The Plasmodium falciparum 3D7 strain was cultivated at 37°C under 5% O₂, 5% CO₂, and 90% N₂ in O+ human erythrocytes at 5% hematocrit in RPMI 1640 medium (Invitrogen) supplemented with 25 mM HEPES, 50 μg/mL hypoxanthine (Sigma-Aldrich), 2 g/L NaHCO₃ (Gibco), 10 μg/mL gentamicin (Gibco), and 0.5% Microbiological Grade Bovine Albumin (MP Biomedicals).

Plasmids were designed to mediate CRISPR-Cas9 transfection of cultivated parasites to produce conditional knockdowns carrying a tandem 3xHA-DDD encoded as a C-terminal tag ( Figure 1A ). The JB113 plasmid ( Supplementary Figure 1B ) was digested at NotI sites flanking an HSP101 gRNA expression cassette. The modifiable sgRNA expression cassette was obtained by PCR using pL6-eGFP (Ghorbal et al., 2014) and inserted into the NotI-digested JB113 plasmid using Gibson Assembly cloning (New England BioLabs). The sgRNA of RON3 was inserted into the BtgZI-digested modified JB113 plasmid using In-Fusion cloning (Clontech). The yeast dihydroorotate dehydrogenase (yDHODH) and Streptococcus pyogenes Cas9 were expressed bicistronically by a 2A viral skip peptide, and sgRNA was expressed from the modified JB113 plasmid. DSM1 was used to retain the modified JB113 carrying yDHODH cassette (Ghorbal et al., 2014). The JB200 plasmid ( Supplementary Figure 1B ) was digested at XhoI and AvrII sites flanking genomic sequences of HSP101 for homologous recombination repair. Genomic sequences of RON3 for homologous recombination repair were obtained by PCR using genomic DNA as template and inserted into the digested JB200 plasmid carrying a tandem 3xHA-DDD tag using Gibson Assembly cloning. Shield mutation was introduced on the donor sequence at sites corresponding to the genomic sgRNA target using a codon-optimized primer. The modified JB200 plasmid carrying a human dihydrofolate reductase (hDHFR) cassette allows the selection of integrants with TMP. The primers used are listed in a Supplementary Table 1 . DNA sequencing was used to confirm all constructed plasmids.

Transfections were initiated by electroporation of 50 µg of modified JB113, and 20 µg of AflII digested linearized modified JB200 plasmids into 1.75 x 10⁹ uninfected erythrocytes prior to use for parasite cultivation. The Gene Pulsar II system (Bio-Rad) was used at 0.31 kV and 975 μF. Immediately after electroporation, 1 x 10⁷ schizonts purified by differential centrifugation on a 70%/40% Percoll-sorbitol gradient were mixed with electroporated erythrocytes and parasite culture was initiated. The transfected culture was selected with 1.5 µM DSM1 and 10 µM TMP to ensure retention of the JB113 plasmid and select the growth of transgenic parasites expressing hDHFR. After 2 to 3 weeks, parasite growth was detected by Giemsa staining of thin blood films. Parasite clones were obtained by limiting dilution (Lyko et al., 2012), and then PCR was used to evaluate the integration of the donor sequence from the modified JB200 plasmid and the unexpected retention of circular modified JB200 plasmid ( Supplementary Figure 1B , Supplementary Table 1 ). All experiments were performed using parasite clones. Conditional knockdown was initiated by TMP removal before RON3 expression or after erythrocyte invasion.

A synthetic peptide was derived from the C-terminal repetitive sequence of ring-infected erythrocyte surface antigen in P. falciparum (PfRESA; PF3D7_0102200, EENVEENVEENVEENVEENV, with an additional N-terminal cysteine) and was conjugated with commercially procured keyhole limpet hemocyanin (KLH) carrier protein (Sigma-Aldrich). A Japanese white rabbit received a total of three immunizations with Freund’s complete and incomplete adjuvant at 3-week intervals (Kitayama Labes, Ina, Japan). Antisera was collected 14 days after the last immunization. The animal work was conducted by Kitayama Labes in compliance with the guidelines based on “Charter for Laboratory Animal Welfare” (Japanese Society for Laboratory Animal Resources). The anti-RESA antibody was verified for specific reaction to native 3D7 wild-type RESA by immunoblotting, indirect immunofluorescence assay, and immunoelectron microscopy ( Supplementary Figure 2 ). The mouse anti-RESA monoclonal antibody 28/2 (Anders et al., 1987) was used as a positive control.

Synchronous schizont-rich parasite cultures were used after membrane fractionation using hemolysis in lysis buffer (7.5 mM Na₂HPO₄, 1 mM EDTA, 1 mM PMSF, pH 7.5) and ultracentrifugation (100,000 x g, 4°C, 1 h). The supernatant was considered the soluble fraction, and the pellet was considered the membrane fraction. Ring-infected erythrocytes were lysed with 0.15% saponin in PBS containing cOmplete Protease Inhibitor Cocktail (Roche) for 10 min on ice. Samples were solubilized and reduced in a modified Laemmli sample buffer at a final 6% SDS concentration. Proteins were separated by electrophoresis in 5–20% polyacrylamide gels (ATTO) and transferred to PVDF membranes. Membranes were blocked using 3% skim milk in 150 mM NaCl, 20 mM Tris-HCl, pH 7.4 with 0.1% Tween 20. Each membrane was then immunostained at RT for 1 h with rabbit anti-HA antibody (Abcam), rabbit anti-RON3N antibody referred to as anti-PfRON3_1 (Ito et al., 2011), rabbit anti-RON2 antibody (Cao et al., 2009; Ito et al., 2011) for a schizont-specific control, rabbit anti-Plasmodium aldolase antibody (Abcam) as a protein loading control, and rabbit anti-RAP1 antibody (Ito et al., 2011) as a membrane protein loading control. Ponceau-S staining of PVDF membranes was used to visualize hemoglobin as a soluble protein loading control. Specific labeling was revealed by probing with horseradish peroxidase-conjugated secondary-antibody (ThermoFisher Scientific), and visualized with Immobilon Western Chemiluminescent HRP Substrate (Millipore) on an LAS 4000 Mini luminescent-image analyzer (GE Healthcare). The relative molecular masses of the proteins were estimated with reference to Precision Plus protein WesternC standards (Bio-Rad). Semi-quantitative analysis of the protein expression levels was performed by ImageJ and the loading controls were used for normalization.

Indirect immunofluorescence assays were performed using air-dried thin blood smears of RON3-DDD parasites with or without TMP prepared side by side on microscope glass slides, fixed with ice-cold 100% acetone for 3 min, and stored at −80°C. The smears were thawed and blocked with PBS containing 3% skim milk at 37°C for 30 min. The antigen samples for IFA described above were stained with primary antibodies diluted at the following concentrations in blocking solution at 37°C for 1 h: rabbit anti-RESA antibody, 1:500; mouse anti-SBP1 antibody, 1:100 (Iriko et al., 2020); rabbit anti-EXP2 antibody, 1:500 (Miyazaki et al., 2021); mouse anti-EXP2 antibody, 1:100 (Morita et al., 2018). Secondary antibodies, Alexa Fluor 488-conjugated goat anti-rabbit IgG and Alexa Fluor 555-conjugated goat anti-mouse IgG (Invitrogen) or Alexa Fluor 488-conjugated goat anti-mouse IgG and Alexa Fluor 555-conjugated goat anti-rabbit IgG (Invitrogen), were used at a 1:500 dilution in blocking solution at 37°C for 30 min. DAPI (4’,6-diamidino-2-phenylindole) at 2 µg/ml was added to the secondary-antibody solution to stain the nuclei. Slides were mounted in ProLong Gold Antifade reagent (Invitrogen) and viewed under a 63× oil immersion lens. Images were collected on a confocal scanning laser microscope (LSM780; Carl Zeiss MicroImaging) with serial 405 nm, 488 nm, or 561 nm excitations with the same detector settings. Images were processed in Adobe Photoshop (Adobe Systems, CA, USA) or ImageJ. For differential counts of protein export pattern, a minimum of 100 infected cells were counted for each parasite stained with anti-RESA or anti-SBP1 antibodies.

Parasites were fixed for 30 min on ice in a mixture of 1% paraformaldehyde and 0.2% glutaraldehyde in HEPES-buffered saline (pH 7.05). Fixed specimens were washed, dehydrated, and embedded in LR White resin (Polysciences, Inc., Warrington, PA) as described (Aikawa and Atkinson, 1990; Ito et al., 2011). Ultrathin sections on the grids were blocked in PBS containing 5% non-fat skim milk and 0.01% Tween 20 (PBS-MT) at 37°C for 30 min. Grids were then incubated at 4°C overnight with rabbit anti-RESA or pre-immune sera in PBS-MT (1:500). After washing with PBS containing 4% BlockAce (Yukijirushi, Sapporo, Japan) and 0.01% Tween 20 (PBS-BT), the grids were incubated at 37°C for 1 h with goat anti-rabbit IgG conjugated to 15 nm gold particles (BBI Solutions) diluted 1:20 in PBS-MT, and rinsed with PBS-BT. The grids were then rinsed with distilled water, dried, and stained with uranyl acetate and lead citrate. Samples were examined with a transmission electron microscope (JEM-1230; JEOL, Tokyo, Japan).

Comparison of growth of RON3-DDD parasites in the presence of TMP or DMSO was determined by microscopy of Giemsa stained thin blood films. For differential counts, a minimum of 2,000 erythrocytes were counted for each time point.

The invasive capacity of RON3-DDD parasites was determined by adding fresh erythrocytes to Percoll-purified schizonts to obtain a parasitemia of 0.3%. Parasitemia was determined by Hoechst 33342 staining (ThermoFisher Scientific) and analyzed by flow cytometry (LSRFortessa X-20; BD Biosciences) performed with the same detector settings. After 24 h, schizont rupture was monitored by microscopy of Giemsa stained thin blood smears and the cultures enabled the percentage of newly ring-infected RBCs to be similarly determined. Data were analyzed using FlowJo software (BD Biosciences).

Comparison of growth of RON3-DDD parasites after the merozoite invasion of erythrocytes in the presence of TMP or DMSO was determined by microscopy of Giemsa stained thin blood films. The viable free merozoites that retained their invasive capacity were purified based on published protocols (Boyle et al., 2010; Ito et al., 2013). Briefly, parasites were synchronized using sorbitol treatment. Late-stage parasites (32 to 36 h post-invasion) were purified using a Percoll-sorbitol gradient, incubated with 10 μM E64 (Sigma-Aldrich) for 6 h, and pelleted. E64-treated schizonts were resuspended in a small volume of incomplete culture medium and filtered through a 1.2-μm Acrodisc 32-mm syringe filter (Pall Corporation) onto fresh erythrocytes.

Synchronized ring-stage parasites were stained with Glucose Uptake Probe-Green (Dojindo, Kumamoto, Japan) and 600 nM MitoTracker Deep Red (ThermoFisher Scientific) for 1 h at 37°C, after which the cells were washed with phenol red-free RPMI containing Washing and Imaging Solution (Dojindo). Cells were dispensed at 0.06% hematocrit to µ-Slide 8 well chamber slides (ibidi, Gräfelfing, Germany) coated with 0.01% poly-L-lysine (Sigma-Aldrich). Prior to imaging, Hoechst 33342 was added to the wells at a final concentration of 2 µM. Cells were imaged within 1 h by Zeiss LSM 780 with the same detector settings, and images were randomly collected by switching every five images between parasites in the presence of TMP or DMSO. The captured images were processed and analyzed using ImageJ. Mean fluorescence intensity (MFI) was measured by the fluorescence intensity attributed to the parasite (red circle as determined by localization with Hoechst/MitoTracker [blue/red] localization) and the erythrocyte (white circle), respectively, as shown in Supplementary Figure 3 . Background MFI in the erythrocyte was subsequently subtracted from the MFI in the parasite to provide the change in MFI (ΔMFI).

Organic solute uptake by infected erythrocytes was continuously tracked as described (Nguitragool et al., 2011; Ito et al., 2017). Trophozoite-stage infected cells were Percoll-enriched after preincubation in 40% Percoll-sorbitol solution for 30 min at 37°C to allow efficient recovery of knockdown parasites that had reduced permeabilities (Ito et al., 2017). Enriched cells were washed, and resuspended in 150 mM NaCl, 20 mM Na-HEPES buffer, 0.1 mg/mL BSA, pH 7.4. Solute uptake was initiated by the addition of an osmotic lysis solution containing 280 mM sorbitol, 20 mM Na-HEPES, 0.1 mg/mL BSA, pH 7.4. The kinetics of osmotic lysis were monitored at 37°C for 1 h by recording the transmittance of 700 nm light through the cell suspension. The addition of 0.5% saponin at the end of each recording was used for the normalization of transmittance values to 100% cell lysis. Sorbitol permeability coefficients were calculated as 1/halftime of osmotic lysis.

All data were presented as mean ± standard errors of means (SEM). All quantitative data were statistically analyzed and graphed using GraphPad Prism 9 for Mac (GraphPad Software, MA, USA). Statistical significance was calculated by unpaired two-tailed Student’s t-test. Significance was accepted at p<0.05.

To investigate the function of the RON3 C-terminal fragment, we tagged RON3 with a DDD at the C-terminal end using the CRISPR-Cas9 system ( Figure 1A ). Integrations were confirmed by PCR ( Figure 1B ). Using clones RON3-DDD carrying the C-terminal tag, we observed the effect of RON3 knockdown on parasite survival. When the DDD-stabilizing small molecule TMP was removed from asynchronous RON3-DDD cultures, a complete block in parasite growth was observed ( Figure 1C ).

We next determined the stage-specific effects of RON3 knockdown on the asexual blood-stage developmental cycle ( Figure 2A ). Upon removal of TMP from the early schizont stage that expresses RON3 (ES, Figure 2A ), RON3-DDD parasites completed their first intracellular development and invaded new erythrocytes accompanied by some reduction of parasitemia, and parasite development was arrested at the subsequent ring stage (ES, Figure 2B ). To quantify the impact of RON3 knockdown on the invasion of erythrocytes, schizonts were isolated from synchronous cultures with or without TMP (con and ES, Figure 2A ) and then added to fresh erythrocytes. The parasite nuclei were stained with Hoechst 33342, and a similar starting parasitemia was confirmed by flow cytometry ( Supplementary Figures 4A, C ). Following further incubation to allow schizont rupture and erythrocyte invasion, the parasitemia of newly-invaded ring-stage parasites was determined by flow cytometry as well as the starting parasitemia ( Supplementary Figures 4B, D ). Consistent with the growth retardation ( Figure 1C ), the parasites from cultures without TMP displayed significantly reduced ring formation, corresponding to 65% of control cultures with TMP ( Figure 2C , Supplementary Figure 3D ). We hypothesized that a reduction of the invasion rate upon RON3 knockdown might arise from the destabilized DDD interfering with the proteolytic processing of RON3-DDD. To investigate this, immunoblotting of the schizont lysates with an antibody against the RON3 N-terminal fragment was conducted ( Figure 2D ) and revealed that both the unprocessed 280 kDa protein (full-length RON3 (260 kDa) + 3HA-DDD tag (20 kDa)) and an 190 kDa cleavage product were similarly identified in the insoluble fraction of both TMP(+) and TMP(-) parasites ( Figure 2E ), but the ratio of full-length RON3 to the RON3 N-terminal fragment was significantly increased in TMP(-) parasites indicating 30% reduction of both N- and C-terminal fragments (ES, Figure 2F ). Accordingly, conditional knockdown alters proteolytic processing but not protein expression level and solubility of RON3 in schizonts.

TMP removal from the early schizont stage (ES, Figure 2A ) induced not only a defect in erythrocyte invasion but also the growth arrest of the subsequent ring-stage parasites (ES, Figure 2B ). To investigate the impact of DDD destabilization on the ring stage parasite growth, we next conducted immunoblotting of ring-stage parasite lysates with antibodies against the RON3 N-terminal fragment and 3xHA epitope tag ( Figure 2G ), and revealed an 80% reduction of the RON3 C-terminal fragment in TMP(-) parasites at the ring stage as compared to a 40% reduction of the RON3 N-terminal fragment (ES, Figure 2H ), under conditions where RON2 was undetectable indicating no schizonts in the ring-stage parasite lysate ( Figure 2G ). In contrast, when TMP was removed in the early ring stage (0 to 30 min post-invasion) (ER, Figure 2A ), RON3-DDD parasites normally grew to the trophozoite stage (ER, Figure 2B ).

As TMP removal after erythrocyte invasion induces no phenotype within the intracellular development (ER, Figure 2B ), we focused on the subsequent ring-stage parasites upon TMP removal from the early schizont stage (ES, Figure 2B ). RON3-deficient parasites have a defect in protein export via PTEX (Low et al., 2019). To investigate the influence of the loss of the RON3 C-terminal fragment in ring-stage parasites on protein export, we examined distinct types of proteins exported by PTEX. Specifically, RESA is secreted from dense granules of merozoite as well as PTEX components at early ring stages (Riglar et al., 2011) and is exported by PTEX; whereas skeleton binding protein 1 (SBP1) is expressed in late ring stage parasites at 12 hours post-invasion (Blisnick et al., 2000) and is exported by PTEX as well as RESA. While RESA is normally distributed throughout the infected erythrocyte and is concentrated underneath the erythrocyte membrane (Export, Figure 3A ), SBP1 is localized to Maurer’s clefts in the cytoplasm of ring-infected erythrocytes (Export, Figure 3B ); however, the signal of each protein remains in close proximity to the parasite and is not translocated to the erythrocyte cytoplasm in 10% of infected cells (No export, Figures 3A–C ) The percentages of export and no export pattern of each protein revealed a 50% reduction of SBP1 export in TMP(-) parasites ( Figure 3C ). Furthermore, the PTEX translocon component EXP2 was detected at the PVM in both parasites ( Figures 3A, B ).

Low et al. (2019) reported that RON3-deficient parasites exhibited a defect in glucose uptake, and so we investigated glucose uptake at different time points in ring-stage parasites. The synchronized second cycle of ring-stage parasite cultures with or without TMP were incubated with a labeled glucose analog. Cells were further stained with Hoechst 33342 and MitoTracker Deep Red, which allowed live parasites to be identified. In both experimental conditions at 0 to 6 hours post-invasion, there was no significant change in mean fluorescence intensity (ΔMFI) in glucose uptake ( Figures 4A, C ). At 12 to 18 hours post-invasion, a 31% significant reduction in glucose uptake into ring-stage TMP(-) parasites was observed ( Figures 4B, C ); indicating that the RON3 C-terminal fragment is important for glucose uptake through the PVM.

Since the deletion of the RON3 C-terminal fragment in the ring stage leads to unusual parasite protein export through the PTEX translocon ( Figures 2G, H , 3 ), we next investigated if the reduction of trafficking of these parasite proteins in the infected erythrocyte leads to reduced nutrient uptake by the erythrocyte membrane PSAC. The kinetics of parasite nutrient uptake was measured by the osmotic lysis of infected erythrocytes in sorbitol, a sugar alcohol with primary uptake via PSAC (Pillai et al., 2010). Using synchronized parasites cultivated to the second cycle of the trophozoite stage with or without TMP, there was a significant 57% reduction in sorbitol uptake into the infected erythrocytes with RON3 knockdown parasites compared with control (p<0.01) ( Figures 5A, B ).