The direct binding of Plasmodium vivax AMA1 to erythrocytes defines a RON2-independent invasion pathway

Significance

Plasmodium parasites invade and replicate within erythrocytes causing the symptoms of malaria disease. The parasite’s erythrocyte invasion machinery could provide targets for vaccines that induce the invasion-blocking antibodies to prevent replication. However, at present our understanding of this complex invasion machinery is incomplete. Here, we describe binding interactions for a Plasmodium vivax protein that is essential for erythrocyte invasion, apical membrane antigen 1 (PvAMA1). We provide evidence that PvAMA1 has two binding interactions for invasion. The first involves direct binding to erythrocyte surfaces and the second is binding to another component of the P. vivax invasion complex rhoptry neck protein 2 (PvRON2). Thus, the role of PvAMA1 is more complex than previously appreciated, and these findings may provide strategies for vaccine design.

Abstract

We used a transgenic parasite in which Plasmodium falciparum parasites were genetically modified to express Plasmodium vivax apical membrane antigen 1 (PvAMA1) protein in place of PfAMA1 to study PvAMA1-mediated invasion. In P. falciparum, AMA1 interaction with rhoptry neck protein 2 (RON2) is known to be crucial for invasion, and PfRON2 peptides (PfRON2p) blocked the invasion of PfAMA1 wild-type parasites. However, PfRON2p has no effect on the invasion of transgenic parasites expressing PvAMA1 indicating that PfRON2 had no role in the invasion of PvAMA1 transgenic parasites. Interestingly, PvRON2p blocked the invasion of PvAMA1 transgenic parasites in a dose-dependent manner. We found that recombinant PvAMA1 domains 1 and 2 (rPvAMA1) bound to reticulocytes and normocytes indicating that PvAMA1 directly interacts with erythrocytes during the invasion, and invasion blocking of PvRON2p may result from it interfering with PvAMA1 binding to erythrocytes. It was previously shown that the peptide containing Loop1a of PvAMA1 (PvAMA1 Loop1a) is also bound to reticulocytes. We found that the Loop1a peptide blocked the binding of PvAMA1 to erythrocytes. PvAMA1 Loop1a has no polymorphisms in contrast to other PvAMA1 loops and may be an attractive vaccine target. We thus present the evidence that PvAMA1 binds to erythrocytes in addition to interacting with PvRON2 suggesting that the P. vivax merozoites may exploit complex pathways during the invasion process.

It was surprising that the replacement of Plasmodium falciparum apical membrane antigen 1 (PfAMA1) with Plasmodium vivax AMA1 (PvAMA1) gene could still allow the parasite to grow in normocytes (1). Many ligands on the parasites have been identified recently, but few erythrocyte receptors for these ligands are known (2). The best-known ligand/receptor pairs for P. vivax are the Duffy-binding protein (DBP) with Duffy blood group antigen (3, 4) and the reticulocyte-binding protein 2b (RBP2b) with transferrin receptor 1 (5), and for P. falciparum the erythrocyte-binding protein (EBP) 175 with glycophorin A (6) and reticulocyte binding protein Homolog 5 with Basigin (7). More recently, it was described that PvRBP2a binds CD98hc on immature erythrocytes (8). In P. vivax, multiple other parasite ligands have been identified for which the erythrocyte receptors are unknown including EBP (a DBP-like ligand) (9–11), merozoite surface protein 1 paralog (12, 13), GPI-anchored micronemal antigen (14, 15), RBP 1a and 1b (16), and reticulocyte-binding surface antigen (17).

AMA1 has a groove surrounded by loops that bind rhoptry neck protein 2 (RON2) that is embedded in the erythrocyte membrane (18–20). As a result, in Toxoplasma gondii (21) and in P. falciparum (19), AMA1 attaches to RON2 on the host membrane, forming the moving junction that brings parasites into the host cell during the invasion. The binding of these proteins is specific for each parasite, e.g., PfAMA1 binds to PfRON2, and thus a PfRON2-derived peptide can block the invasion of P. falciparum (19).

Given that the interaction between AMA1 and RON2 is crucial for the invasion process, another possibility for the successful growth of transgenic P. falciparum expressing PvAMA1 is that PvAMA1 could directly bind erythrocytes and allow merozoites to invade independently of any PvAMA1–PfRON2 interaction. Erythrocyte binding by Loop1a of PvAMA1 has been previously proposed (22), and we also confirmed that recombinant PvAMA1 protein (rPvAMA1) binds erythrocytes, whereas rPfAMA1 binds poorly. Furthermore, we evaluated the invasion inhibition of PvAMA1 parasites in the presence of peptides derived from PfRON2 and PvRON2 and showed that only PvRON2 peptide (PvRON2p) blocked PvAMA1 parasite invasion. The RON2 protein in the PvAMA1 parasites was PfRON2 and its peptides (PfRON2p) blocked PfAMA1 invasion but not PvAMA1, indicating that the PfRON2 could not play the major role in the invasion of PvAMA1. Thus, PvAMA1 mediated the invasion of erythrocytes through direct binding, and furthermore, this binding might be mediated by Loop1a in the RON2 binding groove of PvAMA1.

Results

PvAMA1 Binds to Reticulocytes and Normocytes.

P. vivax AMA1 domains 1 and 2 have been shown to preferentially interact with reticulocytes (22). To further delineate the function of PvAMA1, we expressed recombinant PvAMA1 domains 1 and 2 (rPvAMA1; Fig. 1A) with a His-tag using a human embryonic kidney (HEK) mammalian cell expression system. The HEK cell culture supernatant expressing soluble rPvAMA1 was purified by affinity chromatography (Fig. 1B). The quality of the rPvAMA1 was tested using Sodium Dodecyl-Sulfate Polyacrylamide Gel Electrophoresis (SDS-PAGE) showing the expected size of around 64 kDa (Fig. 1B), and the identity of the protein was further confirmed by Western Blot using anti-His antibodies (Fig. 1C). Next, we performed a flow cytometry-based erythrocyte binding assay to test whether rPvAMA1 binds to reticulocytes and normocytes. We used purified rPfAMA1 protein as a comparator. Our data support previous findings that rPvAMA1 directly binds to erythrocytes (Fig. 1D and SI Appendix, Figs. S1 and S2). Interestingly, rPvAMA1 binding to reticulocytes was higher than normocytes (Fig. 1D). The binding of rPvAMA1 to reticulocytes and normocytes was highly specific as a dose-dependent increase in binding was observed with increasing concentrations of the protein (Fig. 1E and SI Appendix, Fig. S2 B and C). In contrast, only low binding was observed with rPfAMA1 (Fig 1D and SI Appendix, Fig. S1 B–D). We used recombinant D7L1 an Anopheles gambiae salivary protein with His-tag as the negative control to confirm that the binding observed with rPvAMA1 was not due to the his-tag. As expected rD7L1 His-tagged protein showed no binding to normocytes or reticulocytes (SI Appendix, Fig. S1D).

Fig. 1.

PvAMA1 binds to both normocytes and reticulocytes. Schematic structure of AMA1 and RON2p in P. vivax and P. falciparum, and the corresponding recombinant domains 1 and 2 of PvAMA1 [rPvAMA1; amino acids (aa) 43 to 385] and PfAMA1 (rPfAMA1; aa 93 to 442). PvAMA1 (aa 562) and PfAMA1 (aa 622) have one signal peptide (SP), transmembrane (TM) domain and Pro (prodomain). PvRON2 (aa 2203) and PfRON2 (aa 2189) have one SP, TM domain and two cysteine-rich domains. Peptides and proteins made for this study are highlighted in red. PfAMA1 (Pf) and PvAMA1 (Pv) recombinant proteins as indicated by the red line in (A) were expressed in mammalian cells with a 6xHis tag at the C-terminal. The recombinant proteins were visualized in Coomassie Blue staining (B) and Western blot with anti-His antibodies (C). (D) The purified recombinant proteins (40 µg/mL) were incubated with reticulocyte-enriched erythrocyte samples to evaluate binding activity. Erythrocytes were stained with TO to distinguish reticulocytes from normocytes. Values are means of duplicate for the same sample studied (upper and lower end of lines). (E) Binding of rPvAMA1 protein in varying concentration to erythrocytes was determined. The binding results are shown in three independent experiments (Fig. 1 D and E and SI Appendix, Figs. S1 and S2).

Dual Role of PvAMA1

Earlier studies found that the replacement of PfAMA1 with PvAMA1 in P. falciparum did not affect parasite growth or invasion (1). Interestingly, antibodies against PfAMA1 did not block the invasion of the transgenic PvAMA1 parasite. However, PvAMA1 antibodies did block the invasion (1). PfAMA1 interaction with PfRON2 is crucial for invasion in P. falciparum, but the role of PfRON2 in PvAMA1 transgenic parasites is unknown. In vitro data suggested that PfRON2 does not interact with PvAMA1 (23). Here, we aimed to determine if PfRON2 has a role in the invasion of PvAMA1 transgenic parasites. We performed invasion inhibition assays using PfRON2p with transgenic P. falciparum parasites expressing recodonized PfAMA1 (PfAMA1 WT) or PvAMA1 transgenic parasites. As expected, the PfRON2p blocked the invasion of PfAMA1 WT significantly even at the lowest concentration of the PfRON2p used (Fig. 2A). However, the presence of PfRON2p had no effect on the invasion of the PvAMA1 transgenic parasites at any concentration tested (Fig. 2B). These data further confirmed an earlier finding that there was no interaction between rPvAMA1 and PfRON2p by surface plasmon resonance (23). We further tested whether PfRON2p could block rPvAMA1 binding to erythrocytes. In line with the parasite invasion inhibition data, PfRON2p had no effect on rPvAMA1 binding (Fig. 2C and SI Appendix, Fig. S4). These findings suggest that, unlike the PfAMA1-expressing parasites, transgenic parasites expressing PvAMA1 did not utilize PfRON2 for invasion and our binding data further demonstrated that PvAMA1 interacted directly with erythrocytes, indicating that an alternate invasion pathway is used by PvAMA1 transgenic P. falciparum parasites.

Fig. 2.

Invasion inhibition studies of PfAMA1 and transgenic PvAMA1 parasites and erythrocyte binding of rPvAMA1 in the presence of PfRON2p. PfRON2p inhibits parasites expressing PfAMA1, but not PvAMA1. Purified late-stage (A) PfAMA1 and (B) PvAMA1-expressing parasites were mixed with CTFR-labelled erythrocytes and exposed to varying concentrations of PfRON2p (0 µg/mL to 100 µg/mL). Parasites were allowed to undergo egress and invade CTFR⁺ cells overnight. Cells were stained with HO (DNA stain) and double-positive cells were identified (HO⁺CTFR⁺) and measured by flow cytometry. (C) Different concentrations of PfRON2p (0 to 20 µg/mL) and rPvAMA1 (40 µg/mL) were co-incubated for 30 min at room temperature prior to erythrocyte binding assays to evaluate if PfRON2p inhibited the binding of rPvAMA1 to erythrocytes. Dot plot showing the binding inhibition studies is shown (SI Appendix, Fig. S4). Data from three independent experiments were normalized to control (0 µg/mL). Values are means ± 95% CI. Statistical significance was determined by repeated-measures ANOVA, where P < 0.05 is considered significant (ns, not significant; **P < 0.01). (A) p = 0.0003, F = 337.6, df = 5, R² = 0.9941; (B) p = 0.5133, F = 0.6556, df = 5, R² = 0.2469; (C) Reticulocytes: p = 0.7397, F = 0.1614, df = 4, R² = 0.07468, Normocytes: p = 0.5215, F = 0.7516, df = 4, R² = 0.2731. For pairwise comparison, refer to SI Appendix Tables S1 and S2.

As P. vivax long-term culturing in vitro is still not yet possible, functional studies on invasion proteins are greatly hindered. In the case of PvAMA1, we could utilize the transgenic parasites to study its role in invasion. PvAMA1 binds to erythrocytes (22), and it has been shown to interact with PvRON2 (23). We wanted to test if PvAMA1 has two different binding sites for its direct interaction with erythrocytes and PvRON2. Therefore, we assessed the invasion of PvAMA1 transgenic parasites in the presence of PvRON2p. PvRON2p had no significant effect on invasion of the PfAMA1 WT parasite (Fig. 3A); however, significant invasion inhibition was observed in PvAMA1 transgenic P. falciparum parasites with increasing concentrations of PvRON2p (Fig. 3B). To further confirm the specificity of P. vivax AMA1 and RON2 interaction, we tested whether PvRON2p blocked rPvAMA1 binding to erythrocytes. Similar to the invasion inhibition data, binding of rPvAMA1 to erythrocytes was significantly inhibited by PvRON2p with increasing concentrations of the peptide (Fig. 3C). This study suggests that there are two roles played by PvAMA1 during invasion: 1) by interacting with erythrocytes, and 2) by interacting with RON2. Hence PvAMA1 might have two erythrocyte binding sites next to each other, one interacting with a unique erythrocyte receptor directly and another site interacting with the PvRON2. The presence of multiple binding interactions by PvAMA1 has implications for the design of vaccines targeting PvAMA1.

Fig. 3.

Invasion inhibition studies of PfAMA1 and transgenic PvAMA1 parasites and erythrocyte binding of rPvAMA1 in the presence of PvRON2p. PvRON2p inhibits parasites expressing PvAMA1, but not PfAMA1. Purified late-stage (A) PfAMA1 and (B) PvAMA1 expressing parasites were mixed with CTFR-labelled erythrocytes and exposed to varying concentrations of PvRON2p (0 µg/mL to 200 µg/mL). Parasites were allowed to undergo egress and invade CTFR⁺ cells overnight. Cells were stained with HO (DNA stain) and double-positive cells were identified (HO⁺CTFR⁺) and measured by flow cytometry. (C) Serial dilutions of PvRON2p (0 to 20 µg/mL) were co-incubated with rPvAMA1 protein (40 µg/mL) for 30 min at room temperature prior to incubation with erythrocytes to evaluate whether PvRON2p inhibited the binding of rPvAMA1 to erythrocytes. After incubation, erythrocyte binding by PvAMA1 was measured by flow cytometry. Dot plot showing the binding inhibition studies is shown (SI Appendix, Fig. S4). Data from three independent experiments were normalized to control (0 µg/mL). Values are means ± 95% CI. Statistical significance was determined by repeated measures ANOVA, where P < 0.05 is considered significant (ns, not significant; *P < 0.05; **P < 0.01; ***P < 0.001). (A) p = 0.1225, F = 5.602, df = 5, R² = 0.7369; (B) p = 0.0209, F = 29.07, df = 5, R² = 0.9356; (C) Reticulocytes: p = 0.0503, F = 15.73, df = 4, R² = 0.8872, Normocytes: p = 0.0147, F = 18.96, df = 4, R² = 0.9046. For pairwise comparison (SI Appendix Tables S1 and S2).

Structure of PvAMA1 with PvRON2p and the Effect of Loop1a on PvAMA1 Binding to Erythrocytes.

Since PvRON2p could inhibit binding of PvAMA1 to erythrocytes, we investigated the PvAMA1 structure to gain insights into the basis for this effect. In the published structure of the PvAMA1–PvRON2p complex (PDB accession number 5NQG) Loop1a is, one of six loops (1a–1f) surrounding the binding groove of PvAMA1 for PvRON2p, and the Loop1a sequence partially overlaps in sequence with a previously reported PvAMA1 peptide (HABP21270) that has some erythrocyte-binding activity (Fig. 4A) (22). In the molecular structure of PvAMA1 Loop1a forms several contacts with the PvRON2p peptide. The side chain of Arg 88 of PvAMA1 forms a hydrogen bond with the carbonyl oxygen of Asp 2042 of PvRON2p (Fig. 4A). The amide nitrogen and carbonyl oxygen of Arg 88 also form hydrogen bonds with the carbonyl oxygens of Thr 2041, Met 2045 and Gly 2044 of PvRON2p through intervening water molecules. Other close contacts between side chains of the two molecules include Val 82 and Tyr 87 of Loop1a which pack around the side chain of Pro 2047 of PvRON2L (Fig. 4B). Considering that Loop1a has interactions with PvRON2p and that PvRON2p inhibited direct binding of PvAMA1 to erythrocytes, we hypothesized that Loop1a may be involved in erythrocyte binding and that the binding sites for PvRON2 and direct erythrocyte binding may be adjacent to each other. Therefore, we determined whether a Loop1a-derived peptide blocked the binding of PvAMA1 to erythrocytes (Fig. 4C). Interestingly, with increasing concentrations of the peptide, the binding of rPvAMA1 decreases, suggesting that this Loop1a region is required for interaction with the erythrocytes (Fig. 4C and SI Appendix, Fig. S5). However, in repeated studies, Loop1a did not block the invasion of the PvAMA1 transgenic parasites (Fig. 4D). Similarly, recombinant PvAMA1 protein did not inhibit the invasion of wild-type and PvAMA1 transgenic parasites (SI Appendix, Fig. S6).

Fig. 4.

P. vivax AMA1 interaction with the peptide PvRON2p and the effect of PvAMA1 Loop 1a on invasion and erythrocyte binding. (A) Molecular surface of PvAMA1 with PvRON2L-interacting loops (labeled 1a–1f) (23) is shown in ribbon format. Loops 1a (pink), 1b (tan), 1c (yellow), 1d (blue), 1e (cyan) and 1f (green) surround the bound RON2L peptide (purple). Side chains of PvAMA1 Loop 1a that contact the PvRON2L ligand are shown in stick representation; Arg 88 and Tyr 87 are labeled. (B) Space-filling representation of PvAMA1 Loop 1a (carbons colored pink) and the PvRON2L peptide (carbons colored purple) showing the intimate association of the loop with the peptide. Residues Tyr 87 from PvAMA1 and Pro 2047 from RON2L are labeled. Oxygen atoms are colored red, nitrogen atoms blue and sulfur atoms yellow. (C) A peptide based on PvAMA1 Loop 1a inhibits binding of rPvAMA1 to normocytes and reticulocytes. Serial dilutions of PvAMA1 Loop 1a peptide and rPvAMA1 protein (40 µg/mL) were incubated with erythrocytes and binding of PvAMA1 was quantified. (D) Purified late-stage parasites expressing PvAMA1 were mixed with CTFR-labelled erythrocytes in the presence of varying concentrations of PvAMA1 Loop 1a peptide (0 µg/mL to 100 µg/mL). Parasites were allowed to undergo egress and invade CTFR⁺ cells overnight. Cells were stained with HO (DNA stain) and double-positive cells were identified (HO⁺CTFR⁺) and measured by flow cytometry. Data from three independent experiments were normalized to control (0 µg/mL). Values are means ± 95% CI. Statistical significance was determined by repeated measures ANOVA, where P < 0.05 is considered significant (ns, not significant). (C) Reticulocytes: p = 0.0281, F = 32.13, df = 3, R² = 0.9414, Normocytes: p = 0.2644, F = 2.311, df = 3, R² = 0.5361; (D) p = 0.3299, F = 1.568, df = 5, R² = 0.4395. For pairwise comparison (SI Appendix Tables S1 and S2).

Discussion

It is established that PfAMA1 interacts with PfRON2 within a surface groove during invasion of erythrocytes. The binding of AMA1 to RON2, which is part of a larger complex of RON proteins, is thought to be important for the formation of a moving junction that brings the merozoite into the erythrocytes, and this interaction between PfAMA1 and PfRON2 was thought to be an essential step in the invasion process (19). In the case of invasion when PvAMA1 is expressed in a P. falciparum parasite, replacing the endogenous PfAMA1, the parasite still retains expression of PfRON2 (1), but invasion still occurs. Therefore, we investigated how PvAMA1 can complement the function of PfAMA1.

Here, we show that the PfRON2p blocks invasion of PfAMA1 as expected but has no effect on invasion of transgenic parasites expressing PvAMA1 with PfRON2 (Fig. 2 A and B), indicating that PfRON2p does not interact with PvAMA1 that is consistent with a previous study (23). As PfRON2 peptide does not block invasion of PvAMA1-expressing parasite, invasion must then occur independently of PfRON2.

Unlike PfAMA1, rPvAMA1 binds to both normocytes and reticulocytes (Fig. 1 C and D and SI Appendix, Fig. S1) which is partially consistent with the results from the previous report in which the binding of PvAMA1 had been observed with umbilical cord blood (22). Furthermore, PfRON2 peptide could not block the binding of rPvAMA1 and the invasion of PvAMA1 transgenic parasites (Fig. 2), suggesting that invasion by the PvAMA1 transgenic parasite involves PvAMA1 interacting directly with erythrocytes and not with PfRON2. However, it is not clear why low binding was observed for recombinant PvAMA1 binding to erythrocytes. We can speculate that the concentration of PvAMA1 on merozoites is high with multiple AMA1 molecules on the surface of the merozoite that lead to increased binding affinity to erythrocytes for invasion. Indeed, the invasion inhibition of PvRON2p in PvAMA1 transgenic parasites is lower than that of PfRON2p in wild-type parasite even at high concentration (Figs. 2A and 3B), suggesting that the function of PvAMA1 is not only limited to interacting with PvRON2 but also to unknown erythrocyte receptor. Thus, further studies are required to address this question.

Arevalo-Pinzon et al. (22) reported that domains 1 and 2 of PvAMA1 bound erythrocytes and identified a peptide sequence that could bind erythrocytes. Interestingly, only one peptide that contains Loop1a bound to erythrocytes with low activity and inhibited binding of PvAMA1 to reticulocytes by 50%, whereas the peptides derived from the other loops surrounding the PvRON2p binding groove of PvAMA1 did not block binding (22). Furthermore, Vulliez-Le Normand et al. (23) reported that PvRON2 interacted with Loop1a of PvAMA1 through a number of amino acid contacts (Fig. 4 A and B). In the present study, we also confirmed that a Loop1a-derived peptide inhibited binding of rPvAMA1 to erythrocytes, but the peptide did not block invasion (Fig. 4 C and D). Since PvAMA1 Loop1a has low binding activity compared to rPvAMA1 in a previous report (22), it may be that the binding affinity of the peptide is able to block the lower affinity binding of rPvAMA1 but not inhibit merozoites invasion. This may be due to a higher concentration of native PvAMA1 presented on the merozoite surface or the structure of the native PvAMA1, produced by the parasites, achieving a higher affinity interaction with the erythrocyte receptor compared to rPvAMA1. It is also possible that the Loop1a peptide does not fully capture the erythrocyte the erythrocyte-binding residues of PvAMA1, and other sites are involved in binding for invasion function. Further studies should determine the structure of the native PvAMA1 and its interaction with the unknown erythrocytic receptor.

Materials and Methods

Reagents.

Peptides are derived from PvRON2p (Met2034–Leu2072) (PlasmoDB ID: PVP01_1255000) (23) and PfRON2p (Ala2026–Lys2067) (PlasmoDB ID: PF3D7_1452000) (19) that fit within the AMA1 groove of PvAMA1 and PfAMA1, respectively. PvRON2p and PvAMA1 Loop1a (HABP21270) (Glu81–Gly100) (PlasmoDB ID: PVP01_0934200) (22) were synthesized by Bachem, and PfRON2p peptide was kindly provided by David L Narum. The peptides were synthesized according to sequence from Plasma DB. Hoechst 33342 (HO) and Cell Trace Far Red (CTFR) were obtained from Thermo Fisher Scientific. HO was stored as a 20 mM stock solution, while CTFR was reconstituted with dimethyl sulfide according to manufacturer instructions and used as a 1 mM stock solution. Nycodenz (Axis-Shield PoC) was dissolved in water as a 60% stock solution (w/v) at 50 °C with a stir until the solution is transparent. Thiazole Orange (TO) Retic-COUNT reagent was obtained from Becton Dickinson.

Recombinant Protein Expression and Purification.

Gene fragments encoding recombinant (r) domains 1 and 2 of PvAMA1 (Pro43–Asp385) in (PlasmoDB ID: PVP01_0934200) and recombinant domains 1 and 2 of PfAMA1 (PlasmoDB ID: PF3D7_1133400) (Leu93–Pro442) were codon-optimized with point mutations at predicted N-glycosylation sites for a mammalian expression system and synthesized by GenScript in a pcDNA3.3 vector. The synthesized genes were cloned into BMR1_01g020310-bio-His which was a gift from Gavin Wright (Addgene plasmid # 108116) (Addgene, Watertown, MA) between AscI and NotI restriction enzyme sites after digestion (New England Lab, NEB) and ligated with T4 DNA ligase (NEB). The constructs harboring the protein sequences were transfected into HEK cells, HEK293E (American Type Culture Collection), and incubated at 37 °C. Supernatants were harvested 72 h after transfection for protein purification. The protein expression was carried out at the SAIC Advanced Research Facility. For the A. gambie D7L1 protein, the mature cDNA sequence was codon optimized for mammalian expression and synthesized by Bio Basic Inc. in the VR2001-TOPO vector as described in previous reports (24).

The recombinant proteins were purified by affinity chromatography with Nickel-charged HiTrap Chelating HP (GE Healthcare Life Science) followed by size exclusion chromatography with Superdex 200 10/300 GL columns (GE Healthcare Life Science). The column purification was carried out with the AKTA purifier system (GE Healthcare Life Science). The size and purity of recombinant proteins were evaluated on an SDS-PAGE gel in a NuPAGE Novex 4–12% Bis–Tris protein gels (Thermo Fisher Scientific) under denaturing conditions and stained with Coomassie blue by the eStain protein stain system (GenScript) to visualize the proteins. For the Western blots, rPvAMA1 and rPfAMA1 in SDS-PAGE gel were transferred onto Polyvinylidene Difluoride (PVDF) membrane with iBlot2 PVDF Mini Stacks (Invitrogen) according to the manufacturer’s instructions. The membrane was blocked with 1% bovine serum albumin (BSA) in 0.2% Tween 20 in Tris-Buffered Saline (TBS) (TBS-T), followed by 6xHis Tag Monoclonal Antibody (Invitrogen) and KPL Peroxidase-Labeled Affinity Purified Antibody to Mouse IgG (H+L) (SeraCare Life Sciences). The membrane was washed with 1% BSA in TBS-T three times after each step. The mixture of SuperSignal^TM West Femto Trial Kit (Thermo Fisher Scientific) with phosphate-buffered saline (PBS) was spread on membrane to visualize the proteins. The image of membrane was captured with Azure 300 imaging system (Azure Biosystems) according to the manufacturer’s instructions.

Reticulocyte Enrichment from the Buffy Coat.

Fresh reticulocytes were enriched from Duffy antigen-positive buffy coats (Blood bank, NIH) with 19% Nycodenz (w/v) from 60% Nycodenz stock (w/v) dissolved in water (Axis-Shield PoC) in high-KCl buffer (115 mM KCl, 20 mM N-2-hydroxyethylpiperazine-N-2-ethane sulfonic acid (HEPES), 1 mM MgCl₂, 1 mM NaH₂PO₄, 10 mM D-glucose, 0.5 mM Ethylenediaminetetraacetic acid (EDTA), and 12 mM NaCl, pH 7.4) using gradient centrifugation as described previously for cord blood (25). Briefly, leukocytes were removed by a Non-wowen fiber (NWF) filter (Zhixing Bio Co.) after washing with incomplete Roswell Park Memorial Institute (RPMI) 1640 medium (containing 25 mM HEPES, L-glutamine and 5 mg/L hypoxanthine and no other supplements (KD Medical Inc.). The packed cells after centrifugation were resuspended in high KCl buffer at 10 or 20% hematocrit and incubated at 4 °C for 3 h. A total 5 mL of the RBC-high KCl buffer was overlaid on 3 mL of 19% Nycodenz and centrifuged at 3,000×g for 30 min without a brake. The enriched reticulocytes were collected from the interface layer and washed three times with incomplete RPMI 1640. The purity of reticulocytes was determined in thin blood smear after new methylene blue stain for 15 min (26).

Reticulocyte- and Mature Erythrocyte-Binding Assay and Inhibition Assay.

The enriched reticulocytes were diluted to 5 × 10⁵ cells in 200 μL of incomplete RPMI1640 and used for the flow cytometry-based direct-binding assay as described previously (27). Briefly, erythrocytes were incubated with recombinant protein with or without peptide at room temperature for 3 h. The red blood cells were washed with 1% BSA in PBS once after incubation and then incubated with Alexa Fluor 647-conjugated mouse anti-His monoclonal antibody (QIAGEN) in 0.1% BSA in PBS at 4 °C for 1 h in the dark. The red blood cells were washed once with 1% BSA in PBS and incubated with 200 μL of the TO Retic-COUNT reagent (Becton Dickinson) for 30 min at room temperature in the dark. A total of 100,000 events per sample were analyzed with MACSQuant® Analyzer 10 (Miltenyi Biotec), and data were analyzed with FlowJo 10. 8. 1 (FlowJo LLC). Unstained cells and cells singly stained with TO were used for gating the normocytes and reticulocytes, respectively. Binding events of protein to normocyte and reticulocyte were calculated by the following formula (Fig. 1 D and E and SI Appendix, Figs. S1 and S2); Binding events = Binding cells / Total cells (Binding cells + Unbound cells). Q1 and Q2 represent binding cells for normocyte and reticulocyte, respectively. Q4 and Q3 represent unbound cells for normocyte and reticulocyte, respectively. To normalize binding inhibition data, binding events were calculated for each control (0 µg/mL) that is fixed to be 10,000 events (Figs. 2C, 3 C, and 4 C). The number of binding events of PBS samples was subtracted from each sample of experiments.

P. falciparum Culture.

The maintenance of P. falciparum (W2mef strain) expressing recodonized Pf W2 AMA1 (referenced as PfAMA1; 3D7 strain) or transgenic PvAMA1 (encoding the signal sequence and prodomain of P. falciparum AMA1) (1, 28) were maintained as previously described (29). Briefly, parasites were kept in continuous culture with O⁺ human red blood cells (Interstate Blood Bank) at 2% hematocrit in complete medium which consisted of RPMI 1640 incomplete medium (KD Medical Inc.) supplemented with 0.5% albumax (Thermo Fisher Scientific), 2.0 g/L sodium bicarbonate (KD Medical Inc.) and 50 μg/mL gentamicin (KD Medical Inc.). Parasites were grown at 37 °C in a 5% O₂, 5% CO₂, and 90% N₂ gas mixture.

Invasion and Growth Assays Using PfAMA1 and a PvAMA1 Transgenic Line.

Leukocyte-depleted Duffy-positive donor blood (Blood bank, NIH) was washed with incomplete RPMI 1640 medium three times. Uninfected Duffy-positive erythrocytes were labeled with 4 µM of CTFR, prepared in incomplete RPMI 1640 medium with cells at 2% hematocrit. Cells were stained with CTFR for 2 h at 37 °C with shaking in the dark. Cells were subsequently washed twice with incomplete RPMI 1640 medium and stored at 4 °C until used in assays.

Cultures with high parasitemia (~5 to 10%) and a majority of trophozoite-schizont stage parasites underwent magnetic-activated cell sorting (MACS, Miltenyi Biotec) purification to isolate late-stage parasites. MACS purification was done as previously described (30, 31). Briefly, magnetic Large Scale (LS) columns (Miltenyi Biotec) were used with the QuadroMACS® System (Miltenyi Biotec). Incomplete RPMI 1640 medium was allowed to run through the column prior to the addition of parasitized cultures. Columns were then washed with two volumes of incomplete medium, before being removed from the magnet. Another two volumes of incomplete medium were run through the column to elute late-stage parasites. The purified parasites were centrifuged and washed with incomplete RPMI 1640.

For concentration curves, serial dilutions of peptides (PfRON2p, PvRON2p PvAMA1 Loop1a) and recombinant PvAMA1 were prepared in complete RPMI 1640 medium and 50 µL of 400 µg/mL, 200 µg/mL, 100 µg/mL, 50 µg/mL, 25 µg/mL, or 12.5 µg/mL) were dispensed into wells. Fifty microliters of complete RPMI 1640 medium with no peptide was used as a control. Purified late-stage parasites were mixed with CTFR-labelled erythrocytes and the parasitized cell suspension was adjusted to 4% hematocrit, after which 50 µL was dispensed into wells. The final volume of wells was 100 µL, with 2% hematocrit and final peptide concentration of 0 (no peptide added), 200, 100, 50, 25, 12.5, and/or 6.25 µg/mL). Cells were mixed well prior to incubation of plates at 37 °C under gas conditions overnight and analyzed by flow cytometry the next morning.

All test conditions were done in duplicate and additional wells were included to make thin blood smears for verification of parasite blood stages by microscopy. Additional controls also include those for flow cytometry setup and analysis (e.g., wells containing cells stained with CTFR only, and unstained cells for use as an unstained control or HO stained only. All experiments were done independently at least three times.

Measurement of Parasitemia by Flow Cytometry.

Cells were centrifuged at 2,500 rpm for 3 min and supernatant removed. Cells were washed once with PBS and then stained with 4 µM of HO for 30 min in the dark at room temperature. Cells were then washed once with PBS and resuspended in PBS for flow analysis on the MACSQuant® Analyzer 16 (Militenyi Biotec). At least 500,000 events were collected per sample. Data were analyzed using FlowJo software version 10.8.1. Parasitemia was determined from dot plots (SI Appendix, Fig. S3 for example of gating strategy) depicting HO vs. CTFR fluorescence channels. Double-positive cells (HO⁺CTFR⁺) located in quadrant 2 (Q2) were taken as being indicative of newly invaded cells. Unstained, HO-stained only, and CTFR-stained only controls were used to initially determine gating limits, which were then applied to all experiments thereafter. Any background fluorescence occurring in Q2 of the CTFR-stained only control that was run in parallel with the samples of each experiment was subtracted from all samples (SI Appendix, Fig. S3).

Structure.

Structural figures were drawn with Pymol (Schrodinger Inc.).

Data Analysis.

All data were statistically analyzed and graphed using GraphPad Prism 9. For the analysis, count data from cytometric analyses were rescaled against the respective controls (0 μg/mL added component) within a repeat by setting the controls to 10,000 events and multiplying all other data points per repeat with the same factor to overcome inherent data variability between experimental repeats. This step did not alter the data relation between data groups, but improved comparability. Further, it is of note that, within an experimental repeat, the buffy coat preparation was split and used for all applied component concentrations, inherently pairing the data within a repeat. As the rescaled data conformed approximately with underlying assumptions for parametric testing, a repeated measures one-way ANOVA was generally applied (any divisions from this are stated in the text). Graphs show the data mean ± 95% CIs.

Supplementary Material

Appendix 01 (PDF)

Data, Materials, and Software Availability

All study data are included in the article and/or SI Appendix.