Computational protocol: Plasmepsin II–III copy number accounts for bimodal piperaquine resistance among Cambodian Plasmodium falciparum

Similar protocols

Protocol publication

[…] Drug susceptibility was measured using the SYBR Green I method as previously described. In brief, tightly synchronized 0–6 h rings were grown for 72 h in the presence of different concentrations of drugs in 384-clear-bottom well plates at 1% hematocrit, 1% starting parasitemia and 40 μl of 0.5% Albumax culture media. Growth at 72 h was measured by SYBR Green I (Lonza, Visp, Switzerland) staining of parasite DNA. A 24-point dilution of PPQ (Sigma-Aldrich, St. Louis, MO) and a 12-point dilution series of the rest of the drugs (DHA, ART, AS, MEF, and LUM; Sigma-Aldrich, St. Louis, MO) were carried out in triplicate and repeated with at least three biological replicates. Drug stocks were resuspended in dimethyl sulfoxide except for CQ prepared in 0.1% Triton X-100 in water and PPQ prepared in 0.1% Triton X-100 and 0.5% lactic acid in water to ensure complete dissolution, as lactic acid enhanced PPQ solubility. Drugs dispensed by a HP D300 Digital Dispenser (Hewlett Packard Palo Alto, CA), with the Triton X-100 enhancing dispensing of aqueous drug stocks. Relative fluorescence units was measured at an excitation of 494 nm and emission of 530 nm on a SpectraMax M5 (Molecular Devices Sunnyvale, CA) and analyzed using GraphPad Prism version 7 (GraphPad Software La Jolla, CA). EC50 values were determined with the curve-fitting algorithm log(inhibitor) vs. response–Variable slope except for PPQ. PPQs bimodal dose–response would not allow for any curve-fitting hence susceptibility was measured by calculating the area under the second response curve (AUC). AUC was calculated by fitting a six-dimensional least-square polynomial equation to each sample’s dose–response curve and integrating the fitted equation over the region corresponding to the second response curve. These equations were fit to the data using the polyfit module in NumPy, a fundamental package for scientific computing using Python. We chose to use a six-dimensional least-square polynomial equation because it captures the dynamics of the second response curve better than equations with fewer dimensions. Using equations with dimensions > 6 do not result in major differences in calculated AUC values. In fact, AUC values calculated using the boundaries identified by our equation fitting method (0.10 μM – 30 μM) gave roughly the same result. The boundaries of the second response curve were determined by identifying the drug concentration corresponding to the average local minima before and after the second response curve across all drug-assayed samples.Spearman correlation analysis was performed to assess the relationship between the antimalarial EC50 values and PPQ AUC, in vivo clearance half-life, ring survival assay survival rate or pfmdr1 copy number. P values < 0.05 were considered significant. [...] Parasites were stained in 10 × SYBR Green I in 1 × PBS for 30 min in the dark at 37 °C. The staining solution was removed and cells were resuspended in five times the volume of the initial volume of PBS. FACS data acquisition was performed on a MACSQuant VYB (Milteni Biotec) with a 488 nm laser and a 525 nm filter and analyzed with FlowJo 2. RBCs were gated on the forward light scatter and side scatter and infected RBCs were detected in channel B1. At least 100,000 events were analyzed per sample. [...] This publication uses data generated by the Pf3k project ( Except for the KH004-057 and the KH001-053 subclones (data deposited:, we used the variant call format (VCF) files provided by the Pf3k project to perform whole genome analyses. For KH004_057 and the KH001_053 subclones, gDNA was extracted and sheared with a Covaris S220 Focused-ultrasonicator (Covaris, Woburn, MA, USA). Illumina-compatible libraries were prepared on the Apollo 324 (WaferGen Biosystems, Fremont, CA, USA) and sequenced on an Illumina HiSeq 2000 (Illumina, San Diego, CA, USA). P. falciparum populations were sequenced with the goal of reaching over 60 × average fold-coverage across the genome. Reads were aligned to the P. falciparum 3D7 reference assembly (PlasmoDb v 7.1) using the Burrows-Wheeler Aligner (version 0.5.9-r16). Variant calls were determined using the GATK Unified Genotyper using the parameter and quality thresholds described in the supplementary information of a previous paper. The resulting VCF files were then combined with the VCF files downloaded from the Pf3k project and filtered to remove non-variant sites using VCFtools.To determine whether any previously implicated drug resistance SNPs were associated with PPQR, we analyzed the SNPs in the genes listed in the Supplementary Data . Samples were divided into two categories based on their AUC values. Samples with an AUC ≤ 35 were considered PPQS and samples with AUC > 35 were considered PPQR. Based on these categories, we used the software package PLINK (version 1.8) to determine whether any SNP was associated with PPQR.To assess CNV, we quantified the fold-change in the average read depths of non-variant sites within plasmepsin I (PF3D7_1407900), plasmepsin II (PF3D7_140800), and a nearby conserved gene of unknown function (PF3D7_1408200) as compared to the average read depths of non-variant sites in several randomly chosen regions of the genome (nucleotide positions at Chr 3: 353552–361453, Chr 5: 322315–329693, Chr 14: 578376–603856). These regions contain mostly coding sequence and have little variation in read depth. Genes within these regions are annotated as: a putative DNA polymerase epsilon subunit (PF3D7_0308000), a 6-cysteine protein (P38) (PF3D7_0508000), a putative atypical protein kinase, ABC-1 family (PF3D7_1414500), an RNA guanyltransferase (PF3D7_1414600), a putative carboxyl-terminal hydrolase (PF3D7_1414700), a putative small nuclear ribonucleoprotein-associated protein B (PF3D7_1414800), and several conserved proteins of unknown function (PF3D7_0308100, PF3D7_0507800, PF3D7_0507900). P38 is suspected to be involved with gamete fertilization and under some level of immune-mediate balancing selection, but little is known about the functions of the other genes. None have been implicated in drug resistance studies. […]

Pipeline specifications

Software tools Numpy, FlowJo
Applications Miscellaneous, Flow cytometry
Organisms Plasmodium falciparum, Toxoplasma gondii
Diseases Malaria