|Home | About | Journals | Submit | Contact Us | Français|
Merozoite surface proteins (MSPs) of the malaria parasites are major candidates for vaccine development targeting asexual blood stages. However, the diverse antigenic repertoire of these antigens that induce strain-specific protective immunity in human is a major challenge for vaccine design and often determines the efficacy of a vaccine. Here we further assessed the genetic diversity of Plasmodium vivax MSP4 (PvMSP4) protein using 195 parasite samples collected mostly from Thailand, Indonesia and Brazil. Overall, PvMSP4 is highly conserved with only eight amino acid substitutions. The majority of the haplotype diversity was restricted to the two short tetrapeptide repeat arrays in exon 1 and 2, respectively. Selection and neutrality tests indicated that exon 1 and the entire coding region of PvMSP4 were under purifying selection. Despite the limited nucleotide polymorphism of PvMSP4, significant genetic differentiation among the three major parasite populations was detected. Moreover, microgeographical heterogeneity was also evident in the parasite populations from different endemic areas of Thailand.
Plasmodium vivax malaria is considered as the most neglected, highly prevalent, and potentially dangerous disease (Baird, 2007; Price et al., 2007). Currently, it is estimated that 2.6 billion people live at the risk of vivax malaria and there are 130–145 million vivax malaria cases per year (Hay et al., 2004; Guerra et al., 2006). Although rarely fatal, a P. vivax infection is by no means clinically benign, and the episode of debilitating intensity can be repeated several times due to relapses (Price et al., 2007). Recently, several prospective studies have reported that clinical and severe vivax malaria, manifested by severe anaemia, respiratory distress and impaired consciousness, was common in hyperendemic areas of Asia and South America (Barcus et al., 2007; Rodriguez-Morales et al., 2007; Genton et al., 2008; Tjitra et al., 2008). Furthermore, the first-line therapies comprised of chloroquine and primaquine remain unchanged for the past 50 years, and the appearance of resistant parasites to these drugs are of great concern for vivax malaria control. As malaria eradication is once again on the agenda of international malaria control communities (Roberts and Enserink, 2007), these efforts must also include elimination of vivax malaria.
Effective control of malaria requires comprehensive approaches integrating multiple strategies such as chemotherapy, vaccination, and vector control. For vaccine development, antigens located on the surface or apical organelles of merozoites are prime candidates, given their potentials of eliciting immunity that blocks merozoite invasion of host erythrocytes (Richie and Saul, 2002). The effect of vertebrate host immunity on shaping the parasite population structure is well known and considered mostly responsible for the diverse antigenic repertoire of the parasite. There is strong evidence showing the induction of strain-specific protective immunity in human (e.g., Hodder et al. 2001; Cortes et al. 2005). Therefore, the significant levels of polymorphism present in these antigens create a major challenge for the development of a successful vaccine. Since the antigenic repertoire of a specific antigen is shaped by multiple factors such as the selection strength of host immune responses, functional constraints, and frequency of genetic recombination, detailed studies of antigenic diversity of a vaccine candidate in different malaria-endemic areas must be undertaken (Cui et al., 2003).
At least ten distinct P. vivax merozoite surface proteins (PvMSPs) have been identified experimentally or by in silico analysis of the P. vivax genome (Carlton et al., 2008). Among them, PvMSP4 shows significant homology to PfMSP4 in terms of sequence identity and gene structure (Black et al., 2002). In P. falciparum and P. vivax, MSP4 and MSP5 have similar gene organizations and are located in tandem on chromosome 2 and 4, respectively, suggesting that these two genes might have originated from an ancient gene duplication event. In rodent malaria parasites, a single gene MSP4/5 is found with similarity to both MSP4 and 5. The PvMSP4 gene has a two-exon structure, with a 300 bp exon 1 and 387 bp exon 2 separated by a 156 bp intron for the Thai-NYU strain. Like PfMSP4, PvMSP4 is a GPI-anchored surface protein with an epidermal growth factor (EGF)-like domain encoded in the second exon (Black et al., 2002).
PfMSP4 is being considered as a component of a subunit vaccine (Doolan and Stewart, 2007). Immunization of mice with Plasmodium yoelii MSP4/5 is able to confer significant protection against lethal challenges with blood stage P. yoelii (Goschnick et al., 2004). Recombinant PfMSP4 protein is immunogenic in laboratory animals and recognized by antibodies from humans living in malaria endemic areas (Wang et al., 1999; 2001). Limited population studies indicate that PfMSP4 has low sequence diversity (Wang et al., 2002, Benet et al., 2004, Polson et al., 2005), and analysis of PvMSP4 sequences from 30 Colombian samples revealed a similar result of limited polymorphisms (Martinez et al., 2005). Yet, it is not clear whether this reflects a recent expansion of similar strains in this region or global conservation of the PvMSP4 locus. To determine whether this observation can be extended to other vivax endemic regions, we have sequenced 195 samples from several vivax endemic countries. Our analysis supports the finding that PvMSP4 has limited genetic diversity.
We have procured 195 P. vivax finger-prick samples on filter papers from Thailand (n=130), Indonesia (n=34), Brazil (n=24), India (n=1), Papua New Guinea (n=1), Solomon islands (n=1), China (n=2) and Vietnam (n=2). This study was approved by the institutional review board at Chulalongkorn University. Parasite genomic DNA was extracted by using the QIAamp DNA Mini Kit (Qiagen, Hilden, Germany). The DNA purification procedure was essentially as described in the manufacturer’s instruction manual with minor modifications. The purified DNA was dissolved in TE buffer (10 mM Tris-HCl, 1 mM EDTA, pH 8.0) and stored at −20°C until use. All samples were genotyped at the highly polymorphic PvMSP1 locus to confirm that each sample was from a single P. vivax isolate (data not show). Further, samples showing superimposed eletropherogram signals during PvMSP4 sequencing were also excluded from analysis.
The DNA fragment spanning the entire coding region of the PvMSP4 gene was amplified using primers based on the PvMSP4 sequence of the Thai-NYU strain: 5′-ATGAAGGTGGCCTACTTTTTG-3′ and 5′-AGGTGTGTAGGGAGTAGGGA-3′. PCR conditions were denaturation at 94°C for 30 sec, annealing at 60°C for 30 sec and extension at 72°C for 1 min for 35 cycles. To minimize errors introduced in the sequences during PCR amplification, we used ExTaq DNA polymerase (Takara, Shiga, Japan) that has efficient 5′ → 3′ exonuclease activity to increase fidelity and shows no strand displacement. The PCR products were examined by electrophoresis in a 1% agarose gel and visualized with an UV transilluminator.
PCR products were purified using a PCR Purification kit (Qiagen), quantified, and used for direct sequencing. Sequencing was performed in both directions for each template using the BigDye Terminator version 3.1 Cycle Sequencing Kit on an ABI3100 Genetic Analyzer (Applied Biosystems). Overlapping sequences were obtained by using sequencing primers. Whenever singleton occurred, the sequence was re-determined using DNA templates from two independent amplifications from the same DNA samples.
Sequences were aligned by the Clustal X program with minor manual adjustment made by visual inspection. Haplotype diversity, nucleotide diversity, Tajima’s D, Fu & Li’s D* and F*, Fu’s Fs and minimum number of recombination events (Rm) were estimated using the DnaSP 4.1 software (Rozas et al., 2003). The number of nonsynonymous substitutions per nonsynonymous site (dN) and the number of synonymous substitutions per synonymous site (dS) were computed using the Nei-Gojobori method as implemented in the MEGA 4 program (Tamura et al., 2007). Interpopulation variance indices (Fst) and mismatch distribution were analyzed by the Arlequin software version 3.11 (Excoffier et al., 2005). Prediction of HLA-I binding peptides followed the method of Larsen et al. (2005) at http://www.cbs.dtu.dk/services/NetCTL. The PvMSP4 sequences of the Sal-1 and Thai-NYU strains were retrieved from the GenBank under the accession numbers XM001612944 and AF403475 and used for comparison.
To assess the global genetic diversity of P. vivax parasites at the PvMSP4 locus, we determined the PvMSP4 sequence from 195 parasite clinical samples from various vivax malaria-endemic areas. Together with two PvMSP4 sequences from the Sal-1 and Thai-NYU strain, 53 haplotypes were identified (Table 1). Geographically, the 130 Thai, 34 Indonesian, and 24 Brazilian samples contained 35, 15, and 12 PvMSP4 haplotypes, respectively. Hyplotype 1, 3, and 5 were the most abundant and they accounted for 52% of the P. vivax samples from these three countries. There were also shared haplotypes among the regions: 7 (#1, 3, 4, 5, 6, 23, and 26) shared by Thai and Indonesian populations, 6 (#1, 3, 5, 23, 43, and 46) by Indonesian and Brazilian populations, and 5 (#1, 3, 5, 11, and 23) by Thai and Brazilian populations. Previously, a tetrapeptide repeat motif was found in the second exon preceding the EGF domain, which varies in the number of repeats (Martinez et al., 2005). In addition to this repeat region, we have identified another polymorphic repeat array in exon 1 corresponding to amino acids 47–74 of the Sal I strain (Fig. 1). This repeat array consists of six tetrapeptide motifs (SGGA, SGGS, SGDS, SGGL, SGDA and IGDS), which are repeated 5–9 times in different isolates (Table 1). The second repeat array in exon 2, with the majority of the tetrapeptide motifs as DDHG, varies in copy number of 1–10. The variation of sequence and number of repeat array could be generated from slipped-strand mispairing mechanism, resulting in duplication, deletion or mutation of certain repeat units (such as ‘SGGS’ and ‘SGGL’). The polymorphisms in the two repeat arrays have contributed to most of the haplotype diversity and protein length polymorphism of PvMSP4. If the repeat haplotypes were not taken into account, only 9 haplotypes could be assigned to the 197 PvMSP4 sequences. Outside of the repeats, the sequences are highly conserved with only ten single nucleotide polymorphisms (SNPs) (Supplemental Fig. 1).
Nucleotide diversity (π) was limited in both exons of PvMSP4 in all three P. vivax populations (<0.00063). The Indonesian parasite population had the highest nucleotide diversity whereas the Brazilian population had no segregation sites (Table 2). Two silent SNPs and four amino acid substitutions were observed in exon 1, while four nonsynonymous substitutions were found in exon 2. One amino acid substitution (E/H) occurred in the EGF-like motif, akin to what was observed in the EGF-domain of PvMSP1. Intriguingly, none of the substitutions in the non-repeat regions was shared between Colombian isolates (Martnez et al., 2005) and the isolates analyzed in this study (Supplemental Table 1). The intron was conserved both in size and sequence with only one SNP found in samples from Thailand and Indonesia.
Statistical tests were performed to detect departure from neutrality in exon 1, 2, and entire PvMSP4 (Table 2). The Tajima’s D values for either exon 1, 2, or the entire coding region were negative, while significance was detected for exon 1 or the entire coding regions of PvMSP4 in the Thai P. vivax population. This implies that the P. vivax population could have undergone population expansion or have been under purifying selection. A similar trend was observed when the data were analyzed by the Fu and Li’s D* and F* tests. However, these tests cannot be implemented with the Brazilian population due to lack of segregating sites. Mismatch analysis using parameters estimated under the sudden population expansion model has shown that the sum of square deviation and Harpending’s raggedness index for each population shows a significant departure from such an assumption (p < 0.000). Fu’s Fs test for selective neutrality did not generate significant values for any populations. Taken together, results from these tests indicate that exon 1 and the entire coding region of PvMSP4 were under purifying selection.
We have calculated the number of nonsynonymous substitutions per nonsynonymous site (dN) and the number of synonymous substitutions per synonymous site (dS) in each domain of the PvMSP4 gene of isolates from Thailand and Indonesia (Table 3). In the entire coding region, the difference between dN and dS did not significantly depart from zero (Table 3). Despite this, in the region spanning nucleotides 598 to 702 in exon 2 (positions referred to the Thai NYU sequence) where a predicted epitope (YDDAEDDDL) was located, dN significantly outnumbered dS (p<0.05), suggesting micro-scale signature of positive selection similar to what was found in PfMSP4 (Jongwutiwes et al., 2002). Closer look into the substituted residues in exons 1 and 2 of PvMSP4 has identified a number of potential HLA class I-binding epitopes as predicted by the high scores for the C-terminal cleavage and the transporter associated with antigen processing efficiency (Table 4). These epitopes are either clustered upstream of repeat 1 or immediately following repeat 2. It is noteworthy that amino acid substitutions in some of these peptides abolished the predicted property of the epitopes. However, there was no statistical difference between dN and dS among predicted epitopes in exon 1 (data not shown).
The extent of geographic differentiation among the Thai, Indonesian, and Brazilian P. vivax populations was estimated by using Wright’s F-statistics, which revealed genetic differentiation among these three parasite populations (Table 5). The average fixation index of each population was similarly low (0.133–0.144), whereas interpopulation Fst differed significantly. The two Southeast Asian populations (Fst=0.042) were more closely related than they were to the Brazilian population (Fst=0.198 and 0.300, respectively). Despite the small values of the interpopulation Fst indices, they were all significantly different by the permutation test (p = 0.000 – 0.036), indicating that genetic differentiation exists among the P. vivax populations from these countries.
The minimum number of recombination events in the genealogy of the PvMSP4 locus of each population was zero, suggesting that intragenic recombination did not contribute to the microheterogeneity of this gene. Linkage disequilibrium analysis implemented in the DnaSP program did not support recombination either.
In this study, we have characterized the genetic diversity of PvMSP4 gene using 195 field samples representing diverse endemic areas of P. vivax parasite. Our results confirmed the previous finding about the conservation of this gene. Most of the observed haplotype diversity of PvMSP4 is due to polymorphism in repeat copy number of two repeat arrays. The overall nucleotide diversity of PvMSP4 is much lower than what has been determined for other P. vivax antigens such as MSP1 (Putaporntip et al., 2002), MSP3α (Rayner et al., 2002; Mascorro et al., 2005), MSP3β (Rayner et al., 2004), AMA1 (Figtree et al., 2000; Gunasekera et al., 2007; Grynberg et al., 2008), or even TRAP (Putaporntip et al., 2001). Moreover, the genetic diversity of PvMSP4 was almost ten times lower than that for PfMSP4 (Jongwutiwes et al., 2002; Benet et al., 2004; Polson et al., 2005). Our finding from this study is consistent with the notion that haplotype diversity is positively correlated with the level of malaria endemicity. Analyses using both microsatellite and antigenic markers have detected much higher genetic diversity of P. vivax populations in Asia than in South America, which might reflect the population history of the parasite (Gunasekera et al., 2007; Ferreira et al., 2007; Imwong et al., 2007; Karunaweera et al., 2008). Our results are in stark contrast to those from an earlier study of Colombian P. vivax samples although that study also revealed limited polymorphisms within the PvMSP4 gene (Martinez et al., 2005). Not only did the P. vivax population from Colombia have much higher nucleotide diversity than this work, but none of the SNPs were found in the parasite samples in this study. We don’t know whether this discrepancy truly reflects geographical differentiation of the parasite populations in South American regions or is due to sequencing errors.
Different tests for selection suggest the occurrence of opposing forces of selection on the PvMSP4 locus. As seen in the Duffy-binding protein, this is often imposed by the host immune pressure and functional constraint of the protein. Unlike most malaria parasite surface antigens that are under balancing selection, multiple neutrality tests strongly suggest that PvMSP4 has been under purifying selection. Interestingly, several amino acid substitutions in both exons might be major HLA epitopes, suggesting that the polymorphisms observed were likely selected by the host immune systems. The presence of micro-scale signature of positive selection at some clustered sites in exon 2 where a predicted T cell epitope was located and the lack of such signature in T cell epitopes in exon 1 could imply that the former epitope might be under more efficient immune selection than the latter. Nevertheless, further experimental immunological correlates are undoubtedly required to address this issue. The lack of balancing selection on this merozoite surface protein might be due to functional constraints, since PfMSP4 is essential for P. falciparum invasion of red blood cells and genetic knockout of this gene cannot be obtained (Cowman and Crabb, 2006). The limited polymorphism in PvMSP4 in global samples is in sharp contrast to its highly polymorphic paralogue PvMSP5 (Gomez et al., 2005). This could be due to their functional distinctions and/or different locations in the merozoite, which impose different levels of exposure of the molecules to the host immunity (Black et al., 2002).
F statistics detected different levels of significant genetic differentiation among the Thai, Indonesian, and Brazilian parasite populations, which is also reflected in the distribution patterns of individual PvMSP4 haplotypes in different parasite populations. While large geographical (continental) population differentiation is apparent, the populations appear to also have significant differences even within each continent. In particular, significant genetic differentiation exists between the two Asian populations, and between the Brazilian and Colombian populations. Of a similar notice is the different genetic diversity in our Thai samples, where 53 isolates were obtained from Southern Thailand. These samples had very limited haplotype diversity, and most of them were PvMSP4 haplotypes 5 (n=40) and 1 (n=11). This genetic structure resembles clonal expansion of the parasite, which is in striking contrast to the persistent and highly heterogeneous populations we have observed in western Thailand (Putaporntip et al., 2002; Mascorro et al., 2005). Both the macro-and micro-geographical genetic differences may be related to the demographic histories of the parasite populations. In Brazil, the lack of segregating sites at the PvMSP4 locus suggests recent spread of the parasite, a phenomenon also supported by the analysis of the PvAMA1 gene (Grynberg et al., 2008). In southern Thailand, the area has experienced a recent increase in P. vivax prevalence, which might explain the observed extreme paucity of PvMSP4 haplotypes. Taken together, the evolution of parasite antigens as candidates of vaccines needs to be examined on a detailed epidemiological scale.
Since malaria vaccines usually have to take the allele-specific immunity into account, the conservation in PvMSP4 in global parasite populations makes it easier to design a vaccine to cover the antigenic repertoire of PvMSP4. Functional characterization of the immunogenicity of PvMSP4 would allow further evaluation of its vaccine potential. Besides, its similarity to PfMSP4 especially in the EGF-like domain (67% identity) suggests that cross-species immunity to MSP4 may be elicited from natural exposure to either parasite species (Woodberry et al., 2008). Such vaccines may prove to be useful to control malaria in areas where both parasite species are endemic.
We are grateful to all patients who donated their blood samples for this analysis. This work was supported by research grants from The National Research Council of Thailand and The Thai Government Research Budget to S.J. and C.P., a grant from The Thailand Research Fund (RMU5080002) to C.P. and a grant (D43-TW006571) from Fogarty International Center, National Institutes of Health to L.C.
Publisher's Disclaimer: This is a PDF file of an unedited manuscript that has been accepted for publication. As a service to our customers we are providing this early version of the manuscript. The manuscript will undergo copyediting, typesetting, and review of the resulting proof before it is published in its final citable form. Please note that during the production process errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.