Structural Characterization of the Erythrocyte Binding Domain of the Reticulocyte Binding Protein Homologue Family of Plasmodium yoelii

Abstract

Invasion of the host cell by the malaria parasite is a key step for parasite survival and the only stage of its life cycle where the parasite is extracellular, and it is therefore a target for an antimalaria intervention strategy. Multiple members of the reticulocyte binding protein homologues (RH) family are found in all plasmodia and have been shown to bind to host red blood cells directly. In the study described here, we delineated the erythrocyte binding domain (EBD) of one member of the RH family, termed Py235, from Plasmodium yoelii. Moreover, we have obtained the low-resolution structure of the EBD using small-angle X-ray scattering. Comparison of the EDB structure to other characterized Plasmodium receptor binding domains suggests that there may be an overall structural conservation. These findings may help in developing new approaches to target receptor ligand interactions mediated by parasite proteins.

INTRODUCTION

Differences in the ability of the invasive form of the malaria parasite, the merozoite, to recognize and invade red blood cells (RBC) have a direct impact on disease severity. In the case of the human parasite Plasmodium vivax, invasion is restricted to reticulocytes, leading to a lower parasite burden than that of Plasmodium falciparum, which is able to invade RBC of all ages. Differences in the repertoire and expression of parasite ligands on the merozoite are responsible in defining the invasion characteristics of the different parasite species.

Two protein families, called erythrocyte binding-like proteins (EBL) and reticulocyte binding protein homologues (RH), are found in all Plasmodium species and have been shown to be key ligands that enable the parasite to recognize different receptors on the RBC surface (reviewed in references 22 and 34). The total number of EBL varies between different parasite species, with P. falciparum having five members while P. vivax has only a single member (1, 11, 20). All members of the EBL proteins are defined by the presence of the cysteine-rich Duffy binding-like (DBL) domain, with each DBL domain mediating binding to a single receptor on the RBC (1, 2, 26, 40–42). Both in P. falciparum and in P. vivax the RBC receptors recognized by the different members of the EBL family are known. The receptor recognized by each EBL directly correlates with the binding specificity of its DBL domain. As with the EBL, the number of RH varies between different parasite species, ranging from as few as 6 members in P. falciparum to as many as 14 in the rodent malaria parasite Plasmodium yoelii (12, 13, 20). In P. falciparum, different members of the RH are able to recognize different receptors on the RBC, and the combination of RH and EBL expressed in the merozoite defines unique invasion pathways (3, 6–8, 16, 17, 19, 21, 26, 31, 35, 39, 42, 46, 47, 49, 50, 56, 62, 65). Unlike the case of the EBL, there is no easily identifiable domain structure in RH, making it difficult to identify functional domains within these large proteins. Recently, the erythrocyte binding regions of PfRH1, PfRH4, and PfRH5 of the RH family in P. falciparum have been mapped and have shown limited overall sequence conservation between them (5, 19, 23, 32, 50, 63). At this stage, no structural information is available for any members of the RH family.

The RH of P. yoelii are coded for by the 235-kDa rhoptry protein (Py235) multigene family and have been shown to play an important role in parasite virulence, host cell adaptation, and immune evasion (reviewed in references 28, 34, and 55). A single member of Py235 (Py01365) is dominantly expressed in both virulent and avirulent parasite populations (35) and has been shown to directly bind to RBC (44). In addition, Py01365 is recognized by a protective monoclonal antibody, 25.77, and has recently been shown to contain a nucleotide sensing domain (44, 48). Genetic disruption of Py01365 reduces the overall virulence of the P. yoelii YM line by reducing the total repertoire of RBC the parasite is able to invade (4a). This identifies Py01365 as a key mediator of parasite virulence whose binding to a specific RBC receptor leads to increased invasion and thereby parasite burden.

In an effort to further understand the recognition of the RBC receptor by the RH better, we have identified the erythrocyte binding region of Py01365. We show that a recombinant protein containing a region of Py01365, called EBD_1–194, binds mouse RBC with the same specificity as full-length Py235. The homogenous purification of EBD_1–194 enabled us to determine the first low-resolution solution structure of the highly α-helical protein by solution X-ray scattering.

MATERIALS AND METHODS

Gene expression and protein purification.

The reverse primers used for PCR amplification for EBD_1–194 and EBD_1–398 are 5′-AATTACGAGCTCTTAGTCCTTTATATTGTCTATATTAC-3′ and 5′-AATTACGAGCTCTTATCCTAAATTTTCTTTTAAATC-3′, respectively. The forward primer for amplification for both constructs is 5′-GTGAGTCCATGGTATCTGACAAAAATGAATATG-3′. These primers were designed specifically to include SacI and NcoI restriction sites (underlined), respectively. The genomic P. yoelii YM DNA was used as the template. Following digestion with NcoI and SacI, the PCR products were ligated into the pET9d1-His₃ vector (27). The pET9d-His₃ vector, containing the respective gene, was then transformed into Escherichia coli cells [strain BL21(DE3)] and grown on 30 μg/ml kanamycin-containing Luria-Bertani (LB) agar plates. To express EBD_1–194 and EBD_1–398, liquid cultures were shaken in LB medium containing kanamycin (30 μg/ml) for about 20 h at 37°C until an optical density at 600 nm (OD₆₀₀) of 0.6 to 0.7 was reached. To induce production of the recombinant proteins, the cultures were supplemented with isopropyl (thio)-β-d-galactoside (IPTG) to a final concentration of 1 mM. Cells producing recombinant EBD_1–194 and EBD_1–398 were harvested at 8,500 × g for 12 min at 6°C. Subsequently, they were lysed on ice by sonication three times (for 1 min each) in buffer A (50 mM Tris-HCl, pH 7.5, 500 mM NaCl, and 2 mM phenylmethylsulfonylfluoride [PMSF]). Precipitated material was separated by centrifugation at 10,000 × g for 35 min. The supernatant was filtered (0.45 μm; Millipore) and passed over a 3-ml Ni²⁺-nitrilotriacetic acid (NTA) resin column to isolate EBD_1–194, according to the method of Grüber et al. (27). The His-tagged protein was allowed to bind to the matrix for 2.5 h at 4°C and eluted with an imidazole gradient (25 to 400 mM) in buffer A. Fractions containing His₃-EBD_1–194 were identified by SDS-PAGE (37), pooled, and concentrated as required using Centricon YM-3 (3-kDa molecular mass cutoff) spin concentrators (Millipore). In a second chromatographic step, using a size exclusion column (Superdex 75 HR 10/30 column; GE Healthcare), highly pure EBD_1–194 was eluted with buffer consisting of 50 mM Tris-HCl, pH 7.5, 500 mM NaCl, 10 mM EDTA, and 2 mM PMSF. Recombinant NBD94 and NBD94_1–550 were produced and purified as described recently (38). The yeast V-ATPase subunit C was isolated according to the method of Armbrüster et al. (4).

Determination of native molecular mass.

Gel filtration chromatography was performed using a Superdex 75 HR 10/30 column (Amersham Biotech) using a buffer of 50 mM Tris-HCl (pH 7.5) and 500 mM NaCl. To construct a calibration curve, a set of standard proteins (Amersham Biotech and Sigma) and recombinant subunit C of the yeast V-ATPase, for which a solution (4) and crystallographic structure have been determined (15), were analyzed. The K_av parameter (constant for all volumes) was determined (K_av = (V_e − V₀)/(V_t − V₀), where V_e represents the elution volume, V₀ represents the void volume, and V_t represents the total bed volume). The K_av values for standard proteins were plotted as a function of the logarithm of molecular mass (MM), and the resulting calibration curve was used to derive the MM of EBD_1–194.

Mouse blood preparation and enzyme treatment.

Mouse erythrocytes were always freshly obtained just prior to any erythrocyte binding assays. Eight- to twelve-week-old BALB/c mice bred under specific-pathogen-free (SPF) conditions were culled, and blood was collected in acid citrate dextrose (ACD) (Sigma-Aldrich). The blood was thoroughly mixed and left standing for 5 to 10 min and was subsequently washed 3 times at 2,100 rpm for 3 min at 21°C with incomplete RPMI medium.

To study the effect of enzyme treatment on erythrocyte binding, 1 mg/ml trypsin (Sigma-Aldrich), 1 mg/ml chymotrypsin (Sigma-Aldrich), and 25 mU neuraminidase (Roche Diagnostics) were added to the washed mouse blood and incubated for 2 h at 37°C (45). The enzyme-treated blood was then washed 3 times at 2,100 rpm for 3 min at 21°C with incomplete RPMI medium. In the case of trypsin- and chymotrypsin-treated blood, 1 mg/ml trypsin inhibitor (Sigma-Aldrich) was added after the first wash and incubated by rocking at room temperature for 10 to 15 min.

Erythrocyte binding assays.

Erythrocyte binding assays were carried out as previously described (43, 45, 48). In brief, prior to the binding assay, the recombinant proteins were dialyzed against 1× phosphate-buffered saline (PBS) overnight and concentrated to about 1 to 5 mg/ml. One to ten micrograms of dialyzed proteins (NBD_1–550, EBD_1–398, and EBD_1–194) or alternatively 250 μl of supernatant protein was added to 100 μl untreated or enzyme-treated mouse erythrocytes. To this, 50 μl of FBS was added, making up to a final volume of 600 μl using incomplete RPMI. The sample was incubated for 2 h at 37°C before 600 μl of dibutyl pthalate (Sigma) oil was added to separate the supernatant proteins. After centrifugation, the supernatant was removed by aspiration. A 0.5 M concentration of NaCl was used to elute the bound proteins, followed by analysis using SDS-PAGE. The full-length Py235 protein was separated on a 6% SDS-PAGE gel, NBD94, NBD_1–550, and EBD_1–398 were separated on a 12% SDS-PAGE gel, and EBD_1–194 was separated on a 15% SDS-PAGE gel. The full-length proteins was detected by Western blotting using monoclonal antibody 25.77 (18, 33), while the recombinant proteins were detected using mouse anti-penta-His antibody (Qiagen).

CD spectroscopy.

Steady-state circular dichroism (CD) spectra were measured in the far-UV light (180 to 260 nm) using a Chirascan spectropolarimeter (Applied Photophysics). CD spectroscopy of EBD_1–194 (2.0 mg/ml) was performed in a buffer of 50 mM Tris-HCl, pH 7.5, and 500 mM NaCl. CD values were converted to mean residue ellipticity in units of degree cm² dmol⁻¹ using the software program Chirascan, version 1.2 (Applied Photophysics). In order to analyze the CD spectrum, the Dicroprot (14) and NeuralNet (9) software packages were used.

X-ray scattering experiments and data analysis of EBD_1–194.

Small-angle X-ray scattering (SAXS) data for EBD_1–194 were collected by following standard procedures on the X33 SAXS camera (10, 51) of the EMBL Hamburg, located on a bending magnet (sector D) on the storage ring DORIS III of the Deutsches Elektronen Synchrotron (DESY). A photon-counting Pilatus 1 M pixel detector (67 by 420 mm²) was used at a sample-detector distance of 2.4 m, covering the following range of momentum transfer: s > 0.1 and < 4.5 nm⁻¹ (s = 4π sin(q)/l, where q is the scattering angle and l = 0.15 nm, the X-ray wavelength). The s axis was calibrated by the scattering pattern of silver-behenate salt (d-spacing, 5.84 nm). A range of protein concentrations (2.5 to 5 mg/ml) was measured for EBD_1–194 to assess and remove any concentration-dependent interparticle effects. Both proteins were measured in a buffer composed of 50 mM Tris-HCl, pH 7.5, 500 mM NaCl, and 1 mM dithiothreitol (DTT). The protein as well as the buffer samples were injected automatically using the sample-changing robot for solution scattering experiments at SAXS station X33 (51). All the data processing steps were performed automatically using the program package PRIMUS (57). The forward scattering [I(0)] and radius of gyration [R_g] were evaluated using the Guinier approximation (29). These parameters were also computed from the entire scattering patterns using the indirect transform software package GNOM (61).

The molecular masses of both proteins were calculated by comparison with the forward scattering from the reference solution of bovine serum albumin (BSA). From this procedure, a relative calibration factor for the MM can be calculated using the known molecular mass of BSA (66.4 kDa) and the concentration of the reference solution by applying

$${\mathrm{MM}}_p=I{\left(0\right)}_p/C_p\times \frac{{\mathrm{MM}}_{\mathrm{st}}}{I{\left(0\right)}_{\mathrm{st}}/C_{\mathrm{st}}}$$

where I(0)_p and I(0)_st are the scattering intensities at zero angle of the studied and the BSA standard protein, respectively, MM_p and MM_st are the corresponding molecular masses, and c_p and c_st are the concentrations. Low-resolution models of EBD_1–194 were built by the software program DAMMIN (58, 60). Ab initio shape models for EBD_1–194 were obtained by averaging 10 independent DAMMIN reconstructions for each subunit by using the software package DAMAVER (67).

RESULTS

Identification of the erythrocyte binding region of Py235.

Erythrocyte binding assays (EBA) using Py235 show binding of the protein to mouse erythrocytes (Fig. 1B). As shown in Fig. 1B, enzyme treatment of mouse erythrocytes with neuraminidase, trypsin, and chymotrypsin revealed the Py235 receptor to be neuraminidase resistant and trypsin and chymotrypsin sensitive, which is in line with recent data (45). A number of studies have identified the erythrocyte binding regions of two members of the RH family in P. falciparum (19, 23). For both PfRH1 and PfRH4, the binding region is located in the N-terminal half of the protein (Fig. 1A). We have identified a 94-kDa region (NBD94) of Py01365 as containing a functional ATP/ADP binding region and have also shown that either ATP or ADP has a direct impact on Py235 binding to erythrocytes (48). While members of the RH family have limited sequence conservation, we noted that the starting position of NBD94 partially overlaps the erythrocyte binding region identified in PfRH1 of P. falciparum (Fig. 1A); we therefore investigated whether the recombinant expressed protein is able to specifically bind to erythrocytes. Erythrocyte binding assays (EBA) using increasing amounts of NBD94 show concentration-dependent binding of the protein to mouse erythrocytes (Fig. 2 B). Similar concentration-dependent binding was also seen with the truncated form of NBD94 (NBD94_1–550) (Fig. 2C) that was lacking the C-terminal hinge region previously identified (Fig. 2A) (48). To test whether the recombinant proteins NBD94 and NBD94_1–550 retained the same binding properties as the entire Py235 protein, EBA was carried out using neuraminidase-, chymotrypsin-, and trypsin-treated erythrocytes. Both NBD94 (Fig. 2B) and NBD94_1–550 (Fig. 2C) bound efficiently to neuraminidase-treated erythrocytes but showed reduced binding to chymotrypsin- and trypsin-treated erythrocytes. These findings support that NBD94_1–550 contains the minimal erythrocyte binding domain (EBD) of Py235.

Fig. 1.

Identification of the erythrocyte binding region of Py235. (A) Schematic representation of three members of the RH multigene family (PfRH1, PfR4, and Py235). The location of the erythrocyte binding region (EBD) or nucleotide binding domain (NBD) and that the transmembrane domain (TM) are indicated. The amino acid numbers for the beginning and end of the EBD and NBD are indicated. (B) Effect of enzyme treatment on erythrocyte binding of the entire Py235 protein. Five micrograms of Py235 was incubated with untreated (Un), neuraminidase-treated (Ne), chymotrypsin-treated (Chymo), or trypsin-treated (Tryp) mouse erythrocytes. Purified Py235 (left) was detected by anti-His antibody.

Fig. 2.

Delineation of the erythrocyte binding domain of Py235. (A) Schematics of NBD94 and NBD94_1–550. (B) Erythrocyte binding assay with recombinant NBD94. One hundred microliters of mouse erythrocytes were incubated with 0 μg, 5 μg, or 10 μg of recombinant NBD94 (left). Purified NBD94 was detected by anti-His antibody. The effect of enzyme treatment on erythrocyte binding of NBD94 was analyzed by incubating 5 μg of NBD94 with neuraminidase-treated (Ne), chymotrypsin-treated (Chymo), trypsin-treated (Tryp), or untreated (Un) mouse erythrocytes. (C) Erythrocyte binding assay with recombinant NBD94_1–550. One hundred microliters of mouse erythrocytes were incubated with 0 μg, 2.5 μg, 5 μg, or 10 μg of NBD94_1–550 (left). Purified NBD94_1–550 was detected by anti-His antibody. Like NBD94, the recombinant NBD94_1–550 (5 μg) was incubated with untreated (Un), neuraminidase-treated (Ne), Chymotrypsin-treated (Chymo), or trypsin-treated (Tryp) mouse erythrocyte (right).

A 194-amino-acid region of NBD94 specifically binds to mouse erythrocytes.

To further delineate this binding domain, two expression constructs were designed based on the predicted secondary structure of NBD94 and are termed EBD_1–194 and EBD_1–398 (Fig. 3 A). The expressed N-terminal His-tagged proteins were successfully purified using Ni-NTA. Analysis of both recombinant proteins on SDS-PAGE showed that the proteins were purified successfully and contained little contamination (Fig. 3B and C). Erythrocyte binding assays using increasing concentrations of EBD_1–398 (Fig. 3B) or EBD_1–194 (Fig. 3C) revealed concentration-dependent binding of both proteins to mouse erythrocytes. It was important to establish whether EBD_1–194 still bound to mouse erythrocyte receptors with the same specificity as previously described (45). Again, EBA using neuraminidase-, trypsin-, and chymotrypsin-treated erythrocytes was carried out (Fig. 3B and C). As seen for the full-length Py235 protein, neuraminidase treatment of the erythrocyte had no effect on the binding of EBD_1–194 (Fig. 3C), while treatment with trypsin or chymotrypsin (Fig. 3C) significantly reduced the overall binding of the protein to erythrocytes. This confirmed that recombinant EBD_1–194 contains the erythrocyte binding domain of Py235. Amino acid sequence alignment of EBD_1–194 with the previously identified binding region of PfRH1 and PfRH4 (see Fig. S1 in the supplemental material) shows limited amino acid conservation in the different binding regions.

Fig. 3.

Binding specificity of EBD_1–194. (A) Arrangement of EBD_1–398 and EBD_1–194 inside the NBD94 domain. (B) EBD_1–398 was isolated with a reasonable purity suitable for an erythrocyte binding assay. Here, the 125 mM imidazole fraction of the Ni-NTA purification of this protein was obtained and dialyzed with 1× PBS prior to the erythrocyte binding assay. One microgram, 2.5 μg, or 5 μg of the recombinant EBD_1–398 was used in the erythrocyte binding assay, and a concentration-dependent binding of the protein to the red blood cells is indicated. EBD_1–398 was detected by an anti-His antibody. (C) EBD_1–194 after Ni²⁺-NTA affinity chromatography (left) binds in a concentration dependent manner to the red blood cells. Five micrograms of EBD_1–194 was incubated with untreated (Un) or neuraminidase-treated (Ne), chymotrypsin-treated (Chymo), or trypsin-treated (Tryp) mouse erythrocyte. EBD_1v194 was detected by an anti-His antibody.

Secondary structural content of homogenous EBD_1–194.

In order to perform structural studies with the smallest segment of the erythrocyte binding region still binding erythrocytes, EBD_1–194 was further purified to high purity and monodispersity using size exclusion chromatography (see Fig. S2A in the supplemental material). The secondary structure of recombinant EBD_1–194 was determined from circular dichroism spectra, measured between 190 and 260 nm (see Fig. S2B). The maximum at 192 nm and the minima at 208 and 222 nm indicate the presence of α-helical structures in the protein. The average secondary structure content was 75% α-helix and 25% random coil. This result is consistent with secondary structure predictions based on EBD_1–194 amino acid sequence. The molar ellipticity values at 208 nm and at 222 nm are in a ratio of 0.84.

Determination of molecular mass and overall dimensions of EBD_1–194 in solution.

To determine the native molecular mass (MM) of EBD_1–194 in solution, a Superdex 75 gel filtration column was calibrated by determining the K_av values for a set of standard proteins of known molecular masses. A calibration curve based on these K_av values is shown in Fig. S3A in the supplemental material. Comparison of the K_av for subunit C versus the standard proteins suggests a native molecular mass of approximately 45 ± 3 kDa (see Fig. S3B), demonstrating that EBD_1–194 is dimeric in solution. In a complementary approach, SAXS patterns from solutions of EBD_1–194 were recorded and processed as described in Materials and Methods to yield the final composite scattering curve in Fig. 4 A. The data indicate that the protein is monodispersed in solution. Inspection of the Guinier plots of EBD_1–194 at low angles indicated good data quality and no protein aggregation. The radius of gyration, R_g, of EBD_1–194 is 3.59 ± 0.3 nm, and the maximum dimension, D_max, of the protein is 11.8 ± 0.3 nm. Comparison of the forward scattering with the values obtained from a reference solution of bovine serum albumin (BSA) (66.4 ± 2 kDa) yields a molecular mass of 46 ± 2 kDa, in agreement with the results of gel filtration chromatography and supporting that EBD_1–194 is dimeric at the concentrations used. The excluded (Porod) volume of the hydrated particle in solution was 87.9 ± 5 nm³, suggesting a relatively high hydration of EBD_1–194 and characteristic for molecules with larger specific surface accessibility to the solvent. Qualitative analysis of the distance distribution function suggests that EBD_1–194 consists of a major mass, yielding a principal maximum in the p(r) around 2.8 nm (Fig. 4B), with a tailing of the peak between 6 nm and 11.6 nm.

Fig. 4.

Solution X-ray scattering data of EBD_1–194. Experimental scattering data (o) and the fitting curves (—; green, experimental; red, calculated from ab initio model) for EBD_1–194. (B) The distance distribution function of the same protein. (C) Low-resolution structure of EBD_1–194 determined from SAXS data. The average structure of the 10 independent reconstructions, obtained using the program DAMMIN (58), is shown. (D) The right model is rotated by 90° around the y axis. (E and F) Two orientations of the superposition of the solution shape of dimeric EBD_1–194 (brown) with the crystallographic structure of the F1-F2 tandem of P. falciparum EBA-175 (green; PDB entry 1RZL).

Low-resolution structure of EBD_1–194 in solution.

Small-angle scattering of X rays is an established method providing three-dimensional low-resolution structures in solution and is highly complementary to the higher-resolution methods of nuclear magnetic resonance (NMR) spectroscopy and X-ray crystallography as demonstrated for the solution structures of the E. coli F₁-ATPase (59), the chitin binding protein, CHB1 (60), subunit F of the A₁A_O ATP synthase (52), subunit C of the Saccharomyces cerevisiae V-ATPase (4), or the NS3 protease-helicase from dengue virus (38), whose structural features were confirmed by X-ray crystallography (15, 30, 38, 66) or NMR spectroscopy (15, 24, 25, 30, 38, 66). Here, the gross structure of dimeric EBD_1–194 was restored ab initio from the scattering patterns in Fig. 4A. The obtained shape for the protein yields a good fit to the experimental data in the entire scattering range. The corresponding fit, shown in Fig. 4A, has a discrepancy of χ² = 1.032. All 10 independent reconstructions yielded a reproducible shape and have been averaged (Fig. 4C and D). The ab initio modeling produced an elongated structure with two domains of different sizes (Fig. 4C). The larger domain is about 6.0 ± 0.2 nm in length and 4.5 ± 0.3 nm in diameter, whereas the smaller one is about 5.5 ± 0.3 nm in length and 3.0 ± 0.3 nm in width. The smaller domain is tilted by an angle of around 51 (4)° from the larger domain. It is tempting to compare the EBD_1–194 structure with a thumb, consisting of a thenar (larger domain) attached to the phalanx (smaller domain). This gross structure very much resembles the shape of the tandem proteins F1 and F2 of the Duffy binding-like domains of the P. falciparum erythrocyte-binding antigen 175 (EBA-175) (Fig. 4E and F) (PDB entry 1RZL) with an intertwining of the two molecules F1 and F2, as described recently (64). As revealed in Fig. 4E and F, the crystallographic structure of F1 and F2 of the DBL domain fitted remarkably well in the low-resolution solution structure of the dimeric EBD_1–194, with a root mean square (rms) deviation being 0.95 Å. The superposition of the low-resolution structure with the F1-F2 DBL domain was carried out using the software program SUPCOMB, which minimizes the normalized spatial discrepancy measure between the models. The small deviation of 0.95 Å explains that both the structure of EBA-175 and the low-resolution structure of EBD_1–194 fit well. Further, the radius of gyration (R_g) calculated for EBA-175 by the CRYSOL software program is 3.65 nm, which compares well with the R_g value of EBD_1–194 (3.59 nm). These values confirm that the EBA-175 structure fits within the Py235 low-resolution structure.

DISCUSSION

The conservation of the RH superfamily in all Plasmodium species so far analyzed would indicate an essential role of this protein family during host cell invasion. In addition to its role in host cell selection, immune evasion, and virulence during the blood stage of the infection, it is also implicated in playing a role in the sporozoite stage of the infection. Antibodies targeting Py235 of P. yoelii have been shown to successfully protect the host against the virulent strain of this parasite (18, 33), and in vitro culture experiments have also demonstrated that antibodies against RH of P. falciparum can inhibit merozoite invasion (5, 19, 23, 63). While the redundancy of this protein family poses a challenge, the available data suggest that one or more RH may be an ideal component of a future vaccine targeting merozoite invasion. To achieve such a goal, it is important to identify functional domains of this protein that are targets of a protective immune response. The large size and limited sequence homology have made it difficult to identify functional domains, similar to the DBL domains of the EBL, within members of the RH superfamily. In addition to size, it is indeed very challenging to express correctly folded soluble recombinant protein, partially due to the amino acid bias introduced by the high AT content of the P. falciparum sequence.

Previous studies have suggested that RH contain at least two functional domains, one involved in erythrocyte binding while the other is involved in nucleotide binding. Our study now demonstrates that in Py235 the previously identified nucleotide binding domain (NBD94) (48) also mediates specific erythrocyte binding. This close proximity of both nucleotide binding and erythrocyte binding could explain the observation that structural changes induced by ATP or ADP binding to the NBD directly impact on the erythrocyte binding capacity of full-length Py235. It is clear from the work here, though, that the EBD is partially independent from the region involved in nucleotide binding, suggesting that RH are multidomain proteins containing an EBD as well as NBD along with possibly other, so far uncharacterized functional regions. In that sense, the RH proteins are starting to look similar to PfEMP1, a variant surface antigen of P. falciparum that has receptor binding capacity and is arranged as a multidomain protein containing both DBL and cysteine-rich interdomain region (CIDR) domains (54). A key challenge now is to identify whether there are any other functional domains within the RH that provide further insight into the function of this complex protein family.

The overall α-helical content of the EBD of Py235 is around 75% as determined by CD spectroscopy, and it is intriguing to note that this high helical content appears to be a general feature of other so far characterized Plasmodium receptor binding proteins, including the DBL domains of the P. falciparum EBA175 protein (64) and Plasmodium knowlesi Duffy binding protein (53), as well as the CIDR domain of PfEMP1 (36). This suggests that despite limited or lack of apparent sequence conservation, certain structural characteristics are conserved across receptor binding domains of Plasmodium. The single recombinant F1 and F2 proteins of the DBL domain of P. falciparum EBA-175 (64), as well as the recently solved crystal structures of the P. knowlesi DBL domain (53) and the P. falciparum CIDR domain, called MC179 (36), have a similar boomerang shape with two subdomains, one formed by a three-helix bundle. In solution and in the crystal, the recombinant MC179 forms a dimer, resulting in an extended shape in which both MC179 molecules reveal an intertwining (36). By comparison, the related F1 and F2 of the P. falciparum EBA-175 are arranged in tandem with an intertwining of both proteins relative to each other, resulting in an excluded volume of 82.2 nm³, comparable with that determined for the hydrated EBD_1–194 dimer (87.9 nm³) of P. yoelii. The elongated shape of the F1-F2 tandem makes the surfaces for the binding of the glycan molecules of glycoprotein A on human erythrocytes accessible. As demonstrated in Fig. 4E and F, the F1-F2 tandem of EBS-175 superimposes nearly perfectly on the solution shape of dimeric EBD_1–194, resulting in a sufficient contact area of P. yoelii EBD in binding to the red blood cell receptor.

In summary, our findings provide the first structural characterization of the EBD of a member of the RH family. The findings reported here also suggest that despite significant overall sequence diversity, there may be conservation of the overall structure of different receptor binding domains in Plasmodium. The low-resolution solution structure of EBD_1–194 together with the biochemical results presented will provide the foundation for future structural and mutational studies to selectively identify residues and epitopes of EBD_1–194 interacting with the red blood cell receptor protein.