A bacterial type Cls is expressed in P. falciparum

Putative Cls from apicomplexan parasites form a clade distinct from the bacterial and kinetoplastid Cls in phylogenetic analysis of selected PLD-type Cls, with Plasmodium spp. forming a tight cluster (Fig 1C). The putative bacterial-type Cls of P. falciparum (PlasmoDB ID: PF3D7_0609400) has 17% and 23% sequence identity with Streptococcus pneumoniae Cls and E. coli ClsB, respectively (S1A Fig). PfCls contains two highly conserved H(X)K(X)₄D motifs found in the active sites of phospholipase D (PLD), nucleases, and poxvirus envelope proteins [30]. The minimal H(X)K(X)₄D motifs in PfCls are part of two signature PLD domains, R(X)HRK(X)₄D(X)_5–6G(X)₂N and H(X)K(X)₄D(X)₆G(X)₂N(X)D(X)₂S(X)_4–5E, that are common to most bacterial Cls (Figs 1D and S1A) [17,31]. Cls homologs in kinetoplastids such as Trypanosoma brucei and apicomplexan parasites including Cryptosporidium parvum have conserved PLD domains (S1A Fig). The AlphaFold [32] generated model of PfCls indicates conservation of folds with an unstructured loop from residues 222–279 and low model confidence score at the N- and C-termini of the protein (S1B Fig). The active site pocket with the HKD motifs of PfCls is similar to known structures of the PLD family proteins (Fig 1E). A mitochondrial targeting sequence (MTS) is not identified at the N-terminus (MitoProt II score: 0.34).

In order to check expression of the Cls gene in P. falciparum, we generated antibodies against the C-terminal domain (330–590 aa) of the protein in rabbit (Figs 1D and S1C). Specificity of the generated Abs was confirmed by probing lysates from E. coli cells expressing PfCls-CTD and those transformed with the empty vector (S1D Fig). The Abs were used to detect PfCls in western blots of the parasite lysate from asexual blood stages. A major band of ~68 kDa was detected by the anti-PfCls serum together with an upper fainter band, the latter corresponding to the expected size of full-length PfCls (~71 kDa) (Fig 1F). The lower band is likely to be a processed form of the protein although another post-translational modification is not ruled out. The single cls gene identified in P. falciparum lacks an intron, thus minimizing the likelihood of an isoform. No signal was seen in control western blot with rabbit pre-immune serum. PfCls is thus expressed in the parasite blood stages.

Cls is an integral membrane protein in bacteria and eukaryotic cells with distinct transmembrane domains identified in E. coli ClsA [23], Saccharomyces cerevisiae [33] and Trypanosoma brucei Cls [17]. E. coli ClsB and C lack predicted transmembrane helices but seem to be associated with the bacterial membrane [23]. Although a hydrophobic stretch between residues 163–185 is detected in PfCls, a clear prediction is not made for presence of a transmembrane domain (S2A Fig). In immunofluorescence assays (IFA) of E. coli cells expressing recombinant PfCls, the protein localized at the cell periphery suggesting that it is associated with the bacterial membrane (S2B Fig). Differential extraction of PfCls from P. falciparum trophozoites showed partitioning of the protein in the Tris-soluble, membrane peripheral (carbonate soluble), membrane intrinsic and insoluble (Triton X-100 soluble and insoluble) fractions compared to the control protein β-actin which expectedly partitioned in the Tris-soluble and carbonate-soluble fractions alone (S2C Fig). PfCls partitioning across fractions with differential solubility suggests that it primarily exists as a membrane-associated protein with an additional presence outside the cellular membrane component. In vivo DSP-crosslinking followed by DTT-mediated release showed PfCls as part of a large complex (S2D Fig), suggesting that it forms homo- and/or hetero-meric membrane-associated complexes.

PfCls and cardiolipin have stage-dependent distribution in the mitochondria and cellular periphery in P. falciparum asexual and sexual blood stages

Subcellular partitioning of PfCls to the trophozoite organellar fraction was indicated by the thermolysin protection assay (S2E Fig) where PfCls and the apicoplast-resident PfHU were protected from digestion by the protease in digitonin-treated cells, while the cytosolic control protein tubulin was degraded. All three proteins were intact when thermolysin was inhibited by EDTA. The partial thermolysin protection of PfCls compared to PfHU, a resident of the four membrane-bound apicoplast, suggested that PfCls localized to another organelle which was differentially susceptible to digitonin permeabilization. The localization of PfCls across asexual blood stages and gametocytes was determined by immunofluorescence assays (IFA) using anti-PfCls Ab. PfCls signal overlapped with the mitochondrial marker dye (Pearson’s coefficient = 0.756 ± 0.11) in the mid- to late-trophozoite and schizont stages, localizing within the elongated/branched mitochondrion in trophozoites (Fig 2A) and partitioning to individual merozoites in late schizonts (Fig 2B). No overlap of the PfCls signal was seen with the apicoplast marker PfHU (Fig 2B). PfCls thus localized to the mitochondrion in asexual blood stages. In gametocytes (stage III-V), the PfCls signal overlapped with the mitochondrial marker, but prominent signal was also seen in the cytosol and towards the cell periphery (Fig 2C and 2D). PfCls signal partially overlapped with tubulin below the parasite membrane (Fig 2D). These results suggested that PfCls is completely mitochondrial in asexual blood stages but has additional extra-mitochondrial localization in late-gametocytes, where it also accumulates at the cytosolic periphery. The partial thermolysin protection of PfCls compared to apicoplast PfHU in blood stages could be a result of differential permeabilization of the two-membrane bound mitochondrion compared to the four membrane-bound apicoplast by digitonin.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 2. Subcellular localization of PfCls and distribution of CL through asexual and sexual blood stages.

(A, B) Immunofluorescence confocal microscopy for localization of PfCls in asexual blood stages using anti-PfCls antisera. MitoTracker Red is the mitochondrial marker dye in (A). PfHU is the apicoplast marker in (B), DAPI is the nuclear dye. DIC, differential interference contrast; MT, mid-trophozoites; LT, late-trophozoites; S, schizonts; MS, Mid-schizonts; LS, late-schizonts. The co-localization rate (CR, %) and Pearson’s coefficient (PC) between PfCls and MitoTracker Red is given in (A). The scale bar is 4 µm. The images are representative of >20 images scanned from three sets each for (A) and (B). (C) Localization of PfCls at different gametocyte stages (II-V) with MitoTracker Red as the mitochondrial marker dye. The images are representative of >50 scans from two experiments. (D) Partial overlap of PfCls with α-tubulin signal below the parasite plasma membrane in gametocytes. The images are representative of >50 scans from two experiments. (E) Blood stage trophozoites stained with NAO and MitoTracker Red. The scale bar is 2 µm. The images are representative of >30 images. (F) Gametocytes (stage III-V) stained with NAO and MitoTracker Red. The scale bar is 4 µm. The images are representative of >90 images scanned for gametocyte stages.

https://doi.org/10.1371/journal.ppat.1014215.g002

The unusual dual localization of PfCls in gametocytes prompted us to check the cellular distribution of CL in blood stage parasites. 10-N-nonyl acridine orange (NAO), a high affinity probe for CL detection which has low affinity for other anionic phospholipids and does not bind zwitterionic phospholipids, was used with MitoTracker Red. NAO emits green fluorescence at 525 nm at low levels of CL [34–36]. NAO signal overlapped with MitoTracker Red in trophozoites (Fig 2E). However, in gametocytes, NAO fluorescence co-localized with the mitochondrial dye and also appeared at the parasite plasma membrane (PPM) (Fig 2F). A more prominent membrane signal for NAO was seen at late (stage IV/V) gametocytes, suggesting the presence of CL in the PPM in late gametocyte stages. NAO signal in the PPM of gametocytes, but not of trophozoites, indicated that weak NAO binding to anionic phospholipids other than CL contributed only minimally to the observed fluorescence; levels of anionic phospholipids such as PG and PS actually decrease in late gametocytes compared to trophozoites [3,7].

PfCls has phospholipase D activity and can enhance CL levels in E. coli

In order to investigate the biochemical activity of PfCls, we first expressed full-length PfCls (1–603 aa) as 6X-His tagged protein in E. coli. However, the protein was highly unstable and could not be purified. We then expressed a slightly shorter version (1–552 aa) as a 6XHis-fusion protein (S3A Fig), removing the C-terminal end which is not conserved in Cls homologs from other genera (S1A Fig). Although the levels of PfCls expression in E. coli were low, it purified as two closely migrating bands at the expected size (~66 kDa) with a minor degraded product of >25 kDa (Fig 3A). The two bands at ~66 kDa were recognized by ant-6XHis Abs (S3A Fig) indicating that the lower band was an N-terminal degradation product/processed form of the full-length protein. The two bands could not be cleanly separated and the affinity-purified PfCls comprising both forms was used in activity experiments. Upon size exclusion chromatography, PfCls eluted at the expected size of the globular protein, with peak 1 containing the main bands of ~66 kDa and peak 2 with additional degradation products (S3B Fig). Non-aggregation of PfCls was indicated by the absence protein elution in the void volume (before 8 ml) (S3B Fig). Peptide mass spectrometry of the peak 1 sample confirmed the correct identity of the protein (S1 Table). Secondary structure of purified PfCls was confirmed by CD spectroscopy; CD spectrum of PfCls showed a dip at 222 nm indicating the expected presence of α-helices (S3C Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 3. Biochemical characterization of recombinant PfCls.

(A) Coomassie-stained SDS-PAGE of purified recombinant PfCls (i), and the corresponding western blot probed with anti-PfCls serum (ii). (B) Protein-lipid overlay assay to check binding of recombinant PfCls with different concentrations of phospholipids: PG, PE, CDP-DAG and CL. Buffer alone served as negative control; the buffer column is shown in both panels of the blot to show similar exposure. The blot was probed with anti-6XHis Ab. The corresponding graph shows quantification of concentration-dependent binding from three repeat protein-lipid overlay experiments. Mean ± SD are plotted. * indicate P = 0.045 and 0.029 for comparison of binding to CDP-DAG versus PG and PE, respectively. (C) Co-sedimentation of PfCls with different phospholipid (PL) liposomes. The supernatant (S) and pellet (P) were probed with anti-6XHis Ab for detection of PfCls. The graph shows quantification of lipid-bound PfCls (%). Mean ± SD from three repeat experiments is plotted. (D) Conversion of CDP-DAG to PA indicates phospholipase D activity of PfCls. Lanes 1-5 of the TLC are lipid standards. P. syringae Cls was used as positive control. CMP, cytidine monophosphate. (E) PfCls catalyzes formation of PG from PE and glycerol. PsCls was the positive control enzyme. Representative images from three repeat experiments are shown for (D) and (E). (F) CL synthesis by PfCls in E. coli. Fluorescence confocal microscopy of late exponential phase bacteria transformed with pET23a(+)-PfCls (PfCls) or the empty vector pET23a(+) (EV) under identical scan settings. CL was detected by NAO staining. CL-rich regions in PfCls expressing cells are shown by solid arrows. Scale bar is 2 µm. (G) Mean fluorescence intensity calculated for CL-rich regions-- cell septum and cell poles of bacteria in (F). Mean ± SD are plotted for 40 cells each. P value (P < 0.0001) was calculated by the Mann-Whitney non-parametric two-tailed test. (H) 2D-TLC for detection of CL levels in E. coli expressing PfCls or transformed with the empty vector. CL, PE and PG were used as phospholipid standards. Dimensions of the mobile phase are indicated by solid arrows. CL levels are plotted as % of total phospholipids from three experiments; P < 0.01.

https://doi.org/10.1371/journal.ppat.1014215.g003

Recombinant PfCls was used to assess binding with known phospholipid substrates PE, PG, CDP-DAG and the product CL by the protein-lipid overlay assay (Fig 3B) and liposome co-sedimentation assay (Fig 3C). In the former assay, PfCls interacted with all lipids in a lipid concentration-dependent manner with maximal retention of PfCls on the membrane seen with CL and CDP-DAG (Fig 3B). No signals were seen with the buffer alone or with the negative control probe protein PfExo [37] (S3D Fig). Of the three lipid substrates, PfCls showed significantly higher binding to CDP-DAG (Fig 3B). To further confirm PfCls binding to lipid layers, PfCls interaction with liposomes comprising of different phospholipids was checked by co-sedimentation. In the absence of liposomes, most PfCls was in the soluble fraction with low amounts in the pellet (Fig 3C). PfCls co-sedimented with the liposome pellet for all lipids tested, with comparable (78–96%) partitioning in the pellet fraction (Fig 3C). PfCls can thus bind potential lipid substrates as well as its product, cardiolipin.

The biochemical activity of PfCls was tested using different known phospholipid substrates for bacterial Cls in standard assay conditions followed by thin-layer chromatography (TLC) detection [18]. Although CL formation was seen with the positive control P. syringae Cls (PsCls) (S3E Fig), we could not detect CL formation even with five times higher concentration of PfCls in the reactions with PG alone or PG and PE as substrates (S3F Fig). PfCls, however, was able to hydrolyze CDP-DAG to PA, indicating that it had PLD activity as observed in PsCls (Fig 3D) [18]. PfCls could additionally convert PE and glycerol to PG by transphosphatidylation (Fig 3E), a catalytic property also found in E. coli ClsB [22]. PE formation from CDP-DAG and EA was not detected by either purified PsCls or PfCls (S3G Fig). Our failure to detect CL generation by recombinant PfCls from PG and PE in vitro might be attributable to weak activity of the enzyme and/or high CL detection limit in TLC staining.

To determine whether PfCls supports CL synthesis in vivo, we compared CL levels in E. coli cells expressing recombinant PfCls or transformed with the expression vector backbone alone, and harvested in the late exponential phase of bacterial growth. CL was detected by NAO staining. While NAO emits green fluorescence, high CL levels cause aggregation of the dye resulting in Stokes shift of fluorescence from 525 nm to red fluorescence at 640 nm [34]. Bacterial membrane phospholipids comprise of 70–80% PE, 20–25% PG and ~5% CL which show heterogeneous distribution in the plasma membrane. PE and PG localize to cylindrical regions of rod-shaped bacteria such as E. coli [38–40], with CL tending to localize to regions with high negative curvature which are poles and septum in bacteria [12,41]. In comparison to vector control, which showed NAO signal mostly at green emission, PfCls-expressing cells had higher NAO fluorescence at both green and red wavelengths indicative of enhanced CL levels (Fig 3F). NAO fluorescence in PfCls-expressing cells was localized to the plasma membrane with intense signals at the bacterial poles and septum (Fig 3F). Red NAO fluorescence at the bacterial poles was significantly higher in PfCls-expressing cells (Fig 3G). Quantification of bacterial CL after two dimensional-TLC also showed that PfCls-expressing cells had significantly higher CL content (9.5 ± 1.9% of total lipid) compared to bacteria transformed with empty vector (3.3 ± 0.6% of total lipid), P < 0.01 (Fig 3H). Thus, the presence of PfCls in E. coli enhances CL levels in the bacterial cell membrane, suggesting its direct role in CL synthesis. However, the possibility that increase in CL is the result of an indirect effect, such as stimulation of an EcCls, cannot be ruled out.

Generation of PbCls knockout and complemented parasite lines

There is 63% sequence identity between P. falciparum and P. berghei Cls, with conservation of the PLD domain and HKD motifs (S1A Fig). We thus investigated the role of Cls in the Plasmodium life cycle, by generating a knockout parasite line using a double-crossover homologous recombination strategy (S4A Fig). After transfection, recombinant parasites were selected with pyrimethamine, and integration of the GFP selection cassette was confirmed by fluorescence microscopy (S4B Fig). Site-specific disruption of the PbCls locus was validated by diagnostic PCR (S4C Fig). Clonal parasite lines were obtained by limiting dilution from two independent transfections and used for subsequent analyses. To assess the specificity of the knockout phenotype, we generated a complemented line by reintroducing the PbCls coding sequence under its endogenous regulatory elements at the disrupted locus (S4D Fig). Restoration of the wild-type locus was confirmed by diagnostic PCR (S4E Fig). The complemented parasites (PbCls-comp) displayed no detectable defects and phenocopied WT-GFP parasites throughout the parasite life cycle. PbCls in wild type parasites was cross-recognized by the anti-PfCls Ab in western blot (S5A Fig). Fluorescence microscopy revealed strong Cls signals in both blood-stage and late liver-stage parasites. In contrast, no signal was observed in PbCls knockout parasites (S5B and S5C Fig).

PbCls is required for efficient blood-stage propagation and maintenance of mitochondrial respiration

To investigate the role of PbCls during asexual blood-stage development, we intravenously injected Swiss albino mice (n = 5 per group) with 0.5% parasitemia (200 μl) of either WT-GFP or PbCls KO parasites. Parasite proliferation was monitored daily via Giemsa-stained thin blood smears. Compared to WT-GFP parasites, PbCls KO parasites exhibited significantly reduced blood-stage growth (Fig 4A). This impaired replication was also reflected in reduced gametocytemia; however, by day 4 post-infection, gametocyte levels in the KO group approached those of WT-GFP parasites (Fig 4B). We confirmed that the observed phenotype is not due to outgrowth of a minor WT population by genotypic validation of parasites recovered at the end of blood stage development. PCR analysis confirmed the KO locus with no detectable WT contamination (S5D Fig). To determine the effect of Cls depletion on CL levels, we analyzed CL molecular species by targeted lipidomic analysis. There was a significant reduction in total CL in PbCls KO parasites compared to WT (Fig 4C). A number of CL molecular species were reduced in the PbCls KO (Fig 4D). Among the CL species detected, there was no preferential effect on species with short/long chain length or low/high saturation. We also analyzed levels of other phospholipids—PE, PG, PA, PC and PS which have roles in mitochondrial dynamics and function. While there was no significant change in abundance of PG, PA and PC in the PbCls KO, PE and PS abundance was significantly reduced (S6A–S6I Fig); the majority of detected PE and PS molecular species had lower abundance in the PbCls KO (S6G and S6I Fig).

Download:

• PNG
  larger image
• TIFF
  original image

Fig 4. PbCls is required for asexual blood-stage growth and mitochondrial function.

(A) Asexual growth of PbCls KO parasites is significantly reduced compared to WT-GFP and PbCls-complemented lines. Parasitemia was monitored daily following infection; KO parasites exhibited a marked proliferation defect (day 1, P = 0.0308; day 2, P = 0.4687; day 3, P = 0.0978; day 4, P = 0.0009; day 5, P = 0.0001; day 6, P = 0.0009; day 7, P = 0.0002). (B) Gametocytemia in PbCls KO parasites is significantly decreased on day 3 post-infection compared to WT-GFP (P < 0.0001), but recovers by day 4 (P = 0.1471). (C) Relative total CL levels in PbCls KO compared to WT-GFP parasites. P-value was determined by unpaired Student’s t-test (P = 0.0054). (D) Relative abundance of CL molecular species in WT-GFP and PbCls KO parasites. Mean ± SD of six independent biological replicates are plotted. Multiple t-test with 5% Benjamini-Krieger-Yekutieli FDR correction was applied; adjusted P-values are depicted. (E and F) Mitochondrial membrane potential was assessed using MitoTracker Red FM staining. PbCls KO parasites showed a significant reduction in relative fluorescence intensity compared to WT-GFP (P < 0.0001). Error bars represent mean ± SEM; data are representative of three independent experiments. Statistical comparisons were performed using one-way ANOVA. (G) Basal OCR of WT-GFP and PbCls KO parasites from S6F Fig showing mean ± SD of data points for basal respiration (between 1-20 min). P-value was determined by unpaired Student’s t-test (P < 0.0001).

https://doi.org/10.1371/journal.ppat.1014215.g004

To determine whether the impaired blood-stage growth was associated with mitochondrial dysfunction, we detected relative mitochondrial polarization using the cationic dye MitoTracker Deep Red FM. WT-GFP parasites treated with the protonophore carbonyl cyanide m-chlorophenyl hydrazone (CCCP), which collapses mitochondrial membrane potential (ΔΨm), showed near-complete loss of MitoTracker fluorescence and served as a depolarization control (Fig 4E and 4F). PbCls KO parasites displayed significantly reduced mitochondrial fluorescence compared to WT-GFP, though the signal remained higher than in CCCP-treated parasites. In order to determine whether loss of PbCls causes a defect in the mitochondrial electron transport chain (mETC) thus affecting respiration, we measured the oxygen consumption rate (OCR) using malate as substrate. Comparison of OCR of WT and PbCls KO trophozoites assayed in an XFe24 flux analyzer showed significant reduction (P < 0.0001) of basal OCR in PbCls KO parasites (Figs S6J and 4G). The partial loss of MitoTracker fluorescence as well as reduction in basal OCR indicate that mitochondrial respiration is compromised in the PbCls KO.

Parasite development in the mosquito vector is not affected in the PbCls KO line

To assess the role of PbCls during mosquito stage development, we analyzed the phenotype of PbCls KO parasites. Outgrowth of a minor WT population was ruled out by PCR validation of parasites recovered at the end of mosquito-stages (S5D Fig). In initial experiments, we observed a marked reduction in the number of midgut ookinetes in mosquitoes fed on KO-infected mice (Figs 5A and S7A). However, this was accompanied by reduced gametocytemia at day 3 post-infection relative to WT-GFP controls, suggesting that the observed defects may result from delayed parasite growth rather than a direct role of PbCls in mosquito-stage development. To account for this, we repeated mosquito infections using mice with matched gametocytemia at day 4 post-infection. Under these conditions, ookinete numbers, oocyst formation and midgut sporozoite development were comparable between KO and WT-GFP parasites (Figs 5B–5D, S7B and S7C). Furthermore, salivary gland sporozoite loads at day 18 post-feeding did not differ significantly between the KO and WT-GFP parasites (Figs 5E and S7D). These data indicate that PbCls is not required for parasite transmission through the mosquito vector.

Download:

• PNG
  larger image
• TIFF
  original image

Fig 5. PbCls is critical for late liver-stage maturation but dispensable during early exo-erythrocytic development and mosquito stages.

(A) Significant reduction (P = 0.0002) in ookinete numbers in PbCls KO when mosquitoes fed on mice with low gametocytemia at day 3 post infection. (B) No significant difference in ookinete numbers (P = 0.4762) between WT-GFP, PbCls KO, and PbCls-comp parasites when mosquitoes fed on mice at day 4 post infection. (C) Oocyst quantification showed no significant differences among WT-GFP, PbCls KO, and PbCls-comp parasites (P = 0.5272). (D) Midgut sporozoite numbers were comparable between parasite lines (P = 0.0591). (E) Salivary gland sporozoite quantification revealed no significant differences between WT-GFP, PbCls KO, and PbCls-comp parasites (P = 0.1035). Data represent mean ± SEM from three independent experiments. (F and G) HepG2 cells were infected with WT-GFP or PbCls KO sporozoites and fixed at 24, 40, and 65 hpi and immunostained with anti-UIS4 or anti-MSP1 antibodies; nuclei were stained with Hoechst 33342. (H and I) Quantification of EEF number and size revealed no significant difference between WT-GFP and PbCls KO parasites (EEF number: 24 hpi, P = 0.1901; 40 hpi, P = 0.1557; 65 hpi, P = 0.0959; EEF area: 24 hpi, P = 0.0276; 40 hpi, P = 0.0163; 60 hpi, P = 0.0133). (J) Schizont maturation was severely impaired in PbCls KO parasites at 65 hpi, as evidenced by significantly reduced nuclear counts compared to WT-GFP (P < 0.0001). (K) Counting of EEF stages demonstrated an arrest at the cytomere stage in KO parasites, with a significant reduction in mature merozoites (P = 0.0003). Data represent mean ± SEM from three independent biological replicates. Statistical significance was determined using one-way ANOVA. (L) Detached cell formation is significantly reduced in PbCls KO parasites compared to WT-GFP (P = 0.0018).

https://doi.org/10.1371/journal.ppat.1014215.g005

PbCls is critical for mitochondrial biogenesis and maturation of hepatic merozoites

To determine the role of PbCls during liver-stage development, we first examined the infectivity of PbCls KO sporozoites in vivo. C57BL/6 mice were intravenously injected with 5,000 sporozoites, and parasitemia was monitored by Giemsa-stained blood smears. While all mice eventually developed blood-stage infections, the prepatent period was significantly delayed in mice infected with PbCls KO sporozoites. The delay ranged from 2.6 to 4 days compared to WT-GFP and PbCls-complemented control parasites (Table 1), suggesting a defect in liver-stage development or transition to blood stage.

Download:

• PNG
  larger image
• TIFF
  original image

Table 1. Infectivity of PbCls KO sporozoites in C57BL/6 mice. PbCls KO parasites exhibited a significant delay in the pre-patent period compared to controls (Kruskal-Wallis test, P = 0.0313).

https://doi.org/10.1371/journal.ppat.1014215.t001

To investigate the cause of the observed delay in the prepatent period in PbCls KO parasites, we first quantified hepatocyte invasion by sporozoites using differential staining. No significant difference was observed between PbCls KO and WT-GFP parasites, indicating normal invasion efficiency (S8 Fig). We next examined exo-erythrocytic form (EEF) development in vitro. Cultures were fixed at 24, 40, 65, and 72 hours post-infection (hpi) and stained with anti-UIS4 and anti-MSP1 antibodies. Quantification of EEF numbers and area revealed no significant differences between PbCls KO and WT-GFP parasites (Fig 5F–5I). Analysis of anti-MSP1-stained EEFs revealed reduced nuclear segregation and impaired differentiation into merozoites (Fig 5G, 5J and 5K). Quantification of liver-stage forms further demonstrated a significant accumulation of cytomere forms and reduction in mature merozoites in PbCls KO parasites compared to WT-GFP (Fig 5K). To further investigate this, we quantified detached cells formation at 65 and 80 hpi and consistently observed detached cells in WT-GFP parasites, but not in the PbCls KO parasites (Fig 5L). To functionally assess infectivity, detached cells from WT-GFP cultures or equivalent volumes of PbCls KO culture supernatants were injected into Swiss albino mice (n = 5 per group). All mice receiving WT-GFP-derived detached cells developed blood-stage infections, whereas none of the mice injected with PbCls KO supernatants became patent (Table 2). We then assessed mitochondrial biogenesis in PbCls KO EEFs. Mitochondrial branching was impaired in PbCls KO parasites compared to the extensive branching seen in WT-GFP parasites. Quantification of mitochondrial branching in EEFs confirmed a significant reduction in PbCls KO parasites. While a small subset of PbCls KO EEFs at 40 hpi exhibited mitochondrial morphology comparable to WT-GFP, none of the PbCls KO EEFs analyzed at 65 hpi displayed normal mitochondrial branching (Fig 6A and 6B). Finally, parasitophorous vacuole membrane (PVM) integrity at the late liver stage was analyzed using anti-UIS4 staining. At 65 hpi, WT-GFP parasites showed PVM breakdown consistent with maturation and egress, whereas PbCls KO EEFs retained intact PVMs (Fig 6C and 6D), further supporting a defect in maturation and egress. These results collectively demonstrate that PbCls is dispensable for sporozoite invasion and early liver-stage development but is critical for mitochondrial biogenesis, hepatic merozoite maturation and successful transition to the blood stage.

Download:

• PNG
  larger image
• TIFF
  original image

Table 2. Infectivity of detached cells in Swiss mice.

https://doi.org/10.1371/journal.ppat.1014215.t002

Download:

• PNG
  larger image
• TIFF
  original image

Fig 6. Mitochondrial branching and PVM rupture defects in PbCls KO parasites.

(A) Immunofluorescence analysis using an anti-IscU antibody revealed impaired mitochondrial branching in PbCls KO parasites compared to WT-GFP at 40 and 65 hpi. (B) Quantification of mitochondrial branching in EEFs confirmed a significant reduction in PbCls KO parasites compared to WT-GFP. Mitochondria exhibiting ≥5 at 40 hpi or ≥10 branched tubular networks at 65 hpi were classified as normal, while those below these numbers were considered unbranched. Statistical analysis demonstrated a significant decrease in mitochondrial branching in PbCls KO parasites (P = 0.0052 at 40 hpi; P = 0.0003 at 65 hpi). A total of 100 EEFs at 40 hpi and 50 EEFs at 65 hpi were analyzed per group, from two independent biological replicates. (C) HepG2 cells infected with WT-GFP or PbCls KO parasites were fixed at 65 hpi and immunostained with anti-UIS4 antibody to label the PVM. Nuclei were stained with Hoechst. PVM rupture in WT-GFP is indicated by a white line. (D) Quantification of liver-stage parasites shows a significantly higher proportion of EEFs with intact PVMs in PbCls KO parasites relative to WT-GFP (P < 0.005). A total of 120 EEFs were analysed for each group from two independent biological replicates.

https://doi.org/10.1371/journal.ppat.1014215.g006

Ethics statement

All animal procedures were performed in accordance with the guidelines approved by the Institutional Animal Ethics Committee of Council of Scientific and Industrial Research-Central Drug Research Institute, India (Approval no: IAEC/2018/03 and IAEC/2025/124).

Sequence analysis and phylogeny

Sequences of PfCls and its homologs were obtained from OrthoMCL in PlasmoDB [72] and NCBI database (www.ncbi.nlm.nih.gov). Multiple sequence alignments were performed by Clustal Omega [73]. Transmembrane domain prediction was made using TMHMM [74] and mitochondrial targeting score of PfCls was obtained by MitoProt II [75]. Active site motifs and PLD domain were identified via ScanProsite [76].

A maximum-likelihood unrooted phylogenetic tree was constructed for a total of 19 protein sequences (S2 Table) using LG + R + F model by ATGC: SMS (Smart Model Selection) in PhyML 3. 0 [77], with 1000 bootstrap replicates.

Parasites, mosquitoes, and mice

P. falciparum 3D7 (source: MR4, BEI Resources, USA) were cultured in human RBCs (approval CDRI/IEC/2017/A4 by Institutional Ethics Committee- Human Research) in RPMI-1640 medium (HEPES-modified) (Sigma Aldrich, USA) supplemented with 0.2% sodium bicarbonate, 1% glucose, 100 μM hypoxanthine, gentamycin (25 μg/ml) and 0.5% Albumax II (Invitrogen, USA). P. falciparum 3D7 gametocytes (stage I to stage V) were obtained by initial culture at 1% parasitemia and 4% hematocrit for 12–14 days, with supplementation of RPMI culture medium each day without adding fresh RBCs [78].

Parasites were synchronized by treatment with 5% w/v D-sorbitol (Sigma-Aldrich, USA). Parasite lysate was prepared by releasing parasites from RBC by 0.5% saponin lysis, followed by washing with cold phosphate buffered saline (PBS) containing protease inhibitor cocktail (Sigma-Aldrich, USA). The parasite pellet was treated with RIPA buffer (50 mM Tris-HCl pH 7.5, 300 mM NaCl, 0.1% Triton X-100, 0.1% SDS, 0.5% sodium deoxycholate) and sonicated for 5 min at 25% amplitude.

P. berghei ANKA (MRA-311) and P. berghei ANKA GFP (MRA-867, 507 m6cl1; WT-GFP) strains were obtained from BEI Resources (USA). Anopheles stephensi mosquitoes were reared under controlled conditions at 28°C and 80% relative humidity. The parasite life cycle was maintained using 6–8 week-old Swiss albino mice. For in vivo sporozoite infectivity assays, 4–6 week-old C57BL/6 mice were used.

Recombinant protein expression and purification

The 1.6 kb gene segment encoding PfCls (1–552 amino acids) was PCR-amplified using specific primers (S3 Table) with P. falciparum 3D7 genomic DNA as template and cloned into EcoRI and HindIII sites in the pET23a(+) expression vector carrying a 6X-His tag. The clone was confirmed by DNA sequencing. To purify PfCls, E. coli BL21-Codon plus (DE3) were transformed with pET23a(+)-PfCls and grown in LB media supplemented with ampicillin (100 μg/ml) and chloramphenicol (25 μg/ml) at 37˚C until OD₆₀₀ of 0.6 was reached. Culture was then induced with 0.5 mM IPTG at 20˚C for 16 h. The cells were harvested and the induced culture pellet was re-suspended in lysis buffer with an ionic detergent (50 mM Tris-HCl, pH 8.0, 250 mM NaCl, 10% glycerol, 1mM PMSF, 0.3% N-lauryl sarcosine) followed by 30 min incubation at 4 ˚C. Cells were lysed by sonication and centrifuged at 10,733 × g for 45 min at 4˚C. The clarified cell lysate was collected and loaded on pre- equilibrated Ni-nitrilotriacetic acid (Ni-NTA) Superflow (Qiagen, Germany) affinity column followed by sequential washing with buffer (50 mM Tris-HCl, pH 8.0, 250 mM NaCl, 10% glycerol) containing 10 mM and 20 mM imidazole and elution with buffer containing 300 mM imidazole. The eluted protein was checked by SDS-PAGE and western blotting with anti-6XHis Ab (Santa Cruz Biotechnology, USA). After concentration and buffer exchange using Centricon filters (Merck, USA) the protein was quantified by Bradford assay (Sigma-Aldrich, USA). P. syringae Cls [18] was purified from the pET28b-PSPTO_0095 construct by Ni-NTA purification as described for PfCls above.

Affinity-purified PfCls was injected in Superdex 75 (S75) 10/300 GL column (GE Healthcare, USA) on an AKTA (UNICHROMAT 1500; GE Healthcare, USA) for size exclusion chromatography. The column was pre-equilibrated with buffer (50 mM Tris-HCl pH 8.0, 200 mM NaCl, 5% glycerol). Different fractions were collected, concentrated and separated by SDS-PAGE. For sizing, standard protein markers conalbumin (75 kDa), ovalbumin (43 kDa), carbonic anhydrase (29 kDa) and ribonuclease A (13.7 kDa) (GE Healthcare, USA) were separately fractionated on the same column.

Solution peptide mass spectrometry of purified PfCls was carried out at Valerian Chem Private Limited, India. Briefly, the sample was first reduced with 5 mM TCEP and then alkylated with 50 mM iodoacetamide. After digestion with Trypsin for 16 h at 37°C, the mixture was purified using a C18 silica cartridge and concentrated. The dried pellet was re-suspended in buffer A (2% acetonitrile, 0.1% formic acid). Mass spectrometric analysis of the peptide mixture was performed on an Easy-nlc-1000 system coupled with an Orbitrap Exploris 480 mass spectrometer (Thermo Fisher Scientific, USA). The mass/charge values were analysed through Thermo Proteome Discoverer (v2.5) against the Uniprot Reference database. For Sequest and Amanda search, the precursor and fragment mass tolerance were set at 10 ppm and 0.02 Da, respectively.

For generation of antibodies against PfCls, the C-terminal segment (330–590 amino acids) of PfCls (cloned in pET21d(+)-PfCls-CTD, gift from Dr. Niti Kumar) was expressed in E. coli BL21-Codon plus (DE3). Since the protein appeared in the insoluble fraction, inclusion bodies of PfCls-CTD were purified by washing with buffer (50 mM Tris-HCl, pH 8.0, 250 mM NaCl) containing 0.05% TritonX-100. The protein was electrophoresed on 12% SDS-PAGE and stained with Coomassie G-250. The 30.8 kDa PfCls-CTD protein band was cut from the gel followed by destaining and electroelution in 10 mM EDTA in a 14 kDa-cutoff dialysis membrane.

Circular dichroism spectroscopy

CD spectrum of PfCls (5 µM protein in 20 mM Tris, pH 8.0,150 mM NaCl,10% glycerol) was recorded at 25°C in a CD spectropolarimeter (model JASCO-1500). Spectral scans with a total of three accumulations were recorded over a range of 190–250 nm. Scanning speed was set 50 nm/min with a 1 cm path length.

Raising of polyclonal antisera

Purified PfCls-CTD was dialyzed in 1xPBS and to immunize rabbit (New Zealand White) (Institutional Animal Ethics Committee approval no. IAEC/2007/126/(292–5/18) Ren11). Prior to immunization, serum from the rabbit was checked against parasite lysate to confirm absence of cross-reacting antibodies. An emulsion of equal volume of protein (300 µg) and Freund’s complete adjuvant was injected subcutaneously as primary immunization. The first booster dose of PfCls (100 µg) in Freund’s incomplete adjuvant was given 28 days post-immunization with the second booster dose after 10 days of the first booster. The rabbit was bled to obtain anti-PfCls polyclonal antisera after 10 days of the second booster. The antiserum (1:1000) was used to detect PfCls in P. falciparum lysate in western blots with HRP-conjugated goat anti-rabbit IgG (1:5000, Sigma Aldrich, USA) as secondary Ab. The blots were developed using the ECL kit (Merck Millipore, USA).

Differential membrane fractionation

Membrane fractionation was carried out as described by Tanveer et al. (2013) [79]. P. falciparum trophozoites were incubated in lysis buffer (50 mM Tris-Cl pH 7.4, 2 mM EDTA, 1X protease inhibitor cocktail) for 30 min on a tube rotator and centrifuged at 26,500 × g for 40 min to separate soluble proteins from the insoluble fraction. The Tris-insoluble fraction was resuspended in carbonate buffer (0.1 M sodium carbonate pH 11, 1 mM EDTA), incubated for 30 min followed by centrifugation at 26,500 × g for 40 min to extract extrinsic membrane proteins in the supernatant. To extract intrinsic membrane proteins from the carbonate insoluble pellet, the pellet was treated with 2% Triton X-100 in PBS for 30 min. The soluble fraction obtained after centrifugation at 26,500 × g for 40 min represents integral membrane proteins while the pellet fraction has insoluble protein. The pellet was suspended in 1X PBS in the same volume as the supernatant. Equal volume of sample from all fractions were loaded on SDS-PAGE. For western blotting, the membrane was cut and probed separately with anti-PfCls antisera (1:1000) and anti β-actin antibody (1:50000, Sigma Aldrich, USA).

Chemical cross-linking

Parasite proteins were cross-linked in vivo with DSP (dithiobis(succinimidyl propionate); Merck, Germany) as described in [80]. Briefly, infected RBCs were resuspended in PBS containing 2 mM DSP at room temperature. After incubation for 30 min, the reaction was stopped by addition of quenching solution (final concentration: 25 mM Tris-Cl pH 7.5, 100 mM NaCl) and incubated for 15 min. Parasites were released by saponin lysis, resuspended in NRSB buffer [0.05 M Tris-Cl pH 6.8, 10% glycerol, 2 mM EDTA, 2% SDS, bromophenol blue] with different concentrations of DTT, followed by incubation at 80 °C for 5 min. The samples were electrophoresed in a 6–10% gradient SDS-PAG followed by western blotting with anti-PfCls Abs (1:1000).

Thermolysin protection assay

Thermolysin protection assay was performed as described previously [81,82]. P. falciparum 3D7 parasites (at trophozoite stage) were resuspended in assay buffer (50 mM HEPES-NaOH, pH 7.4, 0.5 mM CaCl₂, 300 mM sorbitol) containing no detergent (control), 0.05% digitonin or 0.05% digitonin +10 mM EDTA. After incubation at room temperature for 10 min, thermolysin (Sigma Aldrich, USA) was added to a final concentration of 25 µg per 1 mg of parasite protein. Reactions were incubated at room temperature for 35 min and 15 min on ice. The reactions were stopped by addition of 10 mM EDTA. Western blotting was done using anti-PfCls sera (1:1000), mouse anti-PfHU sera (1:250) [83] and anti-α-tubulin Ab (1:500) (Sigma-Aldrich, USA).

Immunofluorescence assay and confocal microscopy

P. falciparum infected RBCs at the trophozoite, schizont or gametocyte stages were processed for immunofluorescence assay. For mitochondrial staining, cells were incubated in 50 nM MitoTracker Red CMXRos (Invitrogen, USA,) for 30 min at 37°C. Parasite cells were fixed in 1XPBS containing 0.0075% glutaraldehyde (prepared in 4% paraformaldehyde) for 30 min, followed by permeabilization with 0.2% Triton X-100 for 20 min at room temperature. Permeabilized cells were washed with 1XPBS and blocked in 3% BSA (in PBS). Cells were then incubated overnight at 4°C with primary antisera PfCls (1:50), mouse anti-PfHU (1:100) [83] and anti-tubulin Ab (1:250; Sigma-Aldrich, USA). Alexa Fluor 568-tagged anti-rabbit Ab and Alexa Fluor 488-tagged anti-mouse Ab (1:1000; Invitrogen, USA) were used as secondary antibodies. Alexa Fluor 514-tagged anti-rabbit (1:1000; Invitrogen, USA) was used as the secondary antibody for PfCls in MitoTracker labelled cells. DAPI (0.2 µg/ml) was used for nuclear staining. Cells were seeded on poly L-lysine coated coverslips, allowed to adhere for 2 h, washed three times with 1XPBS, and coverslips were mounted in mounting media (Oncogene Research Products, USA). Samples were scanned under a 63X or 100X/1.4 oil immersion objective in Leica-SP8 confocal microscope. Co-localization was analyzed by the Leica Las X software.

For IFAs with P. berghei, parasites at the blood and liver developmental stages were harvested and subjected to IFA as previously described [84]. Briefly, parasites were fixed in 4% paraformaldehyde for 20 min at room temperature. Blood stage smears were permeabilized with 0.1% Triton X-100 (Sigma-Aldrich, USA) for 10 min at room temperature, while liver stage parasites were permeabilized using chilled methanol. Following permeabilization, samples were blocked with 1% BSA for 1 hour at room temperature. Samples were then incubated with anti-PfCls serum (1:100 dilution) for 1 h at room temperature. After washing, samples were incubated with Alexa Fluor 594-conjugated anti-rabbit IgG secondary antibody (1:1000 dilution; Invitrogen, USA). Parasite nuclei were stained with Hoechst 33342.

For IFA in E. coli, bacteria transformed with pET23a(+)-PfCls or the vector alone were fixed with 2.8% formaldehyde and 0.04% glutaraldehyde in1XPBS for 15 min at room temperature and then permeabilized with 0.1% Triton-X-100 for 45 min at room temperature. The cells were washed three times with 1XPBS and incubated in 100 µg/ml lysozyme and 5 mM EDTA for 45 min at room temperature [85]. Cells were incubated with anti-6XHis Ab (1:100, mouse monoclonal; Santa Cruz Biotechnology, USA) overnight at 4°C, followed by incubation with secondary antibody Alexa Fluor 488-tagged anti-mouse Ab (1:1000; Invitrogen, USA) for 2 h at room temperature. DAPI (0.2 µg/ml) was used for nucleoid staining. Cells were seeded on poly L-lysine coated coverslips and scanned on Leica-SP8 confocal microscope under 100X/1.4 oil immersion objective. Identical laser intensity, gain and line accumulation settings were set for comparison of signals in vector alone and PfCls-expressing cells.

Protein lipid overlay assay

Protein lipid overlay assay (dot blot) was performed as described by Lu et al. (2020), and Aktas & Narberhaus (2009) [86,87]. Phospholipids PE (18:1(ρ9-Cis) PE(DOPE)1,2-dioleoyl)-sn-glycero-3-phosphoethanolamine, PG (18:1(ρ9-Cis) PG 1,2-dioleoyl-sn-glycero-3-phospho-(1’-rac-glycerol) (sodium salt), CDP-DAG (18:1 CDP DG 1,2-dioleoyl-sn-glycero-3-(-cytidine diphosphate) (ammonium salt) and CL (18:1,Cardiolipin 1’,3’-bis[1,2-dioleoyl-sn-glycero-3-phospho]-glycerol (sodium salt) (Avanti Polar Lipids, USA) were tested for binding PfCls. 10 µl of different lipid concentrations in buffer (50 mM Tris-Cl pH 8.0, 100 mM NaCl) were spotted onto nitrocellulose membrane (Cytiva, USA), allowed to dry at room temperature for 1 h and blocked with 5% skimmed milk for 3 h. The membrane was washed three times with 1xPBS followed by incubation with 5 µM PfCls overnight at 4°C with gentle rocking. Membrane was washed three times over 30 min in PBST (PBS containing 0.5% Tween 20) followed by detection of bound PfCls using anti-6XHis antibody (1:5000) and HRP-conjugated anti-mouse secondary antibody (1:5000; Sigma-Aldrich, USA). Membrane was washed three times with PBST over 30 min and the signal detected by chemiluminescence (ECL kit, Merck Millipore, USA). Signals were quantified using Image Quant TL (Cytiva, USA).

Liposome co-sedimentation assay

Liposomes were prepared as described by Danne et al. [88]. Briefly, CL, PG, PE and CDP-DAG were dissolved in chloroform and dried under nitrogen gas flow in glass tubes. 1 ml of rehydration buffer (50 mM Tris, pH 8.0, 100 mM NaCl) was added to prepare lipids at a final concentration of 1.5 mM and incubated for 1 h at room temperature. The suspension was bath-sonicated for 5 min. The sonicated lipid suspension was extruded 10 times through an Avanti mini-extruder with 400 nm polycarbonate filter (Avanti Polar Lipids, USA). The particle size of the extruded lipid suspensions was measured by dynamic light scattering using a Zetasizer Nano ZS (Malvern Panalytical, UK). Purified recombinant PfCls was buffer exchanged in rehydration buffer. The interaction between the PfCls and liposomes was measured by incubating 10 µM PfCls with 0.5 mM liposomes of each phospholipid in rehydration buffer at room temperature. BSA (2%) was added to all reaction mixtures to prevent nonspecific interaction of PfCls with liposomes. Negative control tube had PfCls without a phospholipid. Liposome-bound and free proteins were separated by ultracentrifugation at 82,368 × g (Beckmann Optima XPN-100 ultracentrifuge, fixed-angle rotor 70.1Ti) for 1.5 h at 4°C. The supernatant and pellet fractions were separated; to prevent disruption of the pellet, 10 μl volume was left in the pellet fraction. Equal volumes (10 μl) of the fractions were analyzed for the presence of PfCls by western blotting using anti-6XHis antibody.

CL detection by NAO staining

E. coli cells transformed with pET23a(+) or pET23a(+)-PfCls were induced with 0.5 mM IPTG for 12 h at 20°C and incubated with 250 nM of 10-N-nonyl acridine orange (NAO) (Invitrogen, USA) for 1 h at 37°C with gentle shaking [36,41]. Cells were layered on poly L-Lysine coated coverslips, allowed to adhere for 1 h at room temperature, washed three times with 1XPBS, and coverslips were mounted in mounting media. Samples were scanned on Leica-SP8 confocal microscope under 100X/1.40 oil objective. NAO fluorescence was detected at 488 nm/528–563 nm for green fluorescence and 488/600–650 nm for red fluorescence. Identical laser intensity and gain settings were used to compare signals in vector alone and PfCls sets. Line accumulation was set at 1 for green fluorescence and 16 for red fluorescence. Red fluorescence mean intensity at the cell poles was quantified by LAS X Office software.

For CL detection in parasites, iRBCs were incubated with 500 nM of NAO and 50 nM MitoTracker Red CMXRos at 37°C for 1 h with gentle shaking. Cells were seeded on poly L-Lysine coated coverslips, washed three times with 1XPBS and mounted in mounting media. Fluorescence detection was carried out with Leica-SP8 confocal microscope under 63X/1.4 oil immersion objective. NAO fluorescence was detected at 488 nm/528–563 nm.

PfCls in vitro activity assay

Micelle preparation for generating substrates to assay recombinant PfCls activity was carried out as described by Aktas and Narberhaus [87]. Briefly, 10 mg/ml stock solutions of PG, PE, CDP-DAG and CL in chloroform: methanol (1:1) were evaporated with nitrogen gas under vacuum. Dried phospholipids were resuspended in activity buffer (50 mM NaH₂PO₄, 300 mM NaCl, DDM [4CMC], pH-8.0) and Triton X-100 in the Triton: phospholipid molar ratio of 2:1 to solubilize the phospholipids.

For assaying activity with different phospholipid micelle substrates, 5 µM purified recombinant PfCls was incubated with 0.4 mM CDP-DAG, 0.4 mM CDP-DAG with 0.4 mM EA, 0.5 mM PE with 10% glycerol, 0.5 mM PG, or 0.5 mM each of PG and PE for 16 h at 30°C. The total reaction volume was 100 µl. 1 µM of purified P. syringae Cls was used as positive control in the reactions. Total lipids were extracted and analyzed via TLC as described by Vasilopoulos et al. [18]. Lipids were extracted using the Bligh and Dyer method [89]. 375 µl of chloroform: methanol (1:2) was added to the 100 µl reaction volume and vortexed for 30 sec followed by addition of 125 µl of chloroform and mixing by vortexing.125 µl of deionized water was added to separate the organic and aqueous phases. Samples were centrifuged at 100 × g for 10 min. The organic phase was recovered in a fresh tube followed by washing with 125 µl of deionized water. The organic phase was recovered after centrifugation and dried under N₂ gas flow. Dried lipids were resuspended in 20 µl chloroform: methanol (1:1) before spotting on a TLC plate. 5 µl of isolated lipids from activity reactions was spotted on TLC plate (HPTLC silica gel, Merck, USA). Chloroform: methanol: glacial acetic acid: water (3.2:0.2:0.5:0.1 [v/v]) was used as the mobile phase for lipid separation. The TLC chamber was pre-saturation with the mobile phase. After chromatography, the TLC plate was dried at 100°C followed by development with 10% phosphomolybdic acid (Himedia, India) or molybdenum blue spray reagent (Sigma-Aldrich, USA) [18,90].

Lipid extraction from E. coli and two-dimensional TLC (2D-TLC)

E. coli cells transformed with pET23a(+) or pET23a(+)-PfCls were induced with 0.5 mM IPTG and grown to stationary phase (OD_600nm of ∼2.0) in 100ml LB media. Lipid extraction was performed by acidic Bligh and Dyer method [89]. Cells were pelleted and resuspended in 2 ml of 0.1 N HCl followed by addition of 5 ml methanol and 2.5 ml chloroform to form a single-phase solution comprise of chloroform/methanol/0.1 N HCl [1:2:0.8 (vol/vol)]. After incubation at room temperature with gentle mixing, 2.5 ml each of 0.1 N HCl and chloroform were added to convert the single phase to a two-phase solution made up of chloroform/methanol/0.1N HCl [2:2:1.8 (vol/vol)]. The solution was centrifuged at 3000 × g for 25 min at room temperature. The lower organic phase was collected in separate tubes and dried under nitrogen stream. The dried lipids were resuspended in 100 μl of chloroform/ methanol (2:1) and sonicated for 1 min in bath sonicator. 5 µl of each sample was spotted on TLC plate and solvent mixtures of chloroform:methanol:water (65:25:4) and chloroform:methanol:acetic acid:water (90:15:10:3,5) were used as mobile phase for the first and second dimension, respectively. TLC plates were dried at 100°C and developed with 10% phosphomolybdic acid [23,31]. PG, PE and CL were used as phospholipid standards. Phospholipids were quantified by densitometry via Image Quant TL (Cytiva, USA).

Generation of PbCls knockout and complement parasite lines

Targeted disruption and complementation of PbCls in P. berghei ANKA were performed as previously described [91]. For gene disruption, two homology regions were PCR-amplified from P. berghei genomic DNA: a 570 bp upstream fragment (F1) using primers 2101/2102 and a 569 bp downstream fragment (F2) using primers 2103/2104. These fragments were sequentially cloned into the pBC-GFP-hDHFR:yFCU vector. The resulting knockout construct was linearized with XhoI and AscI, and was transfected into purified schizonts using basic parasite nucleofector kit (Lonza) as described previously [92]. Transfected parasites were selected with pyrimethamine, and genomic DNA from drug-resistant populations was screened for correct integration by diagnostic PCR using primer pairs 2183/1225 (5′ integration) and 1215/2184 (3′ integration). Clonal lines were obtained by limiting dilution and screened for the absence of the endogenous PbCls locus using primers 2206/2207. For genetic complementation, a 1,809 bp genomic fragment encompassing the full-length PbCls coding region and flanking sequences was amplified using primers 2101/2104 and reintroduced into the knockout line. Transfectants were selected with 5-fluorocytosine (Sigma-Aldrich, USA) to select against the yFCU marker. Restoration of the wild-type PbCls locus was confirmed by PCR using primers 2206/2207. All primer sequences are provided in S3 Table.

Blood growth assay

To investigate the role of PbCls during the asexual blood stages of P. berghei, female Swiss albino mice (5 mice per group) were intravenously injected with parasitized red blood cells (pRBCs) at 0.5% parasitemia. Parasitemia was monitored daily by preparing Giemsa-stained blood smears.

Mitochondrial membrane potential assay

To assess mitochondrial membrane potential, equal volumes of parasitized blood were collected from each group. Parasites were incubated with 50 μM carbonyl cyanide m-chlorophenyl hydrazone (CCCP), dissolved in DMSO and diluted in 1X PBS, at 37°C for 1 h as a control for mitochondrial depolarization. Following incubation, samples were washed twice with 1X PBS. WT-GFP, PbCls KO, and CCCP-treated parasites were then stained with MitoTracker Deep Red FM at 37°C for 30 min. After staining, cells were washed once with 1X PBS and counterstained with Hoechst 33342 to visualize nuclei. Fluorescence intensity was quantified using Fiji (ImageJ). Mean fluorescence values were calculated per parasite and normalized across biological replicates.

Assessment of PbCls KO in mosquito stage development and sporozoite infectivity

To assess the role of PbCls during mosquito-stage development, A. stephensi mosquitoes were allowed to feed on infected Swiss mice, and parasite development at various stages was analyzed as described previously [93]. At 24 h post-blood feeding, mosquito midguts were dissected, and ookinetes were enumerated by light microscopy. On day 14 post-feeding, midguts were dissected, and the number of oocysts per midgut was counted using a fluorescence microscope. Between days 18 and 23 post-feeding, salivary glands were dissected, homogenized in RPMI-1640 medium, and sporozoites were quantified using a hemocytometer.

To assess the in vivo infectivity of sporozoites, C57BL/6 mice were intravenously injected with 5,000 salivary gland sporozoites. Parasite development was monitored by examining Giemsa-stained thin blood smears. The pre-patent period was determined as the time from injection to the first detection of blood-stage parasites.

In vitro EEF development and IFA

For in vitro EEF development, HepG2 cells were cultured and infected with salivary gland sporozoites. Infected cultures were harvested at various time points, fixed with 4% paraformaldehyde, washed twice with 1 × PBS, permeabilized with methanol, and blocked in 1% BSA (Sigma-Aldrich, USA) for 30 min. Primary antibodies anti-UIS4 (diluted 1:1,000) [94] and anti-MSP1 (diluted 1:5,000) [95] were diluted in blocking solution and incubated for 1 h at room temperature. To reveal UIS4 and MSP1 signals, slides were incubated with Alexa-Fluor 594 conjugated anti-rabbit IgG (diluted 1:000; Invitrogen, USA) and Alexa-Fluor 594 conjugated anti-mouse IgG (diluted 1:1000; Invitrogen, USA), respectively. Nuclei were stained with Hoechst 33342. Nuclei were stained with Hoechst 33342, and slides were mounted using prolong diamond antifade reagent (Invitrogen, USA). Images were captured using a 100X oil immersion objective on an Olympus BX61WI confocal microscope. To assess detached cells formation, 1 × 10⁵ HepG2 cells were seeded in a 24-well plate and infected with 30,000 sporozoites. Culture supernatants were collected at 64 and 80 hpi, and detached cells were quantified using a hemocytometer. To evaluate the infectivity of detached cells, Swiss albino mice were intravenously injected with harvested detached cells or culture supernatant. The onset of blood-stage infection was monitored using Giemsa-stained blood smears. To assess mitochondrial morphology in late liver-stage EEFs, samples were harvested and processed for IFA using an anti-IscU Ab (1:100 dilution) [96]. Signals were revealed using Alexa Fluor 594-conjugated secondary antibody, and nuclei were stained with Hoechst 33342.

Lipidomics

As PbCls KO parasites grow more slowly, comparable parasitemia between WT and KO parasites at harvest was achieved by inoculating mice infected with KO parasites earlier than those infected with WT parasites. Parasite stages (mean±SD %) were ~66 ± 5% rings, 33 ± 4% trophozoites and 1 ± 0.6% schizonts in WT-GFP and ~69 ± 3% rings, 30 ± 4% trophozoites, 0.8 ± 0.6% schizonts in the PbCls KO biological replicates. Parasites were released by saponin lysis, washed and processed for lipid extraction.

Total lipid extraction

Total lipid extractions were done using methods as previously described [97,98]. Briefly, cell pellets resuspended in 1 mL Dulbecco’s Phosphate Buffered Saline (DPBS) were made up to a 4 mL mixture of 2:1:1 choloroform (CHCl₃): methanol (MeOH): DPBS in glass vials [containing 1 nmol of C15:0 monoacylglycerol (C15:0 MAG) and 0.5 nmol of C15:0 fatty acid (C15:0 FA) for semi-quantitative analysis of lipids]. This homogenate mixture was vortexed vigorously and centrifuged at 2500 × g for 15 min to separate the mixture into an organic and an aqueous phase (top) separated by a protein disk. The organic phase (~2 ml) was removed by pipetting and transferred to a new glass vial, while 50 μl of formic acid (MS grade, Fluka-Honeywell, USA) was added to the aqueous phase. This mixture was vigorously vortexed, and 2 ml of CHCl₃ was added, followed again by vortexing. This mixture was centrifuged at 2500 × g for 15 min; the organic layer from this extraction and the previously obtained organic layer were pooled and dried under a stream of nitrogen gas. The dried lipid extracts were re-solubilized in 200 μL of 2:1 CHCl₃: MeOH, and 10 µl of this was used for subsequent LC-MS analysis.

Cardiolipin measurements

Cardiolipin measurement was carried out using multiple reaction monitoring transition-based analysis in an Agilent Technologies 6470 Triple Quadrupole LC/MS. Reverse-phase LC separation was achieved using a Gemini 5U C18 column (Phenomenex, 5 μm, 50 × 4.6 mm, 110 Å) coupled to a Gemini guard column (Phenomenex, 4 × 3 mm). The solvents used for LC separation were buffer A: 95:5 H₂O/MeOH + 0.1% ammonium hydroxide and buffer B: 60:35:5 iPrOH/MeOH/H₂O + 0.1% ammonium hydroxide. LC-MS runs were 30 min post injection, with the following gradient elution parameters: 0.3 ml/min 0% buffer B for 4 min; 0.5 mL/min linear gradient to 100% buffer B from 4-18 min; 0.5 ml/min 100% buffer B from 18-25 min; 0.5 ml/min 100-0% buffer B from 25-25.1 min; and 0.5 mL/min 0% buffer B from 25.1-30 min (re-equilibration). All LC-MS analysis was performed in negative ion mode using ESI source with the following parameters: drying and sheath gas temperature = 350°C and 375°C; drying and sheath gas flow = 11 L/min; nebulizer pressure = 45 psi; capillary voltage = 4000 V; and nozzle voltage = 1000 V. A list of all CL species targeted in this MRM method, describing the precursor parent-ion mass, the product ion targeted, other species-specific voltages and parameters, and the internal standard used can be found in S4 Table. All CL species were quantified by their respective areas under the curve, normalizing it to the area under the curve of the internal standard (C15:0 FA), and then normalizing to cell number.

Untargeted lipid measurements

Untargeted LC-MS runs were performed using the AutoMS/MS acquisition method on an Agilent 6545 Q-TOF mass spectrometer fitted with an Agilent 1290 Infinity II UHPLC system as previously described [99]. Briefly, reverse-phase LC separation was achieved using a Gemini 5U C18 column (Phenomenex, 5 μm, 50 × 4.6 mm, 110 Å) coupled to a Gemini guard column (Phenomenex, 4 × 3 mm). The solvents used for LC separation in positive ionization mode were buffer A: 95:5 H₂O/MeOH + 0.1% HCOOH + 10 mM ammonium formate and buffer B: 60:35:5 iPrOH/MeOH/H₂O + 0.1% HCOOH + 10 mM ammonium formate. The solvents used for LC separation in negative ionization mode were buffer A: 95:5 H₂O/MeOH + 0.1% ammonium hydroxide and buffer B: 60:35:5 iPrOH/MeOH/H₂O + 0.1% ammonium hydroxide. LC-MS runs were 60 min post injection, with the following gradient elution parameters: 0.3 ml/min 0% buffer B for 5 min; 0.5 ml/min linear gradient to 100% buffer B from 5-45 min; 0.5 ml/min 100% buffer B from 45-55 min; 0.5 ml/min 100-0% buffer B from 55-55.1 min; and 0.5 ml/min 0% buffer B from 55.1-60 min (re-equilibration). The autosampler (samples) and column temperatures were maintained 10°C and 40°C, respectively. MS acquisition was conducted using a dual Agilent jet stream electrospray ionization (AJS ESI) source in positive and negative ionization mode. MS parameters: drying gas and sheath gas temperature at 320°C, drying gas and sheath gas flow at 10 L/min, nebulizer pressure at 45 psi, capillary voltage (VCap) at 4000 V, nozzle voltage at 1000 V, and fragmentor voltage at 150 V. Precursor selections were carried out with a maximum of 4 precursor ions per cycle with fixed collision energies set at 5 eV and 15 eV.

For data analysis, a library for different classes of lipids were curated as a Personal Compound Database Library (PCDL) using MassHunter PCDL manager B.08.00 (Agilent Technologies, USA). This library was curated using the METLIN lipid library as reference. The data files were processed in Agilent MassHunter Qualitative Analysis 10.0 using the PCDL library; all the peaks were validated based on relative retention times and fragments obtained, if any. All detected species were within a mass accuracy of 15 ppm and quantified by measuring the area under the curve (AUC) for different lipid species. The values were then normalized to levels of internal standards (C15:0 MAG and C15:0 FA for positive and negative mode respectively) and the cell number for the samples.

Determination of oxygen consumption rate (OCR)

OCR measurement was performed as described [100,101] using Agilent Seahorse XFe24 Analyzer and Agilent XF Cell Mito Stress Kit with some modifications. A sensor cartridge was hydrated at 37°C overnight prior to the experiment. Wild type and PbCls KO parasites primarily at trophozoite stage were harvested with pre-warmed 0.05% saponin and then washed twice with pre-warmed 1XPBS. The cells were resuspended in 1X mitochondrial assay solution (MAS: 220 mM mannitol, 2 mM HEPES, 10 mM malate, 70 mM sucrose, 5 mM MgCl₂,10 mM KH₂PO₄, 1 mM EGTA, 0.2% [wt/vol] fatty acid-free BSA) and cell numbers were adjusted to 5 × 10⁷ cells/ml. Cells were permeabilized with 0.002% (w/v) digitonin (Sigma Aldrich, USA) and 100 µl of this suspension was seeded per well in XFe24 microplate pre-coated with Cell-Tak (Corning, USA). The plate was centrifuged at 800 × g for 10 min. Additional 75 µl of 1XMAS was added without disturbing attached parasites. Inhibitors were prepared in 1XMAS buffer. After placing the sensor cartridge on utility plate, injection ports were filled with 25 µl of inhibitors (Port A: FCCP and Port B: Antimycin at final concentration of 4 µM and 0.5 µM, respectively). The measurement of OCR included 5 cycles of 20 s mixing, 1 min waiting, and 2.5 min for both basal respiration and measurement after injection of inhibitors.

Statistical analyses

All statistical analyses were performed using GraphPad Prism 9 Software. Data were analyzed using unpaired two-tailed Student’s t-test, multiple t-test, one-way ANOVA or Mann-Whitney non-parametric two-tailed test. Statistical significance is indicated as: P < 0.05 (*), P < 0.01 (**), P < 0.001 (***), P < 0.0001 (****); n.s., not significant.