Positive selection on the Plasmodium falciparum clag2 gene encoding a component of the erythrocyte-binding rhoptry protein complex.

A protein complex of high-molecular-mass proteins (PfRhopH) of the human malaria parasite Plasmodium falciparum induces host protective immunity and therefore is a candidate for vaccine development. Clarification of the level of polymorphism and the evolutionary processes is important both for vaccine design and for a better understanding of the evolution of cell invasion in this parasite. In a previous study on 5 genes encoding RhopH1/Clag proteins, positive diversifying selection was detected in clag8 and clag9 but not in the paralogous clag2, clag3.1 and clag3.2. In this study, to extend the analysis of clag polymorphism, we obtained sequences surrounding the most polymorphic regions of clag2, clag8, and clag9 from parasites collected in Thailand. Using sequence data obtained newly in this study and reported previously, we classified clag2 sequences into 5 groups based on the similarity of the deduced amino acid sequences and number of insertions/deletions. By the sliding window method, an excess of nonsynonymous substitutions over synonymous substitutions was detected in the group 1 and group 2 clag2 and clag8 sequences. Population-based analyses also detected a significant departure from the neutral expectation for group 1 clag2 and clag8. Thus, two independent approaches suggest that clag2 is subject to a positive diversifying selection. The previously suggested positive selection on clag8 was also supported by population-based analyses. However, the positive selection on clag9, which was detected by comparing the 5 sequences, was not detected using the additional 34 sequences obtained in this study.

1. Introduction

Plasmodium falciparum, the causative agent of malignant tertian malaria is, for part of its life cycle, an obligate intra-erythrocytic parasite. This stage of the life cycle is characterized by repeating cycles of erythrocyte invasion, followed by growth and schizogonic multiplication within the cell, egress, and re-invasion. The erythrocyte invasive unit, the merozoite, expresses a panel of proteins on its surface, all of which are exposed to the host immune system for the brief time that the parasite is free in the blood plasma. These proteins, exemplified by merozoite surface protein 1 (MSP1), tend to be more polymorphic (showing a higher degree of amino acid diversity between parasite strains) than other non-exposed parasite proteins. A possible explanation for this is that such proteins have undergone “positive diversifying selection”, mediated by host immune pressure [1]. Proteins secreted by the parasite at the point of cell invasion are also exposed, however briefly, to the host immune system, and may be targeted by immune pressures, and polymorphism selected within them, in much the same way as merozoite surface proteins. Apical membrane antigen (AMA1), released from the merozoite micronemes during invasion, for example, is believed to have undergone positive diversifying selection in this manner [2].

The RhopH complex is a high molecular mass erythrocyte-binding protein complex secreted from the merozoite rhoptry during invasion, and antibodies raised against it have been shown to confer anti-parasite protection to the host [3-6]. The precise role of this complex in erythrocyte invasion remains unclear. Components of this complex have been detected on the erythrocyte cytosol side of the parasitophorous vacuole membrane and the parasite-infected erythrocyte membrane, suggesting a role in the formation of the parasite-restructured membranous architecture in the infected erythrocyte [7, 8]. The RhopH complex is itself comprised of three distinct proteins: RhopH1, RhopH2 and RhopH3, each encoded by separate genes [9-11]. RhopH1 is encoded by a multigene family termed the cytoadherence-linked asexual gene (rhoph1/clag) family, that consists of at least five paralogous genes (clag2, clag3.1, clag3.2, clag8, and clag9), and each RhopH complex contains one of the rhoph1/clag gene products [9, 12, 13].

In a previous study, the degree of inter-allelic polymorphism for seven RhopH complex-related genes, 5 rhoph1/clag, rhoph2 and rhoph3, was evaluated by comparing nucleotide sequences from 5 culture-adapted parasite lines. It was found that clag2, clag3.1, clag3.2 and clag8 were highly polymorphic and that amino acid substitutions and insertions/deletions (indels) were found mainly in a region encompassing amino acid positions 1000–1200 of these gene products. The signature of positive selection was detected by an excess of nonsynonymous substitutions over synonymous substitutions for clag8 and clag9. Further evidence for the involvement of positive selection in driving clag8 polymorphism was independently gathered in an analysis that compared the Plasmodium reichenowi orthologous sequence for clag8 to the P. falciparum gene [14].

In this study, we extend our analysis of clag polymorphism further by increasing the number of sequences and employing a population-based approach utilizing sequences obtained from parasites collected in Thailand.

2. Materials and Methods

2.1. Parasite culture and DNA extraction

All 39 P. falciparum lines examined in this study (MS802, MS803, MS804, MS805, MS806, MS807, MS808, MS809, MS810, MS811, MS812, MS814, MS815, MS816, MS817, MS818, MS819, MS820, MS821, MS822, MS824, MS825, MS826, MS827, MS828, MS829, MS830, MS831, MS833, MS834, MS835, MS837, MS838, MS840, MS842, MS843, MS844, MS946, and MS947) were collected in Mae Sot, Thailand from November 21, 1988 to January 16, 1989 and were maintained in vitro essentially as previously described [15-17]. Since MS814 and MS822 showed unclear chromatograms for clag sequences, these lines were cloned by limiting dilution, yielding clones MS814K and MS814R, and MS822B6 and MS822G8, respectively. The human erythrocytes and plasma used for culture were obtained from the Nagasaki Red Cross Blood Center. Parasites were harvested when parasitemia reached about 2%, and parasite DNA was extracted using DNAzol BD (Invitrogen).

2.2. Polymerase chain reaction (PCR) amplification and sequencing

DNA fragments were PCR-amplified twice, independently, with oligonucleotide primers: TATATGGAAAAA GTAGTAATATACAGG and TACTAGTATGTGGTTGAT ATTCTTTTG for clag2 (resulting PCR product with the size of 702 bp); GTTTATGGAAAAAGTGGTAAAATAGG and CTCTTTAAGTTTTCTTCTGAATAGTTC for clag8 (750 bp); and ATAAACTTGATAGAATATATGGTAAAGC and ATTGAATAATCTTTTAATGTACATGCAC for clag9 (764 bp) in a 20 µL reaction mixture using a high-fidelity KOD Plus DNA polymerase (TOYOBO, Japan). The PCR conditions were as follows: 94°C for 2 min; 40 cycles of 92°C for 15 sec, 54°C for 20 sec, 68°C for 1 min 10 sec; final extension step of 68°C for 5 min. PCR products were subjected to 1.5% agarose gel electrophoresis, and when a single band product with no background was observed, PCR-amplified DNA fragments were directly sequenced following treatment of PCR mixture with ExoSAP-IT (GE Healthcare, UK). Two independent PCR products were sequenced using a panel of primers described previously [14]; one in the forward direction and the other in the reverse direction, to ensure the accuracy of the obtained sequences. Sequencing reactions were performed using the BigDye® Terminator v1.1 Kit (Applied BioSystems, UK) with an ABI3730 DNA analyzer (Applied BioSystems). Sequences were manually corrected using BioEdit 7.0.0 software [18]. Sequences of clag2 (AB250822) and clag8 (AB250849) for MS838 were obtained from the database.

2.3. Statistical analyses of genetic diversity

Nucleotide diversity (π) and its standard error (SE) were computed by the Jukes and Cantor method using MEGA 4.0 software [19]. The mean number of synonymous substitutions per synonymous site (dS) and nonsynonymous substitutions per nonsynonymous site (dN) and their standard errors were computed using the Nei and Gojobori method [20] with the Jukes and Cantor correction, implemented in MEGA 4.0. The statistical difference between dS and dN was tested using a one-tailed Z-test with 500 bootstrap pseudosamples in MEGA 4.0. A value of dN significantly higher than dS at the 95% confidence level was taken as evidence for positive selection. The dN/dS ratio was evaluated using a sliding window method (90 bases with a step size of 3 bases) in DnaSP 4.0 [21]. Population genetics tests of neutrality were applied to the 3 clag gene sequences. Tajima’s test was used to test for departure from neutrality as measured by the difference between π (observed average pairwise nucleotide diversity) and θ (expected nucleotide diversity under neutrality derived from the number of segregating sites, S). Under positive diversifying selection, rare alleles are selected and maintained at intermediate frequencies, elevating π above that expected under neutrality and making the value of the test statistic (D) positive [22]. Fu and Li’s test was also used to evaluate positive diversifying selection by comparing estimates of θ based on the number of singletons and that derived from S (the D* index) or π (the F* index). Under positive diversifying selection, an excess of intermediate frequency polymorphisms and lower number of singletons make the value of D* and F* positive [23]. The regions analyzed in this study are shown in Fig. 1.

Fig. 1.

Regions used for the analysis. Bars with arrows under each schematic for Clag protein indicate the regions used for the analysis in this study. Vertical bars and a gap with asterisk above each schematic indicate polymorphic sites among 5 P. falciparum laboratory isolates and an indel reported previously [14]. Amino acid position is according to the 3D7 line sequence.

3. Results and Discussion

3.1. Clag2 sequence is classified into 5 groups with the signature of positive diversifying selection detected on groups 1 and 2.

In this study, we obtained 35 sequences of the most polymorphic region of clag2, 40 sequences for clag8 and 34 sequences for clag9. Iriko et al. (2008) did not detect the signature of positive diversifying selection on clag2 using a set of sequences which excluded the most polymorphic regions with indels, due to difficulty in obtaining a reliable nucleotide sequence alignment [14]. Thus, we classified clag2 into distinct groups using the deduced amino acid sequences from 35 P. falciparum clag2 alleles generated during this work, 31 previously reported alleles and one clag2 ortholog (prclag2) from the chimpanzee malaria parasite Plasmodium reichenowi, based on amino acid similarity and numbers of indels, and attempted to detect the signature of positive selection for each group separately. We classified PfClag2 sequences into five groups (Fig. 2). Among 66 PfClag2 sequences, 44 sequences (including the 3D7 line sequence) were classified in group 1, 11 sequences with a double amino acid insertion in group 2, five sequences with a seven amino acid insertion in group 3, four sequences with an eight amino acid insertion in group 4, and two sequences showing a high degree of similarity to the P. reichenowi Clag2 ortholog sequence in group 5. The observations that 1) amino acid sequences are clearly distinct between groups, 2) the group 5 sequence is similar to the P. reichenowi Clag2 ortholog sequence, and 3) the older origin of clag2 polymorphism than the time to the most recent common ancestor of the extant P. falciparum population was previously proposed based on the analysis using sequences excluding indels [14], raise the possibility that these 5 distinct Clag2 groups may have been generated before the time to the most recent common ancestor of the extant P. falciparum population. This may be clarified by analyzing the sequences of Clag2 orthologs from the malaria parasite species recently discovered in gorillas, and currently thought to be the closest relative species [25].

Fig. 2.

Clag2 is classified into 5 groups based on the similarity in the variable region. Dots and bars indicate identical amino acid residues with the 3D7 line sequence and gaps. Asterisks indicate sequences reported previously [14]. PrCL2 indicate Plasmodium reichenowi Clag2. Amino acid positions shown above the sequences are according to the 3D7 line sequence.

We attempted to detect the signature of positive diversifying selection on clag2 alleles belonging to groups 1 (n = 24) and 2 (n = 8) and clag8 (n = 41) by comparing dN and dS for sequences obtained in Thailand during the 1988–1989 period (“Thai” in Fig. 3), or 44, 11, or 69 sequences, respectively, after combining sequences reported previously (“Thai + others” in Fig. 3) [14]. No significant difference was detected between dN and dS, when the nucleotide sequences corresponding to nucleotide position (nt) 3106–3642 for clag2 or nt 3022–3591 for clag8 (numbered according to the 3D7 line sequence) were analyzed. However, when dN and dS were compared for the region where a high dN/dS ratio was detected by the sliding window plot method, a significant excess of dN over dS was detected for group 1 clag2 (mid point nt position of 3444 for “Thai + others”, dN/dS = 3.13, p<0.02; mid point nt position of 3447–3453 for “Thai + others”, dN/dS = 3.13–3.27, p<0.05), group 2 clag2 (mid point nt position of 3396–3459 for “Thai” only, dN/dS = 5.90–7.65, p<0.02; mid point nt position of 3396–3459 for “Thai + others”, dN/dS = 8.02–10.02, p<0.02), and clag8 (mid point nt position of 3258–3309 for Thai only, dN/dS = 3.49–53.96, p<0.02; mid point nt position of 3258–3309 for Thai + others, dN/dS = 3.21–33.27, p<0.02; mid point nt position of 3531–3546 for Thai + others, dN/dS = 6.99–7.68, p<0.05), suggesting that positive diversifying selection had operated on these regions. Positive selection was not detected for clag9 using sequences from Thai only (n = 34) and Thai + others (n = 39). Thus, in addition to clag8 and clag9, for which a positive selection was detected in a previous study [14], positive diversifying selection was detected here for at least two groups of clag2 sequences (Fig. 3).

Fig. 3.

Sliding window plot of dN/dS ratio for Plasmodium falciparum clag2 (group 1 and 2) and clag8 in Thai isolates and in a set combined with laboratory isolates. Nucleotide positions are according to the 3D7 line sequence. Window length is 90 bp, and step size is 3 bp. n indicates the number of samples analyzed. Asterisks indicate the region where a significant excess of nonsynonymous substitutions (dN) over synonymous substitutions (dS) was observed (single for p<0.05, and double for p<0.02). The statistical difference between dS and dN was tested using a one-tailed Z-test with 500 bootstrap pseudosamples implemented in MEGA4.0. Others indicate sequence reported previously and shown in Fig. 2 [14].

3.2. Population-based analyses also detected positive selection on group 1 clag2 and clag8

Positive diversifying selection was evaluated for group 1 clag2 (n = 24), clag8 (n = 41), and clag9 (n = 34) using sequences obtained from Thailand within two months by a population-based approach. Using Tajima’s test and Fu and Li’s test for nt sequences obtained in this study (nt 3106–3642 of clag2, 3022–3591 of clag8, and nt 2947–3639 of clag9; nucleotide positions are according to the 3D7 line sequence), no significant departure from the neutral expectation was detected. However, the sliding window plot method detected a significantly high Tajima’s D value for group 1 clag2 (mid point nt position of 3492, D = 2.26, p<0.05) and clag8 (mid point nt position of 3363, D = 2.07, p<0.05), Fu and Li’s D* value for clag8 (mid point nt position of 3258–3264, D* = 1.48–1.56, p<0.05; mid point nt position of 3267, D* = 1.61, p<0.02), and Fu and Li’s F* value for group 1 clag2 (mid point nt position of 3375, F* = 1.64, p<0.05) and clag8 (mid point nt position of 3261–3270, F* = 1.75–1.81, p<0.05; mid point nt position of 3363, F* = 1.86, p<0.05) (Fig. 4). Thus, in addition to the detection of positive selection on group 1 clag2 and clag8 detected by an excess of dN over dS, the results of the population-based method also support the finding that both group 1 clag2 and clag8 are subject to a positive diversifying selection.

Fig. 4.

Sliding window plot of Tajima’s D, Fu & Li’s D* and F* tests for Plasmodium falciparum clag2 (group 1) and clag8 genes in Thai isolates. Nucleotide positions are according to the 3D7 line sequence. Window length is 90 bp, and step size is 3 bp. n indicates the number of samples analyzed. Asterisks indicate the region where a significant departure from the neutrality was observed (single for p<0.05, and double for p<0.02).

In summary, two independent tests, one by comparing dN and dS and the other based on the population, suggest that the region of clag2 is under positive diversifying selection. Previously suggested positive selection on clag8 was also supported by population-based analyses. Such observations are consistent with the action of balancing selection maintaining allelic variation in the population. Contrary to a previous report [14], positive selection on clag9 was not detected in this study using an additional 34 sequences. As only 5 clag9 sequences were used to compare dN and dS in this previous study, a low sample number might have generated a false positive result. Further study is required to determine if clag9 is under positive selection.