Genetic diversity of Plasmodium vivax and Plasmodium falciparum lactate dehydrogenases in Myanmar isolates

Background Plasmodium lactate dehydrogenase (pLDH) is a major target in diagnosing the erythrocytic stage of malaria parasites because it is highly expressed during blood-stage parasites and is distinguished from human LDH. Rapid diagnostic tests (RDTs) for malaria use pLDH as a target antigen; however, genetic variations in pLDH within the natural population threaten the efficacy of pLDH-based RDTs. Methods Genetic polymorphisms of Plasmodium vivax LDH (PvLDH) and Plasmodium falciparum LDH (PfLDH) in Myanmar isolates were analysed by nucleotide sequencing analysis. Genetic polymorphisms and the natural selection of PvLDH and PfLDH were analysed using DNASTAR, MEGA6, and DnaSP ver. 5.10.00 programs. The genetic diversity and natural selection of global PvLDH and PfLDH were also analysed. The haplotype network of global PvLDH and PfLDH was constructed using NETWORK ver. 5.0.0.3. Three-dimensional structures of PvLDH and PfLDH were built with YASARA Structure ver. 18.4.24 and the impact of mutations on structural change and stability was evaluated with SDM ver. 2, CUPSAT and MAESTROweb. Results Forty-nine PvLDH and 52 PfLDH sequences were obtained from Myanmar P. vivax and P. falciparum isolates. Non-synonymous nucleotide substitutions resulting in amino acid changes were identified in both Myanmar PvLDH and PfLDH. Amino acid changes were also identified in the global PvLDH and PfLDH populations, but they did not produce structural alterations in either protein. Low genetic diversity was observed in global PvLDH and PfLDH, which may be maintained by a strong purifying selection. Conclusion This study extends knowledge for genetic diversity and natural selection of global PvLDH and PfLDH. Although amino acid changes were observed in global PvLDH and PfLDH, they did not alter the conformational structures of the proteins. These suggest that PvLDH and PfLDH are genetically well-conserved in global populations, which indicates that they are suitable antigens for diagnostic purpose and attractive targets for drug development.

diagnostic methods, microscopic examination of blood smear still remains the gold standard for malaria diagnosis [2]. This method can identify the Plasmodium species and quantify the parasitaemia level at a low cost [1]. However, it is laborious, time-consuming, and requires well-trained and highly qualified technicians. Also, misdiagnoses and incorrect species identification can occur in cases of low parasitaemia, leading to incorrect treatment [3]. To overcome these disadvantages, several alternative methods for malaria diagnosis have been developed.
pLDH is an enzyme essential to a Plasmodium parasite's survival by mediating parasite's anaerobic glycolysis, and it is expressed at high levels during the blood stage [13,14]. Moreover, the level of pLDH in the blood is directly correlated to the parasitaemia level. Due to these advantages, pLDH has been widely examined as a promising diagnostic antigen and has been adopted in many RDTs for malaria. However, few studies have examined pLDH genetic polymorphisms in Plasmodium isolates and the potential influence pLDH genetic variations have on the diagnostic performance of pLDH-based RDTs [15][16][17]. In this study, the genetic diversity of P. falciparum LDH (PfLDH) and P. vivax LDH (PvLDH) in Myanmar parasite isolates was investigated. Moreover, the genetic diversity of global PfLDH and PvLDH populations was also analysed to better understand the genetic structures of the two genes in global P. falciparum and P. vivax populations.

Blood samples and ethics
The blood samples used in this study were obtained from P. vivax and P. falciparum infected patients with symptomatic uncomplicated cases who resided in towns and villages located in Naung Cho and Pyin Oo Lwin townships in Upper Myanmar in 2015. Malaria infection was confirmed by Giemsa-stained thick and thin blood smear examinations. Prior to treatment, 2 ml of venous blood was collected from each confirmed P. falciparum and P. vivax infected patients and placed into ethylenediaminetetraacetic acid (EDTA) tubes for further molecular analysis. All P. falciparum and P. vivax positive samples were further confirmed with polymerase chain reaction (PCR) targeting the 18S ribosomal RNA (rRNA) gene [18,19]. The use of blood samples in this study was approved by the Ministry of Health, Myanmar (97/Ethics 2015) and by the Biomedical Research Ethics Review Board of Inha University School of Medicine, Republic of Korea (INHA 15-013). Written consent was obtained from each individual prior to blood collection.

Amplification and sequencing analysis of the PvLDH and PfLDH
The genomic DNA of parasite was extracted from 200 µl of whole blood using the QIAamp DNA Blood Kit (Qiagen, Hilden, Germany) following the manufacturer's instructions. PvLDH and PfLDH were amplified using PCR with primer sets described previous [15]. Briefly, the forward and reverse primers used to amplify PvLDH were 5′-ATG ACG CCG AAA CCC AAA AT-3′ and 5′-ACC TTT AAA TGA GCG CCT TCAT-3′. The primers for PfLDH amplification was 5′-AGA TGG CAC CAA AAG CAA AAAT-3′ and 5′-ACC TTT AAG CTA ATG CCT TCAT-3′. The thermal cycling parameters for the PCR were as follows: denaturation at 94 °C for 5 min; 30 cycles of 94 °C for 1 min, annealing at 56 °C for 1 min, and extension at 72 °C for 1 min; and a final extension at 72 °C for 10 min. To minimize nucleotide mismatches during amplification, Ex Taq DNA polymerase (Takara, Otsu, Japan), which has proofreading activity, was used in all PCR procedures. Each PCR product was analysed by electrophoresis on 1% agarose gels. The resulting PCR product was purified from the gel and cloned into the T&A cloning vector (Real Biotech Corporation, Banqiao City, Taiwan). Each ligation mixture was transformed into Escherichia coli DH5α competent cells and positive clones with the appropriate insert were screened by colony PCR. The nucleotide sequences of the cloned PCR product were analysed by automatic sequencing using the BigDye ™ Terminator Cycle Sequencing Kit (Applied Biosystems, Foster City, CA, USA) on an ABI3730XL DNA analyzer (Applied Biosystems) with M13 forward and reverse primers. At least two clones from each isolate were sequenced to ensure sequencing accuracy, and some isolates underwent three or fourfold sequence coverage to confirm the existence of rare polymorphisms. The nucleotide sequences reported in this study have been deposited in the GenBank database under the accession numbers KU869730-KU869766 and KX885908-KX885922 for PfLDH and KU895512-KU895548 and KX885923-KX885934 for PvLDH.

Nucleotide sequence polymorphism analysis and neutrality test
The nucleotide and deduced amino acid sequences of Myanmar PfLDH and PvLDH were analysed using Edit-Seq and SeqMan programs in the DNASTAR software package (DNASTAR, Madison, WI, USA). Nucleotide sequence polymorphism analysis was conducted for the 52 PfLDH and 49 PvLDH sequences. The number of segregating sites (S), the number of haplotypes (H), haplotype diversity (Hd), nucleotide diversity (π), and the average number of pairwise nucleotide differences within the population (K) were estimated using DnaSP ver. 5.10.00. [20]. The π was calculated to estimate the stepwise diversity throughout Myanmar PfLDH and PvLDH based on a sliding window of 50 base pairs (bp) with a step size of 10 bp. The values of synonymous (dS) and non-synonymous (dN) substitutions were estimated and were compared using the Z-test (P < 0.05) in the MEGA4 program [21] using Nei and Gojobori's method [22] with the Jukes and Cantor correction. The Tajima's D value [23] and Fu and Li's D and F statistics [24] were analysed using the DnaSP ver. 5.10.00 to evaluate the neutral theory of evolution [20]. Sliding window plot analysis was also performed to analyse the step-wide value of Tajima's D throughout the Myanmar PfLDH and PvLDH based on a sliding window of 100 bp with a step size of 25 bp. The recombination parameter (R), which included the effective population size and probability of recombination between adjacent nucleotides per generation, and the minimum number of recombination events (Rm) were determined using DnaSP ver. 5.10.00 [20].

Genetic diversity of PvLDH and PfLDH in global Plasmodium falciparum and Plasmodium vivax populations
The genetic diversity of global PfLDH and PvLDH were analysed. The publicly available sequences of global PvLDH and PfLDH were used in this study (Additional files 1 and 2: Tables S1 and S2). Nucleotide sequence polymorphism analysis and a neutrality test for global PvLDH (n = 100) and PfLDH (n = 334) population were estimated using DnaSP ver. 5.10.00 [20] and MEGA4 program [21] as described above. To investigate the genetic relationships among global PfLDH or PvLDH haplotypes, the haplotype network for global PfLDH (n = 334) and global PvLDH (n = 100) was separately analysed using NETWORK ver. 5.0.0.3 with the Median Joining algorithm [25].

Homology modelling and mutant analysis of PvLDH and PfLDH
The structural effects caused by amino acid changes identified in global PfLDH and PvLDH were analysed. The default protocol in the YASARA Structure ver.18.4.24 [26] was applied to construct three-dimensional (3D) homology models of PvLDH and PfLDH using the 'hm_ build.mcr' macro file (http://www.yasar a.org/hm_build .mcr). To construct homology models, PSI-BLAST [27] was conducted against PDB entries [28]. The homology models were built based on each template and then were refined using high-resolution energy minimization of the YASARA force field [29]. The best homology model based on the PfLDH template (PDB code: 1T2D) [30] was chosen among the reliable models. The characteristics of the model were evaluated using a Ramachandran plot [31], ERRAT [32] and ProSA [33]. The crystal structure of the PfLDH model (PDB code: 1T2D) [30] was retrieved from the RCSB Protein Data Bank [34]. To determine the impact of mutations on the structural stability, thermodynamic changes in wild-type and mutant structures of PvLDH and PfLDH were predicted using SDM ver. 2 [35] by estimating a stability score. All mutant models were generated using Andante [36] as implemented in SDM ver. 2. CUPSAT [37] and MAESTROweb [38] and were also used to improve the overall prediction accuracy of mutations under consideration. All structural graphics were visualized using an open-source code of PyMOL (The PyMOL Molecular Graphics System, ver. 1.7.2.1, Schrödinger, LLC) in Linux. Structural superpositions onto the wild-type structure were performed and visualized using the 'cealign' method script (https ://pymol wiki. org/index .php/Ceali gn) in PyMOL v1.7.2.1.

Genetic polymorphisms of Myanmar PvLDH
A total of 49 PvLDH sequences were obtained from Myanmar P. vivax isolates. All of the sequences contained 951 nucleotides encoding 316 amino acids, indicating no insertion or deletion existed in the analysed sequences. When the sequences were compared with Sal I PvLDH (GenBank ID: XM_001615570), 45 single nucleotide polymorphisms (SNPs) were identified at 29 positions. Thirty-one of the SNPs were synonymous, and the other 14 were non-synonymous. The non-synonymous SNPs resulted in amino acid changes at 13 positions in Myanmar PvLDH, including 12 di-morphic amino acid changes (A20T, V37A, T62A,  Y67C, V76A, V133A, G180D, K185R, Y250H, I273T,  T277I, and T300A) and 1 tri-morphic amino acid change (K301M/E). Based on these amino acid polymorphic patterns, Myanmar PvLDH was classified into 13 distinct haplotypes (haplotype 1 to 13) (Fig. 1a).

Genetic polymorphisms of Myanmar PfLDH
Fifty-two Myanmar PfLDH sequences were obtained. Comparing those sequences with the reference 3D7 PfLDH sequence (GenBank ID: XM_001349953) revealed 18 SNPs at 17 positions. Of these, 12 SNPs were non-synonymous and resulted in di-morphic amino acid changes at 11 positions (Fig. 1b) 9 different haplotypes (haplotype 1 to 9) based on these amino acid polymorphic patterns (Fig. 1b).

Genetic diversity and neutrality of Myanmar PvLDH and PfLDH
Sequence analysis was performed to examine nucleotide diversity and neutrality of PvLDH  (Fig. 2). Sliding window analysis of Tajima's D values also showed similar patterns in PvLDH and PfLDH.

Genetic polymorphism of PvLDH in the global P. vivax population
Amino acid polymorphism patterns in global PvLDH (n = 100) were analysed. Compared to Sal I PvLDH (GenBank ID: XM_001615570), a total of 39 amino acid changes at 36 positions were identified in the global PvLDH population (Fig. 3). However, these amino acid changes were not commonly identified in global PvLDH, rather they differed by country. The amino acid changes found in Myanmar PvLDH were not identified

Genetic diversity and neutrality of global PvLDH and PfLDH
To further examine whether natural selection had contributed to PvLDH and PfLDH diversity within the global P. vivax and P. falciparum populations, publicly available sequences deposited in the NCBI and PlasmoDB databases were analysed (  (Table 2). Sliding window plot analyses of nucleotide diversity (π) and Tajima's D of the global PvLDH and PfLDH also revealed low genetic diversity and negative selection (Fig. 5). A haplotype network analysis of global PvLDH (n = 100) showed 35 closely linked haplotypes (Fig. 6). Two major haplotypes (H1 and H8) were identified, and the other haplotypes were singletons arising from H1 to H8. The most prevalent haplotype was haplotype 1 (H1), which shared the same nucleotide sequences as Sal I (GenBank ID: XM_001615570) with a frequency of 38.0%. The frequency of haplotype 8 (H8) was 29.0%. A haplotype network analysis of global PfLDH (n = 334) revealed a total of 22 distinct haplotypes (Fig. 7). The most prevalent haplotype was H4, which had the identical nucleotide sequences with 3D7 (GenBank ID: XM_001349953) and was shared by populations from different countries with a frequency of 83.2%. Haplotype 5 (H5) was the second most prevalent haplotype (frequency of 10.4%) and was mainly comprised of African PfLDH. Nineteen haplotypes, except for haplotype 6 (H6), were singletons. These haplotypes were linked to the most frequent H4 by short or long branches. Similar to global PvLDH, haplotype network analysis of global PfLDH revealed a simple   18  19  20  29  37  48  62  67  69  76  78  81  89  90  105  114  117  120  125  133  136  160  180  185  186  195  212  226  250  252  273  277  300  301  310  312  cluster with a small number of haplotypes having low frequencies separated by short branches.

Structural impacts caused by amino acid changes in global isolates of PvLDH and PfLDH
The structural effects of amino acid changes found in global isolates of PvLDH and PfLDH were investigated. Several PvLDH crystal structures (PDB codes: 5HRU at 1.71 Å, 5HTO at 1.90 Å and 2A92 at 2.04 Å) were reported previously (Additional file 3: Table S3), but their resolutions were not high, Accuracy of the template structure such as resolution should be carefully considered to construct the most reliable homology model [39]. Therefore, a homology model of PvLDH was constructed based on a high-resolution crystal structure of PfLDH (PDB code: 1T2D at 1.1 Å resolution), due to the low-resolution crystal structures for PvLDH. The constructed PvLDH model had 94.1% of the amino acid residues in a favorable region according to the Ramachandran plot (Additional file 4: Fig. S1a), a ProSA Z-score of − 9.82 (Additional file 4: Fig. S1b) and an ERRAT overall quality score of 96.74 (Additional file 4: Fig. S1c). Superposition between PvLDH and PfLDH showed a root-mean-square deviation (RMSD) of 0.06, suggesting the two proteins shared the same overall folding pattern (Additional file 5: Fig. S2a). The models predicted a Rossmann fold in all pLDH enzymes and the NADH cofactor-binding site [40,41] that forms a series of alternating 6 β-strands and 5 α-helices (Additional file 5: Fig. S2b). Furthermore, PvLDH and PfLDH shared identical amino acid sequences at catalytic residues (R95, D155, R158, and H182) [42], the active site (K84), cofactor binding sites (P235 and P239) [43], and the substrate-specific loop (DKEWN, amino acid positions 90-94) [44] (Additional file 5: Fig. S2a). When all the amino acid changes found in global PvLDH and PfLDH were applied to the models, it suggested that they were scattered throughout the structures (Figs. 8   25  29  48  87  112  121  136  151  183  186  187  188  191  238  257  259  264  272  295   3D7  Q  G  T  G  H  F  L  G  G M  V  L  R  A  V  I    and 9). The structural impact of all non-synonymous amino acid substitutions of PvLDH and PfLDH were analysed based on the RMSD and the predicted ΔΔG (kcal/mol) values between the wild-type and mutants. Superposition between PvLDH wild-type and 40 mutants generated an RMSD of 0.001-0.059 (Additional file 6: Table S4). PfLDH wild-type and 19 mutants generated an RMSD of 0.001-0.053 (Additional file 7: Table S5). These results suggest that individual amino acid changes may not induce significant structural changes in both PvLDH and PfLDH. To improve the accuracy of potential stability changes (ΔΔG, kcal/mol) with each amino acid change, three different predictors (SDM ver. 2 [35], CUPSAT [37], and MAESTROweb [38]) were applied, but their coincidence results were inconclusive (Additional files 6 and 7: Tables S4 and S5). Although D90V/N of PvLDH was located in the substrate-specific loop, it is predicted not to significantly affect structural stability (Additional file 6: Table S4). These results collectively suggest that all amino acid changes identified in global PvLDH and PfLDH are unlikely to cause significant structural changes in the proteins.

Discussion
Analyses of genetic polymorphisms in Myanmar PvLDH and PfLDH revealed a low level of genetic diversity in Myanmar P. vivax and P. falciparum populations. Both genes showed synonymous SNPs and non-synonymous SNPs, which generated different haplotypes in each gene: 13 haplotypes in PvLDH and 9 haplotypes in PfLDH. Global PvLDH and PfLDH sequence analyses also suggested that amino acid changes caused by diverse nonsynonymous SNPs were identified among the global PvLDH and PfLDH. However, the frequencies of the amino acid changes were not high, and they were not evenly distributed among the global population but were country-specific. Overall genetic diversity was greater in PvLDH than PfLDH. This is consistent with previous findings on the genetic diversity of PvLDH and PfLDH in other countries [15-17, 42, 45]. These results are also consistent with a previous report of comparative genomic analyses suggesting more polymorphic events were identified in the P. vivax genome than P. falciparum [46].
Neutrality tests of Myanmar PvLDH and PfLDH indicated the influence of a strong purifying selection on PvLDH and PfLDH were also in agreement with the dN-dS values. Similar patterns of purifying selection were also identified in global PvLDH and PfLDH [17], suggesting the genetic conservation of global PvLDH and PfLDH was mostly maintained by purifying selection. This is probably due to functional constraints. Haplotype network analyses of global PvLDH and PfLDH were also consistent with purifying selection, each with one or a few modal haplotypes separated by short branches. A total of 39 amino acid changes at 36 positions were identified in the 100 global PvLDH sequences analysed in this study. Regarding PfLDH, 19 amino acid changes at 19 positions were observed in the 334 global sequences. The amino acid changes found in global PvLDH and PfLDH were not all located at essential residues, including catalytic residues (R95, D155, R158 and H182), the active site (K84), cofactor binding sites (P235 and P239) and the substrate-specific loop (DKEWN, amino acid positions 90-94) [47]. One exception to this is D90V/N, which was found in the substrate-specific loop of PvLDH. The mutation D90V/N, however, does not alter the protein's The two views are rotated with respect to each other by 180° around the vertical axis. The 19 amino acid changes identified in global PfLDH are indicated with orange sticks. The mutation residues, which were located on the external or internal regions, were depicted in bold or underlined, respectively. The functional residues of PfLDH sre depicted as colored spheres: catalytic residues (R95, D155, R158 and H182) in red, the active site (K84) and cofactor-binding site (P235 and P239) in blue, and the substrate-specific loop (D90-N94) in green structural stability since no prediction demonstrated a significant impact on protein stability. These results suggest that the amino acid changes in global PvLDH and PfLDH are unlikely to cause significant structural change that may disrupt the function of the enzymes. The 3D structure analyses also supported this notion by suggesting that no significant conformational change is likely to be caused by these amino acid variations.
Concerns have been raised about the genetic diversity found in pLDH and how it might influence the diagnostic efficacy of pLDH-based RDTs [48][49][50]. However, the results of this study suggest that the two genes were well-conserved in global populations even though low frequencies of amino acid changes were identified in the global population. It is unlikely that these polymorphic patterns identified in global PvLDH and PfLDH can influence the diagnostic performance of pLDHbased RDTs. In fact, pLDH-based RDTs performed well in filed performance assessments in India and other countries despite the occurrence of different genotypes [51][52][53]. Blood samples used in this study were also correctly diagnosed to each species by pLDH-based RDT kits (unpublished observations). These collectively suggest that epitopes of PvLDH and PfLDH targeted in RDTs are well-conserved in the global populations and therefore pLDH-based RDTs would be reliable to diagnose malaria in the field conditions. However, it would be noteworthy to mention that accuracy of RTDs can be affected by low parasitaemia and residual parasite antigens after clearance [4,6,54]. And therefore, continuous efforts to develop RDTs with higher sensitivity and specificity should be performed.
Functional and structural constraints of global PvLDH and PfLDH are also highly informative since pLDH is a promising drug target [55]. Plasmodial LDH is an essential enzyme in the central metabolic pathway of malaria parasites [14,[56][57][58]. Due to substantial differences in the human ortholog, pLDH is an attractive target for developing anti-malarial drugs [47,59]. Inhibition of pLDH prevents adenosine triphosphate (ATP) production, which kills malaria parasites [60]. The results obtained in this study may help to identify or design effective inhibitors for anti-malarial drugs targeting pLDH.
The limitation of this study is that only limited numbers of Myanmar and global PvLDH and PfLDH sequences were included in the analysis. To better understand the polymorphic nature of PvLDH and PfLDH, further analysis of PvLDH and PfLDH variations in the global P. vivax and P. falciparum populations would be necessary.

Conclusions
Low levels of genetic diversity, which may be affected by strong purifying selection, were identified in global PvLDH and PfLDH populations. Although non-synonymous SNPs that induce amino acid changes in global PvLDH and PfLDH were found, these amino acid changes were not commonly distributed in the global population, and their frequencies were low. Despite differences in the amino acid sequence, functionally important residues that maintain the structure and function of pLDH are well-conserved. These data collectively suggest that pLDH is a useful target molecule for RDT. Further examination of the genetic diversity of pLDH in diverse global P. vivax and P. falciparum populations with a larger number of isolates is necessary to better understand the polymorphic nature and evolutionary aspects of PvLDH and PfLDH. reaction; pLDH: Lactate dehydrogenase of Plasmodium spp; PvLDH: Lactate dehydrogenase of Plasmodium vivax; PfLDH: Lactate dehydrogenase of Plasmodium falciparum; π: Observed average pairwise nucleotide diversity; rRNA: 18S ribosomal RNA; RMSD: Root-mean-square deviation; S: Number of segregating sites.