A full-length recombinant Plasmodium falciparum PfRH5 protein induces inhibitory antibodies that are effective across common PfRH5 genetic variants

Highlights ► PfRH5 is essential for Plasmodium falciparum erythrocyte invasion. ► Antibodies raised against full-length functional PfRH5 potently blocked invasion. ► Five common PfRH5 polymorphisms were identified across 290 clinical isolates. ► Antibodies raised against one variant inhibited all other common PfRH5 variants. ► Correctly folded recombinant PfRH5 is a strong P. falciparum vaccine candidate.


Introduction
The development of an effective Plasmodium falciparum vaccine is the topic of intense research. Rational arguments can be advanced to support vaccines that target pre-erythrocytic or erythrocytic stages in humans, or parasite development within the mosquito [1]. The number of stages that could be targeted is compounded by the variety of potential approaches [2], which range from genetically or physically attenuated parasites [3], to subunit-based vaccines, of which there are hundreds of theoretical targets within the genome [4]. Regardless of the approach, however, the fundamental requirement for an effective malaria vaccine is the same -to effectively protect the largest number of people against the widest possible range of P. falciparum genetic variants.
Vaccines targeting the erythrocytic stage have their origins in the observations that passive transfer of immunoglobulins from immune to non-immune individuals markedly reduced parasitemia [5], and that immune adults directly inoculated with blood stage parasites avoid clinical symptoms, emphasizing the protective effect of antibodies targeting blood-stage antigens [6]. Several P. falciparum proteins have been suggested as potential bloodstage vaccine antigens, either because of their function, localization on the merozoite surface, or because data from immuno epidemiological studies suggest that they are targets of protective immunity [7,8]. All face the same fundamental challenge, however; by being directly exposed to the human immune system during natural infections, they are under considerable selective pressure, which can result in high levels of polymorphism [9], meaning that vaccine-induced immunity must be able to protect against multiple genetic variants. Although this is a significant challenge, the availability of genome sequences for hundreds of P. falciparum isolates [10] now makes it possible to measure the true scale of the challenge by identifying for each antigen which residues are variable, and in which populations.
P. falciparum Reticulocyte Binding Protein Homologue 5 (PfRH5) is a member of the super family of erythrocyte ligands referred to as the Reticulocyte Binding Like proteins (RBLs), at least one member of which is found in every Plasmodium genome [11]. PfRH5 differs from the other members of the PfRH family in that it is much smaller and the gene encoding PfRH5is refractory to genetic deletion in all tested strains, implying that it is essential for blood-stage growth [12]. PfRH5 bindserythrocytes and is implicated in the species tropism of erythrocyte invasion [12][13][14], and our laboratories have recently established that the receptor for PfRH5 is basigin/CD147 [15]. Of key importance for vaccine development is that unlike all other known erythrocyte-merozoite receptor-ligand interactions, the basigin-PfRH5 interaction was essential for erythrocyte invasion in every P. falciparum strain tested to date [15].
Its role in a universally required merozoite-erythrocyte interaction clearly highlightsPfRH5 as a promising vaccine target. However, early studies using antibodies raised against sub fragments of recombinant PfRH5 expressed in E. coli showed no [14] or very limited [12] efficacy in blood stage growth inhibition assays (GIAs). By contrast, immunization with recombinant viruses encoding full-length PfRH5 produced antibodies that cross-protected against multiple P. falciparum strains [16] suggesting that the conformation of the PfRH5 protein used as an antigen was critical for this activity. Here, we set out to establish whether full-length, functional, PfRH5 protein could induce growth-inhibiting antibodies. At the same time, we performed a systematic investigation of the impact of genetic diversity on invasion inhibitory activity, by testing anti-PfRH5 antibodies against P. falciparum strains that contained all common genetic variants of PfRH5, as defined in sequencing study of 227 P. falciparum clinical isolates [10]. The data has important implications for the vaccine candidacy of PfRH5.

Parasite sequence variant analysis
Single nucleotide polymorphisms (SNPs) and allele frequencies were obtained from a published analysis of 227 clinical P. falciparum infections [10]. MalariaGen SNP catalogue release 1.0 data was used for all analysis (www.malariagen.net/data). PfRH5 sequence was verified for all strains used (see Supplement).

Protein expression plasmids, production and purification
Expression of the P. falciparum 3D7 PfRH5 ectodomain was described previously [15], and 7G8 and GB4 PfRH5 variants and 3D7 AMA1 were made and expressed according to the same principles. The basigin-L ectodomain was expressed as a soluble Cd4d3+4-hexa-his tagged protein, as described [17]. All proteins were produced as secreted proteins by the human HEK293E cell line grown in suspension and transiently transfected at a density of 5x10 5 cells per mL as described [17,18]. His-tagged proteins were purified from spent supernatants after transient transfections on HisTrap TM HP columns (GE Healthcare) using an ÄKTAxpress TM (GE Healthcare).

Surface plasmon resonance
Surface plasmon resonance (SPR) studies were performed using a BIAcore TM T100 (GE Healthcare). Monomeric biotinylated bait proteins were immobilized on a streptavidin-coated sensor chip (GE Healthcare) and increasing concentrations of purified basigin was injected at 20 l/min to determine equilibrium measurements or 100 l/min for kinetic parameters. Both equilibrium and kinetic binding data were analyzed in the manufacturer's BIAcore TM T100 evaluation software (GE Healthcare).

Enzyme-linked immunosorbant assay (ELISAs)
Monobiotinylated proteins were captured on streptavidincoated plates (Nunc) and then incubated for 90 min with 1 g/ml primary antibody from which anti-Cd4 reactivity had been removed by preadsorbing with Cd4 protein. Alkaline phosphataselinked secondary antibodies and appropriate detection reagents (Sigma) were used, and optical density measurements taken at 405 nm (Pherastar plus, BMG laboratories).

Antibodies
Polyclonal antisera against 3D7 PfRH5 and AMA1 were raised commercially and in parallel (Cambridge Research Biochemicals, Billingham, UK; immunization regime listed in Supplement). The use of animals to raise antisera was performed according to UK Home Office governmental regulations and approved by the local Sanger Institute ethical review board. Total IgG from the sera were purified on Hi-Trap TM Protein G HP columns (GE Healthcare) according to the manufacturer's instructions.
2.6. P. falciparum culture and invasion assays P. falciparum parasites were cultured in human O+ erythrocytes at 5% hematocrit. Field strains were confirmed to contain only a single strain using PfMSP1 and PfMSP2 genotyping. Use of erythrocytes and serum from human donors was approved by the NHS Cambridgeshire 4 Research Ethics Committee, and all donors supplied written informed consent. For invasion assays, synchronized parasites were incubated in the presence or absence of antibodies for 24 h. Invasion assays were carried out in 96 well plates, with 3 identical wells set up for each strain/antibody concentration combination. Cultures were then fixed, and parasitized erythrocytes counted using SYBR Green I (Invitrogen) and flow cytometry, as described previously [19]. Invasion efficiency was calculated by comparing invasion in the presence of a given antibody concentration to invasion in the absence of antibodies.

Genetic diversity in PfRH5 is limited
Whole genome sequencing by the MalariaGen consortium of 227 P. falciparum clinical isolates (125 samples from Africa (Mali, Burkina Faso and Kenya), 81 samples from Southeast Asia (Thailand and Cambodia) and 21 samples from Papua New Guinea [10]) identified fourteen polymorphisms in the PfRH5 gene, two of which are synonymous and twelve non-synonymous (Table 1). This excess of non-synonymous over synonymous mutations is consistent with the overall trend of mutation rates in the P. falciparum genome and the specific increase in non-synonymous substitutions in proteins exposed on the merozoite surface [20]. Of the twelve nonsynonymous mutations, only five SNPs reach a frequency greater than 10% in at least one population (Y147H, H148D, S197Y, C203Y and I410 M; marked * in Table 1), and can be considered the most common global variants in PfRH5 that have been currently identified. Of these five SNPs, three are found in previously sequenced laboratory-adapted P. falciparum strains, but Y147H and H148D, which reach 30% frequency in Southeast Asian clinical isolates, are not present in any lab strain sequenced to date (Table 1). We identified two isolates from Cambodia that contained these mutations for use in this study. PfRH5 sequences from all strains was confirmed by sequencing, and the full haplotype of each strain is shown in Supplementary Table 1.

Expression and purification of biochemically active full-length recombinant PfRH5 protein variants
The chances of raising an effective polyclonal antibody response are likely to be significantly higher if the recombinant antigen closely mimics the conformation of the native antigen [21]. Previously, we have shown that biochemically active PfRH5 can be expressed as a secreted full-length protein using a mammalian expression system [15]. The 3D7 variant of PfRH5 was used in these initial experiments, which is notable because it contains six cysteine residues of which C203 is a relatively rare variant, meaning most strains express PfRH5 with five cysteines ( Table 1). As sulfhydryl groups of cysteines are normally oxidized to form structurally critical disulfide bonds in secreted proteins, polymorphism at C203 suggests that cysteine residues in some variants of PfRH5, most likely at positions 203 and 329, are unpaired [12].
To determine whether PfRH5 variants that contain five cysteines were biochemically active, we expressed 7G8 and GB4 variantsin HEK239E cells as Cd4d3+4-tagged monobiotinylated "bait" proteins [17]. Cells were transiently transfected, and all variants produced 2-10 g of protein per ml of cell culture supernatant 6 days post-transfection. These yields were produced using a cost-effective transfection approach, which typically resulted in a transfection efficiency of ∼40%, and could presumably be increased in stably expressing isolates.
To test variants for biochemical activity, we expressed and purified the extracellular region of human basigin as a soluble Cd4d3+4-6H-tagged protein and used surface plasmon resonance to quantify interaction affinities. Serial dilutions of basigin were injected over PfRH5 immobilized on a sensor chip until equilibrium was reached (Fig. 1A). Saturable binding was observed, and the biophysical binding parameters were very similar in all cases (Supplementary Table 2). The association (k a ) and dissociation (k d ) kinetic rate constants were in good agreement with the equilibrium analysis (Supplementary Table 2), providing an independent measure of the interaction parameters. The K D was essentially identical to that reported for the same experiment performed in the reciprocal orientation using purified 3D7 PfRH5 [15] suggesting that the proportion of biochemically active protein was high.

Generation of polyclonal antisera and testing against PfRH5 variants
To raise polyclonal antisera against full-length 3D7 PfRH5, we expressed and purified the protein as previously described [15]. Purified PfRH5-Cd4d3+4-6H runs as two bands at ∼85 and ∼65 kDa (Fig. 1B), which, when the size of the Cd4 tag (∼25 kDa) is taken into account, corresponds well to the two forms of native PfRH5 (63 and 45 kDa) observed in parasite cultures [12], suggesting the same proteolytic processing occurs when expressed in HEK293 cells as in the parasite. We were able to resolve the 85 kDa band from the 65 kDa form by size-exclusion chromatography (Fig. 1B). Purified PfRH5 protein showed no measurable loss of activity when stored at 4 • C for several months (unpublished observations). Rabbits were immunized with 1 mg of purified 3D7 PfRH5-Cd4d3+4-6H protein. To remove anti-Cd4 antibodies, the antisera were preadsorbed with a Cd4-tag-containing protein and shown to then lack anti-Cd4 immunoreactivity by ELISA ( Fig. 2A). Preadsorbed total IgG was then used in ELISAs to quantify the immuno reactivity against both the immunizing 3D7 antigen, and cross-reaction with the 7G8 and GB4 variants. No difference in immuno reactivity to any of the three variants was observed ( Fig. 2A). We heat-treated 3D7 PfRH5 and observed a loss of immuno reactivity suggesting that a good proportion of the antibodies recognized heat-labile conformational epitopes (Fig. 2B).

Antibodies raised against the 3D7 PfRH5 variant inhibit parasites that collectively include all currently identified common PfRH5 genetic variants
To establish whether antibodies raised against the full-length 3D7 variant of PfRH5 could inhibit erythrocyte invasion, total IgG was purified from anti-PfRH5 rabbit polyclonal serum and added in increasing concentrations to late trophozoite stage P. falciparum parasite cultures. Parasites were cultured in the presence of IgG for 24 hours before parasitemia was measured using a fluorescent DNA dye [19]. Anti-PfRH5 IgG had a clear dose-dependent inhibitory effect on the growth of 3D7 parasites, with an IC50 of 1.9 mg/ml total IgG (Fig. 3A). FCR1 was chosen for initial strainspecificity trials because it contains two of the five common PfRH5 SNPs (Table 1). Dose-response curves for inhibition of FCR1 parasites were similar to those against 3D7, with an IC50 of 3.1 mg/ml total IgG. (Fig. 3A). By contrast, antibodies raised against the entire ectodomain of the 3D7 variant of AMA1 had a marked strainspecific effect, with only a 40% reduction in FCR1 growth even at the maximum concentration of 10 mg/ml (Fig. 3B), consistent with the known strain-specific properties of anti-AMA1 antibodies [22]. Neither strain was inhibited by control pre-immune rabbit total IgG (Fig. 3C).
To more comprehensively test the inhibitory effect of antibodies raised against the 3D7 variant of PfRH5, we tested seven additional P. falciparum isolates, that between them included PfRH5 sequences that encompass the five common PfRH5 SNPs. The effect of anti-PfRH5 on 3D7 was re-measured in every replicate as a positive control. The dose-response curves for laboratory (Fig. 4A), Kenyan (Fig. 4B) and Cambodian strains (Fig. 4C) were broadly comparable with that of 3D7. As expected, the inhibitory effect varied slightly Fig. 1. Surface plasmon resonance demonstrates biochemical activity of recombinant full-length 7G8 PfRH5-Cd4. (A) 7G8 PfRH5-Cd4-biotin was specifically captured in a flow cell of a streptavidin-coated chip using an approximate molar equivalent of Cd4-biotin immobilized in a neighbouring flow cell used as a reference. Serial dilutions of purified basigin-Cd4-6H were injected until equilibrium was reached (solid bar, inset) and reference-subtracted data were used to plot the binding responses once equilibrium had been achieved. An equilibrium dissociation constant was calculated using non-linear regression fitting of a simple Langmuir binding isotherm to the data (solid line). Note that the binding was saturable and therefore specific and a KD of 0.9 ± 0.13 M (mean ± SEM) was calculated from two independent experiments (see Supplementary  Table 2). A representative experiment using 7G8 is shown, equilibrium constants for 3D7 and GB4 were calculated in the same way and are reported in Supplementary Table  2. (B) 3D7 PfRH5-Cd4-d3+4-6H was purified using Ni 2+ -NTA-Sepharose and resolved as two bands (∼85 and 65 kDa) using using Coomassie staining after SDS-PAGE under reducing conditions. The 85 kDa protein, which is consistent with the expected size of the full-length PfRH5-Cd4-d3+4-6 H protein, could be separated from the smaller 65kDa band using size exclusion chromatography (post-SEC).
between independent experiments due to biological variation, such as different starting growth rates. To enable a quantitative comparison between experiments, the IC50 for each test strain was compared to that of the 3D7 positive control reference strain performed in parallel. IC50 ratios are listed in the figure legends, and range from a low of 0.53 (Kenya KCC103) to a high of 1.6 (FCR1).
Statistical analysis (f-test) established that no IC50 represents a significant variation from 3D7 (p = 0.623). Serum from a second rabbit immunized with full-length 3D7 PfRH5 were also tested and showed no significant difference in growth inhibition between 3D7 and FCR1 parasites (data not shown). Antibodies raised against the 3D7 PfRH5 variant therefore have broad inhibitory activity across all currently identified common genetic variants, with no evidence for significant variant-specificity.

Discussion
The essentiality of the PfRH5-basigin interaction for erythrocyte invasion across all strains tested [15] have highlighted PfRH5 as a particularly promising blood-stage vaccine candidate. The development of subunit malaria vaccines has, however, been partly limited by the technical difficulties of expressing biochemically active Plasmodium proteins in heterologous expression systems [23]. While viral delivery systems could be considered [24], bloodstage vaccines based on proteins could be more easily combined with vaccines targeting other stages of the Plasmodium lifecycle such as RTS, S [25]. Using a mammalian expression system, we have been able to produce full-length PfRH5 protein and unambiguously demonstrate its biochemical activity by showing saturable binding to its receptor, basigin [15]. While other assays such as erythrocyte binding and immuno reactivity to parasite-exposed sera have traditionally been used to validate the function of recombinant parasite proteins, these assays might not be sufficient to assess correct folding and biochemical activity, which are well-recognized prerequisites of good vaccine antigens. The identification of basigin as the erythrocyte PfRH5 receptor now enables the unequivocal testing of the biochemical activity of any recombinantly produced PfRH5 in a quantifiable manner. It remains to be seen whether recombinant RH5 protein produced either using E. coli with a refolding protocol [12], a wheat germ lysate system [26], or used chemically synthesized peptides [27], has this confirmed biochemical activity. Fig. 2. Immunoreactivities to recombinant PfRH5 7G8 and GB4 are indistinguishable from the 3D7 variant used as an antigen. (A) The 3D7, 7G8 and GB4 PfRH5 variants (colored lines, as indicated in the key) were expressed as Cd4-biotin tagged recombinant proteins, normalized, and specifically captured on a streptavidin-coated microtitre plate. Each variant was equally immunoreactive to a polyclonal antiserum that was raised using the 3D7 variant. Immunoreactivity to the Cd4 tag was removed by preadsorption of the antiserum against Cd4; the preadsorbed antiserum was shown to lack anti-Cd4 immunoreactivity (black line). (B) Recombinant PfRH5 contains heat-labile epitopes. The 3D7 variant of PfRH5-Cd4-biotin was again captured on a streptavidin-coated plate either without treatment or after denaturation by heat (80 • C, 10 min). The two proteins were tested for their immunoreactivity to the anti-RH5 polyclonal antiserum that had been preadsorbed against Cd4 (shown as a control). Data points are means; error bars represent SEM; n = 3. Fig. 3. Polyclonal antibodies raised against 3D7 PfRH5 inhibit both 3D7 and FCR1 P. falciparum parasites. (A) Total IgG purified from a rabbit immunized with the 3D7 variant of PfRH5 was used in invasion assays with the 3D7 and FCR1 P. Falciparum strains. Growth inhibition was observed to a similar extent for both strains with an IC50 ratio for FCR1 of 1.6 (95% CI 1.17-2.18). (B) Total IgG purified from a rabbit immunized with the 3D7 AMA variant was used in erythrocyte invasion assays with the 3D7 or FCR1 P. Falciparum strains and showed strain-specific growth inhibition. (C) Total IgG purified from unimmunized rabbits was used in erythrocyte invasion assays with the 3D7 or FCR1 P. falciparum strains and had no inhibitory effect on erythrocyte invasion. Each IgG concentration/strain combination was assayed in three wells in a 96 well plate; data points are means of all three wells, with standard deviation error bars shown. Fig. 4. Polyclonal antibodies raised against 3D7 PfRH5 inhibit P. falciparum strains that collectively include all common PfRH5 variants. (A) Total IgG purified from a rabbit immunized with the 3D7 variant of PfRH5 was used in invasion assays with Dd2, 7G8, GB4 and 3D7P. falciparum strains. IC50 ratios were 0.5, 1.0 and 1.4 for Dd2, 7G8 and GB4, respectively. 95% CI for IC50s were 0.47-0.63 (Dd2), 0.71-1.36 (7G8) and 1.18-1.71 (GB4). (B) Total IgG purified from a rabbit immunized with the 3D7 variant of PfRH5 was used in erythrocyte invasion assays with two P. falciparum isolates from Kenya (K7336 and KCC103), and the 3D7 P. falciparum strain. IC50 ratios were 1.1 and 0.5 for K7336 and KCC103, respectively. 95% CI for IC50s were 0.81-1.42 (K7336) and 0.35-0.76 (KCC103). (C) Total IgG purified from a rabbit immunized with the 3D7 variant of PfRH5 was used in erythrocyte invasion assays with two P. falciparum isolates from Cambodia (PH21 and PH22) and the 3D7 P. falciparum strain. IC50 ratios were 0.8 and 1.3 for PH21 and PH22 respectively. 95% CI for IC50s were 0.59-1.35 (PH21) and 0.93-1.76 (PH22). Each IgG concentration/strain combination was assayed in three wells in a 96 well plate; data points are means of all three wells, with standard deviation error bars shown.
Using a recombinant PfRH5 based on the 3D7 sequence of P. falciparum, we were able to raise antibodies that have growth inhibitory effects across multiple P. falciparum strains. Genetic diversity in blood-stage vaccine candidates is a concern because of the risk of raising allele-specific antibody responses, and such responses have been observed for multiple P. falciparum merozoite proteins, including those advanced for human trials such as MSP1 [28], MSP3 [28][29][30] and AMA1 [22]. Interrogation of the whole genome sequences of 227 clinical P. falciparum isolates [10] showed that the gene encoding PfRH5 has only a few SNPs present at high frequency. This might be due to the fact that PfRH5 is not a target of natural humoral immunity since, remarkably, sera from Kenyan adults are only weakly immunoreactive to PfRH5 [16]. However, even single polymorphisms can lead to immune evasion, as exemplified by AMA1, where a few high-frequency variants can have a major effect on the specificity of the immune response [31]. However, our study shows that at least in vitro, no currently identified common PfRH5 variant has a significant effect on the ability of anti-PfRH5 IgG to inhibit parasite growth. While the SNPs tested here are based on the largest published whole P. falciparum genome sequence analysis, additional high frequency PfRH5 SNPs may be identified as more samples are sequenced. However, preliminary analysis of more than 1500 global clinical isolates does not reveal any additional PfRH5 SNPs present at higher than 0.10 non-reference allele frequency in any geographic region (Malaria GEN unpublished data). The five SNPs investigated here may therefore represent the most common globally distributed PfRH5 polymorphisms that can be identified using current variation calling pipelines.
Interestingly, previous sequencing of laboratory-adapted isolates [13] identified eight additional PfRH5 SNPs (E48K, I204R, I204K, N347Y, N347D, Y358F, E362D, K429N) that have so far not been seen in any clinical samples, all in strains with an increased ability to invade Aotus erythrocytes [13]. It remains to be seen whether these SNPs have arisen in laboratory strains spontaneously during in vitro culture, or whether they are present in natural isolates at very low frequencies. However, what is clear is that P. falciparum strains containing any of these eight SNPs would be nonrepresentative choices for vaccine challenges or antigen design. In fact the 3D7 PfRH5 sequence would also be a non-representative choice, because it contains the C203Y variant which is relatively rare in field isolates.
In summary, using a mammalian expression system, we have produced full-length biochemically active recombinant PfRH5 that can induce inhibitory antibodies to P. falciparum strains expressing all the currently identified common variants of PfRH5. Broad cross-inhibitory activities of anti-PfRH5 antibodies have previously been demonstrated using virally delivered immunogen [16]. Our study adds significant new data to the vaccine candidacy of PfRH5, by showing that the same strain-transcendent responses can be induced using a protein immunogen, and by extending the range of variants tested. There is therefore now a strong rationale to advance this candidate to Phase I/IIa human trials, which, if successful, could be combined with vaccines targeting other stages of the P. falciparum life cycle, such as the sporozoite stage in the RTS, S, vaccine currently undergoing Phase III trials in Africa [25].