Genetic polymorphism of merozoite surface protein-3 in Myanmar Plasmodium falciparum field isolates

Seventy-two blood samples were obtained from P. falciparum infected symptomatic patients who live in villages located in Naung Cho and Pyin Oo Lwin, Myanmar in 2015 (Additional file 1: Fig. S1). P. falciparum infection was confirmed by Giemsa-stained thick and thin blood smear examination. All P. falciparum positive samples were further confirmed by polymerase chain reaction (PCR) targeting 18S ribosomal RNA (rRNA) gene [21]. Finger-prick blood samples were collected from the patients prior to drug treatment, air-dried, and stored in individual sealed plastic bag at ambient temperature until use. Genomic DNA was extracted from the blood filters using the QIAamp DNA Blood Kit (Qiagen, Hilden, Germany) following the manufacturer’s instructions. Written informed consent was obtained from all patients before blood collection. The study protocol was approved by either the Ethics committee of the Ministry of Health, Myanmar (97/Ethics 2015) and the Biomedical Research Ethics Review Board of Inha University School of Medicine, Republic of Korea (INHA 15-013).

The full-length pfmsp-3 was amplified by a nested PCR method. The primers used in this study were designed as described previously [13]. The thermal cycling parameters for primary and nested PCRs were as follows: one cycle of initial denaturation at 95 °C for 5 min, 25 cycles of 94 °C for 1 min, annealing at 57 °C for 2 min and extension at 72 °C for 2 min, followed by a final extension at 72 °C for 5 min. Ex Taq DNA polymerase (Takara, Otsu, Japan) with proof- reading activity was used in all PCR amplification steps to minimize the nucleotide mis-incorporation. PCR products were resolved on a 1.2% agarose gel, and were visualized under ultraviolet (UV). Each PCR product was purified from the gel, and cloned into the T&A vector (Real Biotech Corporation, Banqiao City, Taiwan). Ligation mixture was transformed into Escherichia coli DH5α competent cells, and positive clones with appropriate insert were selected by colony PCR. The nucleotide sequence of cloned insert was analysed by automatic DNA sequencing with M13 forward and M13 reverse primers. Plasmids from at least two independent clones from each transformation mixture were sequenced in both directions to verify the sequence accuracy. The nucleotide sequences reported in this study have been deposited in the GenBank database under the accession numbers MN787600-MN787671.

The nucleotide and deduced amino acid sequences of pfmsp-3 were analysed using EditSeq and SeqMan in the DNASTAR package (DNASTAR, Madison, WI, USA). The pfmsp-3 sequences of 3D7 (XM_001347593) and K1 (U08851) were used as reference sequences. The value of singleton variable sites, parsimony informative sites, total number of mutations, segregating sites (S), average number pair-wise nucleotide difference (K), haplotype diversity (Hd), and nucleotide diversity (π) were calculated using DnaSP version 5.10.00 [22]. The π was also measured on sliding window plot of 10 bases with a step size of 5 bp to estimate the stepwise diversity across the sequences. The rate of synonymous (dS) and non-synonymous (dN) substitutions were estimated and compared by the Z-test (P < 0.05) in order to test the null hypothesis of strict neutrality of pfmsp-3 in MEGA6 program [23] using the Nei and Gojobori’s method [24] with the Juke and Cantor correction (1000 bootstrap replications). Tajima’s D test [25], and Fu and Li’s D and F statistics [26] were performed to evaluate the neutral theory of natural selection using DnaSP version 5.10.00 [22]. The Tajima’s D values were also calculated on sliding window plot of 10 bases with a step size 5 bp to estimate the natural selection parameter across the gene.

The genetic diversity of global pfmsp-3 from different P. falciparum populations was analysed. The pfmsp-3 sequences included in this study were from Thailand (AM161544-AM161593), India (HM568669-HM568724), Kenya (KU527019-KU527058), and Nigeria (AM161594-AM161644). These sequences covered full-length or partial of pfmsp-3. Genetic polymorphism and tests of neutrality of each population were analysed or calculated using DNASTAR package, DnaSP version 5.10.00 [22] and MEGA6 [23] as describe above. A logo plot was constructed for each pfmsp-3 population to analyse the polymorphic patterns in the B cell epitopes using the WebLogo program (https://​weblogo.​berkeley.​edu/​logo.​cgi).

Seventy-two pfmsp-3 were successfully amplified from Myanmar P. falciparum isolates. Sequence alignment analysis of the sequences suggested that they were classified into 2 allelic types, 3D7 type (n = 43) and K1 type (n = 29). For the 3D7 allelic types, 164 single nucleotide polymorphisms (SNPs) were identified at 87 polymorphic sites, in which 122 SNPs were non-synonymous mutations, resulting in amino acid substitutions at 49 positions. Four amino acid positions showed tri-morphic changes (F12L/S, D142E/G, Y192C/H, and D278E/M) and the others were di-morphic amino acid changes. The most predominant amino acid substitutions identified in Myanmar pfmsp-3 were L68S and D278E/M, which accounted for 90.7% and 58.1% in frequency, respectively. The other amino acid changes occurred at very low proportions. Amino acid changes at 25 positions (G71S, E98G, T117A, D142E/G, A163V, Y192C/H, K236E, E245G, E249G, E253K, S261T, D263E, E267V, E274R, E275K, E276K, N277K, D278M, K281E, K285T, N311D, K331E, G332E, Q335R, and V344I) were novel ones that have not been reported previously. The glutamate rich domain of Myanmar pfmsp-3 was varied because of different repeat numbers of glutamate residues, ranging from 4 to 8 repeats. Deletions of E271 and/or E272 were identified in 12 distinct haplotypes and insertion of 2 additional glutamates was found in haplotype 26. Based on these polymorphic patterns, the 3D7 allelic types of Myanmar pfmsp-3 were further classified into 31 distinct haplotypes, in which haplotype 9 had the highest prevalence (23.3%) (Fig. 1). In the K1 allelic types, a total of 25 distinct haplotypes were identified. Haplotype 21 was the most prevalent, accounting for 17.2%. Sequence analysis revealed that there were 944 SNPs at 117 polymorphic sites, among which 576 SNPs were non-synonymous ones leading to amino acid substitutions at 75 positions. Sixty-nine di-morphic and 6 tri-morphic (P72A/S, T81C/S, K110N/Q, Q114D/E, K188Q/R, A352T/V) amino acid changes were found. High frequency of a glutamate (E) insertion at position 79 was identified at AHR domains. Insertions of 13 amino acids (ETEEEELEEKNEE) and 1–3 glutamate (E) were also detected at positions 280 and 298 in the glutamate rich domain, respectively. Amino acid deletions at 6 positions (G169, E301, N302, E303, K304, and K305) were also found in the K1 allelic types of Myanmar pfmsp-3 (Fig. 2).

Comparative analysis of genetic polymorphism of pfmsp-3 from Myanmar and other different geographical areas was performed. Given the partial sequence information of pfmsp-3 from other global populations, analysis focused on the 3 regions corresponding to AHR, the glutamate rich domain and the leucine zipper domain. For 3D7 allelic types, most amino acid polymorphisms were accumulated at glutamate rich domains, but the frequencies were different by country (Fig. 3). The deletion of E272 (E272-) and an amino acid change (D280E) were found in all analysed populations. Insertions of 1 to 3 glutamate residues at the position of 272 (E272 + , E272 ++, and E272 +++) were more prevalent in African pfmsp-3 than in Asian pfmsp-3. The AHR and leucine zipper domains were relatively well conserved in 3D7 allelic types of global pfmsp-3. Amino acid changes at 9 positions were identified in AHR domains of global pfmsp-3, most of them only in Myanmar pfmsp-3. Only 2 amino acid changes (A103S and K164E) were found in Thailand and Kenya pfmsp-3. The five amino acid substitutions (N311D, K331E, G332E, Q335R, and V344I) in the leucine zipper domain were detected only in Myanmar pfmsp-3. Polymorphic patterns of amino acid changes were more diverse and greater in global K1 allelic types (Fig. 4). Most amino acid changes detected at AHR domains were shared by the global pfmsp-3 population with high frequencies. Only five amino acid substitutions, G75I, T81C, S98F, T165A, and K188R, were uniquely identified in Myanmar pfmsp-3. Although the overall patterns of amino acid polymorphisms identified in global pfmsp-3 were similar, the frequencies of A112S, Q114E, Q117E, S119A, D121E, E123V, K124M, and E130K were relatively higher in Asian pfmsp-3 than in African pfmsp-3. The glutamate rich domain also showed polymorphic patterns in global pfmsp-3. Insertions of glutamate (E) at position 79, 13 amino acids (ETEEEELEEKNEE) at position 280, and glutamate (E) at the position 298 were found in pfmsp-3 from Thailand, Kenya, and Nigeria. Deletion of glutamate (E) at position 297 was also identified in pfmsp-3 from Kenya and Nigeria. Meanwhile, the leucine zipper domain was well-conserved in global pfmsp-3. Only 5 minor amino acid changes (N335S, N336S, T350A, A352T, and A352V) were uniquely identified in Myanmar pfmsp-3. Polymorphic patterns of the B-cell epitopes of 3D7 allelic types of global pfmsp-3 were analysed. The amino acid residues constructing the B-cell epitopes were tightly conserved in global pfmsp-3 allelic types (Fig. 5). However, di-morphic and tri-morphic amino acid changes at the corresponding residues were detected in Myanmar pfmsp-3.

The genetic diversity and tests of neutrality of Myanmar pfmsp-3 were analysed. The average number of nucleotide differences (K) for 3D7 and K1 allelic types were 5.30 and 33.22, respectively. The 3D7 and K1 allelic types showed similar values of haplotype diversity (Hd, 0.990 ± 0.009 for 3D7 and 0.990 ± 0.013 for K1), but nucleotide diversity (π) of K1 allelic types (0.0297 ± 0.0019) was greater than that for 3D7 allelic types (0.0050 ± 0.0006) (Table 1). The sliding window plot of π across Myanmar pfmsp-3 revealed that the highest values of π were identified at AHR domains in the K1 allelic type. In contrast, the highest values of π were detected at the glutamate rich domain for the 3D7 type. The leucine zipper domains showed low values of π in both allelic types of Myanmar pfmsp-3 (Fig. 6a). The dN–dS values of Myanmar pfmsp-3 were negative for both 3D7 (− 0.007 ± 0.002) and K1 (− 0.016 ± 0.009), suggesting that both allelic types of Myanmar pfmsp-3 were affected by negative natural selection. The value of Tajima’s D for the 3D7 allelic types was negative (− 2.692, P < 0.001), suggesting that 3D7 allelic types of Myanmar pfmsp-3 were under purifying selection. Meanwhile, the value was positive for the K1 allelic type (0.373, P > 0.1), implying that the K1 type was under balancing selection (Table 1). The sliding window plot analysis of Tajima’s D across Myanmar pfmsp-3 also indicated that the values were negative across the entire gene for the 3D7 allelic types, whereas positive values of Tajima’s D were predicted at AHR domains and the leucine rich region of K1 allelic types (Fig. 6b).

Influence of natural selection on the genetic diversity of the global pfmsp-3 populations was analysed. Nucleotide diversity and patterns of natural selection predicted in 3D7 and K1 allelic types were different by country (Table 1). In the 3D7 allelic types, the values of nucleotide diversity (π) were similar for the pfmsp-3 of each country, ranging from 0.0030 to 0.0050. The dN–dS value was negative for all countries analysed, indicating that negative natural selection may occur in global pfmsp-3. However, the value of Tajima’s D was different by country: negative for Myanmar and Kenya, but positive for Nigeria and Thailand. The values of Fu and Li’s D and F for each country also showed a similar trend of negative or positive values. Meanwhile, more complex patterns of natural selection were predicted in K1 allelic types (Table 1). The values of nucleotide diversity (π) of K1 allelic types were greater than those of 3D7 allelic types in all countries analysed. The values ranged from 0.0230 to 0.0420. The value of dN–dS was positive in Kenya, but in other countries including Myanmar, Nigeria, and Thailand was negative. Unlike the 3D7 allelic types, the values of Tajima’s D of all countries were positive, indicating they were under balancing selection. The values of Fu and Li’s D and F were also positive in all three countries except for Myanmar. These collectively suggested that the overall genetic diversity in the K1 allelic types was greater than in the 3D7 allelic types in global pfmsp-3. Both allelic types of global pfmsp-3 may be affected by natural selection, but the direction of natural selection differed by country.

The value for the minimum number of recombination events between adjacent polymorphic sites (Rm) for the 3D7 allelic types of Myanmar pfmsp-3 was 1 (Table 2). Potential recombination sites were predicted at the positions between 203 and 906. In the K1 allelic types, the Rm value was 10, in which recombination events were predicted at the positions 121-205, 219-231, 231-333, 345-387, 420-437, 441-505, 505-515, 563-571, 571-587, and 657-1026 (Table 2). The LD index R² plots for both 3D7 and K1 allelic types of Myanmar pfmsp-3 declined across the analysed regions, suggesting that intragenic recombination could be a factor contributing to the genetic diversity of pfmsp-3 found in Myanmar P. falciparum population (Fig. 7).