PyEBL (PY17X_1337400) regions were prepared as described (Otsuki et al., 2009). Briefly, the sequences encompassing the ectodomain or other specific regions and/or domains were amplified from P. yoelii 17X genomic DNA by PCR by using sense primers with XhoI restriction sites and antisense primers with NotI restriction sites ( Supplementary Table S1 ). The amplified fragments were then restricted and ligated into the pEU-E01-GST expression vector (CellFree Science, Matsuyama, Japan) followed by sequencing with an ABI PRISM 3100-Avant genetic analyzer (Applied Biosystems, Foster City, CA). The GST-tagged proteins were expressed with WGCFS (CellFree Sciences) and purified using glutathione-Sepharose 4B columns (GE Healthcare, Camarillo, CA) as previously described (Otsuki et al., 2009).

Expression constructs of putative mouse erythrocyte surface proteins were prepared as previously described (Miura et al., 2018). Membrane and GPI anchored proteins (n=237) were selected from the FANTOM (functional annotation of mouse cDNA) database on the basis of being putative mouse orthologs of the human genes that encode erythrocyte proteins as annotated in the Ensemble (http://asia.ensembl.org/) database ( Supplementary Table S2 ). The proteins were expressed using the WGCFS as N-terminal mono-biotinylated recombinant proteins as previously described (Miura et al., 2018). For further characterization, cDNA of the C57BL/6J mouse basigin gene (FANTOM ID: 0610008G13) was ligated into pEU-E01-GST, and protein was expressed with WGCFS and purified using glutathione-Sepharose 4B column (GE Healthcare) as described (Ito et al., 2013). All specific primers used in this study are presented in Supplementary Table S1 .

Plasmodium yoelii parasite soluble extracts were prepared following intraperitoneal injection of 8-week-old female BALB/c mice (Charles River, Yokohama, Japan) with P. yoelii 17XL parasites as previously described (Rungruang et al., 2005). Briefly, blood from infected mice was depleted of leukocytes using a CF11 cellulose (GE Healthcare, Buckinghamshire, UK) column. Plasmodium yoelii extracts containing PyEBL were prepared by freeze-thawing schizont-infected erythrocyte pellets in the presence of protease inhibitors (1 μg/ml of leupeptin, 1 μg/ml of pepstatin A, 100 μM 4-(2-aminoethyl) benzenesulfonyl fluoride hydrochloride) with 1 mM EDTA. The supernatant containing the soluble extract of approximately 1 × 10⁷ parasites/μl was obtained by centrifugation at 21,600 × g for 10 min and subsequently used for screening assays.

Interactions between PyEBL and the library of mouse erythrocyte proteins was assessed via AlphaScreen (PerkinElmer, Waltham, MA) as previously described (Miura et al., 2018). Briefly, the reaction mixture consisted of 1 μl of P. yoelii extracts, 1.3 μg/ml mouse anti-PyEBL polyclonal antibodies and 1 mg/ml bovine serum albumin (BSA) (Wako, Osaka, Japan) in PBS, for a total volume of 9 μl/well in a 384-well OptiPlate (PerkinElmer). Biotinylated recombinant mouse protein (1 μl) was then added and incubated at 26°C for 30 min. A mixture of streptavidin-coated donor-beads and protein G conjugated acceptor-beads (PerkinElmer) was added to the mixture and incubated for 1 h at 26 °C in the dark to allow the donor- and acceptor-beads to optimally bind to biotin and IgG, respectively. Upon illumination of this complex, a luminescence signal at 620 nm termed ‘raw AlphaScreen Counts’ (ASC) was detected and quantified by an EnVision plate reader (PerkinElmer). Flotillin 2 (FANTOM ID: 4933417M14) or Duffy blood group (FANTOM ID: 2510001J03) that reacts with anti-Pys25 monoclonal antibody #16 (mAb#16) (Tsuboi et al., 1997) was used as a negative control. After the initial down-selection of the top 18 reactive erythrocyte proteins, the assays were repeated twice including 5 additional GPI-anchored proteins. The standardized magnitude of signal increase (AS Signal) between individual mouse target proteins and the negative control was determined using the formula: (ASC - background ASC (ASC of negative control; NC)/ASC of NC to give the AS Signal.

SPR experiments were performed using a Biacore X100 instrument (GE Healthcare) according to manufacturer’s instructions and as previously described (Nagaoka et al., 2019). Briefly, recombinant mouse basigin was immobilized to a CM5 chip (GE Healthcare) by amine coupling. The buffer HBS-EP+ (10 mM HEPES, pH 7.4, 150 mM NaCl, 3 mM EDTA, 0.05% (v/v) surfactant P20) was used as the running buffer for all SPR experiments. Blank flow cells were used to subtract buffer effects on sensorgrams. Recombinant PyEBL regions were assayed as analytes. After subtraction of the contribution of bulk refractive index and nonspecific interactions with the CM5 chip surface, protein-protein association (k _a) and dissociation (k _d) rate constants were obtained by global fitting of the data in a 1:1 binding model equation. Measurement conditions were optimized so that the contribution of mass transport to the observed values of K _D was negligible.

Animal experimental and immunization protocols were approved by the Institutional Animal Care and Use Committee of Ehime University and the experiments were conducted in accordance with the Ethical Guidelines for Animal Experiments of Ehime University (Ref: SU-17-1).

Since EBL proteins play an important role in initiation and formation of the tight junction between the apical end of the merozoite and erythrocyte surface, we sought to determine whether PyEBL binds directly to mouse erythrocytes proteins. First, a recombinant PyEBL ectodomain spanning region (R) 1 to R6 (PyEBL R1-6; amino acids V₂₈-N₇₈₈) was expressed using WGCFS ( Figure 1A ). Expression was confirmed by SDS-PAGE where it resolved as a single band of 120 kDa ( Figure 1B ). Mouse polyclonal antibodies raised against the recombinant PyEBL specifically recognized the native parasite protein as reported previously (Otsuki et al., 2009).

To identify potential receptors for PyEBL on the surface of host erythrocytes, we conducted AlphaScreen assays with the mouse erythrocyte proteins as previously described (Miura et al., 2018). A total of 237 membrane and GPI anchored mouse orthologs of the human genes that encode erythrocyte proteins were successfully expressed using the WGCFS as N-terminal mono-biotinylated recombinant proteins ( Supplementary Table S2 ). The AlphaScreen assay was designed such that the polyclonal antibody binds with native parasite PyEBL which in turn would bind with a recombinant erythrocyte protein ( Supplementary Figure S1 ). After the initial down-selection of the top 18 reactive erythrocyte proteins ( Supplementary Table S3 ), the assays were repeated twice, this time including 5 GPI-anchored proteins and a negative control. Basigin (FANTOM ID: 0610008G13) had the highest and distinct signal among the 23 erythrocyte proteins, suggesting its interaction with parasite native PyEBL ( Figure 1C ; Supplementary Table S4 ).

To validate the interaction between PyEBL and basigin, we analyzed the strength of the interaction between the two proteins using SPR. We observed that recombinant PyEBL at increasing concentration of 6, 12, 24, 48, and 96 nM, directly interacts with recombinant GST-tagged basigin with an equilibrium binding constant (K _D) of 3.3 ×10^-8 M ( Figure 1D ). As a negative control, His-GST did not show specific interactions with basigin ( Figure 1D ). Other parameters determined by the SPR are shown in Supplementary Table S5 . These observations suggest that basigin is the erythrocyte receptor for native PyEBL protein.

To determine the PyEBL region important for its interaction with erythrocyte surface protein basigin, we synthesized PyEBL R1-6 as truncated proteins with sections corresponding to the previously defined regions (Adams et al., 1992) namely region (R)1-2, amino acid V₂₈-G₄₃₇; R2, E₁₁₃-G₄₃₇; R3-5, G₄₃₇-C₇₁₇; and R6, C₇₁₇-N₇₈₈ ( Figures 1A , 2A ). We observed that recombinant PyEBL R1-2 as well as R2, the N-terminal Cys-rich Duffy binding-like region (DBL) interacted with basigin at equilibrium binding constant (K _D) of 1.6 ×10^-8 M and 3.4 ×10^-8 M, respectively ( Figure 2B ). The binding constants were comparable to that of PyEBL R1-6 ( Figure 1D ). In contrast, PyEBL R3-5 and PyEBL R6 did not bind to recombinant basigin ( Figure 2B ), suggesting that the region essential for interaction between PyEBL and basigin is located in R2.

Since a single amino acid substitution in PyEBL R6 domain determines the protein localization and influences the parasite’s virulence (Otsuki et al., 2009) while a mutation in the DBL domain affects both the parasite’s growth rate and virulence (Abkallo et al., 2017), we sought to determine whether the DBL C351Y mutation affects the ability of PyEBL to bind to its putative receptor. Sequences derived from the PyEBL C351Y mutant were cloned, confirmed by sequencing, and used to express recombinant protein. SPR analysis revealed slightly stronger albeit statistically non-significant difference in the binding kinetics between the C351Y mutant (K _D of 2.6 ×10^-8 M) and wildtype derived R1-6 proteins (Student’s t-test, p = 0.23), suggesting the mutation had but a minimal effect on the protein’s binding capabilities to basigin ( Figures 2C, D ). Taken together, the data presented here suggest that PyEBL DBL (R2) is responsible for PyEBL interaction with basigin, and that this binding may not be affected by the C351Y mutation.