Background
Plasmodium vivax is the most prevalent human malaria parasite globally and is responsible for a large proportion of the global malaria burden, especially in regions outside of Africa [
1]. Although it has been neglected as a benign infection,
P. vivax causes serious clinical illnesses including respiratory distress, severe anemia, coma and even death [
2,
3]. Moreover, it has recently re-emerged in many temperate regions from where it had been largely eradicated during global malaria control campaigns. The emergence of drug resistance strains also complicates the burden of the parasite [
4,
5]. Considering the enormous socio-economic impact of
P. vivax on humans, development of an effective vaccine is an important concern in control and elimination strategies. However, no effective vaccine is yet available and the antigenic diversity present in wild-type isolates, which has led to the failure of several licensed and tested malaria vaccines, has been recognized as a major concern in developing a successful vaccine.
Apical membrane antigen-1 (AMA-1) is a type 1 integral membrane protein that is expressed in the late schizont stage of malaria parasites [
6]. It is initially synthesized in the micronemes of the apical complex of merozoites and is transported to the surface of parasite just prior to erythrocyte invasion, where it undergoes proteolytic cleavage [
7,
8]. Although the precise role of AMA-1 is not fully understood, it is believed to be essential in erythrocyte invasion [
9]. AMA-1 consists of a signal sequence, a cysteine-rich ectodomain, a conserved cytoplasmic region and a transmembrane region. The ectodomain is further divided into three distinct domains (domains I, II and III) by disulfide bridges [
10]. The ectodomain of AMA-1 is very immunogenic and a high antibody titer against the domain is produced in humans who are naturally infected with malaria parasites [
11‐
14]. Antibodies against AMA-1 effectively inhibit erythrocyte invasion [
15‐
17]. Therefore, AMA-1 has been considered as a promising candidate antigen for blood stage malaria vaccine [
18,
19].
Although it is recognized that AMA-1 is less variable than the other blood stage malaria vaccine candidate antigens such as merozoite surface protein antigens (MSPs) and circumsporozoite protein (CSP), it also shows sequence variations among global malaria parasites [
20‐
23]. In
P. falciparum, most of the genetic polymorphisms are concentrated in domain I [
20,
21,
24], while the majority of the polymorphic patterns of AMA-1 occur in domains I and II in
P. vivax [
22‐
27]. These polymorphisms result in amino acid changes in the natural population, indicating that PvAMA-1 is under natural selection, may be a result of host immune pressure [
23,
26].
In this study, the population genetic structure and natural selection of PvAMA-1 among Korean P. vivax isolates was analysed. A higher rate of polymorphic patterns and evidence of natural selection were identified in domains I and III. The Korean PvAMA-1 showed different polymorphic patterns compared to other global isolates. Recombination also likely has been important in generating genetic diversity across the PvAMA-1 sequences. These results provide useful information for the understanding of the population structure of P. vivax circulating in Korea and have important implications for the design of a vaccine incorporating PvAMA-1.
Discussion
Vivax malaria re-emerged in Korea in 1993 near the demilitarized zone (DMZ) border with North Korea [
34]. Since its re-emergence, the outbreak has spread into cities and counties adjacent to the DMZ and has persisted until now with fluctuating numbers of annual indigenous cases, with 32,300 official cases [
35]. Understanding the nature of the genetic population of
P. vivax circulating in Korea is beneficial to knowledge of the nationwide parasite heterogeneity and in the implementation malaria control programs in the country. Moreover, this information is helpful for the development of an effective malaria vaccine. Several studies on genetic diversity of polymorphic marker proteins including MSP-1, MSP-3α, MSP-3β and CSP in Korean
P. vivax suggested that rapid dissemination of genetic structure of
P. vivax is processed in recent years [
36‐
42]. Characterization of genetic property of PvAMA-1 in Korean
P. vivax isolates has also been analysed [
43,
44], but these studies analysed sequence polymorphism restricted to partial region corresponding domain I and not the entire PvAMA-1. In the present study, the genetic polymorphisms and natural selection in the entire PvAMA-1 among
P. vivax Korean isolates were analysed to gain an in-depth understanding of the nature of PvAMA-1 in Korean
P. vivax population.
The 66 PvAMA-1 sequences obtained from Korean
P. vivax isolates were classified into 30 different haplotypes, which were further clustered into seven distinct clusters. Compared to Sal I sequence, 66 SNPs, which resulted in 44 dimorphic amino acid substitutions, were identified in Korean PvAMA-1. The majority of the amino acid changes clustered within PvAMA-1 domain I, as has been reported previously for
P. vivax and
P. falciparum AMA-1 [
20‐
23,
26,
27,
45]. The 15 amino acid changes identified in Korean PvAMA-1 are comparable to those previously identified [
22,
23,
26,
27], which suggests that the polymorphic sites are under high natural selection. The 29 amino acid changes were novel and have not been reported, even though they were detected with low frequencies in Korean PvAMA-1. Compared to previously reported North Korea (EU395599) and South Korea (SK0814; GU476844) sequences, which are only two entire PvAMA-1 sequences currently available in Korean Peninsula-origin
P. vivax, Korean PvAMA-1 sequences analysed in this study shared similar patterns of amino acid changes. All amino acid changes reported in the North Korea and South Korea sequences were commonly identified in all Korean PvAMA-1 sequences analysed currently. In particular, the four tightly conserved amino acid changes (D107A, K120R, N132D and E277K), which are the most outstanding characteristic found in Korean PvAMA-1 analysed in this study, were well-conserved in the all the sequences. Based on the sequence polymorphism analyses of the current study, the North Korea sequence was most similar to the sequences belongs to cluster A, while the South Korea sequence was more closely related with the sequences of cluster D or G. Interestingly, Korean PvAMA-1 showed different patterns of polymorphism compared to previously reported sequences from other geographical regions [
22‐
27]. Although K120R/S and E277K were also identified at high frequencies (up to 50 %) in PvAMA-1 from other different geographical areas, D107A/N frequency was low in Iran (29.7 %) and Sri Lanka (17.4 %) and the N132D/G frequency was low in Iran (37.0 %), Sri Lanka (4.3 %), Thailand (17.1 %) and Venezuela (20.6 %) [
22‐
27]. S228D, which showed a high frequency (95.5 %) in Korean PvAMA-1, was identified with a high frequency only in Indian isolates (72.8 %) [
25]. Meanwhile, the amino acid changes (R112K/T/S, L140I, E145A, P201S and R438H), which showed high frequencies in other geographical regions, were not identified or detected only with low levels of frequency in Korean PvAMA-1. Unlike the relatively even global distribution of PfAMA-1 diversity [
46,
47], substantial geographical differentiation between populations was observed for PvAMA-1 [
23]. The results of this study also suggest that Korean PvAMA-1 showed different patterns of polymorphic nature compared to those of other geographical areas and substantial geographic differentiation was observed among global isolates.
Several studies have suggested that the most common polymorphic amino acid residues identified in PvAMA-1 are located on one face of the protein, which suggests that this face is more exposed to the exterior environment and is accessible to host immune responses [
21,
23,
27,
48]. Consistent with the previous studies, the majority of SNPs detected in Korean PvAMA-1 are exposed on one face of the protein. A recent study on the epitope mapping prediction of PvAMA-1 suggested that the potential B cell epitopes across the ectodomain [
27]. Some of the SNPs identified in global
P. vivax isolates, including E145K, P210S, R249H, G253E, K352E, R438H and N445D, overlap with the B-cell epitope regions. These amino acid changes may affect the protein structure by causing changes in charge and polarity of the protein and might help parasites to escape from host protective immunity [
27]. No or few changes were observed in the corresponding amino acids in Korean PvAMA-1, but N445D was identified with a high frequency (74.2 %), which differs significantly from isolates from other different geographical areas. Three additional amino acid substitutions (Q380K, V382E and L384P/R) were also may alter protein structure by causing changes in protein polarity and hydrophilicity [
27], which might decrease epitope binding scores or loss of the predicted linear B cell epitope either together or in combination with proximal polymorphisms [
49]. Q380K and V382E were not identified in Korean PvAMA-1, but L384P was observed at a high frequency (53.3 %). Intrinsically unstructured/disordered regions (IURs), which are widely identified in eukaryotic proteins, play important roles in many fundamental cell functions such as molecular recognition, molecular assembly and protein modification [
50,
51].
Plasmodium spp. also have many proteins containing IURs and these proteins have been considered important in attachment and invasion of the parasite into red blood cells, although their specific functions remain to be defined [
52]. Thereby overlap region of B-cell epitopes and IURs has been postulated more likely to represent a real antigenic/immunogenic region within a protein and the co-occurrence of both regions increases the chance of presenting this region with no secondary structure [
52,
53]. The predicted overlap IUR and B-cell epitope, which comprises 18 amino acid residues (SASDQPTQYEEEMTDYQK), is highly antigenic during natural human infections and is an important antigenic region of the domain II of PvAMA-1 [
27,
53]. This linear epitope was highly conserved in all Korean PvAMA-1 sequences, as it was strictly conserved in currently analysed global isolates [
27]. This suggests a high degree of amino acid sequence conservation of the region among global
P. vivax isolates and further supports the hypothesis that these amino acid sequences in domain II are subjected to strong purifying selection that might be used as a component of a PvAMA-1-based vaccine.
The level of nucleotide diversity at entire PvAMA-1 in Korean
P. vivax isolates (π = 0.00478) was relatively lower than those for isolates from different geographical areas, including Iran (π = 0.00826), Sri Lanka (π = 0.00675), PNG (π = 0.0079), Venezuela (π = 0.0065) and Thailand (π = 0.0089) [
22,
23,
27]. Lower malaria transmission in Korea compared to other endemic countries may contribute to the lower genetic diversity observed in Korean PvAMA-1. These collective results are suggestive of a low level of genetic polymorphism occurring at Korean PvAMA-1. The nucleotide diversity was not evenly distributed across entire PvAMA-1. The 5ʹ-T and 3ʹ-T showed low levels of nucleotide diversity, which suggests that the two regions are relatively well conserved in Korean PvAMA-1. A high level of nucleotide diversity was observed in domain I and domain III in PvAMA-1 of Korean
P. vivax isolates, different from previous reports that nucleotide diversity is highest in domains I and II of PvAMA-1 in Indian, Sri Lankan and Thailand
P. vivax isolates [
22,
26,
27]. This indicates that domain II is more conservative with lower frequency polymorphisms than domains I and III in Korean PvAMA-1, even though a cluster of SNPs was observed in domain II. Natural selection affected by host immune responses and recombination between genetically distinct malaria parasites during meiotic replication in the mosquito midgut have been recognized as the two main mechanisms by which PvAMA-1 genetic diversity is generated and maintained [
49]. The dN/dS for the entire length Korean PvAMA-1 was estimated to 0.703, suggesting that purifying selection pressure may act on the protein [
54]. However, the dN/dS value of domain II was higher than 1 (1.946), which implies that positive natural selection may be occurring in the domain. Negative values of Tajima’s D (−1.4140), Fu and Li’s D (−1.5383) and Fu and Li’s F (−1.2030) for entire Korean PvAMA-1 presently analysed imply an excess of low frequency polymorphisms indicating population size expansion and/or purifying selection [
31]. Each domain of Korean PvAMA-1 also showed negative values of Tajima’s D, Fu and Li’s D and Fu and Li’s F. These results clearly point to stronger diversifying selection, probably by host immune selection pressure, is working at Korean PvAMA-1. Meiotic recombination that occurs between the adjacent polymorphic sites drives allelic diversity of PvAMA-1 [
22,
26]. Evidence of recombination event within the Korean PvAMA-1 was also observed, which supported by the decline of LD index R
2 with increasing nucleotide distance, as consistent with the previous studies.