Chloroquine exposure triggers distinct cellular responses in sensitive versus resistant Plasmodium falciparum parasites

Chloroquine (CQ) treatment failure in Plasmodium falciparum parasites has been documented for decades, but the pharmacological explanation of this phenotype is not fully understood. Current concepts attribute CQ resistance to reduced accumulation of the drug at a given external CQ concentration ([CQ]ex) in resistant compared to sensitive parasites. The implication of this explanation is that the mechanisms of CQ-induced toxicity in resistant and sensitive strains are similar once lethal internal concentrations have been reached. To test this hypothesis, we investigated the mechanism of CQ-induced toxicity in CQ-sensitive (CQS) versus CQ-resistant (CQR) parasites by analyzing the time-course of cellular responses in these strains after exposure to varying [CQ]ex as determined in 72 h toxicity assays. Parasite killing was delayed in CQR parasites for up to 10 h compared to CQS parasites when exposed to equipotent [CQ]ex. In striking contrast, brief exposure (1 h) to lethal [CQ]ex in CQS but not CQR parasites caused the appearance of hitherto undescribed hemozoin (Hz)-containing compartments in the parasite cytosol. Hz-containing compartments were very rarely observed in CQR parasites even after CQ exposures sufficient to cause irreversible cell death. These findings challenge current concepts that CQ killing of malaria parasites is solely concentration-dependent, and instead suggest that CQS and CQR strains fundamentally differ in the consequences of CQ exposure.

available to bind to FPIX and reducing cell damage. A corollary to this hypothesis is that CQ-resistant (CQR) strains of P. falciparum should be killed once CQ concentrations in the DV ([CQ] DV ) that kill CQ-sensitive (CQS) are reached, an equilibrium that requires higher external [CQ]. However, CQR parasites are reported to tolerate higher CQ levels in the DV than CQS strains, an observation that cannot be explained by simple concentration-dependent toxicity effects of CQ. If CQS and CQR parasites are exposed to concentrations of extracellular chloroquine ([CQ] ex ) that generate equal internal CQ-concentrations in the DV ([CQ] DV ), equipotent CQ-dependent killing should be observed. However, CQR parasites have higher survival rates than CQS parasites at the same [CQ] DV 17 . It appears that the PfCRT K76T mutation alone cannot fully explain resistance to CQ from stoichiometric considerations only.
In this study, we report differences in killing kinetics and cell biological consequences in CQS and CQR parasites after exposure to equipotent [CQ] ex , determined based on IC 50 values in assays quantifying growth inhibition. Our findings show that CQS and CQR parasites display distinct cellular responses as a consequence of CQ exposure. These results suggest that our understanding of the mechanism of action of CQ requires refinement, and that cell biological effects of CQ on sensitive and resistant parasites need to be investigated in parallel.

Results
In situ sensitivity of CQ. Since measurements of drug responses can vary among laboratories, partly due to different culture conditions and procedures, we first determined CQ IC 50 values for the strains used in this study in 72 h SYBR Green I viability assays (Table 1), as described previously 18,19 . IC 50 values and their respective SEMs for the CQS parasite strains 3D7 and HB3 were 24 ± 6 nM and 14 ± 1 nM, respectively, and 169 ± 4 nM and 166 ± 9 nM, respectively, for the CQR parasite strains Dd2 and FCB. Resistance to CQ, as indicated by IC 50 values, was partially reversed in CQR strains by co-incubation with 1 µM verapamil (VP), while no significant difference in IC 50 values was observed with VP in CQS parasite strains (p > 0.05), as expected. All IC 50 values are in good agreement with those obtained by other groups [19][20][21][22][23] . Subsequent experiments with CQ were based on these IC 50 values.
Onset of growth inhibition is delayed in CQR parasites. The time course of CQ-induced growth inhibition of CQS and CQR parasites was analyzed using equipotent [CQ] ex as determined above. CQS 3D7 and CQR Dd2 parasite-infected RBCs (iRBCs) were exposed to [CQ] ex equal to 10x, 20x and 30x their respective IC 50 values (to provide equivalently lethal exposures for CQS and CQR parasites) or left untreated (Fig. 1). To confirm that the wash step was sufficient to remove CQ from the culture medium, parasites from the 0 h treatment groups were incubated with the appropriate [CQ] ex and washed immediately. No changes in parasite survival relative to the untreated control group were observed (Fig. 1 The digestive vacuolar membranes of CQS and CQR parasites remain intact. Disruption of the DV membrane and leakage of DV contents into the parasite cytosol have been hypothesized to be the main contributors to parasite death after CQ exposure 24 . We evaluated leakage of DV contents into the parasite cytosol after exposure to regimens of [CQ] ex that elicited irreversibly lethal damage. Prior to CQ exposure, 3D7 and Dd2 parasites were loaded with Fluo-4 AM, a fluorophore that accumulates in the DV and remains trapped within this compartment [25][26][27] . Fluo-4 AM loaded 3D7 and Dd2 parasites were exposed for up to 4 h to 10 μM CQ, which corresponds to 400x and 60x their IC 50 , respectively, as derived from the IC 50 concentrations (Fig. 1). Changes in Fluo-4 fluorescence were monitored in cytosol and DV of control and CQ-exposed CQS and CQR parasites using confocal microscopy. Strong DV and minimal cytosolic Fluo-4 fluorescence was observed in all parasites and remained stable throughout the 4 h time period. Redistribution of Fluo-4 fluorescence from the DV to the cytosol was not detected, suggesting that the DV membrane remained intact in control and CQ-exposed parasites (Fig. 2). The slight increase in DV Fluo-4 fluorescence seen over time in control and CQ-exposed Dd2 parasites (Fig. 2B) is likely due to the known transport of Fluo-4 into the DV of this parasite 27 . Effect of CQ on parasite morphology. Hz crystals are formed and stored in the DV. Hz crystals are easily recognized as dark structures under light microscopy. Live cell imaging reveals that Hz crystals move freely within the DV in an apparently random manner. We observed that Hz crystals appear smaller and move more rapidly in CQR strains (Dd2, FCB, FCR3, K1) compared to CQS strains (3D7, HB3, NF45, D10) (data not shown). Accordingly, exposure of CQR and CQS strains to CQ in culture may affect the movement and localization of Hz within these parasites differently.
To test this hypothesis, synchronized mid-trophozoite stage parasites (CQS strains: 3D7 and HB3; CQR strains: Dd2 and FCB) exposed to equitoxic [CQ] ex were monitored for 4 h using differential interference contrast (DIC) microscopy. The characteristic rapid movement of Hz crystals in CQR strains was not altered in Dd2 (Supplementary Videos S1 and S2) and FCB parasites, in which Hz movement persisted even after 24 h CQ exposure. In contrast, Hz movement was already markedly reduced in the CQS strains 3D7 (Supplementary Videos S3 and S4) and HB3 after 4 h exposure to CQ.
In addition to reduced Hz movement, we also observed the appearance of dark, rapidly moving structures in the cytosol of CQS parasites within 1 h of CQ exposure (Fig. 3A). These compartments appeared to be moving freely and randomly (Supplementary Video S4). The number of CQS parasites containing these dark cytosolic structures (DCS) steadily increased over time. In contrast, very few CQR parasites were observed to have DCS (<10% after 4 h CQ exposure), even at 30x the IC 50 value of CQ ( Fig. 3A; Supplementary Figure S1; Supplementary Video S2). DCS were readily visible when CQ-exposed CQS parasites were methanol-fixed but not Giemsa-stained (Supplementary Figure S2).
Verapamil (VP) has been shown to reverse CQ drug resistance in CQR parasites. To verify if VP affects the appearance of DCS in CQ treated parasites, 10 µM VP was added to CQS and CQR parasites treated with CQ. The addition of VP suppressed the appearance of DCS in CQS parasites and had no effect on CQR parasites (Fig. 3A).
To determine the rate of appearance of DCS in CQS and CQR parasites, 3D7 and Dd2 strains were exposed to 20x their IC 50 [CQ] ex values for up to 4 h. The number of DCS steadily increased in 3D7 parasites within the first 1-4 h (22%, 43%, 56%, and 61%, respectively) until a plateau was reached (Fig. 3B). DCS still appeared in 3D7 parasites at [CQ] ex = 2x the IC 50 value, but the number of parasites containing DCS was reduced (33% versus 61% after 4 h). In contrast, very few CQ-exposed Dd2 parasites had detectable DCS, even when exposed to 30x the IC 50 [CQ] ex value (2%, 6%, 7% and 9% after 1-4 h, respectively; Fig. 3B; Supplementary Figure S1). These values did not change substantially when Dd2 parasites were exposed to 2x their IC 50 [CQ] ex value.
The appearance of DCS was linked with growth inhibition in CQS 3D7 parasites compared to CQR Dd2 parasites (Fig. 3C). These findings support the hypothesis that exposure to equipotent [CQ] causes different cellular responses in CQS and CQR parasites.
Ultrastructural analysis of CQ-exposed parasites. We employed TEM to better characterize the DCS.
3D7 and Dd2 parasites were exposed to CQ for 3 h at 20x their IC 50 value, or left unexposed. To assure the structural preservation of P. falciparum-infected erythrocytes, parasites were enriched using a Percoll gradient. Non-enriched CQ-exposed and non-exposed parasites were also examined to confirm that Percoll enrichment did not affect parasite ultrastructure. No differences in the morphology of CQ-treated 3D7 parasites due to sample preparation were observed (Supplementary Figure S3).
TEM images confirmed the presence of multiple DCS in 3D7 parasites (Fig. 4A). Since no three-dimensional images were taken and the individual TEM sections did not represent the entire parasite cytosol, the percentage of 3D7 parasites with DCS could not be determined in this experimental approach. DCS were surrounded by a membrane bilayer, were between 150-200 nm in diameter, and typically contained 1-5 Hz crystals. No Hz-containing compartments were observed in TEM images of CQ-treated Dd2 parasites (n = 17) or in 3D7 and Dd2 parasites that were not exposed to CQ (3D7, n = 17; Dd2, n = 29) (Supplementary Figure S4).
TEM images also revealed that CQ-exposed 3D7 parasites exhibited additional, previously undescribed, morphological changes after 3 h of 20x CQ IC 50 exposure. In particular, abundant structures that appear to be membrane fragments are evident in the DV. Their lipid bilayer nature is readily visible (Fig. 4B). Furthermore, large vacuoles were present in the cytosol of CQ-exposed 3D7 parasites. These vacuoles were clearly distinct from, and were typically much larger than, the DV (Fig. 4C). These morphological changes were not seen in Dd2 parasites at that time point. No TEM slides were prepared for other time points. Hz-containing compartments likely originate from the DV. To address the origin of the DCS, we pre-loaded 3D7 parasites with Fluo-4 AM before the addition of CQ. We hypothesized that DCS derived from the DV would contain the fluorochrome. Indeed, fluorescently stained compartments containing Hz were seen in the cytosol of 3D7 parasites after CQ exposure (Fig. 5B, Supplementary Video S5), but not in unexposed 3D7 parasites or in Dd2 parasites (exposed or unexposed) (Fig. 5A). These data strongly support the hypothesis that DCS are derived from the DV after exposure of CQS parasites to the drug.

Discussion
Many studies have attempted to resolve the antimalarial mechanism of action of CQ. How exposure to the drug leads to the death of the parasite and how mutations in PfCRT protect against this effect remain unclear. In this study, we show (i) the appearance of dark cytosolic structures (DCS) containing Hz in the cytosol of CQS but not CQR parasites treated with CQ, (ii) the elimination of DCS in CQS parasites treated with CQ in the presence of VP, and (iii) the delayed onset of irreversibly lethal damage in CQR versus CQS parasites exposed to equitoxic [CQ] ex . Together, these and other morphological changes indicate that there is a fundamental difference in the consequences of drug exposure in CQS compared to CQR parasites. Four hypotheses have been formulated to explain CQ resistance in malaria parasites. The first postulates that the cytosol:DV pH gradient is smaller in CQR than CQS parasites, resulting in reduced CQ influx into the DV and a lower equilibrium [CQ] DV at a given [CQ] ex 28,29 . The second hypothesis proposes that the FPIX complex in CQR parasites has reduced affinity for CQ compared to this complex in CQS parasites; therefore, a lower concentration of toxic CQ-FPIX complexes accumulate at a given [CQ] DV . The third hypothesis suggests that PfCRT carrying the K76T mutation causes CQ to be transported out of the DV and reduces the amount of CQ in the DV, again lowering the equilibrium [CQ] DV at a given [CQ] ex . The fourth hypothesis proposes that CQ may be directly responsible for inhibiting PfCRT function. CQ is thought to bind PfCRT and mutations in PfCRT responsible for CQ resistance (defined as increased CQ IC 50 values in 48 or 72 h assays) might impede CQ binding by converting binding into efflux events. While the first hypothesis has not been supported by experimental data, multiple studies support the latter three 14,15,[30][31][32][33][34][35][36] .
There is evidence that CQ resistance cannot be attributed to the K76T amino acid substitution alone. CQR strains incubated with increasing [CQ] ex must accumulate higher internal [CQ] compared to CQS strains to obtain the same efficacy in parasite killing 3,9,16,17,31 . Furthermore, inconsistencies have been found between the presence of the K76T mutation and resistance to CQ in vitro and in vivo [37][38][39][40][41] . In addition, mutant PfCRT contributes to the shift to the right in the concentration-response curve between CQS and CQR strains, but only accounts for 10-20% of the LD 50 in CQR parasites 42 . Even at the same [CQ] DV , CQR strains are relatively protected from the drug and a second mechanism(s) that prevents CQ from causing lethal effects must be in place 9 . To demonstrate if specific PfCRT polymorphisms are required for the formation of DCS, experiments are now in progress using . Electron microscopy images of CQ-exposed 3D7 parasites. 3D7 parasites were exposed to 20x IC 50  isogenic parasites that are genetically modified to express wildtype (CQS) or mutant (CQR) forms of PfCRT and exposing these to CQ. A correlation could suggest that CQ drug resistance is not solely concentration-dependent but that PfCRT mutations may also contribute to distinct cellular responses as a consequence of CQ exposure.
Although the discovery of the PfCRT K76T mutation 14,34 was a milestone in understanding CQ resistance, further studies have revealed that mutant pfcrt cannot confer CQR in all genetic backgrounds 22,43 . A link between PfCRT and/or PfMDR1 mutations to antimalarial drug susceptibility has been suggested [44][45][46][47] , and additional loci other than pfcrt and pfmdr1 may also contribute to CQ resistance 37 .
Other researchers have proposed that a cytosolic target for CQ accounts for the antimalarial activity 48 . Our results suggest that lethal [CQ] ex affect CQS and CQR parasites differently, supporting the idea that the mechanism of CQ resistance includes factors other than selection of alleles of PfCRT. Reduced affinity of CQ for FPIX, the second hypothesis, is a plausible candidate. Although a shift in affinity curves, as reflected by higher IC 50 values required for CQR, does not necessarily affect lethality at saturating [CQ] ex in respective CQS or CQR strains, it may still alter the kinetics of parasite elimination, as indeed shown for the protracted lethality in CQR over CQS.
Increased survival in the presence of equitoxic [CQ] ex of CQR compared to CQS parasites was revealed by incubating trophozoites in CQ for varying lengths of time. Even when exposed to [CQ] ex corresponding to equivalent IC 50 values, CQR parasites were viable longer than CQS parasites. It has been suggested that equal internal [CQ] alone are not sufficient to eliminate CQR parasites compared to CQS parasites 17 . Our results support the possibility that a regulatory mechanism found in CQR parasites reduces the time course of toxicity of CQ in CQR, albeit having the same steady-state endpoint effect, i.e. complete elimination (Fig. 1). Our data suggest that CQR parasites not only reduce CQ accumulation in the DV, but may also delay the initiation and progression of cell death. Additionally to the aforementioned effect, our data also show that the cellular damage elicited by such exposures also differs between CQR and CQS strains.
In human cells, CQ is known to induce programmed cell death (PCD) through both apoptosis and autophagy pathways 49 . Recent investigations of cell death pathways in protozoan parasites have found evidence of PCD 50 . The possibility of CQ-induced apoptotic cell death in malaria parasites is unlikely since Plasmodium spp. lack caspases, a central component of apoptosis. Other classical apoptotic features, such as chromatin condensation, DNA fragmentation, disruption of the plasma membrane and nucleus dissolution, were also not observed in CQ-treated parasites [51][52][53] . In contrast, CQ-induced autophagic-like PCD has been suggested in malaria parasites 51,54 . Although autophagy is a strategic mechanism for cell survival, it can also lead to the death of an organism. Autophagic vacole formation after CQ exposure in malaria parasites was noted in the 1970s 55 . Cytoplasmic vacuolization similar to autophagy was confirmed by EM 51 . These findings are supported by our EM experiments, which revealed morphological effects in CQ-treated CQS parasites that included the accumulation of membranes in the DV and the formation of large vacuoles in the cytoplasm. The accumulation of undigested and unprocessed autophagic vesicles may be caused by defective lysosomal proteolysis or hyperactive autophagy induction 56 . Both possibilities seem plausible in CQ-exposed parasites.
Although some researchers have addressed the biology of cell death after CQ exposure in P. falciparum 24,51,57,58 , a major impediment to such studies is the lack of molecular tools available to study cell death in malaria parasites, e.g., antibodies against regulatory proteins to such as LC3 or caspase-3 that are typically involved in the initiation of PCD. LD 50 -directed quantitative trait loci analysis in CQS versus CQR strains revealed differences in loci that encode proteins linked to autophagy pathways, such as the autophagy-related genes (ATG) 42 , but no genes that are involved in apoptosis 42,54,58,59 . Further work in this area may lead to a better understanding of the antimalarial action of CQ and may reveal additional resistance mechanisms in CQR parasites.
While some PCD-related features have been described in CQ-exposed parasites through analysis of EM images 24,51,60 , no reports describe the appearance of DCS within the cytosol of CQS -and not CQR -parasites after exposure to the drug. The appearance of compartmentalized Hz crystals in our EM images correlated with the DCS observed in DIC images. The CQ-induced redistribution of Hz, or FPIX, from the DV to the parasite cytosol has been previously shown using TEM 61 . However, the effect of CQ on elevated levels of FPIX in the cytosol was only investigated in the CQS 3D7 strain and no comparisons were made between CQS and CQR strains. These findings supported previous work by Ginsburg and coworkers, who hypothesize that the principal toxic effect of CQ is manifest outside the DV and is linked to the binding of CQ to cytosolic FPIX 62 . Free FPIX is thought to easily escape the DV and is rapidly degraded by GSH. As much as 70% of the FPIX generated through hemoglobin digestion was found to be detoxified by cytosolic glutathione 48 . Although the presence of iron in the cytosol has previously been shown, the compartmentalization of Hz in DCS in CQS but not CQR parasites has not been previously described.
Thus, exposure of CQS parasites to lethal [CQ] ex led to the appearance of multiple morphological changes prior to the leakage of DV contents into the cytosol. These changes were not observed in CQR parasites exposed to equally lethal [CQ] ex , suggesting that the consequences of equipotent CQ exposure differ between them within the first 3 h of exposure.
The increase in Hz-containing DCS in the parasite cytosol is not due to rupture of the DV membrane. Experiments with Fluo-4 AM, which is actively transported into the DV by PfMDR1 26 , showed that the DV membrane remained intact when the parasite was exposed to [CQ] ex that cause irreversibly lethal damage, as there was no major leakage of the fluorochrome into the cytosol at a time when the parasites were destined to die. However, fluorescent staining of Hz-containing DCS in the cytosol is most plausible if the compartments formed at the DV and contained DV contents, including Fluo-4. The localization of Fluo-4 fluorescence in DCS provides strong evidence that these structures arose from the DV.
Clearly, further studies are needed to elucidate the implications of the presence of DCS in CQS parasites after CQ exposure and their absence from CQR parasite strains. Understanding PCD in P. falciparum and obtaining insight into how CQR parasites maintain homeostasis even under drug pressure may reveal a unique mechanism in CQR parasites that could be exploited to potentially reverse CQ resistance, opening new possibilities to restore the clinical utility of this drug.

Material and Methods
Parasite strains and culture conditions. CQS (3D7, HB3) and CQR (FCB, Dd2) parasite strains were continuously cultured as described by Trager and Jensen 63 , with modifications. Briefly, parasites at 5% hematocrit were propagated in RPMI 1640 (Life Technologies, Burlington, ON, Canada) supplemented with 25 mM HEPES, 2 mM L-glutamine, gentamicin (20 µg/ml) (Life Technologies, Burlington, ON, Canada), 100 µM hypoxanthine (Sigma-Aldrich, Oakville, ON, Canada), and 0.5% AlbuMAX I (Life Technologies, Burlington, ON, Canada). Parasites were maintained at 37 °C under an atmosphere of 5% CO 2 , 3% O 2 and 92% N 2 . A + red blood cells were obtained from the Interstate Blood Bank (Memphis, TN, USA). Giemsa (Sigma-Aldrich, Oakville, ON, Canada)-stained blood smears were prepared daily to monitor parasite growth. For synchronization, parasites were exposed to 5% D-sorbitol (BioShop Canada, Burlington, ON, Canada) for 10 min at 37 °C; sorbitol was removed and parasites were washed once before placed back into culture. To obtain highly synchronous parasite cultures, this treatment was repeated after 6 h.
Growth inhibition assays. CQ was obtained from Sigma-Aldrich Canada (Oakville, ON, Canada) and dissolved in dH 2 O. Growth inhibition assays were performed as described 19 , with modifications. Highly synchronized ring stage parasites were diluted to a final parasitemia of 0.5% and a hematocrit of 2%. A total of 100 µl culture medium per well was prepared in a 96-well plate assay, with a drug dilution series of 1:3, ranging from 1 µM to 0.15 nM. Plates were incubated at 37 °C, 5% CO 2 , 3% O 2 and 92% N 2 for 72 h, then frozen and stored at −80 °C. Plates were thawed at room temperature and 100 µl 2x lysis buffer (20 mM Tris pH 7.5, 5 mM EDTA, 0.008% saponin, 0.08% Triton X-100, 0.2 µl SYBR Green I/ml) (SYBR Green I: Sigma Aldrich, Oakville, ON, Canada; all other chemicals: BioShop Canada, Burlington, ON, Canada) was added to each well. Plates were incubated in the dark for at least 1 h. Fluorescence intensity was determined using a Synergy H4 plate reader (Fisher Scientific, Nepean, ON, Canada) with 485 nm excitation and 520 nm emission wavelengths. IC 50 values were calculated by fitting concentration response curves with a custom procedure for IGOR Pro 6.2 based on an R script kindly provided by Le Nagard and used as described 18 Chloroquine exposure assay. Wells containing highly synchronized early trophozoite stage parasites were prepared in triplicate at 1% parasitemia and 2% hematocrit in 100 µl culture medium in a 96-well plate. Parasites were either left untreated (control) or exposed to [CQ] ex of 10x, 20x or 30x their IC 50 values (Table 1) and incubated for various lengths of time. After incubation, parasites were washed with culture medium to remove CQ and resuspended in 100 µl fresh culture medium. Parasites were incubated for 48 h (length of time for complete assay) at 37 °C, 5% CO 2 , 3% O 2 and 92% N 2 , after which the plates were frozen at −80 °C. Analysis was performed with SYBR Green I detection as described for the growth inhibition assay.
Live cell imaging. For DCS counts in CQS and CQR parasites, synchronized trophozoite stage parasites were exposed to 2x or 10x their CQ IC 50 , or left unexposed (control). A constant temperature of 37 °C was maintained during imaging using a stage-top incubator (Tokai Hit, Shizuoka-ken, Japan). Infected erythrocytes were thoroughly screened for presence or absence of DCS and grouped in time intervals of 30 min. During each time interval, at least 101 (average 169) parasites were examined. DIC images were taken with a Zeiss LSM 710 confocal microscope equipped with a water-corrected objective (C-apochromat 63x/1.20 W Korr M27) and 488 nm laser (12.5 mW, 3.0%) and analyzed using the ZEN 2010 software (Carl Zeiss MicroImaging, Oberkochen, Germany).
For analysis of DV membrane integrity, synchronized early trophozoite stage 3D7 and Dd2 parasites were loaded with 5 µM Fluo-4 AM (Life Technologies, Burlington, ON, Canada) in Ringer's solution (122.5 mM NaCl, 5.4 mM KCl, 1.2 mM CaCl 2 , 0.8 mM MgCl 2 , 11 mM D-glucose, 10 mM HEPES, 1 mM NaH 2 PO 4 , pH 7.4) for 50 min at 37 °C to allow Fluo-4 accumulation in the DV. Parasites were then washed twice with Ringer's solution to remove excess fluorochrome and left untreated, or incubated with 10 µM CQ (400x 3D7 IC 50 ). For analysis, parasites were transferred onto poly-L-lysine (Sigma-Aldrich, Oakville, ON, Canada) coated cover slips in a microscope chamber with or without 10 µM CQ in Ringer's solution. Parasites were kept at 37 °C using a stage-top incubator (Tokai Hit, Shizuoka-ken, Japan) for the entire experiment. Ringer's solution was replaced every 60 min with or without 10 µM CQ for treated and control groups, respectively. To eliminate the possibility of reduced fluorescence in the DV and increased fluorescence in the cytosol due to lengthy exposure time 65 , the parasites were kept in the dark at 37 °C and single images were taken every 15 min with low laser intensity (488 nm laser; 12.5 mW; 0.8% laser power). Parasites were imaged using a time series function on a Zeiss LSM710 confocal microscope (Carl Zeiss, Oberkochen, Germany) equipped with a water-corrected objective (C-apochromat 63x/1.20 W Korr M27) and 488 nm laser (12.5 mW, 0.8%). Images were analyzed using ImageJ 1.47q (National Institutes of Health, USA). Electron microscopy. CQS and CQR parasites were exposed to 20x IC 50 CQ or left unexposed (control) for 3 h, fixed and prepared for electron microscopy. Parasite enrichment was performed using synchronized trophozoite stage parasites. Cultures at 5% parasitemia were used for Easycoll (Cedarlane Laboratories, Burlington, ON, Canada) enrichment using a density gradient. In a 15 ml tube, a gradient of 4 ml of 80% Easycoll, 3 ml of 70% Easycoll, 3 ml of 60% Easycoll and 3 ml of 40% Easycoll was prepared. The Easycoll gradient was adjusted using warm Ringer's solution. Approx. 500 µl parasite culture was resuspended in 2 ml Ringer's solution and layered on top of the gradient. Tubes were centrifuged at 4,700x g (Thermo Fisher Scientific, Mississauga, ON, Canada) for 45 min at 24 °C with low acceleration and decelerated without brake. Trophozoite stage parasites accumulated at the interphase between 60% and 70% Easycoll. Parasites isolated from this phase were washed twice with Ringer's solution and were typically enriched to 62-93% parasitemia. The enriched trophozoite pellet was resuspended in 500 µl fixation solution containing 50 mM cacodylic acid (Carl Roth, Karlsruhe, Germany), 1% paraformaldehyde (Applichem, Darmstadt, Germany), 2.5% glutaraldehyde (Merck KGaA, Darmstadt, Germany), 2.5 mM MgCl 2 and 50 mM KCl, pH 7.2. Samples were centrifuged at 1,700 x g for 20 min at 4 °C, the supernatant aspirated and another 500 µl fixation solution was added overnight at 4 °C. Fixation solution was removed and the pellet was washed three times with 50 mM cacodylic acid, pH 7.2. Samples were stained with 2% OsO 4 in 50 mM cacodylic acid, pH 7.2 at 4 °C for 2 h, then washed with H 2 O, stained with 0.5% aqueous uranyl acetate at 4 °C overnight and washed with H 2 O. Samples were dehydrated and embedded in epon 812 (Serva, Heidelberg, Germany). Ultrathin sections of 50-70 nm thickness were analyzed using a JEM-2100 TEM (JEOL, Eching, Germany) at 200 kV. Images were taken using a 4 K camera from Tietz Video and Image Processing Systems (TVIPS, Gauting, Germany).