Global analysis of putative phospholipases in the malaria parasite Plasmodium falciparum reveals critical factors for parasite proliferation

For its replication within red blood cells, the malaria parasite is highly dependent on correctly regulated lipid metabolism. Enzymes involved in lipid metabolic processes are therefore potential drug targets. We here provide a functional analysis of the 20 putative phospholipases that are expressed by asexual blood stages of Plasmodium falciparum. We reveal a high level of redundancy among members of this group, but using conditional mislocalization and gene disruption techniques we show that the phosphoinositide-specific phospholipase C (PF3D7_1013500) has a previously unrecognized essential role in intracellular parasite maturation. In addition, we demonstrate that the patatin-like phospholipase PF3D7_1358000 localizes to the mitochondrion. Parasites lacking this enzyme display a severe growth phenotype and defects in mitochondrial morphogenesis and function leading to hypersensitivity towards proguanil and inhibitors of the mitochondrial electron transport chain including atovaquone. This demonstrates that regulated mitochondrial lipid homeostasis is necessary for mitochondrial function and coordinated division during parasite multiplication.


INTRODUCTION 45
With an estimated 228 million cases per year worldwide and more than 400,000 46 deaths, malaria remains one of the most important human health threats (WHO,47 cycles. PI-PLC-null parasites failed to proliferate, confirming our knocksideways-244 based indications that PI-PLC is crucial for viability during asexual blood stage 245 replication of P. falciparum ( Figure 3E). For more in-depth characterization, the F9  invaginations, pointing to the start of merozoite formation. However, we were unable 265 to find more than a few well-segmented schizonts in the PI-PLC-null samples, in 266 contrast to the majority of the mock-treated parasites, which formed well-segmented 267 schizonts with clearly defined merozoites ( Figure 3I,J). These microscopic 268 observations show that most PI-PLC-null mutants show stunted growth in late 269 trophozoite stages while a few of them develop to early schizont stage but stop short 270 of becoming mature schizonts. Taken together, we concluded that lack of PI-PLC 271 caused a severe growth defect during the trophozoite-schizont transition, suggesting 272 that PI-PLC-mediated activity is critical for intraerythrocytic parasite development. We next characterized the growth phenotype of PNPLA2-null parasites in more 293 detail, examining the development of tightly synchronized intracellular parasites over 294 the course of the erythrocytic cycle. This showed that PNPLA2-null parasites 295 exhibited delayed development in comparison to WT parasites, as evident by 296 microscopic quantification of Giemsa-stained thin blood films ( Figure 4D). However, 297 12 this also revealed that in those PNPLA2-null schizonts that developed to maturity, 298 there was no significant decrease in daughter merozoite numbers as compared to 299 WT schizonts ( Figure 4E), indicating that loss of PNPLA2 leads to delayed but not 300 compromised parasite maturation. 301 Given the mitochondrial localization of PNPLA2, we next studied mitochondrial 302 development in the PNPLA2-null parasites. Microscopic examination of mitochondria 303 during trophozoite and schizont development using MitoTracker Red revealed 304 mitochondrial abnormalities in the mutant parasites in the form of accumulations that 305 first became evident following the first round of nuclear division (2 nuclei) and were 306 further pronounced at later stages of schizogony (3 and more nuclei) ( Figure 4F). To 307 further analyze and quantify this phenotype, we performed an end-point analysis by 308 arresting egress of WT and PNPLA2-null schizonts for 8 hours using C2 and then 309 quantifying mitochondrial morphology using MitoTracker Red staining. This showed 310 that whilst most segmented WT schizonts displayed the typical comma-like structure 311 of divided mitochondria, ~80% of PNPLA2-null schizonts showed abnormal 312 mitochondrial accumulations, together indicating that PNPLA2 is involved in 313 mitochondrial morphogenesis ( Figure 4G). 314 315

Confirmation of the PNPLA2-null phenotype by conditional gene disruption 316
To further analyze the function of PNPLA2, and to establish whether the observed 317 defects were detectable immediately following gene disruption, we next targeted the 318 pnpla2 gene using the DiCre-based conditional KO approach. To this aim, we again 319 used Cas9-assisted double homologous recombination to flox the sequence 320 encoding the C-terminal half of the PNPLA2 coding sequence (harboring the catalytic 321 PNPLA domain), simultaneously appending a 3xHA epitope tag to the gene (Figure 322 To establish whether the reduced replication rate in PNPLA2-null parasites was due 345 to inefficient egress or invasion, we isolated schizonts from RAP-and mock-treated 346 PNPLA2:HA:loxPint cultures at the end of the cycle of treatment and incubated them 347 with fresh RBCs under both static and shaking conditions. This showed no significant 348 differences between the resulting increases in parasitemia ( Figure 5F), suggesting 349 that loss of PNPLA2 does not impair egress or invasion. Using two distinct conditional gene targeting approaches we now provide 480 unambiguous genetic evidence that PI-PLC is essential for P. falciparum asexual 481 blood stage proliferation. Both conditional inactivation techniques resulted in a defect 482 in trophozoite to schizont conversion, as well as impaired development of schizonts. 483 In line with this maturation phenotype, our lipidomic analysis showed a significantly 484 reduced lipid content in the PI-PLC deficient parasites, while only a slight impact on 485 the overall lipid profile was visible. It might be interesting, although experimentally 486 challenging, to directly probe the enzymatic activity of PI-PLC by quantitation of PIP 2 487 and inositol 1,4,5-triphosphate (IP 3 ) using our PI-PLC deficient parasites. 488 The maturation phenotype of our PI-PLC deficient parasites is reminiscent of that We further show that PNPLA2-KO parasites are hypersensitive to drugs that target 515 the mtETC and that they have a defect in sustaining ΔΨm, which both together 516 argues for a defect somewhere in the mtETC. Although the molecular dissection of 517 this defect requires further investigation, it is likely that it occurs downstream of 518 complex III, for instance in the electron transport from complex III to complex IV or in 519 the activity of the latter. This assessment is based on the facts that i) we did not 520 observe hypersensitivity towards DSM1 and ii) we could not rescue the KO codon. To target PI-PLC, the above designed construct was synthesized as two parts 607 (2,866 bp and 791 bp) and combined by restriction-ligation (using HindIII and XhoI 608 enzymes) to create pREP-piplc-3HA-loxPint. Similarly, to target PNPLA2, the 609 designed construct was synthesized as two parts (2,421 bp and 1,109 bp) and 610 inserted subsequently into a pCR-blunt vector (Thermo Fisher Scientific) using 611 restriction-ligation with XhoI/ApaI and XhoI/PstI respectively. The synthesized 612 construct did not contain the 3' homology arm, which was therefore amplified from 613 B11 genomic DNA (amplification primers: PF3D7_1358000_3hom_F and 614

coverslip. 700
For IFA of PI-PLC:HA:loxPint parasites, air dried thin blood films were fixed with 4% 701 paraformaldehyde in PBS for 30 min at RT, permeabilized with 0.1% (v/v) Triton X-702 100 in PBS for 10 min, and blocked overnight in 4% BSA/PBS. Samples were probed 703 with rat anti-HA 3F10 (Sigma, 1:500) in 4% BSA/PBS. Bound primary antibodies 704 were detected using biotin-conjugated anti-rat antibody (Roche, 1:1,000) and 705 AlexaFluor594-conjugated streptavidin (Life Technologies, 1:1,000) in 4% BSA/PBS. were adjusted to ~0.1% parasitemia and divided into two 2 ml dishes. To one of 716 these dishes, rapalog (AP21967, Clontech) was added to a final concentration of 250 717 nM (rapalog was stored at −20°C as a 500 mM stock in ethanol, and working stocks 718 were kept as 1:20 dilutions in RPMI at 4°C) while the other dish served as a control. 719 Parasitemia was analyzed by flow cytometry at 1, 3, 5, and 7 days, when most of the 720 parasites were at the trophozoite stage. After analysis on day 5, cultures were diluted 721 10-fold into fresh RBCs to prevent overgrowth. Medium with or without rapalog was 722 changed daily. 723 For growth analysis of TGD-based KO lines, synchronous ring stage cultures were 724 allowed to mature to trophozoites for one day. Parasitemia was then determined one 725 day post-infection by flow cytometry and adjusted to exactly 0.1% starting 726 parasitemia in a 2 ml dish. Medium was changed daily and growth of the parasite 727 28 lines was assessed by flow cytometry after five days (two erythrocytic cycles). As a 728 reference, WT 3D7 parasites were included in each assay. 729

For quantification of developmental stage and schizont analysis of PI-PLC-GFP-KS 730
and PNPLA2-KO lines, synchronous ring stage cultures were diluted to ~1-2% 731 parasitemia in 2 ml dishes, which were either left untreated or treated with rapalog as 732

Statistical analysis 888
For statistical analysis of differences between two groups, paired or unpaired two-889 tailed students t-tests were used. For statistical analysis of differences between more 890 than two groups, a one-way analysis of variance (ANOVA), followed by a Holm-Sidak 891 multiple-comparison test was performed. All statistical tests were done in GraphPad 892 Prism. P values of <0.05 were considered significant. Statistical details (n numbers, 893 tests used, definition of the error bars) are described in the figure legends.