Inhibition of parasite invasion by monoclonal antibody against epidermal growth factor-like domain of Plasmodium vivax merozoite surface protein 1 paralog

The Plasmodium vivax merozoite surface protein 1 paralog (PvMSP1P), which has epidermal growth factor (EGF)-like domains, was identified as a novel erythrocyte adhesive molecule. This EGF-like domain (PvMSP1P-19) elicited high level of acquired immune response in patients. Antibodies against PvMSP1P significantly reduced erythrocyte adhesion activity to its unknown receptor. To determine PvMSP1P-19-specific antibody function and B-cell epitopes in vivax patients, five monoclonal antibodies (mAbs) and 18-mer peptides were generated. The mAb functions were determined by erythrocyte-binding inhibition assay and invasion inhibition assay with P. knowlesi. B-cell epitopes of PvMSP1P-19 domains were evaluated by peptide microarray. The pvmsp1p-19 sequences showed limited polymorphism in P. vivax worldwide isolates. The 1BH9-A10 showed erythrocyte binding inhibitory by interaction with the N-terminus of PvMSP1P-19, while this mAb failed to recognize PkMSP1P-19 suggesting the species-specific for P. vivax. Other mAbs showed cross-reactivity with PkMSP1P-19. Among them, the 2AF4-A2 and 2AF4-A6 mAb significantly reduced parasite invasion through C-terminal recognition. The linear B-cell epitope in naturally exposed P. vivax patient was identified at three linear epitopes. In this study, PvMSP1P-19 N-terminal-specific 1BH9-A10 and C-terminal-specific 2AF4 mAbs showed functional activity for epitope recognition suggesting that PvMSP1P may be useful for vaccine development strategy for specific single epitope to prevent P. vivax invasion.

Sixty-six isolates of pvmsp1p from PlasmoDB (http://plasmodb.org/) originating from 10 countries (Brazil, China, Columbia, India, Mauritania, Mexico, North Korea, Peru, Papua New Guinea, and Thailand) were used for nucleotide diversity analysis. The nucleotide diversity (π) showed 0.00066 within worldwide isolates, thus indicating that pvmsp1p had limited polymorphism (Fig. 1b). The thirty pvmsp1p sequences from Republic of Korea (ROK), Thailand and Myanmar were newly sequenced in this study and sequence alignment indicated a conserved EGF-like domain ( Table 1). The sequence alignfment of thirty isolates are described as Supplementary Data 1. The nucleotide diversity (π) comparison between EGF-like domain of pvmsp1-19 (0.00060) and pvm-sp1p-19 (0.00032) indicated that low polymorphism occurred in pvmsp1p-19 ( Supplementary Fig. S1).
Erythrocyte binding inhibitory ability. Erythrocyte binding inhibitory ability of PvMSP1P-19 specific antibody was confirmed previously 19,20 . The functional activity of five monoclonal antibodies were confirmed by erythrocyte binding inhibition in in vitro system. A serial dilution of each mAb and the non-treated control were used for erythrocyte binding inhibition activity confirmation. The inhibitory effects on mAbs were described as the percentage of erythrocyte binding inhibition compared to the mock control. 1BH9-A10 showed a binding inhibitory effect in a concentration-dependent manner (Fig. 4, Supplementary Fig. S4). However, 2AF4-A2, 2AF4-A6, 3BC6-A5, and 3BC6-B12 did not affect erythrocyte binding inhibition (Fig. 4).
Cross-reactivity with P. knowlesi. P. vivax and P. knowlesi MSP1P EGF domains contained complete conservation of cysteine residues with high amino acids sequence similarity (86%) 24 (Fig. 5a). Immunoblot analysis was performed with P. knowlesi derived recombinant MSP1P EGF domain and naïve parasite to determine cross-reactivity confirmation. Four mAbs clearly recognized the rPkMSP1P-19 and naïve P. knowlesi schizont lysate except 1BH9-A10 which has erythrocyte binding inhibition activity (Figs 5b,c). The naïve P. knowlesi subcellular recognition by the mAbs in the mature schizont were observed on the merozoite surface. However, 1BH9-A10 mAb showed poor recognition with P. knowlesi as it is consistent with immunoblotting ( Fig. 5d, Supplementary Figs S2 and S3). It may suggest the species-specific recognition of 1BH9-A10 mAb to P. vivax.  www.nature.com/scientificreports www.nature.com/scientificreports/ Parasite invasion inhibitory ability. To determine PvMSP1P-19 mAb functions to abrogate merozoite invasion, the human erythrocyte-adapted P. knowlesi parasite was used in ex vivo invasion inhibition assay. P. knowlesi parasites are used for vivax study model, especially during the blood stage, because of the high sequence similarity with conserved cysteine in the EGF-like domain and a feasible in vitro culture 24 .
PvMSP1P-19 linear B-cell epitope in P. vivax patients. The B-cell linear epitopes of PvMSP1P in natural infection were identified by peptide array with vivax patient sera (Table 1) at 1:25 dilution and visualized by goat anti-human Alexa Fluor 546 antibodies. The normalization of the mean fluorescence intensity (MFI) of the average healthy group was compared to that of each peptide group (Table 2, Supplementary Fig. S5). Three peptides, S1764, E1773, and N1791, detected an epitope in the vivax patient that was in line with humoral immune reactivity ( Table 2). The peptides C1809 (p = 0.0272) and C1818 (p = 0.0012) in the C-terminal region also showed significantly high IgG response in patient groups. Among them, E1773 showed the highest sero-positivity of 70% (1.80-fold higher than healthy individuals), followed by S1764 and N1791 at 53.3% (1.63-fold) and 50.0% (1.62-fold), respectively ( Table 2). The C-terminal domain of peptides C1809 and C1818 showed only 20.0% (1.19-fold) and 26.7% (1.44-fold) sero-positivity. The first B-cell epitope in the vivax patient was determined to be at overlapping sites of S1764 and E1773, which are found in the short peptide 1773 EECLCLLNY 1781 . The second linear B-cell epitope was detected at 1791 NEQNSCAVKNGGCDLKAT 1808 . This linear epitope showed that only the complete sequence of N1791 can elicit an antibody response, which is at the central part of the sequence of the B-cell epitope.

Discussion
According to the P. vivax research in previous reports, merozoite surface antigens such as PvMSP1, PvMSP3α, PvMSP3β, PvMSP8, PvMSP9, Pv92, and PvMSA180 elicited high antibodies responses 19,[27][28][29][30][31][32][33] . The functional activity of antibodies in malaria has multiple roles, such as merozoite invasion inhibition, agglutination, growth inhibition, rosetting inhibition, opsonic phagocytosis, and direct killing by complement mediation 34 . Both P. falciparum and P. vivax express MSP1, a major antigen of the merozoite surface that is extensively studied. The antibody against PfMSP1 EGF-like domain (PfMSP1-19) was confirmed to inhibit invasion in previous reports 35,36 . Processing of MSP1 was also found to play role in parasite viability and merozoite egress 9 . Additionally, MSP1 has successfully mapped antibody epitopes and their roles in the two EGF-like domains 37 . In line with the MSP1 study, an antibody against the EGF-like domain of PvMSP1P shows functional activity for erythrocyte binding www.nature.com/scientificreports www.nature.com/scientificreports/ interruption 7,20,36,38 . Thus, as examined in this study, functional epitopes for immune responses and erythrocyte binding ability are important for understanding the biology of P. vivax in patients.
Five monoclonal antibodies were produced and confirmed using indirect fluorescence assay in naïve P. vivax and P. knowlesi parasites. The 2AF4-A2, 2AF4-A6, 3BC6-A5, and 3BC6-B12 staining pattern was scattered on the vivax merozoite surface. It might be due to that epitope does not expose or masking so that monoclonal antibodies could not recognize whole part of PvMSP1P-19 domain as polyclonal antibody does. Meanwhile, PvMSP1P-19 in the vivax patients was highly antigenic at the acute phase in 68.0% of ROK (n = 102) and 72.5% of Thailand (n = 40) isolates 19,39 . These antibodies were stable up to nine months after the patient recovered, and PvMSP1P-19 in the high responder was directly related to the erythrocyte binding inhibition in in vitro system, but not in low responder 39 . Direct erythrocyte binding inhibition ability showed that 1BH9-A10 mAb recognized the C1755 peptide. However, the C1755 peptide position failed to induce an immune response in most vivax patients. This result indicated that the high-responder patient IgG might contain the C1755 recognition antibodies. For parasite invasion inhibition, the effective B-cell epitope was identified in the C-terminal domain as 1818 CICPKGTKPMHEGVVCSF 1835 . In this position, IgG expression was induced in the patient; however, low sero-positivity (26.7%) was observed. Due to the difficulty to perform invasion inhibition assay with P. vivax, the invasion inhibition assay was performed with P. knowlesi. P. knowlesi and P. vivax showed high homology rate in various antigens including msp1p gene which can be used as an alternative way for vivax study model 24,40 . However, 1BH9-A10 which erythrocyte binding inhibitory mAb did not recognize in P. knowlesi. This result www.nature.com/scientificreports www.nature.com/scientificreports/ indicated that 1BH9-A10 recognition site was hampered by the polymorphic residues between orthologue at the N terminus. Additionally, the discrepancy of erythrocyte binding and invasion inhibition result might be caused by following reasons. The invasion inhibitory mAb was also observed in PfMSP1 by inhibition of processing of PfMSP1-42 36 . PvMSP1P also showed processing 19 , and the mAbs 2AF4-A2 and 2AF4-A6 could inhibit the processing of PvMSP1P. Additionally, previous studies indicate that direct binding and antibody recognition sites are not related 36,41 . The parasite growth inhibition ability of mAbs was confirmed by morphological examination. The morphology under the mAb treatment condition was maintained in the healthy condition, which indicates that mAbs against PvMSP1P-19 have no effect on parasite growth.
The merozoite surface antigen has been identified as a promising vaccine candidate from the malaria asexual stage 42 . Generally, merozoite surface antigens have a major role in invasion as a trigger for initial attachment to host target red blood cells 10,30,43 . Because of surface localization, genetic polymorphisms frequently occur by selection from host immune pressure in natural infection 44 . These polymorphisms are raising a problem in vaccine trials 45 . However, a surface antigen remains a promising target for a malaria vaccine because the merozoite surface antigens have a longer exposure time until internalization. To overcome the low efficacy of the malaria vaccine, multiple antigen combinations seem to be a good alternative [46][47][48] . Along with these combinations, the antigen that has resistance to host immune pressure in nature and elicits a high humoral immune response will be designed as a single or multiple vaccine candidate. PvMSP1P-19 confirmed that worldwide isolate sequences were found limited polymorphism. The conserved sequences will avoid vaccine efficiency problems from allele The black arrow head indicates the mature schizont stage of P. knowlesi before invasion. The black arrow points to an unhealthy parasite, and the red arrow indicates the healthy ring stage of the parasite after re-invasion. (b) FACS gating strategy of ring stage parasitaemia evaluation. The schizont stage parasites are considered to have SYBR green signals of more than log 10 6 . (c) The P. knowlesi invasion inhibition efficacy was confirmed by an invasion inhibition assay. Data are shown as the invasion inhibition rate mean ± standard deviation (S.D.) with pre-immune sera (PI) and anti-2C3 (murine anti-Fy6) and PvMSP1P-19 monoclonal antibodies. Significant differences between PI and anti-2C3 or monoclonal antibodies calculated with a one-way ANOVA with the Tukey post-test. ***p < 0.001. www.nature.com/scientificreports www.nature.com/scientificreports/ specificity. Taken together, PvMSP1P-19 elicits stable and high IgG responses in vivax patients regardless of polymorphisms from host immune pressure. The functions of antibodies showed possibility of parasite invasion abrogation. In summary, two effective epitopes were identified at 1755 CRNRKCPLNSFCFIQTIN 1772 and 1818 CICPKGTKPMHEGVVCSF 1835 for disrupting of PvMSP1P with erythrocytes interaction. However, these epitope regions should overcome the lack of natural boosting. Thus, we propose that PvMSP1P can be considered for a vaccine design strategy either for multiple or single vaccine development.

Materials and Methods
Sample collection and ethical clearance. The  Genetic diversity analysis. Thirty isolates from ROK, Thailand and Myanmar were used to determine sequence diversity of pvmsp1p along with PlasmoDB database in this study. P. vivax genomic DNA was extracted from whole blood samples of vivax malaria patients using the QIAamp DNA Blood Mini Kit (QIAGEN, Hilden, Germany) according to the manufacturer's protocol. The pvmsp1p-19 gene was amplified by the forward primer (5′-GACACCCTACACACAATCAACACT-3′) and reverse primer (5′-CTACGCAGTGACGAACGCGAGG-3′) with AccuPower ® Pfu PCR premix (Bioneer, Seoul, ROK) polymerase. The amplicon nucleotide sequence was analysed by the internal primer (5′-TGAAGTGCAACACGTGGAAT-3′) using an ABI 3700 Genetic Analyzer (Genotech, Daejeon, ROK). All the raw sequences were analysed and trimmed using the SeqMan software, Lasergene ver. 7.0 (DNASTAR, Madison, WI, USA).
Sequence diversity (π) and graphical visualization of thirty isolates from newly sequenced in this study and sixty-six pvmsp1p worldwide sequence from PlasmoDB were analysed using the sliding window option with window length 100 and step size 25 sites in DNAsp ver. 5.0 software.

Accession numbers. The nucleotide sequences of PvMSP1P-19 sequences are available under GenBank
Accession Numbers MF968906 to MF968935. Recombinant PvMSP1P-19 (rPvMSP1P-19) was expressed by the wheat germ cell-free system (WGCF) (Cell-free Science, Matsuyama, Japan) and recombinant protein was confirmed by 12% sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) as described previously 19 . Briefly, pEU-E01-His-TEV-MCS (Cell-free Science) vector were cloned with PvMSP1P-19 fragment. The recombinant protein expression was scaled up as large scale WGCF expression system manufacture's protocol and purified with Ni-affinity chromatography. For monoclonal antibody production, female BALB/c mice were immunized intravenously with an rPvMSP1P-19 mixture containing Freund's complete adjuvant (Sigma-Aldrich, St. Louis, MO, USA) and boosted at a two-week interval with an rPvMSP1P-19 mixture containing Freund's incomplete adjuvant (Sigma-Aldrich). Four days later, anti-PvMSP1P-19 polyclonal antibodies were collected, and the spleen was used to produce hybridoma cell lines, as done previously 49,50 . Hybridoma culture supernatants www.nature.com/scientificreports www.nature.com/scientificreports/ were screened for antibody reactivity by enzyme-linked immunosorbent assay (ELISA) and indirect immunofluorescence assay (IFA). Positive cells were cloned by two rounds of limiting dilution, and the antibody isotype was determined using a monoclonal antibody isotyping kit (Santa Cruz Biotechnology, Inc., Santa Cruz, CA, USA) according to the manufacturer's protocol. The cloned cell lines were expanded as ascites in mice primed with pristane (Wako Pure Chemical Industries, Osaka, Japan). Immunoglobulin G (IgG) was purified from the ascitic fluid using the MAbTrap Kit by following the manufacturer's protocol (GE Healthcare, Little Chalfont, UK).

Monoclonal antibody production.
Erythrocyte binding inhibition assay. The PvMSP1P-19 construct and associated primers were described previously 19 . Briefly, the PvMSP1P-19-containing pEGFP-HSVgD vector was used for target expression on the COS-7 cell surface. The PvMSP1P-19 amplicon was ligated using the In-Fusion ® HD Cloning Kit (Clontech, Palo Alto, CA, USA). This construct was cloned in JM109 competent cells (Real Biotech Corporation, Taiwan), and plasmid DNA was purified by the Ultrapure plasmid extraction system (Viogene, Taipei Peptide synthesis. The P. vivax Salvador-I strain sequence was used for synthesis of PvMSP1P-19 peptides. Sequential 18-mer peptides on 9-mer overlapping amino acids were chemically synthesized using Multiple Peptide Synthesis techniques in Solid Phase. The identity and purity of the peptides were analysed by analytical reversed phase-high-performance liquid chromatography (RP-HPLC) and mass spectrometry MALDI-TOF. All peptides solubilized in DMSO with more than 90% purity. The peptide sequence and information are described in Fig. 1a.
Mapping the linear epitope for monoclonal antibodies and clinical isolates. Three-aminopropylcoated glass slides were prepared as described previously 4 . Each peptide was labelled with Cy5 NHS-Ester (GE Healthcare) for array slide printing efficiency calculation. The peptides were diluted with 100 mM sodium bicarbonate buffer (pH 8.3) and 2 µg of Cy5 NHS-Ester (GE Healthcare) and incubated on ice for 2 hours. The reaction mixture was quenched with 1 M Tris-HCl (pH 8.0) solution and purified by Sephadex G-25 columns (GE Healthcare). The labelled peptides were diluted with 400 mM 1-ethyl-3-(3-dimethylaminopropyl) carbodiimide (EDC; Thermo Fisher Scientific Inc., Rochester, NY, USA) and 100 mM N-hydroxysuccinimide (NHS; Thermo Fisher Scientific Inc.) for peptide coupling to the amine slide. A total of 100 ng/μl of each peptide mixture was printed on the amine slide spot and incubated for 30 minutes at room temperature. The peptide printing efficiency was calculated by the Odyssey infrared imaging system (LI-COR Bioscience). The peptide printed array slides were probed with 1:200 dilution monoclonal antibodies for determining the monoclonal antibody binding epitope or probed with a 1:25 dilution of patient sera to identify the PvMSP1P-19 B-cell epitope in vivax patient. The arrays for visualization were incubated with 50 ng/μl goat anti-mouse and -human Alexa Fluor 546 antibodies (Invitrogen) and scanned in an Innoscan-300 (Innopsys, Carbonne, France). Gladbach, Germany) with an LD column (Miltenyi Biotec) and sub-cultured for 10 hours with fresh medium, erythrocytes and 100 μg/ml PvMSP1P-19 monoclonal antibodies. The sub-cultured parasites were adjusted to 1.5% parasitaemia with 2% haematocrit. Additionally, 25 μg/mL DARC monoclonal antibody (2C3) (a kind gift from Renia L, Singapore Immunology Network-BMSI-A STAR) was used as an invasion blocking control 25,26 . After invasion, the ring stage was examined by thin smear with Giemsa staining under light microscopy and fluorescence-activated cell sorting (FACS). For FACS analysis, newly infected RBCs were fixed with glutaraldehyde (0.05%) and stained with SYBR green (Invitrogen). The data were analysed by a FACS Accuri ™ C6 Flow www.nature.com/scientificreports www.nature.com/scientificreports/ for 1 hour at 25 °C. Data were measured using the Odyssey infrared imaging system (LI-COR Bioscience) and analysed with Odyssey software (LI-COR Bioscience).

P. knowlesi in vitro culture and invasion inhibition assay.
Immunofluorescence assay. P. vivax thin smears from Korean isolates schizonts and P. knowlesi A1-H.1 schizonts were prepared with ice cold acetone fixation, and the smears were blocked with 5% BSA in PBS for 30 minutes at 37 °C. The parasite smears were incubated with a 1:100 dilution of mice monoclonal antibodies and anti-PvMSP1-19 antibody from rabbit for 1 hour at 37 °C, followed by incubation with a 1:500 dilution of Alexa Fluor 488 conjugated goat anti-mouse IgG (Invitrogen) and Alexa Fluor 546 conjugated goat anti-rabbit IgG (Invitrogen) as secondary antibodies for 30 minutes at 37 °C. The parasite nuclei were stained with 2 μg/ml DAPI (4′,6-diamidino-2-phenylindole) in the secondary antibody mixture. Slides were mounted by ProLong ® Gold Antifade reagent (Invitrogen) and visualized by a Fluoview ® FV1000 Laser Scanning Confocal Imaging System (Olympus, Tokyo, Japan) under the 60x objective oil-immersion lens. Images were visualized by FV10-ASW 3.0 viewer software.