INTRODUCTION

Top Abstract Introduction Materials & Methods Results Discussion References

The development of a host-protective immune response against Plasmodium falciparum takes several years and many episodes of infection, at least for children living in areas where malaria is endemic. One of the reasons for this is believed to be the large number of distinct parasite strains circulating within an area of endemicity and the assumption that exposure must occur to a sufficiently large sample of these before lasting immunity is induced. However, the detailed epidemiology of endemic malaria infection remains poorly understood at the molecular level, and there is surprisingly little nucleotide sequence data to support the concept of a large repertoire of antigenically distinct strains.

There are at least six antigenically diverse proteins of the asexual stage that are known to be the target of potentially protective host responses. The definition of antigenically distinct strains involves identification of the allelic form expressed at all antigenically diverse locithe extended antigenic haplotype. The loci would include merozoite surface protein-1 to -3 (3), apical membrane antigen-1 (17), S-antigen (6), and P. falciparum erythrocyte membrane protein-1 (PfEMP-1) (5). Such a complete molecular definition of infecting parasites is a highly ambitious task, particularly in the case of blood samples collected from patients harboring mixed infections. Accordingly, most studies focus on one or other of the antigenically diverse antigens. We have elected to study merozoite surface protein-2 (MSP2) (27), a 45- to 50-kDa glycoprotein anchored in the merozoite surface by a glycosylphosphatidylinositol anchor. This surface protein is a promising candidate for inclusion in a malaria subunit vaccine, as both in vitro and in vivo studies have demonstrated the ability of immune responses to MSP2 to inhibit parasite multiplication (23, 25). However, the efficacy of any subunit vaccine containing a single form of MSP2 may be limited by the presence of antigenically distinct parasite strains within an area of endemicity. We will adopt the recently proposed convention for parasite genes and gene products of denoting the gene sequence as MSP2 and the protein as MSP2.

Sequence polymorphism has been described for MSP2 genes of both laboratory-maintained isolates (29) and field isolates (14, 16, 19, 30). Comparison of MSP2 gene sequences from these isolates reveals highly conserved 5' and 3' sequences that flank a central variable region. This central region is composed of repeats flanked by nonrepetitive sequences. The nonrepetitive sequences are one or other of two distinct forms that define two allelic families, FC27 and IC-1/3D7 (29). The central repeats are more variable and define the individual alleles of MSP2. There is a correlation between the general form of the central repeat sequence and the allelic family. For example, FC27 family members have variants of a central 96-bp pair sequence that may be present in one to four copies followed by a 21-bp partial repeat and a variably represented 36-bp sequence that may be present in one to five copies. In contrast, alleles belonging to the 3D7 family show a central repeat region made up of variable numbers of 12- to 24-bp repeats separated by repeating 6-bp sequences.

Field studies aimed at defining the antigenic diversity of MSP2 have approached the problem by determining MSP2 gene structure by various forms of PCR. The rationale for this is that P. falciparum is haploid and MSP2 has been shown to be present in all laboratory and field isolates examined (8-10, 15). Most studies examining the distribution and frequency of different allelic forms of MSP2 have enumerated the presence of the allelic families (11, 12, 14). Whereas a skewed distribution of predominantly 3D7 family alleles exists among laboratory-adapted strains, in the field a more even distribution of FC27 and 3D7 alleles occurs. Often FC27 family alleles are more prevalent than 3D7 alleles, and novel FC27 and 3D7 family alleles have been found in field malaria strains (14, 16). Some field studies examining recurrent MSP2 infections have been performed, but these have classified MSP2 alleles on the basis of family and length of the central repeat region (7, 11, 14). This makes it difficult to form conclusions about the repertoire of repeat sequences in the circulating pool of parasites and to infer the possible action of immune responses to MSP2 repeats. We were interested to examine the sequences of MSP2 alleles circulating in an area of endemicity over time and to determine the persistence of various MSP2 alleles within a localized area. This study describes MSP2 genotypes from malaria-infected inhabitants of the Oksibil region of Irian Jaya and allows comparison of the variation in MSP2 sequences seen over a 2.5-year period within the region as a whole and in particular individuals.

	MATERIALS AND METHODS

Top Abstract Introduction Materials & Methods Results Discussion References

Collection of field isolates and isolation of genomic DNA. The location and characteristics of the study site have been described previously (4). Blood samples were taken with informed consent from individuals living in the Oksibil region of Irian Jaya. Seventeen samples were taken from villagers in June 1993 (Oksibil time point 4), a number of whom had been previously bled in January 1991 (Oksibil time point 1). Subjects were selected at random on the day of sampling and were not selected for symptoms suggesting malaria. Similar blood collections had also taken place in January 1992 (Oksibil time point 2), at which time 15 subjects were venisected, and in September 1992 (Oksibil time point 3), when 14 subjects were bled. These samples were also analyzed. We have adopted a nomenclature that enables us to refer to all MSP2 alleles in a consistent manner. It consists of the letter O followed by the bleed number/sequential numbering of alleles in a particular family/the family name. Therefore, we have renamed the strains found in the first Oksibil sample. OKS1 group is now designated O1/1/FC27, OKS11L is now O1/2/FC27, OKS11U is now O1/1/3D7, OKS10U is now O1/2/3D7, and OKS6 is now O1/3/3D7.

MSP2 gene and allelic family-specific PCR. Amplification of full-length MSP2 genes was carried out in nested PCRs using oligonucleotide primers of the 5' and 3' conserved regions. The initial amplification using the outermost 5' and 3' primers was carried out on 5 µl of genomic DNA in 50-µl reaction volumes containing 5 mM Tris HCl (pH 8.3), 50 mM KCl, 0.5 mg of gelatin per ml 2.5 mM MgCl₂, 200 µM mixture of deoxynucleoside triphosphates (Pharmacia Biotech, Uppsala, Sweden), 1 µM primers p359 and p360 (Fig. 1), and 1.25 U of Taq polymerase (Boehringer, Mannheim, Germany). Forty cycles of amplification composed of 94°C for 1 min, 55°C for 1 min, and 72°C for 45 s, followed by a final prolonged extension step at 72°C for 5 min, were carried out in a C model FTS-320 thermal sequencer (Corbett Research, Sydney, New South Wales, Australia). Second-round amplification was carried out in 50-µl reactions using nested conserved region 5' and 3' primers, p357 and p358 (Fig. 1), to amplify the central region of the MSP2 gene. One microliter of the amplification products from the first-round reaction was used as the template for the nested reaction using the same reaction mix as specified above. Thirty-five cycles of amplification composed of 94°C for 1 min, 55°C for 1 min, and 72°C for 45 s and then one cycle of 72°C for 5 min were carried out for this and all other nested PCRs.

View larger version (23K): [in this window] [in a new window]   FIG. 1.   Locations of MSP2 oligonucleotide primers used in PCR of the MSP2 gene from field isolates. The schematic shows the locations of the primers on representatives of each gene family.

Sequencing methods. PCR products were extracted from the overlying mineral oil, using a chloroform-isoamyl alcohol (24:1) mixture. The amplicon DNA was ethanol precipitated and then prepared for blunt-ended cloning by T4 polynucleotide kinase and T4 DNA polymerase reactions. Product DNA was purified by using a BRESAclean kit (Bresatec, Adelaide, South Australia, Australia). The PCR products were then ligated to pUC18 that had been digested with SmaI. Recombinant plasmids were transformed into Escherichia coli DH5, using a Bio-Rad (Sydney, New South Wales, Australia) Gene Pulser according to the manufacturer's instructions. Transformants were selected on LB medium containing ampicillin and 5-bromo-4-chloro-3-indolyl--D-galactoside. These were further analyzed by restriction endonuclease digestion and prepared for sequencing. If cloned products were used, then at least two independently prepared templates were sequenced for each MSP2 allele, and a third clone was sequenced if any differences were found between the two derived sequences. Alternatively, PCR products were sequenced directly after gel purification. Sequencing reactions were carried out in a thermal cycler using dye terminator reaction and were analyzed with a model 373 DNA sequencer (PE-ABI, Foster City, Calif.).

Nucleotide sequence accession numbers. Nucleotide sequences described in this report have been submitted to GenBank under accession no. U72948 to U72957.

	RESULTS

Top Abstract Introduction Materials & Methods Results Discussion References

Blood samples were collected from 17 individuals living in the Oksibil hamlet in June 1993, and the extracted DNA from each blood sample was subjected to PCR amplification with MSP2-specific primers. As in our previous studies, all field samples required at least two rounds of PCR amplification to produce a visible product by ethidium bromide staining of agarose gels. Three separate types of PCR analysis were then performed on these first-round products (Fig. 1): (i) amplification of near-full-length MSP2, using primers to the conserved region, (ii) DiFS analysis (21) that amplified family-specific products by using 3' primers specific to one or other of the two MSP2 families, and (iii) an amplification of the repeat region by using nested primers that would provide family-specific products. Amplified products of the central repeat regions were either sequenced directly or sequenced following cloning, depending on product yield. Details of the patient population and the MSP2 alleles found are shown in Table 2.

Eleven of the patient samples gave detectable products after two rounds of MSP2 amplification, and the six other samples were repeatedly negative after reactions using a number of primer pairs. A total of 18 MSP2 genes were amplified from the 11 positive blood samples, with seven subjects having dual infections with members of both FC27 and 3D7 allelic families present. There was agreement between the results of the allelic family determinations on the basis of DiFS PCR (21) and deduced protein sequences in all samples studied. All MSP2 sequences could be assigned to either the FC27 or 3D7 allelic family. A schematic illustrating the structures of the various MSP2 alleles found at this time point is shown in Fig. 2. An alignment of the protein sequences is shown in Fig. 3.

View larger version (40K): [in this window] [in a new window]   FIG. 2.   Schematic illustrating the deduced peptide sequences of the central repeats of MSP2 proteins identified in the Oksibil area at time points 1 and 4. Panels A and B show alleles of the FC27 and 3D7 families, respectively, from time points 1 and 4. The number of patients in which each allele was found is shown in parentheses to the right of the allele name. A key listing the sequence of each repeat is shown below. Residues identical to the canonical sequence for each repeat type are indicated by dots.

View larger version (66K): [in this window] [in a new window]   FIG. 3.   Alignment of the sequences of all deduced MSP2 proteins of all Oksibil samples. Gaps have been inserted for the purpose of optimizing alignment and are indicated by dots. Members of the same family are grouped.

Ten FC27 family alleles and eight 3D7 family alleles of MSP2 were detected. Comparison of nucleotide sequences and deduced protein sequences showed a marked preponderance of malaria strains with the same FC27 family MSP2 gene designated O4/1/FC27 (Table 2 and Fig. 2 and 3). This MSP2 allele was present in nine patients (six of seven patients with dual infections and three of four patients with a single infection). The other FC27 family MSP2 allele was found in only one patient sample. Both of these FC27 family MSP2 genes were novel sequences, not having been previously described (Fig. 3). The eight 3D7 family alleles (Fig. 2 and 3) consisted of O4/1/3D7, a sequence identical to the previously described IC-1 allele (29) found in five patient samples, O4/2/3D7, a sequence closely resembling the previously described 3D7 allele in two patient samples (29), and one novel allele, O3/1/3D7 (which was also found in a patient sample from the Oksibil time point 3 [see below]).

The dominant FC27 family strain, O4/1/FC27, was found to have three 32-mers and one 12-mer in the central amino acid repeat region. The first 32-mer differed from the canonical FC27 32-mer by the substitution of a serine for an arginine at residue 10. The other two 32-mers were identical to the canonical FC27 32-mer repeat subunit. A single 12-mer sequence was found downstream of the 32-mer repeats. The other FC27 family allele, O4/2/FC27, had one 32-mer with two substituted residues (asparagine instead of aspartic acid at residue 2 and serine for arginine at residue 10) compared with the FC27 32-mer. This was followed by two 12-mers identical to the 12-mer repeat found in the previously described K1 isolate, which by contrast has five copies of this 12-mer (28). The 3D7-like sequence, O4/2/3D7, found in two patient samples, varied from the published 3D7 protein sequence by the presence of a serine instead of a threonine at the beginning of the 4-mer repeats (GAGGSAGG instead of GAGGTAGG). The other novel 3D7 family allele, O3/1/3D7, had seven 8-mer repeats that consisted of four copies of GAGAVAGS and one copy of GAVAGAGS, GAVAGSGA, and GAGASAGN. These were separated by one GS and two GA dimers (Fig. 2 and 3).

Comparison of MSP2 sequences collected 29 months apart. We had first collected blood in this hamlet some 29 months previously, in January 1991, and have reported on the MSP2 alleles found in those samples (16). At the first collection, we detected 12 MSP2 genes, of which 9 were FC27 family sequences and 3 were 3D7 family sequences. At the first time point, infected individuals carried an average number of 1.2 strains, as judged by MSP2 allelic markers, whereas the average number of strains found in infected individuals at the later time point was 1.6. At both time points, a dominant FC27 allele was detected in patients infected with strains expressing MSP2 of that family, whereas several different 3D7 family alleles were present in smaller numbers of patients. Interestingly, comparison of the MSP2 sequences found at these two time points reveals no sequences in common. Thus, the repertoire of circulating strains had completely changed in this area over the course of 29 months (Table 2 and Fig. 2).

MSP2 diversity at intervening times in the Oksibil area. We were interested to discover if this difference in the MSP2 of reinfecting strains was a more general phenomenon, and so we examined two sets of blood samples that had been collected at time points intermediate between January 1991 and June 1993 (Table 2). We examined MSP2 sequences in a subset of these samples where we had at least one other bleed from the same patient, and the results are presented in Table 2 and Fig. 4. Very few of the patients bled had detectable parasitemias, as conditions had been unseasonably dry that year. Five samples from Oksibil time point 2 were found to contain parasites, and from these it was possible to identify one patient infected with a strain bearing a single FC27 allelic family form of MSP2, one patient with a mixed FC27-3D7 allelic infection, and three patients infected with single 3D7 allelic family strains. All of the individual alleles collected at this time differed from one another. Ten samples were analyzed from Oksibil time point 3, and only three were found to be parasitemic. One single FC27 allelic family infection and two single 3D7 allelic family infections were found. As at the Oksibil time point 2, all collected alleles at this time point differed.

View larger version (54K): [in this window] [in a new window]   FIG. 4.   Schematic showing comparison of the central repeats of MSP2 found in individual patients who were parasitemic at more than one time point in the survey. The representation of individual repeat blocks uses the key shown in Fig. 2.

	DISCUSSION

Top Abstract Introduction Materials & Methods Results Discussion References

Several interesting conclusions may be drawn from this study of malaria strains in the Oksibil valley, set in a highland region subject to hyperendemic malaria (4). Considering all patient samples examined in this study, we have seen a total of 15 different MSP2 alleles, 9 of which have not been previously described, out of a total of 39 parasite isolates. Only four of these alleles were detected at different time points, suggesting that there are many circulating forms of this antigen that can be found in the Oksibil region. A number of studies have now examined the question of repertoire size of MSP2 in different areas of endemicity. The study by Felger et al. (14) identified 38 different MSP2 alleles in 144 infections, of which 33 had not been previously described at the time of the study. Their analysis was performed by using restriction fragment length polymorphism analysis of PCR products and would thus rely on increase or decrease in numbers of repeats or the introduction or removal of one of the five restriction sites examined to recognize differences between alleles. Our sequencing studies demonstrate that there exist allelic variants of the FC27 family of MSP2 that have mutations in the 96-mer nucleotide repeat region that would not be detected by the restriction enzymes used by Felger et al. (14). Thus, that study is likely to represent an underestimate of the MSP2 repertoire. Taking these studies together with others examining MSP2 diversity (11, 18, 19) leads to the conclusion that the total circulating repertoire of MSP2 is likely to be quite large, probably at least 100 to 200 distinct sequences.

In the Oksibil area, the general distribution of families and alleles seems constant. At the two time points analyzed in detail, the majority of patients were infected with FC27 family parasites, and among them, a dominant MSP2 type was evident. Although 3D7 family alleles were detected in fewer patients at both time points, the number of circulating alleles was much greater than for FC27 family alleles. We found no patients who were multiply infected with parasites containing more than one allele from the same family; i.e., all mixed infections were double infections with an FC27 family and a 3D7 family parasite. Both of these results are similar to those of Felger et al. (14) in Papua New Guinea. These studies analyzed samples collected on the same island at roughly the same time, the early 1990s; thus, it is not possible to conclude whether this is a general feature of field infections with P. falciparum. Further studies in other geographical areas will be needed to address this point.

The number of MSP2 alleles found in an individual can be used to estimate the number of distinct parasite strains infecting that individual. This is likely to be an underestimate of the true number of strains, as only one polymorphic locus was examined. On average, individuals carried between 1.2 and 1.6 different strains, depending on the time point examined. This was less than found in a survey of residents of Dielmo in Senegal, where adults carried between 1.9 and 2.3 distinct strains, as judged by detection of differing repeat region fragment lengths of MSP2 (18). Given the greater sensitivity of the sequencing method used in this study, distinctly fewer strains were found in infected individuals in Oksibil. The reasons for this are unknown.

This is the only study to our knowledge that has examined the sequence of MSP2 genes in a single location at different times. A key finding of this study was that the variety of MSP2 genes found in this hamlet change with time. We found four examples where a particular allele could be found at more than one time point. However in general, the repertoire of circulating MSP2 alleles changed with time, and at the last time point sampled, there were no MSP2 sequences in common with those present some 29 months previously. The reasons for this are unknown. It may be due to any or all of a number of factors relying on peculiarities of the host, the parasite, or the vector. The possibility that most intrigues us is that during infection, the patient makes a strain-specific antibody response to the MSP2 of the infecting strain and this has an inhibitory effect on parasite growth. Growth inhibition of P. falciparum by anti-MSP2 antibodies has been found in both in vivo and in vitro studies (23, 25). When a patient is reinfected, growth of a parasite with a different MSP2 would be favored, while that of parasites expressing a previously seen MSP2 would be inhibited. Such selection would influence the repertoire of circulating strains over time, as seen in this study. Clearly the magnitude of this effect would depend on the duration of such strain-specific immunity. If anti-MSP2 immunity was short lived, then reappearance of strains within a community would be expected to be rapid. A corollary of this suggestion is that we would not expect to see patients reinfected with parasites expressing the same form of MSP2 as previously found in that patient. In this study, we have only four examples of infections by parasites containing previously detected MSP2 alleles. In all four cases the infections were not in patients who had seen that form of MSP2 previously.

Studies of naturally acquired antibodies to MSP2 performed in Papua New Guinea demonstrate that significant numbers of people in the study area have antibodies only to the repeat regions of MSP2 and that antibodies to the conserved regions of this protein develop at later ages after more prolonged exposure to malaria (1, 2). Similarly, Taylor et al. reported that the anti-MSP2 response occurring in Gambian adults was directed almost exclusively to polymorphic regions of the protein which consisted of the family-specific regions and the repeats (31). Thus, the repeat regions appear to be the immunodominant regions of the protein. It is clear that anti-repeat regions antibodies may inhibit parasite growth (13), but such antibodies may be much less effective against variant repeats. For example, it has been reported that the growth-inhibitory monoclonal antibody 8G10/48 (13), which recognizes the epitope STNS in some variant 32-mers of the FC27 family sequences of MSP2, is much less effective at inhibiting parasites expressing other forms of the antigen (24). Furthermore, relatively minor sequence changes in MSP2 proteins can have quite major effects on antibody reactivity. For example, Ranford-Cartwright and coworkers (20) demonstrated that patient antibodies react more strongly with FC27 family MSP2 sequences expressing five copies of a 12-mer repeat than with MSP2 sequences in which only one copy of the repeat was found. Such differences in reactivity could well lead to differing degrees of growth inhibition of parasites expressing these variant forms of MSP2. In our study, subject G had at time point 3 an FC27 allele that varied from that present at time point 4 by the deletion of one 32-mer subunit. On the basis of the work of Ranford-Cartwright et al. (20), one would predict that there would be decreased antibody binding in this patient to this form of the antigen. An additional factor favoring reinfection would be the waning of specific antibody levels during that time.

This study was designed to examine the genetic and hence antigenic changes in MSP2 repertoire over time in a particular community. No specific effort was made to repeatedly bleed the same individuals; thus, we do not have a sample for the analysis of reinfecting strains large enough to be statistically significant. Further studies are needed in other areas with differing patterns of malaria transmission to determine if the findings reported here on repertoire change are general. We believe that it would be extremely important to design specific studies of blood samples collected from single individuals over shorter time intervals to determine if there is further evidence for resistance to reinfection by parasites expressing an identical MSP2. Genetic studies such as this need to be complemented by functional studies examining the binding of and effect on parasite growth of antibodies directed to MSP2 variants. Currently, significant technical difficulties hamper the performance of such studies, but they are urgently needed to further our understanding of the distribution of MSP2 alleles circulating in areas where malaria is endemic.