Plasmodium genomes contain a DMT1 homolog, which resides on the vacuole and can functionally complement yeast smf3 for its iron-deficient phenotype

Bioinformatic analysis shows that all Plasmodium genomes encode one gene product with homology to Saccharomyces cerevisiae (the yeast) Smf1-3, Homo sapiens Nramp1 and Dmt1 (Nramp2) (Fig. 1a). This Dmt1 homolog is PF3D7_0523800 in P. falciparum and PY17X_124180 in P. yoelii (PyDMT1). PyDMT1 shares 29.53% amino-acid sequence identity with Homo sapiens Dmt1, and 24.94% with S. cerevisiae Smf3. PyDMT1 is predicted to have twelve transmembrane domains. In order to reveal the expression and subcellular localization of PyDMT1 protein, we fused a HA tag in frame with the endogenous PyDMT1 coding region at the carboxy terminus by CRISPR-Cas9 [17, 18] in P. yoelii 17X, and obtained a recombinant PyDMT1-HA P. yoelii strain (Fig S1a-b). Confocal analysis after immunostaining with HA antibody showed that PyDMT-HA surrounded the hemozoin in the blood stage (Fig. 1b), implying that PyDMT1 is likely localized on the membrane of the food vacuole. The transcription level of DMT1 in the PyDMT1-HA parasites was consistent with that of the wild-type parasites; there was no significant difference in growth between the two strains after infecting mice in the blood stage (Fig S1e-f), which means that the HA fusion has unlikely disrupted the function of PyDMT1.

Fig. 1

PyDMT1 resides in the vacuole and could functionally rescue the iron defect of △smf3 yeast. a Phylogenetic analyses of Plasmodium DMT1 proteins with human and yeast (S. cerevisiae) homologs. The phylogenetic tree was generated with MEGA (Molecular Evolutionary Genetics Analysis). All Plasmodium species examined contain one DMT1 homolog. b Intraparasitic PyDMT1 localizes to Hz-containing DVs in P. yoelii. Shown here are a merge of DIC, Hoechst 33342 nuclear stain (blue), and the signal of tagged PyDMT1(green). c Indirect immunofluorescence assays of PyDMT1-HA and PyCRT-FLAG in the blood stage. The endogenous PyDMT1 gene fused to HA and the food vacuole marker PyCRT-FLAG were imaged. Shown are the HA channel (red, first row), the FLAG channel (green, second row), and a merge of both signals (third row). d Complementation of △smf3 yeast iron deficiency with PyDMT1 and an N-terminal truncated PyDMT1. The regulatory region of FET3 (− 300 bp–6 bp) directed LacZ activity (to indicate FET3 expression response). △smf3 and the parent strain by4742 transformed with the empty vector pYES2-ADH were the controls. Shown are mean values ± SD. ***P < 0.001; one-way ANOVA and Tukey’s multiple comparison test. N = 3. e Whole-cell iron content in △smf3 yeast expressing PyDMT1. Shown are mean values ± SD. *P < 0.05; t-test. N = 3. f Iron content of the vacuole from △smf3 yeast expressing PyDMT1yeast. Shown are mean values ± SD. *P < 0.05; one-way ANOVA t-test. N = 3. g Growth complementation of △smf3 yeast with the expression of PyDMT1 or an N-terminal truncated PyDMT1

Plasmodium CRT (chloroquine-resistant transport) is known to localize on the vacuolar membrane, responsible for chloroquine resistance [19]. To further confirm that PyDMT1 is localized on the food vacuole membrane, we fused a FLAG tag at the C-terminal of CRT by single crossover to label the food vacuole (Fig S1c-d). Confocal imaging revealed that PyDMT1 indeed mainly colocalized with CRT at all blood stages (Fig. 1c), and slightly with the Golgi marker Rab6 (Fig S1h), but completely not with the endoplasmic reticulum marker PMV (Fig S1g). During the schizont stage, a small amount of PyDMT1 signal is found in the merozoites, mainly in the hemozoin region. These data strongly suggest that PyDMT1 mainly localizes on the digestive vacuole membrane.

Sequence comparison indicates that S. cerevisiae Smf3 bears a closer resemblance to PyDMT1 than Smf1 and Smf2 (Fig. 1a), consistent with the finding that Smf3 is located on the vacuole. In the yeast, Smf3 transports iron from the vacuole into the cytosol, and Smf3 deletion makes yeast iron deficient, accompanied by elevated Fet3 expression [20]. The expression of Fet3 can thus act as a sensitive indicator for iron deficiency. To assess whether PyDMT1 could rescue the iron deficiency phenotype in △smf3, a yeast expression vector (pYES2-ADH) with a codon-optimized full-length Pydmt1(696 aa) and a 5’-truncated Pydmt1(470 aa) containing all 12 transmembrane domains were transformed into △smf3 yeast. A parallel transformation with the empty vector was used as the negative control. FET3 promoter activity was monitored by a reporter construct wherein the FET3 gene promoter was fused to LacZ (β-galactosidase gene Z). The results showed that both the full-length Pydmt1 and 5’-truncated Pydmt1 conferred significant rescue to the △smf3 (Fig. 1d). To confirm that expressing PyDMT1 in △smf3 yeast enables iron transport from the vacuole to the cytosol, we quantified the whole-cell iron and vacuole iron contents. Consistently, the whole cell and vacuole iron levels were both lower when expressing a codon-optimized PyDMT1 (Fig. 1e, f). Plate growth assays also showed that expressing PyDMT1 restored the growth of the △smf3 strain (Fig. 1g), indicating that PyDMT1 can function heterogeneously in S. cerevisiae and act as an iron importer there.

PyDMT1 is critical for parasite development in the blood stage

To clarify the physiological function of PyDMT1 in Plasmodium, we attempted to knock out PyDMT1 in P. yoelii17X by CRISPR-Cas9 [17, 18]. Nine guide RNAs were designed to cover the Pydmt1 gene (Fig. 2a, Table S1), and each guide RNA and homologous recombination arms were independently electro-transfected five times. Unfortunately, no Pydmt1 knockout parasites were obtained in all these trials. We, therefore, suspect that PyDMT1 may be indispensable for the viability of parasites.

Fig. 2

PyDMT1 hypomorphs are associated with reduced growth and suppressed pathogenicity. a A schematic showing positions of nine guide RNAs and homologous recombination arms designed for Pydmt1 knockout. Each guide RNA and flanking homologous recombination arms were cloned in the vector expressing Cas9, and parasites were selected with pyrimethamine after transfection. All failed in repeated trials. Different homologous recombination arms were also tried in vain. b Single crossover strategy for PyDMT1 knockdown in P. yoelii. The strategy was designed in such a way that after a successful crossing-over, the 4.1-kb region upstream from the translation start of the neighboring gene would remain intact, leaving only the Pydmt1 regulatory region truncated at the insertion point. Arrows indicate various truncations upstream of Pydmt1 CDS had been tried to make the hypomorphic mutant. c qPCR detection of Pydmt1 transcript levels for PyDMT1 hypomorphic clones A1 and C2. The control parasites of PyDMT1-HA and hypomorphic clones A1and C2 were purified from mouse blood and subjected to qPCR using primers targeting the 3’ segment of PyDMT1. PyDMT1 transcript abundance was normalized to tubulin. Shown are mean values + / − SD. **P < 0.01; N = 3. d Protein levels of Pydmt1-A1 and Pydmt1-C2 as detected by western blots. HA antibody was used for detection. The specificity of the HA antibody was demonstrated with none-HA carrying parasite strain (Figure S2d), suggesting all the three bands from the western blot were derived from PyDMT1. e Reduced growth of the Pydmt1 hypomorphic mutants. Parasitemia of BALB/c mice following infection (i.v.) with 106 PyDMT1 or PyDMT1 hypomorphic iRBC was determined by counting iRBC in Giemsa-stained blood smears(N = 5, one-way ANOVA and Tukey’s multiple comparison test, **P < 0.01, ***P < 0.001). f Loss of pathogenicity for the Pydmt1 hypomorphic mutants. Shown here is the survival of BALB/c mice inoculated with 10.6 control P. yoelii or hypomorphic PyDMT1 parasite-infected iRBC (N = 5, log-rank test, ***P < 0.001)

For an essential gene, the alternative strategy is to generate a partial loss of function mutant (hypomorph). Truncating the regulatory region by inserting a selectable cassette into the upstream of a gene could reduce its expression [21]. We decided to insert an hdhfr(coding for the human dihydrofolate reductase) selectable cassette at the Pydmt1 locus upstream of the coding region of Pydmt1-HA to knock down its expression. In the Plasmodium database, we found that Pydmt1 shares the same regulatory region of approximately 4.1 kb with an adjacent gene, PY17X_1241700, whose function remains unknown. To minimize the undesirable effect on this neighbor gene, we designed a single insertion strategy to shorten the regulatory region of Pydmt1 without altering the entire 4.1 kb regulatory region of PY17X_1241700 (Fig. 2b). Multiple attempts to put the cassette in at − 400, − 600, − 800, − 1200, and − 1600 bp upstream of Pydmt1 were unsuccessful. Nevertheless, we finally obtained a recombinant clone by inserting the hdhfr cassette at the − 2000 bp position (Fig. 2b). After limiting dilution and mice infection, we obtained six independent clones (Fig S2a-b), among which two clones (Pydmt1-A1 and Pydmt1-C2) were selected for Pydmt1 expression analysis. Inserting hdhfr at the − 2000 bp led to a reduction of Pydmt1 expression in the clones, demonstrated at both the mRNA and protein levels (Fig. 2c, d). This insertion, however, had little influence on the mRNA expression of the flanking gene PY17X_1241700 (Fig S2c). Amazingly, this roughly halving of PyDMT1 expression led to a significantly reduction of parasitemia; 3 days after infection the parasitemia for Pydmt1-C2 clone and Pydmt1-A1clone was each reduced by 50 and 60%, respectively (Fig. 2e). Furthermore, BALB/c mice infected with the hypomorphic PyDMT1 parasites survived significantly longer than mice infected with control parasites (Fig. 2f). These results suggest that full PyDMT1 activity is critical for P. yoelii pathogenicity, and even partial reduction of PyDMT1 expression would seriously cripple its lethality. An inference of this finding is that although iron is so critical for the pathogen’s survival, normal P. yoelii does not contain much higher iron levels than needed.

PyDMT1 hypomorph could be completely rescued by ferrous supplementation in vitro

In mammals, DMT1 is the major iron transporter and contributes to non-heme iron uptake in most types of cell. In Dmt1−/− mice, significant hepatic iron stores occurred but these mice died of anemia by day 7 [14, 16]. To test the hypothesis that the slow growth of the hypomorphic PyDMT1 parasite was due to iron deficiency caused by lower iron uptake, we counted the merozoites of the mutant Pydmt1 parasite and the parent strain without or with exogenous iron in vitro by adding Fe2+. After 18 h incubation, the number of merozoites for Pydmt1 mutant parasites (clone A1 and clone C2) was about 9, in comparison to 14 for the parent strain, consistent with the slower in vivo growth of the mutant. Meanwhile, addition of Fe2+ did not significantly increase the number of merozoites for the parent strain, indicating that iron was not a limiting factor for the normal parasite under our experimental condition. In contrast, iron supplements significantly increased the number of merozoites for the Pydmt1 mutant parasite, to a comparable level with that in the control parasite (Fig. 3a, b). The data indicate that it is primarily iron deficiency and not anything else that underlies the growth defects from reduced PyDMT1 expression. Next, we compared the labile iron pool (LIP) in the control parasite and clone A1. To that end, we used a similar strategy adopted by other groups [6, 22]; iRBCs containing both the parent parasite and clone A1 were stained with the iron-sensitive fluorescent probe Calcein-AM, and the signals were analyzed by flow cytometry. The LIP of the Pydmt1 mutant parasite iRBC was significantly lower than that of the control parasite iRBC (Fig. 3c, d, Fig S2e). Noteworthy is that Fe2+ quenches the fluorescence so that higher Fe2+ levels result in lower signals. In order to make sure that these signal changes were primarily from the parasites instead of the RBC, a confocal image was further taken. The primary signal of the Calcein-AM stain indeed appeared from the parasites (Fig. 3e, f). Furthermore, the diffuse and stronger signal across the whole parasite suggests a reduced LIP of Pydmt1-A1 in the cytosol.

Fig. 3

Iron could completely rescue the growth defects of PyDMT1 hypomorph in vitro. a Schizont maturation of Pydmt1-C2 rescued with 50 μM FeSO4. One hundred micromolars ascorbic acid was added to prevent oxidation/maintain Fe2+ status. After 18 h incubation (2 p.m. to 8 a.m., at the atmosphere of 90% N2, 5% O2, and 5% CO2), most parasites in early ring stage developed to mature schizont (one hemozoin is surrounded with merozoites before release). Merozoites were counted under a microscope with blood smear. Shown are values from individual parasites as well as the mean values. One-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001. N = 80 parasites from three independent experiments. b Merozoite numbers of Pydmt1-A1 rescued with 50 μM FeSO4 in the presence of 100 μM ascorbic acid (N = 80 parasites from three independent experiments, one-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001). c A representative histogram showing the Calcein-AM fluorescence intensities of the control PyDMT1-HA(N1) and the mutant Pydmt1-A1(A1) parasites. FITC-A in x-axis represents the Calcein-AM fluorescence intensity. DIP (2,2-dipyridyl) is an iron chelator and the differential ΔMFI after iron chelation is commonly used to measure the labile iron pool. d The LIPs of PyDMT1-HA and Pydmt1-A1 analyzed by flow cytometry, as shown in c. △MFI was determined by evaluating the change in mean fluorescence intensity of Calcein-AM-loaded iRBCs (Hoechst 33,342-positive subset, indicated with PB450A), after incubation with 100 μM DIP (N = 5, Mann Whitney test, *P < 0.05). e Representative images of the control PyDMT1-HA and Pydmt1-A1 parasite-infected RBCs after Calcein-AM treatment. Parasite-infected RBCs were incubated with Calcein-AM and DIP 1 h (the second row) or without (the first row). After three times washing, the iRBCs were transferred to the plate pretreated with poly-lysine, and imaged by Olympus FV3000 laser scanning confocal microscope. Quantitative data shown in f. f The LIPs of PyDMT1-HA and Pydmt1-A1 analyzed by confocal microscopy as shown in e. △MFI was determined by evaluating the change in mean fluorescence intensity of Calcein-AM in the region of parasite, after incubation with 100 μM DIP (N = 15 from three independent experiments, t-test, *P < 0.05.). g The schizont maturation of Pydmt1-A1 rescued with 100 μM Zn2+ (N = 80 parasites from three independent experiments. One-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001). h The schizont maturation of Pydmt1-A1 supplemented with 100 μM Cu2+ or Mn.2+ (N = 80 parasites from three independent experiments. One-way ANOVA and Tukey’s multiple comparison test. n.s. P > 0.05)

We then determined whether the number of merozoites in Pydmt1 mutant parasite could be reverted by other metals. Given the same nature of the different Pydmt1 hypomorph mutants, essentially identical results obtained with A1 and C2, and the laborious amount of work involved, we chose A1 for this work. Zinc addition partially restored the number of merozoites of clone A1 (Fig. 3g), meaning that zinc supplement could, to some extent, rescue the growth defect of Pydmt1 mutant parasites. However, adding copper or manganese ions did not affect the merozoite maturation (Fig. 3h).

Taken together, we conclude that the primary physiological function of PyDMT1 is iron uptake, and decreased PyDMT1 expression causes parasite iron deficiency, leading to severe physiological consequences. The perfect rescue of the PyDMT1 hypomorphs with iron, together with the observation of the inability of iron addition to enhance the growth of the wild-type pathogen, confirms our suspicion that the normal pathogen may obtain just enough iron for its optimal growth.

Inducible ferritin knockout could effectively suppress the growth defect of the hypomorphic PyDMT1 parasite

After achieving complete rescue of the Pydmt1 mutant parasite in vitro by iron addition, we next tried this strategy in vivo. We intraperitoneally injected 100 mg/kg iron dextran into parasite-infected mice. One hundred milligrams per kilogram iron dextran injection did increase the total iron content of the plasma and erythrocytes (Fig S3a), but surprisingly had no effect on the parasitemia of the Pydmt1 mutant parasite (Fig S3b). Concerning that 100 mg/kg iron dextran might not be sufficient to increase the intracellular iron content of the parasite, we increased the concentration of iron dextran to 200 mg/kg. Still, the parasitemia of the Pydmt1 mutant parasite was not restored (Fig. 4a). We subsequently analyzed the LIP of the RBCs and found that it had no change after iron dextran administration (Fig. 4b), meaning elevated plasma iron levels did not translate to an increase of iron in the erythrocytes. This explains why iron dextran injection did not rescue the growth defects of our iron-deficient malaria parasites.

Fig. 4

Inducible knockout of host ferritin could normalize the growth of PyDMT1 hypomorph to that of the control parasite. a Parasitemia of BALB/c mice infected with PyDMT1-HA and Pydmt1-C2 parasites, after injection with 200 mg/kg iron dextran. Iron dextran was injected for 3 consecutive days in advance before PyDMT1-HA and C2 inoculation. Parasitemia were recorded at day 2 and day 3 (the day of inoculation as day 0) (N = 5, one-way ANOVA and Tukey’s multiple comparison test. ***P < 0.001). b LIP of RBC after 5 days iron dextran injection to the mice (N = 3, t-test, n.s. no significance). c The schedule of ferritin (Fth1) inducible knockout. d Body weights and survivals (N = 11, log-rank test, **P < 0.01) of FthR26△/△ (N = 14) and Fthfl/fl mice, after tamoxifen administration (start at day 0). e Expression of Fth1, Fpn1, Dmt1, and Tfr1 mRNA of blood cells in Fthfl/fl (N = 3) and FthR26△/△ (N = 3), 10 days after tamoxifen administration (t-test, ***P < 0.001, **P < 0.01), suggesting an elevation of intracellular iron in blood cells. f Parasitemia of FthR26△/△ (N = 7) and Fthfl/fl (N = 5) mice after 2 days of infection with PyDMT1-HA and Pydmt1-C2 parasites (one-way ANOVA and Tukey’s multiple comparison test, ***P < 0.001, **P < 0.01). g Parasitemia of FthR26△/△ (N = 7) and Fthfl/fl (N = 5) mice 3 days after infection with PyDMT1-HA or Pydmt1-C2 parasites (one-way ANOVA and Tukey’s multiple comparison test, ***P < 0.001, **P < 0.01). h Merozoite numbers of PyDMT1-HA and Pydmt1-C2 parasites cultured in FthR26△/△ or Fth.fl/fl blood cells (N = 80 parasites from three independent experiments, one-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001)

Ferritin, an iron storage protein, is the primary iron storage mechanism critical to iron homeostasis [23, 24]. We reasoned that the removal of ferritin would release the iron and provide an extra iron source to the cell. However, germline ferritin (Fth1) deletion results in embryonic death in mice [25]. We, therefore, tried an inducible genetic manipulation to delete Fth1 to increase intracellular iron levels. Consistent with a previous study [26], induced Fth1 deletion by a ubiquitously expressed Rosa26-Cre (FthR26△/△) resulted in rapid loss of body weight and subsequent death (Fig. 4c, d). This was not observed in the control Fthfl/fl mice, which received the same inducer tamoxifen under the same schedule. Loss of ferritin repressed Dmt1 and Tfr1 mRNA expression of cells in the blood (Fig. 4e), consistent with the expectation of increased intracellular labile iron pool.

To test whether ferritin loss could affect the growth of our iron-deficient malaria parasites, we inoculated clone C2 and control parasite on the eighth day after tamoxifen induction and then determined parasitemia using blood smears from the second day after parasite inoculation. Interestingly, after ferritin was knocked out, the control parasite exhibited growth defects. On the fourth day, the parasitemia of the control parasite reached 5% in the control Fthfl/fl mice, whereas the parasitemia in the FthR26△/△ mice was less than 1%. In comparison, the parasitemias of clone C2 inoculated into the two kinds of mice were similar, but both less than 1% (Fig. 4f, g), comparable to that of the control normal parasite. It is possible that some harmful host effects after ferritin knockout were produced, which suppressed the growth of the control parasites. Why were then the mutant parasites not affected? We speculate that, on the one hand, this harmful effect may also apply to the mutant parasite, but on the other hand, the elevated free iron in the erythrocytes after ferritin knockout may have benefited the mutant growth. These two opposite actions neutralized each other for the mutant, making the difference between the control and Pydmt1 mutant parasites no longer evident.

To put this speculative thinking to the test, we cultured both the parasites with ferritin knockout erythrocytes and analyzed whether ferritin knockout erythrocytes could rescue the growth defect of Pydmt1 mutant. Delightfully, the C2 clone growth was completely normalized after culturing with the ferritin knockout erythrocytes. Additionally, when grown with ferritin mutant erythrocytes, the control parasite’s growth deficiency in FthR26△/△ likewise disappeared, demonstrating that the previously noted additional suppressive effect in vivo may indeed originate from the host independent of the erythrocytes (Fig. 4h).

Ferritin deletion elevated the intracellular free iron level of reticulocytes

To confirm that after ferritin deletion the free iron was indeed elevated in parasite-infected erythrocytes, we first analyzed the LIP in the mice plasma and erythrocytes after inducible ferritin knockout. To our surprise, the plasma iron was significantly reduced (Fig. 5a), while the iron in erythrocytes did not change significantly (Fig. 5b). How could then the Pydmt1 mutant be rescued?

Fig. 5

Ferritin deletion led to iron elevation in the reticulocyte and rescue of PyDMT1 hypomorph. a Plasma free iron of Fthfl/fl (N = 3) and FthR26△/△ mice (N = 3, t-test, *P < 0.05). b Free iron in the red blood cell of Fthfl/fl (N = 4) and FthR26△/△ (N = 4). c Ferritin in the red blood cell of Fthfl/fl (N = 5) and FthR26△/△ (N = 5). d Gating strategy applied to separate parasites infected reticulocytes and infected red blood cell. APC indicates TER-119-APC in X-axis, PE represents CD71-PE in Y-axis. Hoechst33342 (PB450A) to select parasites positive cell when parasitemia is up to 1%. TER-119-APC was used to label erythrocyte; CD71-PE was used to separate infected reticulocytes (iRET) and infected red blood cell (iRBC) with 1–2µL blood. e Cell count of iRET and iRBC (N = 4, t-test, **P < 0.01, *P < 0.05). f Gating strategy applied to select for reticulocytes by flow cytometry. TER-119-APC to separate erythrocyte and non-erythrocyte with 1µL blood, CD71-PE to select reticulocytes in TER-119-APC positive cell. g The LIP of reticulocytes from Fthfl/fl (N = 5) and FthR26△/△ mice (N = 5, t-test, ***P < 0.001). h A proposed model of PyDMT1 action. PyDMT1 reduction leads to iron deficiency, decreasing the number of merozoites in each schizont. Ferritin knockout elevates the intracellular iron of reticulocytes, conferring a rescue on the PyDMT1 hypomorph parasite

After synthesizing erythroid hemoglobin, reticulocytes are formed following enucleation and remain in the bone marrow for 2–3 days before being released into the peripheral blood and then transformed into mature erythrocytes in 1–2 days. In humans, these mature red blood cells lived for about 120 days [27,28,29], and in mice, 38–52 days [30]. Since there is no genome in mature red blood cells, we posit that tamoxifen cannot exert any knockout effect in these cells. On the other side, mature red blood cells constitute a large percentage in the population. Therefore, as a pool, no change of free iron and ferritin in whole red blood cells may be observed only a few days after ferritin removal (Fig. 5b,c), which may only happen in the precursors of erythrocytes before their enucleation process. Moreover, 5 days are sufficient for the reticulocyte development, and we observed, as predicted, that Plasmodium yoelii mainly infected reticulocytes rather than the mature red blood cells (Fig. 5d, e). In other words, the iron level in the reticulocytes, which Plasmodium yoelii infection is susceptible to, can be increased after ferritin removal, enabling a rescue of the growth defect of Pydmt1 mutant parasites. To test this hypothesis, we have to separate the reticulocyte minority from the mature red blood cell majority. We used an antibody to label reticulocytes to divide the blood cells into two populations (Fig. 5f). As expected, the free iron content in the reticulocytes indeed increased significantly after ferritin loss (Fig. 5g, h). This indicates a delicate balance of iron homeostasis exists between the host and pathogen.

Pydmt1 mutant malaria parasites were less sensitive to artemisinin

Artemisinins are first-line antimalarial drugs. Though the precise method by which artemisinins fight malaria is still up for debate, it is widely acknowledged that the endoperoxidic bond must first be activated for the substance to have effect. Several kinds of activation mechanisms have been proposed, but essentially none contradict the involvement of iron in the process [31]. Since Pydmt1 mutations led to serious reduction of iron in the parasites, it would be interesting to test whether these parasites are more sensitive or resistant to artemisinins. Our Pydmt1 mutant parasites were indeed less sensitive to the action of artemisinins, both in vitro (Fig. 6a, b) and in vivo (Fig. 6c, d). Similarly, this also occurred to artesunate (Fig. 6c, d). As a negative control, the Pydmt1 mutation posed no alterations in susceptibility to another common antimalarial drug, chloroquine (Fig. 6c, d).

Fig. 6

PyDMT1 hypomorph parasites were less sensitive to artemisinin. a Merozoite numbers of Pydmt1-C2 treated with 100 nM artemisinin (N = 80 parasites from three independent experiments, one-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001). b Merozoite numbers of Pydmt1-A1 treated with 50 nM artesunate (N = 80 parasites from three independent experiments, one-way ANOVA and Tukey’s multiple comparison test. ****P < 0.0001). c Parasitemia of BALB/c mice treated with 7.5 mg/kg artemisinin, 8 mg/kg artesunate, or 6 mg/kg chloroquine after infected with PyDMT1-HA and Pydmt1-A1 parasites. Drugs were administered when parasitemia was up to 10%. This assay is limited to just 2-day assays due to obvious immune reaction to P. yoelii 17X 5 days after infection. d Inhibition rate of BALB/c mice treated with 7.5 mg/kg artemisinin, 8 mg/kg artesunate, or 6 mg/kg chloroquine after infection with PyDMT1-HA and Pydmt1-A1 parasites (N = 5, t-test, ***P < 0.001, **P < 0.05)