Genetic diversity of the Pvk12 gene in Plasmodium vivax from the China-Myanmar border area

Plasmodium falciparum resistance to artemisinin emerged in the Greater Mekong Sub-region has been associated with mutations in the propeller domain of the kelch gene Pfk13. Here the polymorphisms in Pvk12 gene, the orthologue of Pfk13 in Plasmodium vivax, were determined by PCR and sequencing in 262 clinical isolates collected in recent years (2012–2015) from the China-Myanmar border area. Sequencing of full-length Pvk12 genes from these isolates identified three synonymous mutations (N172N, S360S, S697S) and one non-synonymous mutation M124I, all of which were at very low prevalence (2.0–3.1%). Moreover, these mutations were non-overlapping between the two study sites on both sides of the border. Molecular evolutionary analysis detected signature of purifying selection on Pvk12. There is no direct evidence that Pvk12 is involved in artemisinin resistance in P. vivax, but it remains a potential candidate requiring further investigation. Continuous monitoring of potential drug resistance in this parasite is needed in order to facilitate the regional malaria elimination campaign.


Background
Although global malaria incidence and death rates have decreased in recent years, malaria remains one of the biggest threats to populations living within the tropical and sub-tropical world. In the Greater Mekong Sub-region (GMS) of Southeast Asia, a regional malaria elimination plan has been endorsed, aiming to eliminate Plasmodium falciparum malaria by 2025 and all malaria by 2030 [1]. This ambitious goal is met with challenges, especially the recently developed artemisinin resistance among P. falciparum populations in this region [2][3][4]. Further complication of the situation results from the increasing prevalence of Plasmodium vivax in this region, a parasite that is resilient to current control measures [5]. Although chloroquine (CQ) remains the primary treatment option for P. vivax infections, the emergence of CQ resistance has led to the use of artemisinin-based combination therapy (ACT) for the treatment of vivax malaria in Indonesia [6]. Monitoring drug resistance in P. vivax populations is important for eliminating all malaria from the GMS.
In regions co-endemic for P. falciparum and P. vivax, both species can share the same vectors and human hosts, and are often subject to similar forces of natural selection [7,8]. Several studies showed that anti-malarial drug pressure induces strong selection on both parasite species [9,10]. It is thus assumed that the widespread use of ACT for treating P. falciparum infection may exert similar collateral selective pressure on P. vivax populations. Recently, mutations in the propeller domain of the Pfk13 gene have been incriminated as important determinants of artemisinin resistance in P. falciparum [11][12][13]. Although ACT is still highly efficacious for treating vivax malaria [14,15], vigilance is required on potential emergence of resistance in this parasite [16]. A recent study aiming to determine whether similar mutations mediating artemisinin resistance are also present in the Pfk13 orthologue in P. vivax (here named Pvk12 gene as this gene is located on chromosome 12) found that parasites with a Pvk12 mutation (V552I) was circulating in Cambodia at a very low frequency [17]. To further examine the genetic diversity of this gene in the GMS, recent collections of clinical isolates of P. vivax from both sides of the China-Myanmar border were examined. This region has a very long history of artemisinin deployment and there is an indication that CQ efficacy for treating P. vivax malaria is decreasing [18]. Here Pvk12 genes in 262 P. vivax parasite isolates were sequenced and molecular evolution analyses were performed to determine whether this gene is subject to potential selection.

Collection of parasite clinical isolates
Plasmodium vivax samples were collected from malaria patients with acute P. vivax infections attending malaria clinics located on both sides of the China-Myanmar border. Plasmodium vivax infection was diagnosed by microscopy of Giemsa-stained thick and thin blood films. A total of 162 samples were obtained from 13 clinics around Laiza township, Kachin State, Myanmar during the high transmission season (June to September) in 2015. One-hundred samples were collected in 2012-2013 from Tengchong County, Yunnan Province, China. The standard treatment for P. vivax malaria is CQ (25 mg/ kg, divided into three days) and primaquine (0.25 mg/ kg/day for 14 days in Kachin State; 0.375 mg/kg/day for 8 days in Yunnan Province) [19]. On both sides of the border, malaria transmission has been reduced in recent years [20,21]. Especially, P. falciparum malaria incidence has experienced a sharp decline [22,23], possibly due to the extensive use of ACT. To collect finger-prick blood on filter papers, written informed consent was obtained from the participants or their guardians. This study was approved by the Institutional Review Board of Kunming Medical University.

DNA extraction, gene amplification and sequencing analysis
Genomic DNA was extracted using the High Pure PCR Template Preparation Kit (Roche, Germany) and eluted in 50 µl of H 2 O. Parasite samples were genotyped by PCR/ RFLP of two polymorphic genes msp3α and msp3β to distinguish single from mixed-strain infections [24,25]. Only infections containing single P. vivax strains by genotyping were used for sequencing analysis. The full-length Pvk12 gene was amplified by nested PCR using 2X TSINGKE ™ Master Mix (Beijing Tsingke Biotech, China) containing high-fidelity Pfu DNA polymerase with outer primers K13P1F (5′-CCATACTGGCTGCACCTGCTT-3′) and K13P1R (5′-GTAGTGGCAGTGGAGGAGAG-3′), and nested primers K13P2F (5′-CCACGGAACAGAT-GAATCTTC-3′) and K13P2R (5′-AAACCCGAGAAA-GTTGTAGCA-3′). PCR reactions were performed in 25 with 1 µl DNA template, 0.5 μM of each primer, 12.5 μl Master Mix, and 10.5 μl H 2 O. For the primary reaction the following cycling parameters were used: 5 min at 94 °C, 35 cycles at 94 °C for 30 s, 63 °C for 30 s, 72 °C for 90 s, and final extension for 7 min at 72 °C. For nested PCR, 1 μl primary PCR product was used as template and the following cycling parameters were used: 5 min at 94 °C, 35 cycles at 94 °C for 30 s, 57 °C for 30 s, 72 °C for 90 s, and final extension for 7 min at 72 °C. Sequencing was performed on both strands and sequences were assembled by using DNASTAR (WI, USA) with manual editing. Alignment of DNA sequences were performed using MEGA 6.0 [26] and BioEdit (version 7.2.5) with the Pvk12 sequence of Salvador I (Sal-I) genome (PVX_083080) as the reference. The five haplotypes of Pvk12 gene derived from this study have been submitted to GenBank under the accession numbers (KX961684-KX961688).

Molecular evolutionary analysis of Pvk12
A total of 333 full-length Pvk12 sequences were analysed, of which 262 samples were from the current study. The remaining 71 sequences including four from Oceania, 11 from North America, one from Africa, 34 from South America, and 21 from Southeast Asia were downloaded from PlasmoDB [27], as a result from the P. vivax hybrid selection sequencing project [28]. The sequences were aligned with the Sal-I reference sequence using ClustalW implemented in MEGA 6.0. SNP information was used to generate haplotypes and to calculate haplotype diversity. To determine the genetic diversity of Pvk12 gene, the average pairwise nucleotide diversity (π) and the average number of segregating sites (θ) were calculated. To detect signatures of natural selection, the number of non-synonymous substitutions per non-synonymous site (dN) and synonymous substitutions per synonymous site (dS) were estimated. A codon-based Z test was performed to calculate possible departure from neutrality with the variance of difference between dN and dS being computed by the bootstrap method with 500 replicates [29]. This neutrality test was further corroborated using Tajima's D and Fu's Fs tests. All molecular evolution tests were performed using MEGA 6.0. Pairwise linkage disequilibrium (LD) was also used to determine the degree of random association between different mutations within this gene. The correlation coefficient (R 2 ) between paired alleles was estimated using DNASP v5.0 and Arlequin 3.5, and the significance of each association was determined using the χ 2 test after Bonferroni correction.

Genetic diversity of Pvk12 from the two geographical regions
Two hundred and sixty two full-length Pvk12 sequences (2139 bp) in parasite samples from two regions along the China-Myanmar border were obtained. Compared with the Sal-I reference sequence, one synonymous SNP at nucleotide 2091 was observed in five (3.1%) isolates from the Kachin samples, while two different synonymous SNPs at nucleotide 516 and 1080, respectively, were identified in four (2%) isolates from the Tengchong samples (Table 1; Fig. 1). In addition, one non-synonymous mutation, M124I, was found in two isolates from Tengchong, whereas no non-synonymous mutation was identified in the Kachin samples. Altogether, the frequency of parasites containing Pvk12 mutations was very low, making up of only 4.2%. In comparison, 71 downloaded Pvk12 sequences were all the same as the wild type reference sequence. Consistent with this result, genetic diversity of Pvk12 gene represented by the π and θ values was extremely low for the China-Myanmar border samples (Table 2). Similarly, only five haplotypes were observed with the wild type as the predominant. Among the parasite isolates from Tengchong, four haplotypes were observed, giving a haplotype diversity of 0.116. The haplotypes of these two regions were completely different without overlap. Except for the wild type, other haplotypes containing single mutations all were very rare. Furthermore, pairwise LD analysis revealed no significant associations between different alleles (see Additional files 1, 2; P > 0.05, χ 2 test).

Signature of purifying selection
Several molecular evolutionary analyses were performed to assess the role of natural selection on Pvk12. When all samples were combined, both Tajima's D and Fu and Li's Fs were significantly deviated from neutrality ( Table 2) and suggested of purifying selection. The dN-dS analysis also showed higher prevalence of synonymous than nonsynonymous mutations, albeit the value was not statistically significant. Interestingly, when samples from the two regions were analysed separately, purifying selection on this gene was only evident in isolates from Tengchong, China (Table 2). Although the Tajima's D and Fu and Li's F values for the Kachin samples did not reach significance, they had the same trend as for the Tengchong samples.

Discussion
Mutations in the propeller domain of Pfk13 gene have been associated with artemisinin resistance in P. falciparum, which is manifested as delayed parasite clearance after artemisinin treatment [11,12]. A recent analysis of worldwide P. falciparum populations identified numerous

Point mutations Prevalence [N (%)]
Nucleotide Amino acid Kachin (N = 162) Tengchong (N = 100) non-synonymous mutations with marked geographic disparity in their frequency and distribution, which may reflect the demographic history of the parasite populations and different drug use histories [30]. In sharp contrast, studies conducted to date in the GMS, where artemisinin resistance has emerged in P. falciparum, showed a highly conserved Pvk12 gene [17,31]. In Cambodia where Pfk13 mutations were highly prevalent, analysis of 284 P. vivax isolates identified only two isolates with a non-synonymous mutation, V552I [17], although these parasites may have been exposed to higher drug selection pressure. Similarly, analysis of 66 P. vivax samples from central China identified a G581R mutation in one sample (Wang et al. unpublished). In comparison, this analysis of 262 full-length Pvk12 sequences only identified that two parasite isolates (2%) collected from Yunnan's Tenchong County contained a non-synonymous mutation, M124I, which is located near the N-terminal region. Thus, mutations identified here, being outside of the propeller domain, are unlikely involved in artemisinin resistance even if Pvk12 is presumed to play an analogous role to Pfk13 in resistance. Furthermore, although the recent genomic studies of world P. vivax populations identified five non-synonymous mutations in Pvk12 (N25I, S253T, V552I, V652L, and A710E) in 228 samples, all of these mutations were rare mutations [31]. Specifically, the N25I mutation was identified only in 2% of the western Thai samples (N = 89), whereas the S253T mutation was found only in 8% of the Papua Indonesia samples (N = 55), and each of the three other mutations was present only in one sample. Though some of these mutations are located in the propeller domain, their functions still need to be evaluated individually as mutations within the propeller domains such as A578S in Pfk13 may not necessarily be associated with artemisinin resistance [12]. The observed deficiency of mutations in the Pvk12 gene is in sharp contrast to Pfk13 from the sympatric P. falciparum populations [11,30]. While it is unknown whether Pvk12 and Pfk13 are functionally equivalent, the high level of conservation of Pvk12 suggests that (1) Pvk12 may play a different role in parasite biology even if Pvk12 has a similar function to Pfk13 and its conservation is maintained by functional constraints (purifying selection) and, (2) if Pvk12 is involved in artemisinin resistance, the different mutation rates in Pfk13 and Pvk12 may suggest different drug selection pressures on these two parasites. Likewise, there is evidence that pfk13 is under natural selection as non-synonymous mutations exceed synonymous mutations [32]. In contrast, Pvk12 appears to be under purifying selection, though the limited polymorphisms at four positions of Pvk12 in 11 of 262 samples analyzed here preclude a robust evolutionary analysis. To date the functions of Plasmodium k13 orthologues are still unknown, but their similarity in sequence to the human KEAP1 protein, which is involved in the regulation of the antioxidant responses [33], suggest that they might have similar functions. In P. falciparum, it was found that the Pfk13 is linked to the phosphotidylinositol-3-phosphate kinase (PI3K) pathway and artemisininconferring Pfk13 mutations are associated with increased PI3K activity and higher levels of the product phosphatidylinositol-3-phosphate [34]. If a similar pathway linking k13 and PI3K works in both parasite species, drug selections may be imposed on different enzymes within the same pathway. Interestingly, divergent selections have been identified on the phosphoinositide-dependent kinase PDK-1 among P. vivax populations, an enzyme participating in the PI3K pathway [28]. It is unknown whether this observation suggests that artemisinin family drugs might have acted on the PvPDK-1 gene instead of Pvk12 gene in P. vivax. It is also possible that the highly conserved Pvk12 gene among worldwide P. vivax populations may suggest that Pvk12 may have nothing to do with artemisinin resistance, like that the pvcrt-o gene is not a major player in CQ resistance in P. vivax as compared to the critical role of its orthologue pfcrt in mediating CQ resistance in P. falciparum.
Mutations within Pvk12, though rare, displayed quite significant geographical disparities. Three different non-synonymous mutations have been identified in parasite populations from different parts of the GMS [17,31]. Considering evidence of independent emergence of artemisinin resistance-conferring Pfk13 mutations in different areas of the GMS [35], this difference in Pvk12 may indicate different demographic histories and the presence of gene flow barriers among these parasite populations. It could also reflect different drug use histories for P. falciparum malaria, which may have exerted collateral selection on P. vivax, given that the front-line treatment for P. vivax malaria in most GMS countries remains as CQ and primaquine. Nevertheless, the collateral selection by artemisinins in the case of mixed P. falciparum/P. vivax infections that are treated by ACT should be minimal. Plasmodium vivax parasites reappearing after ACT treatment may be the relapsing parasites resulted from the awakening hypnozoites, and thus further transmission probably will be from these unselected parasites. In this regard, future surveillance on artemisinin resistance in P. vivax may need to focus on areas such as Indonesia where ACT has replaced CQ for treatment of vivax malaria, though recent analysis of a limited number of samples did not detect mutations in the propeller domain of Pvk12 in the Papua Indonesia samples [31].

Conclusions
Pvk12 only displayed very limited genetic diversity in parasite populations from the China-Myanmar border area, as well as from other regions of the GMS, where artemisinin resistance is P. falciparum has emerged recently. Pvk12 is not implicated in drug resistance nor is it recommended for molecular surveillance for drug resistance.