Introduction Malaria remains one of the most widespread infectious diseases of our time. The latest estimates reveal that ∼250 million people are infected with malaria across the globe, of whom ∼800,000 die every year [1], the vast majority being young children. In 2007, the malaria eradication agenda was adopted by many researchers in the antimalarial community and target product profiles for new antimalarial medicines were defined [2]. Most available antimalarials were designed to target the pathogenic blood stages in humans and to address the constant threat of drug resistance [3]. However, to meet the objective of malaria eradication, medicines that block parasite transmission [4] and eliminate the asymptomatic and sometimes cryptic hepatic forms also need to be developed. The bottleneck populations of liver and sexual stage parasites [5] represent potential pathogen vulnerabilities that could be targeted by small molecules; the first such bottleneck is at the liver stage. Within minutes of being released by the bite of an infected female Anopheles mosquito, Plasmodium sporozoites reach the mammalian liver, where they invade hepatocytes and either lie dormant or develop over several days, eventually forming the schizonts that are the prelude to a blood stage infection [6]. Molecules that efficiently target the parasite stages in the liver would offer protection from infection and could theoretically eliminate the cryptic hypnozoite (dormant parasite) infection reservoirs formed by P. vivax and P. ovale. Because only 100 or so sporozoites may be introduced by a bite, there are likely to be many orders of magnitude fewer parasites at this stage than in an active blood stage infection, reducing the possibility of resistance arising. A second bottleneck occurs during sexual development. At each round of schizogony ∼1% of merozoites differentiate into gametocytes [7], and it is these developmentally arrested cells that are transmitted to the mosquito. Mature gametocytes are sexually dimorphic, forming microgametocytes and macrogametocytes that escape the red blood cell (RBC) and produce male and female gametes in the blood meal of the mosquito by processes known as exflagellation [8] and activation, respectively. Following fertilization the zygote differentiates into a motile and invasive ookinete within which the briefly diploid genome undergoes meiosis. These processes occur within an environment almost totally derived from host blood, which can therefore provide a novel and ideal conduit for the delivery of drugs to inhibit parasite transmission to the mosquito. Having crossed the mosquito midgut wall [9], the very few surviving ookinetes differentiate into oocysts, which undergo endomitosis, eventually producing thousands of daughter sporozoites. The sporozoites migrate from the midgut of the mosquito to its salivary glands where the lifecycle begins again. Given that it would be highly desirable for candidate drugs to have activity against hepatic and sexual forms of the malarial parasite, it is surprising that few clinical trials, to date, have examined whether gametocyte carriage can be reduced following drug treatment. The only drugs found to be effective at reducing gametocyte carriage include artemisinin [10], artemisinin combination therapies (ACTs) [11],[12], methylene blue [13], and primaquine [12],[14]. Additionally, few studies have investigated the impact of drugs on the transmission of parasites from human blood to the mosquito vector [15]–[18], nor have many been designed to evaluate antihepatic stage activity. In the context of malaria eradication these gaps in our understanding of the full potential of the drug armoury are problematic. Here we report the development of a series of novel assays against liver, sexual blood, and mosquito stages of the malaria parasite, using both drug-susceptible and drug-resistant parasite strains. We applied these assays to the current portfolio of schizonticidal compounds, consisting of 50 anti-infectives currently in use or under development.. Methods Ethics Statement All work involving laboratory animals for the host-to-mosquito transmission studies was performed in accordance with the European Union (EU) regulations “EU Directive 86/609/EEC” and within the regulations of the United Kingdom Animals (Scientific Procedures) Act 1986, sanctioned by UK Home Office Licence PLL70/6347 awarded in January 2008. Protocol design and implementation was guided by the principle of the three Rs (reduction, refinement, and replacement) and are of mild-to-moderate severity. Protocols are regularly reviewed and revised following approval by the Imperial College Ethics Review Committee. Parasite Maintenance P. berghei parasites constitutively expressing GFP (PbGFPcon) [19]–[21] were routinely maintained as described previously [22]. Only blood showing exflagellating parasites was used in the transmission assays. P. falciparum NF54 strain parasites were maintained in culture as described previously [23]. Gametocyte cultures were produced as described [24]. P. falciparum In Vitro Antimalarial Activity In vitro antimalarial activity was measured using the [3H]-hypoxanthine incorporation assay [25] with various strains of P. falciparum obtained from MR4. Results were expressed as the concentration resulting in 50% inhibition (IC50). P. falciparum Exflagellation Assay Compounds were added to mature gametocyte of drug-sensitive P. falciparum cultures that showed the ability to exflagellate. After 24 h, exflagellation was triggered by a temperature decrease to ∼21°C and observed 20 min later under the microscope. Highly motile exflagellation centres were recorded for ∼50–150 adjacent fields of view and reported per 10,000 RBCs. P. berghei Ookinete Development Assay – Slide Method Compounds and P. berghei gametocyte-infected blood were dispensed in a 96-well plate containing ookinete medium [26]. After 24 h at 19°C, Giemsa-stained ookinetes were counted under the microscope (×40) as previously described [27],[28]. Standard Membrane Feed Assay – P. berghei Membrane feeds were performed as described previously [22]. Briefly, PbGFPcon-infected mouse blood was mixed with compounds, immediately placed into membrane feeders (39°C) and offered for 30 min to 80–100 overnight-starved A. stephensi (SDA 500 strain). After 7–9 d at 19°C/80%RH, mosquito midguts were dissected out. Midguts were fixed with paraformaldehyde and oocyst number was determined microscopically by semi-automated analysis as previously described [29]. Standard Membrane Feed Assay – P. falciparum Gametocyte cultures were produced according to the same protocol as the Pf exflagellation assay. Pooled gametocyte cultures [30] were evenly divided between compounds in fresh medium and incubated at 37°C for 24 h. The parasite pellets were then resuspended in fresh RBCs and human serum treated with the compounds to a 50% haematocrit and immediately fed to mosquitoes as described above and then maintained at 27°C/60%RH for 10–12 d before dissection and counting. Liver Stage Assay 7.5×103 HepG2 cells in 50 µl of medium (1.5×105 cells/ml) were seeded in 384 well plates (Aurora 384 IQ-EB Black/Clear Plates) 20–26 h prior to the actual infection [31]. 2 h prior to infection, 50 nl of compound in DMSO (0.1% final DMSO concentration per well) were transferred with a PinTool (GNF Systems) into the assay plates (10 µM final concentration). P. yoelii sporozoites were freshly dissected from infected A. stephensi mosquito salivary glands and filtered twice with a 40-µm strainer. The HepG2 cells were then infected with 8×103 sporozoites per well. After infection and 1-h incubation at 37°C, the cultures were washed, new media and compound added, and further incubated with 5-fold increased concentration of penicillin/streptomycin for 48 h at 37°C before exoerythrocytic forms (EEFs) quantification of infected cells by immunofluorescence. EEF immunofluorescence quantification After washing with 1×PBS and fixation of the cells with 4% paraformaldehyde solution (EMS), membranes were permeabilised with 0.5% Triton-X-100 (Thermo Fisher Scientific) and EEFs were stained using a mouse polyclonal serum raised against the Plasmodium heat shock protein 70 (HSP70), an Alexa goat antimouse IgG, Fca-specific DyLight 649 secondary antibody (Invitrogen), and the Hoechst 33342 nucleic acid dye (Invitrogen). Stained EEFs were then quantified using the Opera Confocal High Content Screening system (PerkinElmer). Images were collected using a 20× magnification at a binning of 2 using a 365 Xeon arc lamp illumination to detect the nuclei and 635-nm laser line to excite DyLight649. High-content imaging of infected HepG2-CD81 cells was performed as described in Meister et al. [31]. Wells were analysed using a custom Acapella (PerkinElmer) script parameterized for this assay. In brief, images from fields inside the well were first discarded as out of focus when the intensity in the nuclear area was too low. Then hepatic cells were counted by detecting the nuclei labelled with Hoechst 33342 and parasites were segmented using the HSP70 immuno-labelling. Morphology-based (e.g., area, roundness) and intensity-based features were calculated for each object detected including the hepatocyte nuclei and the parasites. Parasitemia was set as the ratio between parasite number (Alexa fluor positive) and the hepatocyte nuclei count, determined at the same time. Results Schizonticidal Activities of Compounds against Strains of P. falciparum with Known Drug Resistance Markers A collection containing all antimalarials approved for use in humans and those in clinical development, anti-infectives, and other controls (Figure 1) was profiled simultaneously on asexual blood stage parasites in a standardized growth inhibition assay (GIA) [25] using seven strains of P. falciparum exhibiting diversity in the molecular causes of resistance and geographical origins (Table S1). The half-maximal inhibitory concentrations (IC50) were determined for each molecule (Table S2). Known mutations in pfcrt, pfmdr1, pfdhfr, and pfdhps correlated with a loss of potency of at least 10-fold for the relevant drugs. As expected the 4-aminoquinolines (4-AQs) chloroquine and hydroxychloroquine showed a 6–100-fold reduction in potency against all drug-resistant strains containing the mutated chloroquine transporter (PfCRT) (Figure 2B). Amodiaquine, tert-butyl-isoquine pyronaridine, piperaquine, and naphthoquine were potent against all parasite lines (IC50 = 2–10 nM), as were the endoperoxides including the natural (artemisinin), the semi-synthetic (artesunate), or the fully synthetic peroxides (ozonides), all with IC50 values of 1–15 nM (Figure 2A). 10.1371/journal.pmed.1001169.g001 Figure 1 The main classes of antimalarials. The chemical structures of all the main classes of antimalarials and other therapeutic and control molecules are assembled according to either the chemical classes they belong to (endoperoxides, 4- and 8- AQs, amino-alcohols) or their function (antifolate, antibiotics), or both (e.g., sulfonamides, a chemical class of antibiotic used in combined antimalarial therapies). The colour code associated with each class is consistent in all the figures in this report. 10.1371/journal.pmed.1001169.g002 Figure 2 The potencies of selected antimalarials against asexual blood stages. Main classes of antimalarials were tested against the blood stage of seven Plasmodium falciparum strains in the [3H]hypoxanthine incorporation assay as described in Vennerstrom et al. [25]. The strains tested here (obtained from MR4) were two drug-sensitive strains NF54 and D6 and five drug-resistant strains: K1, resistant to chloroquine (CQ) and pyrimethamine (PYR), origin Thailand, carries mutations in genes pfmdr1, pfcrt, pfdhfr, pfdhps; W2, resistant to CQ, PYR, quinine, cycloguanil, and sulfadoxine, origin Vietnam; 7G8, resistant to CQ and PYR, origin Brazil; TM90C2A, resistant to CQ, PYR, and MFQ, origin Thailand; and V1/S, resistant to CQ, PYR and cycloguanil, origin Vietnam. Results are expressed as the concentration resulting in 50% growth inhibition (IC50).Values are means of ≥3 independent experiments. (A) Endoperoxides. (B) 4-AQs. (C) 8-AQs and amino alcohols. (D) Antifolates, antibiotics, and others. Compared to the 4-AQs, the 8-AQs (8-AQs) primaquine, diethylprimaquine, and bulaquine were less potent (IC50 = 0.5–2.5 µM) (Figure 2C) against both drug-resistant and -sensitive strains. Tafenoquine and NPC-1161B exhibited IC50 values in the 500-nM range against NF54 and in the 50-nM range against 7G8. IC50 values of both racemic mefloquine and the +RS isomer were below 10 nM. Halofantrine and lumefantrine displayed potencies below 4 nM against the sensitive strains NF54 and D6 and in the case of halofantrine, 0.3 nM against the multi–drug-resistant strains K1, W2, and 7G8. Resistance to pyrimethamine was verified in all drug-resistant strains (Figure 2D). Atovaquone a drug active in the subnanomolar range against P. falciparum blood stages showed at most a 10-fold difference in potency between the strains D6 and V1/S. In contrast, a ≥100-fold loss of potency was observed for cycloguanil between sensitive and resistant strains. The antibiotics azithromycin and trimethoprim, protein synthesis inhibitors, and other molecules such as dehydroepiandrosterone (DHEA), riboflavin, doxycyclin, and the prodrug proguanil showed IC50 values in the 1–10-µM range with no major differences between strains. The Identification of Drugs That Additionally Block Transmission from Human Host to Mosquito Vector Eradicating malaria will require medicines that prevent transmission of the parasite between humans and mosquitoes. Potentially the severe population bottleneck experienced as the parasite progresses from the mature gametocyte in the human host through gametogenesis and fertilization in the mosquito blood meal to the oocyst in the mosquito haemocoele offers the most vulnerable target for intervention. We developed assays for each of these events (Figure 3A). To integrate both the early sexual stages (gametocyte maturation and gametogenesis) and the late vector stage (sporogony) into the drug-testing cascade, we measured the exflagellation of male gametes in vitro (P. falciparum), ookinete formation in vitro (P.berghei), and the production of oocysts in A. stephensi (P. berghei and P. falciparum). Of these assays, we found analysis of P. berghei ookinete production in vitro was the most robust approach to identify molecules potentially targeting the early development of Plasmodium parasites in the mosquito. Forty-six molecules were tested at a concentration of 10 µM (Figure 3B). The most potent molecules were cycloheximide (blood stage IC50 of 25 nM) and atovaquone (IC50 = 65 nM). Thiostrepton (IC50 = 8 µM) and pyronaridine (IC50 = 6 µM) were less potent (Figure 3C). The latter two molecules and pyrimethamine also inhibited P. falciparum exflagellation by more than 80%, as did sulfamethoxazole and mefloquine (+RS). While displaying insignificant activity in the P. berghei ookinete formation assay (Figure 3C), all endoperoxides, with the sole exception of artemether, inhibited P. falciparum exflagellation by >65% (Figure 3D). Similarly all 4-AQs inhibited this event by >60%, with the exception of hydroxychloroquine and chloroquine, which enhanced exflagellation by at least 20% [16]. 10.1371/journal.pmed.1001169.g003 Figure 3 The transmission-blocking potential of selected antimalarials in three bioassays that cover different phases of Plasmodium vector stage development. (A) Assays examined exflagellation (P. falciparum), ookinete formation (P. berghei), or oocyst formation (P. falciparum). All antimalarials were screened at 10 µM. (B) The biological content of the in vitro P. berghei ookinete assay spans gamete formation, fertilization, zygote development, and ookinete formation. Ookinete formation was insensitive to most of the antimalarials tested. Atovaquone, cycloheximide, pyronaridine, pyrimethamine, and thiostrepton all strongly inhibited ookinete formation (p 50% inhibition of exflagellation (p 90% inhibition when tested at 10 µM (Figure S1). The transmission-blocking potential of molecules that were active against exflagellation and/or P. berghei sporogony was then assessed against the production of P. falciparum oocysts, which is the most difficult and lowest throughput, yet highest content analysis by encompassing all vector stages from gametocyte uptake to sporogony (see Figure 3A). In this assay, most endoperoxides inhibited oocyst production by >75%; NPC-1161B and mefloquine (+RS) totally blocked transmission at this stage (Figure 3E). Strikingly, the Coartem component, lumefantrine, and halofantrine impaired sporogony in both P. berghei and P. falciparum, while both mediated only moderate or no inhibition of exflagellation. This finding suggests that these molecules might act specifically on oocysts and not on gametogenesis, a behaviour that could be relevant to transmission reduction in the field given the long half-life of these molecules. As exflagellation is a component process of development within the gut of the mosquito in the P. falciparum oocyst assay (see Figure S2), these results, not unexpectedly, show at least partial concordance with the P. falciparum exflagellation assay. The Identification of Drugs That Suppress Transmission from the Mosquito to the Human Host When an infected mosquito bites a host, ∼100 sporozoites may be injected into the dermis from where they rapidly invade liver cells [32],[33]. This infective step represents the second bottleneck during transmission and therefore another potentially vulnerable point for intervention. In the absence of a practical liver stage assay measuring the formation of P. falciparum/P. vivax liver schizonts, an equivalent assay was developed in P. yoelii and used to assess the activity of the collection of molecules against this specific stage. Specifically, P. yoelii sporozoites were dissected from the salivary glands of infected mosquitoes and were allowed to invade human hepatocarcinoma cells expressing the CD81 protein. The development of the liver schizonts was monitored by immunofluorescence staining using an HSP70 antibody specific to the parasite (Figure 4A). As only 1% of the hepatocytes are infected in these circumstances, high content imaging was used to quantify growth inhibition of parasite schizonts (Figure 4B, 4D). Quantification of the total immunofluorescence per well is shown in Figure 4C. Dose response analysis using serially diluted compounds showed that cycloguanil, pyrimethamine, P218, and atovaquone all displayed IC50 values below 10 nM (Figure 4B, 4D). Methylene blue and artemisone demonstrated IC50 values of 6-d) gametocytes could potentially clear the host of all parasites; i.e., a long half-life would not be required [56]. A sustained stability would however be required for drugs targeting the parasite exclusively in the mosquito. Our study suggests that molecules such as atovaquone that inhibit the electron transport chain in the parasite mitochondria could be suitable candidates, but ideally should lack delayed onset of action and be difficult to raise resistance against [57]. To avoid triple therapies and reduce the risk of drug resistance, dual-activity molecules like amodiaquine, which inhibits haemoglobin digestion in the asexual blood stages and potentially inhibits gametocyte maturation/gamete exflagellation by a different mechanism, could be used in combination with a second antimalarial. Such polyvalent multistage activity has significant benefit to overall drug impact. Our study highlights that molecules such as amodiaquine, naphthoquine, tert-butyl-isoquine, and piperaquine do not lose potency when tested against chloroquine resistant strains. Therefore, to mitigate or defer the risk of drug resistance these molecules might be proposed as potential candidates for partnering new antimalarials such as OZ439. An important consideration would be to favour molecules that have never been used as monotherapy to avoid facing parasites that have previously acquired drug resistance. NPC-1161B inhibited both exflagellation and oocyst production; new molecules with similar properties but devoid of haemolytic liability of the 8-AQs in glucose-6-phosphate dehydrogenase (G6PD)–deficient patients could be interesting candidates as specific transmission-blocking agents. While being strong inhibitors of blood-stage parasites, some molecules such as chloroquine reportedly enhance gametocytogenesis—a property that might have facilitated the spread of drug-resistant parasites [15],[58]. The panel of assays that we applied in this study has confirmed such “collateral” activity, in this case one that could prejudice the rational implementation in any elimination/eradication strategy. Other Human Malarias Although many have argued for the use of drug screens using nonhuman malarias [47], we recognize that to some, including assays with rodent malaria species might be considered suboptimal. Recent observations identifying interspecific variations include cysteine proteases in rodent plasmodia that show subtle active site differences to those in P. falciparum, leading to questions about the use of these models [59]. A critical role of amino acid 23 mediates activity and specificity of vinckepain-2, a papain-family cysteine protease of rodent malaria parasites [60]. P. yoelii 17X strain is intrinsically partially resistant to chloroquine and is therefore a poor model for studying acquisition of P. falciparum chloroquine resistance. Nevertheless we must recall that there are five species of Plasmodium that infect man and their biologies are patently different, therefore detecting drugs that may be active against multiple species in initial screens may offer long-term potential. Our assays provide, to our knowledge, the most comprehensive global overview of antimalarial drug action to date within the constraints imposed by the current state of culture methodologies for all life stages of all mammalian malaria parasites. Ideally, antimalarials developed against P. falciparum would have an even broader clinical usefulness if proven to be as effective against P. vivax [61]. The potencies of some antimalarials against the asexual blood stage of cultivated P. falciparum and P. vivax field isolates show a very good correlation (Table S4) [62]–[67]. This observation suggests that most of the pathways inhibited by antimalarials in P. falciparum are conserved and may offer valid targets in P. vivax. Moreover, the endoperoxide OZ439, which is currently evaluated in phase IIa clinical trials, has recently demonstrated equivalent efficacy in the treatment of P. falciparum and P. vivax patients (personal communication, MMV). Drug Gaps and Future Steps Our work has revealed previously unforeseen opportunities in the current discovery and development pipeline for new antimalarials. We demonstrated that drugs in the current portfolio, like pyronaridine and atovaquone, can also target liver and sexual stages in addition to asexual blood stages. Safe and stable drugs with similar multistage potential are now required. Developing drugs with long half-lives like mefloquine and chloroquine is essential to ensure that in patient blood the exposure of these drugs will remain above the minimum inhibitory concentration for several erythrocytic cycles and should ideally cover the period of gametocyte maturation. Additionally, new chemical scaffolds (e.g., nonendoperoxide) with fast killing potential are needed for a “one dose cure.” The search for new drugs would be enhanced by the continued development of P. falciparum and P. vivax culture systems for every parasite life stage. It is critical to the malaria eradication agenda that these assays are able to identify drugs, such as the 8-AQs, with the capacity to safely eliminate P. vivax hypnozoites from the liver: this objective will also require the implementation and validation of in vitro and in vivo G6PD deficiency-dependent hemolysis assays. In parallel with efforts to discover innovative drugs for radical cure (elimination of P. vivax hypnozoites from the liver), new molecules blocking the onward development of mature (stage V) gametocytes are the other major priority in antimalarial discovery. There is therefore an urgent need to develop and validate high throughput screening assays allowing new libraries to be tested against P. falciparum and P. vivax gametocytes/transmission. These assays could then prioritise compounds for examination in preclinical studies in small mammals and then in standard membrane feeding assays (SMFA) using patient blood to find drugs blocking transmission in the field/clinical situation. Fields studies will remain essential to carefully examine any correlation between activities of new molecules against stabilized laboratory parasite strains and against field isolates. Conclusions For the first time, the main chemical classes of current and future antimalarials have been profiled simultaneously in standardized conditions against three Plasmodium species with respect to every major cellular strategy of the malarial life cycle, e.g., vegetative replication, dispersal, and sex (Figure 5). The present study provides the antimalarial research community with a reference set of methods and data, which may serve as a benchmark for newly discovered molecules when profiled against the entire life cycle of Plasmodium. This information might guide decisions regarding which molecules could be optimally combined to provide the next generation of drugs that will succeed to artemisinin combination therapies (ACTs) [68] and support the eradication of malaria. This comprehensive approach to drug discovery has potential utility for targeting other pathogens with complex life cycles. 10.1371/journal.pmed.1001169.g005 Figure 5 Summary of the activity of the most widely used antimalarials throughout the life cycle of Plasmodium. The three main phases, i.e., liver stage, blood stage, and vector stage, of the life cycle of Plasmodium are shown. The two key entry points leading to transmission of the parasites from vector to host and from host to vector are indicated (green circles). Parasite forms specific to each stage are highlighted and drugs identified as inhibitors of development of these forms are listed in boxes and coloured as described in Figure 1. Stars highlight components of the main artemisinin combination therapies: green, coartem; red, pyramax; orange, eurartesim; blue, ASAQ. Supporting Information Figure S1 The transmission-blocking potential of selected compounds against P. berghei in standard membrane feeding assays. All antimalarials were screened at 10 µM in triplicate in independent experiments. The biological content of this assay spans gamete formation through to occyst development all within the gut of the mosquito. NPC1161-B, lumefantrine, pyrimethamine, and cycloheximide showed most notable transmission-blocking activity. (TIF) Click here for additional data file. Figure S2 Comparative summary of the transmission-blocking potential of selected compounds across vector-stage assays. All antimalarials were screened at 10 µM. By comparing the activities of compounds in assays covering different biological ranges of transmission-stage biology, it is possible to infer the stages at which antimalarial drugs are exerting their effects. Lumefantrine was found to have little activity against exflagellation and ookinete development but showed activity in oocyst assays of P. berghei and P. falciparum. Endoperoxides showed activity as early as exflagellation but not during ookinete development, indicative of action against the mature gametocyte/exflagellation. NPC1161-B showed potency in all assays except the ookinete assay, suggesting that it may have dual actions both in early vector-stage development and later on. (TIF) Click here for additional data file. Table S1 The origin and reported drug resistance of P. falciparum strains used in this study. CQ, chloroquine; PYR, pyrimethamine; CYC, cycloganil; QUI, quinine; SUL, sulfadoxine; MFQ, mefloquine; ATO, atovaquone. (PPT) Click here for additional data file. Table S2 The numerical potencies of selected antimalarials against asexual blood stages. Data corresponding to Figure 2, showing the numerical IC50 values of selected antimalarial compounds against seven P. falciparum strains in the [3H]hypoxanthine incorporation assay. (TIF) Click here for additional data file. Table S3 A comparison of antimalarial exposure in human blood against predicted early vector-stage potency. The IC50 values of selected antimalarials in the P. falciparum exflagellation assay was estimated using additional data generated by screening at 1 µM (*) and compared to drug Cmax values obtained from the literature. The in vitro assay contains both blood and serum. Most of the cells in culture (∼97%) are uninfected RBCs. The medium that the assay is set up in is derived from RPMI and contains 10% human serum. (TIF) Click here for additional data file. Table S4 A comparison of the reported asexual blood stage potencies of selected antimalarials against P. falciparum (field and laboratory isolates) and P. vivax (field isolates). (TIF) Click here for additional data file.
