Introduction

Colorectal cancer is one of the most common malignancy associated death worldwide. Colorectal cancer is defined as an abnormal growth of a polyp on the inner lining of the colon and rectum that can become cancerous and spread into other parts of the body (Koido 2013; Sarvizadeh et al. 2019). Optimization of surgical resection dramatically improves 5- to 10-year survival rates of patients with localized disease, but approximately 25% of them has a possibility to have a relapse due to the presence of metastasis (Tauriello et al. 2017; Temraz et al. 2014). Thus, it is important to develop alternative treatments that can improve survival rate of colorectal cancer patients. One of such approaches is the development of cancer vaccine (Abakushina et al. 2019; Bracci et al. 2020; Wong-Arce et al. 2017).

Vaccines help stimulation of the immune response against particular diseases (Francis 2018; Mason 2016). There are several different systems that can be used for the production of vaccines (Shanmugaraj et al. 2020). Bacterial system has several advantages as a tool for recombinant protein production such as easy manipulation, high expression level, and ease of scale up (Rosales-Mendoza et al. 2016, Shanmugaraj et al. 2020). However, some recombinant proteins produced by bacterial system failed to retain its biological activity due to improper folding or lack of post-translational modification (Tokmakov et al. 2012). Compare to bacterial system, mammalian cell system allows proper folding and authentic post-translational modification of recombinant protein (Lomonossoff and D'Aoust 2016, Shanmugaraj et al. 2020). Based on the advantages, mammalian cell system have contributed to produce many commercially available vaccines (Francis 2018; Rajaram et al. 2020; Zost et al. 2017). Despite of the advantages, high production cost and possible contamination of human pathogens have been regarded as hurdles for application of mammalian cell system for vaccine production (Schillberg et al. 2019). As a alternatives of mammalian system, plant systems have been in spotlight for the production of vaccine candidates (Lomonossoff and D'Aoust 2016, Rybicki 2009, Shanmugaraj et al. 2020, Takeyama et al. 2015). The key advantages of plant system for production of pharmacological recombinant proteins are low production cost, easy cultivation, rapid mass production, inherent safety from human pathogens (Lomonossoff and D'Aoust 2016; Rybicki 2009). Moreover, the use of plant system to produce vaccine candidates allows proper post-translational modifications which are crucial for achieving their desired biological functions. (Blokhina et al. 2020; Brodzik et al. 2008; Daniell et al. 2009; Tusé et al. 2020). For these reasons, there are many attempts testing the feasibility of plants system for production of recombinant proteins (Joung et al. 2016, Shanmugaraj et al. 2020, Tusé et al. 2020). For example, “Elelyso”, a medicine for Gaucher’s disease has been registered as the first plant-based pharmacological recombinant protein (Fox 2012).

Among the plant systems, tobacco system has been extensively used for the production of pharmacological proteins due to its several advantages such as relatively easy transformation, robust protein expression capacity, and short life cycle (Daniell et al. 2001; Grosse-Holz et al. 2018). Golovkin et al. reported the production of the recombinant vaccinia virus B5 antigenic domain (pB5) as a component of vaccine against smallpox in tobacco (Golovkin et al. 2007). Application of plant-derived pB5 conferred protection from lethal dose of vaccinia virus (Golovkin et al. 2007). Similarly, domain III of the dengue 2 envelope protein expressed in tobacco plants induced the production of the anti-dengue virus antibody in mice (Saejung et al. 2007). In addition to virus vaccines, many cancer vaccines have been expressed in tobacco system (Lee and Ko 2017; McCormick 2011). Personalized recombinant vaccines against follicular B cell lymphoma were generated in tobacco system. A phase I safety and immunogenicity clinical trial of the vaccines revealed that almost half of the patients (47%) successfully developed antigen-specific responses (McCormick et al. 2008). The transmembrane human epithelial mucin MUC1 that is over-presented in breast cancer cells was expressed in tobacco system. The plant-derived MUC1 tandem repeat fused to the Escherichia coli enterotoxin B subunit induced production of anti-MCU1 antibodies (Pinkhasov et al. 2011).

For successful development of cancer vaccines, it is necessary to identify and produce tumor-specific antigens that can induce immune response against cancer cells. GA733 proteins have been regarded as potential target antigens that can be used for the production of colorectal cancer vaccines (Fu et al. 2018; Lu et al. 2012; Mamaloudis et al. 2015; Park et al. 2015a). GA733 protein belongs to the group of calcium independent cell adhesion molecules (CAM), which are highly accumulated on the cell surface of colorectal carcinomas (Mamaloudis et al. 2015). Expression of GA733 or its murine homologue in tumor cells inhibited growth of colorectal cancer cells (Armstrong and Eck 2003; Basak et al. 1998; Li et al. 1997). In addition, GA733-2, a member of GA733 family, produced in tobacco or swiss chard also successfully induced immune responses and inhibited tumor growth in mice (Brodzik et al. 2006, 2008; Verch et al. 2004). However, lower expression level of GA733-2 protein was regarded as a constraint for further applications (Lim et al. 2014; Park et al. 2015b). Thus, optimization of recombinant GA733-2 (rGA733-2) protein expression in plant systems is prerequisite for efficient production of GA733 and its application as a cancer vaccine. In this study, we tested several potential factors that affect expression level of rGA733 and rGA733 fused to Fc domain (rGA733-Fc) recombinant protein in tobacco plants. In addition, we generated transgenic tobacco plants that produce rGA733 and rGA733-Fc, respectively, for stable production of the antigens. Finally, the stability and functionality of the recombinant proteins were confirmed by western blot, ELISA, glycosylation pattern, and in vitro analysis.

Materials and methods

Plant expression vector construction

Various plant expression vectors were generated for the expression of the rGA733-2 and the rGA733-Fc protein as described below. The DNA sequence encoding truncated human colorectal carcinoma antigen GA733-2 (aa17-aa266) (Park et al. 2015a; Szala et al. 1990) and GA733-Fc recombinant protein sequence (Fc fragment of human IgG1 (Val97-Gly328, GenBank accession no. AY172957) were amplified with GA733-Fsal (5\(^{\prime}\)-ACGTCGACATGGATACGGCG-3\(^{\prime}\)) and GA733-Rsac (5\(^{\prime}\)-TGAGCTCTCACACGTGGTGGTG-3\(^{\prime}\)) primers. The amplified GA733-2 and GA733-Fc fragments were introduced into the pBI121 vector carrying the 35S promoter, the 5\(^{\prime}\) UTR sequence from the tobacco etch virus and a segment of DNA sequence encoding 30-aa length plant ER signal peptide (MATQRRANPSSLHLITVFSLLAAVVSAEVD, ER signal peptide from Nicotiana plumbaginifolia) through SalI and SacI restriction enzyme sites. The final constructs were named as pBI121-GA733 and pBI121-GA733-Fc, respectively. GA733-2 and GA733-Fc were amplified from pBI121-GA733 and pBI121-GA733-Fc expression vector with GA733-Fsal and KDEL-Rsnab (5\(^{\prime}\)-TGTACGTATCAGAGTTCATCTTT-3\(^{\prime}\)) primers, and cloned into pBINPLUS vector containing 35S promoter, the 5\(^{\prime}\) UTR sequence from the tobacco etch virus and the ER signal peptide through SalI and SnaBI sites. The final constructs were named as pBINPLUS-GA733 and pBINPLUS-GA733-Fc, respectively. To generate pPZP-200-GA733, the GA733-2 expression cassette (p35S:UTR-SP-GA733 -Fc-t35S) was isolated by digestion with HindIII from the pBINPLUS-GA733 plant expression vector, and cloned into the pPZP-200 vectors through a HindIII enzyme site. The pPZP-Bar-GA733 expression vector was constructed by introducing the GA733-2 expression cassettes (p35S:UTR-SP-GA733-t35S) into the pPZP-Bar expression vector through a HindIII enzyme site. The pBINPLUS vector containing the 5\(^{\prime}\) UTR from the tobacco etch virus and a sequence encoding a plant ER signal peptide were obtained from Dr. Kisung Ko of Chung-Ang University. The pPZP-200 and pPZP-Bar vectors were obtained from Dr. Ju-Kon Kim of Seoul National University.

To determine the relationship between rGA733-2 expression and structure of T-DNA, serial truncates of Bacillus subtilis NADPH-cytochrome P450 reductase 102A2V4 gene (GenBank accession no. FJ899088) (200, 400, 600, 800, 1000, 1200, 1400, 1600, 1800, and 2000 bps in length) were amplified with following primers; 2000-F: 5\(^{\prime}\)-TCCGGCCGG AACGGTTCGAG-3\(^{\prime}\), 200-R: 5\(^{\prime}\)-ATAAGGTTTGTTTGATA-3\(^{\prime}\), 400-R: 5\(^{\prime}\)-ATCGCCTTC TGCGGTGC-3\(^{\prime}\), 600-R: 5\(^{\prime}\)-ATGACGCCCTCAAGCTC-3\(^{\prime}\), 800-R: 5\(^{\prime}\)-ATCCGAATGCTTTGATG-3\(^{\prime}\), 1,000-R: 5\(^{\prime}\)-ATGATATTCAACATCCG-3\(^{\prime}\), 1,200-R: 5\(^{\prime}\)-ATTGTGGCTGCTTCCTG-3\(^{\prime}\), 1,400-R: 5\(^{\prime}\)-ATCTCGGTTTTAACGGA-3\(^{\prime}\), 1,600-R: 5\(^{\prime}\)-ATGGTCTTCCGGAAGCT-3\(^{\prime}\), 1,800-R: 5\(^{\prime}\)-ATGTGGACAGTGACGAT-3\(^{\prime}\), 2,000-R: 5\(^{\prime}\)-CCTGCAGATGTCTCAGCCAG-3\(^{\prime}\). Each DNA fragments were sub-cloned into the pGEM-T Easy Vector (Promega, USA). The fragments confirmed by sequence analysis were introduced into the pPZP-200 vector through an EcoRI restriction enzyme site.

Transient expression in tobacco leaves

Six- to 8-week-old N. tabacum cv. Xanthi nc and N. benthamiana plants grown in a growth room (16 h-light/8 h-dark cycle, 24 ℃) were used for transient expression of rGA733-2 and rGA733-Fc proteins. The Agrobacterium tumefaciens (A. tumefaciens) C58C1, LBA4404 and GV3101 strains were incubated in YEP medium (10 g/L yeast extract, 10 g/L peptone, and 5 g/L NaCl, pH 7.2) for 2 days. The bacterial cells were harvested by centrifugation at 13,000 rpm for 20 min, and the harvested cells were resuspended in 10 mM MgCl2 (Final OD600 = 0.9) containing 200 μM acetosyringone (pH 5.6). The tobacco intact leaves were infiltrated with the bacterial suspension using a 1-ml needless plastic syringe. The infiltrated tobacco leaves were harvested at the indicated time points after infiltration and immediately frozen in liquid nitrogen for further analysis.

Agrobacterium-mediated transformation of tobacco plants

Transgenic tobacco plants were generated through the agrobacterium-mediated transformation method as previously described (Horsch et al. 1989). Fully expanded unstressed leaves from a sterile cultured Nicotiana tabacum cv.Xanthi plant were cut into squares with diameter of 0.5–1.0 cm, and the leaf disks were incubated with LBA4404 agrobacterium harboring a plant expression vector for 10 min. After the inoculation, the disks were placed with their adaxial surfaces facing downwards on co-cultivation medium [3% sucrose, 4.4 g/L MS salt (Duchefa, Netherlands), 2 mg/L 2,4-D (Duchefa, Netherlands), and 200 μM acetosyringone (MBcell, Korea), pH 5.8]. After 2 days of co-cultivation, the disks were transferred into shoot induction medium [3% sucrose, 4.4 g/L MS salt, 50 mg/L kanamycin (Biosesang, Korea), 0.01 mg/L NAA (Duchefa, Netherlands), 0.1 mg/L GA3 (Duchefa, Netherlands), 1 mg/L zeatin (Duchefa, Netherlands), and 250 mg/L cefotaxime (Biosesang, Korea), pH 5.8]; the explants were transferred into fresh shoot induction medium every 2–3 weeks. The regenerated shoots were transferred to root induction medium (3% sucrose, 4.4 g/L MS salt, 50 mg/L kanamycin, and 300 mg/L cefotaxime, pH 5.8). The regenerated plantlets were transferred to soils.

Protein extraction and western blot analysis

Total soluble proteins were isolated from tobacco leaves. The harvested tobacco leaves were homogenized using a TissueLyser II (Qiagen, USA) with one volume of Bradley buffer (Bradley and Zaitlin 1971) at 30 Hz for 5 min. The crude extracts were centrifuged at 13,000 rpm for 30 min at 4 °C, and the supernatants were separated on 10% or 12% SDS polyacrylamide gels and transferred to a polyvinylidene fluoride membrane (GE Healthcare, USA). The membrane was incubated with blocking buffer [1% bovine serum albumin (BSA) and 5% skim milk] at 25 °C for 3 h and then washed with TBST (150 mM NaCl, 2.5 mM KCl, 25 mM Tris base, and 0.1% Tween-20). Rabbit rGA733-2 specific antibody was used as a primary antibody (1:5000 dilution) (Park et al. 2015a) and HRP-conjugated goat anti-rabbit IgG was used as a secondary antibody (1:5000 dilution, Thermo Scientific, USA). The proteins were visualized using an ECL system (Millipore, USA) and a LAS-3000 luminescent image analyzer (Fujifilm, Japan).

Enzyme-linked immunosorbent assay (ELISA)

The soluble fraction of crude extracts were diluted in 100 mM carbonate–bicarbonate coating buffer (28.6 mM Na2CO3 and 71.4 mM NaHCO3, pH 9.6) to a final concentration of 10 μg/ml, and 1 μg of soluble protein was used for ELISA analysis. All process was performed in a 96-well immunoassay plate. The samples were blocked with 200 μl of 5% skim milk in PBS buffer (16.34 mM Na2HPO4, 2.68 mM KCl, 0.94 mM K3PO4, 136.9 mM NaCl, pH 7.4) for 2 h and then incubated with 50 μl of 1: 2,000 dilution of rabbit pAb rGA733-2 for 2 h, followed by 50 μl of 1: 2000 dilution of AP-conjugated goat anti-rabbit IgG (Thermo Scientific, USA) for 1 h. 200 μl of p-nitrophenyl phosphate (1 mg/ml; Sigma-Aldrich, USA) was added as a substrate for AP enzyme. The absorbance of the samples at 450 nm was monitored using a VERSA max microplate reader (Molecular Devices, USA).

Purification of rGA733-Fc

The leaves of transformed tobacco plants were ground in 0.5× PBS buffer. The soluble fraction was dialyzed against binding buffer (10 mM NaH2PO4 and 150 mM NaCl, pH8.2). After filtration, total soluble proteins were passed through an Affi-Gel® Protein A Gel column (Bio-Rad Laboratories, CA, USA) at a flow rate of 0.5 ml/min, and the column was washed with 10 bed volume binding buffer. Bound proteins were eluted by elution buffer (0.1 M glycine, pH 3.0) at a flow rate of 0.5 ml/min. The glycosylation patterns of plant-derived GA733-Fc proteins were analyzed by Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric (MS) analysis.

In vitro flow cytometry analysis

To test potential effects of plant-driven rGA733 and rGA733-Fc protein on progress of apoptosis in colorectal cancer cells, we performed in vitro flow cytometry analysis. The MC38 (Mouse colorectal cancer cell line) and the HCT116 (Human colorectal cancer cell line) cells (2 × 106 cells/6 cm2 culture dish) were treated with the crude extract of non-transformed (NT) plants, rGA733OX and rGA733-FcOX transgenic plants. Human Fc protein (h-Fc) was used as a negative control for Fc domain of rGA733-Fc protein. After 48 h incubation, the cells were harvested and used for apoptosis analysis by flow cytometry using Modifit LT 3.0 and CellQuest software program.

Genomic DNA extraction and PCR analysis of transgenic plants

Genomic DNA was extracted from transgenic leaves to confirm the insertion of target DNA. 300 mg of transgenic plant leaves were ground in liquid nitrogen and mixed thoroughly with 700 μl of DNA extraction buffer (0.05 M Tris–HCl, pH 7.6, 0.5% SDS, 0.1 M NaCl, 0.05 M EDTA, and 0.1 M β-mercaptoethanol). The mixtures were incubated at 50℃ for 10 min, and then were thoroughly mixed with 500 μl of phenol: chloroform: isoamylalcohol (25:24:1, Biosesang, Korea). The resulting extracts were centrifuged at 4℃ at 13,000 rpm for 15 min. After centrifugation, the supernatants were transferred to fresh tubes. 2 volumes of ethanol were added to the tubes and the tubes were inverted 5–10 times. The resulting genomic DNA extracts were centrifuged at 4 ℃ at 13,000 rpm for 10 min. The genomic DNA was washed with 75% ethanol and dissolved in distilled water. The presence of the GA733-2 gene was determined by PCR analysis using the specific primers GA733 RT-F (5\(^{\prime}\)-GGC CGT AAA CTG CTT TGT GA-3\(^{\prime}\)) and GA733 RT-R (5\(^{\prime}\)-CGC GTT GTG ATC TCC TTC TG-3\(^{\prime}\)). After preheating the sample at 95 °C for 5 min, 34 amplification cycles of 95 °C for 30 s, 55 °C for 30 s, and 72 °C for 1 min were performed. A final incubation was conducted at 72 °C for 10 min.

RNA extraction and real-time PCR analysis

Total RNAs were extracted from leaves of tobacco plants using an RNeasy® Plant Mini Kit according to manufacturer’s instruction (Qiagen, USA). 1 μg of total RNAs was used for cDNA synthesis using a QuantiTect® Reverse Transcription Kit (Qiagen, USA). A genomic DNA was eliminated by the 7X gDNA Wipeout Buffer (Qiagen, USA) at 42℃ for 2 min. The sample was incubated for 1 min on ice and mixed with reverse-transcription master mix. cDNA synthesis was carried out at 42 °C for 15 min and at 95 °C for 3 min. 1 μl of cDNA was used for quantitative real-time PCR (qRT-PCR) analysis using GA733-specific primers (GA733 RT-F and GA733 RT-R) and actin primers (Actin-F: 5\(^{\prime}\)-TTGGAATGGAAGCTGCTGGA-3\(^{\prime}\) and Actin-R: 5\(^{\prime}\)-TCAGGAGGC GCCACCACCTT-3\(^{\prime}\)). After preheating the sample at 95 °C for 5 min, 34 amplification cycles (95 °C for 30 s, 50 °C for 30 s, and 72 °C for 30 s) were performed using a Rotor-gene Q PCR machine (Qiagen, USA).

Results

Comparison of vector systems for transient expression of rGA733-2 in tobacco plants

To test the effect of expression vector systems for the accumulation of rGA733-2 and rGA733-Fc proteins in tobacco plants, several expression vectors were constructed with Cauliflower Mosaic Virus 35S promoter for overexpression and 5\(^{\prime}\) UTR sequence from the tobacco etch virus for enhancing translation efficiency. Since low protein stability of recombinant proteins in cytosols has frequently emerged as a major limitation for production of recombinant protein in plants, a plant ER targeting signal peptide and ER retention codon were fused to rGA733-2 and rGA733-Fc coding sequence to target rGA733-2 and rGA733-Fc to the ER (Fig. 1a). Agrobacterium-mediated transient expression system was applied to validate efficiency of the constructs for rGA733-2 protein expression in tobacco plants. Western blot analysis showed that rGA733-2 proteins were clearly detected in tobacco plants when infected with the pBINPLUS and the pPZP-Bar vector systems but not by the pBI121 and the pPZP-200 vector systems (Fig. 1b). Similarly, rGA733-Fc protein was highly expressed with the pBINPLUS and the pPZP-Bar vector systems (Fig. 1c). These results suggest that backbone of plant expression vectors affects expression level of rGA733-2/rGA733-Fc proteins in tobacco plants.

Fig. 1
figure 1

Construction of rGA733-2 and rGA733-Fc plant expression vectors. a The schematic diagram of the plant expression vectors constructed for expressing rGA733-2 and rGA733-Fc protein in plants. LB and RB, left and right T-DNA borders; p35S, Cauliflower Mosaic Virus 35S promoters; UTR, 5\(^{\prime}\) UTR from the tobacco etch virus; SP, 30-aa plant ER signal peptide; KDEL, endoplasmic reticulum (ER) retention codon; tNOS, nopaline synthase terminator; NPT II, neomycin phosphotransferase II; Bar, phosphinothricin N-acetyltransferase protein. Expression levels of the rGA733-2 protein (b) and rGA733-Fc (c) in tobacco plants. Agrobacterium tumefaciens GV3101 strain harboring the indicated constructs were infiltrated into tobacco leaves. The level of rGA733-2 and rGA733-Fc was determined by western analysis using the GA733 specific antibody

Full size image

Expression level of rGA733 is affected by length of interval between the rGA733 expression cassette and left border of T-DNA

We found that the rGA733 protein was efficiently expressed by the pPZP-Bar but not by the pPZP-200 vector system even though we introduced the identical expression cassette into the vectors. As the pPZP-Bar vector has identical structure with pPZP-200 except for the existence of the Bar gene, we compared these two systems to understand the observed phenomenon. We first replaced Bar gene in pPZP-Bar-GA733 expression vector with Bacillus subtilis NADPH-cytochrome P450 reductase 102A2V4gene (CYP) to test whether presence of antibiotic resistance gene affects expression of rGA733 protein in plants (Fig. 2a). Western analysis showed that expression level of rGA733 was unaffected by the replacement of the Bar gene to the CYP gene (Fig. 2b), suggesting that the presence of antibiotic resistance gene is unnecessary for the expression of rGA733 protein. We then hypothesized that length of downstream interval between expression cassette and left border of T-DNA might be important for the expression of rGA733 protein in tobacco plant. To test the possibility, we constructed rGA733 expression vectors with insertions of various length of DNA fragments (0.2–2 kb) between rGA733 expression cassette and left border of T-DNA in the pPZP-200-GA733 vector (Fig. 2c, d). Interestingly, we observed a positive correlation in length of inserted DNA fragments between the rGA733 gene and left border of T-DNA and expression level of proteins. Specifically, insertion of DNA fragments longer than 0.8 kb greatly increased rGA733 protein expression (Fig. 2d). Further quantitative real time PCR (qRT-PCR) analysis showed that expression level of rGA733 was two-fold higher in tobacco plants transfected by the pPZP-2 kb-GA733 than pPZP-GA733 vector (Fig. 2c). Collectively, these results suggest that the interval longer than 0.8 kb between rGA733 expression cassette and left border of T-DNA increases rGA733 expression in tobacco plants.

Fig. 2
figure 2

The effect of 3\(^{\prime}\) interval for rGA733-2 protein expression in plants. a The schematic diagram of three plant expression vectors. LB and RB, left and right T-DNA borders; p35S, Cauliflower Mosaic Virus 35S promoters; UTR, 5\(^{\prime}\) UTR from the tobacco etch virus; SP, 30-aa plant ER signal peptide; KDEL, endoplasmic reticulum (ER) retention codon; Bar, phosphinothricin N-acetyltransferase protein; 2 Kb, Bacillus subtilis NADPH-cytochrome P450 reductase 102A2V4 gene. Expression level of rGA733-2 protein (b) and transcripts (c) in tobacco plants infiltrated with the indicated constructs (d) The effect of 3\(^{\prime}\) interval between 35S terminator and left border of T-DNA. Different size of DNA fragments was inserted between the 35S terminator and left border of the pPZP-200 vector. The level of rGA733-2 protein was determined by western analysis using the GA733 specific antibody. Rubisco large subunit (RbcL) was used as a loading control. Relative expression level of rGA733-2 was determined by quantitative real time PCR (qRT-PCR). The SlACTIN gene was used as an internal control. Data represent mean + standard deviation (SD) of three technical replicates

Full size image

Effect of tobacco cultivars (N. tabacum cv.Xanthi nc and N. benthamiana) and Agrobacterium strains (C58C1, LBA4404, and GV3101) on expression of rGA733-2 and rGA733-Fc proteins

Although the tobacco system is suitable for robust expression of heterogeneous protein, it is known that there exit huge variations in the expression levels of recombinant protein depending on the combination of tobacco species and A. tumefaciens strains used for transient expression (Shamloul et al. 2014). Thus, it is important to determine appropriate combination of tobacco species and A. tumefaciens strains to maximize recombinant protein expression. We, therefore, tested the effect of tobacco cultivars (N. tabacum cv.Xanthi nc and N. benthamiana) and Agrobacterium strains (C58C1, LBA4404, and GV3101) on the expression levels of rGA733-2 and rGA733-Fc proteins in tobacco plants (Fig. 3). We also monitored protein levels for 9 days to determine optimal incubation time for robust productions of rGA733-2 and rGA733-Fc protein in tobacco plants. Accumulation of rGA733 and rGA733-Fc protein increased during first 3 days after infiltration, and then continuously decreased in all tested conditions. These results indicate that 3 days after infiltration is optimal time point for robust expression of rGA733-2 and rGA733-Fc in tobacco plants. The type of Agrobacterium strains caused only subtle differences on the expression levels of both rGA733-2 and rGA733-Fc in N. tabacum plants. However, in N. benthamiana plants, the A. thumefaciens LBA4404 strain showed five-fold higher efficiency than C58C1 and GV3101 strains on the expression of rGA733-2 and rGA733-Fc proteins. The highest expression level of the rGA733-2 was 15.92 μg/g and that of the rGA733-Fc was 51.28 μg/g (Fig. 3). Collectively, these data indicate that LBA4404-mediated transient expression of rGA733-2 and rGA733-Fc greatly increases protein expression efficiency in tobacco (N. benthamiana) plants.

Fig. 3
figure 3

Effect of A. tumefaciens strains and tobacco species on rGA733-2/rGA733-Fc expression. Six-week-old Nicotiana benthamiana (N. benthamiana) and Nicotiana tabacum (N. tabacum) plants grown on soils were infiltrated with three different Agrobacterium tumefaciens strains (C58C1, LBA4404, and GV3101) harboring pBINPLUS-GA733 or pBINPLUS-GA733-Fc expression vectors. The infected leaves were harvested at the indicated time points after infiltration. Protein levels were quantified ELISA analysis. Data represent mean ± standard deviation (SD) of three replicates

Full size image

Generation of transgenic tobacco plants expressing rGA733-2/rGA733-Fc

Generation of transgenic plants is important for stable and low-cost production of recombinant proteins (Mason et al. 2002). To achieve stable expression of rGA733-2 and rGA833-Fc protein in plants, we generated transgenic tobacco plants overexpressing rGA733 (rGA733-2OX) (Fig. 4a) and rGA733-Fc (rGA733-FcOX) (Fig. 4b), respectively. PCR analysis using genomic DNA confirmed the existence of transgene cassettes in tobacco plants. Further western analysis showed that rGA733-2 and rGA733-Fc proteins were expressed in all the selected transgenic plants. Among the selected transgenic plants, rGA733-2OX #3 and rGA733-FcOX #2 transgenic tobacco exhibited the highest protein expression level (Fig. 4). The rGA733-2 protein level in rGA733-2OX #3 leaves was 3.12 μg/g fresh weight, and the level of gRNA733-Fc was 5 μg/g fresh weight in the rGA733-FcOX #2 leaves.

Fig. 4
figure 4

Characterization of transgenic tobacco plants expressing rGA733-2/rGA733-Fc protein. Transgenic tobacco plants expressing rGA733-2 (a) and rGA733-Fc (b) were generated using Agrobacterium-mediated transformation method with pBINPLUS-GA733 and pBINPLUS-GA733-Fc plant expression vectors. The transgenic plants were analyzed by genomic DNA PCR and western blot analysis. NT, non-transgenic tobacco plants

Full size image

In vitro analysis for evaluating biological functions of rGA733/rGA733-Fc from transgenic tobacco plants

To confirm the biological functionality of plant-driven rGA733 and rGA733-Fc protein, we performed in vitro flow cytometry analysis using crude extracts of tobacco plants (Fig. 5). The apoptosis progression induced by the crude extracts was monitored in the MC38 (Mouse colorectal cancer cell line) and the HCT116 (Human colorectal cancer cell line) cells. Non-transformed (NT) plants were used as negative controls for rGA733OX and rGA733-FcOX transgenic plants. Human Fc protein (h-Fc) was used as a negative control for Fc domain of rGA733-Fc protein. MC38 and HCT116 cells (2 × 106 cells/6 cm2 culture dish) were treated with the crude extracts of NT, in rGA733-2OX, and rGA733-FcOX and Human Fc protein for 48 h. The extracts of NT plants induced 49% of apoptosis in MC38 cells. Simultaneous treatments of h-Fc caused to induction of apoptosis from 55% of the treated cells. Under same conditions, rGA733-2OX and rGA733-FcOX extracts induced 51% and 68% of apoptosis, respectively. In HCT116 cells, the extracts of NT and rGA733OX plants induced 49% of apoptosis. Under same conditions, rGA733-FcOX extracts induced 54% of apoptosis. As a result, the apoptosis progression of MC38 and HCT116 cells were induced by treatment of plant derived GA733-Fc (Fig. 5). These results suggest that plant-derived GA733-Fc suppresses the tumorigenicity of both MC38 and HCT116 carcinoma cells by inducing apoptosis.

Fig. 5
figure 5

The plant-derived GA733-Fc regulated the apoptosis in MC38 and HCT116 cells. MC38 and HCT116 cells were treated with the extracts of non-transgenic, rGA733-2OX and rGA733-FcOX plants for 48 h. The apoptosis analysis was performed by flow cytometry using a Modifit LT 3.0 and a CellQuest software program. Un, Un-treated; Con, non-transgenic tobacco plants; h-Fc, human Fc protein; GA733-Fc, extracts of rGA733-FcOX plants; GA733, extracts of rGA733-2OX plants

Full size image

Purification and analysis of N-glycosylation of plant-derived rGA733-Fc protein

Purification of plant-derived recombinant protein is necessary for the use of the recombinant protein as a vaccine candidate. The presence of the Fc domain improves stability of fusion protein and allows cost-effective purification during manufacture (Carter 2011; Czajkowsky et al. 2012). Thus, rGA733-Fc recombinant proteins generated by stable transformation were purified using a protein A column. Coomassie brilliant blue staining showed that most of the purified proteins were detected in specific elution fractions (Fig. 6a). The bands were further confirmed as rGA733-Fc proteins by western blot analysis using a rGA733 specific antibody (Park et al. 2015a) (Fig. 6b). In addition to rGA733-Fc band, additional band above 70 kDa was detected by western blot analysis. Similar with our results, Verch et al. reported that rGA733-Fc dimer was detected by western blot analysis (Verch et al. 2004). Final concentration of purified rGA733-Fc was 4.67 ug/g from rGA733-FcOX transgenic plants. As a glycan structure is an important feature for application of a recombinant protein as a vaccine candidate, we next examined glycosylation patterns of the purified rGA733-Fc proteins by MALDI-TOF MS analysis. rGA733-Fc proteins obtained by stable transformation contained oligomannose glycan structures, which are typical patterns detected from ER-associated proteins (Fig. 6c).

Fig. 6
figure 6

Glycosylation patterns of plant-derived rGA733-Fc protein. rGA733-Fc protein was purified from the rGA733-FcOX transgenic tobacco leaves using a protein A column. The fractions that contain rGA733-Fc were determined by SDS-PAGE (a) and western blot analysis (b) using the GA733 specific antibody. T, total soluble protein; FT, flow through; W, flow-through by washing; E1-E13, eluted fractions. c Glycosylation patterns of the purified rGA733-Fc protein. The glycosylation patterns of plant-derived rGA733-Fc proteins were analyzed by Matrix-assisted laser desorption/ionization time-of-flight (MALDI-TOF) mass spectrometric (MS) analysis. Blue square, N-acetylglucosamine (GlcNAc); green circle, mannose

Full size image

Discussion

Plant systems have been extensively used as a biological factory for production of therapeutic recombinant proteins (Shanmugaraj et al. 2020, Takeyama et al. 2015). However, relatively few attempts have been made to develop detailed expression strategies of a particular recombinant protein to maximize its production efficiency in plants (Zahin et al. 2016). In this study, we proposed the optimized expression strategy for the production of rGA733-2 and rGA733-Fc protein in tobacco plants.

The improvement of recombinant protein production by optimization of vector backbone is an ongoing challenge for efficient production of vaccine candidates (Jakob et al. 2013; Mardanova et al. 2017). Significant increase of recombinant protein production was achieved by random mutagenesis of vector backbone or change of vector system (Jakob et al. 2013). These results suggest that evaluation of multiple vector systems or directed evolution of particular vector is necessary for maximizing the production of recombinant protein. We tested four different expression vectors to investigate the effect of vector backbone on rGA733-2 expression in tobacco plants. Among tested expression vectors, pBINPLUS and pPZP-Bar vector systems were efficient for the rGA733-2 expression while pBI121 and pPZP-200 vector systems failed to express the rGA733-2 protein (Fig. 1b). Further analysis using pPZP-200 and pPZP-Bar vector systems revealed that the length of interval between rGA733-2 expression cassette and left border of T-DNA altered expression level of rGA733-2 proteins (Fig. 2d). It is not clear how the interval affects expression of rGA733-2 protein, but the interval may affect transcription efficiency of rGA733-2 transcripts. rGA733-2 expression level was lower in tobacco plants when infected with pPZP-200-GA733 than with pPZP-2 kb-GA733 (Fig. 2b). It will be interesting to further examine effects of the interval on the expression of other recombinant proteins in plants.

The efficiency of transient expression is largely affected by both A. tumefaciens strains and host tobacco species (Nausch et al. 2012; Prado et al. 2019; Shamloul et al. 2014). rGA733-2 protein was expressed at similar levels in N. tabacum plants, regardless of the A. tumefaciens strains used for transient expressions. One the other hand, in N. benthamiana, the A. tumefaciens LBA4404 strain was more efficient than other two strains for rGA733-2 and rGA733-Fc expression. Under the conditions, we detected more than five-fold higher concentration of rGA733-2 and rGA733-Fc proteins, compared with other tested conditions (Fig. 3). Our results suggest ideal combination of A. tumefaciens strain and host tobacco species for transient expression of GA733-2 and GA733-Fc in tobacco plants, which is beneficial for rapid production of vaccine candidates in tobacco plants.

Fc chain has been known to improve stability and yield of recombinant proteins (Czajkowsky et al. 2012; De Buck et al. 2013; Obregon et al. 2006). Fusion of Fc chain to HIV p13 antigen resulted in 13-fold increase of the protein accumulation in tobacco plants (Obregon et al. 2006). Similarly, accumulation of Nanobodies VHH7 in Arabidopsis seeds increased 10- to 100-fold by fusion with an Fc chain (De Buck et al. 2013). Fusion of an Fc chain to the rGA733 resulted in approximately three-fold higher accumulation levels (Fig. 3). Thus, fusion of rGA733-2 with an Fc chain improve stability of rGA733-2 protein, leading to increase accumulation of rGA733-2 in tobacco plants.

Fc fragment fusion to recombinant proteins not only increases protein accumulation but also facilitates one-step purification procedure using the protein A column (Ashkenazi and Chamow 1997; Czajkowsky et al. 2012). Gel staining and western blot analysis confirmed that rGA733-Fc proteins was efficiently accumulated by protein A column (Fig. 6). Potential disadvantage of Fc fusion of recombinant protein is the generation of degradation products (Castilho et al. 2011; De Buck et al. 2013; Lu et al. 2012). Compared to these reports, a negligible amount of degradation fragments was detected in our study (Fig. 6b). The less degradation of rGA733-Fc might be explained by ER retention of rGA733-Fc proteins. C-terminal fusion of KDEL ER retention codon has been proved to trigger localization of fusion protein into ER (Fu et al. 2018). ER targeting of recombinant proteins increases accumulation of recombinant proteins in plants due to relatively high chaperone activity and low protease activity in ERs (Benchabane et al. 2008; Lu et al. 2012; Schouten et al. 1996; Ullrich et al. 2015). In vitro apoptosis test revealed that the rGA733-Fc was more efficient than rGA733-2 protein to induce apoptosis in cells (Fig. 5). It has been reported that fusion of Fc fragment to recombinant proteins stimulated an increase in phagocytosis, the release of inflammatory mediators, antibody-dependent cellular cytotoxicity, endocytosis, and the antigen presentation response in cells (Czajkowsky et al. 2012; Otten and van Egmond 2004). Collectively, fusion of rGA733-2 with both Fc chain and ER retention codon can be used to increase both productivity and functionality of rGA733-2 protein in tobacco plants.

For production of recombinant proteins for vaccine candidates, it is essential to determine glycosylation patterns of the produced recombinant protein (Bosch and Schots 2010; Hariharan and Kane 2020; Van Droogenbroeck et al. 2007), because plant-specific N-glycosylation patterns, such as β-1,2-xylose and α-1,3-fucose, which are generated during protein process in Golgi, are often allergenic for mammalian system (Bakker et al. 2001; Schoberer and Strasser 2018; Wilson et al. 2001). Targeting of recombinant proteins into ER is known to bypass these plant-specific glycosylation patterns (Nausch et al. 2012). Glycosylation pattern analysis revealed that the rGA733-Fc protein fused with KDEL ER retention codon had oligomannose glycan structures, which are typical glycosylation patterns found on ER-associated proteins (Schoberer and Strasser 2018) (Fig. 6c). These results indicate that the KDEL ER retention codon is suitable for preventing unnecessary glycosylation patterns on rGA733 recombinant protein. In this study, we provide useful information to develop efficient strategy to produce rGA733-2 recombinant protein in tobacco systems.