Background
Plasmodium vivax Duffy binding protein (PvDBP) is one of the erythrocyte-binding proteins, which belongs to the large erythrocyte binding protein family [
1]. PvDBP is expressed on the merozoite of
P. vivax and plays an essential role in erythrocyte invasion of the parasite by mediating irreversible binding with its corresponding receptor, the duffy antigen receptor for chemokines (DARC), on the surface of erythrocytes [
1‐
4]. Similar to other plasmodial proteins known to participate in such processes, PvDBP is suggested to be an important vaccine candidate antigen, because it elicits strong immune responses in humans [
5,
6]. Experimental evidences that antibodies against PvDBP inhibit the interaction of this protein with DARC
in vitro and block the invasion of
P. vivax into human erythrocytes also support the notion that this protein is a potential asexual blood stage vaccine candidate antigen against
P. vivax[
7‐
9].
PvDBP is divided into seven regions (regions I-VII), and the amino terminal cysteine-rich region, region II (PvDBPII), contains the central binding motifs necessary for adherence to DARC [
10‐
12]. Critical binding motifs in PvDBPII have been mapped to a 170 amino acid stretch (amino acids 291-460), which includes cysteines 5-8 [
11,
12]. PvDBPII shows the highest genetic diversity compared to the remaining PvDBP regions and appears to be under strong selective pressure [
13,
14]. Analysis of genetic variation of PvDBPII among
P. vivax field isolates from different geographical regions, including Brazil, Colombia, South Korea, Papua New Guinea, Thailand, showed that the PvDBPII is highly polymorphic, but the cysteine residues are conserved within and between
P. vivax populations from different geographic regions [
14‐
21]. Although it has been suggested that these polymorphisms do not significantly alter host-parasite binding [
17,
22], some of them alter immune recognition of PvDBP [
23] and most of the PvDBP-specific antibodies detected in infected individuals recognize PvDBPII, rather than other PvDBP regions [
7,
15]. Consequently, the polymorphic nature of PvDBP, particularly PvDBPII, represents a major impediment to the successful design of a protective vaccine against vivax malaria [
14]. Therefore, understanding the nature and genetic polymorphism in PvDBPII among
P. vivax isolates from distinct geographic areas, particularly where a large proportion of
P. vivax infections occurs, is important for the rational design of vaccines against vivax malaria.
In this study, the genetic polymorphism and natural selection of PvDBPII among P. vivax isolates from Myanmar were analysed. These results suggest that excessive polymorphism of PvDBPII is found in the filed isolates of P. vivax in Myanmar.
Discussion
Malaria is endemic or hypoendemic in Myanmar and is characterized by the occurrence of all four human-infecting
Plasmodium species [
32]. Although morbidity and mortality rates due to malaria have been declining gradually in recent years, Myanmar still contributes to approximately 60% of malaria deaths in the Southeast Asia [
32]. Genetic polymorphisms in the circumsporozoite protein, merozoite surface protein-1 (MSP-1), MSP-3α, and apical membrane antigen-1 (AMA-1) of
P. vivax Myanmar isolates were analysed previously [
24,
33,
34]. As expected, they showed high levels of genetic polymorphisms, but information on the nature and extent of population diversity within malaria parasites in Myanmar is still limited. In this study, genetic polymorphism and natural selection of PvDBPII in Myanmar
P. vivax isolates were analysed to expand our knowledge on population diversity of the parasite in the country.
A total of 54 PvDBPII sequences were obtained from Myanmar
P. vivax isolates. The sequence analysis revealed that the 54 sequences were classified into 12 different haplotypes. A GenBank search for each haplotype revealed that two haplotypes were identical to at least one previously reported PvDBPII sequence found in other regions of the world, whereas the other 10 haplotypes were novel and have not been reported so far. The sequence analysis revealed that a total of 32 point mutations was identified, which resulted in significant amino acid changes (28 non-synonymous and four synonymous changes) through the PvDBPII in Myanmar isolates. Seventeen of the 28 non-synonymous changes were previously reported, whereas the other 11 changes (I310L, F344S, R391H, K455I, K473R, C477G, R490K, D528G, V533M, K541T, and A545V) were unique to Myanmar isolates, which did not hitherto identified. The two highest peaks of nucleotide diversity within PvDBPII in the Myanmar
P. vivax isolates were identified between C5 and C7, which was consistent with previous observations [
14,
15,
17]. Interestingly, a unique amino acid change (C477G caused by TGT to GGT) was identified in three Myanmar isolates. It is well known that the cysteine residues within PvDBPII are well conserved within and between
P. vivax populations from different geographic regions studies [
14‐
21,
31]. It is not certain the nature of this amino acid change. It can be assumed that this mutation may resulted from a PCR or sequencing error, but to eliminate any possible errors, high quality Taq polymerase with a proof-reading function was used in all amplification processes and sequencing reactions were performed on at least two individual clones for each gene in both directions. But, considering the small number of isolates used in this study, further analyses with a large number of isolates is necessary to confirm this change in the Myanmar
P. vivax isolates. Taken together, the results of this study suggest that PvDBPII in the Myanmar isolates showed a high level of genetic polymorphism with a nucleotide diversity (π) of 0.0079 ± 0.0004.
Some polymorphic residues in PvDBPII occurred in only one population or geographic region, but common variant amino acids (K371E, D384G, E385K, K386N, N417K, L424I, W437R and I503K) are found in global isolates [
19‐
21,
31]. A high frequency (> 50%) of D384G (85.2%), R390H (63.0%), L424I (83.3%), W437R (61.1%), and I503K (77.7%) residues were found in Myanmar isolates compared to that in the Sal I sequence. This was highly similar to a report on Thailand isolates (D384G, 76.7%; R390H, 56.7%; L424I, 86.7%; W437R, 63.3%; I503K, 56.7%) [
20] but differed from previous studies showing R308S (67%), D384G (66%), and S447K (59%) in Papua New Guinean isolates; D384G (59%) in Colombian isolates; D384G (85%) and I503K (55%) in Brazilian isolates; D384G (61.3%) and I503K (70.6%) in Iranian isolates; and D384G (94%) and I503K (55%) in Sri Lankan isolates [
19,
21,
31]. F306L, which has only been reported from Asian malaria endemic areas, including Thailand [
20], Iran [
21], and Sri Lanka [
31], was also identified in the Myanmar isolates. Although the Myanmar isolates showed similar amino acid changes compared to those in the Thailand isolates, 9 variants found in the Thailand isolates (R268S, S351C, I367T, S398T, T404R, Q433K, R436T, N507H, and T513K) were not identified in the Myanmar isolates. Meanwhile, 11 variations (I310L, F344S, R391H, K455I, K473R, C477G, R490K, D528G, V533M, K541T, and A545V) found in the Myanmar isolates did not occur in the Thailand isolates. These results indicate that the Myanmar isolates are different from the Thailand isolates even though the two countries are very close geographically.
Although polymorphic residues were widely distributed throughout the PvDBPII sequence, polymorphisms at residues 417, 437, and 503, either in single or in combination, can affect the efficacy of inhibitory antibodies against erythrocyte binding [
23,
35]. As these residues compose an important discontinuous epitope in PvDBP, which might be the main target for inhibitory antibodies, these polymorphisms could be subject to immune pressure responsible for parasite escape from the host immune system. It has been confirmed that this strong positive selection pressure in PvDBPII promotes greater diversity [
14,
30]. The immune pressure drives the generation of new PvDBP variants that are still able to bind erythrocytes but become resistant to inhibitory antibodies, suggesting that this DBP region is under positive pressure at critical residues and under negative pressure at the residues involved in receptor recognition [
22,
23,
35]. A low prevalence of variant N417K (38.9%) was observed among Myanmar isolates, but more than 50% of W437R (61.1%) and I503K (77.8%) were identified. Analyses of the combination of variants revealed that W437R-I503K occurred at a higher frequency (70.4%), whereas N417K-W437K and N417K-I503K occurred at frequencies of 38.9% and 25.9%, respectively. This result suggests that there is a strong association between W437R-I503K in PvDBPII in Myanmar
P. vivax isolates, but not between N417K with either W437R or I503K.
The rate of non-synonymous mutations (
K n) and that of synonymous mutations (
K s) is generally used as an indicator of the action of natural selection in most coding gene sequences [
27]. Negative selection acting on coding genes can usually be identified when non-synonymous mutations are not advantageous, so the rate of synonymous mutations surpasses that of non-synonymous mutations (
K s >
K n). Meanwhile, positive selection is acting on a gene, when non-synonymous mutations can be beneficial (e.g. to avoid the host immune response), and the rate of non-synonymous mutations exceed that of synonymous mutations (
K n
> K s). Previous studies on PvDBPII diversity indicate that the high rate of non-synonymous mutations (
K n) relative to that of synonymous mutations (
K s) reflects positive selection pressure [
14,
31,
36]. The positive value of
K n/
K s (3.299) for all 54 sequences suggest that PvDBPII in Myanmar
P. vivax isolates is under positive natural selection. The positive values of Tajima's D (0.2635, P > 0.10) and Fu and Li's D (1.867, P < 0.02) and F statistics (1.535, 0.1 > P > 0.05) indicate that PvDBPII alleles occur at more intermediate frequencies than expected and that few alleles are rare or near fixation, which is consistent with the action of balancing selection, which maintains allelic variation in a population. These results collectively suggested that strong balancing selection, probably by host immune selection pressure, occurs at PvDBPII in the Myanmar isolates.
Polymorphism in B- and T-cell epitopes of parasite antigens may well enable parasites to escape host immune responses, as a polymorphism in the epitopes can up or down regulate T-cell responses to the index peptide or completely arrest an immune response, assisting escape of the parasite from the host immune system [
37]. The high degree of nucleotide diversity and high rates of non-synonymous to synonymous mutations are observed in known or predicted B- and T-cell epitopes and MHC binding regions of PvDBPII in the Myanmar isolates. Overall nucleotide diversity values for these epitopes and regions were greater than those for the entire PvDBPII. In particular, high levels of nucleotide diversity were identified in peptides 48 and Ia, which is comparable to Brazilian isolates [
30]. Positive Tajima's D values for these epitopes also suggested that positive natural selection preferentially acted on the epitopes in PvDBPII in the Myanmar isolates. These epitopes are predicted to be exposed to the surface of the PvDBP molecule [
30]. The putative changes in protein structure may alter antibody binding efficacy of a particular epitope, thereby allowing escape from the host protective immune response [
13,
30].
Many factors may contribute to genetic diversity in malaria populations, including mutations, intragenic recombination, natural selection, gene flow between different regions, and population size. Although it remains controversial, it has been suggested that recombination also contributes to the diversity of PvDBPII [
30,
36]. The existence of recombination events and the decline in the LD with increasing distance between nucleotide sites suggest that in addition to natural selection meiotic recombination may also contribute to maintain the diversity of PvDBPII among Myanmar
P. vivax isolates, as reported previously in Brazil, Colombia, and Sri Lanka [
30,
31].
Competing interests
The authors declare that they have no competing interests.
Authors' contributions
HLJ and JMK performed all the experiments and analysed the sequence data. SUM, JYK, HWL, KL, and BKN collected the blood samples. SUM performed sequence and phylogenetic analyses. BKN and TSK designed the study and supervised the study process. BKN wrote the paper. TSK, WMS, and JSL assisted in writing and editing the manuscript. All authors read and approved the final manuscript.