Functional analysis of Plasmodium falciparum subpopulations associated with artemisinin resistance in Cambodia

Abstract

Background: Plasmodium falciparum malaria is one of the most widespread parasitic infections in humans and remains a leading global health concern. Malaria elimination efforts are threatened by the emergence and spread of resistance to artemisinin-based combination therapy, the first-line treatment of malaria. Promising molecular markers and pathways associated with artemisinin drug resistance have been identified, but the underlying molecular mechanisms of resistance remains unknown. The genomic data from early period of emergence of artemisinin resistance (2008-2011) was evaluated, with aim to define k13 associated genetic background in Cambodia, the country identified as epicentre of anti-malarial drug resistance, through characterization of 167 parasite isolates using a panel of 21,257 SNPs.

Results: Eight subpopulations were identified suggesting a process of acquisition of artemisinin resistance consistent with an emergence-selection-diffusion model, supported by the shifting balance theory. Identification of population specific mutations facilitated the characterization of a core set of 57 background genes associated with artemisinin resistance and associated pathways. The analysis indicates that the background of artemisinin resistance was not acquired after drug pressure, rather is the result of fixation followed by selection on the daughter subpopulations derived from the ancestral population.

Conclusions: Functional analysis of artemisinin resistance subpopulations illustrates the strong interplay between ubiquitination and cell division or differentiation in artemisinin resistant parasites. The relationship of these pathways with the P. falciparum resistant subpopulation and presence of drug resistance markers in addition to k13, highlights the major role of admixed parasite population in the diffusion of artemisinin resistant background. The diffusion of resistant genes in the Cambodian admixed population after selection resulted from mating of gametocytes of sensitive and resistant parasite populations.

Keywords: Admixed subpopulations; Artemisinin resistance; Cambodia; Malaria; Network based stratification; Plasmodium falciparum; Population fragmentation; Redox metabolism; Shifting balance theory; k13.

Fig. 1

Plasmodium falciparum population structure in Cambodia during early period of emergence of artemisinin resistance (2008–2011). a Map of Cambodia and localization of health center (red dots) from where the 167 samples originate. b Eight groups were identified by hierarchical clustering based on 21,257 non-synonymous SNPs to classify 167 samples. Naming of subpopulations was performed according to previous description of the P. falciparum parasite population structure [12], and the overlap with earlier classification was around 90%. KH5 ART-R subpopulation was introduced. The KH1 and KH2 subpopulations [12] were split in two groups. Group structure was confirmed by network based stratification method, based on co-expression data. Ten samples from the admixed subpopulation KHA were differentially classified (overlap between disks). Small overlap was found between KH3 and KH2.2. One sample in each subpopulation was found in the other group. Surface of the disks is proportional to the number of samples (underlined characters). The color code is reminiscent of the major k13 alleles found in the subpopulation: shades of green: WT, blue: R539T, yellow: Y493H, and shades of red: C580Y

Fig. 2

Flowchart showing the neighboring genes of the k13 gene in co-expression based interaction network. The P. falciparum gene–gene interaction network based on co-expression data is recovered from STRING v10 [32]. The unconnected nodes are not considered in this analysis. The force directed layout is used to plot the network. The genes are colored according to the maximum expression stage data [34] recovered from PlasmoDB [30]. The genes highlighted in “red” color, are the first, second and third neighbor genes of the k13 gene, shown using four networks in clockwise orientation. The neighbor genes are marked using Cytoscape v3.2.1. The barplot with each network represents the number of genes expressed in each stage per diffusion

Fig. 3

Selection of artemisinin resistance associated genes. a Emergence of a common genetic background associated with all five ART-R subpopulations. 5-way Venn-diagram emphasizes great importance of the 265 common set of significant genes in ART-R subpopulations. Genes have at least one significant SNP in the ART-R subpopulations. b Selection and diffusion of the artemisinin-resistance background after crossing with KHA parasites. Red square: genes with significant ALT alleles in the admixed population and in the ART-R subpopulations (57 genes). Alleles for these genes may have been acquired by diffusion after crossing. c The 97 genes (76 with connections) specific to ART-R subpopulation genetic background are connected with ring or schizont stage in P. falciparum. First neighboring genes (in red) in the interaction network based on co-expression data recovered from STRING v10 [32]. The unconnected nodes are not considered in this analysis. The genes are colored according to the maximum expression stage data [34]. The bar plot with each network represents the number of genes expressed in each stage per diffusion. d Conservation of biological pathways during selection and diffusion of significant mutations from original ART-R related genetic background

Fig. 4

Major functions or pathways related with the genetic background of artemisinin resistant Cambodian parasites. a Cell localization of pathways (stars) and significant genes (boxes) with major known biological functions. Annotation terms were taken from genes current description in PlasmoDB [30] and GO terms. Star numbers refer to biological pathways or cell localization: 1, Apical complex; 2, Autophagy; 3, DNA/repair; 4, Drug resistance; 5, Lipid biosynthesis; 6, Mitosis/Meiosis; 7, Oxidation–reduction; 8, RNA metabolism (mRNA); 9, RNA metabolism (tRNA); 10, Cell signalization; 11, Surface antigen; 12, Traffic; 13, Transport; 14, Transcription; 15, Ubiquitination. b Interplay between ubiquitination and cell division in artemisinin resistant parasites. Frequency of isolates having at least one significant mutation or a rare mutation was calculated for the ART-R subpopulations and KHA parasites. Plasma membrane and endoplasmic reticulum membrane are represented by grey bars. Schema was built from protein–protein interactions or relationship given in databases and literature. Orange boxes: genes from the 57 genes set; White boxes: significant genes from the 265 set. AKT has grey box because it contains no significant mutations, but only rare mutations

Fig. 5

Distribution of the significant non-synonymous SNPs along amino acid sequence of K13, PI3K, PDE1, ARK2, ATG18 and ATG7 proteins. Domain structure of Kelch protein (K13), phosphatidylinositol 3-kinase (PI3K), cGMP-specific phosphodiesterase (PDE1), serine/threonine protein kinase (ARK2), autophagy related protein (Atg18 and Atg7), recovered from Pfam database server. Amino acid substitution and position in the protein are given for all non-synonymous mutations. Color code of mutations: black, not significant in any subpopulation; green, significant in only one subpopulation; red, significant mutations observed in several subpopulations