Genetic dissociation of three antigenic genes in Plasmodium ovale curtisi and Plasmodium ovale wallikeri

Plasmodium ovale curtisi and Plasmodium ovale wallikeri are two sympatric human malaria species prevalent in Africa, Asia and Oceania. The reported prevalence of both P. ovale spp. was relatively low compared to other malaria species, but more sensitive molecular detection techniques have shown that asymptomatic low-density infections are more common than previously thought. Whole genome sequencing of both P. ovale spp. revealed genetic dissociation between P. ovale curtisi and P. ovale wallikeri suggesting a species barrier. In this study we further evaluate such a barrier by assessing polymorphisms in the genes of three vaccine candidate surface protein: circumsporozoite protein/ thrombospondin-related anonymous-related protein (ctrp), circumsporozoite surface protein (csp) and merozoite surface protein 1 (msp1). The complete coding sequence of ctrp and csp, and a partial fragment of msp1 were isolated from 25 P. ovale isolates and compared to previously reported reference sequences. A low level of nucleotide diversity (Pi = 0.02–0.10) was observed in all three genes. Various sizes of tandem repeats were observed in all ctrp, csp and msp1 genes. Both tandem repeat unit and nucleotide polymorphism in all three genes exhibited clear dimorphism between P. ovale curtisi and P. ovale wallikeri, supporting evidence of non-recombination between these two species.


Introduction
Plasmodium ovale curtisi and Plasmodium ovale walllikeri are two sympatric species of malaria parasites found across many malaria endemic countries in Africa, Asia and Oceania [1][2][3]. a1111111111 a1111111111 a1111111111 a1111111111 a1111111111 Although morphological features of P. ovale curtisi and P. ovale wallkeri are indistinguishable, these two P. ovale species are genetically distinct, and there is evidence of differences in latency and clinical presentation [4][5][6]. Nuclear genome sequences of P. ovale curtisi and P. ovale wallikeri were recently reported and revealed different expansion in some gene families [7]. Currently the target genes used for discriminating between P. ovale curtisi and P. ovale wallikeri are the SSU rRNA gene [8], tryptophan rich antigen (potra) [9], reticulocyte-binding protein 2 (porbp2) [9], and some sexual stage proteins [9]. Sequence polymorphisms in the cell-surface associated proteins that are candidate targets for vaccine development have only been studied rarely. The current study assessed genetic diversity in a highly polymorphic region of the blood stage merozoite surface protein gene msp1, and in two genes encoding sexual stage and sporozoite proteins, ctrp and csp respectively, in P. ovale curtisi and P. ovale wallikeri.
CTRP is a member of the micronemal and cell-surface associated proteins. In P. falciparum disruption of the ctrp gene prevents oocyst development in the anopheline mosquito [10], indicating that CTRP is important for mosquito midgut development. For this reason CTRP has been proposed as a transmissions-blocking vaccine candidate. CSP is the major surface protein on the Plasmodium sporozoite. It is a candidate target for pre-erythrocytic stage vaccine development. Genetic polymorphism within the csp gene have been investigated in most human malaria species including P. falciparum [11,12], P. vivax [13,14], P. malariae [15,16], and P. knowlesi [17], but not in P. ovale. MSP1 is one of the predominant antigen expressed in the erythrocytic stage of Plasmodium spp. The msp1 gene is highly polymorphic and has been well characterized in P. falciparum [18,19] and P. vivax [13,14]. A study of P. ovale isolates from Thailand revealed low diversity in the msp1 gene [20].
The current study evaluates sequence diversity of ctrp, csp and msp1, in a wider collection of P. ovale isolates collected from Thailand and African countries. Assessing diversity in these surface proteins is important for defining vaccine candidates, and to further assess the species barrier between P. ovale curtisi and P. ovale wallikeri. In the current era of malaria elimination, the better understanding of P. ovale curtisi and P. ovale wallikeri is essential to ensure success against all human malaria species.

Samples
Twenty-five samples of P. ovale (14 P. ovale wallikeri and 11 P. ovale curtisi) were collected from Thailand and African countries during 1995-2010 (S1 Table). All samples were obtained from patients enrolled in previous studies who gave written informed consent to blood sampling. Parasitaemia of these samples varied from 1 per 500 WBC to 198 per 500 WBC. The protocol for this study was reviewed and approved (reference number MUTM2001-049-04) by the ethics committee of the Faculty of Tropical Medicine, Mahidol University, Thailand. Genomic DNA of all samples was confirmed for the present of P. ovale. Nested PCR of the SSU rRNA gene was performed with primer rPLU1/rPLU5 in the primary reaction and with primer rOVA1/rPLU2 in the secondary reaction [21]. A nested PCR protocol based on the linker region of dhfr-ts gene was applied with primer Pla-DHFR-F/Pla-TS-R in the primary reaction and with primer PO-Lin-F/PO-Lin-R in the secondary reaction [22]. In addition, a semi-nested PCR of potra gene was performed with a primer specific to both P. ovale spp. (PoTRA-F/PoTRA rev3) in the primary reaction and with specific P. ovale curtisi (PoTRA-F/ PocTRA-R) and P. ovale wallikeri (PoTRA-F/PowTRA-R) primers in the secondary reaction [23].

Isolation of poctrp and pocsp gene
Specific primers targeting poctrp and pocsp genes were designed to obtain the full length of those two gene sequences (Table 1). A semi-nested PCR approach was used for amplification of each fragment with PCR conditions as presented in Table 1. All PCR reactions were performed with 10 mM Tris-HCl, pH 8.3, 50 mM KCl, 2 mM MgCl 2 , 125 μM dNTPs, 250 nM of

Analysis of variable region in pomsp1 gene
Twelve available pomsp1 sequences from both P. ovale curtisi and P. ovale wallikeri were retrieved from the NCBI database (accession number FJ824670, FJ824671, KC137340-KC137349) and multiple sequence alignments were performed. A highly polymorphic region within pomsp1 was observed between amino acid residues 700 to 1,000. The primers OMSP1. F1, OMSP1.R1, and OMSP1.R2, were designed for a semi-nested PCR approach to analyses this polymorphic domain in 25 P. ovale samples (Table 1). Positive PCR products were then purified by Gel/PCR purification kit (Favogen, Taiwan), before being submitted for DNA sequencing. All pomsp1 sequences obtained in this study were analyzed together with the previous reports.

Isolation and analysis of poctrp
The complete coding sequence of poctrp gene was obtained from 11 P. ovale curtisi and 14 P. ovale wallikeri isolates (accession number MK403987-MK404009). It revealed that the poctrp genes for both species has only one exon encoding for 2,007 to 2,047 amino acids. Sequence alignment of these 25 poctrp sequences, together with another two poctrp sequences (accession number LT594512 and LT594589) available in the NCBI database, and other ctrp sequences from the other Plasmodium spp. that infect humans, showed that poctrp is composed of a signal peptide, six vWA domains, seven TSP1 domains, transmembrane domain, and a cytoplasmic region (Fig 1). Alignment of the CTRP of all human Plasmodium spp. revealed highly conserved transmembrane (TM) and cytoplasmic regions (Fig 1). A conserved amino acid sequence YGYN/K for the tyrosine-based TM motif involved in cellular trafficking, and the cytoplasmic domain tryptophan residue (Fig 1) which is the key interaction to drive parasite motility were conserved between all studied human Plasmodium spp. Multiple alignment of the full-length CTRP among all human-infecting Plasmodium spp. also identified a highly conserved region close to the C-terminus.
All poctrp sequences including the two reference sequences were translated to deduce their corresponding amino acids and analyzed for intra-and inter-specific sequence diversity at this locus. The deduced amino acid alignment of PoCTRP showed two prominent regions. The first region is located around 300-320 amino acids between the vWA1 and vWA2 domains. P. ovale curtisi isolates carries two amino acids repeats "PE" with 7-11 copies, while all P. ovale wallikeri isolates had 4 "PE" repeat units (Fig 2). The second region is located between codons 570 and 600, where a tandem repeat of six amino acids was identified. Three patterns of six amino acids repeats were observed: ENPDSS, EKPGSS, and ENPGSS. Different numbers of repeat units were presented in the P. ovale isolates (Fig 2). The repeat EKPGSS is the most frequent in both P. ovale curtisi and P. ovale wallikeri. This region showed a marked difference in length, providing a potential additional genotypic marker to differentiate P. ovale curtisi from P. ovale wallikeri. Multiple sequence alignment of CTRP of all human Plasmodium spp. revealed species-specific regions for P. ovale spp. at codons 512-538 and codons 573-599. PCR amplification of P. ovale spp. with primers targeting those two regions are useful to distinguish P. ovale curtisi from P. ovale wallikeri.
The available poctrp genes were analysed in a sliding plot for nucleotide diversity between P. ovale wallikeri and P. ovale curtisi (Fig 3). P. ovale curtisi showed higher diversity around the first 1 kb where P. ovale wallikeri showed higher diversity at 4 kb-5 kb of poctrp (Fig 3). For this gene, the average nucleotide diversity of P. ovale curtisi is slightly lower than that of P. ovale wallikeri, and combined analysis of all 27 P. ovale sequences showed a higher diversity value than that calculated from each species alone (Table 2), which indicates distinct distributions of diversity across the poctrp locus in the two species.

Isolation and analysis of pocsp
The pocsp gene was successfully amplified from 14 P. ovale wallikeri and 11 P. ovale curtisi isolates (accession number MK404010-MK404031). The complete pocsp gene varied in size from 1,020 to 1,185 bp, and the size variation resulted from variable tandem repeats in the central repeat region. The pocsp sequences were analyzed together with two other sequences available in NCBI databases and those of the csp of the other human Plasmodium spp. The protein domain architecture of pocsp was determined based on homologous CSP proteins alignment with other human Plasmodium spp. The pocsp structure domain was similar to that of the csp from the other Plasmodium spp. Four domains in the conserved N-terminus domain (conserved region I) and in the conserved C-terminus domain (Th2R, conserved region II, and Th3R) were of particular interest. A summary of the amino acid patterns in each of these domains is presented in Table 3. Overall, a higher number of haplotypes was observed in P. ovale curtisi as compared to P. ovale wallikeri. Whereas P. ovale wallikeri showed only one haplotype in conserved region I and two in conserved region II, three and six, respectively, were observed for P. ovale curtisi. A high number of haplotypes was observed in the Th2R and Th3R domains for both P. ovale curtisi and P. ovale wallikeri, but it is interesting to note that none were shared by both species. The central repeat region of pocsp was also analyzed. Several patterns of nine amino acid repeats were observed. A specific repeat unit was observed for P. ovale wallikeri (DPPAPVPQG), and for P. ovale curtisi NPPAPQGEG, with the latter showing a higher diversity in repeat unit numbers (Fig 2).
The pocsp gene was evaluated for nucleotide diversity between P. ovale curtisi and P. ovale wallikeri. Sliding plots of nucleotide diversity revealed overall higher nucleotide diversity in P. ovale curtisi than in P. ovale wallikeri (Fig 3). The estimated synonymous (dS) and nonsynonymous (dN) substitution was also found at higher value in P. ovale curtisi than that of P. ovale wallikeri ( Table 2). Combined analysis of both P. ovale spp. showed significantly positive values (p<0.05) for Fu and Li's D and Fu and Li's F tests, suggesting population bottlenecks or balancing selections in these two species (Table 2).

Genetic analysis of pomsp1
The sequences for the variable regions within the pomsp1 gene covering amino acids 710 to 1,020 were obtained from 14 P. ovale wallikeri and 11 P. ovale curtisi isolates (accession number MK404032-MK404049). Apart from this, sixteen sequences of pomsp1 gene (accession number LT594511, LT594588, KX672044, KX672045, FJ824670, FJ824671, KC137340-KC137349) were available in the NCBI database. Taken together, 41 PoMSP1 sequences were used in the alignment. A clear dimorphic pattern was observed between P. ovale curtisi and P. ovale wallikeri. Amino acid tandem repeat patterns were found in P. ovale spp. The tandem repeats are characteristic for the two different P. ovale spp. There were three arrangement patterns of three 5-amino acid repeat units (PGAGG, PGAAG, and PGVPG) found exclusively in P. ovale wallikeri isolates Whereas, nine arrangement patterns of six 4-amino acids repeat units (QAAT, QTAT, HAST, QATT, QVTT, QSAT) were observed specifically in the P. ovale curtisi isolates (S2 Table).
Analysis of gene diversity and haplotype diversity at the pomsp1 locus showed that P. ovale curtisi has higher diversity than that of P. ovale wallikeri ( Table 2). Sliding window plots showed higher overall nucleotide diversity in P. ovale curtisi than in P. ovale wallikeri (Fig 3). The ratio of synonymous (dS) and nonsynonymous (dN) substitutions was higher in P. ovale wallikeri than P. ovale curtisi ( Table 2).

Comparative analysis of P. ovale curtisi and P. ovale wallikeri
Genetic analysis of P. ovale curtisi and P. ovale wallikeri based on three surface protein genes revealed clear dissociation between these two species. Analysis within each species was performed though sequence diversity and amino acid patterns. The sequence polymorphism in  poctrp, pocsp and pomsp1 showed more divergence in P. ovale curtisi than in P. ovale wallikeri (Fig 3, Table 2). The test for neutrality (Tajima's D, Fu and Li's D, and Fu and Li's F tests) was applied to poctrp, pocsp and pomsp1 to compare observed polymorphism frequencies with expected frequencies.
Significantly positive values were obtained from Fu and Li's D and Fu and Li's F when P. ovale spp. were analyzed as one group (Table 2). These statistics reflect higher than expected frequencies of alleles, which might have resulted from population bottlenecks or balancing selections. P. ovale curtisi had a higher number of different haplotypes in all conserved domains of the CSP (Table 3). This suggests that P. ovale curtisi is intrinsically more genetically diverse than P. ovale wallikeri, but may also represent limitations of our sample. Some of the studied P. ovale curtisi and P. ovale wallikeri infections were mixed with other human malaria spp., which might have impacted the characteristics of CTRP, CSP and MSP1. Therefore the genetic analysis of CTRP, CSP and MSP1 was compared between single and mixed infections. There were four mixed infection found in P. ovale wallikeri, in which all four samples were collected from Thailand. For P. ovale curtisi, most samples were collected from Africa with four single infections and five mixed infections. There was no significant difference in average nucleotide diversity (Pi), haplotype diversity, and dN/dS between single and mixed infections. Statistical testing for neutrality (Tajima's D, Fu and Li's D, and Fu and Li's F tests) was also not significantly different (S3 Table). In addition, the pattern of tandem repeats in CTRP, CSP and MSP1 in P. ovale spp. showed no difference either between single and mixed infections or between Asia and Africa isolates.
To infer genetic relationships of P. ovale curtisi and P. ovale wallikeri, a phylogenetic tree was reconstructed based on the three cell-surface associated proteins, CTRP, CSP and MSP1. The Neighbour-Joining method [27] was used to infer the evolutionary history of each species. Based on CTRP, CSP and MSP1 (Fig 4), P. ovale curtisi and P. ovale wallikeri clustered according to species, and the tree topologies inferred from each gene showed similar features of grouping into P. ovale curtisi and P. ovale wallikeri. This suggested that the genes of P. ovale curtisi and P. ovale wallikeri do not recombine and show distinct characteristics (Fig 4).

Discussion
In addition to the earlier described polymorphisms in pomsp1, the current study provides the genetic characterization for two more cell-surface associated proteins: ctrp and csp. Analysis of the complete ctrp gene from all human Plasmodium spp., including the P. ovale species presented here, revealed a strongly conserved region in the CTRP protein, likely related to its importance for parasite survival. The highly conserved transmembrane and cytoplasmic regions are likely associated with cellular trafficking and parasite development. The C-terminal part containing residues and domains crucial for CTRP function for all human Plasmodium spp. were also conserved. This protein could therefore be a candidate target for vaccine development. Analysis of the csp gene in P. ovale curtisi and P. ovale wallikeri revealed a similar gene structure compared to that of the other human malaria species. The amino acid haplotypes observed in the conserved region of the csp gene were nearly all specific to either one or other species, with only 1/3 and 1/7 shared for conserved region I and conserved region II, respectively. Interestingly, no overlap was observed for the 17 Th2R and 13 Th3R haplotypes detected. This could imply a species-specific immune interactions with these T helper epitopes, or indicate a distinct biologically functional constraint. As the two species harbor distinct pocsp repeat regions, NPPAPQGEG and DPPAPVPQG, respectively, these peptides may provide useful species-specific targets for the development of antibody reagents for serological distinction of sporozoites from the two ovale species.
The study also provided addition information on the cell-surface associated protein, MSP1. Analysis of the variable region within pomsp1 of 25 P. ovale samples in this study supplemented with 16 pomsp1 from previous reports showed a clear distinction between P. ovale curtisi and P. ovale wallikeri. Alignment of the MSP1 from all human Plasmodium spp. showed the interspecies conserved blocks corresponding to previous characterizations [28]. Sequence polymorphisms of CTRP, CSP and MSP1 from each P. ovale spp. can be used for determination of parasite evolutionary relationships. Phylogenetic tree reconstruction based on concatenated CTRP, CSP and MSP1 clearly showed that P. ovale curtisi and P. ovale wallikeri are cluster separately, consistent with previous reports [29,30].
Analogies in the reported surface proteins in P. ovale with other human Plasmodium species could help selecting potential vaccine candidates. For instance, CTRP affects oocyst development of P. falciparum in Anopheles mosquitoes [10], and conserved regions within CTRP across human Plasmodium spp. could provide candidate targets for transmission-blocking vaccine. In MSP1, domain architectures are similar between all human Plasmodium spp., and our study of PoMSP1 revealed an interspecies conserved domains 6 (residues 812-911) Genetic dissociation of Plasmodium ovale curtisi and Plasmodium ovale wallikeri between the Plasmodium spp., which could be candidates for a trans-species malaria vaccine. Our data could also provide the basis for development of new serological reagents for distinguishing the two species, and for identifying individuals with a history of exposure to P. ovale spp. carrying species-specific serum antibodies. In addition to the genes evaluated in this study, other important polymorphic genes have been used for discrimination between the two P. ovale spp., including the surfin variant gene family and the Plasmodium interspersed repeat (pir) superfamily, which showed expansion in both P. ovale spp. [7]. Additional genes encoding potential targets for vaccine development warrant further study, including genes encoding reticulocyte binding proteins and tryptophan-rich domains [31].
In summary, this study showed conserved domains in the poctrp and pocsp genes which code for potential targets for future vaccines. Quantifying polymorphism in nucleotide sequences and the tandem repeat diversity between P. ovale curtisi and P. ovale wallikeri showed absence of recombination, supporting their designation as distinct species. Within the three analysed genes, diversity was higher in P. ovale curtisi than in P. ovale wallikeri. However, this will need to be confirmed in a larger sample size with better comparison between the geographical areas where the strains were collected. In the current sample most P. ovale curtisi was collected from highly endemic African countries whereas most P. ovale wallikeri were collected in Thailand which has low endemicity.
Supporting information S1