The RBD of SARS-CoV-2 (SARS-CoV-2 RBD) (Genbank accession number: YP_009724390.1; F318-C617) was designed to anneal with Fc region of human immunoglobulin G1 (IgG1) (GenBank accession number: 4CDH_A) by peptide linker 3XGGGGs at the C-terminus (Figure 1). The nucleotide sequence of SARS-CoV-2 RBD was codon optimized and commercially synthesized (Genewiz, Suzhou, China) with XbaI and BamHI restriction sites for cloning with the Fc region, that contains BamHI and SacI restriction sites at the 5′ and 3′ ends, respectively. The SARS-CoV-2 RBD and human Fc region were ligated into a geminiviral vector (pBYR2eK2Md; pBYR2e) (Chen et al., 2011; Diamos and Mason, 2018) using XbaI and SacI restriction sites to construct the plant expression vector pBYR2e-SARS-CoV-2 RBD-Fc. The murine leader sequence (Shanmugaraj et al., 2020b) and ER retention signal peptide (SEKDEL) was included at N-terminus and C-terminus of the gene construct, respectively (Figure 1).

The plant expression vector pBYR2e-SARS-CoV-2 RBD-Fc was transformed into Agrobacterium tumefaciens strain GV3101 cells by electroporation (MicroPulser, Bio-Rad, United States). The recombinant Agrobacterium clones were confirmed by polymerase chain reaction (PCR) using the RBD gene-specific primers. Agrobacterium containing pBYR2e-SARS-CoV-2 RBD-Fc was resuspended with 1xinfiltration buffer (10 mM 2-(N-morpholino) etanesulfonic acid (MES), 10 mM MgSO4, at pH 5.5) to get final OD₆₀₀ of 0.2 prior to agroinfiltration. The Agrobacterium suspension was injected into the adaxial side of 6-week-old N. benthamiana leaves. The infiltrated plants were maintained in an optimal 16-h light/8-h dark condition at 28°C and harvested after 4 days post infiltration (dpi).

The infiltrated leaves were harvested and extracted with 1xPBS (phosphate-buffered saline: 137 mM NaCl, 2.68 mM KCl, 10.1 mM Na₂HPO₄, 1.76 mM KH₂PO₄ pH 7.4) and clarified by centrifugation at 26,000 g for 45 min at 4°C. The supernatant was filtered by using 0.45 μm S-Pak membrane (Merck, Massachusetts, United States). The clarified supernatant was purified by affinity column chromatography with protein-A beads (Expedeon, Cambridge, United Kingdom). The purified column was equilibrated and washed by 1xPBS pH 7.4 followed by elution with 0.1 M glycine buffer pH 3. Subsequently, the pH of the eluted proteins was neutralized by using Tris-HCl pH 8.8. The purified SARS-CoV-2 RBD-Fc was concentrated by using Amicon^® ultracentrifugal filter (Merck, Massachusetts, United States) and filtered through 0.22 μm syringe filter (Merck, Massachusetts, United States). The purified plant-produced SARS-CoV-2 RBD-Fc fusion protein was analyzed by using sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) and western blotting under reducing and non-reducing conditions. The SARS-CoV-2 RBD-Fc samples were subjected to 8% sodium dodecyl sulfate polyacrylamide gel electrophoresis and stained by Coomassie staining solution for protein visualization. For western blot analysis, the separated proteins were transferred to nitrocellulose membrane (Biorad, United States) and detected with a 1:5,000 sheep anti-human gamma chain-HRP conjugate antibody diluted in 1xPBS (The Binding Site, United Kingdom). The yield of purified plant-produced RBD-Fc fusion protein was estimated by direct ELISA assay using human IgG (Abcam, United Kingdom) as protein standard. The samples were probed by using anti-human gamma chain-HRP fusion (The Binding Site, United Kingdom) at the dilution of 1:1,000 in 1xPBS. A 3,3,5,5′-Tetramethylbenzidine (TMB) solution (Promega, United States) was added into the plate as a colorimetric developer followed by the addition of 1M H₂SO₄. The absorbance at 450 nm was read using a 96-well plate reader (Molecular Devices, United States).

The purified proteins were separated on SDS-PAGE. The targeted protein band was excised and sent to National Center for Genetic Engineering and Biotechnology, Pathum Thani, Thailand for LC-MS analysis. The protein was enzymatically digested with trypsin and injected into Hybrid quadrupole Q-Tof impact IITM (Bruker Daltonics Ltd., Hamburg, Germany) equipped with a Nano-captive spray ion source was coupled to a nanoLC system: Ultimate 3000 LC System (Thermo Fisher Scientific, United States). Equipment operation was controlled by Compass 1.9 software (Bruker Daltonics Ltd., Hamburg, Germany). The resulting MS/MS spectra were searched using the Mascot Sever (Matrix Science) against SwissProt database. For Mascot searches, the peptide mass tolerance was set at 0.6 Da and the fragment mass tolerance was set at 1.2 Da.

The binding activity of plant-produced SARS-CoV-2 RBD-Fc fusion protein to ACE2 protein was demonstrated by ELISA. Briefly, 96-well ELISA plate was coated by 100 ng of two different ACE2 protein derived either from HEK293-cells (Abcam, United Kingdom) or CHO-cells (InvivoGen, California, United States). For blocking, 5% skim milk in 1xPBS was added into the wells and incubated for 2 h at 37°C. After blocking, the plate was washed three times with 1xPBST (1xPBS plus 0.05% Tween-20) and incubated with various concentrations of plant-produced SARS-CoV-2 RBD-Fc fusion protein in 1xPBS. The SARS-CoV-2 RBD proteins in the wells were detected by addition of 1: 100 dilution of plant-produced anti-SARS-CoV-2 (H4) mAb (Shanmugaraj et al., 2020b) and followed by a 1:1,000 dilution of anti-human Kappa chain-HRP fusion (SouthernBiotech, United States) in 1xPBS for 1 h at 37°C. For colorimetric development, a TMB solution (Promega, United States) was added into the wells followed by addition of 1M H₂SO₄ for terminating the enzymatic reaction. The absorbance at 450 nm was measured using a 96-well microplate reader (Molecular Devices, United States).

Mice immunization protocols were approved by the Institutional Animal Care and Use Committee, Faculty of Medicine, Chulalongkorn University (Protocol review No. 012/2563). Seven-week-old female ICR mice (n = 5 per group) were intramuscularly (IM) immunized via., anterior tibialis with 10 μg of plant-produced SARS-CoV-2 RBD-Fc fusion protein without adjuvant or formulated with 0.1 mg alum (InvivoGen, California, United States) on days 0 and 21. Mice sera were collected prior to the first immunization (pre-bleed, day 0) and 14 days post-vaccination to assess the SARS-CoV-2-specific antibody response. The mice were sacrificed on day 35 (14 days after second booster) to collect the spleen for quantitative measurement of SARS-CoV-2 RBD-specific T-cell responses.

For non-human primate immunogenicity studies, all procedures were reviewed and approved by the National Primate Research Center of Thailand-Chulalongkorn University (NPRCT-CU) Animal Care and Use Committee (Protocol review No. 207512) and the facility has been AAALAC International Accredited (1752). Thirteen SPF juvenile cynomolgus macaques (Macaca fascicularis), aged between 2.5 and 3.5 years old, and body weight between 2.18 and 3.17 kg, were assigned into three groups in which the control group (n = 3) was immunized by PBS adjuvanted with 0.5 mg alum, and other two groups (n = 5 per group) were administered with 25 and 50 μg of plant-produced SARS-CoV-2 RBD-Fc fusion protein along with 0.5 mg alum adjuvant, respectively. Monkeys were received two intramuscular injections on day 0 and 21. The blood samples were collected on day 0 (pre-immunization) and 14 days after each vaccination (days 14 and 35) to assess SARS-CoV-2 RBD specific IgG, neutralizing antibody and cell-mediated immune responses.

96-well plate was coated with 100 ng of SARS-CoV-2 spike protein RBD derived from Sf9 insect cells (GenScript, United States) and incubated overnight at 4°C. Then, the wells were blocked with 5% skim milk powder in 1xPBS for 2 h at 37°C. Subsequently, the animal sera were twofold serially diluted with 1xPBS starting at 1:100 was loaded on the wells and incubated for 2 h at 37°C. Goat anti-mouse IgG HRP conjugate antibody (Jackson ImmunoResearch, Pennsylvania, United States) and goat anti-monkey IgG HRP conjugation (Abcam, United Kingdom) diluted 1:2,000 in 1xPBS were added into the wells for detecting mouse and monkey antibodies, respectively, and the plate was incubated for 2 h at 37°C. TMB substrate (Promega, United States) was added into the plates for colorimetric development. The enzymatic reactions were terminated by adding 1M H₂SO₄. The absorbance was measured at 450 nm using a microplate reader (BMG Labtech, Germany). Between each step, the plates were washed by 1xPBST for three times. For mouse IgG1 and IgG2a analysis, the mice sera were twofold serially diluted with 1xPBS starting 1:100 in the same fashion and detected by 1:2,000 goat anti-mouse IgG1 (HRP) and goat anti-mouse IgG2a heavy chain (HRP) antibody (Abcam, United Kingdom), respectively, diluted in 1xPBS. The endpoint titer of IgG1 and IgG2a were also computed for monitoring Th2 and Th1 lymphocyte responses, respectively. The endpoint titers were determined as the highest dilution of immunized sera, which had A₄₅₀ more than cut off calculated from A₄₅₀ of pre-immunized sera in the dilution of 1:100 in 1xPBS (Frey et al., 1998). All experiments were performed in duplicates and 1xPBS was used as a control. The statistical analyses of immunological data were performed using GraphPad Prism software version 8.0. Statistical significance was calculated by two-way analysis of variance (ANOVA). All data in each group were compared by using Tukey’s multiple comparisons test and the values of p < 0.05 were considered as statistically significant.

Microneutralization assay was performed using Vero E6 cell line and live SARS-CoV-2 virus isolated from a COVID-19 patient and conducted in a certified biosafety level 3 facility of Microbiology department, Faculty of Science, Mahidol University, Bangkok, Thailand. The cells were plated in 96-well plate at 1 × 10⁴ cells/well in DMEM (Dulbecco’s Modified Eagle’s medium supplemented with 10% heat-inactivated FBS, 100 U/mL of penicillin and 0.1 mg/mL of streptomycin) and incubated for overnight. The immunized sera and positive convalescent serum from COVID-19 patient were heat-inactivated at 56°C for 30 min. The immunized sera were twofold serially diluted in duplicates and incubated with 100 of 50% tissue culture infective dose (TCID₅₀) of the SARS-CoV-2 virus in DMEM at 37°C for 1 h. Virus control at 100TCID₅₀ and uninfected cell control wells were included in all plates. Subsequently, the mixture of diluted serum and virus was transferred to the cell monolayer and incubated at 37°C with 5% CO₂ for 2 days. After incubation, the cells were washed once with 1xPBS and then fixed and permeabilized with ice-cold 1:1 methanol/acetone fixative solution at 4°C for 20 min. The cells were washed 3 times with 1xPBST and blocked with 2% BSA at room temperature (RT) for 1 h. After washing, the viral nucleocapsid was detected using 1:5,000 of SARS-CoV/SARS-CoV-2 nucleocapsid (N) monoclonal antibody (SinoBiological, United States) and incubated at 37°C for 1 h followed by adding 1:2,000 of HRP-conjugated goat anti-rabbit polyclonal antibody (Dako, Denmark) diluted with 1xPBS and incubated at 37°C for 1h. The KPL Sureblue^TM TMB substrate (SeraCare, United States) was added for colorimetric development. Then the reaction was stopped by the addition of 1N HCl. The absorbance at 450 and 620 nm was read by a Sunrise^TM microplate reader (Tecan, Switzerland). The absorbance differences between 450 and 620 nm (A₄₅₀–A₆₂₀) of diluted samples were compared with the 50% specific signal of the cut point, which was calculated by the following equation to determine the potent neutralization titers of the immunized sera.

Where A_{Virus control} and A_{cell control} are the average of A₄₅₀-A₆₂₀ of virus control wells and cell control wells, respectively. The neutralizing titers were defined as the reciprocal highest dilution providing the average of A₄₅₀–A₆₂₀ of the diluted serum well more than the cut point. The neutralizing antibody titers of each experimental group were compared by GraphPad Prism 8.0 using Mann-Whitney test. The significant differences were considered when p < 0.05.

For mouse splenocyte preparation, the spleen cells were aseptically plated in the petri-dish and dissociated into single-cell suspension using needle#21 for 2–3 times. The separated splenocytes were cultured in R5 medium (RPMI1640 with 100 U/mL penicillin, 100 U/mL streptomycin, 5% heat-inactivated fetal bovine serum (FBS, Gibco, United States) and 2-mercaptoethanol) and centrifuged at 1,200 g 4°C for 5 min. Subsequently, splenocytes were lysed with 1xACK lysis buffer, added R5 medium and collected the cell pellet by centrifugation at 1,200 g 4°C for 5 min. Finally, splenocytes were counted and adjusted for using in ELISpot assay. The IFN-γ secreting cells were quantified by using mouse IFN-γ ELISpot assay kit (Mabtech, Stockholm, Sweden). Briefly, splenocytes were resuspended at 5 × 10⁶ cells/mL in R5 medium. The 96-well plates (Millipore, Bedford, MA, United States) were coated with 500 ng of anti-mouse IFN-γ (AN18) monoclonal antibody (mAb) (Mabtech, Stockholm, Sweden) in 1xPBS at 37°C with 5% CO₂ for 3 h. Then, the plates were washed six times with 1xPBS and blocked with R10 medium at RT for 1 h. A quantity of 5 × 10⁵ splenocytes per well were activated by lyophilized SARS-CoV-2 peptide pools (BioNet-Asia, Thailand, and Mimotopes, Australia) at a final concentration of 2 μg/mL at 37°C with 5% CO₂ for 40 h. R5 medium and concanavalin A (ConA) were used as negative and positive controls, respectively. After incubation, the plates were washed six times with 1xPBST followed by three times with 1xPBS. Then, the plates were incubated with anti-mouse IFN-γ-biotinylated mAb (Mabtech, Stockholm, Sweden) diluted in 1xPBS at RT for 3 h. After washing, streptavidin-alkaline phosphatase (ALP: Mabtech, Stockholm, Sweden) was added and incubated at RT for 1 h. The substrate solution (5-bromo-4-chloro-3-indolyl-phosphate/nitro blue tetrazolium; BCIP/NBT) were added into the wells and incubated until distinct spots emerge. Reactions were stopped by washing extensively in tap water and rinsing the underside of membrane. Inspect and count spots were performed with an ELISpot reader (ImmunoSpot^® Analyzer, United States). Results were expressed as spot-forming cells (SFCs)/10⁶ splenocytes. The positive responses were defined as > 50 SFCs/10⁶ splenocytes over the background signal. The result analyses were conducted using Kruskal-Wallis test in GraphPad Prism 6.0. All p-values < 0.05 were defined as significant.

For peripheral blood mononuclear cells (PBMCs) preparation, the cells were isolated by density gradient separation using Isoprep (Robbins Scientific Corporation, CA). Briefly, after removal of plasma, EDTA-treated whole blood was diluted (1:1) with RPMI1640 medium containing 2 mM L-Glutamine (Gibco, United States) and layered over Isoprep. Samples were then centrifuged at 1,200 g for 30 min. The PBMC layer was harvested and washed twice with RPMI1640. Then, the cells were resuspended in R10 medium (RPMI1640 supplemented with 100 U/mL of penicillin, 100 U/mL of streptomycin and 10% heat-inactivated fetal bovine serum (FBS, Gibco, United States) for applying in ELISpot assay. The antigen-specific cells secreting monkey IFN-γ were enumerated by using monkey IFN-γ ELISpot assay kit (Mabtech, Stockholm, Sweden). Briefly, PBMCs were stimulated with SARS-CoV-2 spike peptides pools (BioNet-Asia, Thailand, and Mimotopes, Australia) at 37°C with 5% CO₂ for 40 h. R5 medium and phytohemagglutinin (PHA) were served as negative and positive control, respectively. The secreted monkey IFN-γ were detected by anti-monkey IFN-γ-biotinylated mAb (Mabtech, Stockholm, Sweden) and followed by addition of ALP solution (Mabtech, Stockholm, Sweden). For spot development, BCIP/NBT-plus substrate solution was added into the wells and incubated until distinct spots emerge. The spots were inspected and counted by ELISpot reader (ImmunoSpot^® Analyzer, Germany). Results were expressed as spot-forming cells (SFCs)/10⁶ PBMCs following the subtraction of the negative control. The positive responses were defined > 100 SFCs/10⁶ PBMCs over the background. Statistical analyses were presented using GraphPad Prism 8.0. Comparison of frequencies of populations in each group was made using Friedman and Mann-Whitney tests. All p-values < 0.05 were defined as significant.

The nucleotide sequence of RBD of SARS-CoV-2 was codon-optimized and fused with Fc region of human IgG1 at the C terminus of the RBD gene construct. The codon-optimized SARS-CoV-2 RBD-Fc fusion gene was cloned into the geminiviral plant expression vector pBYR2e. For expression of RBD-Fc fusion protein in plants, N. benthamiana plants were infiltrated with Agrobacterium containing pBYR2e-SARS-CoV-2 RBD-Fc (Figure 1). The leaves infiltrated with Agrobacterium containing pBYR2e-SARS-CoV-2 RBD-Fc showed significant phenotypic necrosis compared to the leaves infiltrated by Agrobacterium without the plant expression vector (Figure 2A).

Plant-produced SARS-CoV-2 RBD-Fc fusion protein was extracted and purified from crude extract by single-step protein A affinity chromatography. The purified SARS-CoV-2 RBD-Fc fusion protein was concentrated and filtered by using 0.22 μm syringe filter. The purity of the purified plant-produced SARS-CoV-2 RBD-Fc was analyzed by SDS-PAGE gel stained with Coomassie blue staining under reducing and non-reducing condition. The purity of SARS-CoV-2 RBD-Fc fusion protein was found to be higher than 90% based on the visual inspection of a Coomassie-stained SDS gel (Figure 2B). The SARS-CoV-2 RBD-Fc fusion protein was further analyzed by western blot probed with anti-human gamma chain-HRP conjugate antibody. The protein band corresponding to the molecular weight of 65 kDa was observed in reducing condition (Figure 2C; lane 1). In addition, the plant-produced SARS-CoV-2 RBD-Fc fusion protein under non-reducing condition was observed at approximately 150 kDa (Figure 2C; lane 2), which implies that the SARS-CoV-2 RBD-Fc fusion protein could be linked by disulfide bond into dimeric form. The expression level of plant-produced SARS-CoV-2 RBD-Fc was quantified by ELISA and found to be 25 μg/g fresh weight. The authenticity of purified plant-produced SARS-CoV-2 RBD-Fc fusion protein was further confirmed by using a high-resolution LC-TOF MS/MS analysis as shown in Supplementary Document.

The binding of plant-produced SARS-CoV-2 RBD-Fc fusion protein was confirmed by ELISA by using commercial HEK293 and CHO-produced ACE2 protein as the capture reagent. The various dilutions of purified plant-produced SARS-CoV-2 RBD-Fc was incubated with commercial ACE2 proteins. For detection of SARS-CoV-2 RBD, anti-SARS-CoV-2 (H4) mAb (Shanmugaraj et al., 2020b) and anti-human kappa chain-HRP conjugate antibody were added into the wells. The results showed that plant-produced SARS-CoV-2 RBD-Fc fusion protein produced saturable binding to both commercial ACE2 proteins with substantially high affinity in comparison to PBS control (Figure 3), which confirms the authenticity of plant-produced SARS-CoV-2 RBD-Fc.

Mice immunogenicity was assessed in 7-week-old female Mlac:ICR mice by immunizing intramuscularly on days 0 and 21 with 10 μg of plant-produced SARS-CoV-2 RBD-Fc fusion protein with or without alum adjuvant. Mice sera were collected on days 0, 14, and 35 (Figure 4A). SARS-CoV-2 RBD-specific antibodies were evaluated by ELISA using commercial SARS-CoV-2 RBD-His protein produced from Sf9 cells as a capture antigen. The SARS-CoV-2 RBD-specific immune responses were observed after first immunization of plant-produced SARS-CoV-2 RBD-Fc alone, whilst a slightly increased specific-mouse total IgG response was observed in SARS-CoV-2 RBD-Fc immunized with alum. All mice immunized with plant-produced RBD-Fc elicited significantly higher antibody titer after second booster compared with the control group (Figure 4B). Plant-produced SARS-CoV-2 RBD-Fc was found to be immunogenic, while the addition of alum adjuvant could significantly improve its immunogenicity. In addition, we appraised the levels of SARS-CoV-2 RBD-specific IgG1 and IgG2a subclasses, which are indicators of Th2 and Th1 lymphocyte responses in mice, respectively. The results demonstrated that plant-produced SARS-CoV-2 RBD-Fc induced both SARS-CoV-2 RBD-specific IgG1 (Figure 4C) and IgG2a (Figure 4D) with the IgG1 bias. The in vitro neutralizing ability of the immunized sera was evaluated. The SARS-CoV-2 RBD-Fc without adjuvant induced neutralizing antibody against SARS-CoV-2 after the second dose at a dilution of 1:1,280 (Figure 4E). Interestingly, mice sera immunized by SARS-CoV-2 RBD-Fc adjuvanted with alum displayed maximum SARS-CoV-2 neutralization with a dilution more than 1:10,240 (Figure 4E). IFN-γ levels of splenocytes isolated from mice was tested by IFN-γ ELISpot assay. The results showed that plant-produced RBD-Fc elicited IFN-γ secretion which was statistically significant compared with mock control group (Figure 4F).

Cynomolgus macaques (Macaca fascicularis) were intramuscularly immunized with 25 and 50 μg of plant-produced SARS-CoV-2 RBD-Fc with the presence of alum on day 0 and 21. Monkey sera were collected on day 0, 14, and 35 (Figure 5A). Plant-produced SARS-CoV-2 RBD-Fc protein adjuvanted with alum was capable of inducing dose-independent SARS-CoV-2 RBD-specific IgG antibodies in monkeys after first and second immunization with the dilution 1:800, and 1:51,200, respectively (Figure 5B). Specifically, the microneutralization assay was performed by using Vero E6 cell line to evaluate the level of neutralizing antibodies against live SARS-CoV-2. Sera collected 14 days after the first immunization (day 14) showed neutralizing activity against the SARS-CoV-2 and increased at day 35 (14 days after the second immunization) with the neutralization titer approximately 5,120 (Figure 5C). In addition, cell-mediated immune responses in cynomolgus monkeys were monitored by IFN-γ ELISpot assay by using peripheral blood mononuclear cells isolated from immunized sera on day 35. The specific IFN-γ expression was detected in monkeys immunized with two doses of plant-produced SARS-CoV-2 RBD-Fc, and a significant difference from the control group was detected compared to immunized monkeys (Figure 5D).