Transmission of Artemisinin-Resistant Malaria Parasites to Mosquitoes under Antimalarial Drug Pressure

P. falciparum clinical isolates that successfully adapted to long-term culture (Chotivanich, unpublished data) were derived from a previous, multicenter, open-label, randomized trial collecting samples from patients with acute, uncomplicated malaria (16). All isolates were sequenced and assessed for multiplicity of infection (MOI) at point-of-care and determined to be sufficiently homozygous to make cloning (and potentially loss of ability to form gametocytes) unnecessary (23). Among isolates, five were followed further based on their ability to form functional mature gametocytes in vitro. These were compared to the standard laboratory control parasite strain NF54. Each was validated by PCR, confirming the five clinical isolates have variant polymorphisms in PfK13 (Table 1). Each P. falciparum isolate was tested ex vivo for sensitivity to the artemisinin derivative artesunate using the 24-h trophozoite maturation inhibition assay (TMI). The TMI assesses the effect of artemisinin derivatives on the maturation of asexual ring stages to trophozoites in comparison to an artemisinin-sensitive control (24). While PfK13^var isolates presented a wide range of TMI 50% inhibitory concentration (IC₅₀) values, they all showed increased resistance to artesunate compared to a PfK13^WT culture-adapted Thai and non-gametocyte-producing laboratory strain, TM267 (Table 1).

In addition to the PfK13 genotype, the genetic background of each parasite isolate was investigated to explore whether additional mutations might be present, such as those associated with other drug resistance phenotypes. Whole-genome sequencing analysis was completed for each, confirming different PfK13 genotypes (Table 2). In addition, multiple previously reported mutations in genes associated with various drug sensitivities were found among PfK13^var isolates (as reviewed in reference 2) (Table 2). Mutations were found in the chloroquine resistance transporter (pfcrt), agreeing with reported mutations found in some parasites following ACT treatment (25). None of the isolates, however, carried mutations in pfcrt associated with increased DHA-piperaquine treatment failure (7, 8, 26) or piperaquine resistance in vitro (27, 28). In addition, all isolates carried mutations in four genes that are associated with the PfK13^C580Y haplotype, namely, fd (ferredoxin), arps10 (apicoplast ribosomal protein S10), mdr2 (multidrug resistance protein 2), and crt (chloroquine resistance transporter) (Table 2) (17).

Increased copy numbers for the multidrug resistance transporter pfmdr1 and enzymes plasmepsin II/plasmepsin III are known to be associated with enhanced survival of parasites exposed to mefloquine or piperaquine, respectively (29–32). To test if any of the field isolates harbored copy number variations, we created sashimi plots of the next-generation sequencing coverage (Fig. S1 in the supplemental material) and found that plasmepsin II/plasmepsin III are duplicated for isolate APL4G (Table 2). APS3G and APL5G both had copy number variants of pfmdr1 (Table 2). This suggests that APS3G and APL5G will also likely show resistance to mefloquine, while APL4G might also show resistance to piperaquine (30, 33).

PfK13^var that are associated with artemisinin resistance are known to also show reduced asexual blood-stage growth (34, 35). To validate this in selected isolates, parasites were set up in synchronized ring-stage cultures, at a starting parasitemia of 2%, and followed over the course of 8 days. Parasitemia was analyzed every second day by flow cytometry and cultures rediluted to 2%. NF54 parasites showed a cumulative parasitemia as expected under standard laboratory conditions (Figure 1a). Parasites with a PfK13^var showed a significantly reduced replication rate in in vitro compared to NF54 (Figure 1a), agreeing with previous studies (34, 35). To explore the underlying mechanism of slowed growth, we measured the number of merozoites per schizont, which will directly determine potential growth rates (36). Late synchronized schizonts were blocked from merozoite egress using the protein-kinase G (PKG) inhibitor, compound 2 (37). Thin smears, 12 h later, were then made of each culture and stained with the nuclear stain 4′,6-diamidino-2-phenylindole (DAPI) to count nuclei per schizont. PfK13^var isolates displayed fewer nuclei per schizont than the NF54 control, suggesting that the observed reduced growth rate may at least be partially explained by a reduction in the number of progeny (Figure 1b).

To investigate the transmission capability of each P. falciparum field isolate, we induced gametocytes at a starting parasitemia of 2% (38) and, 14 days post induction, fed cultures to Anopheles stephensi mosquitoes by standard membrane feeding assay (SMFA) (39). No significant differences were noted in the stage V (mature) gametocytaemia between different isolates. NF54 showed higher overall stage V gametocytaemia compared to Pf13^var (Figure S2a). Male gametocyte exflagellation rates (exflagellation/μl of culture) showed some variability between isolates. Although all the isolates were seeded at equal hematocrit and asexual parasitemia percentages, we cannot directly compare those rates (Figure S2a). Rather, they indicate male gametocyte viability of each individual culture before SMFAs (Figure S2b and c). At 10 days postfeeding, mosquito midguts were dissected and oocyst numbers recorded. All field isolates were found to be capable of infecting mosquito midguts at varied intensity levels, i.e., oocyst counts per midgut, as reported previously for Cambodian field isolates (40). To test if PfK13^var led to a reduced replication rate in mosquito stage growth (following the reduced merozoite count), we measured the diameter of each oocyst in these infections as a proxy for replication. Oocyst size showed no consistent pattern of variation compared to NF54, other than a PfK13^R539T isolate, which displayed significantly larger oocysts (Figure S2c). This shows that while variations in PfK13 may reduce parasite multiplication rate in the asexual blood stage, they do not appear to directly influence transmission and growth in the mosquito stage.

It has previously been shown that artemisinin and its derivatives have an inhibitory effect on male gametocyte exflagellation, sterilizing male gametocytes from activation, but have no effect on female gametocyte activation (21, 22). To explore whether PfK13^var isolates were resistant to this sterilizing effect, we applied a modified dual gamete formation assay (PfDGFA) (22) measuring only the male gamete activation, i.e., male gamete formation assay (MGFA) for this purpose. A 24-h incubation of cultures with the artemisinin derivative DHA was found to be insufficient to elicit a complete inhibition of exflagellation for NF54 parasites (Fig. S3), likely as a result of the instability of the drug, which has been shown to have an in vitro half-life of 2.3 h in human plasma at pH 7.4 and 37°C (41). To improve activity and allow for a comparative analysis between isolates, gametocytes were exposed to a second compound dose at 24 h after the first, resulting in a double-dose regimen with a readout after 48 h. Double exposure consistently gave complete inhibition of male activation with DHA at the highest concentration tested (Fig. S3). In parallel, two other artemisinin derivatives and four other antimalarial drugs were tested by the MGFA (Fig. S4 and S5). Exflagellation rates for PfK13^var isolates varied in the presence of drug, with the two PfK13^R539T and PfK13^C580Y isolates consistently showing tolerance to artemisinin derivatives (Fig. 2) but not to other compounds, i.e., UCT048 and NITD609. These two pipeline antimalarials are known to target PfPI4K (42) and PfATP4 (43), respectively. The lack of a consistent pattern of reduced sensitivity to artemisinin-based drugs across PfK13^var isolates suggests that K13 polymorphisms alone likely do not completely explain sensitivity of sexual stages to artemisinin treatment. This observation is corroborated by similar findings with piperaquine resistance, which is mostly, but not always, associated with copy number variants in plasmepsin II/plasmepsin III (30), but has now been associated with specific pfcrt mutations in vitro (27, 28). In addition, the two isolates exhibiting reduced exflagellation sensitivity to artemisinin-based drugs, namely, APS3G and APS5G, are the only two isolates in our study exhibiting amplification of pfmdr1 (Fig. S1). This is especially interesting as a reduction in copy number variations of pfmdr1 were previously found to increase DHA sensitivity in asexual parasites (44, 45). However, the presence of even a single isolate with reduced sensitivity to artemisinin-based drugs in both asexual and sexual stages suggests there is the potential that a resistant strain might be favored for transmission to the mosquito.

To mimic the impact of DHA’s short elimination half-life (46) on its exflagellation effect, NF54 stage III (early) and stage V (mature) gametocytes were incubated with a single DHA concentration, and the drug replaced with half of the concentration every 50 min (range 3.5 μM to 0.027 μM DHA, a total of eight incubations equaling 6.67 h). A 50-min exposure time was selected as a trade-off between the fast elimination half-life of DHA and the minimum time required for gametocytes to settle by gravity under the assay conditions. DHA was then washed out and gametocytes matured until control cultures reached stage V or an additional 48 h, respectively. When mature stage V gametocytes were exposed to rapidly reduced DHA concentrations, exflagellation was still effectively inhibited (54.72%). Exflagellation of early-stage gametocytes was almost entirely inhibited (Figure 3a and c). This is likely explained through a killing effect, as gametocytes disappear from the culture when early stages are exposed to DHA (Figure 3b and d). Interestingly, the sterilizing effect of DHA on stage V gametocytes appeared to be irreversible, as it was sustained even after the drug had been washed out and was independent from an immediately visible reduction in gametocytaemia (Figure 3d).

To explore the hypothesis that PfK13^var-associated resistance might allow resistant parasites to efficiently infect mosquitoes under drug coverage, we selected the PfK13^C580Y isolate APL5G. This isolate showed a comparable level of mosquito infection to NF54 (Fig. S2) and also showed a high level of male gamete activation resistance to DHA (Fig. 2). Gametocyte cultures of both parasites, APL5G and NF54, were exposed to a range of DHA concentrations for 48 h using our double-dosing regimen, before feeding to A. stephensi mosquitoes by SMFA. The DHA concentration range was selected to ensure exflagellation IC_50s for both lines were incorporated in the SMFA three-point dose-response. At day 10 postfeeding, mosquitos were dissected and midguts examined for oocyst load (Fig. S6). Generalized linear mixed effects models were used to analyze infection intensity (number of oocysts per midgut) and infection prevalence (proportion of midguts with oocysts) in response to treatment with DHA, in order to incorporate data from 18 individual SMFA experiments within the same modeling framework. A decrease in both intensity and prevalence of infected mosquitoes was observed for both parasite isolates with increasing DHA concentration (Figure 4a and b). A significant decrease in both the oocyst intensity (ratio of oocyst counts = 0.73, 95% confidence interval [CI]: 0.66 to 0.80) and prevalence (odds ratio = 0.46, 95% CI: 0.41 to 0.52) of mosquito infection was observed for NF54 with increasing drug concentration. In contrast, APL5G parasites (Pf13^C580Y) showed no evidence for a significant decrease in oocyst intensity with increasing DHA (ratio of oocyst counts = 0.84, 95% CI: 0.65 to 1.07). Increasing concentrations of DHA did still reduce the oocyst prevalence for APL5G (odds ratio = 0.72, 95% CI: 0.56 to 0.96); however, this effect was significantly less than the effect seen for WT parasites (Table 3).

To position these findings in the context of likely transmission events, we further explored the impact of DHA on transmission at a single and conservatively selected concentration of DHA (2 μM) (46) in comparison with dimethyl sulfoxide (DMSO). The peak of DHA serum concentration following the recommended WHO ACT oral dosing regimen for artesunate exhibits a wide range with the maximum drug concentration (C_max) up to 7.5 μM (46). Since DHA sterilizes male gametocytes irreversibly, the presence of even a transient spike of 2 μM in serum will, therefore, lead to transmission blocking. In the absence of DHA, NF54 parasites were consistently observed to have a higher infection prevalence (55.1%, 95% CI: 50.5% to 57.8%) compared to APL5G parasites (19.2%, 95% CI: 13.3% to 26.0%) (Figure 4c). This suggests that a fitness cost is associated with the APL5G genotype, which causes a sizeable reduction in the onward probability of infection relative to wild-type (WT) parasites in the absence of DHA in A. stephensi mosquitoes. However, this changed significantly in the presence of DHA. At a concentration of 2 μM DHA, no significant difference was observed between NF54 parasites (20.6%, 95% CI: 27.7% to 14.3%) and APL5G parasites (12.1%, 95% CI: 4.1% to 25.4%) (Figure 4c). The equivalence of infection demonstrates a profound impact of DHA on NF54 but not APL5G transmission, meaning that under clinically relevant artemisinin concentrations, APL5G gametocytes are less affected than NF54 gametocytes for successful onward transmission. This observed transmission-resistance phenotype may offset any fitness costs (such as growth) observed in the absence of drug.

Asexual blood stage and gametocytes were cultured as previously described (38) with the following modifications. Asexual blood-stage cultures were maintained in asexual culture medium (RPMI 1640 with 25 mM HEPES [Life Technologies], 50 μg liter⁻¹ hypoxanthine [Sigma], 5% A+ human serum [Interstate Blood-Bank], and 0.5% AlbuMAX II [Life Technologies]). Gametocyte cultures were maintained in gametocyte culture medium (RPMI 1640 with 25 mM HEPES [Life Technologies], 50 μg liter⁻¹ hypoxanthine [Sigma], 2 g liter⁻¹ sodium bicarbonate [Sigma], 5% A+ human serum [Interstate Blood-Bank], and 0.5% AlbuMAX II [Life Technologies]).

All parasite isolates were sequenced and assessed for multiplicity of infection at point-of-care baseline before ACT treatment. Selected isolates were sufficiently homozygous (based on the value F_WS delineated in reference 23), which allowed us to proceed without additional parasite cloning steps. Interestingly, all copy number variations for APS3G, APL5G (mdr1), and APL4G (plasmepsin II/III) were maintained after prolonged culture cultivation and several cryopreservation steps, as the same copy number variations were found at baseline (52) and during our most recent sequence analysis in 2019.

Anopheles stephensi mosquitoes were reared under standard conditions (26°C to 28°C, 65% to 80% relative humidity, 12 h:12 h light/darkness photoperiod). Adults were maintained on 10% fructose.

Genomic DNA isolation, whole-genome sequencing, and calling of single-nucleotide variants were undertaken essentially as recently described (52). To determine gene amplification copy number variants, sashimi plots were created and visualized using the integrated genomics viewer (IGV) (53) comparing aligned bam files. Sashimi plots were visually inspected for increased read coverage over genes of interest.

To prepare parasites for the growth assay, asexual parasites were sorbitol-synchronized at least twice 16 h apart to create an 8 h growth window. Briefly, cultures containing mainly ring-stage parasites were incubated with 5% sorbitol at 37°C for 5 min, spun down, and resuspended in culture medium. The second synchronization step was repeated 16 h later, resulting in a culture where parasites were between 16 h to 24 h ring stages. To start the growth assay, parasitemia was seeded in 2 ml total volume at 2% hematocrit and 1 to 2% early ring stages in triplicates that were treated separately. The assay was performed twice with 3 replicates each. Every 48 h and for a total of 8 days, parasitemia was determined using flow cytometry and each well was diluted back to 1 to 2% parasitemia as follows. One milliliter of each culture was fixed in 4% formaldehyde and 0.2% glutaraldehyde for at least 10 min. After washing with phosphate-buffered saline (PBS), DNA was stained with SYBR green I (diluted 1:10,000) in the dark for 20 min at room temperature. After incubation, cells were washed three times with PBS and resuspended in 80 μl PBS. Flow cytometry was performed counting a total of 100,000 cells per well. Cumulative growth was calculated based on absolute parasitemia and dilution factor for each day.

Parasites were synchronized twice using 5% sorbitol to obtain a 10-h life cycle window. An aliquot of 10 μM compound 2 was added to late trophozoite stages for a maximum of 12 h to block egress of the RBCs (37). Resulting segmented schizonts were thinly smeared, then fixed with 4% formaldehyde and 0.2% glutaraldehyde for 20 min. Smears were then stained with 1 μg ml⁻¹ DAPI for 5 min and mounted in Vectashield (Vector Laboratories). Z-stacks were taken using a Leica microscope at 100× magnification. Nuclei of arrested segmented schizonts were counted using the plugin tool “Manual counting” on ICY (54). Only singly invaded RBC were counted.

The trophozoite maturation inhibition assay (TMI) was performed as described (24). Briefly, P. falciparum-infected blood was collected into heparin-coated vacutainer tubes and centrifuged at 800 × g at 4°C for 5 min to allow the removal of the plasma and buffy coat. This was followed by three washes in RPMI 1640 (without serum supplement) and adjusted to 3% cell suspension in 10% A+ human serum-supplemented RPMI 1640. Then, 96-well microtiter plates (Nunc MicroWell 96-well microplate; Thermo Fisher Scientific) were predosed with artesunate dissolved in 5% NaHCO₃ (Guilin Pharmaceutical Co., Ltd., China), ranging from 0.01 to 400 ng ml⁻¹ final concentration or no drug as negative control. A 75-μl P. falciparum ring stage infected RBC cell suspension was added to the test plate and incubated for 24 h at 37°C in 5% CO₂. All samples were tested in triplicate. Upon completion of drug exposure, thick and thin blood smears were prepared of all wells and the number of 24- to 30-h trophozoites (55) was counted per 100 infected RBCs. To identify the inhibition activity of artesunate, the percentage of trophozoite maturation compared to the negative control was assessed. The IC₅₀ (50% inhibitory concentration) was calculated as the drug concentration causing 50% inhibition of P. falciparum maturation from ring stage to trophozoite stage and normalized to the negative-control wells. All IC_50s were determined by sigmoid curve fitting using WinNonlin computer software (version 3.1; Pharsight Corporation, USA). As a technical control, all ex vivo assays were performed in parallel to the standard laboratory Thai strain TM267, which is a non-gametocyte-producing line.

The double dose male gamete formation assay (MGFA) was adapted from previously described methods (22) to incorporate an additional drug dosage, accounting for the low compound half-lives of artemisinin and its derivatives. Briefly, compounds were prepared in 10 mM DMSO stocks and dispensed in serial dilutions into multiwell plates using an HP D300 digital dispenser. All drugs were supplied by the Medicines for Malaria Venture (MMV), including UCT048 (MMV048), NITD609 (Cipargamin), gentian violet, methylene blue, and DHA. Samples were normalized to 0.25% DMSO and contained 0.25% DMSO and 12.5 μM gentian violet as negative and positive controls, respectively. Half the maximal DMSO content was plated per plate, accounting for the accumulation of DMSO over two dosages. Mature gametocytes with an exflagellation rate of >0.2% of total cells were diluted in gametocyte culture medium to 25 million RBCs ml⁻¹. Mature gametocyte culture was plated in drugged 96-well plates and incubated in a humidified chamber under 92% N₂/5% CO₂/3% O₂ (BOC special gases) at 37°C for 24 h. For the second drug dosage at 24 h, the drugged culture was transferred to a second drugged well plate and incubated for a further 24 h under 92% N₂/5% CO₂/3% O₂ at 37°C in a humidified chamber.

At 48 h, gametogenesis was induced with ookinete medium (RPMI 1640 with 25 mM HEPES [Life Technologies], 50 μg liter⁻¹ hypoxanthine [Sigma], 2 g liter⁻¹ sodium bicarbonate [Sigma], and 100 μM xanthurenic acid). Plates were immediately incubated at 4°C for 4 min and then 28°C for 5 min before transferring to a Nikon Ti-E widefield microscope. Exflagellation events were recorded by automated phase-contrast microscopy, in 96-well plates. Twenty-frame time lapses were recorded at 10× magnification and 1.5× zoom. Exflagellation events per field were derived using an automated ICY Bioimage Analysis algorithm. Resulting counts were converted to percentage inhibition values, calculated relative to positive (C1) and negative (C2) controls:

Raw data demonstrated a Z-score ≥ 0.4 and was derived from n ≥ 2 and n ≥ 3 technical and biological replicates, respectively. GraphPad Prism (version 8) was used to calculate IC_50s from the dose response data using with the log(inhibitor) versus response–variable slope (four parameters) function. IC_50s were derived from curves demonstrating R² ≥ 0.95.

P. falciparum NF54 gametocyte cultures were seeded according to the MGFA protocol on the same day from the same inoculum and maintained in Nunc EasYFlask cell culture flasks (Thermo Fisher Scientific, Nunclon Delta surface treated, 25 cm²). DHA was kindly provided by Medicines for Malaria Venture and prepared at a 10 mM stock solution in DMSO and stored at −20°C until further use.

On the day of the half-life assay, nonpurified stage III (culture day 9) and stage V (culture day 14) gametocyte cultures were assessed for their gametocytaemia with Giemsa smears and stage V also checked for male gametocyte exflagellation. Each culture was split into two Nunc EasYFlask cell culture flasks (25 cm²), exposed to 92% N₂/5% CO₂/3% O₂, and cells were allowed to settle for 1 h at 37°C. Culture supernatants were removed and replaced with freshly prepared and prewarmed 10 ml gametocyte culture RPMI containing 0.25% DMSO or 3.5 μM DHA (0.25% DMSO final), exposed to 92% N₂/5% CO₂/3% O₂, and cells were allowed to settle at 37°C. After 50 min, culture supernatants were aspirated and replaced with 10 ml of freshly prepared and prewarmed gametocyte culture RPMI containing DMSO in the control culture or 1.75 μM DHA for the treatment culture (half of the initial DHA concentration). This step was repeated until a final exposure of 0.027 μM DHA was reached (8 exposure steps). Culture supernatants were then replaced with 10 ml of prewarmed gametocyte culture RPMI and incubated for 2 h before replacement of supernatants with fresh gametocyte culture RPMI. This washing step was repeated 3 times in total. Gametocyte cultures were then matured until day 15 (stage III) and day 16 (stage V) culture. Exflagellation was assessed according to the MGFA and inhibitions quantified to DMSO controls.

For single-dose DHA wash-out assays, nonpurified stage III, IV, and V (culture days 9, 11, and 14, respectively) gametocytes were each split into two Nunc EasYFlask cell culture flasks (25 cm²) and incubated with 10 ml of freshly prepared and prewarmed gametocyte culture RPMI containing either 0.25% DMSO or 1.6 μM DHA, exposed to 92% N₂/5% CO₂/3% O₂, and kept at 37°C for 24 h. Supernatants were aspirated and all cultures were washed 3 times for 2 h each time, according to the half-life washing steps above. Gametocytes were then further cultured until day 14 (stage III and stage IV) and day 15 (stage V). Exflagellation levels were measured according to the MGFA. Gametocytaemia was counted per 1,000 RBC.

Gametocytes were induced and maintained as described above. At day 14 postinduction, gametocytes were spun down at 38°C and resuspended in 5 ml of suspended animation buffer (SA) (56). To ensure that a consistent number of RBCs and gametocytaemia were used for drug incubation for each isolate, gametocytes were magnetically activated cell sorting (MACS)-purified and resuspended in gametocyte medium with 25 × 10⁶ fresh RBCs. DHA or DMSO was added to the desired end concentration into a 10-ml gametocyte culture and added again 24 h later (double-dosing within 48 h). After 48 h, the parasite culture was mixed with fresh blood and human serum and fed to adult A. stephensi mosquitoes using a 3D-printed feeder (39).

At day 10 postfeeding, mosquitoes were dissected, midguts were stained in 0.1% mercurochrome, and then inspected using light microscopy with 10× magnification to count oocysts.

To measure oocyst size, midguts of A. stephensi fed on P. falciparum-infected blood were dissected and fixed with 4% formaldehyde, permeabilized with 0.1% Triton X-100 for 1 h, blocked with 3% bovine serum albumin (BSA) for 30 min, and stained with 1 μg ml⁻¹ in DAPI for 3 min. Midguts were washed with 1× PBS and mounted in Vectashield. Images were acquired on a Nikon Ti-Eclipse inverted fluorescence microscope. Images of P. falciparum-infected midguts were captured using the DAPI channel and z-stack imaging to obtain greater depth of oocysts. These stacked images were then processed in ND Processing using the Maximum Intensity Projection option, which then created an image with brighter intensity of the oocysts in every midgut. Oocyst detection was automated by using the Automated Spot Detection program based on the intensity of the oocysts compared to midgut cells (NIS-Elements). The size, diameter, and intensity of each selected oocyst were recorded in an MS Excel file for analysis.

To assess the impact of artemisinin on the ability of each parasite line to form oocysts, we used generalized linear mixed effects models to incorporate data from different experimental replicates within the same modeling framework. These models have previously been used to model transmission-blocking interventions (57). We modeled either infection intensity or prevalence as the response with treatment (DHA concentration) included as a fixed effect and 0 μM DHA represented by control groups treated with DMSO. The parasite line that was treated (PfK13^WT or PfK13^C580Y) was included as a fixed effect to assess the differential impact of artemisinin on transmission success. The impact of treatment between experimental replicates was allowed to vary at random between replicates. A logistic regression (binomial error structure) was used to model the prevalence of mosquito infection, i.e., the presence or absence of oocysts, and a zero-inflated negative binomial distribution was used to model the intensity of infections, i.e., the numbers of mosquito oocysts (58). The 95% confidence interval estimates were generated for the impact of drug concentration by bootstrapping methodology (with 100,000 replicates).

Raw experimental data are available on request, while genomic data are publicly available at the European Nucleotide Archive (https://www.ebi.ac.uk/ena). Accession numbers, group ERP121586 and individual: ARN1G (ERS3395814); APS2G (ERS3395811); APS3G (ERS3395812); APL4G (ERS3395813); and APL5G (ERS3395810). The NF54 reference genome is available through https://plasmodb.org/plasmo/.