Analysis of blood-induced Anopheles gambiae midgut proteins and sexual stage Plasmodium falciparum interaction reveals mosquito genes important for malaria transmission

Plasmodium invasion of mosquito midguts is a mandatory step for malaria transmission. The roles of mosquito midgut proteins and parasite interaction during malaria transmission are not clear. This study aims to identify mosquito midgut proteins that interact with and affect P. falciparum invasion. Based on gene expression profiles and protein sequences, 76 mosquito secretory proteins that are highly expressed in midguts and up-regulated by blood meals were chosen for analysis. About 61 candidate genes were successfully cloned from Anopheles gambiae and expressed in insect cells. ELISA analysis showed that 25 of the insect cell-expressed recombinant mosquito proteins interacted with the P. falciparum-infected cell lysates. Indirect immunofluorescence assays confirmed 17 of them interacted with sexual stage parasites significantly stronger than asexual stage parasites. Knockdown assays found that seven candidate genes significantly changed mosquitoes' susceptibility to P. falciparum. Four of them (AGAP006268, AGAP002848, AGAP006972, and AGAP002851) played a protective function against parasite invasion, and the other three (AGAP008138, FREP1, and HPX15) facilitated P. falciparum transmission to mosquitoes. Notably, AGAP008138 is a unique gene that only exists in Anopheline mosquitoes. These gene products are ideal targets to block malaria transmission.

www.nature.com/scientificreports/ cells using a serum-free medium to produce the recombinant proteins. FREP1 and HPX15 were expressed in the same system as internal controls. Seven genes, including AnAPN1, were not cloned due to incorrect annotation or other reasons. Eight genes were removed from further study because of their low expression. The remaining sixty-one genes ( Fig. 2A) were expressed at the ELISA-detectable level (Fig. 2B). About 23% of the candidate proteins were highly expressed (> 50 μg/10 6 cells), about 30% (18/61) were moderately expressed (> 10 μg/10 6 cells), and the rest were expressed at a detectable level (Fig. 2B). Each of the recombinant protein had a 6 × His tag at its C-terminal that would be probed with anti-His monoclonal Ab.
With these 61 insect cell-expressed recombinant proteins, we examined their interaction with P. falciparum. The P. falciparum-infected cell culture contains cells in various stages: uninfected, asexual stage, and sexual stage parasites (gametocytes and ookinetes). ELISA assays are more sensitive and easier than other methods. Thus, it was used as the screening approach to find candidate proteins that interacted with P. falciparum. The 15-day cultured P. falciparum that contained more than 10% gametocytes were collected by centrifugation and lysed. The lysate was used to coat the plates, followed by blocking with BSA. The recombinant proteins were diluted, and 0.2 pmol of each recombinant protein was added to each well ( Fig. 2A). The retained recombinant proteins in wells were detected by anti-His monoclonal Ab. Recombinant FREP1 (Fig. 2C, A8) and heat-inactivated FREP1 (Fig. 2C, F8) were used as the positive and controls, respectively. The results (Table S2) showed that A 405 of the 25 proteins were over triple of the negative control (Fig. 2C, Table 1). Out of these 25 proteins, 11 proteins (44%, including FREP1) showed > tenfold of binding signals of the negative control, and 9 proteins (36%) showed medium binding signs (≥ fivefold of the negative control;). It is worth noting that the Histidine tag does not bind   003781 005416 009101 005686 008176 001183 003010 FREP1  012263 006425 010820 007262 004900 001508 003499 002851  009102 007197 003573 001245 013078 011006 011371 001193  006268 001956 HPX15 013377 011379 008138 011378 003926  001203 002848 004054 005093 006972 005899 007924 004918  001005 001873 006400 011442 001509 007077 Figure 2. Cloning and expression of the selected midgut proteins and their interaction with P. falciparuminfected cell lysates by ELISA assays. In total, sixty-one genes were expressed at the detected level. (A) Candidate gene IDs. The access number is AGAP(ID)-PA. (B) Heterologous expression of the midgut genes with a baculovirus expression system. The color intensity highlights a different expression level for each gene. The concentration of each expressed recombinant protein in each well in panel (B) corresponds to the genes at the same position in panel A. (C) P. falciparum-infected cell lysate was used to coat ELISA plates. The same amount of insect-expressed proteins was added to analyze their interaction with the lysate by standard ELISA. A 405 values were measured. The signals in panel C correspond to proteins at the same positions in panel A, plus a negative control (heat-inactivated FREP1) at the well of F8. The intensity of the binding signal was represented by the corresponding color: the dark brown color indicated no binding signal was detected, and the yellow color indicated the most substantial binding signal. The binding threshold was that A 405 of the candidate protein was ≥ threefold over the negative control (heat-inactivated FREP1). www.nature.com/scientificreports/ to P. falciparum lysate based on our previous publication 9 . Indeed, the heat-inactivated 6xHis FREP1 did not show any interaction signal (Fig. 2C, F8). This experiment was conducted twice, and the results were consistent. The function of the 25 P. falciparum-binding proteins was summarized in Table 1. Ten out of 25 parasitebinding proteins are involved in digestion, including five serine-type peptidases (AGAP004900, AGAP008176, AGAP005686, AGAP001245, and AGAP011920). Ten out of 25 parasite-binding proteins were related to protection, including four niemann-Pick Type C-2 (NPC-2) proteins and three fibrinogen-related proteins (AGAP011371, FREP1, and AGAP004918). Five of 25 parasite-binding proteins were unknown. functional analyses of the candidate An. gambiae midgut genes on P. falciparum infection in mosquito midguts by RnAi. For the P. falciparum-binding midgut proteins, we examined their effects on P. falciparum infection in mosquitoes. Because FREP1 12 and HPX15 18 were previously examined with knockdown assays, we focused on the remaining 23 candidate genes. Gene-specific dsRNA was injected into the mosquito hemolymph to knock down the expression of a candidate gene. Then we infected the treated mosquitoes with the cultured P. falciparum. Seven days after the infection, mosquitoes were dissected to count the oocysts in the mosquito midguts. GFP dsRNA was used as a control. The number of oocysts in each mosquito of the experimental group was compared to that of the control group by a Mann-Whitney-Wilcoxon test. The results are summarized in Table 2 and shown in Fig. 3A. Out of 23 genes, five (AGAP006268, AGAP002848, AGAP006972, AGAP008138, and AGAP002851) showed significant effects on P. falciparum transmission to mosquitoes, e.g., the number of oocysts was significantly different between the experimental group and the control group (p < 0.05, significance score > 3).
If injecting a specific dsRNA into a mosquito did not change the mosquito phenotypes of infection for some reason, such as low knockdown efficiency, we would not analyze it in this report. To determine the knockdown efficiency, we examined the expression level of the candidate genes in the treated mosquitoes using quantitative RT-PCR 48 h after injecting dsRNA. The results confirmed that the expression of five candidate genes was down-regulated by gene-specific dsRNA (Fig. 3B). The expression of AGAP008138 is about 40% of the control, while only 5% of AGAP002848 mRNA was left after dsRNA injection compared to the control. Notably, knocking down either one of AGAP006268 (peritrophin), AGAP002848 (NPC-2), AGAP006972 (keratin-associated protein 16-1), or AGAP002851 (NPC-2) significantly (p < 0.05) increased the number of P. falciparum oocysts in mosquito midguts compared to the control with 1.6-folds, 2.5-fold, 4.5-fold, and 1.9-fold changes, respectively (Fig. 3C), indicating that these four proteins inhibited malaria infection in mosquitoes. However, knockdown www.nature.com/scientificreports/ of AGAP008138 (uncharacterized) expression significantly decreased the number of P. falciparum oocysts in An. gambiae, from 9.7 to 3.9 per midgut ( Fig. 3C), suggesting that the AGAP008138 gene product facilitated Plasmodium invasion in mosquitoes. The RNAi experiments of the five candidate genes were conducted at least twice, and the results were reproducible (Fig. 3C).

Confirmation of the insect cell-expressed candidate proteins by western blotting and their
interaction with the sexual stage P. falciparum. First, we confirmed a set of insect cell-expressed candidate proteins, including the five P. falciparum-transmission related proteins, which were specifically recognized by anti-His antibody using Western blotting assays. Samples containing recombinant proteins were separated by 12% SDS-PAGE and analyzed by western blotting assays with anti-His antibodies. Seven recombinant proteins, including the five recombinant proteins, AGAP006972, AGAP008138, AGAP002581, AGAP002848, and AGAP006268, were specifically detected by anti-His monoclonal antibodies (Fig. 4, lanes 1, 3, 4, 6, and 7, respectively). The size of AGAP006972 protein with the signal peptide was predicted about 15 kDa, similar to the observed band (Fig. 4, lane 1). The molecular masses of the mature proteins of AGAP008138 and AGAP002581 were expected to be 55 kDa and 16 kDa, matching the observed bands (Fig. 5, lanes 3 and 4, respectively). The predicted sizes of the precursor and mature AGAP002428 are 18 kDa and 16 kDa, respectively, matching the observed two bands (Fig. 4, lane 6). The size of the AGAP006268 product with the signal peptide was expected to be 12 kDa, matching the observed band (Fig. 4, lane 7). The data supported that the recombinant proteins were expressed by the insect cells in full length with or without signal peptides. This data also demonstrated that the anti-His monoclonal antibody is specific, and the ELISA signals in Fig. 2C corresponded to the bound recombinant proteins. Next, an indirect fluorescence assay (IFA) was used to examine the interaction between a candidate protein and sexual stage P. falciparum parasites, because only sexual stage parasites are relevant to parasite infection in mosquito midguts. Following our previous published methods to generate ookinetes 9,19 , we collected the latestage cultured P. falciparum cells and suspended them in an ookinete culture medium and incubated for 16 h at RT. The cell mixture, containing uninfected cells, asexual stage (trophozoites and schizont), and sexual stage parasites (gametocytes and ookinetes), were deposited onto glass coverslips and fixed with 4% paraformaldehyde. These cells were then incubated sequentially with the insect cell-expressed candidate protein and detected by fluorenes-labeled antibodies. The unrelated insect cell-expressed protein chloramphenicol acetyltransferase (CAT) with 6 × His tag at its C-terminal was used as a negative control. We also added anti-His antibody directly Table 2. The data of dsRNA-mediated knockdowns of the tested midgut gene mRNA on Plasmodium falciparum infection in mosquitoes. The access number of each gene was obtained from Vectorbase.org and followed the format of AGAP(ID)-PA. N: the number of mosquitoes for each treatment. Mean: The average number of oocysts per mosquito; SD: standard deviation; P values were calculated using the Mann-Whitney-Wilcoxon test. Significance score equals to -ln(p).  www.nature.com/scientificreports/ immediately after blocking to confirm the anti-His antibodies did not bind to cells as a negative control. Under a fluorescent microscope, the bound recombinant proteins that were recognized by the anti-His Ab showed a red color, and cell nuclei of parasites appeared blue. Sexual stage and asexual stage parasites were different in shape.
The results (Fig S1) showed 24 candidate proteins bound to sexual and asexual P. falciparum at various levels.
After subtracting the background pixel intensity, we found that all of the ELISA-positive proteins interacted with P. falciparum-infected cells (Fig S1A). Out of 24 proteins (FREP1 was excluded), 17 interacted with sexual parasites significantly stronger than asexual stage parasites (Fig S1C). For the five candidate proteins that are related to parasite transmission (Fig. 5A), we found that they significantly bound more to sexual P. falciparum parasites than to asexual stage parasites (Fig. 5B, p < 0.05). In addition, the interaction signals to sexual stage parasites were also significantly stronger than the negative control (Fig. 5C, p < 0.05). It is worth noting that both gametocytes and ookinetes are relevant to malaria transmission. Therefore, we did not distinguish one over the other at this time. Together, our Western blotting data and IFA demonstrated that the five candidate proteins specifically interacted with Plasmodium sexual stage parasites (gametocytes and/or ookinetes).

Discussion
It is essential for Plasmodium parasites to infect mosquitoes to complete a malaria transmission cycle. After a blood meal, mosquito midgut epithelial cells secrete materials such as proteins, chitin to form PM, which wraps blood bolus inside. Thus, parasites must overcome the PM first to infect mosquitoes. Only the molecules inside or within PM can contact large molecules such as Ab in the host blood 9 . To identify some midgut proteins that are important for malaria transmission, we focused on proteins that were highly expressed and secreted in midguts and up-regulated three hours post blood meal. We selected these genes as our candidates for further analysis. Consistent with our expectations, functional analysis in our results revealed that most of these candidate proteins were involved in PM formation, microbe recognition, and digestion. We successfully PCR-cloned and insect cell-expressed > 80% of the candidates. Nearly half (41%) of candidate proteins bound to the cultured P. falciparum parasite lysate. The detailed analysis found two known proteins, FREP1 and HPX15, bound to parasites, which was consistent with the previous report 9, 16 . They participated in Plasmodium transmission to mosquitoes. Seven parasite-binding proteins were peptidases or proteases that should bind to the substrate proteins. Four NPC-2 proteins bound to P. falciparum lysates. Two NPC-2 proteins (AGAP002848 and AGAP002851) were immunoglobulin-like secreted proteins, containing MD-2-related lipidrecognition (ML) domain. ML domain is conserved from fungi to plants and animals 20 . NPC-2 proteins play important roles in lipid metabolism and immune signaling pathway 21 . NPC-2 proteins were named because its identification is related to Niemann-Pick disease type C2. In invertebrates, evidence shows that NPC-2 proteins may be involved in innate immune signaling pathways, such as the immune deficiency (Imd) pathway. Three Drosophila NPC2 proteins can bind the bacterial cell wall components of peptidoglycan (PG) and lipoteichoic acid (LTA) to serve as a pathogen recognition receptor (PRR) to initiate the innate immune responses [22][23][24] . The 25 proteins that bound to the Plasmodium-infected cell lysate also included six unknown proteins.
Indeed, all the 25 candidate proteins were confirmed to bind to P. falciparum-infected cells by IFA assays. About 68% of these proteins showed significantly higher interaction signals with sexual stage parasites than asexual stage parasites. Since both gametocytes and ookinetes are related to malaria transmission, we did not distinguish them in this report. Among the 25 parasite-binding proteins, 28% (n = 7) significantly affected parasite transmission when the gene expression was knocked down. Two (HPX15 and FREP1) were previously known, and five genes had not been reported as related to Plasmodium transmission. By silencing any of AGAP006268 (peritrophin), AGAP006972, AGAP002848 (NPC-2 related), or AGAP002851 (NPC-2 related), the number of P. falciparum oocysts in mosquito midguts significantly increased compared to that of the control, suggesting they protected mosquitoes from infection. Specifically, AGAP006268 is a structural protein that is secreted by Anopheles midgut epithelial cells after a blood meal to form PM. It contains a chitin-binding domain at its N-terminus 25 , which allows peritrophin to bind to chitin fibers and form the structural basis of PM 26 . Reducing  www.nature.com/scientificreports/ peritrophin expression might destroy the integrity of PM, and facilitates parasite penetration of PM. The function of AGAP006972 is unknown. There are several repeats in the 96-amino acid AGAP006972 peptide, including eight GGY repeats and three GGF. The two NPC-2 proteins (AGAP002848 and AGAP002851) may serve as a PRR to trigger the innate immune response. Different from the other four genes, but similar to FREP1, knocking down the expression of AGAP008138 significantly reduced the number of P. falciparum oocysts in the mosquito midguts. The mature AGAP008138 has 492 amino acids in length. However, AGAP008138 does not have any known motifs. Its expression was induced 7.5-fold 3 h post-blood meal and decreased close to or less than naive mosquitoes after 24 h. IFA showed that it bound to ookinetes. Notably, AGAP008138 is a unique gene, existing only in the Anopheles species and not in Aedes or Culex, which is consistent with the fact that only Anopheles mosquitoes transmit malaria. Therefore, AGAP008138 might be an ideal target for malaria transmission-blocking vaccine development, although the detailed mechanisms of these mosquito genes on parasite transmission need further investigation.
This study identified a set of new midgut proteins that interact with P. falciparum ookinetes and affect malaria transmission. Further investigation of these proteins will help to elucidate the molecular mechanisms of a Plasmodium invasion in mosquito midguts as well as the innate immunity in mosquitoes.

Materials and methods
Rear mosquitoes. An. gambiae (G3 strain) eggs were obtained from BEI Resources (https ://www.beire sourc es.org/). The insectary was set at 27 °C, 80% relative humidity, 12-day/night cycles. Larvae were fed with grounded fish food, and adult mosquitoes were maintained with 8% sucrose solution.
culture P. falciparum gametocytes and ookinetes. P. falciparum (NF54) was obtained from BEI Resources, and cultured with RPMI-1640 complete medium, containing 4% new O + human red blood cells, 10% human AB + serum, and 12.5 μg/mL of hypoxanthine in a candle-jar at 37 °C. The culture started at 0.1% parasitemia, and day 15-17 P. falciparum cultures were used to infect the mosquitoes or to be lysed, as described previously 9 . To prepare P. falciparum ookinetes, we transferred 5 mL of day-15 cultured P. falciparum containing ~ 2% stage V gametocytes into a 15 mL centrifuge tube and centrifuged at room temperature at 650 g for 5 min (m). The pellet was washed with RPMI-1640 three times and then resuspended in 500 µL of sterile ookinete culture medium (RPMI-1640, 20% human serum AB + , 50 µg/mL of hypoxanthine, 2 g/L NaHCO 3 ). The resuspended cells were transferred into a well of a 12-well plate and incubated at room temperature on a shaker (50 rpm) for 24 h to generate ookinetes. Finally, the cell mixtures of the ookinetes, gametocytes and the asexualstage P. falciparum were collected by centrifugation at 650 g for 5 m at RT. Selection of candidate An. gambiae midgut proteins. We focused on the proteins that may contact parasites directly in mosquito midgut lumen. PM starts to form inside of the midgut from secretory materials immediately after a blood meal, reaching its thickest 18 h post blood meal. Therefore, we chose the proteins that have signal peptides, and the corresponding genes are immediately up-regulated by a blood meal, return to regular expression at 24 h after a blood meal, and expressed higher in midguts than in other tissues. The candidate An. gambiae midgut proteins were selected from the databases of ReAnoXcel 13 , vectorbase 14 , and AgExpression 15 according to the following four criteria: (1) candidate proteins contain N-terminal signal peptides based on An. gambiae protein databases from previous annotation 13 and the vectorbase (https ://www.vecto rbase .org) 14 ; (2) the expression of candidate genes is > 1.2-fold higher 3 h post bloodmeal comparing to naïve mosquitoes; (3) the expression of candidate genes should return to the expression level of naïve mosquitoes 24 h post-bloodmeal; and (4) candidate genes are expressed > 1.2-fold higher in midguts than in other tissues. The gene expression data were obtained from the vectorbase database (https ://vecto rbase .org/downl oad/anoph elesgambi aeexp r-stats vb-2019-06txt gz) and AgExpression of a previous publication 15 . The protein functions were obtained from our previous annotation 13 and the uniprot database (https ://www.unipr ot.org). cloning and expression of candidate genes in insect cells. Total RNA was extracted from adult female mosquito midguts using RNAzol (Sigma-Aldrich, MO). About 20 midguts were dissected from threeday-old female An. gambiae and put into a 1.5 mL plastic tube that contained 200 μl of RNAzol. The tissue was grounded by a micro pestle (Sigma-Aldrich). Eight hundred µL RNAzol was added. After centrifugation (13,000 g for 10 m), RNA in the supernatant was precipitated by isopropanol (1:1 v/v). The cDNA was synthesized with the Superscript First-Strand Synthesis System (Invitrogen, CA). Gene fragments were amplified by PCR using the DNA Engine Dyad Thermal Cycler (Bio-Rad, CA) with gene-specific primers (Table S3). DNA fragments were purified through a GeneJet PCR Purification Kit (Thermo Fisher Scientific) and cloned into modified donor plasmid pFastBac1 that contains a 6 × His tag at the C-terminal of multiple cloning sites. Recombinant plasmids were transformed into competent DH5α cells. Positive recombinant plasmids were confirmed by sequencing and then transformed into DH10Bac to obtain recombinant Bacmids that showed white colonies on culture plates. The positive white colonies were picked and confirmed by PCR. One μg of a recombinant Bacmid in 100 μL un-supplemented grace medium (Invitrogen) was mixed with 100 μL of un-supplemented Gibico grace medium (Thermal Fisher Scientific), containing 5 μl of cellfection II (Thermal Fisher Sci), and incubated for 30 m at RT. The mixture was then added into 1.5 mL of unsupplemented grace medium that contained 800,000 sf9 cells in a well of a six-well culture plate and incubated for 4 h at 26 °C. The supernatant was replaced with 2 mL of complete race medium with 10% FBS (Invitrogen) and incubated at 26 ºC for three days to generate recombinant baculovirus viruses. A 100 μL culture supernatant that contained the recombinant virions were added in 2 mL of Express Five SFM medium (Gibco) supplemented with 20 mM of L-glutamine and 1 million High Five cells in a well of a six-well culture plate. The cells were incubated at 26 ºC for three days to express the Determination of the effects of An. gambiae genes on P. falciparum transmission to mosquitoes using dsRnA mediated gene expression silencing assays. The procedure was similar to the previous report 27 . Gene-specific primers that contained a T7 promoter and candidate gene sequences were designed with the E-RNAi web server 28 . Double-stranded RNAs (dsRNA) were synthesized with a MEGAscript Kit (Thermo Fisher Scientific, MA). GFP dsRNA was used as a control. About 67 nl of 3 µg/µL dsRNA was injected into 1-day old An. gambiae hemocoel using Nanoject II (Drummond Scientific, PA), and 100 mosquitoes were injected for each candidate gene. Thirty-six hours after treatment, the treated mosquitoes were fed with 0.2% P. falciparum gametocytaemia. The midguts of the infected mosquitoes were dissected seven days postinfection and stained with 0.1% mercury dibromofluorescein disodium in PBS. Oocysts were counted under a light microscope. The Mann-Whitney-Wilcoxon test was used to determine the statistical difference of the oocysts between the experimental group and the control. The transcript knockdown efficiency was confirmed by the quantitative RT-PCR in five of treated mosquitoes that were taken randomly 24 h after infection. GelDoc (UVP, Upland, CA) was used to take photos of the gels. PCR reactions (to detect specific genes and loading control S7) from one sample were run in the same gel. The bands from a particular gene and S7 were cropped for easier interpretation and smaller file sizes.

Detection of recombinant proteins with a monoclonal anti-His antibody with Western Blotting.
To maximize the yield of recombinant proteins, we collected the expressed proteins in both culture medium and cells. The culture cell medium was separated by centrifugation (650 g for 4 m). The supernatant was concentrated 100-fold by Centricon 10 Centrifugal Filter Device (Millipore, Danvers, MA). The cell pellets were lysed in native cell lysis buffer (Clontech). The cell lysis and concentrated medium were mixed and loaded on 12% SDS-PAGE and separated by electrophoresis until the Bromophenol blue reached the bottom. The proteins on the gel were then transferred on nitrocellulose membrane. The membrane was blocked with 2% BSA in PBST, followed by incubation in mouse monoclonal anti-His antibody (Sigma, 1:2000 dilution with blocking buffer) at RT for 1 h, goat anti-mouse IgG-alkaline phosphatase conjugate (Sigma, 1:10,000 dilution in blocking buffer ) for 1 h. The membrane was washed three times with PBST between incubations. Finally, nitroblue tetrazolium (Sigma-Aldrich, MO) was added to develop.
Indirect immunofluorescence assays to determine midgut protein-ookinete interactions. The cultured P. falciparum gametocytes with ookinetes were deposited onto premium cover glass slips to make blood smears. Before the smears were completely dry, the semi-dry smears were fixed in 4% paraformaldehyde in PBS for 30 m at RT to keep the cell membrane intact. Then, the cover glass slips were sequentially incubated in 1 mL of PBS that contained 10 mM glycine for 20 m and the blocking buffer for 1.5 h at RT. After blocking, cells on a coverslip were incubated with a candidate recombinant protein (10 μg/mL) in 200 μl of PBS containing 0.2% BSA (blocking buffer) for 1 h, followed by sequential incubation with 4 drops of enhancer (Alexa Fluor ® 594 Goat Anti-mouse SFX kit, Invitrogen) for 30 m, 200 μl of anti-His monoclonal Ab (1:2,500 dilution in the blocking buffer, 2 μg/mL) for 1 h, and 200 μl of secondary Ab (Alexa Fluor ® 594 Goat Anti-mouse SFX kit, Invitrogen; 1:1,000 dilution in the blocking buffer) for 30 m. Between each incubation, the slip was washed three times with 1 mL of blocking buffer for 3 m each. In the end, the coverslip was rinsed in distilled water for 20 s, and then coated with 4,6-diamidino-2-phenylindole (DAPI; Sigma-Aldrich). About 20 μl of VECTASHIELD anti-fade mounting media (Vector Laboratories, Burlingame, CA) was added onto the coverslip and mounted onto a slide.
After incubation for at least 2 h in the dark, the cells were examined under fluorescence microscopy (Nikon Eclipse Ti-S fluorescence microscope). An unrelated protein chloramphenicol acetyltransferase was used to replace a candidate protein as a negative control, as demonstrated previously 19 . We also added anti-His antibody directly immediately after blocking to confirm the anti-His antibodies did not bind to cells as a negative control.
To quantify the binding proteins, we measured the red pixel intensity values of the parasites at three randomly selected regions using Adobe Photoshop (version 2018, San Jose, CA). All values subtracted the background value. Then, we calculated the difference of the mean red pixel intensity values between sexual parasites and asexual parasites or the negative control and experimental groups.