Additional PfCRT mutations driven by selective pressure for improved fitness can result in the loss of piperaquine resistance and altered Plasmodium falciparum physiology

Abstract

Our study leverages gene editing techniques in Plasmodium falciparum asexual blood stage parasites to profile novel mutations in mutant PfCRT, an important mediator of piperaquine resistance, which developed in Southeast Asian field isolates or in parasites cultured for long periods of time. We provide evidence that increased parasite fitness of these lines is the primary driver for the emergence of these PfCRT variants. These mutations differentially impact parasite susceptibility to piperaquine and chloroquine, highlighting the multifaceted effects of single point mutations in this transporter. Molecular features of drug resistance and parasite physiology were examined in depth using proteoliposome-based drug uptake studies and peptidomics, respectively. Energy minimization calculations, showing how these novel mutations might impact the PfCRT structure, suggested a small but significant effect on drug interactions. This study reveals the subtle interplay between antimalarial resistance, parasite fitness, PfCRT structure, and intracellular peptide availability in PfCRT-mediated parasite responses to changing drug selective pressures.

Keywords: PfCRT; Plasmodium falciparum; drug resistance evolution; fitness; malaria.

Fig 1

Mutations reduce fitness cost of Dd2 + F145I PfCRT in vitro. (A) Isogenic pfcrt-modified parasite lines, each generated in the Dd2 strain, were co-cultured at an approximately 1:1 ratio with a reference Dd2 line expressing GFP (Dd2-GFP). The labeling indicates the PfCRT isoform expressed by each gene-edited line. The pfcrt-edited (ed) lines expressing Dd2 + F145I + F131C, Dd2 + F145I + C258W, or Dd2 + F145I + I347T were generated using customized zinc finger nucleases. Cultures were analyzed by flow cytometry every 2–3 days up to day 21 (10.5 parasite generations), and the mean ± SEM percentages of GFP⁺ parasites were plotted over time. Values above the 50% dashed line indicate that the test line was less fit than Dd2-GFP. (B) Mean ± SEM relative fitness cost shown as a percent reduction in parasite growth per 48-h generation is shown, relative to the isogenic Dd2^Dd2 line. N = 2–3; n = 2.

Fig 2

Dd2^{Dd2+F145I+C258W} and Dd2^{Dd2+F145I+I347T} display increased susceptibility to PPQ and altered susceptibility to CQ due to altered drug transport via PfCRT. (A) Survival rates of isogenic pfcrt-modified lines, each generated in the Dd2 strain, were determined using the in vitro PSA (starting with 0- to 6-h rings and treated for 72 h). The 10% cutoff represents a standard 200 nM threshold for PPQ resistance in this assay (30). The labeling indicates the PfCRT isoform expressed by each recombinant line. “ed” refers to lines where the novel PfCRT mutation (F131C, I347T, or C258W) was gene edited into the Dd2^Dd2+F145I parasite line. Survival values are shown in Table S1 as means ± SEM. N = 7; n = 2. (B and C) Mean ± SEM IC₉₀ values (Table S2) were calculated from 72-h concentration-response assays for (B) PPQ and (C) CQ. N = 4–7; n = 2. Significance was determined using Mann-Whitney U tests. Comparisons are shown between pfcrt-edited lines and their isogenic controls. (D and E) Mean ± SEM uptake of 100 nM (D) ³H-PPQ or (E) ³H-CQ was measured for 1 min in proteoliposomes containing the indicated PfCRT variants or in control liposomes. Data are means of n = 4–6. *P < 0.05; **P < 0.01; ***P < 0.001. (B–E) Individual circles indicate values from each independent experiment.

Fig 3

Molecular dynamics of the various PfCRT isoforms in this study. (A) Average energy minimized from cryoEM template (EMMD; see reference 32) structures of Dd2 + F145I, Dd2 + F145I + F131C, Dd2 + F145I + C258W, and Dd2 + F145I + I347T, with mutations highlighted relative to Dd2. (B) Root mean square deviation (RMSD) comparison of Dd2 PfCRT EMMD (32) vs Dd2 + F145I, Dd2 + F145I + F131C, Dd2 + F145I + C258W, or Dd2 + F145I + I347T EMMD structures (from left to right). Differences are highlighted using ColorByRMSD where blue indicates minimum RMSD and red indicates maximum RMSD. Gray regions were not compared. The structures are all highly similar with minimum RMSDs (0.18, 0.41, 0.11, and 0.39 Å from left to right), low average RMSDs (2.22, 2.44, 1.81, and 2.54 Å from left to right), and small regions of maximum RMSDs (11.91, 10.42, 9.44, and 9.73 Å from left to right) corresponding to particularly flexible PfCRT domains. (C) The previously proposed CQ binding site A in Dd2 showed direct interactions between the anilinyl CQ nitrogen (middle blue on the CQ molecule) and the oxygen of N84 (red) that is ~3.2 Å away. F145, positioned directly between the proposed CQ binding sites A and B, is likely to interact with the drug’s quinolinyl ring throughπ-π interactions during transit between the two sites. Notably, these π-π interactions were predicted to disappear upon mutation to isoleucine (F145I). The distance between the drug quinolinyl and F145 phenyl rings in site A is 7.4 Å. (D) The predicted three-phenylalanine π-π bundle for Dd2 PfCRT formed between the side chains of F131, F251, and F252 near the cytosolic ends of helices 3 (aqua) and 7 (yellow-green), respectively, is lost in the Dd2 + F131C + F145I isoform.

Fig 4

Dd2^{Dd2+F145I+C258W} results in decreased accumulation of peptides. (A) Heatmap of 134 peptides (detected in positive or negative modes and present in metabolite extracts from any of the parasite lines) classified by peptide length. These data are presented as the log₂ fold change in the average abundance of each peptide identified compared to Dd2^Dd2+F145I. N = 3; n = 3. Oligo refers to peptides eight or more amino acids in length. PfCRT haplotypes of the isogenic lines are listed. (B and C) Significant differences in peptide levels between lines with a given mutation and Dd2 + F145I are shown (P < 0.05 in unpaired t-tests). Peptides are classified by (B) peptide length or (C) isoelectric point. Details are provided in Tables S4 and S5.