A patatin-like phospholipase is crucial for gametocyte induction in the malaria parasite Plasmodium falciparum

Patatin-like phospholipases (PNPLAs) are highly conserved enzymes of prokaryotic and eukaryotic organisms with major roles in lipid homeostasis. The genome of the malaria parasite Plasmodium falciparum encodes four putative PNPLAs with predicted functions during phospholipid degradation. We here investigated the role of one of the plasmodial PNPLAs, a putative PLA2 termed PNPLA1, during blood stage replication and gametocyte development. PNPLA1 is present in the asexual and sexual blood stages and here localizes to the cytoplasm. PNPLA1-deficiency due to gene disruption or conditional gene-knockdown had no effect on erythrocytic replication, gametocyte maturation and gametogenesis. However, blood stage parasites lacking PNPLA1 were severely impaired in gametocyte induction, while PNPLA1 overexpression promotes gametocyte formation. The loss of PNPLA1 further leads to transcriptional down-regulation of genes related to gametocytogenesis, including the gene encoding the sexual commitment regulator AP2-G. Additionally, lipidomics of PNPLA1-deficient asexual blood stage parasites revealed overall increased levels of major phospholipids, including phosphatidylcholine (PC), which is a substrate of PLA2. Because PC synthesis is pivotal for erythrocytic replication, while the reduced availability of PC precursors drives the parasite into gametocytogenesis, we hypothesize that the high PC levels due to PNPLA1-deficiency prevent the blood stage parasites from entering the sexual pathway.

We here report on the functional characterization of one of the four plasmodial PNPLAs, termed PNPLA1. The protein is found in the cytosol of the asexual and sexual blood stages of P. falciparum. Asexual blood stages lacking PNPLA1 exhibit upregulated phospholipid levels as well as impaired gametocyte induction both under normal growth conditions and following growth in phospholipid precursor-deficient medium, suggesting that the deregulated phospholipid levels in the asexual blood stages render these less sensitive to triggers of gametocytogenesis. 8 wildtype (WT) was subsequently demonstrated by IFA (Fig. S4C). Similarly, WB using the mouse anti-PNPLA1rp1 antisera, could detect a PNPLA1-positive protein band in blood stage lysates of the WT, but not in lysates of the two PNPLA1-KO lines (Fig. S4D).
In addition, we generated a conditional PNPLA1-knock down (KD) line (using the pARL-HA-glmS vector), by fusing the sequences coding for a hemagglutinin A tag and for the glmS element to the 3´-region of pnpla1 (Fig. S6A). In the PNPLA1-KD, the pnpla1 mRNA is expressed under the control of the glmS ribozyme (Prommana et al., 2013), which catalyzes its own cleavage in the presence of glucosamine (GlcN) and thus triggers degradation of the pnpla1 mRNA. Vector integration was confirmed by diagnostic PCR and sequencing of the integration sites ( Fig. S6A, B; S7). Immunolabelling with an anti-HA antibody confirmed the synthesis of a HA-tagged PNPLA1 fusion protein in the untreated PNPLA1-KD line, while no signal was detected in WT (Fig. S4C). Similarly, immunoblotting with the anti-HA antibody detected the PNPLA1 fusion protein running with a molecular weight of 82.2 kDa and confirmed protein expression in the asexual blood stages as well in gametocytes during development and following activation (Fig. 1C). Subcellular fractioning, using an anti-HA antibody, further demonstrated that PNPLA1-HA locates to the cytosolic fraction of the blood stage parasite, while the protein was absent in the peripheric, the integral and the insoluble factions (Fig. 1D).
Immunoblotting with antibodies against the cytoskeletal element actin1 (Ngwa et al., 2013; reviewed in Baum et al., 2008;) the cytosolic protease falcilysin (Weißbach, Golzmann, Bennink, Pradel, & Ngwa, 2017), the endoplasmic reticulum-associated protein Pf39 (Simon et al., 2009) and the transmembrane protein AMA1 (Boes et al., 2015;Peterson et al., 1989) were used for fraction controls. To verify the conditional down-regulation of PNPLA1 synthesis, asexual blood stage parasites were treated with 2.5 mM GlcN for 72 h. Upon GlcN addition, PNPLA1 levels were significantly reduced to 40.00 ± 13.4% compared to the untreated control ( Fig. S6D, E). Initial phenotype analyses of the two PNPLA1-KO lines showed that intraerythrocytic development of the asexual blood stages was normal compared to WT and no differences in parasitemia over a period of 120 h as well as in blood stage morphology and progression and in the numbers of merozoites formed per schizont were observed ( Fig. 3A-C; S8A).
Furthermore, no significant differences in the exflagellation behavior as well as in the formation of macrogametes and zygotes following in vitro activation of the PNPLA1-KO gametocytes was seen (Fig. 3D). When gametocytemia was compared between the PNPLA1-KO lines and the WT over a period of 14 d, however, significant reductions in the numbers of gametocytes were documented (Fig. 3E).
To investigate the effect of PNPLA1-deficiency on gametocytes in more detail, we used the PNPLA1-KD line in the follow-up experiments. We confirmed that reduced PNPLA1 synthesis did not lead to any differences in parasitemia over a period of 120 h as well as in blood stage morphology and progression, when these were treated or not treated with GlcN ( Fig. 4A; S8B, S9). The low PNPLA1-levels, however, significantly reduced gametocyte numbers in the GlcN-treated PNPLA1-KD line, when this line was compared to untreated PNPLA1-KD parasites and treated or untreated WT parasites (Fig. 4B). To determine, if the PNPLA1-deficiency effects gametocyte induction or rather early gametocyte development, we treated the blood stages with GlcN either before or after inducing sexual commitment by the addition of lysed RBCs. We observed that PNPLA1 down-regulation prior to gametocyte induction resulted in significantly reduced gametocyte numbers, while PNPLA1 downregulation after gametocyte induction had no effect (Fig. 4C). Untreated PNPLA1-KD parasites as well as GlcN-treated and untreated WT were used for controls in these experiments.

2.3/ PNPLA1 deficiency counteracts gametocyte induction by serum-free medium
In a recent study, Brancucci et al. (2017) reported that serum-free cell culture medium (-SerM), which only contained the minimally required fatty acid species C16:0 and C18:1 necessary for the growth of P. falciparum blood stages (Mi-Ichi, Kano, & Mitamura, 2007;Mi-Ichi, Kita, & Mitamura, 2006), triggers sexual commitment. This effect could be reverted by addition of lysoPC to the serum. We thus repeated the above experiments and induced gametocyte formation by cultivation of the blood stage parasites in -SerM. We confirmed that in the WT treatment with -SerM resulted in an increased formation of gametocytes compared to medium supplemented with lysoPC, choline or human serum (Fig. 5A). The two PNPLA1-KO lines, however, did not respond to -SerM and gametocyte numbers were significantly reduced ( Fig. 5A). Similarly, the PNPLA1-KD line did not respond to -SerM, when treated with GlcN prior to gametocyte induction, while the untreated PNPLA1-KD line formed higher numbers of gametocytes in -SerM compared to medium supplemented with lysoPC, choline or human serum (Fig. 5B). Brancucci et al. (2017) further described that lysoPC-depletion triggers an immediate transcriptional upregulation of a variety of genes, including such encoding transcriptional regulators of sexual commitment like ap2-g (Brancucci et al., 2017;Kafsack et al., 2014;Sinha et al., 2014) and essential gametocytogenesis factors like gdv1 (Brancucci et al., 2017;Eksi et al., 2012). Furthermore, genes coding for enzymes that are known to be involved in the utilization of ethanolamine to generate PC were upregulated (Bobenchik et al., 2010;Brancucci et al., 2017;Pessi, Choi, Reynolds, Voelker, & Mamoun, 2005;Pessi, Kociubinski, & Mamoun, 2004). We chose eight genes that were found to be upregulated under lysoPC-limiting conditions and investigated their transcriptional levels at 46 ± 2 h.p.i. (hours post-invasion) to compare induced to non-induced controls. For all of the eight genes, a 2-to 5-fold transcriptional upregulation could be observed in the -SerM-cultivated WT (Fig. 5C, D). A similar transcriptional up-regulation was seen in the PNPLA1-KO lines A12 and C11, but the fold-changes were decreased. For ap2-g, and the phosphoethanolamine N-methyltransferase PMT that catalyzes the conversion of PE into PC, significantly reduced transcript levels compared to the -SerM-induced WT control were observed.
In a next step, the effect of commercial PLA2 inhibitors on gametocyte induction was tested. First, the antimalarial effect of the four inhibitors 4-bromophenacyl bromide (4-BPB), bromoenol lactone, ASB14780 and N-(p-amylcinnamoyl) anthranilic acid on erythrocytic replication was investigated, using the Malstat assay. The highest antimalarial effect was detected for 4-BPB with an IC50 value of 5.6 ± 0.82 µM (Table 1). This concentration was used to treat asexual blood stages when inducing gametocytes by treatment with -SerM. When the asexual blood stages were treated with 4-BPB, significantly reduced gametocyte numbers were observed in the WT upon -SerM incubation compared the untreated WT and the solvent control ( Fig. 5E). On the other hand, in the two PNPLA1-KO lines, no significant differences were observed in 4-BPB-treated parasites compared to the controls. 4-BPB treatment, however, had no significant effect on gametocyte maturation and gametocyte stage progression, when added to early stage II gametocytes, while epoxomicin used for positive control (Aminake et al., 2011) killed the gametocytes (Fig. S10A, B).
We also investigated the reverse effect, i.e. the effect of PNPLA1 overexpression on gametocyte induction. The sequence for full-length PNPLA1 was cloned into the vector pARL-GFP, which was driven either under the control of the ubiquitously active crt promotor (Külzer et al., 2010;Przyborski et al., 2005;Spork et al., 2009) or the gametocyte-specific fnpa promotor (Bennink et al., 2018;Pradel et al., 2004) (Fig. S11A). Presence of the plasmids in the transfectants was verified by diagnostic PCR, purified plasmid DNA was used for control ( Fig. S11B). Live imaging of corresponding lines showed that the crt promotor allowed the episomal expression of PNPLA1 in the transfected asexual blood stages (line crt-PNPLA1-GFP), while the protein was expressed in transfected gametocytes when controlled by the fnpa promotor (line fnpa-PNPLA1-gfp) (Fig. S11C). WT parasites did not show any GFP expression ( Fig. S11D). Asexual blood stages of the transfectant lines crt-PNPLA1-GFP and fnpa-PNPLA1-GFP were subsequently induced by -SerM, which resulted in significantly increased gametocyte numbers in the crt-PNPLA1-GFP line at day 5 post-gametocyte induction (p.g.i.) compared to the WT and fnpa-PNPLA1-GFP (Fig. 5F). No differences in the morphology of the asexual and sexual blood stages, though, were observed in lines crt-PNPLA1-GFP and fnpa-PNPLA1-GFP compared to WT (Fig. S12).

2.4/ PNPLA1-deficiency results in an overall increase of major phospholipids
Since the deficiency of PNPLA1 impaired the proper gametocytogenesis pathway, we investigated whether the parasite lipid composition was altered in the PNPLA1-KO lines, accordingly to the putative role of the enzyme and the role of phospholipids for maintaining normal asexual blood stage growth. Total lipids from synchronized ring stages of WT and the PNPLA1-KO lines A12 and C11 were extracted and mixed with an internal lipid standard mix (LIPDOMIX standard, Avanti). Each lipid class was separated by high performance thin layer chromatography and then extracted from the silica gel. Total fatty acids from each lipid class were methanolized to fatty acid methyl ester and then quantified by GC-MS. The general phospholipid levels of the two PNPLA1 lines was slightly higher compared to the WT (Fig.   6A). This increase was not particularly associated to one specific lipid class but more to general lipid increase. The phospholipid changes, however, were particularly drastic for PC, which was significantly increased in the PNPLA1 lines compared to the WT (Fig.6 B). Noteworthy, he difference in PC levels did not alter the overall phospholipid composition (Fig.6C), probably for the parasite to keep its membrane integrity. Solely for lysobisphosphatidic acid (LBPA) levels, a decrease in the PNPLA1-KO lines compared to WT could be observed (Fig. 6B, C). The lysoPC content in the different lines was below detection level.

3/ DISCUSSION
During infection with P. falciparum, gametocytes begin to develop approximately one week after the appearance of parasites in the human blood. A small proportion of committed parasites leaves the asexual blood stage cycle and enters the sexual pathway on a continuous basis, an event probably driven by endogenous factors. The gametocyte commitment rate, however, increases drastically in response to environmental stress signals, including parasite density, anaemia, host immune response or drug treatment (reviewed in Alano, 2007;Kuehn and Pradel, 2010). Increasing data suggest that the lack of phospholipid precursors in the human serum triggers gametocyte induction. Particularly, low serum levels of lysoPC, which was recently shown to act as the main exogenous precursor of PC, cause the blood stage parasites to enter the sexual pathway (Brancucci et al., 2017;Wein et al., 2018). Despite previous assumptions that choline is acquired from serum, new data indicate that the majority of choline to be incorporated in PC is generated from imported lysoPC (Fig. 7).
Similar to the asexual blood stages, gametocytes have a roughly 6-fold higher phospholipid level compared to the non-infected RBC. In gametocytes, however, the PC levels are lower than in the asexual blood stages (Tran et al., 2016). Under normal serum conditions, PC is generated from choline by the de novo cytidine diphosphate (CDP)-choline (Kennedy) pathway (reviewed in Tischer et al., 2012;Flammersfeld et al., 2018). In the absence of lysoPC in the serum, however, PC is mainly synthesized via triple methylation of phosphoethanolamine using ethanolamine (from serum) or serine (from hemoglobin) as external precursors (Fig. 7). This pathway involves the activity of PMT and parasite lines deficient of PMT die in serum lacking these precursors (Wein et al., 2018). This alternative route of PC synthesis via the PMT pathway appears to be upregulated in gametocytes, explaining the higher tolerance of these stages towards serum-free medium (Brancucci et al., 2017).
While these data demonstrate the high sensitivity of the plasmodial blood stages to the availability of phospholipids, we now show that PNPLA1 of the parasite is involved in the phospholipid-dependent switch from asexual blood stage replication to gametocyte development. PNPLA1 is a cytosolic PLA2 of the asexual blood and gametocyte stages with a particular role in gametocyte induction. Asexual blood stage parasites lacking PNPLA1, due to conventional gene-KO or conditional gene-KD, are less responsive to triggers of gametocyte commitment, including lysoPC deficiency. This loss-of-function phenotype can be mimicked by the PLA2 inhibitor 4-BPB, pointing to direct link between PNPLA1 activity and gametocyte induction. In contrast, overexpression of PNPLA1 leads to higher gametocyte commitment rates.
The lack of PNPLA1 further results in a deregulation of the major phospholipid biosynthesis pathway in the asexual blood stages. Lipidomics indicated increased overall levels of phospholipids and more particularly PC in the absence of PNPLA1. Noteworthy, PLA2 activities were predicted to be important for two reactions of this pathway; the degradation of PC to lysoPC and the degradation of PE to lysoPE (see http://mpmp.huji.ac.il/). In consequence, the loss of PNPLA1 activity would evidently result in an accumulation of PC, PE, and PS, and could also lead to an increased conversion of PC into SM (Fig. 7). PNPLA1 is one of four PNPLAs identified in P. falciparum, and to date for none of these functional data have been reported. However, in the Apicomplexan parasite Toxoplasma gondii, three of the roughly six annotated PNPLAs have meanwhile been functional characterized. TgPL1 and TgPLA2 localize to vesicles and are discharged upon immune stress and during invasion, respectively; Furthermore, TgPL2 is crucial for apicoplast integrity (Cassaing et al., 2000;Lévêque et al., 2017;Mordue, Scott-Weathers, Tobin, & Knoll, 2007;Tobin & Knoll, 2012). These findings indicate diverse roles of the enzyme family for parasite survival. Similar diverse functions were reported for other eukaryotic PNPLAs. For example, in tobacco, PNPLAs are usually synthesized upon virus infection to generate signaling molecules that induce programmed cell death of the infected tissue (e.g. Dhondt et al., 2000;Dhondt et al., 2002;reviewed in Ryu, 2004). Mammalian PNPLAs, in contrast, are mostly used in lipid metabolism and turnover, e.g. the triglyceride synthesis and lipolysis (reviewed in Wilson and Knoll, 2018). Noteworthy, PNPLAs can also be found in prokaryotic microbes like Legionella pneumophila. Legionella injects the virulence factor VipD into the host cell, which has PLA1/2 activity (Gaspar & Machner, 2014;Ku et al., 2012;Shohdy, Efe, Emr, & Shuman, 2005;Zhu, Hammad, Hsu, Mao, & Luo, 2013). VipD localizes to early endosome membranes.
Dependent on its PLA1 activity, it alters associated lipids and proteins and thereby avoids endosomal fusion (Gaspar & Machner, 2014). Another report described that PLA2-acting VipD hydrolyses PC and PE of the mitochondrial membrane, which leads to the formation of reactive lysophospholipids. These, in consequence, help to destabilize the mitochondrial membrane, resulting in the programmed killing of the infected host cell (Zhu et al., 2013).
In view of our data, we postulate that that the high PC levels due to PNPLA1-deficiency prevent malaria parasites from entering the sexual pathway. In consequence, the molecular switch from asexual blood stage replication to gametocyte formation would be flipped by decreasing PC levels. Due to reduced need of PC and the activation of the PMT pathway, gametocytes have a higher viability in serum with low levels of phospholipid precursors. Future studies need to identify, how the malaria parasite monitors its phospholipid levels and unveil the downstream signaling pathways leading to the activation of transcriptional regulators of sexual commitment.

4.1/ Gene Identifiers
The following PlasmoDB gene identification numbers are assigned to the genes and proteins investigated in this study: ActinI

4.5/ Recombinant protein expression
Recombinant proteins corresponding to PNPLA1rp1 (spanning aa 498-670) and PNPLA1rp2 (spanning aa 335-496) were expressed as fusion proteins with an N-terminal maltose binding protein (MBP)-tag using the pMAL TM c5x-vector (New England Biolabs). Cloning was mediated by the addition of the restriction sites XmnI/PstI to the ends of gene fragments PCRamplified from P. falciparum gDNA, using PNPLA1rp1 forward primer and PNPLA1rp1 reverse primer for PNPLA1rp1 and PNPLA1 rp2 forward primer and PNPLA1 rp2 reverse primer for PNPLA1rp2 (for primer sequences, see Table S1). Recombinant proteins were expressed using E.coli BL21 (DE3) RIL according to the manufacturer's protocol (Stratagene).
The recombinant fusion proteins were purified via affinity chromatography from bacterial extracts using amylose resin (New England Biolabs) according to manufacturer's protocol. The full-length recombinant protein corresponding to PNPLA1rp3 (spanning aa 001-679) was expressed as fusion protein with an N-terminal glutathione-S-transferase (GST)-tag using the pGEX-6P-1 vector (GE Healthcare). A synthetic PNPLA1 gene, codon-optimized for recombinant protein expression in E. coli with the GeneOptimizerTM software (Thermo Fisher Scientific), was used as template DNA (for synthetic sequence, see Fig. S13). Cloning was mediated by the addition of the restriction sites BamHI and XhoI to the ends of gene fragments PCR-amplified from codon-optimized DNA-template using the PNPLA1rp3 forward primer and PNPLA1rp3 reverse primer (for primer sequences, see Table S1). Recombinant proteins were expressed using E.coli BL21 (DE3) RIL according to the manufacturer's protocol.

4.6/ Generation of mouse antisera
Recombinant fusion proteins PNPLA1rp1-MBP and PNPLA1rp2-MBP were purified via affinity chromatography as described above followed by PBS buffer exchange via filter centrifugation using Amicon Ultra 15 (Sigma-Aldrich) according to the manufacturer's protocol. Protein concentrations were determined via Bradford assay. Immune sera were generated by immunization of 6 weeks-old female NMRI mice (Charles Liver Laboratories) via subcutaneous injection of 100 µg recombinant protein emulsified in Freund's incomplete adjuvant (Sigma-Aldrich) followed by a boost after 4 weeks with 50 µg of recombinant protein.
Mice were anesthetized 10 days after the boost by intraperitoneal injection of ketamine-xylazine mixture according to the manufacturer's protocol (Sigma-Aldrich). Polyclonal immune sera were collected via heart puncture and pooled from three mice immunized with the same antigen.

4.7/ Generation of the PNPLA1-KO parasite lines
PNPLA1-KO parasite lines were generated via single cross-over homologous recombination using the pCAM-BSD-KO vector (Bennink et al., 2018;Dorin-Semblat et al., 2007;Wirth, Bennink, Scheuermayer, Fischer, & Pradel, 2015;Wirth et al., 2014). A 567 bp gene fragment homologous to the PNPLA1 gene coding for the N-terminal part was amplified via PCR using the PNPLA1-KO forward and reverse primers (for primer sequences, see Table S1). Ligation of insert and vector backbone was mediated by BamHI/ NotI restriction sites. A WT culture synchronized for 5% ring stages was loaded with 100 µg vector in transfection buffer via electroporation (310 V, 950 µF, 10 ms; Bio-Rad gene-pulser) as described Wirth et al., 2014). Blasticidin (InvivoGen) (2), pCAM-BSD forward primer (3) and pCAM-BSD reverse primer (for primer location see Fig. S4; for primer sequences, see Table S1). Successful confirmation of vector integration was followed by clonal dilution of a >3% ring stage culture in a 96-well plate. After three weeks of cultivation, clonal sub-cultures were identified via Malstat assay as described below and subsequent diagnostic PCR was performed to confirm correct integration and absence of the WT pnpla1 gene locus. Two clonal lines, PNPLA1-KO A12 and PNPLA1-KO C11, were isolated.

4.8/ Generation of the PNPLA1-KD parasite line
PNPLA1-KD parasite lines were generated via single cross-over homologous recombination using the pARL-HA-glmS vector (Fig. S3). An 867-bp gene fragment homologous to the PNPLA1 gene coding for the C-terminal part was amplified using the PNPLA1-KD forward primer and the PLPLA1-KD reverse primer (for primer sequences, see Table S1). The stop codon was excluded from the homologous gene fragment. Ligation of the insert with the vector backbone was mediated by NotI and AvrII restriction sites. Transfection of parasites was performed as described above. For selection of parasites carrying the vector, WR92210 (Jacobus Pharmaceutical Company) was added to a final concentration of 2.5 nM and successful integration of the vector was confirmed by diagnostic integration PCR using PNPLA1-KD 5'integration primer (1), PNPLA1-KD 3'-integration primer (2), pARL-HA-glmS forward primer (3) and pARL-HA-glmS reverse primer (4) (for primer location see Fig. S6; for primer sequences, see Table S1).

4.9/ Generation of the PNPLA1-GFP episomal expression parasite lines
The crt-PNPLA1-GFP and the fnpa-PNPLA1-GFP parasite lines were generated using the pARLII-GFP vector (Fig. S10A). The pnpla1 full-length gene was amplified from cDNA using the PNPLA1-gfp forward primer and the PNPLA1-gfp reverse primer (for primer sequences, see Table S1). The stop codon was excluded from the full-length gene sequence. Ligation of the insert with the vector backbones was mediated by XhoI and AvrII. The plasmid was sequenced to confirm that the encoding segment was inserted in frame with the GFP encoding sequence. Transfection of parasites was performed as described above. For selection of parasites carrying the vectors, WR92210 (Jacobus Pharmaceutical Company) was added to a final concentration of 2.5 nM and successful uptake of the vector was confirmed by diagnostic PCR using the PNPLA1-gfp episome forward primer and the PNPLA1-gfp episome reverse primer (for primer location see Fig. S10; for primer sequences, see Table S1). was amplified in 25 cycles using PNPLA1 RT forward and reverse primers (for primer sequences, see Table S1). To confirm purity of the asexual blood stage and gametocyte samples, transcript amplifications of ama1 (407 bp) using AMA1-RT forward and reverse primers and of pfccp2 (286 bp) using CCP2-RT forward primer and reverse primers were performed (Peterson et al., 1989;Pradel et al., 2004). Amplifications of the housekeeping gene aldolase (378 bp) using Aldolase-RT forward and reverse primers were used as loading control and to test for residual genomic DNA in the negative controls without reverse transcriptase. PCR products were separated by 1.5% agarose gel electrophoresis.

4.13/ Real-time RT-PCR
RNA from asexual blood stages was isolated as described above and 2 µg of each total RNA sample was used as template for cDNA synthesis using the Super Script IV First-Strand Synthesis System (Invitrogen) following the manufacturer's protocol. The synthesized cDNA was first tested by diagnostic RT-PCR for purity of the asexual blood stage and gametocyte samples, using ama1-and pfccp2-specific primers, respectively. Controls without reverse transcriptase were used to exclude potential genomic DNA contamination by using aldolase  (Livak & Schmittgen, 2001) method using the endogenous control gene encoding the P.

4.14/ Subcellular fractioning
The protocol for subcellular fractionation of parasite lysates using the solubility of proteins in different buffers was modified from Grüring et al. (2012). In brief, ~3*10^6 schizonts of the PNPLA1-KD line were purified via Percoll gradient and liberated from the host cells with 0.03% saponin/PBS for 3 min at 4°C. After washing with PBS, parasites were hypotonically lysed in 100 µl of 5 mM Tris-HCL (pH 8), supplemented with protease inhibitor cocktail (complete EDTA-free, Roche) for 10 min at RT followed by freezing at -20°C and thawing on ice. Soluble proteins (Tris-HCL fraction) were separated from non-soluble constituents by centrifugation (5 min, 16,000xg, 4°C). This was followed by sequential pellet extraction with 100 µl each of freshly prepared Na2CO3 on ice for 30 min (peripheral fraction), ice-cold 1% Triton X-100 (integral membrane fraction), and 0.5 x PBS/4% SDS/ 0.5% Triton X-100 (insoluble fraction) at RT. Between each extraction, the remaining pellet was washed with 1 ml ice-cold PBS to remove any impurities. Supernatants were centrifuged to remove residual material. The individual fractions were subjected to WB as described below. binding was blocked by incubation of the membranes in Tris-buffered saline containing 5% w/v skim milk, pH 7.5, followed by immune recognition overnight at 4°C using polyclonal mouse immune sera specific for Pf39 (dilution 1:5000), PfAMA1 (dilution 1:1,000), falcilysin

4.16/ Asexual blood stage replication assay
To compare asexual blood stage replication between the parental WT and the PNPLA1-KO lines A12 and C11, synchronized asexual blood stage cultures were set to an initial parasitemia of 0.25% ring stages und cultivated as described above. Giemsa-stained thin blood smears were

4.18/ Comparative exflagellation assay
At day 14 p.g.i., mature gametocytes of the WT and the PNPLA1-KO lines A12 and C11 were adjusted with erythrocytes to a final gametocytemia of 1% with a total hematocrit of 5% in cell culture medium. A volume of 100 µl of each gametocyte culture was activated in vitro with 100 µM XA for 15 min at RT. At 15 min p.a., the numbers of exflagellation centers were counted at 400-fold magnification in 30 optical fields using a Leica DMLS microscope. Four independent experiments were conducted in quadruplicate; data analysis was performed using MS Excel 2014. Exflagellation was calculated as a percentage of the number of exflagellation centers in the PNPLA1-KO lines in relation to the number of exflagellation centers in the WT control (WT set to 100%).

4.19/ Comparative macrogamete and zygote formation assay
For the analysis of macrogamete and zygote formation, mature gametocytes of the WT and of the PNPLA1-KO lines A12 and C11 were enriched via Percoll purification and adjusted to a final gametocytemia of 2% with a total hematocrit of 5% in cell culture medium. Gametocyte cultures were activated in vitro with 100 µM XA for 30 min (macrogametes) or 6 h (zygotes) at RT. An equal volume of each sample was coated on Teflon slides and subjected to IFA as described above using rabbit anti-Pfs25 antisera for macrogamete and zygote labelling.
Parasites were counted microscopically using a Leica DM 5500B fluorescence microscope with 600-fold magnification. Numbers of macrogametes and zygotes were counted for a total number of 1000 erythrocytes in triplicate and calculated in relation to the WT control (set to 100%). For each assay, three experiments were performed, each in triplicate. Data analysis was performed using MS Excel 2014.

4.21/ Gametocyte toxicity assay
WT parasites were grown at high parasitemia to induce gametocyte formation. When stage II gametocytes were formed, 1 ml of the culture was plated in triplicate in 24-well plate and incubated with 4-BPB at IC50 (5.62 µM) concentration. Chloroquine at IC50 (12.1 nM) concentration and 0.5% v/v DMSO served as negative controls. The proteasome inhibitor epoxomicin (60 nM) was diluted in DMSO and used as positive control (Aminake et al., 2011;Ngwa et al., 2017). The parasites were treated with inhibitors for 2 d and the medium was replaced daily. The parasites were cultivated for 10 d and Giemsa-stained blood smears were taken at day 7 and day 10 to determine numbers of gametocytes of stages IV and V per 1,000 RBCs. Three experiments were performed, each in triplicate. Data analysis was performed using MS Excel 2014 and GraphPad Prism 5.

4.22/ Lipidomics analysis
Analysis was performed on four independent cell harvests of the WT and the PNPLA1-KO lines A12 and C11. Highly synchronous parasite cultures were harvested by transferring the cultures to pre-warmed 50-ml tubes and the total amount of parasites were determined by Giemsa staining and via counting on a haemocytometer. The cultures were metabolically quenched via rapid cooling to 0°C by suspending the tube over a dry ice/100% ethanol slurry mix, while continually stirring the solution. Following quenching, parasites were constantly kept at 4°C and liberated from the enveloping RBCs by mild saponin treatment with 0.15% w/v saponin/0.1% w/v BSA PBS for 5 min. After three washing steps with 1 ml ice-cold PBS followed by centrifugation each time, the parasite pellet was subjected to lipid extraction using chloroform and methanol (Amiar et al., 2016)

4.23/ Statistical analysis
Data are expressed as mean ± SD. Statistical differences were determined using One-Way     were immunolabelled with mouse anti-PNPLA1rp1 antisera (green). The asexual blood stages were stained with EB; the gametocytes were counterlabelled with rabbit anti-Pfs230 sera (red).
The parasite nuclei were highlighted by Hoechst33342 nuclear stain (blue). DIC, differential interference contrast. Bar, 5 µm. Results are representative of three independent experiments.