Ca2+ monitoring in Plasmodium falciparum using the yellow cameleon-Nano biosensor

Calcium (Ca2+)-mediated signaling is a conserved mechanism in eukaryotes, including the human malaria parasite, Plasmodium falciparum. Due to its small size (<10 μm) measurement of intracellular Ca2+ in Plasmodium is technically challenging, and thus Ca2+ regulation in this human pathogen is not well understood. Here we analyze Ca2+ homeostasis via a new approach using transgenic P. falciparum expressing the Ca2+ sensor yellow cameleon (YC)-Nano. We found that cytosolic Ca2+ concentration is maintained at low levels only during the intraerythrocytic trophozoite stage (30 nM), and is increased in the other blood stages (>300 nM). We determined that the mammalian SERCA inhibitor thapsigargin and antimalarial dihydroartemisinin did not perturb SERCA activity. The change of the cytosolic Ca2+ level in P. falciparum was additionally detectable by flow cytometry. Thus, we propose that the developed YC-Nano-based system is useful to study Ca2+ signaling in P. falciparum and is applicable for drug screening.


Results
Establishment and calibration of the Ca 2+ biosensor YC-Nano in the malaria parasite P. falciparum. To monitor the changes of cytosolic free Ca 2+ in P. falciparum, we generated transgenic lines expressing fluorescent protein-based Ca 2+ biosensors YC-Nano15 or YC-Nano50 driven by the P. falciparum heat shock protein 86 (PfHSP86) constitutive promoter (Fig. 1a). The difference in the sensitivity to Ca 2+ between these biosensors results from different lengths of the linker peptide between CaM domain and M13 peptide. The generated transgenic lines showed strong fluorescence signals in the parasite cytosol (Fig. 1b). In order to evaluate Ca 2+ sensing capacity, we generated calibration curves using live trophozoite stages of these transgenic parasites. To exclude the indirect influence of Ca 2+ in the iRBC cytosol and PV space, iRBC were treated with saponin to permeabilize the RBC and parasitophorous vacuole membrane (PVM) surrounding the parasite. The ratio of YFP and CFP was determined by confocal microscopy with Tyrode's buffer containing different concentrations of Ca 2+ (0-500 nM). The obtained calibration curves revealed a dissociation constant value of 15.5 nM and 45.8 nM for YC-Nano15 and YC-Nano50, respectively (Fig. 1c). These values are in agreement with the observed values by in vitro Ca 2+ titration for those biosensors expressed in Escherichia coli; 15.8 nM and 52.5 nM, respectively 20 . Thus, our results indicate that malaria parasites efficiently express functional YC-Nano biosensors in the parasite cytosol. Because reports showed that the concentration of free Ca 2+ in the parasite cytosol was roughly 40-100 nM by using synthetic chemical fluorescent Ca 2+ indicator Fura 2 21 , consistent with our preliminary data using generated transgenic parasites, we selected YC-Nano50 for further experiments to monitor Ca 2+ in P. falciparum cytosol, which has a more suitable dynamic range than YC-Nano15 for this purpose.
Resting cytosolic Ca 2+ concentration of intraerythrocytic P. falciparum. To describe the constitutive expression of the fluorescent proteins, we examined the YC-Nano50 fluorescence throughout the malaria parasite life cycle in RBC in addition to the trophozoite stage. We detected fluorescent signals from all blood stages (amoeboid ring, schizont and merozoite stages) and gametocytes, at higher FRET signals than in the trophozoite stage (Fig. 2a). To estimate the resting Ca 2+ concentration of the parasite cytosol from FRET signals, we used the in situ Ca 2+ calibration method with the Grynkiewicz equation 22 (Supplementary Fig. S1). Addition of 10 mM Ca 2+ containing the calcium ionophore A23187 to the Ca 2+ free parasite culture increased YFP/CFP ratio from 1.36 (Ca 2+ = 2.93 nM (median); minimum between 0-30 sec) to 2.66 (Ca 2+ = 895.2 nM; maximum between 120-180 sec) in trophozoite stage parasites, confirming that YC-Nano50 has a large dynamic range in P. falciparum. The calculated median cytosolic Ca 2+ concentration in the trophozoite stage was 30.0 (interquartile range: 5.6-55.0) nM. We found calculated Ca 2+  P. falciparum cytosolic Ca 2+ level is not modulated by thapsigargin, a mammalian SERCA inhibitor. The endoplasmic reticulum is an important Ca 2+ storage compartment to maintain and regulate the cytosolic Ca 2+ concentration in eukaryotic cells, and uptake of Ca 2+ from cytosol to ER is regulated by SERCA. In Plasmodium conflicting reports describe the responses of malaria parasites against SERCA inhibitors, specifically thapsigargin (TG) 21,23 . We therefore revisited the effect of TG for parasite cytosolic Ca 2+ homeostasis, and found that 15 μM CPA, a SERCA specific inhibitor consistently reported to inhibit P. falciparum SERCA (PfSERCA) 24 , increased the cytosolic Ca 2+ (Fig. 3a); whereas 7.6 μM TG, a concentration reported to inhibit PfSERCA pump activity 23 , did not change the cytosolic Ca 2+ concentration (Fig. 3b). The effect of TG on the cytosolic Ca 2+ level was not observed even when 76 μM TG was applied ( Supplementary Fig. 2). The positive control calcium ionophore A23187 increased the cytosolic Ca 2+ , and a solvent control DMSO showed no effect (Fig. 3c,d).
Because the parasite is surrounded by Ca 2+ rich environments in the human body and in the culture -for example, 45-86 nM in the RBC cytosol, ~40 μM in the PV space, and ~1 mM in the human plasma 25,26 -we further evaluated the effect of CPA and TG in Ca 2+ -free medium after selective membrane permeabilization. Firstly, iRBCs were treated with streptolysin O (SLO) to selectively permeabilize the RBC membrane, but not the PVM and parasite plasma membrane (PPM). When TG was added to SLO-treated iRBC, no effect was observed, but Purple to red color scale in the YFP/CFP panel represents low to high FRET efficiency (0 to 2.5). Scale bar, 2.5 μm. (c) The normalized fractional changes of the FRET signals (ΔR/R 0 ) are plotted against the different Ca 2+ concentration (0, 20, 40, 60, 80, and 100 nM). The curves represent the averaged data of ten parasites from 3 independent experiments.
Scientific RepoRts | 6:23454 | DOI: 10.1038/srep23454 subsequent addition of CPA increased the cytosolic Ca 2+ (Supplementary Fig. S3a), consistent with the previous result, thus indicating that the observed effect of CPA and TG was not due to Ca 2+ in the RBC cytosol or medium. Next, iRBCs were treated with saponin to permeabilize the RBC membrane and PVM, but not the PPM. Again CPA increased cytosolic Ca 2+ , but TG did not ( Supplementary Fig. S3b), indicating the effect of CPA and TG was not due to Ca 2+ in the PV space. These results confirmed that the parasite cytosolic Ca 2+ concentration changed in response to CPA in the presence or absence of RBC membrane and parasitophorous vacuole membrane, indicating that CPA targeted intracellular Ca 2+ storage.
To gain insights into the difference between human SERCA and PfSERCA in the response against TG, we constructed a model structure of human SERCA and PfSERCA based on the co-crystal structure of rabbit SERCA with TG using homology modeling. Docking simulation with 200 individual genetic algorithm from homology modeling resulted in an estimated binding free energy and inhibitory constant (K I ) for TG and PfSERCA of − 9.58 kcal/mol and 95.62 nM, respectively; whereas those for TG and human SERCA were − 10.82 kcal/mol and 11.77 nM, respectively. A more negative free energy value indicates a stronger molecular interaction, and thus the results suggest a 8.1-fold weaker interaction between PfSERCA and TG than that between human SERCA and TG. Next, to optimize the interaction between protein and TG in detail, we performed energy minimizations The total value between PfSERCA and TG was − 99.0, a value higher than that between human SERCA and TG (−100.8). Because a more negative value of the estimated van der Waals force indicates more stability, these results also suggest that the interaction between PfSERCA and TG is less stable than that between human SERCA and TG. These homology modeling and binding energy calculations between PfSERCA and TG support the hypothesis of P. falciparum insensitivity against TG.
Dihydroartemisinin does not alter the cytosolic Ca 2+ homeostasis of P. falciparum. Artemisinin (ART) derivatives are currently the most commonly used anti-malarial drugs, due to their low cost and because P. falciparum has not developed resistance against these drugs outside of Southeast Asia 1 . In spite of their importance for treatment of malaria, the mechanism of action of the active metabolite of the ART derivatives, dihydroartemisinin (dART), in the malaria parasite is still not clearly understood 27 . Two mechanisms of action have been proposed, the first that dART targets parasite hemoglobin metabolism 28 and the second that dART targets SERCA 29 . Because both dART and TG are sesquiterpene lactones and might act towards SERCA in a similar manner, we evaluated if dART could disturb cytosolic Ca 2+ homeostasis as implicated from the latter hypothesis. We exposed parasites to different concentration of dART (1, 10, and 100 μM) and found that dART had no effect on the parasite cytosolic Ca 2+ concentration even at 100 μM ( Fig. 4a-c). Sequential exposure of the dART-treated parasites to CPA increased cytosolic Ca 2+ , thus validating that the parasites were responsive to inhibitors. The IC 50 of dART used in the above experiment against P. falciparum was 1.5 nM for 24 hours 30 , confirming the pharmacological activity of dART. These results suggest that dART does not target PfSERCA.
Docking simulation of dART with PfSERCA or human SERCA were performed with 200 individual genetic algorithm from homology modeling resulted that the estimated free energy and K I for PfSERCA and dART were − 6.96 kcal/mol and 7.85 μM, respectively. For human SERCA and dART the values for estimated free energy and K I were − 7.47 kcal/mol and 3.37 μM, respectively; suggesting, firstly, a 2.3-fold weaker interaction between PfSERCA and dART than that between human SERCA and dART; and, secondly, a much weaker interaction of dART with both human and PfSERCA than the case for TG. These modeling calculations support the observed inactivity of dART against cytosolic Ca 2+ homeostasis in P. falciparum. Ca 2+ is stored in compartments other than the ER at the trophozoite stage of P. falciparum. In the present study we detected an increase of cytosolic Ca 2+ level in the Ca 2+ -containing medium with the calcium ionophore A23187, but not CPA, even after SERCA was inhibited with CPA as a SERCA specific inhibitor ( Fig. 5a, Supplementary Fig. S6a), thereby indicating that the second peak of Ca 2+ level was not derived from the ER. In contrast, addition of CPA to the A23187-pretreated iRBCs in the Ca 2+ -containing medium did not increase the cytosolic Ca 2+ (Supplementary Fig. S6b). Although the ER is the main site of intracellular Ca 2+ storage, other intracellular compartments are known to contribute to this process in the phylum Apicomplexa 4 , and therefore these results suggest that Ca 2+ flows into the parasite cytosol either from other Ca 2+ -storing compartments or from outside of the parasite. To investigate if we could detect the existence of non-ER Ca 2+ storage sites using the transgenic reporter parasites, we excluded the Ca 2+ source in the iRBC cytosol and medium. iRBC were selectively permeabilized with SLO or saponin and the effect on the cytosolic Ca 2+ level was evaluated by sequential addition of CPA and A23187 in the Ca 2+ free medium. When A23187 was added to SLO-treated and CPA-pretreated iRBC, cytosolic Ca 2+ was increased (Fig. 5b), suggesting that second elevation of Ca 2+ level was not due to the influx from medium or RBC cytosol. Next, to exclude the Ca 2+ source in the PV space, cytosolic Ca 2+ level was evaluated using saponin-treated iRBC in the Ca 2+ free medium. Addition of A23187 still increased the cytosolic Ca 2+ level in the CPA-pretreated parasites (Fig. 5c), indicating that the second elevation of Ca 2+ was not due to influx from the PV. Furthermore, we found that the second peak of cytosolic Ca 2+ in the parasites treated with saponin was significantly lower than the peak with SLO (13% reduction; n = 4; p < 0.0001, Mann-Whitney U-test), suggesting that this reduction reflect Ca 2+ influx from PV into parasite cytosol in SLO-treated iRBC. These results indicate that our system is able to detect the existence of parasite Ca 2+ in the PV space and intracellular Ca 2+ storage compartments in addition to the ER.
Flow cytometry-based system is applicable for drug screens targeting P. falciparum Ca 2+ homeostasis. In order to develop a high-throughput method to screen panels of compounds, we examined if flow cytometry could be used to detect FRET signals in the YC-Nano50-expressing P. falciparum (Fig. 6a). In this assay we included another SERCA inhibitor, 2,5-di-tert-butylhydroquinone (BHQ), in addition to CPA and TG, to compare the FRET signals obtained by confocal microscopy (Fig. 3a,b and Supplementary Fig. 6c) and those by flow cytometry. BHQ is known for its structural simplicity and low cost in comparison to other SERCA inhibitors 31 . The R post /R pre values by flow cytometry were 1.24 ± 0.02 (mean ± standard error of the mean (s.e.m.)), 1.13 ± 0.03, 0.99 ± 0.02, and 0.98 ± 0.02, for CPA, BHQ, TG, and DMSO, respectively (Fig. 6a); indicating that CPA and BHQ, but not TG, affect the cytosolic Ca 2+ homeostasis. In the above experiments the IC 50 values against  Table). Clear correlation (R 2 = 0.9956) existed between FRET signals obtained from confocal microscopy and flow cytometry (Fig. 6b).

Discussion
In this study we generated transgenic P. falciparum lines which stably express genetically encoded YC-Nano Ca 2+ biosensors in the cytosol, and a robust system to monitor Ca 2+ concentrations under physiological conditions. This technology enabled us to evaluate the cytosolic Ca 2+ concentration of the parasite cytosol at different developmental stages, and to monitor the change in cytosolic Ca 2+ levels caused by a panel of compounds proposed to act against the ER-residing Ca 2+ -ATPase, SERCA. As an initial attempt, to avoid cell damage and obtain reproducible FRET signals without photobleaching, we used a 1% (< 3 μW) power of 457 nm laser beams for excitation. Introduction of the Perfect Focus System and galvano scanner enabled stable capture images every 1 second at a 512 × 512 pixel resolution, which is critical to monitor changes in organisms of sizes less than 10 μm diameter, such as the malaria parasite. With these optimizations the FRET signals from this organism became stable for 10 minutes or more.
To our knowledge this is the first report to estimate cytosolic Ca 2+ concentrations throughout the blood stages of the malaria parasite. The cytosolic Ca 2+ concentration is high for all stages (values for amoeboid ring, schizont, merozoite, gametocyte stage III and gametocyte stage IV-V are 373, 310, 949, 131 and 522 nM, respectively), with the exception of the trophozoite (30 nM) (Fig. 7). The trophozoite is metabolically the most active stage, which may favor a lower cytosolic Ca 2+ in order to respond to subtle changes in Ca 2+ concentrations. The trophozoite stage parasite is able to quickly recover the cytosolic Ca 2+ level after the artificial increase of the Ca 2+ level with SERCA inhibitors (Fig. 5), indicating the existence of the SERCA-independent mechanisms to maintain the cytosolic Ca 2+ level less than 100 nM. PfATP4 (PF3D7_1211900), a non-SERCA-type Ca 2+ -transporting P-ATPase which is located on the parasite plasma membrane, may participate to this process 32 . Because the cytosolic Ca 2+ level significantly increased in the schizont stage, both mechanisms appear to be less active in this mature parasite form. The estimated cytosolic Ca 2+ concentration was highest at the merozoite stage, significantly higher than the schizont stage (n = 10; p < 0.0001, Mann-Whitney U-test). Ca 2+ signaling is known to be involved in the egress of the merozoites from the RBC, as well as the invasion into new RBC by triggering the secretion of microorganelles such as exonemes and micronemes in the merozoite stage parasite 7,9 . Thus we consider that the observed highest cytosolic Ca 2+ level at released merozoites indicates that the Ca 2+ secretion signals have been initiated. Glushakova et al. reported that cytosolic Ca 2+ level increased at the schizont stage, reaching to 1-10 μM range just prior egress using a chemical indicator Fura Red, which is consistent to our estimated cytosolic Ca 2+  In the trophozoite stage parasite, P. falciparum SERCA (PfATP6) is inhibited by cyclopiazonic acid (CPA) and 2,5-Di-t-butyl-1,4butylhydroquinone (BHQ), but not by thapsigargin (TG) and dihydroartemisinin (dART). CPA and BHQ may also affect PfATP4, a Na + -ATPase located on the parasite plasma membrane (PPM), and inhibit Ca 2+ influx from parasite cytosol to parasitophorous vacuole. Expected non-ER Ca 2+ -containing compartment(s) with potential integral membrane Ca 2+ transporters are indicated. Inositol trisphosphate receptor (IP 3 R) on the ER membrane has been proposed, but the encoding gene has not been identified 4 .
Scientific RepoRts | 6:23454 | DOI: 10.1038/srep23454 concentration of 949 nM at the merozoite stage 33 . P. falciparum calcium-dependent protein kinase 5 (PfCDPK5) and PfCDPK1 are plant-like protein kinase family members and have been proposed to act during these steps 34 .
Although it is not clear if these PfCDPKs are activated at Ca 2+ concentrations as high as the 310 nM estimated at the schizont stage, one report proposed that the K d value for Ca 2+ of a purified beetroot CDPK is 770 nM 35 . In this regard, Carey et al. successfully observed Ca 2+ oscillation during P. berghei sporozoite movement using another calcium biosensor, TN-XXL, which has a K d of 830 nM 36 . The higher Ca 2+ level in the released merozoites was noted by Biagini et al. using Fluo 4-AM, but the concentration was not determined due to the limited resolution of the system 37 . After completing RBC invasion, the parasite appears to gradually establish mechanisms to regulate cytosolic Ca 2+ level less than 100 nM during ring stage development. Because CDPK1 is also expressed in both male and female gametocytes 38 , high Ca 2+ concentration estimated at the gametocyte stage is consistent with possible CDPK1 activity in the high Ca 2+ level environment.
To validate the robustness of our established system to monitor the Ca 2+ concentration in P. falciparum, we conducted four experiments to evaluate: 1) the effect of TG on the P. falciparum cytosolic Ca 2+ homeostasis, 2) the effect of dART on the P. falciparum cytosolic Ca 2+ homeostasis, 3) the feasibility to detect Ca 2+ storage(s) beside ER, and 4) the feasibility to establish a high-throughput method to detect the change of the Ca 2+ level by flow-cytometry. Uptake of Ca 2+ from cytosol to ER is largely regulated by SERCA; however, there are conflicting observations regarding malaria parasite responses against the SERCA inhibitor TG. One report concluded that PfSERCA was TG-insensitive 21 , but another reported TG-sensitivity 23 . This controversy may be in part due to the employed method to monitor the cytosolic Ca 2+ with synthetic chemical indicators Fura 2-AM or Fluo 3-AM. Our analysis using a parasite line expressing a biosensor revealed that the Ca 2+ concentration at the trophozoite stage was 30 nM, which was much lower than the dissociation constant of Fura 2-or Fluo 3-based indicators (K d = 140 and 325 nM, respectively). Using the YC-Nano50 biosensor with a K d value of 45.8 nM, which is superior than the chemical indicators to evaluate Ca 2+ concentration at the trophozoite stage of P. falciparum, we were able to clarify that TG had no effect on the Ca 2+ homeostasis of P. falciparum. SERCA of apicomplexan parasites, including Plasmodium, is evolutionally more closely related to one of the two types of plant SERCA than to mammalian SERCA 39 . TG is a plant-derived compound and the plant Ca 2+ -ATPases have developed insensitivity to TG 40 , which is in agreement to our observation that PfSERCA is TG-insensitive. Docking models of TG with PfSERCA and mammalian SERCA also suggest a clear difference in the shape of the TG binding pocket between the two SERCA structures. Based on these differences in the sensitivity against TG and the structure of the TG binding pocket, TG may serve as a seed compound for a structure-based drug design to develop selective anti-malarial compounds.
Because both TG and ART are composed of sesquiterpene lactone, and since TG is a highly selective inhibitor for mammalian SERCA 41 , it was therefore reasoned that both TG and ART would behave in a similar manner towards SERCA. Consistent to this expectation, studies analyzed Ca 2+ -ATPase activity using PfSERCA expressed on Xenopus oocyte membrane proposed that ART had effect on PfSERCA 42,43 . However, other experiments did not support that PfSERCA was a target of ART 44,45 . In this study, we clearly showed that dART had no effect on Ca 2+ concentration in the parasite cytosol. Docking models of dART and both SERCA showed that the affinity of dART to both SERCA is at the micromolar level, suggesting that dART may not be effective against both SERCA. Together, our data indicate that dART plays at most a minor role to modulate P. falciparum Ca 2+ homeostasis.
The ER is the most important organelle storing Ca 2+ in the malaria parasite 12,21 ; but other compartments, such as DV 37 , mitochondrion 46 , acidocalcisome 10 , and PV space 47 have also been proposed to act as Ca 2+ storage sites (Fig. 7). In this study we indicate the existence of Ca 2+ in the P. falciparum PV space by comparing SLO-treated iRBC and saponin-treated iRBC. In P. falciparum, two Ca 2+ ATPases, PfSERCA (PfATP6) and PfATP4, have been annotated among the 13 P-type ATPases. PfATP4 is localized on the PPM and is considered to transport not only Na + but also Ca 2+ 32,48 . This ATPase is potentially responsible for the difference in the observed higher level of the cytosolic Ca 2+ increase after calcium ionophore stimulation in SLO-treated iRBC than that in saponin-treated iRBC. The DV was reported to contain only moderate amounts of Ca 2+ and no dynamic changes of the Ca 2+ concentration were observed in DV following induced cytosolic Ca 2+ bursts 37 . Although there are some reports of the mitochondria and acidocalcisome as Ca 2+ storages, active participation of these compartments to maintain cytosolic Ca 2+ homeostasis of malaria parasites is still unclear 49 .
In conclusion, we generated a transgenic P. falciparum expressing YC-Nano50 biosensor and showed that this parasite is a suitable and powerful tool in which to study Ca 2+ homeostasis in the trophozoite stage of P. falciparum. We determined, for the first time, that the resting Ca 2+ concentrations at schizont, merozoite, ring, and late gametocyte stages are higher than 300 nM. We also showed that TG and dART did not affect the cytosolic Ca 2+ level of trophozoite stage of this parasite. FRET signals are detectable by flow cytometry and correlate with a microscope-based assay, indicating that the developed flow cytometry-based system is applicable for drug screens targeting mechanisms which maintain P. falciparum Ca 2+ homeostasis.
Generation of expression plasmids for Ca 2+ biosensor. Plasmids for P. falciparum transfection were constructed based on the Invitrogen Multisite Gateway ® system (Invitrogen, Carlsbad, CA). DNA fragments encoding YC-Nano15 and -50 were amplified from corresponding plasmid templates by PCR amplification and recombined with pDONR ™ P2R-P3 to generate pENT23-YC-Nano15 and -50, respectively. Expression vectors, pLN-YC-Nano15 and 50 (Fig. 1a), were generated by LR reaction from pENT23 plasmids described above, Scientific RepoRts | 6:23454 | DOI: 10.1038/srep23454 pENT41 plasmid containing P. falciparum HSP86 promoter region, pENT12-linker, and pLN-DEST-R43(II) containing a blasticidin-s deaminase (BSD) selectable marker 50 . Nucleotide sequence data reported are available in the DDBJ Sequenced Read Archive under the accession numbers LC028929 and LC075581. All experiments conducted in this study were approved by the committee for recombinant DNA experiment, Nagasaki University. The methods were carried out in accordance with the approved guidelines.
Parasite lines, culture, and transfection. The P. falciparum Dd2 parasite line was originally obtained from National Institute of Health, USA. The parasites were maintained with O + human RBC at 2% hematocrit in fibrinogen-free human plasma-containing complete RPMI medium and transfection was performed as described 51 . At days 4-5 post transfection, drug selection with 2.5 μg/mL BSD (InvivoGen, San Diego, CA) was started and culture was maintained until drug-resistant parasites appeared. The usage of human RBC and plasma was approved by the ethical committee, Institute of Tropical Medicine, Nagasaki University.
Cytosolic Ca 2+ measurements. YC-Nano-expressing P. falciparum parasites (3-6% parasitemia) were used for live cell imaging experiments. Ca 2+ measurements were performed using trophozoite parasites which were obtained by 5% sorbitol synchronization before 18-24 hr experimentation 48 . On the day of imaging, parasite cultures were collected and washed twice with 1 ml of 37 °C warmed plasma-free incomplete RPMI medium (ICM). Then 1 ml of 0.25% hematocrit parasite infected-RBC (iRBC) was plated on the glass bottom 35-mm cellview TM TC treated hydrophilic coated dish (Grenier bio-one, Germany). After keeping the iRBC in the dish for 30 min, ICM were replaced with phenol red-and plasma-free RPMI medium containing 0.5% AlbuMAX ® I. Time-lapse imaging was performed at 37 °C using an A1R confocal microscope system configured with an inverted microscope (Ti-E; Nikon, Japan) with 60× or 100× oil objective lens (PlanApo, NA 1.4, Nikon). The inverted microscope configuration acts as a stable system with the Perfect Focus System (PFS, Nikon). The water chamber stage and the objective lens were kept at 37 °C with a temperature controller (Tokai-Hit, Japan). The fluorescence resonance energy transfer (FRET) image analysis between cyan fluorescent protein (CFP) and yellow fluorescent protein (YFP) was performed by confocal microscopy. YC-Nano was excited at 457 nm for both CFP and YFP, and emissions were detected for CFP (482/35 nm) and YFP (525/50 nm). Time-lapse images were captured every 1 sec at a 512 × 512 pixel resolution by confocal microscopy. For the first 30 sec, time-lapse images were taken without chemical compounds. Chemical compounds (TG, CPA, dART, A23187, BHQ, and DMSO control) were added directly to the edge of the chamber containing transgenic parasites. The parasite cytosolic region was used for the analysis as a region of interest (ROI) and background fluorescence was subtracted. The imaging analysis was carried out using NIS-Element Advanced Research imaging software (Nikon). The R/R 0 value was calculated for each parasite, where R is the YFP/CFP ratio and R 0 is the mean YFP/CFP ratio before adding the drug as baseline (time between 0-30 s).

Permeabilization of parasite-iRBC with streptolysin O or saponin.
To selectively permeabilize the RBC membrane only, iRBC were treated with 20 U/ml streptolysin O (SLO; Sigma Aldrich Chemical Co, St. Louis, MO) in PBS for 6 min at room temperature, washed three times with ICM, and kept in ICM 52 . To permeabilize RBC and PVM, iRBCs were treated with 0.01% saponin (Wako Pure Chemical Industries, Ltd, Japan) in PBS for 10 min at room temperature, washed three times with PBS, and kept in ICM 53 . SLO-or saponin-treated iRBCs were transferred to hydrophilic-coated dish and kept for 30 min to let the iRBCs adhere to the glass dish bottom. The iRBCs were washed three times for 10 min each with Ca 2+ free Tyrode's buffer (140 mM NaCl, 10 mM glucose, 10 mM HEPES, 4 mM KCl, 1 mM MgCl 2 , pH 7.4) to remove Ca 2+ from the extracellular medium. Finally, Ca 2+ free Tyrode's buffer was used for time lapse imaging. YC-Nano calibration curve for P. falciparum. To generate the calibration curve, iRBCs were permeabilized with saponin and prepared for analysis as described above. Tyrode's buffer containing different concentration of Ca 2+ (0, 10, 20, 40, 60, 80, 100, and 500 nM) were prepared with calcium chloride. Parasites were re-suspended in the different concentrations of Ca 2+ -containing buffer, kept for 10 min, and observed under the confocal microscope. Images were obtained from 10 independent parasites for each Ca 2+ concentration and the fractional change of the YFP to CFP ratio (ΔR/R 0 where ΔR = R − R 0 ) was calculated. The ΔR/R 0 values were normalized by dividing by the ΔR/R 0 value with 500 nM Ca 2+ buffer and plotted using GraphPad Prism6 software (GraphPad Software, Inc., La Jolla, CA).

Calculation of the resting cytosolic Ca 2+ concentration.
To estimate the resting cytosolic Ca 2+ concentration in the different stages of P. falciparum, we employed an in situ Ca 2+ calibration method with time course change of Ca 2+ concentration in the individual intact iRBC, which is more precise than single images of the parasite 54 . The area of the parasite desired for quantification of the fluorescence signal, and the control area were selected as ROIs. YFP/CFP was determined in a given time course. The YFP/CFP value was converted to calcium concentration value using the following equation 22  schizont was calculated as 345.6 nM using above equation from YFP/CFP value of 2.34. Obtained Ca 2+ concentration using Grynkiewicz equation with above condition was limited to 1000 nM to avoid unacceptably large fluctuations by NIS-Element Advanced Research imaging software.
Flow cytometry-based Ca 2+ measurement. The FRET signal of parasites was measured by flow cytometry (Gallios TM , Beckman Coulter, Inc., Brea, CA). Before assay parasite-iRBCs were washed twice in phenol redand plasma-free RPMI medium containing 0.5% AlbuMAX ® I. To measure CFP and FRET signals, iRBC were excited with a 405 nm laser and fluorescence was collected in the CFP channel with a standard 450/50 filter, while the FRET signal was measured with a 525/20 filter. To measure YFP signal, parasite-iRBCs were excited with a 488 nm laser and emission was taken with 525/20 filter. For each sample a minimum of five thousand YFP positive iRBCs were evaluated and non-infected RBCs were used as a baseline for signal detection. To measure the changes of cytosolic Ca 2+ , baseline Ca 2+ levels were first measured for 60 sec for each sample, followed by addition of different Ca 2+ inhibitors (CPA, TG and BHQ) for 3 min. FlowJo software (FlowJo LLC, OR, USA) was used to analyze the obtained data. The mean fluorescence intensities of the iRBC before and after adding inhibitors were obtained and the R post ratio of iRBC after adding inhibitors were normalized by R pre of iRBC before adding inhibitors to obtain the FRET signal changes of each inhibitor.
Homology modeling, docking simulation, and fragment molecular orbital calculation. The coordinates of the crystal structure of complex between rabbit SERCA and TG was downloaded from the Protein Data Bank (http://www.rcsb.org; 2AGV). A model structure of human SERCA and PfSERCA were generated by a homology modeling based on the rabbit SERCA structure using Modeller9.14 (https://salilab.org/modeller/) 55 and PyMOL (http://www.pymol.org). Binding free energies of TG (PubChem CID: 446378) and dART (PubChem CID: 71939-50-9) with human/PfSERCA were estimated by docking simulations using AutoDock4.2 56 . In these simulations, 200 individual genetic algorithm calculations were run in each of which 25 × 10 6 energy evaluations were performed. Other model structures of TG and selected 144 amino acid residues located near the binding region were constructed by 2000 steps energy minimizations using AMBER99SB force field 57 . Using these structures, we performed fragment molecular orbital (FMO) calculations 58 at second order Møller-Plesset perturbation theory with resolution of the identity approximation for analysis of van der Waals interactions. In these FMO calculations, cc-pVDZ basis set 59 was employed and PAICS program 60 was used.
Statistical analyses. All statistical analysis was performed by Graphpad Prism 6 software (GraphPad Software, Inc. CA. USA).