Introduction

Hepatitis A virus (HAV), a picornavirus with a single-stranded, positive-sense RNA genome, encodes a single polyprotein that is subsequently processed into structural and nonstructural proteins. The structural proteins of HAV are divided into the capsid polypeptides VP1, VP2, VP3, and VP4. VP1 appears to be the dominant structural protein among the capsid proteins. VP1 is suggested to contain a major immunodominant epitope of HAV, based on studies of the reaction of HAV with monoclonal antibodies and the use of isolated structural proteins or synthetic peptides for induction of neutralizing antibodies (Hughes et al. 1984; Stapleton and Lemon 1987; Hughes and Stanton 1985; Emini et al. 1985).

Use of recombinant subunit vaccines comprised of antigenic peptides (epitopes) is a particularly attractive strategy for wide-scale immunization when compared to formalin-inactivated HAV vaccines. Recombinant subunit vaccines offer potentially equivalent efficacy but are much safer and, in some cases, easier to produce (Arnon 1991). Plants offer unique advantages for production of subunit vaccines in terms of scale, speed, costs, yield, and safety (Fisher et al. 2004). The use of transgenic plants as factories for producing recombinant proteins has economic and qualitative benefits compared to bacterial and mammalian expression systems (Giddings 2001). This includes the potential to use existing infrastructure for growth and harvesting (Daniell et al. 2001; Fisher and Emans 2000). While recombinant VP1 has successfully been produced in Escherichia coli and insect cell systems (Ostermayr et al. 1987; Hamon et al. 1988; Lee et al. 2009), its expression in transgenic plants has not been investigated. In this study, HAV VP1 fused to the immunoglobulin Fc fragment (HAV VP1-Fc) was expressed in transgenic tomato leaves and purified using a simple one-step Protein A Sepharose affinity fractionation. Using purified recombinant protein, we performed an immunological assessment via intraperitoneal immunization of mice with the transgenic plant-derived vaccine antigen.

Materials and methods

DNA constructs

VP1 cDNA from HAV was amplified from pGEM-T/HAV-VP1 (Lee et al. 2009) via a polymerase chain reaction (PCR) using sense (5′-GCGGCCGCAGTTGGAGATGATTCAG-3′) and antisense (5′-AGATCTCTCAAATCTTTTATCTTCCTC-3′) primers. The Fc fragment of human IgG cDNA (Gene Bank accession no. AY172957) was amplified from total RNA extracted from human B cells IM-9 (Korean Cell Line Bank, Korea) using reverse transcription-PCR with oligonucleotide primers (sense: 5′-AGATCTGCGGCCGCAGTTGAGCCCAAATCTTGTGAC-3′; antisense: 5′-AGATCTACCCGGGGACAGGGAGAGGC-3′). The amplified VP1 and human Fc sequences were then cloned into the pGEM-T Vector (Promega, Madison, WI, USA). The NotI–BglII fragment of pGEM-T/VP1 and the BglII fragment of pGEM-T/Fc were inserted between the NotI and BglII sites of ImpactVector 1.3-tag (Plant Research International, Wageningen, Netherlands), which includes an N-terminal signal peptide, C-terminal tags (c-myc and His6), and an ER retention signal (KDEL). The C-terminus of VP1 was fused to the Fc fragment to yield ImpactVector 1.3-tag/VP1-Fc/c-myc-His6-KDEL. The VP1-Fc/c-myc-His6-KDEL sequences were amplified from ImpactVector 1.3-tag/VP1-Fc/c-myc-His6-KDEL via PCR using sense (5′-GATATCATGTCTCTTAGCCAGAAC-3′) and antisense (5′-ACTAGTTTAAAGTTCGTCCTTGTG-3′) primers and subcloned into the EcoRV and SpeI sites of pCsVMV-BCTVR (Kim et al. 2007), thereby yielding pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL (Fig. 1a). All of the constructs were verified using restriction enzyme mapping and DNA sequencing.

Fig. 1
figure 1

a Schematic diagram of the vector construct (pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL). Ovals mark the duplicated viral origin of replication, and REP indicates the viral gene block containing the viral L1 (REP), L2, L3 (REn), and L4 genes of Beet curly top virus. The region between the right (T R ) and left (T L ) borders also carries VP1-Fc (the sequence of VP1 fused to the Fc fragment of human IgG), the CsVMV promoter from the Cassava vein mosaic virus, the nopaline synthase (NOS) transcription terminator, the signal peptide, C-terminal tag (c-myc-His6), the ER retention signal (KDEL), and the selection marker cassette (HPT) conferring resistance to hygromycin. T-DNA regions are not drawn to scale. b Western blot analysis of protein extracts from transgenic tomato leaves. One hundred micrograms of proteins from extracts of transgenic tomato leaves were separated on 10 % SDS-polyacrylamide gel, and the expression of recombinant VP1-Fc was determined by Western blot analysis using anti-human IgG (Fc-specific). WT means protein extracts of wild type tomato leaves. Lanes 1–14 mean protein extracts isolated from putative transgenic tomato leaves transformed with recombinant Agrobacterium harboring pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL. SDS-PAGE followed by silver staining (c) and western blot analysis (d) of the recombinant VP1-Fc purified from transgenic tomato leaves. M indicates molecular weight markers. Before and after affinity purification is shown in lanes (1) and (2), respectively. The arrows indicate the recombinant VP1-Fc

Full size image

Recombinant Agrobacterium tumefaciens strains and transformation

The expression vector pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL was introduced into Agrobacterium tumefaciens LBA4404 using electroporation, as described by Ainsworth et al. (1996). Tomato (Lycopersicon esculentum) seeds were germinated on Murashige and Skoog (MS) medium at 25 °C under constant light. Explants of tomato leaves were incubated for 2–3 min with A. tumefaciens LBA4404 harboring pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL. The explants were transferred to MS basal medium without antibiotics. After two days of co-cultivation in the dark, the explants were transferred to MS shooting medium containing 1 mg/L zeatin, 0.02 mg/L indole-3-acetic acid (IAA), 250 mg/L cefotaxime, and 1 mg/L hygromycin B under constant light. After 4–5 weeks, regenerated shoots were cut from the explants and dipped into MS rooting medium containing 250 mg/L cefotaxime and 1 mg/L hygromycin B. After generation of roots, plants were transferred into soil.

Western blot analysis

Proteins were extracted from transgenic tomato leaves, as described by Ainsworth et al. (1996) followed by electrophoresis on 10 % SDS-polyacrylamide gel. The fractionated proteins were transferred to nitrocellulose membrane (Hybond-C extra; Amersharm Pharmacia Biotech, Pittsburgh, PA, USA). The membranes were pre-incubated with blocking solution [5 % (w/v) non-fat dry milk in TBS containing 0.1 % Tween-20] for 1 h, incubated with goat anti-human IgG (Fc-specific, diluted in blocking solution to 1:10,000), and probed with peroxidase-conjugated anti-goat IgG (diluted in blocking solution to 1:5,000). Protein bands were detected using an enhanced chemiluminescent (ECL) western blotting detection reagent (GE Healthcare, Sweden).

Purification of recombinant VP1-Fc

All steps were performed at 4 °C. Recombinant VP1-Fc was purified using affinity fractionation with Protein A Sepharose 4 Fast Flow (GE Healthcare) according to the manufacturer’s recommendations. Protein extracts were dialyzed with binding buffer (20 mM sodium phosphate buffer, pH 7.0) and applied to a Protein A Sepharose 4 Fast Flow column. Weakly bound contaminating proteins were washed from the beads using binding buffer. Recombinant VP1-Fc was then eluted using 0.1 M glycine (pH 3.0), neutralized with 1 M Tris–HCl (pH 9.0), and dialyzed in PBS. Protein concentrations were determined using a Bradford Protein Assay Kit (Bio-Rad, Hercules, CA, USA) with BSA as a standard.

Animal experiments and antibody detection

Five-week-old female BALB/c mice were divided into 2 groups (5 mice per group) and immunized intraperitoneally with PBS and VP1-Fc purified from transgenic tomato leaves. Mice were injected with 20 μg of VP1-Fc (or PBS as a control) with Freund’s adjuvant (FA) four times at 2-week intervals. Freund’s complete adjuvant (FCA) was used in the first immunization, and Freund’s incomplete adjuvant (FIA) was used for subsequent booster injections. All animal experiments were performed in accordance with protocols approved by the Institutional Animal Care and Use Committee of Kyung Hee University.

To determine specific antibodies produced in serum after intraperitoneal immunization, blood was collected from the retro-orbital plexus at one week after the third and fourth immunization (days 42 and 56), and each serum sample was separated using centrifugation at 10,000×g and stored at −70 °C until use. Anti-VP1-Fc IgG antibodies in serum were assayed using enzyme-linked immunosorbent assay (ELISA). Briefly, a 96-well ELISA plate was coated overnight at 4 °C with 100 μL of purified VP1-Fc (2 μg/mL) per well in a coating buffer (0.05 M carbonate–bicarbonate buffer, pH 9.0). The plate was washed three times with 200 μL of PBST (PBS with 0.05 % Tween-20). Serum diluted in PBST was added to each well, and the plate was incubated for 1 h at RT. After washing with PBST, 100 μL of peroxidase-conjugated anti-mouse IgG (diluted to 1:30,000; Sigma, St. Louis, MO, USA) was added per well and incubated for 1 h at RT. Finally, the plate was washed and developed for 30 min with TMB (3,3′,5,5′-tetramethylbenzidine) in phosphate–citrate buffer (pH 5.0) containing 0.002 % hydrogen peroxide. The reaction was then stopped by the addition of 50 μL of 2 M H2SO4 to each well. The absorbance was determined at 450 nm using an ELISA reader (Bio-Tek Inc., Winooski, VT, USA).

Cross-reactivity of anti-HAV VP1-Fc antibodies to native HAV antigens

Cross-reactivity of the antibodies induced by intraperitoneal immunization with VP1-Fc in BALB/c mice and HAV antigens (Meridian Life Science, Memphis, TN, USA) was investigated. A 96-well ELISA plate was coated with 100 μL of individual HAV antigen (1 μg/mL) per well in a coating buffer and left overnight at 4 °C. The plate was washed three times with 200 μL of PBST. Sera were then diluted to 1:300 in PBST, added to each well, and incubated for 1 h at RT. After washing with PBST, the plate was incubated with 100 μL of peroxidase-conjugated anti-mouse IgG (diluted in PBST to 1:5,000) per well for 1 h at RT. The plate was washed and color-developed according to the procedures described above. Cross-reactivity of the antibodies raised after intraperitoneal immunization with VP1-Fc in BALB/c mice and HAV antigens was also investigated using western blot analysis. One microgram of HAV antigen was resolved on SDS-polyacrylamide gels and transferred onto nitrocellulose membranes. The membranes were blocked with blocking solution and then incubated with PBS-immunized or VP1-Fc-immunized serum (diluted in blocking solution to 1:100). This was followed by incubation with peroxidase-conjugated anti-mouse IgG (diluted in blocking solution to 1:10,000); the band was detected using the ECL western blot detection reagent.

Splenocyte culture and cytokine detection

The procedure for isolating cells from the spleen was carried out as described previously (Lee et al. 2009). Briefly, spleens were surgically removed and gently crushed using the plunger of a disposable syringe on a Falcon cell strainer (BD Bioscience, Franklin Lakes, NJ, USA). Red blood cells from the spleen were removed using ACK buffer [0.15 M NH4Cl, 10 mM KHCO3, 0.1 mM ethylenediaminetetraacetic acid (EDTA) at pH 7.3]. After three washes in RPMI-1640 medium, cells were resuspended in RPMI-1640 medium containing 10 % FBS, 2-mercaptoethanol, and antibiotics, and were seeded at a density of 5 × 106 cells/mL onto a 24-well plate (Nunc, Denmark) for cytokine assay. Splenocytes were cultured for 48 or 72 h in the presence of an appropriate antigen (VP1-Fc or PBS as a control). Cell culture supernatants were applied to a sandwich ELISA system (BD Bioscience) for detection of IFN-γ and IL-4 according to the manufacturer’s instruction. The levels of IFN-γ and IL-4 were measured from 48 to 72 h supernatant samples, respectively.

Statistical analysis

All data are presented as mean ± standard error (SE). Student’s t test was used to compare different data groups (*p < 0.05, **p < 0.01, ***p < 0.001).

Results and discussion

A total of 14 independent transgenic tomato plants were regenerated after transformation with recombinant Agrobacterium tumefaciens LBA4404 harboring pCsVMV-BVTVR-VP1-Fc/c-myc-His6-KDEL. The pCsVMV-BCTVR-VP1-Fc/c-myc-His6-KDEL vector system includes two Beet curly top virus (BCTV) replication origins flanking the intact viral Rep and REn genes and VP1-Fc cDNA under control of the strong constitutive CsVMV promoter (Fig. 1a). The BCTV-based transient expression vector system has successfully been developed to express a high level of recombinant green fluorescence protein (GFP) and human HAV VP1 in Nicotiana benthamiana (Kim et al. 2007; Chung et al. 2011). The Fc fragment of the antibody and the ER retention signal KDEL were fused to the C-terminus of VP1 for facilitating expression, detection, and purification of antigens. Western blot analysis to screen transgenic tomato plants expressing recombinant VP1-Fc among the 14 lines showed that recombinant VP1-Fc was expressed from leaves of nine T0 transgenic plants (Fig. 1b). Seven of them showed strong recombinant VP1-Fc expressions. Recombinant VP1-Fc was expressed as a band with molecular mass of approximately 68 kDa, which is higher than the predicted molecular weight (63 kDa) for the recombinant protein. Analysis of the chimeric VP1-Fc peptide sequence using the NetGlyc 3.1 server (http://www.cds.dk/services) and the OGPET 1.0 server (http://ogpet.utep.edu/OGPET) indicates that the chimeric VP1-Fc protein has two potential sites (N278and N429) for N-linked glycosylation, but does not contain any consensus sites for O-linked glycosylation. In our previous work (Chung et al. 2011), the molecular weight of recombinant VP1-Fc transiently produced from Agrobacterium-infiltrated leaves of Nicotiana benthamiana was reduced to approximately 63 kDa by digestion using PNGase F, an endoglycosidase releasing N-linked oligosaccharides (data not shown). Therefore, the difference between expressed and predicted molecular weights is most likely a consequence of glycosylation.

Recombinant VP1-Fc was rapidly purified to near homogeneity using a Protein A Sepharose affinity purification procedure. The purity of the protein was analyzed using SDS-PAGE and silver staining (Fig. 1c). Western blot analysis using an anti-human IgG (Fc-specific) further confirmed the identity of the purified protein (Fig. 1d). No contaminating proteins were visible on silver nitrate-stained SDS-PAGE gel (Fig. 1c, lane 2). We obtained approximately 0.57 μg of purified recombinant VP1-Fc from one gram of transgenic tomato leaf material. The Fc fragment of the chimeric VP1-Fc allowed the efficiency of purification to be improved using the Protein A-based method, as compared to the Ni-NTA resin-based method of a low yield <10 % recovery (data not shown). Therefore, the Fc-tag can be used as an alternative to a His-tag for single-step purification of recombinant proteins from other plant expression systems. Using non-reducing conditions, recombinant VP1-Fc migrated at a size consistent with its assembly into dimers (data not shown). The assembly status of a chimeric peptide has been suggested to be a determinant of stability for a recombinant HIV-1 p24-immunoglobulin fusion molecule (Obregon et al. 2006). The ability of these molecules to dimerize might confer some structural advantage in terms of recombinant protein stability when expressed in transgenic plants. The immunogenicity of the recombinant VP1-Fc derived from transgenic tomato plants was examined in animal experiments using BALB/c mice. ELISA revealed that transgenic plant-derived recombinant VP1-Fc, when administered intraperitoneally with FA, produced specific IgG antibodies in serum (Fig. 2a). Levels of IgG antibodies specific for the transgenic plant-derived recombinant VP1-Fc were significantly increased after the third injection, indicating the effective stimulatory effect of VP1-Fc on B cell differentiation and maturation in mice. To our knowledge, this indicates that transgenic plant-derived VP1 is capable of inducing a strong IgG antibody response in mice after intraperitoneal immunization. The cross-reactivity of antibodies developed in intraperitoneally immunized BALB/c mice against HAV antigens (formalin-inactivated HAV pHM175 strain; Meridian Life Science) was also investigated to determine the differences in the physical properties of the recombinant and viral proteins. Two serum groups were used for non-competitive indirect ELISA analysis. One group was from mice immunized with recombinant VP1-Fc and FA while the other group (as a control) was from mice immunized with PBS and FA. When serum samples at 1:300 dilution were applied to wells coated with HAV antigens, the mean ELISA values were 0.61 and 0.44, respectively, for the two groups (Fig. 2b). Two serum groups described above were further used for western blot analysis to determine whether the antibodies developed in intraperitoneally immunized BALB/c mice can detect HAV antigens. As shown in Fig. 2c, a 33-kDa band corresponding to a band for VP1 protein of HAV antigens was detected when sera from mice immunized with recombinant VP1-Fc and FA was used in western blot analysis. Taken together, these results indicate that antibodies produced by mice against transgenic plant-derived recombinant VP1-Fc most likely recognize epitopes in the target viral proteins.

Fig. 2
figure 2

a Serum antibody response to VP1-Fc in BALB/c mice. BALB/c mice were immunized with 20 μg of purified VP1-Fc (or PBS as a control) and FA. Seven days after the third immunization, specific IgG antibodies in antisera were measured using ELISA. b ELISA titers of sera from BALB/c mice immunized with VP1-Fc antigen against hepatitis A viral antigens. c Western blot analysis of HAV antigen using sera from BALB/c mice immunized with VP1-Fc antigen. Left panel is the western blot analysis using PBS-immunized serum. Right panel is the western blot analysis using VP1-Fc-immunized serum. The arrow indicates the location of a band that reacted with VP1-Fc-immunized sera against HAV antigen

Full size image

Splenocytes prepared from immunized mice were cultured in the presence of each antigen, and synthesis of IFN-γ and IL-4 was examined. As shown in Fig. 3a, splenocytes from mice immunized with VP1-Fc and cultured in the presence of the antigen (VP1-Fc) secreted large amounts of IFN-γ. This result indicates that VP1-Fc promotes formation of Th1-type cytokines. When splenocytes from mice immunized with VP1-Fc were cultured with VP1-Fc, production of IL-4 was increased approximately twofold more than when cultured in the absence of VP1-Fc (Fig. 3b). This indicates that VP1-Fc also promotes formation of Th2-type cytokines. The Th1 response can lead to a cytotoxic T cell response and virus clearance. IFN-γ plays an important role in the VP1-mediated T cell response since elevation of the IFN-γ level is considered to be related to Th1-type immunity. The Th2 response can eventually lead to a B cell response and production of specific antibodies. The mutually antagonistic effects of IL-4 and IFN-γ can regulate the Th1/Th2 balance and subsequent polarization (Rengarajan et al. 2000). For example, exposure of naive helper T cells to IL-4 causes differentiation into Th2 cells at the beginning of an immune response. Also, under a threshold level of IL-4, Th2 development is greatly favored over Th1, even if IFN-γ and IL-12 are present (Seder and Paul 1994). Although cell-mediated immunity is known to be responsible for controlling intracellular infections such as HAV infection, humoral immunity is essential for controlling extra-cellular infections. Therefore, co-stimulation of both cellular and humoral immunity has been proposed to be necessary for a host to control infections (Bertoletti and Gehring 2006). We observed a significant increase in the secretion of both IFN-γ (Th1-like) and IL-4 (Th2-like) with the recombinant VP1-Fc protein compared with PBS, indicating co-stimulation of both Th1 and Th2 immune responses. Our limited analysis of cytokine profiles from spleen cells implies that the recombinant VP1-Fc protein derived from transgenic plants can be an ideal candidate as an effective vaccine for control of HAV infection.

Fig. 3
figure 3

The concentrations of IFN-γ and IL-4 in splenocyte cell culture supernatants. Splenocytes from mice were stimulated with purified VP1-Fc antigen (or PBS as a control). After 48 or 72 h, supernatants were collected to examine the levels of IFN-γ (a) and IL-4 (b) using commercially available mice cytokine ELISA kits. The levels of IFN-γ and IL-4 were measured from 48 and 72 h supernatant samples, respectively

Full size image

In summary, we report a strategy for the production of safe and inexpensive VP1 vaccine antigens against HAV. In this study, VP1 fused to the immunoglobulin Fc fragment (VP1-Fc) was expressed in transgenic tomato plants using a BCTV-replicating vector system. The high-yield potential of the BCTV-replicating vector system enhances the feasibility of production of plant-derived vaccine in transgenic tomato leaves. This indicates that the BCTV-based expression system can be useful for expression of recombinant vaccine antigens such as HAV VP1-Fc in transgenic plants. Recombinant chimeric proteins of HAV capsid protein VP1 and the Fc antibody fragment from transgenic tomato plants elicited production of specific IgG in serum after intraperitoneal immunization. Our findings show that functional expression of the major immunogen VP1-Fc using a replicating vector system based on BCTV in transgenic plants provides a convenient source of recombinant VP1-Fc for research into vaccine development.