Gene sequence data and protein domain information of pkmsp1-19 (accession nos. PKNH_0728900 and XP002258582) and pkmsp1p-19 (PKNH_0728800 and XP002258581) with orthologous genes were obtained from the PlasmoDB website and GenBank, respectively. The amino acid sequence data of P. knowlesi MSP1-19, MSP1P-19, and other Plasmodium species were aligned using the CLUSTAL-W program in MegAlign Lasergene software ver 7.0 (DNASTAR, Madison, WI).

Rhesus macaque (Macaca mulatta) blood was provided by the National Primate Research Center (NPRC) in the Korea Research Institute of Bioscience and Biotechnology (KRIBB) of Korea. The rhesus macaque blood and human umbilical cord blood samples were collected in a 10-ml sodium heparin tube (BD Vacutainer, Franklin Lakes, NJ). All experiments were performed under relevant guidelines and regulations, and the study was approved by the Kangwon National University Hospital Ethical Committee (IRB no. KNUH-B-2021-06-034). Human umbilical cord samples were obtained from the by-product of routine clinical diagnostic procedures of healthy donors. The written informed consent was obtained from all subjects.

Genomic DNA from the mixed stage of P. knowlesi A1-H.1 culture was extracted with QIAamp DNA Blood Mini Kit (Qiagen, Hilden, Germany) and used as a PCR template. The PkMSP1-19- and PkMSP1P-19-tagged glutathione S-transferase (GST) were expressed and purified in an Escherichia coli system as follows. Segments of pkmsp1p-19 and pkmsp1-19 were amplified using a high-fidelity KOD-plus Kit (TOYOBO, Osaka, Japan) with the following pairs of gene-specific primers: pkmsp1p-19 forward 5′-ggatccccag gaattc atGATCGTGTGAAAAATAACTGCAGA-3′ and reverse 5′-gatgcggccg ctcgag GCACACGACCCCCTCATATA-3′; pkmsp1-19 forward 5′-ggatccccag gaattc atTACAGATACTTGGACGGAACGG-3′ and reverse 5′-gatgcggccg ctcgag GCTACAGAAAACTCCCTCAAAAAG-3′. Small letters indicate the plasmid-derived sequence for In-Fusion cloning (Clontech, Kusatsu, Japan), and restriction enzymes, EcoRI and XhoI, are indicated as italicized and underlined letters. The amplicons were cloned into the pGEX-4T-2 expression vector (GE Healthcare, Wauwatosa, WI) by the In-Fusion cloning method, and the sequences of cloned plasmids were confirmed. The positive clone was cloned to recombinant proteins induced with 0.1 mM isopropyl-β-D-thiogalactopyranoside (IPTG, Sigma-Aldrich, St. Louis, MO) and purified using glutathione sepharose 4B resin (GE Healthcare) following the manufacturer’s protocol with the modified elution buffer (50 mM HCl–Tris, 20 mM reduced glutathione, 300 mM NaCl, 2% glycerol, and 200 mM imidazole, pH 8.0) (Spielmann et al., 2003).

Female BALB/c mice (6 weeks old; Dae Han Bio Link Co., Korea) and Japanese white rabbits were immunized with recombinant PkMSP1P-19 and PkMSP1-19 proteins to generate polyclonal antibodies. In total, 20 μg or 250 μg of recombinant protein was injected with Freund’s complete adjuvant (Sigma-Aldrich), followed by Freund’s incomplete adjuvant (Sigma-Aldrich) three times at a 3-week interval into mouse or rabbit, respectively, and the sera were collected at 2 weeks after the final boost. Total IgGs were purified from 1 ml immune rabbit serum using a protein G HP column following the manufacturer’s protocol (GE Healthcare). The eluted immunoglobulin G (IgG) fractions were dialyzed to incomplete RPMI 1640 (Invitrogen/Gibco, Grand Island, NY) or phosphate-buffered saline (PBS) using centrifugal devices with a 30-kDa cutoff value (Millipore, Billerica, MA). Purified IgGs were pre-adsorbed with human O⁺ erythrocytes (25 μl of RBCs per 4 mg of IgGs) for 30 min to remove any non-specific inhibitory effect (Muh et al., 2018a).

Recombinant proteins were analyzed by 13% SDS-PAGE under reducing conditions. The separated proteins were transferred onto a 0.45-μm PVDF membrane (Millipore) in semidry transfer buffer (50 mM Tris, 190 mM glycine, 3.5 mM SDS, 20% methanol) at a constant 400 mA for 40 min using a semidry blotting system (ATTO Corp., Tokyo, Japan). The PVDF membrane (Cytiva, Marlborough, MA) containing recombinant protein was blocked with 5% skim milk in PBS-T (0.5% v/v Tween-20 in 1× PBS) and then incubated with anti-GST antibody (Novagen, Reno, NV) and mouse and rabbit immune sera diluted 1:2,000 in PBS-T. After the primary antibody reaction, the membrane was incubated with the secondary IRDye^® goat anti-mouse (1: 5,000 dilution) or IRDye^® goat anti-rabbit (1: 5,000) (Li-COR^® Bioscience, Lincoln, NE) antibodies to detect antigens. The results were visualized in the Odyssey infrared imaging system and analyzed with Odyssey software (Li-COR^® Bioscience) (Lee et al., 2016).

Reticulocytes were enriched from heparinized umbilical cord blood with 19% Nycodenz solution (Accurate Chemical & Scientific Co. Carle Place, NY) in high-KCl buffer using gradient centrifugation as described previously (Sorette et al., 1992). Briefly, fresh cord blood was washed with incomplete RPMI 1640 medium, and leukocytes were removed by Acrodisc^® white blood cell syringe filter (Pall Life Sciences, Port Washington, NY). The packed cells were resuspended in high KCl buffer (pH 7.4) and incubated at 4°C for 3 h while rotating. Five milliliters of the RBC-high KCl buffer mixture was overlaid on 3 ml of 19% Nycodenz solution and centrifuged at 3,000 ×g for 30 min. Reticulocytes were harvested from the interface layer, and the concentration and purity were calculated by observing more than 2,000 RBCs in thin blood smear samples stained with new methylene blue stain solution (Sigma-Aldrich).

The flow cytometry-based direct-binding assay with 5 × 10⁵ RBCs from rhesus macaque or human in 200 μl of incomplete RPMI 1640 medium (Invitrogen/Gibco) was carried out as described previously (Ito et al., 2013). Briefly, RBCs were washed three times with incomplete RPMI 1640 medium and incubated with 0 to 40 µg/ml of GST-tagged recombinant proteins for 4 h at 25°C. Region II of PkDBPα (PkDBPα-RII) and GST-His were used as the positive and negative controls, respectively. After incubation, RBCs were washed twice with phosphate-buffered saline containing 1% BSA (PBS-BSA) and incubated with Alexa Fluor 647-conjugated mouse anti-GST monoclonal antibody (Cell Signaling, Danvers, MA) at 4°C for 1 h in the dark. The samples were washed three times with PBS-BSA and incubated with the thiazole orange (TO) Retic-COUNT reagent (BD Biosciences, Franklin Lane, NJ) for 30 min at 25°C in the dark. A total of 100,000 events were analyzed per sample using FACS Accuri™ C6 Flow Cytometer (Becton-Dickinson, Franklin Lakes, NJ) using FlowJo (v10.9.0, LLC, Ashland, OR). Unstained cells and cells with single TO staining were used for gating the normocytes and reticulocytes, respectively.

Asexual stage parasites of P. knowlesi A1-H.1 were maintained with fresh human erythrocytes in RPMI 1640-based complete medium (Invitrogen Life Technologies, Grand Island, NY) with 10% horse serum as described previously (Moon et al., 2013). To make a homogenous stage, the culture with the mixed stage was synchronized to generate protein extracts using 5% D-sorbitol (Sigma Aldrich) treatment. Ten percent parasitemia of mature schizont-enriched red bood cells was treated with 0.15% saponin (Sigma Aldrich) for 1 min on ice to lyse the RBC membrane. The pellet was washed twice with ice-cold PBS containing 1× protease inhibitor, mixed 1:1, and boiled with reducing buffer. Extracts were separated on 13% polyacrylamide gel and then blotted onto the PVDF membrane following the immunoblotting protocol. Membranes were probed with total IgG from PkMSP1-19- and PkMSP1P-19-immunized rabbits at 1 µg/ml as primary antibodies.

To confirm the localization of PvMSP1-19 and PkMSP1P-19, blood smears for indirect immunofluorescence assay (IFA) were prepared as described previously (Li et al., 2012). Slides smeared with P. knowlesi schizont-enriched blood were fixed with 4% paraformaldehyde and blocked with 5% skim milk in PBS. Rabbit anti-PkMSP1-19 and mouse anti-PkMSP1P-19 diluted in 1:50 were used as primary antibodies. Alexa Fluor 488-conjugated goat anti-rabbit IgG or Alexa Fluor 546-conjugated goat anti-mouse IgG secondary antibodies (Invitrogen Life Technologies, Carlsbad, CA) and 4′,6-diamidino-2-phenylindole (DAPI; Invitrogen Life Technologies) were used for secondary antibody staining. The slides were mounted with ProLong Gold Antifade reagent (Invitrogen) and visualized under immersion oil using the FluoView^® FV1000 Laser Scanning Confocal Imaging System (Olympus, Tokyo, Japan) equipped with a 60× oil objective. Images were captured using the FV10-ASW 3.0 Viewer software (Olympus). The fluorescence graphic containing more than 500 pixels was calculated by ImageJ (NIH, Rockville, MD).

Plasmodium knowlesi wild-type (WT) and knockout (KO) strain schizonts enriched by magnet separation (MACS, Miltenyi Biotec, Rhine-Westphalia, Germany) were cultured in a 96-well plate at 100 μl per well (Ribaut et al., 2008). Initial parasitemia and hematocrit were adjusted to 1%–2% and 2%, respectively. Purified anti-PkDBPα-RII (Muh et al., 2018a), PkMSP1P-19, PkMSP1-19, GST-His, pre-immune rabbit IgG (2 mg/ml), and anti-Fy6 monoclonal antibody against DARC (2C3; 25 μg/ml) (Absolute antibody, Wilton, UK) were mixed to a 96-well plate as tested wells. A positive invasion well was set as the parasite’s normal rupture and reinvasion ability. The parasites were incubated at 37°C in a culture chamber for approximately 10 h until newly invaded ring-stage parasites appeared. After centrifugation, the pellets were washed twice with PBS and fixed with 0.05% glutaraldehyde (Sigma-Aldrich) for 10 min. After washing twice again, the red blood cells were stained with SYBR Green I (Sigma-Aldrich) at 0.2× dilution for 10 min and washed twice with PBS. A total of 200,000 events per sample were analyzed with an Accuri C6 flow cytometer (BD Biosciences). Error bars indicate mean ± SEM for duplicate twice measurement.

Schizonts were enriched with 55% Nycodenz solution in RPMI 1640, followed by a 2-h incubation with 4-[7-[(dimethylamino)methyl]-2-(4-fluorophenyl)imidazo[1,2-a]pyridin-3-yl] pyrimidine-2-amine (compound 2, kindly provided by Dr. Michael Blackman; Francis-Crick Institute, London, UK) to prevent parasite egress (Collins et al., 2013). Compound 2 was removed from the parasite by washing with RPMI 1640 and then incubated with new red blood cells to allow invasion. The ring stage of infected RBCs was purified from the bottom layer after the 55% Nycodenz enrichment method. Newly purified ring-stage parasites were incubated for up to 24 h for maturation. The tightly synchronized schizonts were enriched again with 55% Nycodenz solution, as described above. After compound 2 removal, the parasites were incubated in RPMI 1640 for 15 min to allow initiation of egress, and then 10 μl of packed cells, including the parasites, were used for each transfection.

The tightly synchronized mature schizont-stage parasites of P. knowlesi were transfected using the Amaxa 4D electroporator (Lonza, Basel, Switzerland) and the P3 Primary cell 4D Nucleofector X Kit L (Lonza) following previous reports (Moon et al., 2016). Briefly, a 20-μg repair template and 20 μg pCas9/sg (Mohring et al., 2019) containing sgRNA sequences for pkmsp1p were mixed with P3 Primary Cell nucleofector solution, including supplement 1 (Lonza), and transferred to a Nucleocuvette™ Vessel (Lonza), followed by nucleofection with program FP158. Transfected parasites were immediately transferred to complete media with RBCs and incubated at 550 rpm for approximately 30 min at 37°C to allow invasion before transferring to standard culture conditions. After 24 h, transfected parasites were selected by drug pressure with 100 nM pyrimethamine (Sigma-Aldrich), and the medium, including pyrimethamine, was replaced at daily intervals for 5 days. The transfected parasites were cloned out by limiting dilution and confirmed by genotyping with diagnostic primers of extracted genomic DNA ( Supplementary Tables 1 , 2 ).

PCR-based repair templates were generated by a two-step nested PCR method to fuse the homology regions (HRs) for pkmsp1p gene knockout (KO) as described in our previous study (Mohring et al., 2020). Each HR was amplified with at least 25 nucleotides of homologous barcode overhangs at 3′- or 5′-terms for HR1 or HR, respectively. The HRs were then fused by nested PCR. PCR was carried out with CloneAmp (TaKaRa, Kusatsu, Japan) using the following cycle conditions: 35 cycles of 5 s at 98°C, 20 s at 55°C, and 15 s at 72°C. Seven of 50 μl reactions (total 350 μl) of the final templates were pooled, and ethanol precipitated in 10 μl. All primers used for two-step nested PCR are listed in Supplementary Tables 1 and 2 .

sgRNA candidates were screened using CRISPR RGEN Tools (Bae et al., 2014), and those with a high score of Out-of-frame Score and only one on-target effect were selected for pkmsp1p gene KO ( Supplementary Table 3 ).

Ten milliliters of cultured parasites with 2% blood were used for mRNA extraction using the AccuPrep Universal RNA Extraction Kit (Bioneer, Daejeon, Korea) following the manufacturer’s manual. cDNA was amplified from 1 μg of extracted RNA using TOPscript RT DryMIX (dT18 plus) (Enzynomics, Daejeon, Korea), and the synthesized cDNAs were used as a template for RT-PCR with GoTaq qPCR master mix (Promega, Madison, WI). RT-PCR was carried out using the following cycles: hot-start polymerase activation of 2 min at 95°C, followed by 40 cycles of 3 s at 95°C, 30 s at 60°C in AriaMx Real-Time PCR system (Agilent, Santa Clara, CA). All primers used for RT-PCR were listed in Supplementary Table 1 .

The data were analyzed with GraphPad Prism version 9.5.1 (GraphPad Software Inc., San Diego, CA) and Microsoft Excel 2016 (Microsoft, Redmond, WA). Unpaired t-test was used to compare the statistical difference values of each group of binding by protein and P. knowlesi invasion inhibition result (*p < 0.05; **p < 0.01; ***p < 0.001). Difference value p < 0.05 was considered significant.

The putative pkmsp1p gene (GenBank accession no. XP002258581 and PlasmoDB ID no. PKNH_0728800) is located upstream of pkmsp1 (XP002258582 and PKNH_0728900, respectively) on chromosome 7 and encodes a 1,874-amino-acid-long protein with predicted high molecular mass (approximately 220 kDa). The predicted primary structure of PkMSP1P consists of signal peptide (aa 1–36), serine-rich domain (aa 942–970), EGF domain 1 (aa 1,772–1,807), EGF domain 2 (aa 1,814–1,848), and GPI-anchor (aa 1,852–1,874) ( Figure 1A ). Orthologues of MSP1P were found in all human infective species except P. falciparum (P. knowlesi, P. vivax, P. ovale, and P. malariae) and two avian malaria (P. gallinaceum and P. relictum) according to the PlasmoDB database. Interestingly, the paralogue is also absent among the rodent malaria clade (including P. berghei and P. yoelii).

Because, in P. vivax, only the 19-kDa EGF-like domains of MSP1P (PvMSP1P-19) at the C-terminal region showed binding activity to human red blood cells and can be recognized by malaria-exposed patients (Egan et al., 1995; Cheng et al., 2013; Han et al., 2018), in this study, we focused on its orthologous target in P. knowlesi. Sequence alignment of the MSP1P-19 protein showed the conserved 12 cysteine residues across the species, 10 of which are similar in PkMSP1-19 ( Figure 1B ), suggesting that the function of PkMSP1P-19 may be conserved. To address this hypothesis, the recombinant PkMSP1P-19, along with PkMSP1-19 as a control protein, was expressed and purified in soluble form under non-denaturing conditions in bacterial protein expression system, and the purity was assessed by SDS-PAGE. Both proteins migrated as expected molecular weights of approximately 37.2 kDa and 35.3 kDa, and GST protein was observed at 25 kDa ( Figure 1C , arrowhead). We failed to cleave the GST-tag from the proteins because of precipitation; thus, we generated immune sera from mice and rabbits of the recombinant proteins with the GST-tag. The polyclonal antibodies against PkMSP1P-19 and PkMSP1-19 reacted to their respective recombinant proteins, but no bands were observed with pre-immune sera ( Figure 1D ). However, mutual interaction of the antibody to both proteins was observed, suggesting that antibody cross-reactivity may occur (MSP1 and MSP1P in Figure 1D ; Supplementary Figure S1 ). MSP1P, like MSP1, is predicted to be attached to the plasma membrane of the parasite via a GPI anchor. The GPI anchor motif requires both an N-terminal signal peptide and also a C-termini GPI anchor motif. The addition of a tag to either the N or C termini would disrupt this signal and remove its GPI anchor. So while tagging of the gene would have additionally helped to understand MSP1P localization, it was not possible without a more complex, internal tagging approach.

An indirect immunofluorescence assay was carried out to determine the localization of the PkMSP1-19 and PkMSP1P-19 in the asexual blood stage of P. knowlesi parasites. The fluorescence signal of native proteins was observed, indicating that both proteins are expressed at the mature schizont stage, surrounding the periphery of developing merozoites ( Figure 1E ; Supplementary Figure S2 ). The PkMSP1P-19 antibody gave signal that was similar to that seen for the established merozoite surface marker, MSP1-19; furthermore, free merozoites after rupture were surrounded with the signal. While this suggested that PkMSP1P-19 is present as a merozoite surface protein before and after rupture, the potential cross-reactivity with MSP1-19 meant that no firm conclusions could be drawn.

Plasmodium knowlesi proteins involved in erythrocyte invasion are critical for determining host cell specificity, and several have distinct binding characteristics between macaque and human erythrocytes. For example, whereas P. knowlesi PkDBP-α interacts with both macaque and human erythrocytes, its paralogues, PkDBP-β and PkDBP-γ, only bind macaque erythrocytes (Adams et al., 1992). For those reasons, the binding of PkMSP1P-19 to rhesus macaque and human erythrocytes was investigated using a FACS-based binding assay system. In a previous study, PvMSP1P-19 showed the binding activity to human reticulocytes specifically (Han et al., 2018), so we enriched reticulocytes from human umbilical cord blood and used for reticulocyte binding assay. In the result, PkMSP1P-19 significantly bound to human reticulocytes (8.1% ± 2.8%) but neither macaque erythrocytes nor human erythrocytes. Due to the resource limitation of macaque erythrocytes for the reticulocyte enrichment, it was impossible to specifically test purified macaque reticulocytes. In contrast to PkMSP1P-19, no significant binding activity of PkMSP1-19 to erythrocytes was observed. PkDBPα-RII, a well-known binding partner of the Duffy Antigen Receptor for Chemokines (DARC) on RBCs, showed strong binding activity with human reticulocytes (40.2% ± 8.6%) and erythrocytes (2.3% ± 0.4%) as well as rhesus macaque erythrocytes (46.0% ± 4.6%) ( Figure 2A ), consistent with a previous report (Chitnis and Miller, 1994). We observed a gradual increase in the binding activity of PkMSP1P-19 to human reticulocytes in a concentration-dependent manner ( Figure 2B ), and the binding was significantly inhibited by the purified IgG against PkMSP1P-19 ( Figure 2C ). These results suggest that PkMSP1P-19 binds to human reticulocytes and, therefore, could play a role specifically during invasion. However, in order to mitigate the limitations caused by antibody cross-reactivity, we decided to study the function of PkMSP1P in human erythrocyte invasion by CRISPR-Cas9 genome editing.

To validate the role of PkMSP1P during the invasion process, gene deletion was carried out using the CRISPR/Cas9 system (Mohring et al., 2019). PCR-based donor DNA to knockout pkmsp1p gene was co-transfected with pCas9/sg, which provides spCas9 nuclease, sgRNA, and a positive selection marker ( Figure 3A ). The transfected parasites were cloned out by limiting dilution, and the pkmsp1p gene disruption was confirmed by genotyping ( Figure 3B ). This was further confirmed via transcriptional analysis of the pkmsp1p gene mRNA from wild-type (P. knowlesi A1-H.1) and knockout (pkmsp1p-KO) parasites, revealing loss of the pkmsp1p transcript in KO lines ( Figure 3C ).

We evaluated the parasitemia of pkmsp1p-KO parasites compared with the wild type (WT) over 4 days to determine whether the absence of PkMSP1P impacts growth in human erythrocytes. The initial parasitemia of both parasite lines was diluted to be less than 0.1%, and that of the WT parasite line reached up to approximately 6% after 4 days as expected, but the pkmsp1p-KO line was just around 1% ( Figure 3D ; Supplementary Figure S3 ), and a comparison of the average fold multiplication rate between WT and KO parasite indicates that growth of the transfected parasite was significantly impaired due to pkmsp1p gene disruption with WT averaging fivefold growth and KO only twofold ( Figure 3E ). Furthermore, an “invasion assay” was conducted to specifically determine whether pkmsp1p gene disruption affects the conversion of schizonts to the new ring stage in infected human erythrocytes. These results showed that the relative invasion rate of the pkmsp1p-KO line was significantly decreased, at around half of the invasion efficiency observed for the WT parasites ( Figure 3F ). These results demonstrated that PkMSP1P is important in one or more of the key processes during the schizont to ring stage transition, which may include steps such as egress, invasion, or early ring development.

Given that the deletion of pkmsp1p led to a significant growth defect of P. knowlesi merozoites in human RBCs, we next aimed to better validate the specificity of our vaccine-induced polyclonal antibodies against PkMSP1P-19 by testing them on P. knowlesi A1-H.1 and pkmsp1p-KO lines. To verify that PkMSP1P-19 expression was absent in the KO lines, IFA and immunoblotting with native antigens were performed. It was immediately clear that despite the loss of the MSP1P protein, the signal for the PkMSP1P-19 antibody remained—demonstrating apparent evidence of background cross-reactivity of the antibody ( Figure 4A ; Supplementary Figure S4 ). Despite this, a clear quantitative change in signal was observed. The IFA result was compared using the pixel signal strength of PkMSP1-19 (Red) with PkMSP1P-19 (Green) by two-dimensional scatter data. It showed a high overlap of green and red pixel strength (R ^{2 =} 0.97) in the wild type, in clear contrast to the pkmsp1p-KO parasite that showed a lower specific signal with a correlation R ^{2 =} 0.67 ( Figure 4A ). The subcellular localization assay of pkmsp1p-KO with anti-PkMSP1P-19 did not disappear but only decreased the signal compared with WT. We then determined whether antibodies raised against PkMSP1-19 and PkMSP1P-19 can specifically recognize their antigens using WT (P. knowlesi A1-H.1) and KO (pkmsp1p-KO) schizont-stage parasite lysates; a normal rabbit antibody was used as a negative control ( Figure 4B ; Supplementary Figure S5 ). Both antibodies detected several bands, suggesting that posttranslational modification or proteolytic process of the MSP family commonly occurs in Plasmodium proteins (Blackman et al., 1991; Crosnier et al., 2013). The signal strength of the knocked-out strain is weaker than WT in both-antigen detection due to the expression level in the schizont stage being affected after the gene ablation. Although mRNA and DNA level data confirm that we succeeded in knocking out the pkmsp1p gene, antibody specificity assays supported the hypothesis that anti-PkMSP1-19 and anti-PkMSP1P-19 antibodies have clear cross-reactivity.

The inhibitory effect of those antibodies during the invasion process into human erythrocytes was evaluated by FACS using PkA1-H.1 and pkmsp1p-KO strains [the FACS gating as described previously (Muh et al., 2018b)]. Invasion inhibition of rabbit polyclonal IgGs against PkDBPα-RII, PkMSP1-19, and PkMSP1P-19, as well as monoclonal IgG against human DARC (2C3), was investigated over one replication cycle (24 h). An anti-Fy6 (α-2C3) was used for a positive inhibitory control for the assay with significantly hindered WT and KO line invasion with 94.6% ± 4.2% and 94.2% ± 4.4% [mean ± SEM], respectively. Following, the antibody raised from PkDBPα-RII also showed a high invasion inhibition effect with 62.9% ± 3.3% and 60.9% ± 5.1%, respectively. Comparing pre-immune (PI) sera (13.4% ± 2.9% and 7.3% ± 2.0%) and anti-GST (7.0% ± 2.1% and 2.9% ± 3.8%) invasion inhibition effect with anti-PkMSP1-19 (24.3% ± 2.1% and 15.5% ± 1.6%) and anti-PkMSP1P-19 (18.9% ± 6.0% and 10.7% ± 3.7%), there was no significant difference, except anti-PkMSP1-19 (24.3% ± 2.1%) with anti-GST (7.0% ± 2.1%) observed in WT strain ( Figure 4C ). In conclusion, anti-PkMSP1-19 and anti-PkMSP1P-19 have relatively low inhibitory effects, suggesting that even though PkMSP1P is essential for parasite blood-stage development, the polyclonal antibody specifically targeting PkMSP1P-19 may lack key inhibitory epitopes to block the parasite invasion to human host erythrocytes.