M sodium acetate in formamide (pH 5.0)] , and was further incubated for 30 min at room temperature (22°C). These samples were spotted on to filter papers and precipitated by 5% TCA (trichloroacetic acid), and the unbound [35S]cysteine was washed off from the paper pads. The concentration of aminoacylated tRNA at each time point was quantified by determining the radioactivity on each paper pad in a liquid scintillation counter (Beckman LS6000SC) and correcting for quenching effects. Steady-state kinetic parameters were determined by fitting initial aminoacylation rate against tRNA substrate concentration to the Michaelis–Menten equation.
RESULTS AND DISCUSSION
P. falciparum encodes a single eukaryotic-type CysRS
Bioinformatic analysis of the P. falciparum genome has revealed only a single copy of the CysRS gene (PF3D7_1015200.1) (Figure 1A). No other CysRS genes are encoded by either the apicoplast or the mitochondrial genomes . Alignment of the amino acid sequence of this gene (hereafter called PfCysRS) to other bacterial and eukaryotic CysRS sequences revealed an N-terminal extension, approximately 75 amino acid in length, not conserved in non-apicomplexan CysRSs. The characteristic KMSKS (Lys-Met-Ser-Lys-Ser) and HIGH (His-Ile-Gly-His) sequences for substrate binding  have been conserved in Plasmodium species. Unlike yeast and human CysRS , Plasmodium lacks the insertion domain found within the Rossmann fold (~100 amino acids). Interestingly, this insertion is not evenly distributed across all eukaryotes, but is well conserved in the opisthokonts. Previous analyses indicated that a C-terminal extension in the human CysRS, not found in bacteria, is involved in enhancing specificity for eukaryotic tRNA . This extension is also found in Plasmodium and other apicomplexan CysRSs (Figure 1B).
PfCysRS exists as two isoforms and is alternatively spliced
Phylogenetic analysis of CysRSs shows a strong alliance between enzymes responsible for plastid aminoacylation and bacterial sequences (Figure 1C). This suggests a bacterial origin for plastid CysRS congruent with the primary endosymbiotic origin of the plastid. Trees inferred using multiple methods place the CysRSs from plant and green algal chloroplasts with cyanobacteria, the bacterial ancestor of plastids, albeit with bootstrap values consistently lower than 50. The plastid CysRS from the diatom Thalassiosira is grouped with the plastid CysRSs from a red alga (Cyanidioschyzon), consistent with the known red algal origin of the diatom plastid (Figure 1C). Mitochondrial CysRSs (with the exception of the slime mould Dictyostelium) form a distinct clade, although these sequences do not group with the α-proteobacterial ancestors of mitochondria. Like bacterial sequences, mitochondrial and chloroplast CysRS sequences lack the C-terminal extension found in other eukaryotic CysRSs.
The P. falciparum CysRS forms a clade with other apicomplexan sequences, including the CysRS of Cryptosporidium parvum, which lacks an apicoplast and has no mitochondrial translation , so the Cryptosporidium CysRS must be a cytosolic enzyme. The apicomplexan clade is nested within the radiation of other cytosolic eukaryotic CysRSs. Together with the C-terminal eukaryote-specific extension, the phylogenetic analysis indicates a vertical eukaryotic (nuclear) ancestry for PfCysRS, and no evidence for an origin by endosymbiotic gene transfer that is seen for the plastid CysRSs. Diatoms, relatives of the Apicomplexa, have two CysRSs: one that groups with the cytosolic eukaryotic sequences, and a second CysRS with a bipartite plastid-targeting leader, which groups strongly with the CysRS from a red alga (Cyanidioschyzon) related to the ancestor of diatoms’ secondary endosymbiotic plastid (Figure 1C). A single secondary endosymbiosis is argued to have given rise to the plastids of Chromalveolata, which includes diatoms and Apicomplexa , so the ancestor of Apicomplexa could originally have had a plastid-targeted CysRS derived from the red algal endosymbiont, as well as a cytosolic enzyme related to the cytosolic CysRSs of other eukaryotes. No mitochondrial CysRSs from any chromalveolate have been characterized, so their evolutionary origin and fate is unknown.
PfCysRS is alternatively spliced to produce forms with and without a bipartite apicoplast-targeting leader
The single PfCysRS gene is annotated to contain three exons (Figure 1A), similar to the corresponding syntenic orthologues from all other annotated Plasmodium species. The initial exon consists almost entirely of coding sequence corresponding to a predicted bipartite apicoplast-targeting leader sequence with a signal peptide and transit peptide [29,53]. This predicted bipartite leader sequence stops at the start of the predicted second exon, at which point the protein sequence begins to exhibit a strong similarity to other CysRS sequences. This second exon also commences with a start codon with a preceding AAA nucleotide triplet characteristic of translation start sites in Plasmodium. RNA-Seq data from a whole genome analysis identifies a considerable pool of alternatively spliced transcripts that either lack all or part of the first exon, or which have an alternative splice event within the first intron. These alternative transcripts lack the complete open reading frame that commences with the methionine residue in the first annotated exon, and must therefore commence translation in the annotated second intron .
To confirm the existence of these alternatively spliced transcript forms we performed RT (reverse transcriptase)–PCR on mixed blood-stage P. falciparum parasites using a variety of different primer combinations to sequence the resulting cDNAs. Sequencing confirmed the existence of at least two transcript variants, one of which contained all three exons with the two introns spliced out as annotated in PlasmoDB v8.2 (GenBank® accession number KF056896), and the other (GenBank® accession number KF056897) showed only a single intron removed, whose open reading frame commenced with the start codon near the beginning of the annotated exon 2 (Figure 1A). These data confirm that this gene is alternatively spliced, producing alternative versions of the open reading frame, one commencing with the predicted bipartite apicoplast-targeting sequence, and the other lacking this putative trafficking information.
The exon structure of the CysRS gene is conserved between diverse Plasmodium species, and because all apicomplexans possess only a single CysRS protein and most possess an apicoplast, this suggests that similar alternative splicing is conserved throughout the genus. However, the gene structure is quite different in other apicomplexans, such as Toxoplasma and Theileria, so it is unclear if the mechanism of targeting via alternative splicing is shared for all apicomplexan CysRS enzymes.
The bipartite leader sequence is sufficient for targeting GFP to the apicoplast in P. falciparum
To test whether the bioinformatically predicted bipartite apicoplast leader acts as a true apicoplast-targeting sequence, we tested whether this sequence was sufficient to target a reporter protein to the apicoplast. The first 153 amino acids of the protein sequence, consisting of the predicted bipartite leader, was cloned from P. falciparum cDNA, and fused in-frame in front of GFP in an overexpression plasmid (pGlux.1) controlled by the heterogeneous PfCRT (CRT is chloroquine-resistance transporter) promoter. Parasites transfected with this construct (PfCyRS-A1–153GFP) were obtained approximately 26 days after transfection and were assayed by Western blotting and fluorescence microscopy. Live-cell imaging revealed the GFP signal for PfCyRS-A1–153GFP to be localized to a single organelle with morphology typical of apicoplast targeting (Figure 2A). Parasites had a single spherical punctum in early stages, which branched and divided in schizont stages before segregating into individual merozoites (results not shown). In immunofluorescence assays of PfCyRS-A1–153GFP, all of the GFP signal co-localized with a lumenal marker of the apicoplast, the ACP (Figure 2C). Western blot analysis confirmed overexpression of a protein of the expected mass for the fusion protein lacking the signal peptide and apicoplast transit peptide (Figure 2D). An additional form ~4 kDa larger was also observed after longer exposures (results not shown). These data are consistent with a model in which the reporter was imported to the apicoplast and the N-terminal extension removed in the apicoplast. This model is consistent with observations of nearly all native and exogenously apicoplast-targeted proteins (see, for example, [16,53]). These data confirmed targeting of this reporter (PfCyRS-A1–153GFP) to the apicoplast and confirm that the first exon of this gene is a true apicoplast-targeting leader sequence.
Localization of PfCysRS isoforms using GFP fusions in transgenic Plasmodium parasites
To test the targeting characteristics of protein commencing with the alternative translation start site, a GFP-fusion protein construct was also generated using an N-terminal fragment that started at the methionine residue located at the start of the annotated exon 2. A coding sequence corresponding to the 70 amino acids following this internal methionine residue was cloned from P. falciparum cDNA and this fragment was cloned into pGlux.1 as described above. Parasites were transfected and positive transfectants were established approximately 23 days later. Live-cell microscopy of this line (PfCyRS-B1–70GFP) demonstrated that this GFP-fusion protein was retained in the cytosol, with no apparent organellar accumulation (Figure 2B). These parasites were also assayed by Western blotting, which revealed tagging of a protein of the expected mass for the unprocessed protein fused to GFP (34.6 kDa). These data indicate that the internal alternative translation initiation site does not correspond to an organellar targeting sequence, consistent with the observation that protein initiating from this site is retained in the cytosol.
Endogenous PfCysRS tagged with a HA epitope is targeted to the cytosol and apicoplast: a model of dual targeting
Despite the use of the deliberately weak PfCRT promoter, episomally expressed transgenes in the pGlux.1 expression system may not faithfully reflect native protein localization. Such episomal transfectants also lack chromosomal context that may be necessary to recapitulate true splicing events. Subcellular localization of the native PfCysRS was further studied using an affinity-purified rabbit antibody raised against a peptide from PfCysRS. Immunofluorescence assays using these antisera also showed that the major fraction of the PfCysRS was localized in the cytosol, but also identified a further minor fraction in the apicoplast (Figure 3A). Western blot analysis showed specific recognition of the ~69 kDa protein, consistent with the predicted mass of both the cytosolic protein and the truncated apicoplast protein (Figure 3B).
Co-localization of CysRS by immunofluorescence assays in P. falciparum
We also followed the targeting of the endogenous copy of PfCysRS by introducing the sequence of a triple haemagluttinin tag (3×HA) into the 3′ end of the gene using the pHA3 recombination vector. Parasites were recovered after 15 days, then subsequently cycled on and off selective agent (WR99210) to select for parasites that had integrated the plasmid into the PfCysRS locus. Stable transgenic integrants were obtained after the third cycle and were assessed by Western blot analysis and immunofluorescence assays. Western blot analysis revealed that the HA tag had integrated into the expected gene, producing a band at ~72 kDa (Figure 3B). This is congruent with a protein with translation initiation at the internal methionine residue with the added 3×HA tag (73 kDa). However, this would also be consistent with a processed product that initiated at the first methionine residue and had the bipartite apicoplast leader removed, which could be between ~72 and 76 kDa.
The PfCysRS–HA-tagged parasites were used to perform immunofluorescence assays, which revealed that most of the protein was localized to the parasite cytosol, whereas some PfCysRS–HA co-localized with the ACP apicoplast marker (Figure 3C). However, because of the strong cytosolic background labelling of the PfCysRS–HA, and the small diameter of the apicoplast (~200 nm), it was difficult to be sure that this was not fluorescence from protein just outside of the apicoplast. Analysis of labelling using confocal microscopy (Figure 3D) and image deconvolution (Supplementary Movie S1 at http://www.biochemj.org/bj/458/bj4580513add.htm) were congruent with co-localization, but did not prove this explicitly. We addressed this uncertainty in apicoplast localization through a differential permeabilization approach previously used for apicoplast labelling . This demonstrated that when the cytosolic fraction of PfCysRS–HA was extracted, a specific punctum remained, co-localized with the apicoplast marker ACP. The results indicate that for this tagged native gene product, a major isoform exists in the parasite cytosol, but a minor isoform also exists in the apicoplast lumen (Figure 3E). No such apicoplast fraction remained using the GFP-tagged truncated cytosolic isoform (results not shown).
The single TgCysRS has previously been investigated using reporter genes not in the native genomic context and using exogenous promoters . We investigated trafficking of this protein from the endogenous locus to more faithfully recapitulate native splicing and expression of the gene. We introduced a 3×HA tag into the 3′ terminus of the single chromosomal TgCysRS gene in a ΔKU80 T. gondii strain. Only a single protein species was present at ~75 kDa (Supplementary Figure S1A at http://www.biochemj.org/bj/458/bj4580513add.htm). This is smaller than the predicted mass from the gene model based on RNA-Seq data (~90 kDa) and our sequence of cDNA obtained derived from a 5′ RACE pool (GenBank® accession number KF056898). This discrepancy suggests translation from an internal methionine residue or alternative splicing as for the Plasmodium orthologue, although we found no alternative transcripts by 5′ RACE. We attempted to perform solubilization on these transfectants to further investigate the apicoplast isoform, but were unable to find any conditions that gave clean differentiation between plastid and cytosolic fractions [55,56]. Immunolocalization of this HA-tagged TgCysRS showed a similar pattern to the Plasmodium orthologue, with a dominant cytosolic localization, as well as some signal overlapping with the Cpn60 apicoplast marker  (Supplementary Figure S1B), although the small size of the organelle means that we cannot distinguish whether this overlap is specific apicoplast localization, or spatially unresolved fluorescence from the surrounding cytosol.
Taken together, the results of the present study show that the PfCysRS is alternatively spliced to produce isoforms that are respectively targeted to the apicoplast and cytosol. Targeting of individual proteins to multiple organelles has also been observed for several other P. falciparum proteins [18,58,59], although a molecular mechanism for targeting to multiple destinations has not been directly demonstrated in any of these cases.
PfCysRS complements an E. coli cysRS temperature-sensitive mutant
The E. coli K-12 strain UQ818 is a temperature-sensitive mutant isolated as a spontaneous mutant that lacks CysRS activity above the permissive temperature of 40°C . The CysRS activity in this mutant is defective at non-permissive temperatures due to a single amino acid change (V27E) that disrupts cysteine binding . A truncated version of the PfCysRS lacking the apicoplast-targeting sequence was expressed in UQ818 and tested for growth in both liquid and solid media. The UQ818 mutant transformed with empty pTrc99a vector was able to grow at the permissive temperature of 37°C (Figure 4A), but showed no growth at 42°C (Figure 4B). The PfCysRS-containing vector restored growth at the restrictive 42°C on solid media (Figure 4B) and in liquid media (results not shown). Expression in a CysRS-wt (wild-type) DH5α strain revealed no growth differences. Complementation of the cysRS defect in UQ818 demonstrates that the PfCysRS has aminoacylation activity and is able to recognize E. coli tRNACys.
Complementation of a temperature-sensitive cysRS E. coli strain with PfCysRS
Analysis of the cytosolic and apicoplast P. falciparum tRNACys molecules
Sequence analysis of P. falciparum tRNACys revealed that although the cytosolic version possesses all of the canonical features as in E. coli tRNACys (Figure 5A), the apicoplast version possesses unusual features (Figure 5B). Although all three tRNAs contain U73 as the discriminator base and the 5′-GCA sequence as the anticodon, which are major determinants for aminoacylation with cysteine , the apicoplast tRNACys is distinct in the following ways. First, it has A1–U72 and A2–U71 as the first two base pairs in the acceptor stem, which differ from the canonical G1–C72 and G2–C71 base pairs (Figure 5B). The higher A–U content in the acceptor end may confer unusual conformations to the CCA terminus . Secondly, it has an A15–U48 tertiary base pair, in contrast with the more common G15–C48 base pair as in the cytosolic version (Figure 5B). Interestingly, E. coli tRNACys has an unusual G15–G48 base pair (Figure 5C), which modulates the activity of aminoacylation [45,62]. Thirdly, the apicoplast tRNACys contains A19, rather than the conserved G19, in the D loop that forms a tertiary base pair with C56 in the T loop in the canonical structure (Figure 5B). The possession of A19 may permit the formation of a novel A19–U56 base pair in the apicoplast form. Other unusual features of the apicoplast tRNACys are the anticodon stem, made up of entirely A–U base pairs, and the use of an A30–U40 base pair in the anticodon stem, in contrast with the more common G30–C40 sequence.
Cloverleaf structures of tRNACys
Enzyme kinetics of P. falciparum CysRS
To evaluate the activity of PfCysRS for aminoacylation of the cytosolic and apicoplast versions of tRNACys, we carried out steady-state kinetics analysis using purified recombinant PfCysRS enzyme and tRNA transcripts. The aminoacylation activity for each tRNA was examined by steady-state kinetic analysis under conditions where tRNA was in excess of the enzyme so the synthesis of cysteinyl-tRNACys was linear with time (Supplementary Figure S2). Each tRNA transcript was prepared by transcription using T7 RNA polymerase and heat-cooled annealed to serve as the substrate for aminoacylation. Plateau levels of aminoacylation for cytosolic and apicoplast tRNACys were at 19 and 11% in extended time courses respectively. We also tested E. coli tRNACys, which has been well characterized [4,63,65], as a substrate for comparison.
The kcat and Km (tRNA) values for each tRNA were measured from fitting the data of the initial rate of aminoacylation as a function of tRNA concentration to the Michaelis–Menten equation (Figure 6). These data are summarized in Table 1. The Km (tRNA, 0.25±0.05 μM), kcat (0.24±0.01 s−1) and kcat/Km (0.96±0.6 μM−1·s−1) values of the PfCysRS enzyme for the cytosolic tRNA were closely similar to known values of CysRS, where Km is typically ~1 μM, kcat is ~1 s−1 and kcat/Km is ~1 μM−1·s−1 , indicating that the recombinant enzyme was active and that it exhibited a normal activity with the cytosolic tRNA. The recombinant enzyme also efficiently aminoacylated the transcript of E. coli tRNACys, with kinetic parameters equivalent to those of the cytosolic tRNA, but with an overall 3.4-fold decrease in the catalytic efficiency of aminoacylation kcat/Km. This indicates that the PfCysRS enzyme is similar to eukaryotic enzymes (e.g. human CysRS), which readily accommodate E. coli tRNACys [50,66,67]. The quality of transcription was verified using E. coli CysRS as a control to show that aminoacylation of E. coli tRNACys by the homologous E. coli enzyme exhibited normal kinetic parameters (Table 1).
|tRNACys||Km (tRNA) (μM)||kcat (s−1)||Kcat/Km (μM−1·s−1)||Ratio|
|P. falciparum tRNACys, cytosolic||0.25±0.05||0.24±0.01||0.96±0.29||1.0|
|P. falciparum tRNACys, apicoplast||1.45±0.13||0.086±0.003||0.060±0.005||1:16|
|E. coli tRNACys||0.65±0.15||0.19±0.01||0.28±0.04||1:3.4|
|(b) E. coli CysRS|
|Km (tRNA) (μM)||kcat (s−1)||Kcat/Km (μM−1·s−1)||Ratio|
|E. coli tRNACys||1.2±0.01||1.2±0.07||1.1±0.07||1:3.9|
|tRNACys||Km (tRNA) (μM)||kcat (s−1)||Kcat/Km (μM−1·s−1)||Ratio|
|P. falciparum tRNACys, cytosolic||0.25±0.05||0.24±0.01||0.96±0.29||1.0|
|P. falciparum tRNACys, apicoplast||1.45±0.13||0.086±0.003||0.060±0.005||1:16|
|E. coli tRNACys||0.65±0.15||0.19±0.01||0.28±0.04||1:3.4|
|(b) E. coli CysRS|
|Km (tRNA) (μM)||kcat (s−1)||Kcat/Km (μM−1·s−1)||Ratio|
|E. coli tRNACys||1.2±0.01||1.2±0.07||1.1±0.07||1:3.9|
Steady-state kinetic parameters of Plasmodium and E. coli tRNACys
Interestingly, aminoacylation of the apicoplast tRNACys by PfCysRS was less efficient by 16-fold as compared with the cytosolic tRNA (Table 1). The defect was found in both the Km for tRNA, showing an increase to 1.45±0.13 μM, and the kcat of aminoacylation, showing a decrease to 0.086±0.003 s−1, with an overall decrease in kcat/Km to 0.06±0.005 μM−1·s−1.
Together, these aminoacylation activity results confirm that the PfCysRS is able to recognize and charge the tRNACys molecules encoded by the nucleus and the apicoplast. Although the cytosolic tRNACys is efficiently recognized by PfCysRS (Figure 5 and Table 1), the apicoplast tRNACys is recognized less efficiently (Table 1). The discrimination in aminoacylation of the apicoplast tRNA by PfCysRS may reflect the presence of unusual features in the apicoplast tRNA, or the decreased aminoacylation activity could be related to unknown base modifications that may exist in the apicoplast tRNA and are not recapitulated in our heterologous transcription system. We note that the apicoplast encodes a mere 30 proteins in comparison with over 5000 by the nucleus. Furthermore, apicoplast-encoded proteins are translated in modest amounts; none are readily detectable without resorting to 14C/35S amino acid labelling [68,69]. The requirements for tRNA charging and turnover in the apicoplast, although absolute, are therefore likely to be very minimal, such that a PfCysRS with weak activity for the apicoplast tRNA might suffice.
In the present study we show that PfCysRS mRNA is alternatively spliced, producing transcript versions with and without apicoplast trafficking sequences, resulting in accumulation of protein in both the apicoplast and the cytosol (Figures 2 and 3). This produces enzymes that recognize and charge both nuclear and apicoplast tRNACys. The presence of an aaRS that is likely to be essential for translation in all compartments of Plasmodium presents an attractive drug target. aaRSs are being keenly pursued as drug targets in many organisms, with several compounds shown to be selective inhibitors targeting aaRSs in Plasmodium [70–72]. Inhibition of Plasmodium CysRS would probably have similar actions as borrelidin, an inhibitor of the dual-targeted Plasmodium threonyl-tRNA synthetase, which leads to widespread translational arrest and death [18,73]. In contrast, antimalarial drugs that affect apicoplast housekeeping processes often result in a ‘delayed death’ phenomenon whereby parasites die only in the subsequent proliferative cycle after drug treatment [74,75]. It is conceivable that compounds inhibiting both apicoplast and cytosolic translation may have a double effect, killing some parasites through inhibition of cytosolic translation, but additionally disrupting the apicoplast so that escapees of the first effect do not proliferate. This dual action may therefore act to slow drug resistance. Given these qualities, inhibitors of the dual-targeted CysRS should possess rapid antimalarial activity, thereby combining the advantages of rapid parasite death with a delayed disruption of organellar division and segregation that could ‘mop up’ survivors of the first mode of action.
James Pham, Reiko Sakaguchi, Nilushi De Silva, Lee Yeoh and Stuart Ralph designed and performed experiments, analysed data and wrote the paper. Ya-Ming Hou and Geoffrey McFadden designed experiments, analysed data and wrote the paper.
We thank Giel G. van Dooren (Research School of Biology, Australian National University, Canberra, Australia) for his reagents and helpful discussions. The Australian Red Cross Blood Service kindly supplied blood.
This work was supported by the Australian Research Council [grant numbers FT0990350, DP130103236], the Australian NHMRC (National Health and Medical Research Council) [grant numbers 520700, 628704 and 637406 (to S.A.R.)] and an NHMRC program grant (to G.I.M.). J.S.P. and L.M.Y. are supported by Australian Postgraduate Awards. The research stay of J.S.P. with the Hou laboratory at Thomas Jefferson University, Philadelphia, PA, U.S.A. was supported by a J.D. Smyth Postgraduate Student Travel Award from The Australian Society for Parasitology.
Present address: Department of Microbiology and Immunology, Columbia University Medical Center, New York, NY 10032, U.S.A.