The COE nucleotide sequence encoding for amino acid 499–638 of the COE in the S protein of the attenuated PEDV DR13 strain (NCBI accession number JQ023161.1) was synthesized, codon optimized commercially in tobacco (Genebank accenssion number BankIt2361779 Optimized MT761690), and then inserted in a pEZ cloning vector (POCH, Life Science Missouri City, Texas 77489). The COE gene was amplified from the vector, and replaced for the H5 sequence present in pRTRA-CaMV35S-H5-GCN4pII-cmyc-his-KDEL (22) to via the BamHI and Bsp120I sites. The resulting expression cassette (Figure 1A) was inserted into the expression vector pCB301-Kan (27) via HindIII cleavage, then transformed into the Agrobacterium tumefaciens [pGV2260 in C58C1; (28)] strain via electroporation at 2.5 kV, 25 μF capacitance, and 400 Ohm resistance.

To express the recombinant protein in planta, the agro-infiltration protocol was performed as described by Pham et al. (29), with some modifications. Briefly, bacteria containing the expression vector pCB301-COE-GCN4pII-cmyc-his-KDEL and plant vector including HcPro (23), that was used as a gene silencing suppressor for enhancing the expression levels of targeting proteins in plants (30, 31), were mixed and diluted in an infiltration buffer [10 mM 2-(N-morpholino) ethanesulfonic acid (MES), 10 mM MgSO₄, pH 5.6]. N. benthamiana plants (5 weeks old) were completely infiltrated in an Agrobacterium solution, and maintained in a greenhouse. Six days after agro-infiltration, plant leaf samples were collected and stored at −80°C.

Recombinant COE-GCN4pII protein was purified via immobilized metal affinity chromatography (IMAC) as described by Pham et al. (29). The oligomeric form of purified COE-GCN4pII protein was determined by a cross-linking reaction that was described by Weldon et al. (32) and a native PAGE.

Leaf samples were ground in liquid nitrogen, mixed with PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO4, 1.8 mM KH₂PO4, pH 7.4). The crude extract was clarified by centrifugation twice at 13,000 rpm for 30 min at 4°C. The COE-GCN4pII protein in leaf crude extract and a number of known concentrations of anti-TNFα-nanobody-ELP standard protein [100, 200, 400, 800 ng, (33)] were separated in a 4–12% SDS-PAGE gel and transferred to a PVDF membrane (Millipore). The protein expression was determined via Western blot that was performed as described by Pham et al. (29) using monoclonal anti-c-myc antibody. The COE-GCN4pII protein expression level in the leaf crude extract was semi-quantified via Western blotting by comparison of the Western blot signal intensities and those of the anti-TNFα-nanobody-ELP standard protein using ImageJ software.

The study was approved by the ethical committee of the Institute of Biotechnology, Academic of Science and Technology Vietnam (VAST), Hanoi, Vietnam. The crude extracts after 6 days of the storage were mixed with the Emulsigen®-D adjuvant (MVP Technologies, 4805 G Street, Omaha, NE 68117, USA) with a ratio of 4:1 (v/v), respectively. Three groups of 6–8-week-old female BALB/C mouse (five per group) respectively numbered G1, G2, G3 were subcutaneously vaccinated at days 0, 14 and 28 with 200 μl of Emulsigen®-D adjuvant-formulated crude plant extracts of non-transgenic plants as negative control) or 200 μl of the commercial vaccine against the PEDV DR13 strain (4 × 10⁶ TCID50/dose, CTC Vacc PED, Korea) as positive control or 200 μl of Emulsigen®-D adjuvant-formulated crude plant extracts containing 18.76 μg of COE-GCN4pII protein that was semi-quantified by Western blotting. The bloods of mice were collected at seven days after the second and the third immunization via the retro-orbital sinus. All mouse sera were collected separately by centrifugation. To inactivate the non-specific complement, all mouse sera were incubated at 56°C for 30 min before being stored at −20°C until used.

PEDV propagation and purification were carried out as described by Hofmann et al. (34), with modifications. The Vero E6 cell line (ATCC® CRL-1587TM) was propagated and incubated at 37°C in Dulbecco's Modified Eagle Medium (DMEM) including 10% fetal bovine serum (FBS) and antibiotics (100 μg/mL penicillin/streptomycin). The cells were cultured in a 5% CO₂ at 37°C. Then, the PEDV-DR13 strain was propagated in Vero cells with 10 μg/mL trypsin treated- tosyl phenylalanyl chloromethyl ketone (TPCK) (Worthington, Lakewood, NJ, USA). After 36 h of cultivation, when all cells showed 100% cytopathogenic effects with morphological changes using cell morphology evaluation by inverted light microscopy, the infective culture fluid was harvested andfreeze-thawed three cycles. Next, cellular debris was pelleted by centrifugation at 10,000 × g for 30 min. The clarified supernatant was then enriched by ultracentrifugation at 30,000 × g. Sucrose density gradient centrifugation (20, 40, 60%) was then used to purify PEDV.

To detect PEDV-specific IgG mouse antibodies, 1 μg of purified PEDV DR13 was loaded onto 3-wells of one SDS-PAGE gel. The virus was separated and transferred to a PVDF membrane. The membranes were blocked with 5% (w/v) fat-free milk in PBS buffer (137 mM NaCl, 2.7 mM KCl, 10 mM Na₂HPO4, 1.8 mM KH₂PO4, pH 7.4) for 2 h. To separate the three lanes, the membrane was cut. Next, the single lane was incubated with a mixture of five mouse sera from each group (G1, G2, or G3) with a dilution of 1:200 at room temperature for 2 h, followed by the incubation with goat anti-mouse IgG secondary antibody-conjugated HRP with a dilution of 1:5,000.

The indirect ELISA was performed as described by Pham et al. (29), with some modifications. Hundred microliter of purified PEDV DR13 (5 ng/μl) in PBS (100 mM NaCl, 32 mM Na₂HPO4, 17 mM Na₂HPO4, pH 7.2) were added in a microtiter plate (ImmunoPlate Maxisorp, Nalgen Nunc International, Roskilde, Denmark) that was then incubated overnight. The plate was blocked, and incubated with 100 μl of each serially diluted mouse serum in 1% (w/v) BSA in PBST at ratios of 1:100, 1:200, 1:400, 1:800, 1:1,600, 1:3,200, 1:6,400, 1:12,800, 1:125,600, 1:51,200, 1:102,400 with three replications, followed by the addition of 100 μl of a 1:10 000 dilution of goat anti-mouse IgG-conjugated HRP (Invitrogen) or 1:2,000 dilution of goat anti-mouse IgA cross-adsorbed secondary antibody-conjugated HRP (Invitrogen) or 1:1,000 dilution of goat anti-mouse IgM secondary antibody-conjugated HRP (Invitrogen) in 1% (w/v) BSA in PBST. An internal control was generated by mixing 5 μl of each serum of all mouse groups, and put in all plates. OD450 signals of internal control were used to normalize the ELISA data. The cut-off value of ELISA was determined as (MEAN + 3.848 × SD) of negative control (35). BSA background was subtracted.

Two-fold serial dilutions of serum samples after the 2nd immunization were prepared in α-Minimum Essential Medium (MEM) including 1% antibiotic-anti-mycotic solution (Invitrogen, USA). Then, 10³ TCID50/0.1 ml of PEDV DR13 was added with an equal volume of diluted serum, and the virus-serum mixture was maintained at 37°C for 1 h. Next, 100 μl of each virus-serum mixture was introduced onto Vero cell monolayers in 96-well plates. The virus-serum mixture was removed after adsorption at 37°C for 1 h. The plates were then washed for 10 min with PBS. Finally, 200 μl of serum-free α-MEM medium containing trypsin were placed into each well and maintained at 37°C for 6 days. For controlling this assay, the virus control, positive serum control, negative serum control and blank control were used. The serum neutralization titres (SN titres) were defined as the highest serum dilution and consequent on inhibition of the cytopathic effect.

Statistical analyses for the ELISA test and virus neutralization assay were carried out in Sigma Plot software using a t-test. The difference between sample data mean was compared and is showed as the X ± standard deviation (SD). P-values that were <0.05 were determined to be significantly different.

The expression of the COE-GCN4pII protein in planta was successful demonstrated via separation of SDS-PAGE under reducing conditions, blotted and detected by Western blot using an anti-c-myc monoclonal antibody (Figure 1B).

The apparent band shown in Figure 1B with COE molecular weight was larger than the expected COE size predicted from the COE polypeptide sequence (26 kDa). One of the possible reasons for the increase in the molecular weight of COE protein is that the N-glycosylation sites located within the COE-S protein of PEDV at amino acids 511 and 533 may influence the electrophoretic behavior during the PAGE separation (16, 17). No COE protein was detected in wild type crude extract. The COE-GCN4pII protein expression level in leaf crude extract was semi-quantified by Western blotting. The COE-GCN4pII protein was quantified with a high expression level of ~4% of the total soluble protein. The amount of plant-produced COE-GCN4pII protein was found to be 234 mg/kg wet weight. Several publications have reported the accumulation of COE in transgenic plants (15, 17, 19); however, the expression level was still lower than that in our report. We demonstrated that the accumulation of COE in tobacco leaves could be significantly improved by codon optimization for plant expression by using a strong expression system, such as agro-infiltration.

The purification process was validated by collecting samples from each step of the purification procedure to analyse via SDS-PAGE and Western blot using a monoclonal anti-his antibody (Figures 1C,D). These results indicate the enrichment and successful purification of COE-GCN4pII protein from N. benthamiana leaves. The oligomeric state of the COE-GCN4pII protein was successfully determined by a cross-linking reaction with BS3 and a separation under native condition by native-PAGE. A band with a molecular weight of approximately 100 kDa corresponding to molecular weight of trimeric COE form was detected (Figures 2A–C). These results revealed that the trimeric COE protein was successfully generated in planta by the fusion of COE with GCN4pII motif.

Since animal vaccine development should minimize downstream processing in pig immunizations, the crude plant extract was chosen for testing immunogenicity in mice. Interestingly, after storing the crude extract containing trimeric COE at 4°C for six days, the COE content was still stable, as revealed by Western blotting (see Supplementary File).

The antibody- mediated humoral immune responses from vaccinated mice were first examined against the purified PEDV DR13 strain by Western blot (Figure 3A). Before vaccination, PEDV-specific IgG antibody responses were not detected in mice. The Western blots in Figure 3A showed that there was a band with molecular weight of over 245 kDa detected in mice groups G2 (vaccinated with the commercial vaccine against PEDV) and G3 (vaccinated with crude extract containing COE-GCN4pII) that was larger than the expected size of S protein PEDV (151.38 kDa). The larger band size obtained in Western blot might be explained that might be due to the influence on electrophoretic behavior by the 29 potential N-glycosylation sites locating within the S protein of PEDV during the PAGE separation (36). The results indicate that PEDV-specific IgG antibody responses were induced in mice groups G2 and G3 after the 2nd immunization and the 3rd immunization, and the antibody responses were strongly increased in both mice groups G2 and G3 after the 3rd immunization. In contrast, no PEDV-specific IgG antibody response was detected in mice group G1 vaccinated with crude extracts of non-transgenic plants.

Different levels of PEDV-specific IgG, IgA and IgM antibodies in each mouse sera group were calculated as the reciprocals of the geometric mean titer of the five mice of each group. End-point titer of each mouse sera group was compared by the t-test. The results showed that after the 2nd immunization, crude extract containing COE-GCN4pII (G3) elicited elevated levels of IgG, IgA and IgM antibody responses against PEDV (Figures 3B–D), reaching end-point antibody titres of 1:216 178, 1:45 429, 1:37 801, respectively. The levels of IgG, IgA and IgM antibody responses in mice group G3 were strongly enhanced after the 3rd immunization, having end-point antibody titres of 1:383 711, 1:195 985, 1:65 198, respectively. Notably, no statistically significant difference in PEDV-specific IgG antibody responses was obtained between mice group G3 and those in mice group G2, with P-values of 0.056 after the third immunization (P < 0.05). Therefore, plant crude extract containing the trimeric COE protein had the level of IgG antibodies similar to that of commercial vaccines against the PEDV DR13 strain after the third injection. However, level of PEDV-specific IgA and IgM antibody responses in mice group G3 were lower than those in mice group G2 after the second and the third immunization. Levels of IgG, IgA and IgM antibody responses against PEDV found in the sera of negative control mice group G1 were very poor.

The cytopathic effect caused by the wild-type PEDV DR13 virus of all serum samples was determined using a microscope and a representative cytopathic effect result observed under microscope of a single dilution of a serum from each group was presented (Figure 4A). The highest dilutions of sera that caused cytopathic effect inhibition were defined as the serum neutralizing antibody titres. The serum neutralizing antibody (SN) titer of every single mouse serum was presented in each dot (Figure 4B). As expected, cytopathic effect inhibition of sera from the negative control group was observed in a low serum dilution, which produced neutralizing geometric mean titres of this group as low as 1. In contrast, sera of mice vaccinated with the crude extract containing COE-GCN4pII and the commercial PEDV vaccine (the positive control) had the ability to neutralize PEDV with high neutralizing geometric mean titres: 57.6 and 61.86, respectively. These titres were significantly different from those of the negative control group. Notably, the serum neutralization titer in the crude extract containing COE-GCN4pII (G3) was not significantly different compared to that of sera from mice vaccinated with the commercial vaccine, with a P-value of 0.187.