Open-source discovery of chemical leads for next-generation chemoprotective antimalarials

To discover leads for next-generation chemoprotective antimalarial drugs,we tested more than 500,000 compounds for their ability to inhibit liver-stage development of luciferase-expressing Plasmodium spp. parasites (681 compounds showed a half-maximal inhibitory concentration of less than 1micromolar).Cluster analysis identified potent and previously unreported scaffold families as well as other series previously associated with chemoprophylaxis. Further testing through multiple phenotypic assays that predict stage-specific and multispecies antimalarial activity distinguished compound classes that are likely to provide symptomatic relief by reducing asexual blood-stage parasitemia from those which are likely to only prevent malaria. Target identification by using functional assays, in vitro evolution, or metabolic profiling revealed 58 mitochondrial inhibitors but also many chemotypes possibly with previously unidentified mechanisms of action. INTRODUCTION Malaria remains a devastating disease, affecting 216 million people annually, with 445,000 deaths occurring primarily in children under 5 years old. Malaria treatment relies primarily on drugs that target the diseasecausing asexual blood stages (ABS) of Plasmodium parasites, the organisms responsible for human malaria. Whereas travelers may rely on shortterm daily chemoprotective drugs, those living in endemic regions require long-termmalaria protection such as insecticide-treated nets (ITNs) and vector control. However, ITNs do not fully shield individuals from malaria, may lose potency with time, and can be bulky and difficult to use. Another concern is that mosquitosmay become resistant to the active insecticides that are used in ITNs and vector control. RATIONALE As the possibility of malaria elimination becomesmore tangible, the ideal antimalarial medicine profile should include chemoprotection. Chemoprotectivemedicines typically work against the exoerythrocytic parasite forms that invade and develop in the liver and are responsible for the earliest asymptomatic stage of the infection. Such medicines could be formulated to provide long-acting prophylaxis, safeguarding individuals that are living near or traveling to areas that have been cleared of parasites. Long-acting chemoprotection in endemic regions could also greatly reduce circulating parasite numbersandpotentially replace a vaccine in an elimination campaign. Although millions of compounds have been screened for activity against parasiteABS, and some have been subsequently tested for potential prophylactic activity, large-scale searches that beginwith prophylactic activity have not been performed because of the complexity of the assay: This assay requires the production of infected laboratory-rearedmosquitoes and hand-dissection of the sporozoiteinfected salivary glands frommosquito thoraxes. A Plasmodium vivax liver-stage schizont on a lawn of hepatocytes. The parasite schizont has been stained with antibodies to parasite HSP70 (red) and UIS4 (yellow). Cell (parasite and hepatoma) nuclei are shown in blue. This study identifies compounds that can prevent the development of these liver-stage parasites and may function as chemoprotective drugs for malaria. RESULTS To discover leads for next-generation chemoprotective antimalarial drugs, we used luciferase-expressing Plasmodium spp. parasites, dissected from more than a million mosquitoes over a 2-year period, to test more than 500,000 compounds for their ability to inhibit liver-stage development of malaria (681 compounds showed a half-maximal inhibitory concentration of<1 µM). Cluster analysis identified potent and previously unreported scaffold families, as well as other series previously associatedwith chemoprophylaxis. These leads were further tested through multiple phenotypic assays that predict stagespecific and multispecies antimalarial activity. This work revealed compound classes that are likely to provide symptomatic relief from bloodstage parasitemia in addition to providing protection. Target identification by use of functional assays, in vitro evolution, or metabolic profiling revealed 58 mitochondrial inhibitors but also many chemotypes possibly with previously unknownmechanisms of action, somewhichmay disrupt the host pathogen signaling. CONCLUSION Our data substantially expands the set of compounds with demonstrated activity against two known targets of chemoprotective drugs, cytochrome bc1 and dihydroorotate dehydrogenase. These present a rich collection of chemical diversity that may be exploited by members of the community seeking to accelerate malaria elimination with chemoprotection and chemoprophylaxis through open-source drug discovery.


Compound Libraries
The compound library was obtained from Charles River. The library consists of 538,273 compounds in DMSO at concentrations ranging from 0.5 to 7.84mM (average 2mM). Compounds were screened at 0.250µM to 39µM (average 10µM). Library characteristics are listed in Table S1.

Compound Suppliers
Compounds that were further evaluated for metabolomic studies and DHODH inhibition were purchased from the following supplier: MMV1068987 (MolPort-009-487-577), MMV1490397 Compound purity was assessed by mass-spectrometry and is shown in Table S8.

Plasmodium berghei exoerythrocytic parasites
Anopheles stephensi mosquitoes infected with P. berghei-ANKA-GFP-Luc-SMCON (Pb-Luc)(15) sporozoites bearing the antifolate-resistance-conferring Toxoplasma gondii dihydrofolate reductase-thymidylate synthase selectable marker, pyR2 (15) and which is integrated into the d-ssu-rrna fragment that serves as a target for integration into the genome in the nonessential c-or d-ribosomal rna gene unit (36). The GFP-Luc reporter is driven by the constitutively active pbeef1aa promoter (15). These were received from the Insectary Core Facility at New York University. The Pb-Luc sporozoites were obtained by dissection of infected Anopheles stephensi mosquito salivary glands. Dissected salivary glands were homogenized in DMEM media (Life Technology, CA) using a glass tissue grinder, filtered twice through a 20 M nylon net filter (Steriflip, Millipore), and counted using a Neubauer hemocytometer (C-Chip, InCyto, Republic of Korea). The sporozoites were kept on ice until needed.
The Dd2-attB and Dd2-ScDHODH parasite cell lines have been described previously (18) and were maintained under the same culture conditions as described above for wild-type Dd2 parasites.
Luciferase-based high throughput screening: P. berghei-Luciferase liver stage assay (Pbluc) and HepG2 cytotoxicity assay (HepG2tox) For the Pbluc high-throughput screen, we used P. berghei because of its higher infection rates in immortal human hepatocyte cell lines, making it more conducive to high-throughput screening than the human malaria parasites. HepG2-A16-CD81EGFP(37) cells were cultured at 37°C in 5% CO2 in DMEM media as described in Cell Lines section above. For both Pbluc and HepG2tox assays, 20-26 hour prior to sporozoites infection, 3x10 3 of the HepG2-A16-CD81EGFP20 cells in 5 μl of assay medium (DMEM without Phenol Red (Life Technologies, CA), 5% FBS, and 5x Pen Strep Glutamine (Life Technologies, CA)) at concentration 6x10 5 cells/ml were seeded in white solid bottom 1536-well plates (custom GNF mold ref# 789173-F, Greiner Bio-One). For primary single dose screens, 18 hours prior to infection, 50nL of compound in DMSO (0.5% final DMSO concentration per well) was transferred with a PinTool (GNF Systems) into the assay plates. For IC50 determinations in second and third round of the screenings, 18 hours prior to infection 50nL of compounds in 1:3 serial dilutions in DMSO (0.5% final DMSO concentration per well) were transferred with either Echo Liquid Handler (Labcyte) (for the second round of screening) or Acoustic Transfer System (ATS) (Biosero) (for the third round of screening). Atovaquone (5µM) and puromycin (10μM) at a single concentration were used as positive controls for the Pbluc and HepG2tox, respectively. 0.5% DMSO was used as negative control for both assays. Pbluc sporozoites were freshly dissected from the infected A. stephensi salivary glands, filtered twice through a 20μm nylon net filter (Steriflip, Millipore), counted in a hemocytometer, and adjusted to final concentration 200 sporozoites per 1μl in the assay media (DMEM without Phenol Red (Life Technologies, CA), 5% FBS, and 5x Pen Strep Glutamine (Life Technologies, CA). To infect the HepG2-A16-CD81EGFP cells, 1x10 3 sporozoites per well (5μl) were added with a single tip Bottle Valve liquid handler (GNF), and the plates were spun down at 37°C for 3 minutes with a centrifugal force of 330xg on normal acceleration and brake setting (Eppendorf 5810 R centrifuge). The HepG2-A16-CD81EGFP cell designated for toxicity studies were left uninfected, with 5μl of additional assay media added to each well to maintain equal concentrations of compounds with Pbluc infected plates. The plates were then incubated at 37°C for 48 h in 5% CO2 with high humidity to minimize media evaporation and edge effects.
Bioluminescence quantification for Pbluc and HepG2tox assays After incubation at 37°C for 48 h, the EEF growth and HepG2-A16-CD81EGFP cells viability were assessed by a bioluminescence measurement as follows: Media was removed by spinning the inverted plates at 150xg for 30 seconds; 2µl per well of BrightGlo (Promega) for quantification of Pbluc EEFs or CellTiterGlo (Promega) reagent (diluted 1:2 with deionized water) for quantification of HepG2-A16-CD81EGFP cell viability (HepG2tox) were dispensed with the MicroFlo (BioTek) liquid handler. Immediately after addition of the luminescence reagent, the luminescence was measured by the Envision Multilabel Reader (PerkinElmer).
Hit analysis and IC50 determination for Pbluc inhibition and HepG2tox assays For the first round of screening for 538,273 compounds in single point concentration in the 1536well plate, the lack of inhibition (baseline) was defined as the average bioluminescence of 64 wells with 0.5% DMSO, and the background inhibition was defined as average luminescence of 64 wells with 10μM atovaquone. The Pbluc growth inhibition was determined as a ratio of luminescence reads from every well with a compound minus the background inhibition readings with this difference divided by the baseline minus background inhibition readings. For the second round of screening first round of reconfirmation for 10,000 compounds in 8 points (1:3) dilution series, the background for the EEF inhibition was defined as the average of 16 wells with atovaquone at concentration of 10μM, and the background for the HepG2tox was defined as the average of 16 wells with puromycin at concentration of 10μM. IC50 values were determined using the average normalized bioluminescence intensity of 2 wells per concentration from 2 technical replicates of the 1536-well plates, and a nonlinear variable slope four-parameter regression curve fitting model in Prism 6 (GraphPad Software Inc.).
For the second reconfirmation round of screening (third round of screening for 631 compounds resourced from powder), the background for the Pbluc inhibition was defined as average bioluminescence of the 16 wells with atovaquone at single concentration of 5M, and the background for the HepG2tox was determined as the average of the 12 wells with puromycin at single concentrations of 5M. The baseline was defined as average bioluminescence of the 150 well with 0.5% DMSO. The data was normalized against positive and negative controls and IC50 were determined with Levenberg-Marquardt algorithm for curve fitting of the dose response data (fit parameters as min = 0.0, max ≥ 80.0, slope ≥ 0.0) using CDD Vault from Collaborative Drug Discovery (Burlingame, CA. www.collaborativedrug.com) All data in this study were archived using the CDD Vault from Collaborative Drug Discovery (Burlingame, CA, www.collaborativedrug.com).

Recombinant Photinis pyralis luciferase (Ffluc) inhibition assay
The luciferase inhibition assay was performed for the hits of the primary screen in the first and second rounds of reconfirmation (second and third rounds of screening). In the 1 st round of reconfirmations (second round of screening), 50 nL of compounds in 1:3 serial dilutions in DMSO (0.5% final DMSO concentration per well) was transferred with an Echo Liquid Handler (Labcyte) into white, solid-bottom 1536-well plates (custom GNF mold ref# 789173-F, Greiner Bio-One). Next, 8l of the Quantilum Luciferase (catalog #E1701) in 1:1000 dilution in DMEM (Life Technologies, CA) was dispensed with the MicroFlo (BioTek) liquid handler. As positive control 100nL resveratrol (Sigma catalog #5010) was added to each well to a final concentration of 500μM. Plates were incubated at 37°C for 3 hours. After the incubation 2μL of BrightGlo (Promega) was added to the wells with the MicroFlo liquid handler (BioTek). Immediately after addition of Bright Glo (Promega), the plates were read by an EnVision Multilabel reader (PerkinElmer). In the third round of screening (second round of reconfirmations for the 631 compounds resourced from powder), 40nL of compounds in 1:3 serial dilutions in DMSO (0.5% final DMSO concentration per well) were transferred in white solid bottom 1536-well plates (custom GNF mold ref# 789173-F, Greiner Bio-One) with Acoustic Transfer System (ATS) (Biosero). 8 l of 20 pM QuantiLum Recombinant Luciferase (Promega, catalog #E1701A) dilution in assay medium (DMEM without Phenol Red (Life Technologies, CA), 5% FBS, and 5x Pen Strep Glutamine (Life Technologies, CA)) were dispensed with the MicroFlo (BioTek) liquid handler. Luciferase Inhibitor-I (non-competitive) (Calbiochem, Ref # 119113) and Luciferase Inhibitor-II (competitive) (Calbiochem, Ref # 119114) at single concentration of 10μM and 9.8μM respectively were used as positive controls. As a negative control 0.5% DMSO was used. The plates containing tested compounds and QuantiLum Luciferase were incubated at RT for 15 min. After the incubation, 2μL of BrightGlo (Promega) was added to the wells with the MicroFlo liquid handler (BioTek). Immediately after addition, the plates were read by an EnVision Multilabel reader (PerkinElmer).

IC50 determination for Ffluc
To determine IC50 for luciferase inhibition in the luciferase inhibition assay in the first round of reconfirmations (second round of screening), the background luminescence was defined as the average of 48 wells with 500μM resveratrol and baseline was defined as DMSO readings in 64 wells. The data was normalized to positive and negative controls, and IC50 values were obtained using the average normalized luminescence intensity of 2 wells per concentration and a nonlinear variable slope four-parameter regression curve fitting model in Prism 6 (GraphPad Software Inc.).
To determine IC50 for luciferase inhibition in the third round of screening (631 compounds resourced from powder), the background for the luciferase inhibition was defined as average luminescence of the 16 wells with luciferase Inhibitor-I at single concentration of 10M, and the baseline was defined as average luminescence of the 152 well with 0.5% DMSO. The data was normalized against positive and negative controls and IC50 were determined with Levenberg-Marquardt algorithm for curve fitting of the dose response data (fit parameters as Min = 0.0, Max ≥ 80.0, Slope ≥ 0.0) using CDD Vault from Collaborative Drug Discovery (Burlingame, CA. www.collaborativedrug.com). All data in this study were archived using the CDD Vault from Collaborative Drug Discovery (Burlingame, CA. www.collaborativedrug.com). P. falciparum Dd2 asexual blood stage (ABS-Sybr) assay P. falciparum Dd2 strain was cultured as described above in Culturing Asexual Erythrocytic-Stage P. falciparum parasites. For high throughput screening, 40nL of compounds in 1:3 serial dilutions in DMSO (0.5% final DMSO concentration per well) was transferred with Acoustic Transfer System (ATS) (Biosero) into black, clear-bottom 1536-well plates (custom GNF-mold ref# 789092-F, Greiner Bio-One). Artemisinin (5µM) at single concentration was used as positive and 0.5% DMSO was used as negative control. A parasite suspension with 0.3% parasitemia and 2.5% hematocrit in screening medium (SM) (RPMI 1640 with l-glutamine, without phenol red (Life Technologies, CA) supplemented with 0.2% AlbuMAX II lipid-rich BSA, 0.014mg/mL hypoxanthine, 3.4mM NaOH, 38.4mM Hepes, 0.2% glucose, 0.2% sodium bicarbonate, and 0.05mg/mL gentamicin) was prepared and dispensed into the 1536-well black, clear bottom plates with pre-spotted compounds using the MultiFloTM Microplate dispenser (BioTek) at volume 8μl per well. The plates containing the compounds and parasites were incubated at 37°C for 72 h with water-soaked tissue in a double closed Ziploc bag gassed with 1% oxygen, 3% carbon dioxide, 96% nitrogen.

Fluorescence quantification P. falciparum asexual blood stage parasites (ABS-Sybr)
After incubation at 37°C for 72 h, the P. falciparum erythrocytic forms growth was assessed by fluorescence measurement as follows: detection reagent mixture (10x SYBR Green I (Invitrogen) in Lysis buffer (20mM Tris/HCl, 5mM EDTA, 0.16% Saponin wt/vol, 1.6% Triton X vol/vol) was added using MultiFloTM Microplate dispenser (BioTek) at volume 2μl per well. The plates were incubated in the dark at RT for 24 h. The fluorescence was read from the bottom of the plates by using the EnVision® Multilabel Reader (PerkinElmer) (485 nm excitation, 530 nm emission).

IC50 determination for P. falciparum ABS-Sybr
The background for the EF inhibition was defined as average fluorescence of the 16 wells with Artemisinin at single concentration of 5M, and the baseline was defined as average fluorescence of the 96 well with 0.5% DMSO. The data was normalized as a percentage of negative control (0.5% DMSO) and IC50 were determined with Levenberg-Marquardt algorithm for curve fitting of the dose response data (fit parameters as Min = 0.0, Max ≥ 80.0, Slope ≥ 0.0) using CDD Vault from Collaborative Drug Discovery (Burlingame, CA. www.collaborativedrug.com) All data in this study were archived using the CDD Vault from Collaborative Drug Discovery (Burlingame, CA. www.collaborativedrug.com).

P. vivax high content liver stage assay
The P. vivax liver stage assay used in this study has been described elsewhere (38). In brief, collagen coated 384-well plates (Grenier, Monroe, NC, USA) were seeded with 18,000 live primary human hepatocytes (BioIVT, Westbury, NY, USA) 1-4 days prior to sporozoite infection. Plating media containing serum (BioIVT) and supplemented with 50μg/mL penicillinstreptomycin, 100μg/mL of neomycin, and 10μg/mL of gentamicin was changed prior to sporozoite infection, the day after sporozoite infection, and between daily drug additions. The Human Subjects protocols for this study was approved by the Institutional Ethics Committee of the Thai Ministry of Public Health and the Oxford Tropical Medicine Ethical Committee, Oxford University, England (TMEC 14-016 and OxTREC . Blood containing at least 5 P. vivax gametocytes per 500 leukocytes (counted by Giemsa-stained blood smear) was collected by venipuncture into heparin tubes from malaria patients reporting to malaria clinicals after giving written informed consent. Collected blood from anonymized donors was pelleted and serum replaced (1 volume to pellet volume) with O+ serum and then fed to day 5-7 old female mosquitoes via artificial membrane feeder (Hemotek, Great Harwood, England) (39). Fed mosquitoes were separated from unfed and kept on 10% sucrose via wetted cotton in a environmental chamber set to maintain 26 ºC and 80% humidity. Salivary glands were dissected into Schneiders' insect medium (pH at 7.1) at day 14 post feed, counted by hemocytometer, diluted in media, and 7,000 sporozoites added to each assay well prior to centrifugation of the assay plate at 200 G for 5min (40). Compounds were supplied pre-plated at 10mM in DMSO (Evotec). To add compound a custom-ordered hand-held pin-tool, designed to transfer 40nL (VP Scientific, San Diego, CA, USA), was dipped in drug plates and then inserted into assay plates with 40μL media, achieving a 1000x dilution of drug. The pin-tool was washed and blotted twice into each of 5 washing solutions: 50% DMSO, 70% ethanol, VP clean solution, water, and methanol, prior to transfer for additional drug plates.
Each assay plate contained 12 DMSO control wells, 8 KDU691 positive controls wells, and 8 atovaquone reference wells. At 6 days post infection cultures were fixed with 4% paraformaldehyde for 10 min and washed twice with PBS prior to overnight staining at 4 ºC with recombinantly expressed mouse anti-PvUIS4 diluted to 40 ng/mL in dilution buffer (0.03% TritonX-100 and1 % (w/v) BSA in 1 PBS). Following three PBS washes, the primary stain was detected with goat anti-mouse Alexa Fluor® 488 (Molecular Probes, ThermoFisher, Waltham, MA, USA) diluted to 2µg/mL in dilution buffer, and counterstained with 10µg/mL Hoechst for 1hr at 37ºC. Assay plates were imaged and analyzed by the Operetta Imaging System and Harmony software 4.1 (Perkin Elmer, Waltham, MA, U.S.A.). Images were acquired with FITC, DAPI, and brightfield channels using a 20 objective. Host cell nuclei were quantified using DAPI counts while parasites were identified and classified by growth area. Raw parasite growth metrics were loaded into CDD Vault for normalization and hit identification. Z' factor for controls were calculated for both biological replicates of all three plates containing 631 screen compounds (41). Schizont count Z' ranged from -0.02 to 0.35 and schizont growth area -0.39 to 0.19.

Computational compound clustering
All 538,273 compounds in the primary screen were hierarchically clustered based on structure similarity using JChem fingerprint. Compound clusters are separated based on a minimum Tanimoto coefficient requirement of 0.85. Primary hits are defined as compounds with >82.98% inhibition. Hit enrichment probability of each cluster is estimated by the redundant siRNA activity (RSA)(42) p-value using inhibition percentage.

Metabolomic profiling
Metabolic response to selected compounds was determined using whole cell hydrophilic extraction followed by ultra-high precision liquid chromatography mass-spectrometry (UHPLC-MS) as previously described (19). Briefly, synchronous, staged trophozoite infected cells (24-36 hpi) were magnetically separated from culture. This was followed by quantification of cells via hemocytometry, subsequent aliquoting of 1x10 8 cells a well, and recovery for 1-2 hours under incubation. The wells were then treated with 10xIC50 of compound for 2.5 hours, with each condition performed in triplicate, and included a no drug control and a positive control of atovaquone in each experiment. Cells were then collected by centrifugation and washed using PBS. Hydrophilic extraction of metabolites was performed using 90% methanol containing isotopically-labeled aspartate (0.5μM) as an internal standard, and dried under nitrogen (20). Metabolite extracts were then measured by UHPLC-MS and the subsequent data was analyzed as previously described (20). The averaged metabolite data for each drug treatment was directly compared to an untreated control and the resulting fold-changes (Data S5) were hierarchically clustered based on a Pearson distance and visualized using MetaPrint self-organizing maps (20,44).

Low throughput ScDHODH assay and heatmap generation
Drug susceptibility in the Dd2-attB and Dd2-ScDHODH cell lines was measured by growth assay as previously reported (44). Briefly, synchronized ring stage parasites were cultured in the presence of twelve-point serial dilutions of the test compounds in triplicate in 384-well black clear-bottom plates for 72 h. Lysis buffer and SYBR Green I fluorescent dye was added as above, and after overnight incubation at room temperature fluorescence was read on a Molecular Devices SpectraMax M5 plate reader (bottom read, excitation and emission wavelengths of 485 nm and 530 nm). IC50 values were calculated using a nonlinear regression curve fit in Prism Software version 7 (GraphPad). The assay was replicated four to eight times. The extent of cross-resistance was determined by dividing the IC50 of the Dd2-ScDHODH line by the IC50 of the parent Dd2-attB strain, and taking the log10 of this IC50 ratio. These values were visualized on a heatmap generated in Prism Software version 7 (GraphPad). In cases where the IC50 value was outside the range of the highest dose tested, the maximum dose tested (10 µM) was used as a placeholder value to generate the heatmap.

well ScDHODH assay with proguanil (high-throughput mode)
To identify and differentiate between Cytb and DHODH inhibitors in the set of 631 compounds from the third round of screening, drug susceptibility in D10-attb and D10-ScDHODH bloodstage parasites in the presence and absence of proguanil was determined. The assay was performed in 1536-well format with the same experimental conditions as described in the section: P. falciparum Dd2 Asexual blood stage assay (ABS-Sybr), with the exception that the compounds were tested in eight-point serial dilutions and in duplicate. In the absence of proguanil (1µM), the D10-ScDHODH strain is resistant to mitochondrial electron transport chain inhibitors. In the presence of 1M proguanil, sensitivity to Cytb inhibitors is restored but resistance to DHODH inhibitors remains. This distinction allows identification of the specific mitochondrial electron transport chain target for each compound.

In vitro evolution experiments
For MMV1432711, three independent resistant cultures were created by exposing a clonal Dd2 strain to sublethal drug pressure for 16 weeks at a starting concentration of 30nM using the "ramp-up" method to a final concentration of 600nM. Parasite growth rates were monitored daily and IC50 curves collected every 2 weeks. Dose-response assay showed a shift of ~8-fold for the two replicates compared to the parental clone (106nM and 149.6nM relative to the parental value of 15.48nM). These replicates also showed a phenotypic change since MMV1432711 typically arrested parasites in late trophozoite phase. DNA was isolated from the resistant nonclonal cultures and submitted for whole genome sequencing.
For MMV1454442, approximately 5×10 8 Dd2 parasites in three independent flasks were treated at an initial concentration of 2µM MMV1454442 until parasites could not be seen by microscopy. Following treatment, compound pressure was removed and cultures were allowed to recover in the absence of drug. Once healthy parasites were seen and parasitemia reached ~2%, compound pressure was reinstated at an increased dose of 3µM. Cultures were again allowed to recover in the absence of compound and then re-pressured with 5µM MMV1454442. Parasites that recovered from the 5µM treatment were phenotypically resistant by dose-response assay. Resistant parasites were cloned by limiting dilution in the absence of drug pressure. Individual clones were then phenotyped again and submitted for whole genome sequencing.
For MMV1427995, approximately 10 9 parasites were exposed to 1.3μM of the compound (approximately 4x IC50). Parasites that recrudesced on day 21 showed no IC50 shift and were then subjected to a ramping procedure starting at 650nM and gradually building up to 1.5μM (approximately 4.75 x IC50) over a course of 40 days. The resulting resistant populations were cloned as described above and, phenotyped for MMV142795 susceptibility and clones with varying IC50 shifts were submitted for whole genome sequencing.
DNA Sequence analysis of MMV1454442, MMV1432711, and MMV1427995-resistant clones Infected RBCs were washed with 0.05% saponin and genomic DNA (gDNA) was isolated using the DNeasy Blood and Tissue Kit (Qiagen) standard protocol. gDNA was tagmented and amplified with the Nextera XT Kit (Cat. No FC-131-1024, Illumina) following the standard dual index protocol to prepare the sequencing libraries. Samples were sequenced on the Illumina HiSeq 2500 in RapidRun mode, generating paired-end reads 100bp in length. Following sequencing, reads were aligned to the P. falciparum reference genome (PlasmoDB v13.0) using the previously described Platypus pipeline (45). Sequences contained an average of 24,811,649 reads of 100bp with 95.6% mapping to the reference genome. An average coverage of 72.88x coverage was obtained with 95% of the genome covered by 5 or more reads. SNVs and INDELs were called using GATK's HaplotypeCaller and filtered using GATK's best practice recommendations (46). CNVs were detected using GATK's DiagnoseTargets, which computes average coverage using an input of gene lengths and locations as the analyzed intervals. Genes with known high variability in copy numbers, such as those in the var, rifin, and stevor families, were removed from the interval list to avoid potential misalignment. Read coverage was normalized first by log transforming and centering genes to the mean and then centering sequence samples to the sample means. CNVs were called if they contained >3x coverage compared to the average over the length of the CNV and were visually confirmed in Integrative Genomics Viewer (IGV).

Figure S1
Figure S1. Library characteristics. Differences between the initial library (538,273 compounds, blue) and hits (9,989, red). Comparison of library characteristics between full compound library (blue) and primary screen hits (red). Each bar represents the mean value for each characteristic with error bars indicating the standard deviation, as reported in Tables S1 and S2.       Table S4. Mitochondrial inhibition data with proguanil rescue. Shown is dose response activity of select compounds in Dd2-scDHODH parasites with addition of proguanil (PG), which exhibits a synergistic effect specifically with Cytbc1 inhibitors. Compounds with a DHODH+PG/Dd2 ratio of 1 are Cytbc1 inhibitors (indicating a restoration of sensitive phenotype) while those with a ratio >3 (resistance phenotype) are DHODH inhibitors.  Table S5. Clone characteristics for resistant P. falciparum lines. IC50 of each compound was determined in dose-response format using a SYBR Green-I cell proliferation assay (ABS-Sybr).   Table S7. Identified copy number variants in evolved, compound-resistant clones. Variants were detected as described in methods and are based on increased read coverage per gene in evolved clone relative to sensitive parent. Start and end positions on the indicated chromosome are approximate and include the indicated gene.