Skip to main content
NCBI home page
Plant Biotechnology Journal logoLink to Plant Biotechnology Journal
. 2023 Aug 12;21(12):2546–2559. doi: 10.1111/pbi.14152

Rice‐produced classical swine fever virus glycoprotein E2 with herringbone‐dimer design to enhance immune responses

Qianru Xu 1,2,3, † , Fanshu Ma 2,4, † , Daichang Yang 5,6, Qingmei Li 1, Liming Yan 7, Jiquan Ou 6, Longxian Zhang 2,8, Yunchao Liu 1, Quan Zhan 6, Rui Li 1, Qiang Wei 1, Hui Hu 2, Yanan Wang 1, Xueyang Li 2, Shenli Zhang 2, Jifei Yang 1, Shujun Chai 1, Yongkun Du 2, Li Wang 1, Erqin Zhang 2,8,✉, Gaiping Zhang 1,2,8,9,✉
PMCID: PMC10651154  PMID: 37572354

Summary

Pestiviruses, including classical swine fever virus, remain a concern for global animal health and are responsible for major economic losses of livestock worldwide. Despite high levels of vaccination, currently available commercial vaccines are limited by safety concerns, moderate efficacy, and required high doses. The development of new vaccines is therefore essential. Vaccine efforts should focus on optimizing antigen presentation to enhance immune responses. Here, we describe a simple herringbone‐dimer strategy for efficient vaccine design, using the classical swine fever virus E2 expressed in a rice endosperm as an example. The expression of rE2 protein was identified, with the rE2 antigen accumulating to 480 mg/kg. Immunological assays in mice, rabbits, and pigs showed high antigenicity of rE2. Two immunizations with 284 ng of the rE2 vaccine or one shot with 5.12 μg provided effective protection in pigs without interference from pre‐existing antibodies. Crystal structure and small‐angle X‐ray scattering results confirmed the stable herringbone dimeric conformation, which had two fully exposed duplex receptor binding domains. Our results demonstrated that rice endosperm is a promising platform for precise vaccine design, and this strategy can be universally applied to other Flaviviridae virus vaccines.

Keywords: transgenic rice, rational vaccine design, herringbone‐dimer, Pestivirus, E2 protein

Highlights.

  • A simple vaccine design of a herringbone‐dimer rE2 is expressed in rice.

  • The dimeric antigen elicits robust immune responses at minimum dose.

  • Ht‐dimer rE2 is likely to be a native‐like, antigenic and immunogenic homodimer.

Introduction

Viruses in the Pestivirus genus pose a global animal health burden and cause huge economic losses (Leyssen et al., 2000). Classical swine fever virus (CSFV) was the first reported virus caused by a Pestivirus and is a single‐stranded positive‐sense RNA‐enveloped virus (Becher et al., 1999). Most of the currently approved vaccines against CSFV are live attenuated vaccines, which are hindered by safety concerns, lack of rapid and long‐term protection, and failure to achieve sterilization (Blome et al., 2017; Hobson‐Peters et al., 2019; Postel et al., 2018). Additionally, most of the marketed subunit vaccines require high doses and multiple injections, which may cause B‐cell exhaustion, have potential side effects, and immune responses generally remain weak (Corbett et al., 2020; Holla et al., 2019; Tian et al., 2021; Wu et al., 2022). Thus, the development of a safe and highly efficient vaccine with low‐dose and better stability is urgently needed.

Diverse vaccine strategies have been explored, including plasmid DNA, adenovirus vectors, protein subunit, and inactivated viruses (Krol et al., 2019). In particular, subunit vaccines possess superior safety (Strauch et al., 2017). Using plants for antigen production is an attractive platform (Margolin et al., 2018; Mason et al., 1992). Plant‐based production systems appear to be optimal in terms of safety, cost, yield, and physical stability to extreme temperatures compared with other systems such as bacterial cultures, yeast, and mammalian cell cultures (Desai et al., 2010). Several plant‐based vaccines against different viruses, such as hepatitis C virus (HCV), porcine circovirus type 2 and influenza A H6N2, have been previously explored and have demonstrated the great potential of plant‐based biopharmaceuticals (Clarke et al., 2017; Dobrica et al., 2021; Gunter et al., 2019; Smith et al., 2020). The production of CSFV E2 glycoproteins in plants has been reported for plant cells and Nicotiana benthamiana (Jung et al., 2014; Laughlin et al., 2019; Legocki et al., 2005; Park et al., 2019, 2020, 2021; Yiu et al., 2013). Despite some success, these either did not report the protective effect or there is still room for optimization in terms of immunization dose. A molecular engineering approach was employed to construct a polymeric immunoglobulin G scaffold (PIGS) that incorporates multiple copies of an antigen that enhances vaccine uptake by the immune cells in vivo (Kim et al., 2017). Hence, vaccine efforts should focus on the production of redesigned immunogens with plant expression platforms to optimize antigen presentation.

As the primary target for neutralizing antibodies (NAbs), E2 alone is protective in animals and it is the major focus of subunit vaccine design (Bardina et al., 2017). Although the crystal structures of bovine viral diarrhea virus (BVDV) E2 ectodomain have been resolved from two different groups, efficient vaccines remain elusive due to limited structural information and because it is unclear how structural flexibility relates to membrane fusion. Optimizing E2 antigen to improve its immunogenicity and elicitation of NAbs by rational design may lead to an effective vaccine (Kong et al., 2015). New antigen design strategies have been sought to produce potential vaccines with a combination of native conformation and immunogenicity. On the surface of HCV, several different conformations are in dynamic equilibrium with the dose‐dependent neutralization of NAbs, suggesting that the predominant conformation with high neutralization activity at low doses is a preferred structure (Meola et al., 2015). In the case of dengue virus, EDE MAbs were found to bind virions only to dimers exposed on the viral surface that display a smooth, mature herringbone arrangement (Lin et al., 2018; Slon‐Campos et al., 2019; Zhang et al., 2013). Additionally, the previously described E–E glycoproteins in Flavivirus or E1–E2 glycoproteins in HCV designs include additional inter‐subunit disulfide bonds or covalently linked proteins, which result in structurally stable and homogeneous antigens (Ruwona et al., 2014). Hence, an alternative would be to redesign the E2 dimer, similar to the mature viron‐surfaced proteins arrangement to enhance B‐cell immune responses.

Here, we propose a simple, potent, dimer vaccine design strategy, which involves a ‘head‐to‐tail’ dimer (ht‐dimer) consisting of a flexible peptide linker [GGGGS)3]. We first validated the feasibility of vaccine design using CSFV E2 as an example and expressed the redesigned rE2 protein in a rice endosperm expression system. Vaccine potency was confirmed in mice, rabbits and pigs. Immunization of piglets with ht‐dimer vaccines induced a 32‐fold higher titer of NAbs at a minimum dose than the live attenuated vaccine, and the longevity of the E2‐specific antibodies exceeded 6 months. We further show that the vaccine elicited immune responses were not affected by pre‐existing antibodies. After the crystal of the redesigned antigen was resolved, we confirmed a herringbone conformation of rE2, with two fully‐exposed duplex receptor binding domains. As Pestivirus and HCV share many structural and functional properties, Pestivirus viruses have been widely used as surrogates for human HCV. We therefore provide a simple and useful strategy for efficient vaccine design against HCV and other viruses.

Results

Design and expression of the ‘head‐to‐tail’ (ht) dimer E2 in rice endosperm

To obtain a herringbone arrangement immunogen that facilitates the elicitation of high‐affinity antibodies, we designed and constructed a ‘head‐to‐tail’ E2 dimer. CSFV E2 (GenBank Acc. No.: AAK21202.1) was selected as the target antigen because it was recognized by potent nAbs. To increase the stability and spacing of the dimer, we fused the N terminus of one molecule with the C terminus of the other molecule via a flexible linker (GGGGS)3 (Figure 1a). The introduced flexible linker increased the spacing and the tandem dimer exposed the antigenic epitopes more fully to verify whether the design enhanced the immune response.

Figure 1.

Figure 1

Open in a new tab

Design and expression of the ‘head‐to‐tail’ dimer rE2 in rice endosperm. (a) Herringbone pattern of E dimers on the surface of mature virus particles. The target antigen E2 was linked by a flexible peptide to form a stable ‘head‐to‐tail’ E2 dimer. (b) Plasmids used for the expression of ht‐dimer E2. Different stages of transgenic rice obtained by Agrobacterium‐mediated transformation, including rice gene transformation, screening of positive plants, rooting and rice maturation. (c) SDS‐PAGE of purified ht‐dimer rE2 protein. M represents a molecular marker. (d) The dilution of purified‐rE2 was determined by a CSFV E2 antigen rapid test strip. The dilution ranged from 1: 1000 to 1: 32 000. Negative control, TP309. (e) SEC chromatograms of rE2, showing effective recombination into a dimeric state (12–13 mL). Peak 1 marks the rE2 protein. Native PAGE of rE2 protein shows the homogeneity of the immunogen. (f) Sedimentation velocity analytical ultracentrifugation (SV‐AUC) of rE2 dimer protein. C(s) distributions from sedimentation velocity (SV) runs for rE2 (1.1 mg mL˗1).

To obtain the ht‐dimer E2 antigen, the ht‐dimer E2 gene was expressed in rice endosperm (rice‐derived E2, rE2). First, optimization of the ht‐dimer E2 gene with rice‐preferred gene codons and transfer of the gene along with an endosperm‐specific promoter (Gt13a and its signal peptide; Figure 1b) into an Agrobacterium both contributed towards high expression levels of rE2 (He et al., 2011). A total of 119 independent positive transgenic plants were obtained via Agrobacterium‐mediated transformation. Then, stable expression lines of transgenic rice were screened by qPCR (Wang et al., 2015), and five homozygous lines with high expression were selected from 119 strains of the T3 generation for further study, of which 119‐5‐1 from the T2–T4 generation was highly stable. The rE2 protein expressed in rice endosperm was analysed by SDS‐PAGE and had a molecular weight (MW) of ~63 kDa (Figure 1c). The band was specifically recognized by antibodies against CSFV by Western blots. In addition, we tested the antigenicity of the rE2 protein by the CSFV E2 antigen rapid test strip, which is a fast and efficient tool for screening large samples of plant expression (Figure 1d). The results showed that we had successfully obtained transgenic rice capable of stably expressing ht‐dimer E2 protein and preliminarily validated the biological activity of rE2 protein.

Separation and formation of ht‐dimer rE2

To explore whether the rE2 protein could dimerize stably and homogeneously, we investigated the behavior of rE2 in solution. The rE2 protein was produced by positive transgenic rice and extracted into the supernatant. No additional purification tags were introduced to ensure proper protein folding. The soluble rE2 protein was harvested and purified to 95% purity by ion exchange chromatography, hydrophobic chromatography and size‐exclusion chromatography (SEC). Based on the analysis of peak 1 by SEC, it could be concluded that the restructured protein molecule was a dimer with a protein size of ~63 kDa (Figure 1e; Table S2). To confirm the MW of rE2, we performed a MALDI‐TOF‐MS assay and the results were consistent with the predicted protein size. In addition, native PAGE demonstrated that the rE2 protein was dominated by one formation, indicating that the protein is homogeneous.

To further verify the polymerization state of rE2, we investigated the behavior of rE2 in solution. As judged by sedimentation velocity analytical ultracentrifugation (SV‐AUC), two distinct species were observed, with most proteins forming rod‐like dimers (apparent sedimentation coefficient, sw = 2.7 S) with MWs of ~63 kDa, and a few forming multimers (sw = 5.1 S) with MWs of ~165 kDa (Figure 1f; Table S1). The two species could be accounted for by the presence of protein aggregation, reaching a dimer–multimer equilibrium.

Validation of vaccine protection by ht‐dimer rE2 in mice and rabbits

To assess the immunogenicity of rE2 protein, we immunized Balb/c mice (n = 5 in each group) with 5 μg of rE2 in the adjuvant 50 V (SEPPIC) at Days 0 and 28, and measured the E2‐specific antibody levels for 8 weeks (Figure S1a). The rE2 group elicited higher antigen‐specific antibody levels compared with the live attenuated vaccine 7 days after the first immunization and the antibody response reached a maximum at day 56 (P < 0.01; Figure S1b). Then, a rabbit model of CSFV infection was used to evaluate the ht‐dimer rE2 as an immunogen. New Zealand White rabbits (n = 4 in each group) were vaccinated only once with 20 μg of rE2, which resulted in the induction of E2‐specific antibodies as early as 7 days of post‐immunization and sustained high antibody responses (P < 0.001; Figure S1c,d). By comparison, rE2 elicited responses roughly similar to those of the positive controls, both of which were higher than the theoretical threshold of protection. A typical post‐challenge fever was not detected in rabbits vaccinated with rE2 or the positive control, but was recorded for the negative control (Figure S1e). These findings suggested that rE2 immunogen elicited protective responses against CSF virus.

Utility of rE2 as a vaccine antigen in pigs with or without pre‐existing immunity

To further evaluate the effects of the designed ht‐dimer rE2 on the elicitation of protective humoral responses, we immunized pigs (n = 5 in each group) with 30 μg of rE2 mixed with an equal volume of 50 V adjuvant at days 0 and 28. Sera were collected weekly, and humoral immune responses were measured for specific antibodies (Figure 2a) and NAbs that inhibit viral infection. Consistent with the results in mice and rabbits, rE2 induced CSFV‐specific antibodies as early as 7 days of postimmunization. At Day 56, the rE2 group elicited neutralizing antibody up to an endpoint titre of 4096, which was higher than the neutralizing antibody titre of 64 induced by a commercial live attenuated vaccine (P < 0.001; Figure 2b). In addition, Th1 cytokines (IL‐2 and IFN‐γ) were measured in peripheral blood, and a significant induction was observed after the second vaccination (P < 0.05; Figure 2e,f). On Day 56, the pigs were attacked and there was a significant difference (P < 0.05) between the negative control group and the rE2 group in terms of weight (Figure 2g). Body temperatures were monitored daily up to 9 days of postchallenge, and the negative control showed a significantly elevated body temperature while the other groups showed normal body temperatures (Figure 2h). Survival rates for the rE2 group and the positive control group were 100% and 60%, respectively; by contrast, all negative controls died within 6 days (Figure 2i). To further describe the protective effect, tissues were observed by histopathological sectioning at 18 days of postchallenge. As shown in Figure 2j, all lymph node, spleen and ceca samples harvested from the negative controls exhibited significant haemorrhaging and necrosis. Milder lesions were observed in the positive controls, as the spleen showed slight necrosis. Notably, no histopathological changes were observed in the rE2 group, along with increased numbers of T and B cells in the spleen and lymph nodes (Figure 2h), indicating a surge in the immune response.

Figure 2.

Figure 2

Open in a new tab

Protective immunity induced in pigs with or without pre‐existing immunity. (a) Immunological trend map without maternal antibodies. (b) CSFV neutralizing antibody titers of pigs without maternal antibodies. (c) Immunological trend map with maternal antibodies. (d) CSFV neutralizing antibody titers of pigs without maternal antibodies. (e) IL‐2 concentration in the peripheral blood were measured. (f) IFN‐γ concentration in the peripheral blood were measured. (g) Animals were monitored for weight increase and daily temperatures after challenge. (h) Rectal temperatures were recorded over 9 days. (i) Survival curves after challenge with Shimen strain, with the death of five pigs from the negative controls and two pigs from the positive controls. (j) H&E staining of the spleen and lymph nodus in pigs after challenge (400× magnification). The tissue regions of the lymph nodes and spleen showing a surge of lymphocytes that is demarcated by the black dashed circles (100× magnification). All data are the mean ± SEM. ****P < 0.0001, ***P < 0.001, **P < 0.01, *P < 0.05. Statistical analysis was performed by one‐way ANOVA with a Bonferroni post‐test.

Pre‐existing antibodies are one of the major challenges affecting vaccine development, including some vaccines of the Flaviviridae family. To verify whether ht‐dimer rE2 is affected by pre‐existing antibodies, we selected pigs with maternal antibodies. Due to the use of CSFV live attenuated vaccines, most piglets possess maternal antibodies. The results showed that the antibody responses in the rE2 group with maternal antibodies were consistent with those of pigs without maternal antibodies, with antibody levels reaching a maximum at Day 56 and low dispersion (n = 5 in each group; Figure 2c,d). However, interference by maternal antibodies resulted in a decline in antibody levels in the live attenuated vaccine group over 21 days following the first immunization, increasing the likelihood of wild virus infection. In summary, we demonstrated that ht‐dimer rE2 is a protective antigen against CSFV that can rapidly and efficiently activate the humoral response without interference from pre‐existing antibodies.

Optimization and comparison of minimum dose rE2 immunization regimen

To optimize the immunization procedure of ht‐dimer rE2, the different inoculation doses, duration of immunization and single‐dose immunization were investigated. The results showed that all CSF vaccines were effective in inducing E2‐specific antibodies; however, compared with live attenuated vaccines, ht‐dimer rE2 had a higher antibody response at a dose of 5.12 μg; while at doses of 2.56 μg and 853 ng, antibody levels were consistent for both vaccines. This result was consistent with the presence or absence of pre‐existing antibodies (Figure 3a,b). The low‐dose rE2 groups still elicited high levels of NAbs at Month 6 (Figure 3b). When pigs were immunized with a single dose of 5.12 μg of rE2, the antibody blocking rate reached 60% and lasted for 49 days (Figure 3c). Figure 3d shows that pigs immunized twice with 284 ng were able to achieve protection that lasted for at least 6 months. The 5.12 μg group had the highest antibody levels and the 284 ng group had the lowest levels, which was consistent with dose‐dependence. These results demonstrated that rE2 vaccine could produce a sustained and highly effective acquired immune response at minimum dose. To investigate the possible contributions of ht‐dimer, we also tested Ig isotypes in animal serum. The IgG isotypes were detected on Days 7 and 35 postimmunization and the antibody response was characterized by IgG 2a (Figure 3e).

Figure 3.

Figure 3

Open in a new tab

Optimization and comparison of minimum dose for rE2 immunization. (a) The antibody blocking rates of the rE2 groups were compared with the live attenuated vaccine groups. Pigs without maternal antibodies were immunized with 284, 853 ng, 2.56 and 5.12 μg of rE2 (Upper). Pigs with maternal antibodies were immunized with 284, 853 ng, 2.56 and 5.12 μg of rE2 (Lower). (b) CSFV neutralizing antibody titers after minimum dose immunization with rE2 vaccine. (c) Immune trend map of single immunization of rE2. (d) Immune trend map of duration with minimum dose rE2 vaccine. (e) The concentrations of cytokines in the peripheral blood. (f) Immunity difference between the rE2 groups and subunit vaccine groups using different dosages. Differential maps of 284, 853 ng, 2.56 and 5.12 μg. All data are the mean ± SEM.

Given that the designed immunogen induces high nAb levels and complete immune protection, we tested the immune potential of ht‐dimer rE2 at low doses. Vaccine efficacy was also compared among ht‐dimeric rE2 and recombinant E2 (Figure 3f). Pigs were immunized with doses of 284, 853 ng, 2.56 and 5.12 μg of the recombinant E2 or ht‐dimer rE2 at days 0 and 28. Compared with the recombinant E2, ht‐dimer rE2 protein significantly increased antibody responses at all four doses.

Structural characterization of the herringbone conformation of rE2

To gain insight into the shape of rE2 in solution, we performed small‐angle X‐ray scattering (SAXS) to determine the low‐resolution molecular structure of rE2 using the DAMMIF and DAMAVER programs. The P(r) function suggested that the maximum dimension of rE2 was close to 198 Å and was a dumbbell‐like assembly dimer at 5 mg/mL, which was consistent with the AUC results. The dimensions of the rE2 protein in solution were approximately 198 Å × 100 Å × 65 Å. These results indicated that the ht‐dimer rE2 protein is a stable dimer in liquid, while the exact dimer arrangement remains unclear.

To confirm that the molecular structure of the rE2 protein was as expected, the rE2 dimer was further analysed by its crystal structure. We determined the crystal structure of CSFV rE2 from strain Shimen. The rE2 construct lacked the C‐terminal transmembrane anchor and the structure was refined at a 4.4‐Å resolution by single anomalous diffraction (Table S3). The asymmetric structural unit of rE2 consisted of two molecules, forming a single stable parallel conformation. The overall structural configuration of rE2 dimer showed certain significant differences to the structure of BVDV E2 (PDB ID: 4JNT and 2YQ2) dimer determined previously, exhibiting a unique structure. In contrast to the tail‐to‐tail E2 dimer, the head‐to‐tail dimer matched the central rod‐shape region of the SAXS model in solution (Figure 4a). The ht‐dimeric arrangement revealed that rE2 sits ‘flat’, with each pair of epitopes arranged in anti‐parallel, exposed in the same plane (Figure 4b). In the ht‐dimer rE, the epitopes located in the C terminus would be highly conserved and fully exposed, which is important for the subsequent study of its structure and function (Figure 4c).

Figure 4.

Figure 4

Open in a new tab

Structural characterization of the ‘head‐to‐tail’ dimer rE2. (a) The rE2 envelope accommodates the crystallographic E2 structure in which the flexible (GGGGS)3 connects the N‐C of two E2 molecules. The experimental envelope is shown as a white rod‐like surface. Domain A (DA; grey) was modelled with BVDV E2 (PDB: 4JNT) DA as a template. (b) Surface representation of the rE2 dimer. The four major antigenic epitope regions are indicated as Ep_1 (N‐terminus; purple), Ep_2 (CTAVSPTTLRTEVVK; yellow), Ep_3 (FRREKPFPHRMDCVTTTVENED; magenta) and Ep_4 (NKYYEPRDSYFQQY; cyan). (c) The conservation analysis of rE2 protein, shown in two orientations. Blue paths are variable and pink paths are conserved.

The monomers of CSFV rE2 and BVDV E2 (PDB: 4JNT) were similar, with the main difference being between the CSFV rE2 and BVDV E2 (PDB: 2YQ2) monomers in the C segment (Figure 5a, left and middle). Comparing the above three monomers (Figure 5a, right), it can be seen that the main difference is the rotation of the N‐ and C segments. This suggests that the variations in the N‐ and C segments may play an important role in virus development and require further in‐depth study. In the rE2 monomer, the N‐terminal domain is a disordered region represented in grey. This region has been previously reported to be poorly resolved in low pH BVDV E2, likely facilitating the hinge‐like conformation movements. Excluding the disordered regions noted above, Domain II (D II) to Domain III (D IIIB) of CSFV rE2 protein showed an overall shape and size similar to that of BVDV E2 proteins (with a root‐mean‐square deviation of 0.847 and 2.407 Å), where D IIIc is in a nonexchangeable structure, a state in which energy is at its lowest and conformational stability is relatively high. The six disulfide bonds on rE2 were presented as intramolecular (Figure 5b). Comparing the monomers of rE2 and BVDV E2, the main difference was that the surface electrostatic potential of the rE2 active site was mostly surrounded by negatively charged patches (Figures 5c and S3). Furthermore, the loop in D III, which is the connection, was not clear here. The flexibility of the linker is probably responsible for the low resolution and twining of the crystals. Although no electron clouds were detected in the N terminus, the linker (GGGGS)3, the loop in DD or the C terminus, they were confirmed by mass spectrometry with 78% coverage of the total sequence. Collectively, these data showed that the structure of the rE2 protein is consistent with the stable head‐to‐tail dimer configuration and provide valuable information for further understanding the E2 protein in the genus Pestivirus.

Figure 5.

Figure 5

Open in a new tab

Structure comparison and glycan analysis. (a) Global superimposition of the rE2‐monomer (red), BVDV E2‐monomer (PDB ID: 2yq2; cyan) and BVDV E2‐monomer (PDB ID: 4JNT; green) illustrates their conformational similarity, with the shape represented as a rod. The dimerization site DD of rE2 rotated 180° counterclockwise to coincide with E2‐monomer (PDB ID: 2YQ2). (b) Disulfide bond of rE2. Disulfide bonds are labelled as green spheres; the cysteines forming the disulfide bonds are C103–C139, C129–C167, C180–C188, C204–C225, C207–C241 and C256–C277. (c) Molecular surfaces are coloured according to their electrostatic surface potential and range from red to blue. (d) Site‐specific glycan analysis of rE2 glycoprotein expressed in rice endosperm (Upper). Schematic diagram of the CSFV rE2 glycoprotein organization (amino acids 358–687). Red, yellow and blue denote Domain regions DI, DII and DIII, respectively. The structural components with N‐linked glycans N542, N586 and N617 are indicated by grey, and conserved disulfide bonds are indicated by blue dashed lines. The yellow ball denotes a conserved and unpaired cysteine (C651) in the Domain‐swapped region (Lower).

N‐glycan modifications of rE2 produced in rice expression system

Glycan modifications are important for protein folding and stability, and different expression platforms enable various levels of glycosylation modifications (Lindner et al., 2015; Nagashima et al., 2018). The N‐glycan compositions of rE2 were treated with trypsin to obtain the N‐glycans present on rE2 protein. The N‐glycans were then purified and analysed using reversed‐phase liquid chromatography tandem mass spectrometry (RPLC‐MS/MS). As a highly glycosylated protein, various N‐glycan compositions of E2 were identified. Three of the most potential N‐linked glycosylation were sited on N542, N586 and N617 with 73%, 18%, and 74% complex/hybrid N‐glycans and 27%, 82%, and 26% high‐mannose N‐glycans, respectively (Figure 5d).

Antigenic stability of ht‐dimer rE2 in a cost‐effective transgenic rice system

Although plant‐derived subunit vaccines have intrinsic advantages over other vaccines in terms of safety and scale‐up, they must meet stability requirements prior to clinical production. As described above, we first screened stably expressed transgenic plants by quantitative (Q) PCR. Then, a high‐yield gene of low‐gluten rice was introduced into E2 homozygous rice parents by hybrid technology and plants with significantly increased yields were selected (Figure S3). The highest expression level of rE2 in the hybrid plants reached 480 mg/kg. Moreover, transgenic plants are stable during manufacture, storage and distribution. In addition to the stability of transgenic rice, we also verified the stability of the purified antigen. After 6 months of storage of purified ht‐dimeric rE2 protein at 2–8 °C, the protein size and the results of animal experiments were consistent with previous results. In conclusion, these data suggest that ht‐dimer expressed in transgenic rice has high antigenic stability and that this is a cost‐effective method for antigen production.

Discussion

Pestiviruses remain a source of major financial and production losses in the livestock industry worldwide, making their eradication an important objective of veterinary research (Newcomer, 2021). Vaccines approved for pestiviruses are mostly live attenuated vaccines. Considering the safety of these vaccines, current vaccine platform technologies are constantly evolving and recombinant envelope proteins are the first choice of antigens for the development vaccines. However, some recombinant proteins are manufactured in mammalian systems, resulting in poor yields, high cost and the need for multiple immunizations; some multivalent antigen vaccines are unstable and subject to incorrect folding (Hobson‐Peters et al., 2019; Urakami et al., 2017). To address these issues, we focussed on designing native‐like and stable antigens—with duplex B‐cell epitopes that drive affinity maturation to elicit potent and long‐lasting protection—that are safe, while offering a high yield and high efficiency at a minimum dose (Rappuoli and Aderem, 2011).

Here, using a flexible linker, the antigen has greater freedom in the direction of the connecting functional domains, forming a stable herringbone configuration with fully exposed paired antigenic epitopes. This is consistent with previous reports that the ideal scaffold would be to present the E dimer as on the virion, assembled into a herringbone pattern (Rey et al., 2018). This structural data allowed us to find a high degree of similarity between ht‐dimer E2 monomer and BVDV E2 monomer (PDB: 4JNT), and the C segment of both proteins rotates back to a lower energy state, suggesting that the flexible linker peptide of ht‐dimer E2 does not disrupt the natural state of the protein. Compared with the design of many vaccines, which present functional epitopes with high structural complexity, the simple design using native proteins better reflects their structure in a natural state, thus increasing the likelihood of recognition by B cells and enhancing their neutralization potential.

A deep understanding of the transition between different envelope glycoprotein states is important for better knowledge of the entry mechanism and provided further insights for a molecular understanding of the immune recognition. During infection by an enveloped virus, there are different conformational states of the viral envelope glycoproteins to mediate receptor binding of the virus with the host cell (Urakami et al., 2017). Structural analysis showed that the CSFV E2 ectodomain likely plays a large role in aggregation via additional hydrophobic interactions or disulfide cross‐linking, as its ectodomain contains conserved and surface‐exposed phenylalanines and tyrosines. These residues are reported to be critical for coreceptor interactions but could readily mediate HCV E1E2 aggregation without transmembrane regions present (Owsianka et al., 2006; Pierce et al., 2016; Tzarum et al., 2018). Previous structural analysis of BVDV E2 also showed that there is a β‐hairpin loop on D IIIc and the C segment of E2 may be responsible for viral fusion (Li et al., 2013). From the analysis of the Flavivirus E protein results, structural domain II rotation is considered as the first step in fusogenic conformational rearrangement. Moreover, a head‐to‐tail dimer structure has been reported to offer greater plasticity and adhesion, and is biologically important for understanding the processes of viral assembly, maturation, and disassembly (Berthet‐Colominas et al., 1999). Therefore, the exposed C segment will be important in further investigating the membrane fusion mechanism of Pestivirus, and even HCV.

Following the design of the ht‐dimer, we tested this immunogen to stimulate robust B‐cell response upon vaccination. Consistent results obtained in mice, rabbits and pigs indicated that the redesigned antigen provided rapid and effective specific antibodies, and conferred complete protection in rabbits and pigs. Three different mammalian experiments showed that the immunogen was species‐independent. To further validate the advantages of the ht‐dimer design, we compared vaccine potency among the ht‐dimer rE2, homodimer E2 and live attenuated virus at low dose. An assay of the dose‐dependent effects on the antibody response showed that ht‐dimeric E2 produced higher levels of E2‐specific antibodies than homodimeric E2 immunization. By comparing ht‐dimer with live attenuated vaccine, we found that an 853 ng dose of ht‐dimer elicited specific antibodies comparable to those elicited by live attenuated vaccine, whereas the neutralizing antibody titre of ht‐dimer was 32 times higher than that of live attenuated vaccine. In addition to eliciting a strong immune response, ht‐dimer was not susceptible to interference by maternal antibodies. Notably, minimum‐dose immunization experiments have previously been performed in a mouse model. However, this study successfully validated that a single immunization with 2.56 μg of ht‐dimer E2 was sufficient to induce protective and stable antibodies in pigs. Enhanced potency and dose‐sparing may be critical for human health and animal productivity, both to satisfy multiple vaccines in one shot and to decrease unnecessary B‐cell exhaustion (Kim et al., 2006).

This report has also shown that plants have huge potential for the production of highly efficient vaccines (Webb et al., 2020). Glycosylation modifications of plant vaccines have been shown to considerably enhance immune responses (Arcalis et al., 2013). Although glycosylation modification is regarded as the maintenance of conformational stability and high antigenicity, there are still concerns that the glycans from plants may cause allergies. Medicago's HA‐based vaccine was reported to be safe, even in volunteers with pre‐existing plant allergies, suggesting that these concerns were largely unwarranted (Ward et al., 2014). Interestingly, different plant polysaccharides are being explored as adjuvants, with several being used already (Rosales‐Mendoza et al., 2016). Although the feasibility of plant‐made CSFV E2 vaccine has been demonstrated, there is still room for improvement in vaccine efficiency and dose used. And challenges that limit the development of traditional plant vaccines are immunogen instability, low yield and transgenic product security. Our vaccine platform addresses these issues by (a) linker design, which improves the stability of the antigen; (b) high efficiency, by designing an ht‐dimer that cross‐links BCRs; (c) minimum dose, a small amount of production can meet high demands for immunity; and (d) lower planting range, which lowers the biosafety risk. In our study, the expression levels of ht‐dimer rice reached 480 mg per kilogram. Since the immunization dose was as low as 284 ng, 1 kg of transgenic rice produced enough antigen to immunize 800 000 piglets. Thus, the immunogen design of the ht‐dimer and the rice endosperm expression system enabled us to obtain a safe and efficient vaccine.

Compared with more conventional systems, plant vaccines are safe, cheap, can be scaled‐up and avoid the risk of spreading animal pathogens or contaminants. The advantage of rice would also be to orally deliver an antigen that could remove the purification or concentration steps (Kashima et al., 2016). However, a major challenge for mucosal vaccines and therapeutics is the need for them to withstand degradation, especially following oral delivery (Kumar et al., 2020)—unless the proteins naturally contain epithelial receptor‐binding ligands or possess the ability to cross using another mechanism, such as using the nontoxic cholera toxin B subunit to facilitate oral delivery of SARS‐CoV‐2 spike protein to the immune system (He et al., 2021; Singh et al., 2023). In addition, another limitation of molecular farming for oral administration is the requirement for high‐dose administration of recombinant proteins from edible plant tissues, which makes it difficult to validate the antigen design. Thus, this trial used injectable immunization to validate the structural design, and our next studies will explore oral immunization in depth.

In summary, we describe a vaccine design with a flexible linker and provide proof that this approach can be used to develop a native‐like, antigenic and immunogenic E2‐homodimer. As a candidate vaccine model, we demonstrated that the redesigned CSFV E2 protein can be efficiently produced in rice endosperm and elicit a rapid and strong immune response. In addition to the CSFV E2 vaccine, this vaccine design approach has also been validated against Newcastle disease virus caused by a paramyxovirus and COVID‐19 caused by a coronavirus. This strategy represents a unique and effective platform for the production of safe, efficient, minimum dose, high‐yield and cost‐effective vaccines, including the Flaviviridae virus vaccines, among others.

Experimental procedures

Design of the ht‐Dimer rE2

Two identical DNA fragments encoding the ectodomain of CSFV E2 (GenBank Acc. No: AAK21202.1; residues 1–342) were linked by (GGGGS)3, a flexible chain that connects the ‘head’ N terminus and ‘tail’ C terminus of the two molecules to form a ht‐dimer rE2, which can be introduced into protein design at the genetic level via a simple step.

Plasmid construction and rice transformation

The ht‐dimer rE2 gene was optimized with a rice codon bias and synthesized by Genscript Biotechnology Inc. (Genscript, Nanjing, China). The synthesized gene was digested by MlyI and XhoI and subcloned into plasmid pOsPMP3 (Healthgen Biotechnology Ltd. Co., Wuhan, China) digested with NaeI/XhoI, resulting in the generation of pOsPMP3‐rE2. To construct the Agrobacterium binary vector, pOsPMP3‐rE2 was digested with HindIII/EcoRI and cloned into the binary vector pCAMBIA1300. The entire expression cassette (containing the Gt13a promoter, its signal peptide, the codon optimized rE2 gene and Nos terminator) was generated. Then, the binary plasmid pCAMBIA1300‐rE2 was finally introduced into Agrobacterium strain EHA105.

The plasmid pCAMBIA‐1300‐rE2 was cotransformed into calli derived from scutellum of the variety TP309 via Agrobacterium‐mediated transformation. After the calli developed into complete plants, PCR was used to identify successful co‐transformants. A forward primer (5′‐CGATTCCGGAAGTGCTTGAC‐3′) from the Gt13a promoter and a reverse primer (5′‐GTTGAGGAGGGTGGTGTTGTACT‐3′) from the rE2 gene were used to identify positive co‐transformations by PCR. The rE2‐positive plants were transplanted and grown in a greenhouse until maturation.

Quantitative PCR analysis

Total RNA was extracted from plant leaves with TRIzol reagent (Invitrogen) and used for reverse transcription‐PCR (RT‐PCR) using a 5× Primescript RT Master Mix (Takara). After cDNA was synthesized, qPCR amplification was used to identify homozygous plants, with a forward primer (5′‐GACCAGCTGCACCTTCAACTA‐3′) and a reverse primer (5′‐AGTACTGGTACTCGCCCTTGAG‐3′) specific to the rE2 gene. The qPCR program was as follows 95 °C for 10 min, followed by 32 cycles of 95 °C for 15 s, 53 °C for 20 s.

Identification of rE2 protein

The total soluble protein from five seeds was extracted using extraction buffer [100 mm Tris–HCl (pH 8.5), 1 mm EDTA] at room temperature (RT) for 2 h. Approximately 5 μg of protein was loaded and separated on a 10% polyacrylamide gel. Western blots were performed as described previously (Lin et al., 2000). The CSFV‐positive swine serum (provided by Henan Provincial Key Laboratory of Animal Immunity, Henan Academy of Agricultural Sciences, Zhengzhou, China) was 500‐fold diluted and used as primary antibody, and goat anti‐swine IgG antibody conjugated to horseradish peroxidase (Jackson Immuno Research) was 5000‐fold diluted and used as secondary antibody. Subsequently, antibody binding was detected with enhanced chemiluminescent (ECL) reagent (NCM Biotech, China).

Next, a CSFV E2 antigen rapid detecting strip (prepared by Henan Provincial Key Laboratory of Animal Immunity, Henan Academy of Agricultural Sciences) was used to identify and screen the transgenic plants expressing rE2 protein. Each sample extract was diluted 1 : 100 with physiological saline, and 100 μL of the sample was added to the sample pool. Nontransgenic rice strain TP309 was used as a negative control. After 10 min, the results of the Test‐line (T) and the Control‐line (C) were recorded. To screen for high‐expression rE2 transgenic plants, the extracted soluble protein was diluted 1 : 100, 1 : 200, 1 : 400, 1 : 800, 1 : 1600, 1 : 3200 and 1 : 6400, and then tested with the strip, as described above. The plants with the maximum dilution were screened as an alternative to the high‐expression transgenic plants.

Protein purification

Rice seeds were ground into powder and homogenized in extraction buffer [25 mm Tris–HCl (pH 8.9), 1 mm EDTA] at a ratio of 1 : 5 (w/v) at RT for 2 h. The crude extract was clarified by centrifugation at 12 000  g for 20 min and passed through a 0.22 μm filtration membrane (Millipore). The clarified extract was loaded onto a Toyopearl Q‐650M (TOSOH) at pH 8.6. The resin was washed with buffer [10 mm Tris–HCl (pH 8.3), 150 mm NaCl, 1 mm EDTA]. The resulting collected fractions were adjusted to pH 7.5 and further purified with Butyl Sepharose High Performance (GE Healthcare) using 1 M ammonium sulfate. The rE2 protein from Butyl Sepharose was then purified through an SP Sepharose High‐Performance column (GE Healthcare), and eluted with PB buffer [10 mm PB (pH 6.5), 50 mm NaCl]. Further purification of rE2 was performed by SEC on a Superdex 200 16/60 column (GE Healthcare) with buffer [10 mm Tris–HCl (pH 7.5), 25 mm NaCl]. The purified protein was verified by non‐reducing SDS‐PAGE and native PAGE.

Recombinant rE2 quantification assay

Crude extracts of rE2 protein (200 μL per seed) were obtained using extraction buffer [100 mm Tris–HCl (pH 8.5), 1 mm EDTA]. The protein concentration was measured with the BCR protein assay (Pierce). Purified rE2 protein was measured with the BCA protein assay kit (Pierce) and a standard curve was plotted with different dilutions of rE2 protein using a direct ELISA kit (CSFV Ag Elisa kit, Median, Korea). To determine the expression levels, the crude protein from homozygous and hybrid plants was diluted at ratios of 1 : 200, 1 : 400, 1 : 800 and 1 : 1600 and evaluated using a direct ELISA kit.

Analytical ultracentrifugation (AUC) analysis

A sedimentation velocity (SV) experiment was carried out on a Beckmann XL‐A analytical ultracentrifuge using a rotor speed of 1 29 000 g at 20 °C. Protein samples were studied at a concentration of 1.1 mg/mL in buffer [10 mm Tris–HCl (pH 7.5), 25 mm NaCl]. Sample solution (380 μL) and reference solution (400 μL) were loaded. A total of 444 scans were collected. Equilibrium sedimentation and data analysis was performed in the preliminary experiments (Chaudhry et al., 2009; Oliva et al., 2010). SV data were analysed using the SEDFIT program. The molecular masses calculated from the amino acid sequences were 63.6 and 165 kDa.

Mouse immunizations

Balb/c female mice (6 weeks of age; Zhengzhou University Animal Center) were divided randomly into four groups (n = 5 in each group) and vaccinated with a prime immunization, then 28 days later boosted with a second vaccination. Immunogen rE2 was mixed 1 : 1 (w/v) with ISA 50 V2 adjuvant (SEPPIC) and mice were injected intramuscularly with 5 μg protein (200 μL). The positive and the negative groups were also injected with 200 μL of commercial C‐strain live attenuated vaccine and non‐transgenic plant protein (TP309) with 50 V adjuvant, respectively. Whole blood (10 μL) was collected at 7‐day intervals and added to 190 μL of physiological saline. Serum was stored at −20 °C and tested using a Classic Swine Fever virus Antibody Test Kit (IDEXX, Mine).

Rabbit immunizations and challenge

New Zealand White rabbits (n = 4 in each group) at 3 months of age were purchased. After a 7‐day acclimation time, rabbits were vaccinated once, and then 21 days later, the rabbits were challenged. Prior to inoculation, immunogen rE2 was mixed 1 : 1 (w/v) with ISA 50 V adjuvant and rabbits were injected intramuscularly with 20 μg protein (1 mL). The positive and the negative groups were also injected with 1 mL of commercial C‐strain live attenuated vaccine and TP309 with 50 V adjuvant, respectively. Blood was collected at 7‐day intervals, and serum was separated and stored at −20 °C until use. Three weeks of postprime, all rabbits were challenged intravenously with a 15 000 50% rabbit infectious dose (RID50) of the commercial C‐strain live attenuated vaccine for the evaluation of vaccine efficacy. After infection, body temperature was monitored and recorded every 8 h until 3 days post‐infection. Fever was considered as a rectal temperature ≥40 °C that lasted for at least 18 h.

Piglet immunizations and challenge

Fifteen piglets (a mix of males and females, 30 days of age) were divided randomly into three groups, five pigs per group, and vaccinated with a prime immunization and 28 days later with a second vaccination. Immunogen rE2 was mixed 1 : 1 (w/v) with ISA 50 V adjuvant and each piglet was immunized intramuscularly with 30 μg of rE2 protein (1 mL). The positive and the negative groups were also injected with 1 mL of commercial C‐strain live attenuated vaccine and TP309 with 50 V adjuvant, respectively. Sera were collected at 7‐day intervals as previously described. At 28 days of postboost, each pig was challenged intramuscularly via the neck with 106 of the tissue culture infective dose (TCID50) of the highly virulent CSFV Shimen strain in a volume of 1 mL. After infection, clinical signs were observed, rectal temperatures and body weight were measured daily. Blood samples were collected at 0, 3, 6, 9, 12, 15, and 18 days of postchallenge. All surviving pigs were euthanized at 21 days of postchallenge, and lymph nodes, ceca, spleens and tonsils were observed and collected for H&E analysis. Additionally, 15 piglets (30 days of age) with maternal antibodies were divided randomly into three groups, five pigs per group, and vaccinated with a prime immunization and 28 days later with a second vaccination. Immunogen rE2 was mixed 1 : 1 (w/v) with ISA 50 V adjuvant and each piglet was immunized intramuscularly with 30 μg of rE2 protein (1 mL). The positive and the negative groups were also injected with 1 mL of commercial C‐strain live attenuated vaccine and TP309 with 50 V adjuvant, respectively. Sera were collected at 7‐day intervals to test antibodies levels.

The immunizations in piglets were administered as split doses at nanoscale amounts to examine vaccine potency. The 30‐day‐old piglets, five per dosing group, were immunized intramuscularly twice with an interval of 28 days. Immunogen rE2 was mixed 1 : 1 (w/v) with ISA 50 V adjuvant and immunizations consisted of four dose groups: 284, 853 ng, 2.56 and 5.12 μg of rE2 protein. Positive controls included two groups, subunit vaccine and live attenuated vaccine, where the subunit vaccine was also divided into four dose groups (284, 853 ng, 2.56 and 5.12 μg of subunit vaccine). The live attenuated vaccine group and the negative group were injected with 1 mL of commercial C‐strain live attenuated vaccine and TP309 with 50 V adjuvant, respectively. Serum were collected and stored as previously described.

A single‐dose immune test was conducted. For this, 21‐day‐old piglets, eight per dosing group, were immunized intramuscularly once with 2.56 and 5.12 μg of rE2 protein. The negative control (TP309) was the same as that described above. Sera were collected for 49 days.

Serum neutralization assays

The neutralization activity of the serum was measured in a microneutralization assay as previously described (Hu et al., 2013). PK‐15 cells were used to propagate CSFV from the Shimen strain in 96‐well plates. Serum samples were heat‐inactivated at 56 °C for 30 min, then a 50‐μL volume of serum was diluted in 1640 medium with 10% FBS (Hyclone) by serial twofold dilutions (from 2−1 to 2−20). Next, 50 μL of the 200 median TCID50 of the virulent strain was added to the serial dilution and incubated at 37 °C for 1 h, and 100 μL of the sera/virus mixture was then added to the prepared PK‐15 cells in 96‐well plates. Following a 40‐h incubation, the cells were fixed with 4% paraformaldehyde at −20 °C for 30 min and incubated with CSFV‐positive serum (1 : 400) as the primary antibody and HRP‐goat anti‐swine antibody (1 : 5000) as the secondary antibody. Antibody binding was detected with ECL reagent (NCM Biotech) and visualized under the light microscope. The neutralizing titre was determined by the highest dilution of serum that inhibited virus infection and was expressed as 2x.

Enzyme‐linked immunosorbent assay (ELISA)

Serum samples were collected and tested by a CSFV antibody test kit (IDEXX). The 96‐well microplate in the commercial kit was coated with the CSFV E2 protein. After being incubated with serum for 2 h at RT, the plates were washed three times with phosphate‐buffered saline‐Tween 20 (PBS containing 0.05% Tween‐20, PBST) and incubated with HRP‐conjugated anti‐swine antibody for 30 min at RT. After washing three times with PBST, the absorbance of the serum samples was read at 450 nm using a POLARstar Omega microplate reader (BMG, Germany). When the calculated S/P ratio was over 40%, the results were determined to be seropositive.

Serum concentrations of IL‐2 (Mlbio, China), IL4 (Invitrogen), IFN‐γ (BD Bioscience), IgG 2a (MEIMIAN, China) and IgM (MEIMIAN) were determined in duplicate with commercially available ELISA kits for the detection of porcine cytokines.

Crystallization, data collection and structure determination

Crystals of rE2 were concentrated to 7.3 mg/mL and grown by sitting drop vapour diffusion in drops containing protein added to the mother liquor at a 1 : 1 ratio with a well solution comprising 0.1 M magnesium formate and 15% (w/v) PEG 3350 at 20 °C. Diffraction data were recorded with the BL19U2 beamline at the National Center for Protein Sciences Shanghai (NCPSS) and processed with HKL2000 (Otwinowski and Minor, 1997). The structure was determined with the program PHENIX AUTOSOL by multiwavelength anomalous dispersion (MAD) phasing using crystals. Manual building was performed with the program COOT (Emsley and Cowtan, 2004). Because of the low resolution (4.4 Å) of the data, refinement was limited to rigid‐body and restrained refinements. Final model geometry was checked with MolProbity. For data collection and refinement statistics, see Table S3.

Small‐angle X‐ray scattering (SAXS)

SAXS data for rE2 were collected on the SIBYLS Beamline 12.3.1 at the Advanced Light Source (Shanghai Synchrotron) using protein samples at concentrations between 1 and 5 mg/mL. All data sets were measured at three exposure times (0.5, 1 and 6 s) at 283 K. Multiple curves with different concentrations and different exposure times were scaled and merged to generate an ideal average scattering curve. Data handing and reduction were performed as described previously. The quality of the scattering curves was analysed using the program PRIMUS to ensure that there was no obvious aggregation or radiation damage before further analysis. The initial R g values were calculated based on the Guinier plots. Only data from low q values were used for the calculation. The P(r) distribution function was calculated with the program GNOM. The molecular mass was estimated directly from the SAXS MoW server using the P(r) distribution function. The low‐resolution rod shape of rE2 protein in solution was modelled by the program DAMMIF in the asymmetric unit and P2 symmetry using both the original scattering curve and the calculated P(r) distribution curve. Using the program DAMAVER, 20 continuous and meaningful individual DAMMIF calculations were aligned, combined, and filtered to generate a final mode (Otwinowski and Minor, 1997). The crystal structure of rE2 was docked into the resulting ‘most probable’ envelope, and the outputs were displayed and manually refined with the UCSF Chimera software.

Intact N‐glycopeptide analysis

To confirm the N‐glycan compositions in the rE2 protein, the purified rE2 (1 mg) was digested with trypsin (enzyme:substrate w/w, 1 : 50) at 37 °C for at least 16 h. Then, site‐specific N‐glycosylation profiling of rE2 was determined by RPLC‐MS/MS analysis, as described previously (Bi et al., 2022; Qin et al., 2022). For graphical illustration of N‐glycan differences between samples, the spectra were processed using PEAKS Studio 8.5 software (Bioinformatics Solutions Inc., Waterloo, Canada) and evaluated manually. The results of N‐glycosylation sites, characteristic peptides and complete N‐glycoproteins obtained from mass spectrometry data matching by GPSkeeker and pGlyco 3.0 database. The relative abundances (r.a.) of individual glycans were calculated from their normalized intensities using Microsoft Excel.

Data analysis

Graphical and statistical analyses were performed using GraphPad Prism 5.04 (GraphPad Software Inc., La Jolla, CA). Data are represented as the mean and standard error of the mean (SEM) quoted to indicate the uncertainty around the estimate of the group mean. The differences between the mean values were analysed by one‐way analysis of variance (ANOVA) followed by Bonferroni post‐test, and a P‐value of 0.05 was considered significant. All structural renderings of proteins were generated using the UCSF Chimera and PyMOL Molecular Graphics System.

Author contributions

G.Z. and E.Z. were involved in conceptualization. Q.X., F.M., Q.L., L.Y., J.O., Y.L., Q.Z. and J.Y. were involved in methodology. Q.X., X.L., S.Z. and S.C. were involved in investigation. Q.X., R.L., Q.W., Y.W., Y.D. and L.W. were involved in visualization. D.Y. was involved on supervision. Q.X. and F.M. were involved in writing—original draft. L.Z., H.H. and G.Z. were involved in writing—review & editing.

Conflict of interest

All authors declare no competing interests.

Ethical approval

Animal experiments were performed under licences granted by the Henan Academy of Agricultural Sciences (Approval Number SYXK 2014‐0007) and followed the animal welfare guidelines of the Institutional Animal Care and Use Committee (IACUC).

Accession numbers

Coordinates and structure of the rE2 homodimer have been deposited in the Protein Data Bank under accession number 7EGY. All other data are available from the authors upon reasonable request.

Supporting information

Figure S1 Immune protection induced in mice and rabbits.

Figure S2 SAXS data of ht‐dimer rE2.

Figure S3 Electrostatic‐potential surface maps of CSFV rE2 monomer and BVDV E2 monomer.

Figure S4 Enhancing the expression of rE2 protein in transgenic rice.

PBI-21-2546-s002.docx (402.1KB, docx)

Table S1 Parameters of sedimentation velocity analysis of rE2.

Table S2 SAXS parameters for rE2.

Table S3 Data collection and refinement statistics.

PBI-21-2546-s001.docx (32.1KB, docx)

Acknowledgement

We are grateful to Professor Hongde Liang of Henan Agricultural University for his help in observing the pathological sections. We also acknowledge our debt to the Key Laboratory of Animal Immunology, Henan Academy of Agricultural Sciences for financial support, and to the staff of the BL18U and BL19U2 beamline at the National Center for Protein Sciences Shanghai (NCPSS) and Shanghai Synchrotron Radiation Facility (SSRF), Shanghai, People's Republic of China, for their assistance during data collection. This study was funded by the Key Projects of Science and Technology of Henan Province (Grant no. 192102110007) and the Key Projects of Science and Technology of Henan Province (Grant no. 221100110600). We thank International Science Editing (http://www.internationalscienceediting.com) for editing this manuscript. All completed graphics were created with BioRender.com.

[Correction added on 23 May 2024, after first online publication: the affiliations 1 and 3 are swapped and updated the author byline in this version.]

Contributor Information

Erqin Zhang, Email: zhangerqin76@163.com.

Gaiping Zhang, Email: zhanggaiping@pku.edu.cn.

References

  1. Arcalis, E. , Stadlmann, J. , Rademacher, T. , Marcel, S. , Sack, M. , Altmann, F. and Stoger, E. (2013) Plant species and organ influence the structure and subcellular localization of recombinant glycoproteins. Plant Mol. Biol. 83, 105–117. [DOI] [PubMed] [Google Scholar]
  2. Bardina, S.V. , Bunduc, P. , Tripathi, S. , Duehr, J. , Frere, J.J. , Brown, J.A. , Nachbagauer, R. et al. (2017) Enhancement of Zika virus pathogenesis by preexisting antiflavivirus immunity. Science, 356, 175–180. [DOI] [PMC free article] [PubMed] [Google Scholar]
  3. Becher, P. , Orlich, M. , Kosmidou, A. , Konig, M. , Baroth, M. and Thiel, H.J. (1999) Genetic diversity of pestiviruses: identification of novel groups and implications for classification. Virology, 262, 64–71. [DOI] [PubMed] [Google Scholar]
  4. Berthet‐Colominas, C. , Monaco, S. , Novelli, A. , Sibai, G. , Mallet, F. and Cusack, S. (1999) Head‐to‐tail dimers and interdomain flexibility revealed by the crystal structure of HIV‐1 capsid protein (p24) complexed with a monoclonal antibody Fab. EMBO J. 18, 1124–1136. [DOI] [PMC free article] [PubMed] [Google Scholar]
  5. Bi, M. , Bai, B. and Tian, Z. (2022) Structure‐specific N‐glycoproteomics characterization of NIST monoclonal antibody reference material 8671. J. Proteome Res. 21, 1276–1284. [DOI] [PubMed] [Google Scholar]
  6. Blome, S. , Staubach, C. , Henke, J. , Carlson, J. and Beer, M. (2017) Classical swine fever—an updated review. Viruses, 9 4, 86. [DOI] [PMC free article] [PubMed] [Google Scholar]
  7. Chaudhry, C. , Weston, M.C. , Schuck, P. , Rosenmund, C. and Mayer, M.L. (2009) Stability of ligand‐binding domain dimer assembly controls kainate receptor desensitization. EMBO J. 28, 1518–1530. [DOI] [PMC free article] [PubMed] [Google Scholar]
  8. Clarke, J.L. , Paruch, L. , Dobrica, M.O. , Caras, I. , Tucureanu, C. , Onu, A. , Ciulean, S. et al. (2017) Lettuce‐produced hepatitis C virus E1E2 heterodimer triggers immune responses in mice and antibody production after oral vaccination. Plant Biotechnol. J. 15, 1611–1621. [DOI] [PMC free article] [PubMed] [Google Scholar]
  9. Corbett, K.S. , Edwards, D.K. , Leist, S.R. , Abiona, O.M. , Boyoglu‐Barnum, S. , Gillespie, R.A. , Himansu, S. et al. (2020) SARS‐CoV‐2 mRNA vaccine design enabled by prototype pathogen preparedness. Nature, 586, 567–571. [DOI] [PMC free article] [PubMed] [Google Scholar]
  10. Desai, P.N. , Shrivastava, N. and Padh, H. (2010) Production of heterologous proteins in plants: strategies for optimal expression. Biotechnol. Adv. 28, 427–435. [DOI] [PubMed] [Google Scholar]
  11. Dobrica, M.O. , van Eerde, A. , Tucureanu, C. , Onu, A. , Paruch, L. , Caras, I. , Vlase, E. et al. (2021) Hepatitis C virus E2 envelope glycoprotein produced in Nicotiana benthamiana triggers humoral response with virus‐neutralizing activity in vaccinated mice. Plant Biotechnol. J. 19, 2027–2039. [DOI] [PMC free article] [PubMed] [Google Scholar]
  12. Emsley, P. and Cowtan, K. (2004) Coot: model‐building tools for molecular graphics. Acta Crystallogr. D Biol. Crystallogr. 60, 2126–2132. [DOI] [PubMed] [Google Scholar]
  13. Gunter, C.J. , Regnard, G.L. , Rybicki, E.P. and Hitzeroth, I.I. (2019) Immunogenicity of plant‐produced porcine circovirus‐like particles in mice. Plant Biotechnol. J. 17, 1751–1759. [DOI] [PMC free article] [PubMed] [Google Scholar]
  14. He, Y. , Ning, T. , Xie, T. , Qiu, Q. , Zhang, L. , Sun, Y. , Jiang, D. et al. (2011) Large‐scale production of functional human serum albumin from transgenic rice seeds. Proc. Natl. Acad. Sci. U. S. A. 108, 19078–19083. [DOI] [PMC free article] [PubMed] [Google Scholar]
  15. He, W. , Baysal, C. , Lobato Gomez, M. , Huang, X. , Alvarez, D. , Zhu, C. , Armario‐Najera, V. et al. (2021) Contributions of the international plant science community to the fight against infectious diseases in humans‐part 2: affordable drugs in edible plants for endemic and re‐emerging diseases. Plant Biotechnol. J. 19, 1921–1936. [DOI] [PMC free article] [PubMed] [Google Scholar]
  16. Hobson‐Peters, J. , Harrison, J.J. , Watterson, D. , Hazlewood, J.E. , Vet, L.J. , Newton, N.D. , Warrilow, D. et al. (2019) A recombinant platform for flavivirus vaccines and diagnostics using chimeras of a new insect‐specific virus. Sci. Transl. Med. 11, 522. [DOI] [PubMed] [Google Scholar]
  17. Holla, P. , Ambegaonkar, A. , Sohn, H. and Pierce, S.K. (2019) Exhaustion may not be in the human B cell vocabulary, at least not in malaria. Immunol. Rev. 292, 139–148. [DOI] [PubMed] [Google Scholar]
  18. Hu, W. , Chen, A. , Miao, Y. , Xia, S. , Ling, Z. , Xu, K. , Wang, T. et al. (2013) Fully human broadly neutralizing monoclonal antibodies against influenza A viruses generated from the memory B cells of a 2009 pandemic H1N1 influenza vaccine recipient. Virology, 435, 320–328. [DOI] [PMC free article] [PubMed] [Google Scholar]
  19. Jung, M. , Shin, Y. , Kim, J. , Cha, S.B. , Lee, W.J. , Shin, M.K. , Shin, S. et al. (2014) Induction of immune responses in mice and pigs by oral administration of classical swine fever virus E2 protein expressed in rice calli. Arch. Virol. 159, 3219–3230. [DOI] [PubMed] [Google Scholar]
  20. Kashima, K. , Yuki, Y. , Mejima, M. , Kurokawa, S. , Suzuki, Y. , Minakawa, S. , Takeyama, N. et al. (2016) Good manufacturing practices production of a purification‐free oral cholera vaccine expressed in transgenic rice plants. Plant Cell Rep. 35, 667–679. [DOI] [PubMed] [Google Scholar]
  21. Kim, Y.M. , Pan, J.Y. , Korbel, G.A. , Peperzak, V. , Boes, M. and Ploegh, H.L. (2006) Monovalent ligation of the B cell receptor induces receptor activation but fails to promote antigen presentation. Proc. Natl. Acad. Sci. U.S.A. 103, 3327–3332. [DOI] [PMC free article] [PubMed] [Google Scholar]
  22. Kim, M.‐Y. , Van Dolleweerd, C. , Copland, A. , Paul, M.J. , Hofmann, S. , Webster, G.R. , Julik, E. et al. (2017) Molecular engineering and plant expression of an immunoglobulin heavy chain scaffold for delivery of a dengue vaccine candidate. Plant Biotechnol. J. 15, 1590–1601. [DOI] [PMC free article] [PubMed] [Google Scholar]
  23. Kong, L. , Jackson, K.N. , Wilson, I.A. and Law, M. (2015) Capitalizing on knowledge of hepatitis C virus neutralizing epitopes for rational vaccine design. Curr. Opin. Virol. 11, 148–157. [DOI] [PMC free article] [PubMed] [Google Scholar]
  24. Krol, E. , Brzuska, G. and Szewczyk, B. (2019) Production and biomedical application of flavivirus‐like particles. Trends Biotechnol. 37, 1202–1216. [DOI] [PubMed] [Google Scholar]
  25. Kumar, S.R.P. , Wang, X. , Avuthu, N. , Bertolini, T.B. , Terhorst, C. , Guda, C. , Daniell, H. et al. (2020) Role of small intestine and gut microbiome in plant‐based oral tolerance for hemophilia. Front. Immunol. 11, 844. [DOI] [PMC free article] [PubMed] [Google Scholar]
  26. Laughlin, R.C. , Madera, R. , Peres, Y. , Berquist, B.R. , Wang, L. , Buist, S. , Burakova, Y. et al. (2019) Plant‐made E2 glycoprotein single‐dose vaccine protects pigs against classical swine fever. Plant Biotechnol. J. 17, 410–420. [DOI] [PMC free article] [PubMed] [Google Scholar]
  27. Legocki, A.B. , Miedzinska, K. , Czaplinska, M. , Plucieniczak, A. and Wedrychowicz, H. (2005) Immunoprotective properties of transgenic plants expressing E2 glycoprotein from CSFV and cysteine protease from Fasciola hepatica. Vaccine, 23, 1844–1846. [DOI] [PubMed] [Google Scholar]
  28. Leyssen, P. , De Clercq, E. and Neyts, J. (2000) Perspectives for the treatment of infections with Flaviviridae. Clin. Microbiol. Rev. 13, 67–82 table of contents. [DOI] [PMC free article] [PubMed] [Google Scholar]
  29. Li, Y. , Wang, J. , Kanai, R. and Modis, Y. (2013) Crystal structure of glycoprotein E2 from bovine viral diarrhea virus. Proc. Natl. Acad. Sci. 110, 6805–6810. [DOI] [PMC free article] [PubMed] [Google Scholar]
  30. Lin, M. , Lin, F. , Mallory, M. and Clavijo, A. (2000) Deletions of structural glycoprotein E2 of classical swine fever virus strain alfort/187 resolve a linear epitope of monoclonal antibody WH303 and the minimal N‐terminal domain essential for binding immunoglobulin G antibodies of a pig hyperimmune serum. J. Virol. 74, 11619–11625. [DOI] [PMC free article] [PubMed] [Google Scholar]
  31. Lin, H.H. , Yip, B.S. , Huang, L.M. and Wu, S.C. (2018) Zika virus structural biology and progress in vaccine development. Biotechnol. Adv. 36, 47–53. [DOI] [PubMed] [Google Scholar]
  32. Lindner, H. , Kessler, S.A. , Muller, L.M. , Shimosato‐Asano, H. , Boisson‐Dernier, A. and Grossniklaus, U. (2015) TURAN and EVAN mediate pollen tube reception in Arabidopsis Synergids through protein glycosylation. PLoS Biol. 13, e1002139. [DOI] [PMC free article] [PubMed] [Google Scholar]
  33. Margolin, E. , Chapman, R. , Williamson, A.L. , Rybicki, E.P. and Meyers, A.E. (2018) Production of complex viral glycoproteins in plants as vaccine immunogens. Plant Biotechnol. J. 16, 1531–1545. [DOI] [PMC free article] [PubMed] [Google Scholar]
  34. Mason, H.S. , Lam, D.M. and Arntzen, C.J. (1992) Expression of hepatitis B surface antigen in transgenic plants. Proc. Natl. Acad. Sci. U.S.A. 89, 11745–11749. [DOI] [PMC free article] [PubMed] [Google Scholar]
  35. Meola, A. , Tarr, A.W. , England, P. , Meredith, L.W. , McClure, C.P. , Foung, S.K. , McKeating, J.A. et al. (2015) Structural flexibility of a conserved antigenic region in hepatitis C virus glycoprotein E2 recognized by broadly neutralizing antibodies. J. Virol. 89, 2170–2181. [DOI] [PMC free article] [PubMed] [Google Scholar]
  36. Nagashima, Y. , von Schaewen, A. and Koiwa, H. (2018) Function of N‐glycosylation in plants. Plant Sci. 274, 70–79. [DOI] [PubMed] [Google Scholar]
  37. Newcomer, B.W. (2021) 75 years of bovine viral diarrhea virus: current status and future applications of the use of directed antivirals. Antiviral Res. 196, 105205. [DOI] [PubMed] [Google Scholar]
  38. Oliva, M.A. , Halbedel, S. , Freund, S.M. , Dutow, P. , Leonard, T.A. , Veprintsev, D.B. , Hamoen, L.W. et al. (2010) Features critical for membrane binding revealed by DivIVA crystal structure. EMBO J. 29, 1988–2001. [DOI] [PMC free article] [PubMed] [Google Scholar]
  39. Otwinowski, Z. and Minor, W. (1997) Processing of X‐ray diffraction data collected in oscillation mode. Methods Enzymol. 276, 307–326. [DOI] [PubMed] [Google Scholar]
  40. Owsianka, A.M. , Timms, J.M. , Tarr, A.W. , Brown, R.J. , Hickling, T.P. , Szwejk, A. , Bienkowska‐Szewczyk, K. et al. (2006) Identification of conserved residues in the E2 envelope glycoprotein of the hepatitis C virus that are critical for CD81 binding. J. Virol. 80, 8695–8704. [DOI] [PMC free article] [PubMed] [Google Scholar]
  41. Park, Y. , An, D.J. , Choe, S. , Lee, Y. , Park, M. , Park, S. , Gu, S. et al. (2019) Development of recombinant protein‐based vaccine against classical swine fever virus in pigs using transgenic Nicotiana benthamiana . Front. Plant Sci. 10, 624. [DOI] [PMC free article] [PubMed] [Google Scholar]
  42. Park, Y. , Lee, S. , Kang, H. , Park, M. , Min, K. , Kim, N.H. , Gu, S. et al. (2020) A classical swine fever virus E2 fusion protein produced in plants elicits a neutralizing humoral immune response in mice and pigs. Biotechnol. Lett. 42, 1247–1261. [DOI] [PMC free article] [PubMed] [Google Scholar]
  43. Park, Y. , Oh, Y. , Wang, M. , Ganges, L. , Bohorquez, J.A. , Park, S. , Gu, S. et al. (2021) A novel E2 glycoprotein subunit marker vaccine produced in plant is able to prevent classical swine fever virus vertical transmission after double vaccination. Vaccines, 9, 5. [DOI] [PMC free article] [PubMed] [Google Scholar]
  44. Pierce, B.G. , Keck, Z.Y. and Foung, S.K. (2016) Viral evasion and challenges of hepatitis C virus vaccine development. Curr. Opin. Virol. 20, 55–63. [DOI] [PMC free article] [PubMed] [Google Scholar]
  45. Postel, A. , Austermann‐Busch, S. , Petrov, A. , Moennig, V. and Becher, P. (2018) Epidemiology, diagnosis and control of classical swine fever: recent developments and future challenges. Transbound. Emerg. Dis. 65(Suppl 1), 248–261. [DOI] [PubMed] [Google Scholar]
  46. Qin, S. , Qin, S. and Tian, Z. (2022) Comprehensive site‐ and structure‐specific characterization of N‐glycosylation in model plant Arabidopsis using mass‐spectrometry‐based N‐glycoproteomics. J. Chromatogr. B Analyt. Technol. Biomed. Life Sci. 1198, 123234. [DOI] [PubMed] [Google Scholar]
  47. Rappuoli, R. and Aderem, A. (2011) A 2020 vision for vaccines against HIV, tuberculosis and malaria. Nature, 473, 463–469. [DOI] [PubMed] [Google Scholar]
  48. Rey, F.A. , Stiasny, K. , Vaney, M.C. , Dellarole, M. and Heinz, F.X. (2018) The bright and the dark side of human antibody responses to flaviviruses: lessons for vaccine design. EMBO Rep. 19, 206–224. [DOI] [PMC free article] [PubMed] [Google Scholar]
  49. Rosales‐Mendoza, S. , Salazar‐Gonzalez, J.A. , Decker, E.L. and Reski, R. (2016) Implications of plant glycans in the development of innovative vaccines. Expert Rev. Vaccines, 15, 915–925. [DOI] [PubMed] [Google Scholar]
  50. Ruwona, T.B. , Giang, E. , Nieusma, T. and Law, M. (2014) Fine mapping of murine antibody responses to immunization with a novel soluble form of hepatitis C virus envelope glycoprotein complex. J. Virol. 88, 10459–10471. [DOI] [PMC free article] [PubMed] [Google Scholar]
  51. Singh, R. , Lin, S. , Nair, S.K. , Shi, Y. and Daniell, H. (2023) Oral booster vaccine antigen‐Expression of full‐length native SARS‐CoV‐2 spike protein in lettuce chloroplasts. Plant Biotechnol. J. 21, 887–889. [DOI] [PMC free article] [PubMed] [Google Scholar]
  52. Slon‐Campos, J.L. , Dejnirattisai, W. , Jagger, B.W. , Lopez‐Camacho, C. , Wongwiwat, W. , Durnell, L.A. , Winkler, E.S. et al. (2019) A protective Zika virus E‐dimer‐based subunit vaccine engineered to abrogate antibody‐dependent enhancement of dengue infection. Nat. Immunol. 20, 1291–1298. [DOI] [PMC free article] [PubMed] [Google Scholar]
  53. Smith, T. , O'Kennedy, M.M. , Wandrag, D.B.R. , Adeyemi, M. and Abolnik, C. (2020) Efficacy of a plant‐produced virus‐like particle vaccine in chickens challenged with Influenza A H6N2 virus. Plant Biotechnol. J. 18, 502–512. [DOI] [PMC free article] [PubMed] [Google Scholar]
  54. Strauch, E.M. , Bernard, S.M. , La, D. , Bohn, A.J. , Lee, P.S. , Anderson, C.E. , Nieusma, T. et al. (2017) Computational design of trimeric influenza‐neutralizing proteins targeting the hemagglutinin receptor binding site. Nat. Biotechnol. 35, 667–671. [DOI] [PMC free article] [PubMed] [Google Scholar]
  55. Tian, J.H. , Patel, N. , Haupt, R. , Zhou, H. , Weston, S. , Hammond, H. , Logue, J. et al. (2021) SARS‐CoV‐2 spike glycoprotein vaccine candidate NVX‐CoV2373 immunogenicity in baboons and protection in mice. Nat. Commun. 12, 372. [DOI] [PMC free article] [PubMed] [Google Scholar]
  56. Tzarum, N. , Wilson, I.A. and Law, M. (2018) The neutralizing face of hepatitis C virus E2 envelope glycoprotein. Front. Immunol. 9, 1315. [DOI] [PMC free article] [PubMed] [Google Scholar]
  57. Urakami, A. , Ngwe Tun, M.M. , Moi, M.L. , Sakurai, A. , Ishikawa, M. , Kuno, S. , Ueno, R. et al. (2017) An envelope‐modified tetravalent dengue virus‐like‐particle vaccine has implications for flavivirus vaccine design. J. Virol. 91(23), e01181–17. [DOI] [PMC free article] [PubMed] [Google Scholar]
  58. Wang, X. , Jiang, D. and Yang, D. (2015) Fast‐tracking determination of homozygous transgenic lines and transgene stacking using a reliable quantitative real‐time PCR assay. Appl. Biochem. Biotechnol. 175, 996–1006. [DOI] [PubMed] [Google Scholar]
  59. Ward, B.J. , Landry, N. , Trepanier, S. , Mercier, G. , Dargis, M. , Couture, M. , D'Aoust, M.A. et al. (2014) Human antibody response to N‐glycans present on plant‐made influenza virus‐like particle (VLP) vaccines. Vaccine, 32, 6098–6106. [DOI] [PubMed] [Google Scholar]
  60. Webb, S.R. , Twyman, R.M. and Moloney, M. (2020) Agtech infrastructure for pandemic preparedness. Nat. Biotechnol. 38, 1025–1027. [DOI] [PubMed] [Google Scholar]
  61. Wu, F. , Luo, S. , Zhang, Y. , Ou, Y. , Wang, H. , Guo, Z. , He, C. et al. (2022) Single‐shot AAV‐vectored vaccine against SARS‐CoV‐2 with fast and long‐lasting immunity. Acta Pharm Sin. B, 13, 2219–2233. [DOI] [PMC free article] [PubMed] [Google Scholar]
  62. Yiu, J.C. , Liu, C.W. , Su, R.Y. , Lai, W.J. , Tseng, M.J. and Ou, S.J. (2013) Immunogenicity of a lettuce‐derived vaccine candidate expressing the E2 protein against classical swine fever virus. Plant Cell Tissue Org. 113, 483–490. [Google Scholar]
  63. Zhang, X. , Ge, P. , Yu, X. , Brannan, J.M. , Bi, G. , Zhang, Q. , Schein, S. et al. (2013) Cryo‐EM structure of the mature dengue virus at 3.5‐A resolution. Nat. Struct. Mol. Biol. 20, 105–110. [DOI] [PMC free article] [PubMed] [Google Scholar]

Associated Data

This section collects any data citations, data availability statements, or supplementary materials included in this article.

Supplementary Materials

Figure S1 Immune protection induced in mice and rabbits.

Figure S2 SAXS data of ht‐dimer rE2.

Figure S3 Electrostatic‐potential surface maps of CSFV rE2 monomer and BVDV E2 monomer.

Figure S4 Enhancing the expression of rE2 protein in transgenic rice.

PBI-21-2546-s002.docx (402.1KB, docx)

Table S1 Parameters of sedimentation velocity analysis of rE2.

Table S2 SAXS parameters for rE2.

Table S3 Data collection and refinement statistics.

PBI-21-2546-s001.docx (32.1KB, docx)

Articles from Plant Biotechnology Journal are provided here courtesy of Society for Experimental Biology (SEB) and the Association of Applied Biologists (AAB) and John Wiley and Sons, Ltd

RESOURCES