Artemisinin kills malaria parasites by damaging proteins and inhibiting the proteasome

Artemisinin and its derivatives (collectively referred to as ARTs) rapidly reduce the parasite burden in Plasmodium falciparum infections, and antimalarial control is highly dependent on ART combination therapies (ACTs). Decreased sensitivity to ARTs is emerging, making it critically important to understand the mechanism of action of ARTs. Here we demonstrate that dihydroartemisinin (DHA), the clinically relevant ART, kills parasites via a two-pronged mechanism, causing protein damage, and compromising parasite proteasome function. The consequent accumulation of proteasome substrates, i.e., unfolded/damaged and polyubiquitinated proteins, activates the ER stress response and underpins DHA-mediated killing. Specific inhibitors of the proteasome cause a similar build-up of polyubiquitinated proteins, leading to parasite killing. Blocking protein synthesis with a translation inhibitor or inhibiting the ubiquitin-activating enzyme, E1, reduces the level of damaged, polyubiquitinated proteins, alleviates the stress response, and dramatically antagonizes DHA activity.

M alaria remains a scourge of humanity, affecting hundreds of millions of people and causing~440,000 deaths each year 1 . Current antimalarial control is highly dependent on artemisinin (ART) combination therapies (ACTs). Thus, it is very concerning that parasites with decreased sensitivity to this drug class have emerged in South East Asia, delaying the clearance of parasites from patients 2 . In areas with concomitant partner drug resistance, up to~50% treatment failure is now observed, and the decreased efficacy of ARTs is putting further pressure on the partner drugs 3,4 . Even more worryingly, the first indigenous African parasite isolate with mutations in the resistance marker (K13) has been reported 5 . Thus, it is critically important to understand the mechanism of action of ARTs, with a view to overcoming resistance.
DHA is known to cause widespread damage to parasite proteins in different cellular compartments [6][7][8] . These damaged proteins would be expected to be targeted for ubiquitination and degradation by the proteasome; and changes that affect this pathway might be expected to help enable parasites to survive DHA exposure.
Mutations in a Kelch domain protein (K13-propeller; PF3D7_1343700) are associated with decreased ART sensitivity in vitro; 9 and genetic manipulation of the K13 locus has confirmed a role in the decreased sensitivity phenotype, although other genetic factors also contribute 10 . K13 shows sequence similarity with a class of substrate adaptors that facilitate ubiquitin ligation 9 , suggesting that protein ubiquitination plays a role in ART action or resistance 11 and mutations in ubiquitination machinery have been reported to be associated with ART resistance 12 . Indeed, we have previously shown that proteasome inhibitors act synergistically with ARTs 13 , further implicating the ubiquitin/proteasome pathway in ART action and resistance. Until now, however, the molecular details have remained unclear.
Here we identify proteasome function as a key target of DHA and pinpoint the build-up of polyubiquitinated proteins and the consequent endoplasmic reticulum (ER) stress response as the key toxic events. We show that inhibition of protein polyubiquitination protects parasites against DHA-mediated killing. The work reveals new points of vulnerability that could be targeted with chemotherapies against malaria parasites.

Results
DHA treatment induces an ER stress response. We previously reported that rings and early trophozoites exhibit growth retardation following exposure to very short (sub-lethal) pulses of DHA 13 . This result suggested DHA may initiate a stress response, so we examined the ER stress response hallmarks, eIF2α phosphorylation and translational arrest. We subjected ART-sensitive (K13 wildtype, 3D7 line) trophozoites to pulsed exposure to DHA (60-90 min), an exposure duration designed to reflect the short (~1 h) in vivo half-life of DHA 14 . We found that eIF2α is phosphorylated in a concentration-dependent manner upon exposure to DHA (Fig. 1a). A similar response was observed in mid-ring stage ( Supplementary Fig. 1a). The level of phosphorylation is similar to that observed with the known inducer of ER stress, dithiothreitol (DTT; Fig. 1a). By contrast, exposure to 1 μM chloroquine (60 min), which has a slower onset of killing 15 , did not induce eIF2α phosphorylation ( Supplementary Fig. 1b). Upon removal of DTT, eIF2α phosphorylation is rapidly reversed, indicating recovery from the stress, whereas DHA-induced phosphorylation is still evident 6 h later (Fig. 1b), indicating prolonged ER stress.
Consistent with ER stress slowing (but not arresting) protein translation 16 , we found that parasite protein synthesis was decreased by 65% in the presence of DTT, while cycloheximide (CHX) caused more complete inhibition (Fig. 1c). Similarly, exposure to 100 nM DHA (1 h) reduced protein translation by5 0% (Fig. 1c). By contrast, a lethal pulse of the antifolate antimalarial, WR99210 which exhibits a slow onset of killing 15 (see Supplementary Fig. 1c for kill curves), had no immediate effect on protein translation. The partial arrest of translation was indicated by lack of complete inhibition, even at higher DHA concentrations ( Supplementary Fig. 1d).
PK4 is responsible for the DHA-initiated ER stress response.
To determine the kinase responsible for eIF2α phosphorylation, we examined the effects of genetically modifying the three eIF2α kinases, eIK1, eIK2 and PK4, that have been identified in P. falciparum 17 . eIK2 (which controls the latency of sporozoites) 18 and, the GCN2 homolog, eIK1 (which regulates eIF2α phosphorylation in response to amino acid starvation) 19 have been shown to be dispensable in P. falciparum 20 . We found that eIF2α is still phosphorylated following DHA and DTT treatment in eIK1-or eIK2-knock-out lines (Fig. 1d), indicating that they are not responsible for DHA-induced eIF2α phosphorylation.
The PERK homologue, PK4, has been reported to control global protein translation levels in late schizonts, and to be essential for growth 20,21 . We generated a hemagglutinin (HA)tagged inducible knockdown transgenic parasite line, PK4 mRNA fused to the glmS riboswitch 22 . Upon the addition of glucosamine (GlcN), the glmS ribozyme self-cleaves at the 3′ UTR and the target mRNA is degraded. As anticipated, upon addition of 10 mM GlcN, marked depletion of PK4-HA was observed (Fig. 1e). While 10 mM GlcN had no effect on eIF2α phosphorylation levels in the 3D7 parent, it ablated DHA/DTT-induced eIF2α phosphorylation in the PK4 knock-down line. These data are consistent with PK4 being the kinase responsible for eIF2α phosphorylation, and further validates the suggestion that DHA initiates an unfolded protein stress response in P. falciparum. To further investigate the role of PK4, we made use of a potent cellpermeable, inhibitor of mammalian PERK ((PERK Inhibitor-I or GSK2606414) 23 ). Exposure of P. falciparum trophozoites to PERK Inhibitor-I (20 μM) for 4 h had no effect on parasite development; however, DHA/DTT-induced eIF2α phosphorylation was prevented (Fig. 1g), again supporting the role of this kinase in DHA-induced eIF2α phosphorylation.
Accumulated polyubiquitinated proteins underpin DHA toxicity. Next, we examined the consequences of DHA exposure on global ubiquitination. Exposure of trophozoites to DHA for 90 min ( Fig. 2a; Supplementary Fig. 2a) or mid-rings to DHA for 3 h ( Supplementary Fig. 2b), resulted in a concentration-dependent increase in polyubiquitinated proteins, consistent with damage to proteins and overwhelming of the protein stress response. By contrast, exposure to chloroquine had little effect on polyubiquitination levels ( Supplementary Fig. 2c). A similar build-up of polyubiquitinated proteins was induced by the proteasome inhibitor, epoxomicin (3 h) (Fig. 2b, Supplementary Fig. 3a, b.). This accumulation of ubiquitinated proteins appears to be toxic to parasites, as exposure to b-AP15, an inhibitor of the deubiquitinases that remove ubiquitin from substrates before degradation 24,25 , or RA190, an inhibitor of the ubiquitin receptor 26 , both caused a build-up of ubiquitinated proteins (Fig. 2b, Supplementary Fig. 3a, b) and led to parasite death (Supplementary Fig. 2d).
DHA treatment inhibits proteasome function. The rapid accumulation of ubiquitinated proteins in DHA-treated parasites led us to consider the possibility that DHA inhibits the activity of the proteasome. Increasing concentrations of DHA and epoxomicin decreased proteasome activity (as monitored using the proteasome-GLO system 27 ) (Fig. 2c). Additionally, when extracts of treated parasites were subjected to native gel electrophoresis and incubated with a fluorogenic peptide substrate 28 , a significant decrease in 26S activity was evident, upon even a 60-min exposure to DHA (Fig. 2d, e), while a lethal dose of WR99210 (see kill curve in Supplementary Fig. 1c) did not inhibit activity (Fig. 2d, e). This effect did not result from a decrease in the abundance of the proteasome (as indicated by Western blotting; Supplementary Fig 2e), suggesting that DHA targets Anti-Pf BiP Anti-eIF2α (total) Anti-Pf GAPDH proteasome function, rather than causing proteasome turnover.
The loss of proteasome activity may occur indirectly via clogging of the proteasome with ubiquitinated proteins, or directly via protein damage or dissociation of subunits, in a manner similar to that for oxidative damage of the mammalian 26S proteasome 29 , and consistent with a study showing that an ART derivative becomes cross-linked to a P. falciparum proteasome subunit 8 .
To further examine the in vivo activity of the proteasome in DHA-treated parasites we used transfectants expressing GFP fused to a destabilisation domain (DD) 30 (Fig. 2f), which is Error bars represent s.e.m. **P < 0.01 between mock and 10 μM DHA-treated samples, ***P < 0.001 between mock and epoxomicin treated samples (n = 4), ANOVA. f Schematic of the GFP-DD reporter system. g, h GFP-DD transfectants (in a 3D7 background) were maintained in Shield-1 for 24 h, before wash-out. Trophozoites were re-exposed to Shield-1 or indicated compounds for 4 h. targeted for degradation by the proteasome in the absence of the protective ligand, Shield-1. Inhibitors of proteasome-mediated degradation (epoxomicin, b-AP15 or RA190) prevented GFP-DD degradation in the absence of Shield-1 (Fig. 2g, Supplementary  Fig. 2f). Similarly, treatment with DHA, but not WR99210, caused a build-up of full-length protein (Fig. 2g), confirming that DHA disrupts proteasome-dependent degradation.
DHA treatment causes a build-up of damaged proteins. We used the fluorescence signal from GFP-DD to investigate protein unfolding (or misfolding or damage). In the absence of Shield-1, or upon treatment with WR99210, fluorescence was low but not abolished, consistent with degradation of part of the population of the fluorescent protein (Fig. 2h, Supplementary Fig. 2g). Treatment with Shield-1 increased the fluorescence signal, indicating an increase in full-length, properly folded GFP-DD. Treatment with epoxomicin, b-AP15 or RA190 also increased GFP-DD fluorescence (Fig. 2h, Supplementary Fig. 2g), indicating that the GFP remains folded when the proteasome is inactivated by these compounds. By contrast, upon DHA exposure, the GFP-DD fluorescence was reduced to below background (minus Shield-1) levels ( Fig. 2h, Supplementary Fig 2g), consistent with protein unfolding/damage. Taken together, our findings show that DHA induces protein damage and prevents degradation via the proteasome, leading to an accumulation of damaged, polyubiquitinated proteins.
Inhibiting polyubiquitination antagonises DHA activity. To investigate further the nature of the toxic events leading to parasite killing, we examined different strategies for preventing the accumulation of proteasome substrates. CHX arrests protein translation, a state that protects mammalian cells from the lethality induced by proteasome inhibition 31 . We found that treatment of trophozoites with a pulse of CHX (up to 20 μM for 6 h) had no effect on viability ( Supplementary Fig. 4a). Exposure of trophozoites to CHX decreased the base-line level of polyubiquitinated proteins (Fig. 3a, Supplementary Fig. 3c, d). Similarly, exposure to a combined pulse of CHX and DHA markedly reduced the DHA-mediated build-up of polyubiquitinated proteins (Fig. 3b, Supplementary Fig. 3e-f), and decreased DHAmediated unfolding/damage of GFP-DD (Fig. 3c, Supplementary  Fig. 4b). Importantly, CHX co-treatment strongly inhibited DHA-mediated eIF2α phosphorylation (Fig. 3d) and strongly antagonised DHA-mediated killing (Fig. 3e, Supplementary  Fig. 4c, d), confirming that accumulation of unfolded, polyubiquitinated proteins is a key toxic event initiated by DHA treatment. Compound 1 (5′-O-sulphamoyl-N(6)-[(1S)−2,3-dihydro-1Hinden-1-yl]-adenosine), hereafter referred to as C1, is a paninhibitor of ubiquitin-activating enzymes (E1), which depletes intracellular pools of activated ubiquitin 32 . Treatment of trophozoites with a pulse of C1 (6 h up to 20 μM) had no effect on parasite viability (Supplementary Fig. 5a). However, it decreased the base-line level of polyubiquitinated proteins (Fig. 3a,  Supplementary Fig. 3c, d) and increased the GFP-DD fluorescence in the absence of Shield-1 ( Supplementary Fig. 5b), consistent with accumulation of proteasome substrate that is not recognised/degraded as it is not ubiquitinated. Additionally, C1 treatment reduced DHA-mediated accumulation of polyubiquitinated proteins to below background levels (Fig. 3b,  Supplementary Fig. 3e-f) and strongly inhibited the DHAmediated unfolding/damage of GFP-DD (Fig. 3c, Supplementary  Fig. 4b). Accordingly, C1 reduced the DHA-induced build-up of phosphorylation of eIF2α (Fig. 3d) and strongly antagonised DHA-mediated killing (Fig. 3f, Supplementary Fig. 5c, d). C1 also antagonized the build-up of polyubiquitinated proteins and killing induced by epoxomicin ( Supplementary Fig. 5e, f), but did not prevent DHA-mediated damage to proteasome function ( Supplementary Fig. 5g).

Discussion
Consistent with recent reports showing that ARTs induce widespread cellular damage 6-8 , our data indicate that DHA causes protein damage/unfolding and inhibits folding of newly synthesised proteins. This likely leads to global perturbation of cell function and viability, compromising protein functionality 8 and inducing protein aggregation 34 .
We show that the DHA-initiated proteostatic stress induces eIF2α phosphorylation via activation of PK4, the Plasmodium homologue of mammalian PERK. By analogy with PERK, PK4 is likely an integral membrane protein in the ER with an N-terminal sensor of unfolded proteins protruding into the ER lumen (the sequence has one strongly predicted transmembrane domain and several other putative transmembrane helices). Upon accumulation of damaged proteins, the cytoplasmic C-terminal catalytic domain is activated and phosphorylates eIF2α. Indeed, we show that DHA treatment leads to prolonged eIF2α phosphorylation and attenuation (but not arrest) of protein synthesis, consistent with stressinduced, eIF2α-mediated disruption of translation. The decreased protein translation initiated by DHA likely underpins reports of growth retardation 35 and quiescence 36 upon exposure to ARTs.
Indeed, while this manuscript was under review, another study 37 showed that DHA treatment leads to an increase in eIF2α phosphorylation and repression of translation and that overexpression of a PK4 dominant-negative mutant blocks translational arrest and enhances the ability of DHA to kill parasites. This work provides complementary evidence for an important contribution of the stress response to the killing mechanism.
We propose that an unresolved ER stress response underpins DHA-mediated toxicity, compounded by DHA-mediated (direct or indirect) inhibition of proteasome function, which would prevent removal of damaged proteins, compromise amino acid homeostasis 31 and inhibit processing of important cellular factors 38 , leading to eventual proteostasis network collapse 39 . Importantly, we show that following short duration exposure to DHA, damage to specific protein targets occurs, but is not, in and of itself, the toxic event. That is, we show that in the presence of a potent inhibitor of ubiquitin activation (C1), DHA treatment still leads to decreased proteasome function, but parasite killing is largely avoided. Similarly, inhibition of translation by CHX, which as expected decreases protein ubiquitination, strongly antagonises the action of DHA. The antagonism of DHAmediated killing by an inhibitor of NEDD8 activation (MLN4924) suggests the involvement of a cullin E3-mediated ubiquitin ligation event, although additional work is needed to confirm the involvement of NEDD8.
Taken together our results lead us to propose that DHA packs a double whammy, damaging and unfolding proteins and inhibiting proteasome function, with the consequent build-up of ubiquitinated proteins as the key event that leads to cell death, likely via the unresolved ER stress pathway 35 . As illustrated in Fig. 4, this work defines the critical events in the mechanism of killing by ARTs.
Our work provides insights into the possible role of K13 in ART action. K13 shares sequence similarity with BTB/Kelch substrate adaptors that facilitate ubiquitin ligation by cullin-3 ligases 9 , such as the putative P. falciparum cullin-3 (PF3D7_0629800). Human cullin-3 works with~70 different BTB/Kelch adaptors to ubiquitinate different substrates. By contrast, K13 is the only BTB/Kelch protein encoded by P. falciparum, consistent with it playing a major role in targeting a substantive protein sub-set to the proteasome. We propose that K13 is an adaptor that plays a critical role in presenting damaged proteins for ubiquitination, leading either directly or indirectly to enhanced ubiquitination levels, and that the consequent build-up of polyubiquitinated proteins clogs the proteasome 40 , eventually leading to cell death (Fig. 4). The enhanced survival of K13 mutants may be due to a defect in presenting damaged proteins for polyubiquitinatation. Increased expression of chaperones and proteostasis pathway proteins 41 , and an increased basal level phosphorylation of eIF2α in early ring stage parasites 37 likely also contribute to survival.
Our work identifies the ubiquitin-proteasome system as a point of critical vulnerability in P. falciparum, suggesting a number of potential targets, including the proteasome itself 38 and the deubiquitinating enzymes that feed proteins to it 42 . A recent saturation mutagenesis showed that 54 of the 72 genes in the proteasome degradation pathway can be classified as essential and that single-insertion mutants of an ART sensitivity cluster of genes showed increased sensitivity to the proteasome inhibitor, Bortezomib 43 . It is of particular interest that mutations in the deubiquitinase, UBP1, have been observed in ART-pressured parasites 12 , and that UBP1 mutations are also associated with decreased ACT effectiveness in Africa 44 and South East Asia 45 . Mutations in different components of the ubiquitin-proteasome pathway may enhance ART resistance by helping to reduce accumulation of polyubiquitinated proteins. Obversely, targeting components of the ubiquitin-proteasome system provides a means of overcoming resistance-inducing mutations, thus maintaining the impact of control programs.
Culturing and cell lines. P. falciparum parasites used in this study were propagated in O+ human RBCs (Australian Red Cross Blood Service) in RPMI-1640, supplemented with GlutaMAX TM , 25 mM HEPES (ThermoFisher), 5% (v/v) human serum (Australian Red Cross Blood Service), 0.25% (w/v) AlbuMAX II (Life Technologies), 10 μM D-glucose, 22 μg mL −1 gentamycin, and 0.5 mM hypoxanthine, and incubated at 37°C in an atmosphere of 1% O 2 , 5% CO 2 and 94% N 2 . Cultures were monitored by Giemsa staining of methanol-fixed blood smears. Culture media was replaced at least every 48 h and parasitemia was maintained below 10% to ensure health of cultures.
P. falciparum strains used in this study were: 3D7 laboratory strain, Cam3.II rev (K13 WT ) 10 and a GFP-DD reporter strain generated with plasmid pEF-GFP-HA-DD24 (ref. 46 ), kindly provided by Dr Mauro F Azevedo and Dr Paul Gilson (Burnet Institute). eIK1 and eIK2 knock-out lines were kindly provided by Prof Christian Doerig (Monash University) 20 . PK4-HA-glmS ribozyme-inducible knockdown parasites were generated by PCR amplification of the~850 bp region preceding the 3′ end of pk4 (PF3D7_0628200, gDNA isolated from 3D7). PCR products were ligated into pGlmS-3HA to generate final targeting vectors 22,47 . Cell lines were verified to be mycoplasma free.
Protein synthesis assays. Trophozoite-stage cultures at 2% haematocrit and 10% parasitemia were enriched by magnet purification and allowed to recover at 37°C for 1 h prior to experimentation. 1 × 10 7 cells were subjected to the drug treatment indicated in the figure legend in the presence of 3 H-tryptophan (20.1 mCi mmol −1 stock, Perkin Elmer) at 37°C. Samples were washed with ice-cold PBS and the acidinsoluble radioactive protein products were precipitated in 30% (w/v) TCA. The precipitated protein pellet was washed once with 5% (w/v) TCA and once with MilliQ ddH 2 O. Precipitated proteins were solubilised by incubation in 1 M NaOH at 50°C for 30 min. The mixture was decolourised by incubation in 12.5% (w/v) NaClO for 20 min at room temperature. Solubilised radiolabelled proteins were then diluted in MilliQ ddH 2 O and transferred to vials containing scintillation fluid for β-scintillation counting using a Tri-Carb 4810TR Liquid Scintillation Counter (Perkin Elmer). Data were processed and visualised using GraphPad Prism software (version 5). Values presented are normalised to mock (0.1% DMSO)-treated sample (100% 3 H-Trp incorporation).
Analysis of proteasome activity by proteasome-GLO system. Trophozoitestage parasite cultures at 3% haematocrit and 4% parasitemia were subjected to the drug treatment indicated in the figure legend. RBC cytoplasmic components were released by incubation in 0.05% (w/v) saponin and parasite pellets were washed with ice-cold PBS three times. The parasite pellet was lysed in ice-cold water (pH 4.5) and the supernatant was cleared by centrifugation at 760×g for 10 min. One   Fig. 4 Model of DHA-mediated killing of P. falciparum. Activated ART (ART*) initiates a multi-pronged assault on protein homeostasis by damaging and unfolding proteins, preventing folding of newly synthesised proteins and inhibiting the proteasome (toxicity indicated with red bar). Activation of the ER Stress response, attenuates translation. Prolonged activation of the ER Stress response, and accumulation of toxic polyubiquitinated substrates (S) eventually leads to cell death. Other proteasome inhibitors also cause accumulation of polyubiquitinated proteins and parasite killing (Epoxomicin (Epo), b-AP15, and RA190). Inhibition of the E1/E2/E3 ubiquitin machinery (C1 or MLN4924) or ablation of protein synthesis (cycloheximide, CHX), prevents accumulation of polyubiquitinated proteins and enables cell survival (rescue indicated with green bar). In ART resistant parasites, defective presentation of proteins for polyubiquitination, may enable cell survival microgram cell lysate was mixed with an equal volume of the Proteasome-GLO (chymotrypsin-like) assay reagents (Bio-Rad). Luminance was measured using CLARIOstar microplate reader (BMG Labtech). Measurements were taken every 2 min for 3 h. Average readings between 60 and 90 min (after the signal reached equilibrium) were used for proteasome activity analysis. Values are normalised to mock (0.1% DMSO)-treated sample (100% Proteasome-GLO signal).
Analysis of proteasome activity by native gel electrophoresis. Trophozoitestage parasite cultures at 2.5% haematocrit and 5% parasitemia were enriched by magnet purification and subjected to the drug treatment indicated in the figure legend at 37°C. Cell pellets were washed with ice-cold PBS and RBC cytoplasmic components released by incubation in 0.15% (w/v) saponin. The parasite pellet was solubilised in hypotonic lysis buffer (25 mM Tris-HCl, 10 mM MgCl 2 , 1 mM ATP, pH 7.4) on ice for 30 min. The supernatant was cleared by centrifugation at 13,000×g for 15 min and analysed for protein content by BCA assay (Pierce). Samples were analysed for proteasome activity using an in-gel assay 49 . Fifty microgram cell lysate was mixed with native loading dye (250 mM Tris-HCl pH 7.4, 50% (v/v) glycerol, 0.0007% xylene cyanol) and loaded onto a 4% acrylamide gel. Electrophoresis was carried out at 130 V for 90 min at 4°C. Following electrophoresis, the gel was incubated in developing buffer (50 mM Tris-HCl pH 7.4, 5 mM MgCl 2 , 1 mM ATP) supplemented with 100 μM suc-LLVY-AMC (Enzo Life Sciences) for 30 min at 37°C, after which time the signal was captured under UV light using the Bio-Rad ChemiDoc TM MP imaging system. All reagents contained 1 mM ATP to maintain proteasome activity.
For Coomassie staining, following UV imaging, the gel was washed three times in MilliQ ddH 2 O and stained with SimplyBlue TM SafeStain (Invitrogen) according to the manufacturer's instructions. Densitometric analysis was carried out using ImageJ software (Gel Analysis method). The integrated pixel intensity for each entire lane was corrected for background signal and normalised to the sum of the signal produced by all samples (100% proteasome activity). Brightness and contrast were not adjusted in the main figure.
Measurement of GFP fluorescence using flow cytometry. Trophozoite-stage parasite cultures at 5% haematocrit and 5% parasitemia were subjected to the drug treatment indicated in the figure legends at 37°C. Culture samples were stained with 2 μM SYTO-61 50 and adjusted to 0.1% haematocrit in PBS. SYTO-61 and GFP fluorescence was recorded using the FACSCanto™ II cytometer (Becton Dickinson). Analysis was performed in FlowJo (version10). Parasites were gated based on the SYTO-61 signal, and parasite GFP fluorescence values reported are mean fluorescence values. Data were fit by nonlinear regression using GraphPad Prism Software (version 5).
Assessment of parasite viability following drug treatment. Drug pulse assays were performed using established methods 13, 51-53 . Tightly synchronised parasites (0-2 h post invasion (h.p.i.) for ring stage or 24-27 h.p.i. for trophozoite stage) at 0.2% hematocrit and 1-2% parasitemia were subjected to the indicated drug pulse at 37°C. Cells pellets were washed thoroughly to remove drug and returned to culture until parasites reached trophozoite stage of the following cycle. Cells were then stained with 2 μM SYTO-61 and their fluorescence measured by flow cytometry (FACSCanto™ II cytometer; Becton Dickinson). Data were gated and analysed using FlowJo software (version 7.6.5) to determine parasitemia of each sample. Parasitemia values were then graphed by nonlinear regression using GraphPad Prism software (version 5). Survival (%) represents the parasitemia normalised to untreated (100% survival) and kill-treated (0% survival) controls, where kill-treated refers to samples treated with 2 μM DHA for 48-72 h.
Statistical analysis. Statistical significance between groups was determined using one-way ANOVA with Dunnett's Multiple Comparison Test (Figs. 1c, 2e) or unpaired two-tailed t-test ( Supplementary Fig. 6c) (with GraphPad Prism Software, version 5). P values were considered significant if P < 0.05. In all cases, n, refers to the number of biological replicates.