Preliminary characterization of Plasmodium vivax sporozoite antigens as pre-erythrocytic vaccine candidates

Plasmodium vivax pre-erythrocytic (PE) vaccine research has lagged far behind efforts to develop Plasmodium falciparum vaccines. There is a critical gap in our knowledge of PE antigen targets that can induce functionally inhibitory neutralizing antibody responses. To overcome this gap and guide the selection of potential PE vaccine candidates, we considered key characteristics such as surface exposure, essentiality to infectivity and liver stage development, expression as recombinant proteins, and functional immunogenicity. Selected P. vivax sporozoite antigens were surface sporozoite protein 3 (SSP3), sporozoite microneme protein essential for cell traversal (SPECT1), sporozoite surface protein essential for liver-stage development (SPELD), and M2 domain of MAEBL. Sequence analysis revealed little variation occurred in putative B-cell and T-cell epitopes of the PE candidates. Each antigen was tested for expression as refolded recombinant proteins using an established bacterial expression platform and only SPELD failed. The successfully expressed antigens were immunogenic in vaccinated laboratory mice and were positively reactive with serum antibodies of P. vivax-exposed residents living in an endemic region in Thailand. Vaccine immune antisera were tested for reactivity to native sporozoite proteins and for their potential vaccine efficacy using an in vitro inhibition of liver stage development assay in primary human hepatocytes quantified on day 6 post-infection by high content imaging analysis. The anti-PE sera produced significant inhibition of P. vivax sporozoite invasion and liver stage development. This report provides an initial characterization of potential new PE candidates for a future P. vivax vaccine.


Introduction
Malaria is a major public health concern with approximately 247 million cases and 619,000 deaths occurring in 2021 in 84 countries worldwide [1].P. vivax infections account for a large portion of reported malaria cases in the Americas (71.5%),South-East Asia (39.7%),Western Pacific region (31.9%), and the Mediterranean Region (18%) [1].Clinically, P. vivax has been mischaracterized as a benign disease, since infections often cause intense tertian malaria fever or paroxysm, anemia, repeated relapse infections, and low birthweight [2,3].Moreover, some liver stages can pause development to become dormant hypnozoite stages that later can activate to complete development to blood stages leading to clinical relapse infections furthering disease morbidity and transmission.
Development of an effective pre-erythrocytic (PE) stage vaccine that targets the infecting sporozoites from the mosquito and development of liver stages offers the potential to neutralize a new infection of P. vivax sporozoites before any clinical illness or subsequent transmission to mosquitoes.In particular, the non-replicating short-lived sporozoites are directly exposed to antibodies during their migration to the liver while the resident liver stages have prolonged vulnerability to immune effector cells.Therefore, a PE vaccine can possibly prevent clinical disease, hypnozoite formation, and transmission [4,5].However, functional protective responses of P. vivax PE antigens are poorly characterized relative to P. falciparum [6].Currently, this is a major gap in our understanding of potential P. vivax vaccine candidates.
The Plasmodium circumsporozoite protein (CSP) is the dominant sporozoite surface protein and the long-time leading PE vaccine candidate based on early studies that linked immune responses to CSP to the induction of sterile immunity by irradiated P. berghei sporozoites [7][8][9][10].RTS,S/AS01 and R21/MM are virus-like nanoparticle vaccines containing components of the P. falciparum circumsporozoite protein (PfCSP) [11][12][13].When used seasonally in combination with chemoprevention, RTS,S reported efficacy is 70.6% against severe malaria while R21 is 81% effective against clinical malaria in children [13,14].However, similar progress toward a P. vivax CSP-based vaccine were not successful in initial clinical trials [15][16][17][18][19][20][21][22].Additional studies suggest a lack of long-lived protective responses of PvCSP may be due to poor immunogenicity [23,24].Thus, there is a need to identify other potential PE candidates for inclusion with PvCSP in a multivalent vaccine design to prevent infection and subsequent progression to clinical disease and mosquito transmission.
Ideal antibody-based PE vaccine candidates would be surface-exposed sporozoite antigens accessible to immune antibodies that could act synergistically with PvCSP or other antigens during the infection process.The ideal PE vaccine would inhibit gliding as the sporozoite migrates from the skin to the liver, block sporozoite cell traversal when it reaches the liver parenchyma, inhibit invasion of and development in hepatocytes, and block egress to initiate blood stage development.Previously, we identified numerous P. vivax PE antigens that were upregulated in response to mammalian host-like microenvironmental changes that enhanced sporozoite infectivity [25].These potential PE candidates that associated with infectivity were analyzed for other favorable properties, such as other evidence for essentiality to invasion and/ or liver stage development to select to a short list of four antigens to evaluate as potential PE vaccine candidates.
Sporozoite surface protein 3 (SSP3) was first characterized by Harupa et al in P. yoelii.As gliding plays an important role in sporozoite infectivity [26] it was found that PySSP3 − mutant sporozoites were gliding deficient [27].In a related study, the PbSSP3 − mutant sporozoites had normal gliding patterns, yet mutants were unable to cause a blood-stage (BS) infection [28].This discrepancy demonstrates that mutant SSP3 orthologs in rodent species of Plasmodium have varying phenotypes.A second vital PE antigen is the sporozoite surface protein essential for liver stage development (SPELD).In P. berghei, PbSPELD − mutants resulted in early PE developmental arrest, yet other invasive phenotypes such as gliding were unaffected [29].Another essential sporozoite microneme protein, MAEBL, is vital for sporozoite salivary gland invasion, and hepatocyte invasion [30][31][32][33].Antisera targeting the M2 domain of MAEBL demonstrated protection against inhibition of PE development and lethal P. yoelii infection [30,34].Lastly, we selected the sporozoite protein essential for cell traversal (SPECT1), which is another microneme protein that facilitates sporozoite transmigration across host cells [35][36][37].These antigens were chosen for likelihood of antibody accessibility, their important roles in sporozoite invasion, and subsequent liver stage development.Here, we partially characterize these novel P. vivax PE antigens providing preliminary results to support their potential as P. vivax PE vaccine targets.

Ethics statement
Protocols for human subjects were approved by the Ethics Committee on Human Rights Related to Human Experimentation, Mahidol University Thailand [MU-IRB 2012/079.2408].Written informed consent was obtained from all participants.BALB/c mice aged 4-6 weeks were maintained under pathogen-free conditions per Institutional Animal Care and Use Committee (IACUC) protocol R IS00007010 approved by the University of South Florida Ethics Committee.

Antigen production, expression, and purification of antigens
Plasmodium vivax gene sequences coding for SSP3 (PVX_123155), SPECT1 (PVP01_ 1212300), SPELD (PVX_092505), and M2-MAEBL (PVP01_0948400) were acquired from PlasmoDB [38].The signal peptide of PvSSP3 and PvSPECT1 as well as the transmembrane domain for PvSSP3 and PvSPELD were excluded in expressed recombinant proteins (Fig 1b).Modifications were done to ensure that the functional domains remained intact while hydrophobic and disordered regions were removed (Fig 1b and S1 Dataset).Relative molecular weight and predicted isoelectric points were estimated using ExPASy (Compute pI/Mw tool).The modified coding sequences were codon-optimized for E. coli expression (Genscript, USA) and cloned into a pET21a+ expression vector with a C-terminal 6xHis-tag.Recombinant antigens were expressed in One Shot BL21 Star (DE3) Escherichia coli, henceforth referred to as BL21 (Invitrogen).Starter cultures (350ml) of E. coli expressing clones were grown overnight.The next day, the starter culture was diluted to 20% in 4.2 L of warm LB broth.The bacterial broth was incubated at 37˚C, 250 rpm until an OD value of 0.6 to 0.7 and induced with IPTG (Goldbio) to a final concentration of 1 mM.The induced culture was left to incubate at 30˚C, 250 rpm for 6-8 hours, and the bacterial pellet was harvested by centrifuging at 10,000 rpm.
To determine the solubility of the expressed protein, the bacterial pellet was lysed with phosphate buffer (0.5M NaCl, 20 mM sodium phosphate buffer, and 25 mM Imidazole) and purified by affinity chromatography using the HisTrap HP columns (GE) on the Akta Pure purification system (Cytiva).Western blots were performed with an anti-His tag antibody to confirm expression of each recombinant protein.The insoluble fractions were purified from inclusion bodies (IB).Briefly, the remaining bacterial pellet was lysed using two separate buffers, buffer 1 (50 mM Tris, 0.5M NaCl, 0.2 mM EDTA, 3% sucrose, and 1% Triton X-100) and buffer 2 (50 mM Tris, 0.5M NaCl, 0.2 mM EDTA and 3M urea) and IB was recovered by centrifugation.The IB was solubilized in a 20mM phosphate buffer containing 6M guanidine hydrochloride and recombinant protein purified by affinity chromatography and refolded by rapid dilution.SDS-PAGE was used to analyze the native formation of refolded protein by observing protein migration in reducing and non-reducing conditions.SeeBlue Plus2 Pre-Stained Protein Standard (ThermoFisher) and Precision Plus Protein Dual Color Standard (BioRad) were used to evaluate relative molecular weight.Endotoxins were removed from proteins using a ToxinEraser Endotoxin Removal kit (GenScript).

Variant calling and epitope analyses
High-quality variant calls were extracted from a previously published dataset of Southeast Asian P. vivax samples as described [39] (S2 Dataset).FASTA sequences for all antigens were sourced from NCBI via a MegaBLAST restricted to P. vivax (taxid:5855).To expand our dataset, we searched available genome assemblies for antigen sequences.4 published draft genomes had corresponding genes extracted and 19 genome assemblies which had been created as described in [40] up through correction were searched via blast for similarity to PVP01_0938800, PVP01_0948400, PVP01_1212300, and PVP01_1427900.Other published genomes including PvC01, PvT01, PvW1, PvCMB-1, and Pv_N_Korean_v1 were obtained from respective repositories and similarly BLAST searched [41][42][43].Sequences were aligned in MEGA 11 [44] and formatted to remove gaps and ambiguous sequences.Site-specific nonsynonymous and synonymous substitution rates (dN/dS) were analyzed for selective pressure on Datamonkey webserver (Datamonkey Adaptive Evolution Server) (S3 Dataset).

Enzyme-Linked immunosorbent assay (ELISA)
Nunc 96-well flat-bottom microtiter plates were coated with 2 μg/ml of the desired antigen in carbonate coating buffer (0.0125 M NaHCO 3 , 0.0875 M Na 2 CO 3 at pH 9.6) overnight at 4˚C.Plate wells were blocked with 5% non-fat milk in PBS with 0.05% Tween (PBST) for 2 hours at room temperature (RT).Plates wells were then incubated in duplicate for 2 hours at RT with serial dilution of mouse sera (starting at 1:100).Bound immune sera were detected using Blue-Phos Microwell Phosphatase Substrate System (KPL) measured at 650 nm absorbance.
To determine total IgG antibody levels in patient plasma, the plasma was tested at 1:200 dilutions.The protocol used was similar to that used for the mouse sera except goat antihuman IgG-peroxidase conjugated secondary antibody (Sigma) was used for detection.A baseline OD was established using plasma samples from the 30 healthy control (HC) volunteers.Total IgG antibody levels were reported as reactivity index (RI), calculated by dividing OD values of tested samples by a cut-off value (mean + 2 s.d.) from HC samples.An RI � 1 was considered responders while RI < 1 were considered non-responders.
Sporozoite lysate was also tested for reactivity with immune mouse sera raised against the various antigens by ELISA as described for the recombinant antigens above.

Immunofluorescence assay (IFA)
Cryopreserved P. vivax sporozoites were thawed at RT, and quantified using a hemocytometer.Sporozoites were spotted on a specimen slide (Tekdon) at a density of 1x10 4 sporozoites per well.The slides were air dried and fixed with 4% paraformaldehyde (PFA) for 15 minutes and washed thrice with PBS.The fixed-sporozoites were blocked in 5% fetal bovine serum (FBS) in PBS for 45 minutes at RT and then incubated overnight at 4˚C with mouse polyclonal antibodies (pAbs) of rPvSPECT1, rPvSSP3 or rPvM2-MAEBL at 1:500 dilution.The slides were washed thrice and incubated with Alexa Fluor 488 goat anti-mouse IgG secondary antibody (1 μg/ml, Invitrogen) and Hoechst 33342 (1:1000 dilution, Fisher Scientific) for 1 hour at RT. ProLong Glass Antifade Mountant (Thermo Scientific) was used to preserve signal intensity during image acquisition.pAbs to Duffy binding protein (DBPII) served as an isotype control and anti-PvCSP VK210 (2F2) mAbs served as a positive control.Sporozoites were examined under oil immersion at x100 magnification using the Zeiss AXIO Observer.Z1 (ZEISS Group).

Study participants
Regions of low malaria transmission along Southern Thailand in the Ranong and Chumphon provinces were chosen to recruit P. vivax-infected volunteers (S1 Table ).The volunteers (n = 52) were diagnosed by microscopic examination of both thin and thick blood films and confirmed by nested PCR.Plasma samples were collected to test the reactivity of total IgG to recombinant antigens.Volunteers were a mixed group of Thai (n = 41) and Myanmar (n = 11) nationals, ages ranging from 18 to 63 years old (S1 Table ).Naïve control plasma (n = 30) was obtained from volunteers in a non-malaria endemic region (Bangkok, Thailand).

Inhibition of liver stage development assay (ILSDA) with cryopreserved P. vivax sporozoites
Mouse sera inhibition of P. vivax sporozoite invasion and development was assessed using primary human hepatocytes (PHH) donor lot YNS (Bioreclamation IVT).Cryopreserved P. vivax sporozoites were thawed as previously described in hepatocyte culture medium (HCM) (InVitroGro CP Medium) at RT [49].2x10 4 sporozoites were incubated with pooled and heat inactivated mouse sera at 1:50 dilution in HCM for 20 minutes prior to adding to wells of PHH seeded plate in triplicates.Infected plates were centrifuged at 200 x g for 5 minutes and sporozoites allowed to invade at 37˚C for 16 hours before changing the medium.PHH medium was changed every alternate day until day 6 post infection and the cells were fixed with 4% PFA.Non-treated sporozoites served as infection control for normalization while anti-Duffy Binding protein polyclonal mouse sera served as a non-specific control.
Liver-stage (LS) forms high content imaging were accomplished using the Cell Insight CX7 system and the HCS studio software via a modified Target Activation Advanced Bioapplication (Thermo Fisher Scientific).LS forms were counted using the Alex Fluor 488 channel and identified by mean intensity and cell roundness as previously described [50].Wells with no test sera exposed sporozoites served as controls for % inhibition calculations.
Where 'x' is average number of LS forms in wells with sera exposed sporozoites and 'y' is average number of LS forms in control wells previously described [50].

Statistical analyses
ELISA data were analyzed using GraphPad Prism v9, GraphPad (San Diego, USA).A baseline OD was established using plasma samples from the 30 healthy controls (HC).Total IgG for each antigen was standardized to a reactivity index (RI), calculated by dividing the OD values of tested samples by a cut-off value (mean + 2SD) from HC samples.An RI � 1 was positive for specific antibodies; those with an RI < 1 were considered negative.RI was tested for normality using the Shapiro-Wilk test.A Kruskal-Wallis one-way ANOVA with a Dunn's multiple comparisons test was used to compare seropositivity between groups.Mouse endpoint titers were tested for normality using D'Agostino & Pearson test.Analysis of mouse endpoint titers to sporozoites, antisera inhibition, and LS form area were done using Kruskal-Wallis one-way ANOVA with Dunn's multiple comparisons test as the groups were not normally distributed.For selection pressure analyses, we used the fixed effects likelihood (FEL) [51] and mixed effects model of evolution (MEME) [52] with a threshold p-value of p� 0.05 after removing identical sequences.

Expression of P. vivax PE antigens in bacteria
The selected P. vivax sporozoite antigens represent the migratory-to-hepatocyte-infecting phases of the PE infection processes (Fig 1a and Table 1).Signal peptides and other potentially hydrophobic regions were not included as part of the recombinant proteins.Coding sequences of the PE candidates were expressed in bacteria as near full-length recombinant proteins (rPvSPELD, rPvSPECT1) or the primary functional domains of the PE candidates (rPvMAEBL, rPvSSP3) (Fig 1b).In particular, the expressed fragments of rPvSSP3 included the predicted S48/45 functional domain, and rPvMAEBL contained the M2 PAN-like putative ligand domain.The full-length ectodomain of rPvSPECT1 retained all four helix bundles and the ectodomain of rPvSPELD contained its tyrosine-rich region (Fig 1b).Further, we tested feasibility for future product development as a manufactured recombinant protein product using the established protocol developed for production of a conformationally-correct cysteine-rich ligand domain of the P. vivax Duffy binding protein (DBP), also known as DBPII, a blood-stage vaccine candidate [53].In bacterial expression rPvSPECT1 could be directly purified from the soluble fraction of clarified bacterial lysate, while rPvSSP3, rPvSPELD and rPvM2-MAEBL were purified from inclusion bodies, and refolded to native conformation by rapid dilution.The purified recombinant proteins of rPvSSP3, rPvSPELD and rPvSPECT1 expressed at the expected sizes of 23.0 kDa, 20.3 kDa and 25.7 kDa compared to standard protein ladder when analyzed by SDS-PAGE (Fig 2a -2c) with average yields of 12 mg/L, 7.25 mg/ L, and 2.85 mg/L, respectively.Some rPvSPECT1 was present in inclusion bodies, but most of the expressed recombinant antigen was in the soluble fraction.rPvM2-MAEBL migrated at ~64 kDa, over 30% higher than its predicted 48.3 kDa (Fig 2d) and had an average yield of 1.9 mg/L.Unlike rPvM2-MAEBL and rPvSSP3, rPvSPELD aggregates after purification and refolding from inclusion bodies, leading to its exclusion from further characterization.

Immunogenicity studies generated high-titer antibodies that recognized antigens on cryopreserved P. vivax sporozoites
To determine whether the immune sera raised against the recombinant proteins recognized P. vivax sporozoites, we performed an ELISA with crude sporozoite lysate as well as indirect immunofluorescence with sporozoites.The murine antisera to the recombinant PE antigens recognized epitopes on P. vivax sporozoites (Fig 3).Antibody endpoint titers were determined (sera from day 56) by reactivity to whole sporozoite lysate (Fig 3a).Titers ranged from 1:200 to 1:6400 and PvSPECT1 had significantly higher reactivity to sporozoites than PvM2-MAEBL Table 1.Key attributes of P. vivax PE antigens to prioritize as potential vaccine candidates.

Mouse antisera inhibit invasion and attenuate growth of P. vivax sporozoites in PHH
The potential vaccine efficacy of the PE antigens was evaluated using an in vitro assay for inhibition of liver stage development (ILSDA).The 384-well liver assay uses primary human hepatocytes to support high level of infections with P. vivax and P. falciparum sporozoites leading to complete liver development to blood stage breakthrough [50].Each antiserum to the recombinant PE antigens demonstrated significant inhibitory activity against P. vivax sporozoite invasion into hepatocytes as quantified by high content imaging at day 6 post-infection relative to control (Fig 4a).Antisera to PvSPECT1, PvSSP3 and PvM2-MAEBL inhibited P. vivax sporozoite invasion by 31%, 30%, and 25% respectively (Fig 4a).
In addition, a decreased growth phenotype was observed in developing LS forms relative to the no serum control for all antisera tested (Fig 4b).Surprisingly, sporozoite pre-incubation with the anti-DBPII serum also resulted in growth attenuation of the P. vivax liver stages, which might reflect some toxicity in the mouse serum or cross-reactivity with a sporozoite antigen.Additionally, the proportion of developing LS small forms (<100 μm 2 ) to large forms (>100 μm 2 ) were similar between groups in the donor lot tested for all experimental groups (S3 Fig).

Post-infection serological responses to PE candidate antigens
We next evaluated whether residents in a malaria endemic region in Thailand with natural exposure to P. vivax had serological reactivity against the recombinant PE candidate antigens versus healthy controls (HC).Of 52 samples analyzed by ELISA, seropositivity was observed in 11 (21.1%), 15 (28.8%), and 17 (32.7%)against rPvM2-MAEBL, rPvSSP3, and rPvSPECT1, respectively (Fig 5).However, significant differences were not observed between antigen responses (Fig 5).Interestingly, only 7.6% of the samples were seropositive for all three antigens and both rPvSSP3 and rPvSPECT1, while 3.8% were reactive to both rPvM2-MAEBL and rPvSSP3 and 1.9% reactive to both rPvM2-MAEBL and rPvSPECT1 (Fig 5).

Variant calling of genes identifies polymorphisms in putative epitopes
We next identified the polymorphisms by variant calling among sequence information available from the published sequences to characterize residues in regions that may be under immune selection.There were only 5 polymorphisms in the M2 domain apart from the 19 other SNPs in the PvMAEBL CDS.Likewise, we observed 4 (PvSSP3) and 1 (PvSPELD) mutation (S2 Dataset).Surprisingly, no mutations were identified within the CDS of PvSPECT1.Therefore, based upon frequency of synonymous to nonsynonymous mutations (p� 0.05) this analysis suggests on this small sample size that none of the genes for these PE candidates are under strong immune selection (Table 2 and S3 Dataset).Though the observed variation was relatively limited compared to many merozoite surface antigens, further analysis indicated some of these variant residues were observed in putative B cell epitopes in the expressed portions of the PE candidates.In PvSSP3, PvM2-MAEBL, and PvSPELD, we identified 8 (PvSSP3 & PvSPECT1), 12 (PvM2-MAEBL), and 3 (PvSPELD) (S4 Dataset).There were 2 and 114 (PvSSP3), 12 and 140 (PvSPELD), 12 and 146 (PvSPECT1), 13 and 106 (PvM2-MAEBL) putative T cell epitopes for MHC Class I and II, respectively (S4 Dataset).A single point mutation was observed in SPELD (I73T), PvM2-MAEBL (I981S) for the Class II and I epitopes, respectively (Table 2).PvSSP3 had a single point mutation Q174E in a putative B cell epitope, with an uncharged glutamine replaced by a charged glutamic acid (Table 2).

Discussion
The skin-to-hepatocyte stage transition (Fig 1a) represents a critical bottleneck in the life cycle of Plasmodium spp., since sporozoites are non-replicating and relatively few are injected by a mosquito bite [4,5].Nonetheless, completion of development in the liver by a single sporozoite can result in a clinical malaria and potentially death.The poor track record in bringing to market effective anti-sporozoite vaccines that can block infection clearly demonstrates the challenge inducing an immune response capable of neutralizing sporozoites.Multiple immune interventions need to be considered to exploit this bottleneck for a designing an effective PE vaccine candidate.Therefore, our goal is to prioritize PE targets critical for sporozoite invasion that induce a broadly neutralizing, strain-transcending, functionally inhibitory antibody response.Previously, we identified P. vivax PE antigens that were upregulated in an infective microenvironment [25].We further evaluated key characteristics of vaccine candidates with the objective of down-selecting potential PE targets suitable to be incorporated into multivalent antibody-based vaccine effective against P. vivax (Table 1).Based on these criteria, PvSSP3, PvM2-MAEBL, PvSPECT1, and PvSPELD were chosen for further characterization in our study.
Here, we successfully expressed recombinant PvM2-MAEBL, PvSSP3 and PvSPECT1 in an established recombinant protein expression system that allows for high-quality cost-effective protein production and is suitable for clinical studies [3,59,60].Bacterial expression also prevents unwanted post-translational modifications and is relatively easy to manipulate [61].Previous studies demonstrated that apart from transmembrane and signal peptide regions, greater protein disorder, and a more basic isoelectric point contribute to reduced soluble expression in E. coli [62].Therefore, recombinant antigens were designed by eliminating hydrophobic regions without truncating putative and confirmed domains for PE protein function (Fig 1b).
In light of these antigens' essentiality to Plasmodium spp.liver stage we generated antisera in mice for characterization.High titers were observed against all the recombinant antigens (S2 Fig) , although reactivity to whole sporozoites lysate was much lower, except for PvSPECT1 (Fig 3a).The basis of the high seroprevalence to PvSSP3 was unexpected as it is a minor surface protein relative to the immunodominant CSP.However, despite the possible immune distraction by the immunodominant antigen CSP, in vitro studies with P. yoelii revealed when PyCSP was shed PySSP3 was unmasked [27].Therefore, we hypothesize that the high seropositive rate we observed indicates that the PvSSP3 could be activated in a similar manner leading to its  [27,28,30,32,35,36,57] Consistent with this IFA localization, immunoelectron microscopy demonstrated that PbSPECT1 is associated with sporozoite micronemes [35].Likewise, previous studies have demonstrated MAEBL to be closely associated with sporozoite micronemes in both P. berghei and P. falciparum [31,32].
To further characterize the immune sera raised against the PE candidate antigens, we evaluated if the anti-PE antisera had inhibitory activity against P. vivax sporozoites.Orthologues of these PE antigens have been shown to play important roles in gliding, traversal and invasion [27,28,30,31,33,35].The ILSDA used primary human hepatocytes and P. vivax sporozoites as described previously [49,50].Each PE-specific antiserum significantly inhibited P. vivax hepatocyte invasion (Fig 4a).Interestingly, the mouse antisera to PvM2-MAEBL demonstrated a similar level of inhibition as a previously reported rabbit antiserum to P. yoelli M2-MAEBL [30].These data suggest the presence of functional inhibitory epitopes within the expressed recombinant antigens.Comparative analysis showed an overall decreased growth phenotype for LS forms in all antisera tested (Fig 4b).These data are similar to our previous study that reported a decreased growth phenotype after incubating sporozoites with monoclonal antibodies to CSP repeats of both P. vivax and P. falciparum as well as antisera to CSP-based P. vivax nanoparticle vaccine [50].These results hint at a post-invasion antibody-mediated mechanism not yet understood.
We next asked the question if our recombinant P. vivax PE antigens are naturally immunogenic in P. vivax infections by testing post-infection plasma from residents living in a vivaxendemic region of Thailand.PvM2-MAEBL had the lowest overall seropositivity, much lower than PvSSP3 and PvSPECT1, but was significantly different from healthy controls (Fig 5).This was surprising since MAEBL was highly immunogenic in controlled human experimental infection by P. falciparum sporozoite inoculation [72].PvSPECT1 had the highest seroprevalence at ~32% followed by PvSSP3 at 28.8% seropositivity (Fig 5).The high level of seropositivity to PvSPECT1 may be related due to its relative abundance, since it has been identified as one of the most abundant proteins detected in P. vivax and P. falciparum sporozoites [73][74][75].
Surface-exposed B cell epitopes are often highly polymorphic as they can be accessible to neutralizing immune antibodies [76,77].Moreover, such polymorphisms occur as escape mutants are driven by immune selection and the basis of strain-specific immune responses, which is better characterized for P. vivax blood-stage antigens [78][79][80][81] compared to most P. vivax PE antigens.Similarly, variation can be driven in T cell epitopes by abrogating anamnestic responses from prior infections.Additionally, recent systems serology studies highlighted the crucial role of Fc-effector mechanisms in controlling malaria infections [82,83].Since humoral effector mechanisms can be bolstered by cell-mediated responses as both play a critical role in protection [84][85][86][87] we next evaluated the prevalence of variant epitopes that would be indicative of immune selection pressure.FEL was used to detect a diverse selection in a small sample set.Additionally, we also wanted to determine if there is pervasive positive selection by using MEME [51,52].Positive selection was not detected in our study, probably due to the small sample size.However, a single mutation (D236N), was observed for PvSPECT1 in the MalariaGEN P. vivax Genome Variation project's genome browser [88].We also, identified putative immunogenic epitopes that were consistent with immune selection, using the Immune Epitope Database.For PvM2-MAEBL, a single polymorphic mutation (I981S) was observed in putative B and class I epitopes.Previously for PfMAEBL, the epitope, 'YVSSFIRP-DYETKCPPRYPL', was validated to have high affinity to HLA-DRB1*0301 and *1101 and was mostly conserved [89,90].We identified a similar epitope 'YEEKCPPRFPL' that had an intermediate binding affinity with an IC 50 of 101nm & 110 nm for Class I & II, respectively (S4 Dataset).
In this study, we evaluated P. vivax PE antigens for an antibody-based vaccine.We provide preliminary characterization of three potential candidates (SSP3, SPECT1, M2-MAEBL).We successfully demonstrated that these P. vivax PE antigens expressed as subunit proteins in bacteria are immunogenic in laboratory animals, eliciting functionally inhibitory antibodies that inhibited P. vivax liver-stage invasion, and the PE proteins are naturally immunogenic.Altogether our data indicate that these conserved antigens represent suitable targets for P. vivax vaccine development.Ultimately, the further evaluation of immunogenicity and protective efficacy of each of these antigens alone or in combination could aid in the development of an effective vaccine against P. vivax.

Fig 1 .
Fig 1. Critical functions of sporozoite invasion facilitated by antigens of interest.a) Schematic of skin-to-hepatocyte stage demonstrating the functional role of antigens in invasive sporozoites.Created with BioRender.comb) P. vivax PE antigens chosen for overexpression.Schematic of protein constructs (not drawn to scale) with putative domains and secondary structures identified.Highlighted sections in blue boxes represent the estimated domain/region chosen for recombinant expression in BL21.https://doi.org/10.1371/journal.pntd.0011598.g001

(
p = 0.005) (Fig 3a), which might reflect relative abundance or epitope accessibility in the crude antigen extract instead differences in titer.Indirect immunofluorescent revealed the expected surface-association with PvSSP3 and PvSPECT1 (Fig 3b and 3c) while the more intense anterior end fluorescence observed for PvM2-MAEBL was consistent with localization to the micronemes (Fig 3d).

Fig 4 .
Fig 4. Inhibition of P. vivax liver stage development.a) Mouse antisera to each PE vaccine candidate significantly inhibited P. vivax sporozoite invasion in primary human hepatocytes versus no antibody control and antiserum to P. vivax DBP.b) P. vivax sporozoites incubated with anti-PE sera significantly decreased in size post-invasion.Small LS forms (<100 μm 2 ) were removed from all populations prior to analysis.Results were determined on day 6 post-infection by high content imaging of infected hepatocytes by antibody staining of intracellular parasites with UIS4 and are expressed as mean with error bars indicating the standard error of the mean.Statistical significance was determined using a non-parametric Kruskal-Wallis one-way ANOVA with a mean rank multiple comparisons test and represented as p<0.05 (*), p = 0.0002 (***), and p<0.0001 (****).https://doi.org/10.1371/journal.pntd.0011598.g004

Fig 5 .
Fig 5. Seropositivity of plasma samples from P. vivax-infected patients (n = 52) in Ranong and Chumphon provinces of Thailand.The reactivity index (RI) of total IgG responses to rPvM2-MAEBL, rPvSSP3, and rPvSPECT1.Dotted line indicates the cutoff for RI samples � 1. Error bars indicate the s.e.m.Statistical significance between responses was tested using a Kruskal-Wallis one-way ANOVA with a mean rank multiple comparisons test and represented as not significant (ns).https://doi.org/10.1371/journal.pntd.0011598.g005

Table 2 . Polymorphic residues of PvSSP3, PvSPELD, and PvM2-MAEBL in putative epitopes.
Variants in epitopes were determined after a MUSCLE alignment using MEGA.Codon-by-codon selective pressure was statistically tested by FEL and MEME (p�0.05) in putative linear B-cell and MHC I/II T-cell epitopes.Indirect immunofluorescence stained sporozoites revealed a surface association of PvSSP3 and PvSPECT1 (Fig 3b and 3c).PvM2-MAEBL was associated with the apical end of the P. vivax sporozoites (Fig 3d), similar to other studies.