Clinically relevant atovaquone-resistant human malaria parasites fail to transmit by mosquito

P. falciparum cultivation and mutant selection

Studies were conducted with low passage P. falciparum NF54 (MRA-1000 BEI Resources, NIAID, NIH), grown asynchronously at 37 °C under candle jar atmosphere ( ~ 13% O₂, 3% CO₂, 84% N₂) in RPMI supplemented with 2.1 mM L-glutamine, 25 mM HEPES, 0.25% NaHCO₃, and 0.37 mM hypoxanthine, in 10% heat-inactivated human serum (Interstate Blood Bank) and O+ erythrocytes at 2 or 4% hematocrit²⁷. Periodic testing for mycoplasma was negative (MP0025, Sigma-Aldrich). Erythrocytes obtained weekly from healthy donors under a Johns Hopkins Institutional Review Board-approved protocol were provided without identifiers to the laboratories. Prior to atovaquone pressure, parasites were adapted over 14 d to 5% CO₂ in air, as described previously²⁶.

For mutant selection, atovaquone (PHR1591, Sigma Aldrich) in DMSO (D128500, ThermoFisher) was further diluted in a medium containing 10% serum such that the final DMSO was ≤0.1%. Cultures were exposed in parallel to atovaquone or no drug (Supplementary Fig. S1a,b). Medium was replaced every other day; cultures were split 1:2 every 7 d; only the first split was retained. Cells were maintained under constant drug pressure for 60 days or until healthy parasites appeared. Upon confirmation of Y268S mutation, atovaquone was removed and drug-resistant lines and their concurrent no-drug controls were returned at once to candle jar conditions, amplified, and aliquots were frozen (Glycerolyte 57, McKesson General Medical). In all subsequent experiments WT and mutants were manipulated and studied identically and in parallel.

To obtain gametocytes, cultures were initiated at 0.5% parasitemia, and 4% hematocrit, and the medium was replaced daily for 15-19 d until gametocytes reached stage V. Mutant parasites are being deposited with ATCC.

In vitro drug sensitivities

Asynchronous P. falciparum (0.25% parasitemia, 1.2% hematocrit) in microtiter plates was exposed to serial dilutions of the drug (in quadruplicate) for 72 h. [³H(G)]-hypoxanthine (16 Ci/mmol, 1 mCi/mL NET177001MC, Perkin Elmer) was added for the final 24 h, and incorporated radiolabel was evaluated in harvested cells, as described previously⁴¹. EC₅₀ values were obtained by nonlinear regression analysis (Prism 6; GraphPad).

cytB and whole genome sequences

For cytB sequence in parasites surviving drug pressure, lysates were prepared from cultures at 1% parasitemia (90 °C, 5 min), and an 1848 bp fragment (containing the 1131 bp cytB open reading frame) was amplified with primers (Supplementary Fig. S2a, b) and Phusion DNA polymerase (NEB). PCR product was verified by gel electrophoresis as single band of predicted size, and sequenced. Sequences were aligned to P. falciparum 3D7 cytochrome b (PF3D7_MIT02300) using NCBI blast, and positions on all trace files were assessed in SnapGene 7.0.

For whole genome sequences to identify the causative mutation⁴², DNA was isolated from 10⁸ asexual parasites; QiAamp DNA Mini Blood kit, Qiagen), quantified (Qubit HS ds DNA assay (Qubit Flex Fluorometer, ThermoFisher)), and quality assessed by Genomic Screentape analysis (TapeStation 2200, Agilent). Barcoded libraries for DNA-Seq were synthesized from 100 ng DNA (Celero EZ-Seq kit with NuQuant, Tecan Genomics), quality assessed by High Sensitivity DNA Lab Chips (BioAnalyzer 2100, Agilent), quantified by Qubit HS dsDNA assay, and sequenced on Illumina’s MiSeq platform, (2 x 300 bp v3 with 5% PhiX; JHMI Synthesis and Sequencing Facility). Using Partek Flow 10.0.22.0410 with sequence defaults, analyses included pre-alignment QA/QC, adaptor/read trimming, reference genome alignment to P. falciparum NF54 (PlasmoDB) using Bowtie2 2.2.5, post-alignment QA/QC, FreeBayes variant calling, variant filtering (VarQual ≥ 30), and variant annotation based on precedence rules. To identify changes unique to atovaquone-resistant parasites, sequences were compared to that of the reference strain P. falciparum NF54 GCA009761475.1 (PlasmoDB).

Gametocyte exflagellation

Just prior to mosquito feed, 200 μL of gametocyte culture was taken to assess male exflagellation^43,44. Briefly, cultures were centrifuged in a warm tube (500 x g, 5 min), 2 µL of cell pellet was placed on a glass slide, covered, and incubated (room temperature, 15 min). The number of exflagellation centers in 11 independent 400X fields was counted and normalized to 1.5% gametocytemia. In three of four experiments, counts were obtained in a blinded fashion.

Mosquito maintenance and infection. All experiments were conducted with An. stephensi (Liston, NIH) or An. gambiae (Keele, NIH) reared and maintained in the Johns Hopkins Malaria Research Institute insectary (27 °C, 80% humidity, 12/12-h light/dark cycle), and provided 10% sucrose solution-soaked cotton pads⁴³. For each experimental group, 100–200 mosquitoes at 3–7 d post-emergence were fasted overnight prior to 37 °C water-jacketed glass membrane feeding, for one hour, by established protocols⁴³. All bloodmeals were adjusted to 0.5% stage V gametocytemia. Thereafter, 10% sucrose was withheld for 48 h to eliminate unfed mosquitoes.

Ookinete, oocyst and sporozoite counts

Ookinetes were counted by an adaptation of established protocols⁴⁵. At 20-26 h post bloodmeal, midguts were dissected from ten visibly blood-fed mosquitoes, placed in 20 μL 3% acetic acid, and vigorously pipetted to lyse midguts. An aliquot (2 μL) was loaded on a hydrophobic slide (VWR, 100488–892), spread to fill a 14 mm circle, dried, fixed with methanol, and stained with Giemsa. For each experimental group, the number of mature and immature ookinetes in sixty 1000X fields was recorded.

Oocyst and sporozoite counts were obtained by adaptation of established protocols⁴³. At 9–10 d post feed, 15–30 midguts per group were dissected into 1X PBS on a hydrophobic slide (VWR, 100488-892). Samples were stained with 0.1% mercurochrome (M7011, Sigma Aldrich), covered, and examined (100X) for oocysts. Oocyst diameters were analyzed in ImageJ version 1.53. At 17–20 d post feed, salivary glands from 20-38 mosquitoes per group were collected, transferred to 100 μL 1X PBS on ice, sedimented (5000 g, 30 sec), and homogenized (sterile pestle, 30 sec). Parasites were counted by hemocytometer to obtain an average sporozoite count per mosquito.

Transmission from An. stephensi to huHep mice

All animal studies were conducted under a protocol approved by the Johns Hopkins Animal Care and Use Committee with huHep mice (FRG KO (Fah–/–, Rag2–/–, Il2rg–/–) on NOD background) engrafted with human hepatocytes and capable of maintaining circulating human erythrocytes (Yecuris^46,47). Female mice (27–28 wk old) were housed with sterile water, chow (ad libitum), and bedding, at 20-23 °C and 30-70% humidity, with 12-hr light/dark cycles. Until ≥8 d prior to infection they were maintained on 1 mg/mL 2-(2-nitro-4-trifluoro-methylbenzoyl)1,3-cyclohexedione (CuRx™ Nitisinone, 20-0027) in sterile 10% dextrose. Given the rarity and expense of huHep mice, the subtle, if at all, difference in infectivity of female versus male animals (including humans), these studies were conducted with female mice. Mosquitoes (17–18 d post-infection with WT or isogenic atovaquone-resistant parasites) were fed through a mesh-covered container on mice anesthetized with 4–6 μL/g ip of 14 mg/mL ketamine (67457010810, Myland Pharmaceuticals) plus 1.2 mg/mL xylazine hydrochloride (X1251-1G, Sigma Aldrich) in PBS. Mosquitoes fed for 14 min, with interruption every two min to limit blood loss in mice; over 80% evidenced having taken blood. At 5.5 and 6 d post mosquito probe, mice were injected intravenously with 600 μL human blood (50% red cells, 25% O+ serum, 25% RPMI). Human erythrocyte engraftment was assayed by flow cytometry of peripheral blood incubated 1:50 with APC-conjugated TER-119 rat antimouse monoclonal antibody (BD Biosciences catalog no 557909, lot 9213638), with gates at APC⁺ and APC⁻ (mouse and human erythrocytes, respectively) (Attune NxT Flow Cytometer, BD Biosciences) (Supplementary Fig. S11).

Starting 6.5 d after infection, Giemsa-stained blood films were monitored twice daily. At patency, all mice were sacrificed. Blood was collected by cardiac puncture into citrate/phosphate/dextrose (Sigma, C7165), for PCR (150 µL aliquot) and in vitro culture (the ≥1 mL residual blood volume). After perfusion with 15 mL 0.9% NaCl, the liver and spleen were harvested, weighed, and stored at −70 °C. Blood cultures were maintained for three weeks or until a robust parasitemia was evident.

PCR assays

Mouse tissues were assayed for STEVOR or cytB by nested PCR. DNA was isolated from blood (QiaAmp DNA Mini Blood Kit, Qiagen; or Monarch Genomic DNA purification kit, NEB) or from perfused organs homogenized (Fisher Bead Mill 4, 2.4 mm metal beads) prior to extraction (DNA Maxi Kit; Qiagen). For STEVOR assay, eluates containing DNA (from 25 µL whole blood, or 2 µg DNA from liver or spleen (Nanodrop 2000, ThermoFisher)) were analyzed as described previously^48,49. For cytB analysis, the 25 μL primary amplification reaction contained 50 nM each primer (Supplementary Fig. S2a,c), 200 μM each dNTP (NEB), 1 U Phusion DNA polymerase (NEB), and 10 ng template DNA. The 25 μL nested reaction contained 200 nM of each primer (Supplementary Fig. S2a,c), 200 μM each dNTP, 1 U of Phusion DNA polymerase, and 1 μL of the primary reaction product. Thermocycling conditions for both reactions were: denaturation (98 °C, 3.5 min); 25 cycles of (denaturation (98 °C, 30 sec), annealing (56 °C, 40 sec), and extension (72 °C, 45 sec)); and final extension (72 °C, 5 min). Products were separated by agarose gel electrophoresis and visualized by ethidium fluorescence. Serial dilutions into whole blood of a known number of P. falciparum-infected erythrocytes indicated both methods detect ≥10 parasites per 25 µL blood sample.

Statistical analyses

Data were collected in Excel 16.73. Dose-response data were analyzed as described in detail previously⁴¹. Unless indicated otherwise, differences between Y268S mutants and WT controls were assessed by the Mann−Whitney test or Fisher’s exact test, in GraphPad Prism 9.

Reporting summary

Further information on research design is available in the Nature Portfolio Reporting Summary linked to this article.