Disruption of the Plasmodium falciparum PfPMT Gene Results in a Complete Loss of Phosphatidylcholine Biosynthesis via the Serine-Decarboxylase-Phosphoethanolamine-Methyltransferase Pathway and Severe Growth and Survival Defects*

Biochemical studies in the human malaria parasite, Plasmodium falciparum, indicated that in addition to the pathway for synthesis of phosphatidylcholine from choline (CDP-choline pathway), the parasite synthesizes this major membrane phospholipid via an alternative pathway named the serine-decarboxylase-phosphoethanolamine-methyltransferase (SDPM) pathway using host serine and ethanolamine as precursors. However, the role the transmethylation of phosphatidylethanolamine plays in the biosynthesis of phosphatidylcholine and the importance of the SDPM pathway in the parasite's growth and survival remain unknown. Here, we provide genetic evidence that knock-out of the PfPMT gene encoding the phosphoethanolamine methyltransferase enzyme completely abrogates the biosynthesis of phosphatidylcholine via the SDPM pathway. Lipid analysis in knock-out parasites revealed that unlike in mammalian and yeast cells, methylation of phosphatidylethanolamine to phosphatidylcholine does not occur in P. falciparum, thus making the SDPM and CDP-choline pathways the only routes for phosphatidylcholine biosynthesis in this organism. Interestingly, loss of PfPMT resulted in significant defects in parasite growth, multiplication, and viability, suggesting that this gene plays an important role in the pathogenesis of intraerythrocytic Plasmodium parasites.

The protozoan parasite Plasmodium falciparum is the causative agent of the most severe form of human malaria, resulting in 300 -500 million cases and 1-3 million deaths per year (1). The pathogenic stages of P. falciparum are those that invade mature erythrocytes, which are devoid of internal organelles and incapable of de novo lipid biosynthesis. Despite this, P. fal-ciparum undergoes dramatic morphological and metabolic developmental changes and asexually divides to form up to 36 new daughter cells within 48 h of invading a human erythrocyte (2). This rapid multiplication of P. falciparum within host erythrocytes entails the active production of new plasma membranes in which phospholipids are major architectural and functional components. Phosphatidylcholine (PtdCho) 2 and phosphatidylethanolamine (PtdEtn), comprise 40 -50% and 35-45%, respectively, of the parasite's total plasma membrane phospholipid content (3).
Genome data predict that P. falciparum possesses enzymatic pathways for the synthesis of all the necessary phospholipids from precursors transported from host milieu, such as serine, choline, inositol, glycerol, and fatty acids (3,4). The synthesis of PtdCho in P. falciparum takes place via two metabolic pathways, the CDP-choline pathway (Kennedy pathway) and the serine-decarboxylase-phosphoethanolamine methyltransferase (SDPM) pathway (Fig. 1). In the SDPM pathway, ethanolamine (formed by the decarboxylation of serine by a parasite serine decarboxylase) is converted into phosphoethanolamine. The latter serves as a precursor for the synthesis of PtdEtn through the CDP-ethanolamine pathway and for the synthesis of PtdCho through the SDPM/CDP-choline pathways. Serine is readily available in the parasite cytosol due to the active degradation of host hemoglobin as well as uptake from the host milieu (5,6).
The PfPMT gene in P. falciparum encodes the phosphoethanolamine methyltransferase that specifically methylates phosphoethanolamine to phosphocholine (p-Cho) via the SDPM pathway (7-9) (Fig. 1). p-Cho is then utilized by the CDP-choline pathway to synthesize PtdCho. The CDP-choline pathway also utilizes p-Cho synthesized from choline. PfPMT is a member of the PEAMT family of phosphoethanolamine methyltransferases that are also found in worms, plants, and other protozoa (7)(8)(9). The restricted phylogenetic distribution of this class of enzymes suggests that they could be potential targets for the development of drugs to treat malaria and other parasitic diseases.
The physiological role PfPMT plays in parasite membrane lipid biogenesis and in the development and survival of intraerythrocytic P. falciparum is not fully understood. Here, we provide evidence that disruption of the PfPMT gene in P. falciparum completely abrogates the biosynthesis of PtdCho from serine and results in severe parasite growth and survival defects. Our findings show for the first time that the in vivo methylation of PtdEtn to PtdCho does not take place in P. falciparum, thus making the SDPM and CDP-choline pathways the main routes for PtdCho biosynthesis. Furthermore, the data define a novel role for PfPMT in the proliferation and survival of intraerythrocytic malaria parasites.

EXPERIMENTAL PROCEDURES
Plasmid Construction and Parasite Transfection-The plasmid pRZ-TK-BSD2 (10) was used to construct the targeting vector pW-pfpmt⌬ (Fig. 2). The pRZ-TK-BSD2 plasmid encompasses the positive selectable marker BSD (blasticidin S deaminase) (11) from Aspergillus terreus that confers resistance to blasticidin and whose expression in P. falciparum is under the regulatory control of the PcDT (Plasmodium chabaudi DHFR/ TS) promoter. The plasmid also harbors the negative marker TK (thymidine kinase) from herpes simplex that confers sensitivity to ganciclovir and whose expression is under the regulatory control of the P. falciparum CAM promoter. To construct the targeting vector pW-pfpmt⌬ for PfPMT gene disruption, a 601-bp fragment (nucleotides 18 -619 of the unspliced PfPMT gene sequence) (MAL13P1.214 PlasmoDB) was amplified by PCR and cloned at the HindIII/BlpI site upstream of the PcDT promoter in the pRZ-TK-BSD2 plasmid. A second PCR was used to amplify a 505-bp fragment (nucleotides 650 -1155 of the unspliced PfPMT gene sequence) for directional cloning at the EcoRI/NarI site downstream of the HrpII terminator in the pRZ-TK-BSD2 plasmid. P. falciparum strain 3D7 was cultured in human red blood cells (RBCs) by the method of Trager and Jensen (12). Parasite transfection was done as previously described (13). After transfection, the cultures were maintained for an initial 48 h without blasticidin, and, thereafter, the drug was introduced in the cultures at 2.5 g/ml final concentration. Cultures were continuously supplemented with 200 M choline. The overall knock-out strategy ( Fig. 2A) involved a twostep process by which the chromosomal PfPMT locus was first disrupted with the BSD cassette, after double cross-over homologous recombination, followed by elimination of the episome by selecting against expression of the TK gene with ganciclovir, a subversive substrate of the TK enzyme. To avoid loss of the knock-out parasites within the population (in the event that the SDPM pathway is an essential route for PtdCho biosynthesis), transfected parasites were continuously cultured in the presence of 200 M choline, a concentration of choline ϳ20-fold greater than that present in human serum (14,15). Transgenic parasites, selected on blasticidin and ganciclovir, were cloned by limiting dilution and genomic DNA from cloned parasites purified for molecular analysis using PCR and Southern and Western blotting to confirm the disruption of the PfPMT gene. PfPMT gene disruption at the chromosomal locus was analyzed using the primer pair PfPMT-F (5Ј-ATGACTTTG-ATTGAAAACTTAAACTCTG-3Ј) and PfPMT-R (5Ј-TTTGG-TGGCCTTAAAATAACCCCATCTTTGCA-3Ј).
A PfPMT add-back vector for pfpmt⌬ complementation was generated by PCR amplification of the 801-bp full coding sequence of PfPMT from P. falciparum total cDNA. The fragment was directionally cloned at the XhoI site of a pHC1 plasmid downstream of the P. falciparum CAM promoter. The pHC1 vector harbors a Toxoplasma gondii dehydrofolate reductase gene that confers resistance to the selection drug, Pyrimethamine (16).
Southern Hybridization-Genomic DNA was extracted from wild type and transfected parasites, as previously described (11). About 2.5 g of the genomic DNA was digested with HindIII, separated on a 1% agarose gel, and blotted onto a Nytran SuperCharge nylon membrane (Whatman Schleicher and Schuell). Membranes were probed with [␣-32 P]dCTP-labeled PfPMT and BSD gene probes.
Western Blot Analysis-Parasites were extracted from infected erythrocytes by treatment with 0.15% saponin and sonicated in PBS. The soluble fraction was separated on a 12% SDS-polyacrylamide gel and transferred to a nitrocellulose membrane. Immunoblotting was done using affinity-purified antibodies against PfPMT and the P. falciparum translation elongation factor-1␣ (PfEF-1␣) as previously described (17).
Reverse Transcriptase-PCR-RNA was extracted from asynchronous parasites by TRIzol reagent (Invitrogen). Exactly 1 g of total RNA was treated with DNase I (Invitrogen) to remove any contaminating DNA, followed by reverse transcription. PfPMT was amplified from total cDNA using the primer pair PfPMT-F and PfPMT-R. As a loading control, the cDNA of the PfTPXI gene encoding the P. falciparum thioredoxin peroxidase (PF14_0368 in PlasmoDB 4.4) was amplified.
PfPMT Enzyme Activity Assay-Synchronous parasites were grown to 12% parasitemia and harvested at the trophozoite stage. Total extracts were prepared following sonication in extraction buffer (100 mM Hepes-KOH, pH 7.8, 5 mM dithiothreitol, 2 mM Na 2 EDTA, 10% glycerol (v/v)) and centrifugation. The reaction mixture contained 100 mM Hepes-KOH, pH 8.6, 2 mM EDTA, 10% glycerol, 2 mM S-adenosyl-L-methionine (AdoMet), 1 mM phosphoethanolamine, 80 M [methyl-14 C]AdoMet (400 nCi), and 50 g of parasite protein extract in a total reaction volume of 100 l. A blank reaction mixture lacked protein extract. The reaction mixtures were incubated for 30 min at 30°C and terminated by the addition of 1 ml of ice-cold water. The products of the methylation reaction were purified through a AG (H ϩ ) resin, following the method of Nuccio et al. (18) and quantified by liquid scintillation counting.
Labeling Assays and Phospholipid Analysis-Synchronized parasites cultured in medium without choline at 2% hematocrit were grown to 10% parasitemia at the early trophozite stage. The parasitized erythrocytes were resuspended in fresh medium containing either 1.6 Ci of [ 14 C]ethanolamine hydrochloride (51 mCi/mmol) or 0.1 Ci of [ 14 C]choline chloride (57 mCi/mmol) and cultured for 12 h. The infected erythrocytes were washed twice with PBS, and the pellet from a 60-ml culture volume was resuspended in a 20ϫ volume of chloroform/ methanol (2:1, v/v). Lipids were extracted by the Folch method (19), and the organic phase of the extract was evaporated and redissolved in 600 l of chloroform/methanol (2:1). The organic extract (150 l) was fractionated by two-dimensional TLC on Silica Gel-60 20 ϫ 20-cm plates (Merck). For the first dimension, a solvent containing chloroform/methanol/ammonium hydroxide (84.5:45:6.5) was used. The second solvent was composed of chloroform/glacial acetic acid/methanol/water (90:30:6:2.6).
Parasite Growth Assays-Synchronized parasites at early ring stage were seeded at 2% parasitemia and cultured in the presence of 0, 10, 50, 100, 200, or 500 M choline chloride for 96 h. Giemsa smears of the cultures were prepared at time points of 0, 12, 18, 32, 42, 56, and 96 h of culture, and the parasite counts and morphological state were determined by light microscopy.
Terminal Deoxynucleotidyltransferase (TdT)-mediated dUTP Nick End Labeling (TUNEL) Assay-The TUNEL kit (Roche Applied Science) was used to monitor cell survival following the manufacturer's instructions, with some modifications. Briefly, asynchronous cultures were grown to 10% parasitemia, and 1 ml of the culture aliquots was washed twice with PBS, and the cell number was adjusted to 1 ϫ 10 6 cells/ml. Fixation was achieved by resuspending the cells in 2% paraformaldehyde solution in PBS and incubated at room temperature for 1 h. The fixed cells were washed twice with PBS, resuspended in 100 l of 0.1% Triton X-100 in 0.1% sodium citrate solution, and incubated for 2 min on ice. Cells were then washed twice in PBS and kept on ice until labeling. A positive control sample was prepared by incubating fixed and permeabilized wild type (3D7) cells with 3000 units/ml of DNase I for 15 min at room temperature, followed by two washes with phosphate-buffered saline. To label the cells, 50 l of the TUNEL reaction mixture was added to the cell pellet and mixed gently. A negative control sample was included by treating fixed, permeabilized wild type (3D7) cells with the labeling solution in the absence of the terminal TdT enzyme. The reaction mixtures were incubated in the dark at 37°C for 1 h and then washed twice in PBS and resuspended in 250 l of PBS. Cells were analyzed by fluorescence microscopy at 525 nm with the fluorescein isothiocyanate filter. To confirm that TUNEL staining correspond to cell death, parasites were grown in medium containing either 100 nM chloroquine or 100 nM pyrimethamine for 24 h and then processed for the TUNEL assay as described above.
Nuclei Counting by Spinning Disc Confocal Microscopy (SDCM)-Counting of nuclei by SDCM was done essentially as previously described for other strains of P. falciparum (20). In brief, all cultures were first synchronized three times by 5% D-sorbitol treatment, since multiple synchronization treatments successively improve the ring/early trophozoite ratio (20). Data were then routinely obtained 35-38 h after the last synchronization step, wherein segmented schizonts are clearly visible by light microscopy. We have previously described our customized SDCM apparatus in detail (20,21). This instrument acquires "z stacks" for live cells at ϳ210 -220 nm resolution (x, y, and z) in less than 1 s, which eliminates blurring due to parasite movement within the infected RBC. Along with reduced photobleaching and improved deconvolution procedures (20,21), the method provides a reliable and convenient way to count nuclei for live intraerythrocytic schizonts using a Merzhauser motorized MS-2000 XY translation stage and an additional piezo table-optimized z-movement (5 mm/s over a range of 100 m). Oil was DF-type (n ϭ 1.515, low background fluorescence), and the camera was a Hamamatsu ORCA ER-cooled CCD with 1.3-megapixel full frame and 8.9 frames/s full rate. Excitation of SybrGreen-labeled nuclei (20) was with a Coherent Innova 300 I argon/krypton laser (300 milliwatts at 488 nm). Exposure was for 100 ms at 100 milliwatts laser power, and z-spacing was 200 nm (appropriate for iterative deconvolution as in Ref. 20. After iterative deconvolution using experimentally derived point spread functions, the x, y resolution of fluorescence SDCM data is only slightly lower than that of classical LSCM confocal data (measured on our instrument to be 213 nm using 520 nm light). z stack data were transferred to a Dell mini tower customized with three 750-Gb RAID array hard drives, and restoration was done using an Imaris 5.0.1/Auto-Quant X software package from Bitplane Inc. (Saint Paul, MN), as described (20). z-Series of optical sections were deconvolved using the MLE method and a fixed point spread function routine with 15 iterations. Point spread function were obtained by mixing subresolution (d ϭ 0.17 m) fluorescent beads in cell culture. Restored images were transferred to Imaris and sorted into freely rotating three-dimensional objects in the "Surpass" mode, and nuclei counts were done with the aid of the "Spots" routine, which locates fluorescence peaks in three-dimensional space at operator-defined contrast half-width and intensity.
Assignment of three-dimensional peaks as due to nuclei was confirmed by eye and then done in a semiautomated fashion to average across Ͼ200 schizonts in each case. Counts were exported to Excel or SigmaPlot software 9.0 for further statistical analysis.

RESULTS
Genetic Disruption of the PfPMT Gene in P. falciparum-To assess the physiological role of PfPMT in P. falciparum, we disrupted the PfPMT gene by homologous recombination. PCR analysis of genomic DNA from the cloned transgenic parasites demonstrated the replacement of the PfPMT gene by the BSDcontaining targeting cassette in 10 individual clones. Southern blotting of HindIII-digested genomic DNA isolated from one clone (Fig. 2B) and using a 601-bp 5Ј-end PfPMT gene fragment probe (identical in sequence to that cloned in the pW-pfpmt⌬ vector) showed the presence of 5.6-and 9.3-kb fragments in the pfpmt⌬ and wild type parasites, respectively, confirming the disruption of the PfPMT gene locus in the pfpmt⌬ parasites (Fig. 2, B and C). Additionally, probing with the BSD gene fragment revealed the expected 5.4-kb fragment in the pfpmt⌬ parasites but not in the wild type (Fig. 2C), further confirming that gene replacement had occurred.
To test the loss of PfPMT expression in the selected pfpmt⌬ clone, reverse transcription-PCR analysis was performed using primers within the PfPMT open reading frame. PfPMT cDNA could be detected using wild type but not pfpmt⌬ RNA (Fig.  2D). As a control, the expression of the PfTPX1 gene was detected in both wild type and knock-out parasites. Analysis of the loss of PfPMT expression by immunoblotting using affinitypurified anti-PfPMT antibodies (17) showed that, although a 30 kDa PfPMT band was detectable in the wild type strain, there was no signal in the knock-out strain (Fig. 2E). As a loading control, antibodies raised against the parasite translation elongation factor, PfEF-1␣ (22), were used, and the expected ϳ49-kDa PfEF-1␣ band was found in both the wild type and pfpmt⌬ strains (Fig. 2E).
pfpmt⌬ Parasites Lack PMT Activity and Cannot Synthesize PtdCho From Ethanolamine-To determine whether PfPMT encodes the only phosphoethanolamine methyltransferase activity of P. falciparum, the three-step methylation of phosphoethanolamine to phosphocholine and the intermediate products was examined in protein extracts from either wild type or pfpmt⌬ parasites using radiolabeled AdoMet as a methyl donor. As a control, we constructed a pfpmt⌬ ϩ PfPMT complemented strain, which expresses wild type PfPMT episomally in the pfpmt⌬ strain. Whereas the formation of radiolabeled phosphocholine was readily detectable in reactions with extracts from the wild type and pfpmt⌬ ϩ PfPMT strains, phosphocholine could not be detected in the reaction mixture containing extracts from the pfpmt⌬ parasites (Fig. 3A). Together, these results confirmed the loss of PfPMT enzymatic activity in the pfpmt⌬ parasites and demonstrated that PfPMT encodes the only phosphoethanolamine methyltransferase activity in P. falciparum.
To determine how the loss of PfPMT affected SDPM-dependent synthesis of PtdCho, in vivo labeling assays with [ 14 C]ethanolamine in the wild type, pfpmt⌬, and complemented strains were performed in medium lacking choline. Two-dimensional TLC analysis of the parasite phospholipids following incubation with [ 14 C]ethanolamine revealed, as expected, the formation of radiolabeled PtdEtn in all three parasite strains (Fig. 3B, top panels). In contrast, the formation of radiolabeled PtdCho from [ 14 C]ethanolamine was only detectable in the wild type and pfpmt⌬ ϩ PfPMT parasites, not in the pfpmt⌬ strain (Fig. 3B, top panels). Thus, disruption of PfPMT function results in a complete loss of the SDPM pathway. This result also demonstrates that P. falciparum, unlike mammalian  OCTOBER (Fig. 3B, bottom panels), demonstrating that the CDP-choline pathway is not altered by the loss of PfPMT.

Physiological Function of PfPMT in P. falciparum
Loss of PfPMT Results in Growth and Morphological Defects That Are Only Partially Complemented by Choline-To examine the importance of PfPMT in the P. falciparum intraerythrocytic life cycle, the growth rate of pfpmt⌬ parasites during their intraerythrocytic development and multiplication was compared with that of the wild type and pfpmt⌬ ϩ PfPMT strains. Since the SDPM and the CDP-choline pathways converge at the synthesis of phosphocholine from phosphoetha-nolamine and choline (Fig. 1), respectively, we envisaged that choline supplementation could complement any defects resulting from the loss of the PfPMT gene. The three strains were synchronized and cultured at 1% parasitemia in media containing increasing concentrations of choline for 48 h. Giemsastained parasite smears were analyzed by light microscopy (Fig.  4). Although similar parasite counts were found in the wild type and pfpmt⌬ ϩ PfPMT strains, the parasitemia in the pfpmt⌬ culture was reduced by at least 50% at choline concentrations of 0, 10, and 50 M in the culture medium (Fig.  4A). The addition of choline at 200 or 500 M (10-and 25-fold higher than the physiological level, respectively) had only a marginal effect on growth of the pfpmt⌬ (Fig. 4A) (data not shown). Wild type and pfpmt⌬ ϩ PfPMT parasites inoculated at 3% starting parasitemia and cultured for two consecutive generations in medium lacking choline increased their parasitemia to ϳ35%, whereas pfpmt⌬ parasites reached only 10% parasitemia (Fig. 4B) under the same conditions (RBCs contain residual choline that can support limited parasite growth). These results indicated that choline supplementation at physiological levels could not rescue the loss of PfPMT, although a marginal rescue effect was notable at concentrations that were at least 10-fold the physiological level. Microscopic analysis of highly synchronized cultures indicated that, although the wild type and the pfpmt⌬ ϩ PfPMT strains progressed normally, the pfpmt⌬ strain formed trophozoites and schizonts that were smaller in size and depicted a delayed progression rate (Fig. 4, C and  D). Determination of the relative size of the mature schizonts by measuring the average RBC area occupied by the parasites showed that mature schizonts of the pfpmt⌬ parasites occupied only about 33% of the RBC area, whereas their wild type and pfpmt⌬ ϩ PfPMT counterparts occupied ϳ75% of the RBC area (Fig. 4E).
pfpmt⌬ Parasites Have Altered DNA Replication-The low number of daughter parasites produced during the intraerythrocytic life cycle of pfpmt⌬ parasites could be due to a low number of nuclei produced during schizogony (and thus a low number of merozoites). Therefore, we quantified the number of nuclei from synchronized populations of live pfpmt⌬ and pfpmt⌬ ϩ PfPMT parasites under continuous perfusion using DNA-targeted fluorophores and SDCM. The complemented strain (Fig. 5A) showed a Gaussian distribution of numbers of new nuclei per schizont that is similar to that measured previously for wild type parasites (20). In contrast, the pfpmt⌬ strain grown in the absence of exogenous choline (Fig. 5B) showed a distinctly biphasic distribution of nuclei counts with the majority of schizonts showing a reduced number of nuclei consistent with microscopic analyses. Interestingly, when the pfpmt⌬ strain was grown in the presence of exogenous choline (Fig. 5C), the distribution remained broadened but shifted to higher numbers of nuclei. These results suggest that efficient production of phosphocholine plays a role in the formation of new nuclei during schizogony. Similar results were obtained by flow cytometry (data not shown).
pfpmt⌬ Parasites Have Decreased Cell Viability-To assess the effect of PfPMT disruption on the viability of intraerythrocytic parasites, asynchronous wild type, pfpmt⌬, and pfpmt⌬ ϩ FIGURE 3. A, evidence for the loss of PfPMT enzymatic activity in the pfpmt⌬ strain compared with the WT and pfpmt⌬ ϩ PfPMT strains. PfPMT activity in P. falciparum extracts (50 g of total protein) was measured at 30°C using phosphoethanolamine as substrate and AdoMet as co-substrate (methyl-donor). The product, phosphocholine, was purified using an AG-50(H ϩ ) ion exchange resin. The formation of radiolabeled phosphocholine was quantified by scintillation counting, and the equivalent methyltransferase activity was derived. A reaction mixture without protein extract was included as a blank (Ϫ). Data are means Ϯ S.D. derived from three independent experiments. B, synchronized WT, pfpmt⌬, and pfpmt⌬ ϩ PfPMT parasites grown to 10% parasitemia at early trophozoite stage in the absence of choline were labeled with either [ 14 C]ethanolamine (top panels) or [ 14 C]-choline (bottom panels) and incubated for 12 h. The infected erythrocytes were washed twice in PBS, and lipids were extracted by the Folch method (17). The organic phase of the lipid extracts were resolved by twodimensional TLC, and signals were generated by autoradiography. E, sample application point; PtdEtn and PtdCho, positions of phosphatidylethanolamine and phosphatidylcholine, respectively. P-Etn, phosphoethanolamine.
PfPMT parasites were analyzed by the TUNEL assay, which measures DNA fragmentation that occurs as an early event in cells undergoing cell death. Although a substantial number of pfpmt⌬ parasites cultured in 0, 20, or 200 M choline were positively stained with the TUNEL stain, wild type and pfpmt⌬ ϩ PfPMT parasites cultured under similar conditions were not (Fig.  6A). As a control, wild type parasites treated with DNase I to induce chromosomal DNA strand breaks stained positively in the presence of the TdT enzyme but negatively when the TdT enzyme was omitted from the reaction (Fig. 6B). To ensure that TUNEL staining correlates with parasite death, wild type, pfpmt⌬, and pfpmt⌬ ϩ PfPMT parasites were treated with 100 nM chloroquine or 100 nM pyrimethamine (concentrations ϳ10fold above their IC 50 values) for 24 h and subjected to the TUNEL assay. In all cases, the drug-treated parasites showed a positive TUNEL staining (Fig. 6C). Taken together, these results suggest that PfPMT plays an important role in the parasite's viability.

DISCUSSION
The results presented here show that P. falciparum parasites lacking PfPMT are unable to synthesize their major phospholipid, PtdCho, via the SDPM pathway, and display major alterations in parasite development, multiplication, and survival. Furthermore, this study provided the first genetic evidence that, unlike mammalian and yeast cells, P. falciparum parasites lack the ability to form PtdCho from PtdEtn.
PfPMT is a member of a new class of AdoMet-dependent methyltransferases found in Caenorhabditis elegans and plants. No homologs exist in mammals (23)(24)(25)(26)(27). Although they share significant homology in their primary structure, these enzymes differ in the structural organization of their catalytic domains. The plant PEAMTs have two tandem catalytic domains, with the N-terminal domain catalyzing methylation of phosphoethanolamine into monomethylphosphoethanolamine and the C-terminal domain acting in the last two methylation reactions to form phosphocholine (23,24). C. elegans PEAMTs contain only one methyltransferase domain located in either the N-terminal end of the protein, in the (Based on t test analysis of the data, the difference between 0 and 200 M choline was not significant.) C, progression of synchronized wild type, pfpmt⌬ ϩ PfPMT, and pfpmt⌬ parasite strains grown at different concentrations of choline. The Giemsa-stained smears were prepared at different time points of culture (h postinvasion) as indicated and examined by light microscopy. One thousand RBCs were counted, and the percentage of parasitized cells was determined. The wild type and pfpmt⌬ ϩ PfPMT parasites that had reached the schizont stage at 42 h progressed to release new daughter parasites (merozoites) that then invaded previously uninfected RBCs to form the Ring stage of the parasites at 48 h. It should be noted that at this stage, there could have been some merozoites that were in the preinvasive stage and thus could not be easily detected microscopically. To be representative, at each time point and for each sample, 200 infected RBCs were examined, and the parasite morphology that was found to be consistent in at least 80% of the infected RBCs was taken as the representative morphology. D, representative images of schizonts for the wild type (WT), pfpmt⌬ ϩ PfPMT, and pfpmt⌬ parasites. Five hundred parasites were analyzed for each form. E, comparison of the RBC area (percentage) occupied by a mature schizont stage of the WT, pfpmt⌬ ϩ PfPMT, and pfpmt⌬ (gray, white, and black columns, respectively). The area was determined using ImageJ software (National Institutes of Health). Values are means of 10 different schizont-infected RBCs examined for each parasite strain, with S.D. values shown as error bars. OCTOBER 10, 2008 • VOLUME 283 • NUMBER 41 case of Pmt1, or in the C-terminal end of the protein in Pmt2 (25,26). Pmt1 catalyzes only the first methylation reaction, whereas Pmt2 catalyzes the last two methylation reactions. The Plasmodium PfPMT shares homology with both plant and C. elegans PEAMTs (Pmt1 and Pmt2) but is only half the size of these proteins and catalyzes all three methylation steps (7). The properties of these enzymes and their absence in mammals suggests that, if validated, they would be viable targets for developing antiprotozoan and antihelmintic therapies (7,25). Metabolic studies using radiolabeled ethanolamine showed that although wild type, knock-out, and complemented parasite strains synthesized PtdEtn from ethanolamine, only wild type and complemented parasites formed PtdCho from this precursor. Interestingly, radiolabeled choline was incorporated to similar levels into PtdCho in the three strains, suggesting that neither the CDP-choline pathway nor the transport of choline is affected in the pfpmt⌬ knock-out parasite. Together, the metabolic studies demonstrate that the transmethylation of PtdEtn to form PtdCho does not occur in P. falciparum and that the SDPM pathway is the sole pathway used by the parasite to synthesize PtdCho from ethanolamine. This is further supported by the fact that no intermediates of the transmethylation of PtdEtn (monomethyl-PtdEtn and dimethyl-PtdEtn) can be identified in wild type parasites labeled with ethanolamine and that PfPMT in vitro or expressed in yeast cells lacks the ability to catalyze the transmethylation of PtdEtn (7,8).

Physiological Function of PfPMT in P. falciparum
Together the data suggest that the SDPM pathway might be a significant source of phosphocholine used for the biosynthesis of PtdCho in P. falciparum, as is the case in plants (28). This is consistent with the previous observation that P. falciparum develops and propagates normally in culture medium without choline (29,30). Interestingly, C. elegans with a knocked down phosphoethanolamine N-methyltransferase had severe developmental defects and a drastic reduction in fertility that could not be rescued by the addition of superphysiological levels of choline (26), suggesting that the  The seemingly higher than 10% parasitemia is due to permeabilization conditions that reduce the number of uninfected red blood cells. B, positive (DNase-treated) and negative (no DNase) controls for TUNEL staining in WT parasites. C, WT, pfpmt⌬ ϩ PfPMT, and pfpmt⌬ parasites were cultured in the presence of either 100 nM pyrimethamine or 100 nM chloroquine for 24 h and tested for apoptosis by the TUNEL assay.
physiological importance of PEAMTs is conserved across kingdoms.
Whereas the lipid labeling studies confirmed the essential function of the SDPM pathway in the synthesis of PtdCho in the absence of choline, the loss of PfPMT was not lethal when choline was not added to the culture medium. The survival of knock-out parasites under these conditions is most likely due to the presence of residual choline in human erythrocytes. Nevertheless, loss of PfPMT resulted in significant reduction (ϳ50%) in parasite progeny that was only marginally rescued by choline supplementation even when choline was added at 25 times the physiological concentration (7-20 M) (12,13).
Analysis of parasite progression into schizogony indicated that pfpmt⌬ parasites had a delayed rate of nuclear division and formed fewer nuclei per infected erythrocyte than the wild type and complemented parasites (Fig. 5). Furthermore, of the parasites that progressed through the intraerythrocytic life cycle, a substantial percentage underwent DNA degradation, as revealed by TUNEL staining (Fig. 6). Treatment of wild type parasites, which are uniformly TUNEL-negative, with lethal doses of chloroquine or pyrimethamine (all three strains are sensitive to these compounds) resulted in positive TUNEL staining, indicating that this assay detected dying and/or dead parasites. Together, these findings indicate that disruption of PfPMT results in defective parasite growth and multiplication as well as reduced parasite viability.
In S. cerevisiae, the loss of the PEM1 and PEM2 genes encoding phospholipid methyltransferases responsible for the synthesis of PtdCho from PtdEtn results in choline auxotrophy. Choline supplementation restores the growth of the mutant to wild type levels (31,32). In contrast, whereas choline incorporation into PtdCho was similar in the wild type, knock-out and complemented parasites, choline supplementation only partially complemented the growth, multiplication, and survival defects of pfpmt⌬ parasites. Two possible hypotheses could account for these major differences. First, PfPMT might contribute to the synthesis of a pool of PtdCho that is structurally different from that synthesized via the CDP-choline pathway and that might play a critical role in membrane biogenesis and/or signaling during the parasite's intraerythrocytic growth and proliferation.
Alternatively, the intermediates of the transmethylation of phosphoethanolamine (monomethyl-phosphoethanolamine and dimethyl-phosphoethanolamine) could function as signaling molecules during the parasite's life cycle, controlling various cellular processes, such as parasite size, DNA replication, multiplication, and survival. Future cell biological and genomic studies will aim to distinguish between these hypotheses and unravel the mechanism by which PfPMT controls growth, proliferation, and survival of P. falciparum.
Altogether, the studies presented here show that PfPMT is important for membrane biogenesis, development, survival, and propagation of the parasite. Compounds that target PfPMT could thus be combined with those that specifically target choline transport, choline phosphorylation, or other steps of the CDP-choline pathway in order to completely block PtdCho biosynthesis and kill the parasite.