Evaluation of Plasmodium vivax HAP2 as a transmission-blocking vaccine candidate

Abstract

Transmission-blocking vaccine (TBV) is a promising strategy to interfere with the transmission of malaria. To date, only limited TBV candidate antigens have been identified for Plasmodium vivax. HAP2 is a gamete membrane fusion protein, with homology to the class II viral fusion proteins. Herein we reported the characterization of the PvHAP2 for its potential as a TBV candidate for P. vivax. The HAP2/GCS1 domain of PvHAP2 was expressed in the baculovirus expression system and the recombinant protein was used to raise antibodies in rabbits. Indirect immunofluorescence assays showed that anti-PvHAP2 antibodies reacted only with the male gametocytes on blood smears. Direct membrane feeding assays were conducted using four field P. vivax isolates in Anopheles dirus. At a mean infection intensity of 72.4, 70.7, 51.3, and 15.6 oocysts/midgut with the control antibodies, anti-PvHAP2 antibodies significantly reduced the midgut oocyst intensity by 40.3, 44.4, 61.9, and 89.7%. Whereas the anti-PvHAP2 antibodies were not effective in reducing the infection prevalence at higher parasite exposure (51.3–72.4 oocysts/midgut in the control group), the anti-PvHAP2 antibodies reduced infection prevalence by 50% at a low challenge (15.6 oocysts/midgut). Multiple sequence alignment showed 100% identity among these Thai P. vivax isolates, suggesting that polymorphism may not be an impediment for the utilization of PvHAP2 as a TBV antigen. In conclusion, our results suggest that PvHAP2 could serve as a TBV candidate for P. vivax, and further optimization and evaluation are warranted.

1. Introduction

The HAP2/GCS1 protein, originally identified in Arabidopsis thaliana, and later in Lilium longflorum pollens [1–3], is present in a wide range of eukaryotic species, including plants, metazoans, invertebrate animals, algae, as well as pathogenic and non-pathogenic protists [4–9]. HAP2 has a highly conserved structure, with a single membrane-spanning domain, an extracellular domain and a relatively short cytoplasmic tail [6, 10]. Structure and functional studies on HAP2 in Chlamydomonas, Arabidopsis, and Tetrahymena indicate that it is a gamete membrane fusion protein, with homology to the class II viral fusion proteins [10–15]. A hydrophobic cd loop consisting of a ~40 amino acid (aa) conserved region in domain II plays an important role in the membrane fusion process [10]. HAP2 is present in the genomes of all malaria parasites. Functional studies in the rodent malaria parasite Plasmodium berghei showed that PbHAP2 disruption blocked fertilization of the gametes in mosquito midgut and subsequent transmission in mosquitoes [5, 16]. Consistent with the surface expression of PbHAP2 on male gametes, PbHAP2 is a male fertility factor, and is needed for membrane fusion of the male and female gametes [5, 16].

Plasmodium vivax is now the predominant Plasmodium parasite in Asia and the Americas, putting about 2.85 billion people at risk of infection worldwide [17]. P. vivax infections have raised considerable concerns, since the infection can cause severe and fatal pathologies, especially when associated with co-morbidities [18, 19]. Several biological features including the development a long-lasting, dormant liver stage (the hypnozoites) endow this parasite a much longer period of transmission and resilience to conventional control measures mostly designed for P. falciparum [20–22]. In addition, there is strong evidence that P. vivax has evolved resistance to the frontline treatment chloroquine/primaquine in many endemic areas [23, 24]. The increased dominance of P. vivax in P. falciparum co-endemic areas highlights the need for integrated innovative interventions targeting this parasite, including the development of effective vaccines [25, 26]. However, compared to P. falciparum, the development of vaccines against P. vivax is still in early pre-clinical stages and the identification of new candidate antigens is a high priority [27, 28]. The early and continuous production of gametocytes during P. vivax infections, required for transmission of the parasites through mosquitoes, suggests that transmission-blocking vaccine (TBV) is a promising strategy for the elimination of P. vivax.

TBVs target antigens in sexual and early sporogonic stages of the malaria parasite and midgut proteins of the mosquito vectors [29, 30]. To date, a number of TBV candidates have been investigated, but only a handful of antigens show clear transmission-reducing activity (TRA) [29–33]. TBV development has focused on a countable few antigens including the pre-fertilization antigens HAP2 [34, 35], Pfs230 [36] and Pfs48/45 [37, 38], the post-fertilization antigens Pfs25 [39] and Pfs28 [40], and the mosquito antigen AgAPN1 [41]. Although PbHAP2 is involved in gamete fusion, and is not needed for attachment of the male gamete to the female gamete, antibodies against PbHAP2 still showed strong in vivo and in vitro TRA [34]. Mouse antibodies against recombinant PfHAP2 expressed in the wheat germ cell-free system also showed strong TRA in a standard membrane feeding assay (SMFA) [35]. Recently, Angrisano et al. have shown that antibodies raised against the conserved HAP2 cd loop peptides exhibited potent TRA in both P. berghei and P. falciparum systems [42], further underlining that the class II fusion step in Plasmodium gamete fertilization is a viable target for TBV [43].

Here we report the characterization of the HAP2 ortholog in P. vivax, referred to as PvHAP2. We confirmed PvHAP2 expression and localization in P. vivax. The substantial reduction of the number of P. vivax oocysts in mosquitoes conferred by the anti-PvHAP2 antibodies suggests that PvHAP2 could serve as a potential TBV candidate for P. vivax elimination.

2. Materials and methods

2.1. In silico analysis of PvHAP2

To identify the HAP2 ortholog in the P. vivax genome, the protein sequence of PbHAP2 was used to BLASTP-search in PlasmoDB (http://plasmodb.org). Protein pattern and architecture were examined using the Simple Modular Architecture Research Tool (SMART, http://smart.embl-heidelberg.de). Multiple sequence alignment of HAP2 genes from Chlamydomonas reinhardtii (ABO29824.2), P. falciparum (XP_001347424.1), P. berghei and P. vivax was performed using ClustalW (https://www.genome.jp/tools-bin/clustalw).

2.2. Expression of PvHAP2 in baculovirus

A 229 aa fragment of PvHAP2 spanning aa 231–459 was expressed using the Bac-to-Bac® Baculovirus Expression System (Thermo Fisher Scientific, USA). The PvHAP2 fragment was codon-optimized for baculovirus expression in Spodoptera frugiperda (Sf9) cells and the DNA sequence was synthesized (GenScript Biotech Corp., China). The N-terminus of PvHAP2 was fused with a 6×His tag and a glutathione S-transferase (GST) tag. The synthesized gene was cloned into pFastBac1 at the EcoRI and HindIII sites, which was transformed into MAX Efficiency®DH10Bac™ cells (Life Technologies, USA) to obtain the recombinant Bacmid.

Sf9 cells were maintained in Sf-900 II SFM medium (Thermo Fisher Scientific). Transfection of the Sf9 cells with the recombinant Bacmid was performed using Cellfectin® II Reagent according to the manufacturer’s instructions. Transfected cells were incubated in Sf-900 II SFM for 5–7 days at 27°C and the supernatant was collected as the P1 virus stock, which was used to infect new batches of Sf9 cells to produce P2 viruses. For purification of the recombinant protein, a 1.5-L culture of Sf9 cells was infected with the P2 viruses at a multiplicity of infection of 5. Cells were incubated in Sf-900 II SFM for 72 h at 27°C before harvest. The expression of PvHAP2 in Sf9 cells was analyzed by SDS-PAGE and Western blotting. Cells were harvested and lysed in a lysis buffer (50 mM Tris, 300 mM NaCl, 1% NP-40, 5% Glycerol, pH 8.0). The insoluble fraction was resuspended in a binding buffer (50 mM Tris-HCl, 7 M GuHCl, 0.5 mM TCEP, pH 8.0) and incubated with Ni-NTA resin (Novagen) to capture the target protein. The purified protein was refolded by dialysis in Tris-buffered saline (50 mM Tris-HCl, 6 M Urea, 150 mM NaCl, pH 8.0), and then in phosphate-buffered saline (PBS, pH 7.4). The purified proteins were then passed through a Superdex 200 prep grade column (GE Healthcare Life Sciences), followed by sterilization through a 0.22 μm filter. The protein concentration was determined by the Bradford assay (Bio-Rad), and the protein was subjected to electrophoresis under reducing or non-reducing conditions [44] on 10% SDS-PAGE gels. Western blotting was performed using an anti-His monoclonal antibody (mAb) and anti-GST polyclonal antibodies (Genscript), respectively.

For negative control, the GST-tag protein was expressed from the empty vector pGEX-4T-1 (+) in Escherichia coli Rosetta-gami B (DE3) cells after induction with 0.5 mM isopropyl-D-1-thiogalactoside (Sigma) at 37°C for 5 h. The proteins were purified with the GST-tag Protein Purification Kit (Beyotime, China) and used for immunization.

2.3. Generation of anti-PvHAP2 polyclonal antibodies

To generate antisera against recombinant Pv-HAP2 and the GST tag, female New Zealand white rabbits were immunized subcutaneously with 250 μg of the proteins in Freund’s complete adjuvant, followed by two booster immunizations at three-week intervals with 250 μg of proteins each in Freund’s incomplete adjuvant. Antisera were collected 14 days after the last immunization. The polyclonal antibodies were purified from both pre-immune and immune sera using Protein A columns. Concentrations of anti-PvHAP2 and anti-GST antibodies were determined by using the BCA Protein Assay Kit (TaKaRa, Japan) and were at 13.4 and 11.7 μg/μl, respectively. The IgGs were adjusted to a concentration of 11.7 μg/μl using PBS for mosquito feeding. Animal procedures were conducted according to the guidelines of The Animal Usage Committee of China Medical University.

2.4. Enzyme-linked immunosorbent assay (ELISA).

To estimate the antibody titers of purified IgGs, ELISA was performed as previously described [45]. The microtiter plate was coated with purified recombinant proteins (5 μg/mL) in 0.05 M sodium carbonate buffer (pH 9.6) at 4°C overnight. Then the plate was washed with PBS-T (0.05% Tween-20 in PBS) three times and blocked with 1% bovine serum albumin (BSA, Sigma) at 37°C for 1 h. To each well of the plate, 100 μl of purified IgGs, serially diluted with 1% BSA in PBS from 1:1000 to 1:512000, were added and incubated at 37°C for 2 h. After three washes with PBS-T, 100 μl HRP-conjugated goat anti-rabbit secondary antibodies (Invitrogen, 1:5000) were added and incubated at 37°C for 2 h. After additional six washes, 100 μl of tetramethyl benzidine substrate solution (Amresco, USA) were added and incubated in the dark for 5 min. The reaction was stopped by adding 50 μL of 1 mM H₂SO₄ and the absorbance at 490 nm was read immediately. The value for the final dilution of the antisera was defined as that above the average value of the GST control antisera + 3 × standard deviation (cut-off value).

2.5. Sequencing of Pvhap2 genes in clinical isolates

Genomic DNA from dried filter-paper blood spots of P. vivax patients used for the membrane feeding assay was extracted by using a QIAamp DNA Blood Mini kit (Qiagen, Germany). The full-length Pvhap2 gene was amplified by nested PCR with the following primers: Pvhap2-primary F [nucleotide position (nt.) −129 to −110] 5’-GAGGCGTACATGCAGATAGA-3’ and Pvhap2-primary R (nt. 2723–2742) 5’-ACCCTGGTGGTGCGTAGAG-3’; Pvhap2-nested F (nt. −30 to −13) 5’-ACGTAGGAGGAGAAGGAG −3’and Pvhap2-nested R (nt. 2705–2722) 5’-ATTCGCCTGGAGGTTCTA-3’. The primary amplifications were performed in 10 μl containing 1× KOD-Plus-Neo buffer, 200 μM dNTPs, 1 mM MgSO₄, 250 nM of each primer, 0.2 units of KOD Plus neo polymerase (Toyobo, Japan), and 1 μl of genomic DNA as template. The reactions were run with an initial denaturing at 94°C for 2 min, followed by 30 cycles of 98°C for 10 s, 56°C for 30 s, and 68°C for 2 min, and a final extension at 68°C for 5 min. Nested PCR was performed in 30 μl containing 1× KOD-Plus-Neo buffer, 200 μM dNTPs, 1 mM MgSO₄, 250 nM of each primer, 0.6 units of KOD Plus neo polymerase, and 1 μl of primary PCR product as template. The cycling conditions for the nested PCR were the same as for the primary PCR. PCR fragments were analyzed by electrophoresis on a 1.2% agarose gel. DNA fragments encoding the PvHAP2 (aa 231–459) were sequenced using primers Pvhap2 F1 (nt. 653 to 677) 5’-TATGCTATTCTCCGAATCGGCCACC-3’and Pvhap2 R1 (nt. 1565–1589) 5’-CACCATTACGCATATTACTACTCCC-3’. The PCR products were cloned in the pMD18-T vector (Takara Biotechnology, China) for sequencing.

2.6. Indirect immunofluorescence assay (IFA)

To obtain parasite specimens for IFA, a P. vivax patient with gametocytes detected in a blood smear by microscopy was recruited in a malaria clinic located at the China-Myanmar border. After obtaining written informed consent, 5 ml of venous blood were drawn and immediately mixed with two volumes of suspended animation buffer (10 mM Tris, pH 7.4, 170 mM NaCl, 10 mM glucose) at 37°C to avoid exflagellation. After centrifugation, the supernatant was removed and the red blood cell pellet was resuspended in 10 ml of RPMI 1640 medium. Then the mixture was layered onto 47% Nycodenzs/RPMI 1640 and centrifuged at 500 × g for 25 min. The gametocyte-rich fraction at the grey interface was collected and washed three times with RPMI 1640. The isolated gametocytes were spotted onto multi-well slides and fixed with 4% paraformaldehyde and 0.0075% glutaraldehyde (Sigma). The slides were blocked with 3% BSA in PBS at 37°C for 30 min and incubated with rabbit polyclonal anti-PvHAP2 antibodies (1:500) at 37°C for 1 h. After washing with PBS, the slides were incubated with Alexa Fluor® 488 goat anti-rabbit IgG antibodies (1:500; Invitrogen) and the nuclear stain 4′,6-diamidino-2-phenylindole dihydrochloride (DAPI) at 37°C for 15 min. After washing with PBS, the slides were mounted with a ProLong antifade kit (Invitrogen) and images were obtained using an epifluorescence microscope (Olympus, Japan). IgGs isolated from anti-GST immune sera were used as the negative control for IFA.

2.7. Direct membrane feeding assay (DMFA)

Single-species infection of P. vivax was diagnosed by conventional microscopy. Four P. vivax-infected patients were recruited at the Tha Song Yang hospital, located at the Thai-Myanmar border. The parasite densities in the blood samples were estimated in Giemsa-stained thin films from 1 μl of blood. After obtaining informed consent, 10 ml of blood were collected from each patient. For DMFA, purified IgGs against PvHAP2 were diluted with heat-inactivated (complement minus) AB⁺ human serum from malaria naive Thai volunteers at the ratio of 1:1 in a total of 180 μl. Meanwhile, purified IgGs from a GST-immunized rabbit were diluted in the same manner. These diluted IgG samples were then mixed with erythrocytes from the patients (1:1, v/v) and the reconstituted blood samples were introduced into membrane feeders after incubation at 37°C for 15 min. For each sample, 100 starved Anopheles dirus mosquitoes from a laboratory colony were allowed to feed on the blood samples through the feeders for 30 min. After feeding, non-fed mosquitoes were removed and fully engorged mosquitoes (typically 60% of total mosquitoes used) were maintained at 27°C and 80% relative humidity on 10% sucrose. Twenty mosquitoes from each group were dissected on day seven and the numbers of oocysts developed on the mosquito midguts were counted under a light microscope after staining with 0.5% mercurochrome. Mann-Whitney U test was used to examine the differences in the number of oocysts per mosquito between the groups of mosquitoes fed with purified IgGs from the GST-immunized rabbit (control) and IgGs from the PvHAP2-immunized rabbit. The mosquito infection rate was calculated as the percentage of oocyst-positive mosquitoes out of the total mosquitoes dissected. Fisher’s exact test was used to compare the difference of the infection rate between control and PvHAP2-immunization groups. P values of less than 0.05 were considered statistically significant.

3. Results

3.1. Identification of PvHAP2

A putative PvHAP2 gene (PVX_094925), located on chromosome 8 at nt 651,355–654,032 (+) containing an intron, was identified in the P. vivax Sal-I genome (Fig. 1A). PvHAP2 encodes an 862 aa protein with a predicted molecular mass of 97.3 kDa. A 22 aa signal peptide and a 23 aa transmembrane helix at the C-terminus were predicted (Fig. 1B). Between the signal peptide and the transmembrane region, is the HAP2 ectodomain, which shows a similar three-dimensional fold as the class II viral fusion proteins [10]. From the functional studies of PbHAP2, the cd loop and PFAM 10699 motif in domain II are possibly two important functional regions of the HAP2 ectodomain [10, 46]. The ~40 aa long cd loop is essential for the insertion of the HAP2 ectodomain into lipid bilayers. The cd loop is highly variable in the cysteine-rich region among four out of the five eukaryotic kingdoms [6]. Compared with the cysteine-rich region from C. reinhardtii HAP2 (CrHAP2), thirteen residues were conserved among Plasmodium spp., including five cysteines participating in disulfide bonds, a glutamate and an arginine making a salt-bridge in three-dimensional fold of the HAP2 ectodomain [10] (Fig. 1C). The PFAM 10699 motif is the HAP2/GSC1 domain, a ~50 aa domain with several conserved residues from plants to lower eukaryotes. Multiple sequence alignment with the 53 aa PFAM 10699 motif from CrHAP2 revealed that the Plasmodium PFAM 10699 motif was 54 aa long with a glycine insertion (Fig. 1D).

Fig. 1. Pvhap2 gene.

A. Localization of Pvhap2 (PVX_094925) on chromosome 8 in the P. vivax genome. B. Domain organization of PvHAP2. SP, signal peptide; DI, DII, DIII, the domain I-III; L, linker (L); TM, transmembrane segment; C-rich region, the cysteine-rich region; PFAM10699, the HAP2/GSC1 domain. C. Alignment of the cysteine-rich region of HAP2 proteins from Plasmodium spp. and Chlamydomonas reinhardtii (CrHAP2). Identical residues are shadowed in black, while conserved residues in Plasmodium spp. are shadowed in gray. Arrowheads above the conserved residues present the conserved salt bridge between R185 and E126 in CrHAP2. The cd loop involved in gamete fusion in P. berghei is indicated. D. Alignment of PFAM10699 domain of HAP2 proteins from Plasmodium spp. and CrHAP2. Arrowhead marks the absence of a residue in CrHAP2. The sequences used in C and D are from: C. reinhardtii, ABO29824; P. vivax, PVX_094925; P. falciparum, XP_001347424; P. ovale wallikeri, SBT34166; P. ovale curtisi, SBS82838; P. malariae, SBS84478; P. knowlesi, XP_002258781; P. berguei, EU369602; P. chabaudi, XP_739579; and P. yoelii, ETB58232.

3.2. Baculovirus expression of the PvHAP2 HAP2/GCS1 domain

The HAP2/GCS1 domain has been shown to be critical for the function of PbHAP2 [46]. We have thus expressed a 229 aa fragment of PvHAP2 spanning aa 231–459, which corresponds to one domain I with seven cysteines but without the cd loop (Fig. 1). Sf9 cells were infected by the recombinant baculovirus expressing the 229 aa PvHAP2 fragment, and cells were harvested at 72 h post-infection. Although a GST tag was fused at the N-terminus of the recombinant PvHAP2, the protein was found mainly in the insoluble fraction. Thus, the recombinant PvHAP2 was purified under denaturing conditions by one-step affinity chromatography on Ni-NTA resin and then refolded. After refolding, the purified protein presented as a single prominent band. Densitometric analysis of the Coomassie-stained protein gel showed that the density of the PvHAP2 protein band was more than 85% of the density of the entire lane (not shown), suggestive of greater than 85% purity of the recombinant protein. It was confirmed under non-reducing and reducing conditions by SDS-PAGE and Western blotting using both anti-His tag and anti-GST antibodies (Fig. 2A,B). The GST control protein was confirmed by western blotting and SDS-PAGE, which showed a molecular mass of ~26 kDa.

Fig. 2. SDS-PAGE and Western blot analysis of recombinant PvHAP2.

A. SDS-PAGE analysis of purified PvHAP2. M, protein marker; 1, BSA; 2, PvHAP2 (reducing conditions); 3, PvHAP2 (native conditions); 4, GST (reducing conditions); 5, GST (native conditions). B. Western blot analysis of purified PvHAP2. M, protein marker; 1, PvHAP2 (reducing conditions); 2, PvHAP2 (native conditions); 3, PvHAP2 (native conditions); 4, PvHAP2 (reducing conditions). The blots were probed with an anti-His monoclonal antibody (left panel) and rabbit anti-GST polyclonal antibodies (right panel).

Purified recombinant proteins were used to immunize New Zealand rabbits to produce polyclonal antibodies for detection and TRA studies. ELISA with purified antibodies showed that rabbit developed a strong antibody response against recombinant PvHAP2. Antibody titers were significantly higher than those from GST control and reached a titer of 1:256000 after the third immunizations (Fig. 3). The results indicated that the recombinant PvHAP2 was highly immunogenic to induce specific antibody responses in rabbits.

Fig. 3. Antibody titers of purified anti-PvHAP2 IgG.

Female New Zealand rabbits were immunized with the rPvHAP2 protein, and total IgG titer in the purified antibodies was measured by ELISA. Error bars indicate mean ± standard deviation. ** P < 0.01 represents the difference between anti-GST and anti-PvHAP2 groups (Student’s t test).

3.3. Detection of PvHAP2

To determine the expression and localization of PvHAP2, IFA was performed using the gametocyte-enriched fraction of a clinical P. vivax isolate. Whereas the control IgGs did not show any reactivity with the gametocytes, the anti-PvHAP2 antibodies only reacted with the male gametocytes (Fig. 4).

Fig. 4. Detection of PvHAP2 in gametocytes.

Gametocytes were fixed and stained with the PvHAP2 antibodies (1:200) and then with Alexa Fluor® 488-conjugated goat anti-rabbit IgG antibodies (green). Anti-GST antibodies used as negative control. Nuclei were stained by DAPI (blue). BF, bright field; AF488, Alexa Fluor® 488. The filled and open arrows point to male and female gametocytes, repsectively.

3.4. TRA of antibodies against PvHAP2

To evaluate whether rabbit antibodies against PvHAP2 could reduce P. vivax transmission to the mosquitoes, DMFA was carried out using laboratory-reared An. dirus mosquitoes and P. vivax isolates collected from four P. vivax patients. The densities of male and female gametocytes in the four vivax blood samples were 9–25 and 21–48 parasites/μl, respectively. The DFMA results showed that the anti-PvHAP2 antibodies possessed modest but significant TRA as compared to the control group. DFMA from three blood samples produced moderate levels of infection in the control group of mosquitoes fed on blood mixed with the anti-GST antibodies, with a mean midgut oocyst intensity of 72.4, 70.7, 51.3, respectively, while mosquitoes fed on the blood from patient 4 had a much lower infection intensity with 15.6 oocysts/midgut (Table 1). Compared with mosquitoes from the control group, the anti-PvHAP2 antibodies significantly reduced the midgut oocyst intensity by 40.3, 44.4, 61.9, and 89.7%. In addition, there was an inverse trend showing the higher level of parasite exposure (oocyst intensity in the control group), the lower level of the inhibition of infection intensity. Whereas the anti-PvHAP2 antibodies were not effective in reducing the infection prevalence at higher parasite exposure (control group oocyst density of 51.3–72.4), the anti-PvHAP2 antibodies reduced infection prevalence by 50% at a lower challenge (control group oocyst density of 15.6) (Fig. 5, Table 1).

Table 1.

Transmission-reducing activities (TRA) of the anti-PvHAP2 antibodies.

P. vivax isolates	Antibodies	Na	Oocyst number/midgut		% inhibition of oocyst densityd	P valuee	Infection rate (%)f	% inhibition of prevalenceg	P valueh
			Median (IQR)b	Mean±SDc
Case 1	GST	20	65.5 (56.0 – 92.3)	70.7 ± 31.1			100
	Pv HAP2	20	30.5 (9.3 – 77.8)	39.3 ± 34.2	44.4	0.0151	90	10.0	0.4872
Case 2	GST	20	77.5 (57.0 – 85.3)	72.4 ± 16.1			100
	Pv HAP2	20	45.5 (9.5 – 69.5)	43.2 ± 32.6	40.3	0.0014	90	10.0	0.4872
Case 3	GST	20	53.0 (28.3 – 74.8)	51.3 ± 28.4			100
	Pv HAP2	20	16.5 (6.3 – 33.0)	19.6 ± 16.6	61.9	0.0004	90	10.0	0.4872
Case 4	GST	20	16.5 (3.0 – 25.8)	15.6 ± 11.6			90
	PvHAP2	20	0 (0.0 – 1.8)	1.6 ± 3.5	89.7	0.0001	45	50.0	0.0057

^aN, the number of mosquitoes dissected in each group.

^bIQR; inter-quartile range.

^cSD; Standard deviation

^d% inhibition of oocyst density was calculated as (mean _control – mean _test)/mean _control × 100%.

^eFor comparing the TRA between the two groups (anti-PvHAP2–1 antibodies and anti-GST antibodies), the median number of oocysts was used (Mann–Whitney U test).

^fThe infection prevalence was calculated as the number of oocyst-positive mosquitoes per 20 mosquitoes dissected in each group.

^g% inhibition of prevalence was calculated as (% prevalence _GST – % prevalence PvHAP2)/% prevalence _GST× 100%

^hDifference in TRA between the two groups (anti-PvHAP2 antibodies and the anti-GST antibodies) was compared by Fisher’s exact test.

Fig. 5. TRA of anti-PvHAP2 antibodies.

DMFA were carried out using four P. vivax isolates with rabbit IgGs mixed with heat-inactivated (complement minus) AB⁺ human serum in the ratio of 1:1. Numbers of oocysts formed in mosquito midguts are plotted as scattered dots. The long horizontal bar designates the median number of oocysts in each group, and the two short horizontal bars designate the interquartile ranges. Statistical difference of the median number of oocysts between the recombinant PvHAP2-immuned group and GST-immunized control group was analyzed by the Mann–Whitney U test. *, **, ***, and **** denote p < 0.05, 0.01, 0.001, and 0.0001, respectively.

To examine whether the variation in TRA with different parasite isolates might be due to variation in pvhap2 gene sequence, the partial genes coding 231–459 aa of PvHAP2 in the four clinical P. vivax isolates used in DMFA were amplified and sequenced. Multiple sequence alignment showed 100% identity among these P. vivax isolates (Fig. S2).

4. Discussion

TBV development focuses on surface antigens of gametes and ookinetes. P25 and P28 are the lead targets for the ookinete stage, and antibodies raised in animals against heterologously expressed P25 or P28 completely inhibit parasite development in mosquitoes [47, 48]. In a recent clinical trial, however, even with EPA conjugation, Pfs25 elicited antibodies that were short-lived and required four doses to induce functional TRA by SMFA, but not by direct skin feeding [49] It has been long recognized that antibodies raised against gametes possessed strong TRA [50], despite the relatively short period of exposure of the gamete surface antigens to host antibodies (probably less than 1 h). In particular, molecules involved in the fertilization step such as P230 and P48/45 as well as the gamete fusion factor HAP2 are among the top TBV candidates expressed on the gamete surface. TBV development for P. vivax has lagged far behind of that for P. falciparum, largely due to the lack of a continuous P. vivax culture system; thus most DMFA are performed with fresh field isolates, which is logistically demanding. To date, only a few of the TBV candidates in P. vivax have been tested – and almost all such tests are based on the studies of orthologous proteins in P. falciparum. TBV studies for P. vivax have focused on Pvs25 and Pvs28 [51, 52], and to a much less extent on Pvs230 [53] and Pvs48/45 [54]. Here, our results demonstrated that PvHAP2 should be considered for inclusion in the TBV design for P. vivax.

HAP2 is a male-specific factor expressed in male gametes of sexually dimorphic species [5]. HAP2 has three domains, and domains II and III mediate membrane fusion. Further scrutiny identified a short ~40 aa peptide within domain II as the fusion loop (cd loop) [10]. It has been shown that polyclonal antisera raised against the recombinant PbHAP2 protein (aa 355–609) expressed in E.coli was able to inhibit P. berghei ookinete conversion in vitro and oocyst development in vivo [34]. Polyclonal antisera raised against the recombinant PfHAP2 (corresponding to the PbHAP2 fragment) expressed as a GST fusion protein in a wheat germ cell-free expression system also displayed high levels of TRA in SMFA [35]. Recently, the putative fusion loop was identified in the domain II of PbHAP2 and PfHAP2, and antibodies against a small peptide of the cd loop conjugated to a carrier protein displayed potent inhibition of ookinete conversion in vitro and transmission to mosquitoes in ex vivo SMFA [42]. Since HAP2 is a class II fusion protein, which is not involved in gamete binding [5], the blockade of ookinete formation by anti-HAP2 antisera may be due to the inhibition of the membrane fusion during the fertilization process. It is also possible that antibodies against HAP2 may prevent male-female gamete interactions by spatial hindrance. In the present study, a fragment of the PvHAP2 encompassing 231–459 aa was expressed in the baculovirus system and used to produce polyclonal antibodies in rabbits. The IgGs purified from the immunized rabbits recognized epitopes in male gametocytes of P. vivax, and showed significant TRA in DMFA as determined by oocyst counts. For the four clinical samples used for DMFA, regardless of the exposure level (reflected in the oocyst intensity in control group mosquitoes), the anti-PvHAP2 antibodies conferred significant, 40–90% reductions in oocyst density. This is consistent with the conclusions derived from analysis of SMFA data conducted on P. falciparum and P. berghei [55] However, a substantial reduction by 50% in infection prevalence was observed only with the lower level of parasite exposure with the control group infection intensity of 15.6 oocysts/midgut. There are multiple lines of evidence showing that the reduction in the number of infected mosquitoes depends on the parasite exposure, with it being much easier to show reductions in oocyst prevalence if oocyst density is low or intermediate in the control group [55, 56]. Thus, our data showing significant reductions in both infection prevalence (50%) and oocyst density (90%) in a case with lower parasite exposure are encouraging, given that wild caught mosquitoes with natural Plasmodium infections typically have very low oocyst intensity [57].

It is noteworthy that our current PvHAP2 fragment encompassing aa 231–459 within domain I and II does not include the cd loop region. Thus, further optimization of the PvHAP2-based vaccine is needed to determine the best combination of different epitopes. In addition, this study did not investigate the glycosylation status of the recombinant PvHAP2 produced in baculovirus. Since Plasmodium parasites lack most of the enzymes needed for N-glycosylation [58], it is highly desirable that non-glycosylated antigens are used for malaria vaccines. The wild-type SF9 cells used in our study usually produce very low levels of glycans and do not produce complex glycans [59]. Further, another TBV candidate, Pfs230C, produced in the baculovirus system had a low level of undesired glycosylation, but the glycosylation did not appear to impact induction of functional antibodies [60]. Regardless, future efforts should elucidate whether glycosylation in baculovirus-expressed PvHAP2 affects the antigenicity of the protein. Whereas this study showed strong immunogenicity of the recombinant PbHAP2 when immunized with the Freund’s adjuvant, adjuvants suitable for clinical studies need to be evaluated in future studies. In addition, conjugation of PvHAP2 to carrier proteins to enhance immunogenicity also needs to be explored.

Highlights.

• Plasmodium vivax HAP2 protein was expressed in baculovirus for immunization

• Transmission-reducing activity was evaluated with four P. vivax clinical isolates

• Anti-PvHAP2 antibodies (Ab) significantly reduced infection intensity in mosquitoes

• Anti-PvHAP2 Ab substantially reduced infection prevalence at low parasite exposure