Single-chain variable fragments (scFvs), as well as other recombinant antibodies, have become powerful tools for therapy and analysis in both human and veterinary medicine. Transgenic plants have been developed as an efficient production system for these recombinant proteins, achieving high yields at moderate cost. Seeds are especially useful organs for molecular farming because of their high protein content and their ability to keep proteins functional during extended storage in ambient conditions. Several concepts have been applied to optimize protein accumulation in seeds, such as the use of specific promoters and transcription enhancers. In this paper, we present a new strategy to enhance substantially the expression of scFvs in transgenic tobacco seeds, based on C-terminal fusions to elastin-like polypeptides (ELPs). This strategy resulted in a 40-fold increase in scFv accumulation, with levels approaching 25% of total soluble protein. The fusion proteins show specific activities and affinities comparable with the properties of the corresponding scFvs. This strategy thus opens up new ways to improve greatly the production of recombinant antibodies in plant seeds.
Recombinant antibodies are valuable tools for analytical purposes and therapy, with several applications in human and veterinary medicine (Reiter and Pastan, 1998). In addition to complete immunoglobulins (Igs), several types of recombinant antibody, such as Fab fragments, single-chain variable fragments (scFvs), minibodies and single-domain antibodies, have been developed (Bird et al., 1988; Muyldermans, 2001). Since the first report on the expression of recombinant antibodies in transgenic plants (Hiatt et al., 1989), several expression systems have been tested to produce large amounts of recombinant antibodies in transgenic plants (Ma et al., 2003; Schillberg et al., 2003). Endoplasmic reticulum retention seems to be important for the optimal accumulation of correctly folded active recombinant antibodies (Conrad and Fiedler, 1998). In plant seeds, high expression, driven by seed-specific promoters, combined with the good stability of scFvs, even after long-term storage at ambient temperature, has resulted in high-level accumulation (Fiedler and Conrad, 1995; De Jaeger et al., 2002).
In addition to the enhancement of expression by the use of optimized promoters, fusions to peptides or proteins may result in higher levels of accumulation of transgenic products. As we have demonstrated a twofold increase in spider silk protein accumulation in tobacco leaves after fusion to 100×ELP (Scheller et al., 2004), we were interested to determine whether this could be transferred to other proteins, such as scFvs. Two scFvs, one binding to 2-phenyl-oxazol-5-one (oxazolone, Ox) and the other to kresoxim-methyl (Kres), were chosen, because these antigens do not appear in plants, and therefore plant development and growth are not influenced by the high expression of these scFvs, as already shown for both scFvs (Fiedler and Conrad, 1995; Leps, 2002). Synthetic elastin-like polypeptides (ELPs) consist of repeats of the pentapeptide Val-Pro-Gly-Xaa-Gly, where Xaa is any amino acid except proline (Meyer and Chilkoti, 1999). This sequence is derived from the characteristic repeat motif, VPGVG, found in native mammalian elastin. This protein is soluble in water below its transition temperature, and becomes reversibly insoluble if the temperature is raised above its transition temperature. This property can be used for the purification of the soluble fusion protein by ‘inverse transition cycling’ (Meyer and Chilkoti, 1999; Scheller et al., 2004).
As it is more convenient to accumulate scFvs in seeds rather than leaves, we used seed-specific promoters. We generated fusion proteins of scFvs and ELPs, and found that the accumulation of active scFv fusions dramatically increased, up to a level 40-fold greater than that of unfused scFvs, when expressed in tobacco seeds under the control of a seed-specific promoter. The strategy proposed here could improve greatly the production of recombinant antibodies in plant seeds for use in human and veterinary medicine.
scFv-ELP fusion proteins are expressed in large amounts in seeds of transgenic tobacco plants
The expression constructs for scFv antibodies, one binding to Ox and the other recognizing Kres, under the control of the seed-specific promoters legumin B4 (LeB4) or unknown seed protein (USP), were constructed. For some constructs, the scFv antibody was C-terminally fused to 100 repeats of the ELP (Figure 1). ELPs are composed of repeats of the pentapeptide Val-Pro-Gly-Xaa-Gly, where Xaa is valine, glycine or alanine (Meyer and Chilkoti, 1999). The LeB4 signal peptide and the KDEL signal at the C-terminus should cause endoplasmic reticulum retention, which has been shown to be optimal for seed-specific expression of scFvs (Fiedler et al., 1997). A c-myc tag was used for the detection of scFvs in Western blotting and enzyme-linked immunosorbent assay (ELISA). Transgenic tobacco plants were generated by leaf disc transformation (Zambrinski et al., 1983), and seeds were harvested and tested by Western blotting (data not shown).
High-expressing plants, as detected by Western blotting, were chosen (seven of 50 for USP-SS-anti-Ox-ELP; eight of 50 for LeB4-SS-anti-Ox-ELP; six of 50 for USP-SS-anti-Kres-ELP; and 11 of 50 for LeB4-SS-anti-Kres-ELP), and proteins were extracted from seeds with 50 mm Tris-HCl, 200 mm NaCl, 5 mm ethylenediaminetetraacetic acid (EDTA), 0.1% Tween 20, pH 8.0. Extracts were analysed by sodium dodecylsulphate-polyacrylamide gel electrophoresis (SDS-PAGE) and silver staining (Figure 2) and by Western blotting (Figure 3). A remarkable enhancement of seed-specific accumulation of the fusion proteins was observed compared with the corresponding unfused scFvs. The prominent band of about 75 kDa was demonstrated to be the scFv-ELP fusion protein by Western blotting. Using this extraction method, scFv is present in the seed extracts up to a concentration of 25% of total soluble protein (TSP) (Figure 3A, lanes 4–7; Figure 3B, lanes 1 and 2), as determined by the use of interleukin-6 (IL-6)-c-myc as a standard (Figure 3A, lanes 1–3; Figure 3B, lanes 3–5). The intensity of the scFv-ELP band corresponds in concentration to the main legumin band of tobacco seeds (Figure 2A). Comparable results were achieved with both anti-Kres-scFv-ELP and anti-Ox-scFv-ELP and with the two different promoters.
For controls, scFvs were expressed in seeds using identical constructs but lacking ELP coding sequences. These plants were grown at the same time and in the same position in the glasshouse as scFv-ELP plants. The seeds were harvested, extracted and tested by Western blot analysis. In this experiment, much lower expression levels were achieved, which did not exceed 0.5% of TSP (Figure 3). scFv bands could not be identified in silver-stained gels (Figure 2B).
scFv-ELP fusion proteins from plant seeds show similar binding behaviour to the corresponding unfused scFvs
A major question to be answered was whether such fusion proteins produced in plant seeds showed the same antigen-binding properties as the corresponding scFvs produced in plant seeds. For LeB4-SS-anti-Ox-scFv-ELP (Z3L), USP-SS-anti-Ox-scFv-ELP (Z3U), LeB4-SS-anti-Ox-scFv (V3L) and USP-SS-anti-Ox-scFv (V3U) seed extracts, the binding to oxazolone-bovine serum albumin (Ox-BSA) was measured by an indirect ELISA (Figure 4A,B). In Figure 4A, the binding of anti-Ox scFv, expressed in seeds under the control of the LeB4 promoter (V3L) or USP promoter (V3U), is shown. In Figure 4B, the binding of anti-Ox scFv-ELP fusion proteins, expressed in seeds under the control of the LeB4 promoter (Z3L) or USP promoter (Z3U), is shown. Excessive ELISA values beyond the range of linearity were included in the figures to show the difference between scFv and scFv-ELP extracts. All scFv-ELP-containing extracts showed higher activity, corresponding to the larger amount of recombinant antibody fragments in these extracts. However, the differences in ELISA activity between different seed extracts containing scFv-ELP fusions, shown in Figure 4, were not paralleled in Western blotting analysis (Figure 3) or silver staining (Figure 2). Therefore, we determined whether ELP fusion may change the binding behaviour of scFv molecules, or whether a part of the recombinant scFv is inactive. For this purpose, we diluted the seed extracts (V3U/20 and Z3U/13) so that, under the ELISA conditions used, comparable basal values (no competition, OD495 = 2.00 and OD495 = 1.47, respectively) could be measured. Highly comparable competition curves were measured for both the scFv and scFv-ELP fusion proteins. Fifty per cent inhibition was achieved at 22 nm and 28 nm free antigen for scFv-ELP and scFv, respectively. This shows that the 100×ELP does not generally influence the antigen-binding behaviour after seed-specific expression. The proportion of inactive antibodies therefore seems to be different in individual transgenic lines. Similar results were achieved for anti-Kres-scFv-ELP and anti-Kres-scFv produced in tobacco seeds (data not shown).
An essential improvement in transgenic seeds as a production system for recombinant scFvs was achieved by the ELP fusion method demonstrated here. The high activity measured by indirect ELISA and the comparable affinity constants measured by competitive ELISA emphasize that transgenic seeds can serve as a reliable source of therapeutic recombinant antibodies. Sub-bands visible only in some lines (Figure 2), and the rather high activity in the seed extracts of one line (Figure 4B), showed that suitable production lines must be selected carefully. By simple selection, lines with a low level of inactive antibodies could be isolated and propagated further. The technology may possibly be further developed by combination with optimized seed promoters (De Jaeger et al., 2002). Future studies should also include the production of complete antibodies as ELP fusions, as well as other types of recombinant antibody and protein, to develop this method further as a general plant production tool. ELPs are animal-derived proteins. Therefore, they should not cause problems of unwanted immunogenicity, but this must be tested in detail in further experiments. The attachment of 100×ELP to spider silk proteins led to a twofold increase in spider silk protein accumulation in tobacco leaves (Scheller et al., 2004).
The technology described here can enhance the accumulation of transgenic scFv proteins in tobacco seeds by a factor of 40. In the future, it will be interesting to determine whether this strategy of enhanced seed-specific accumulation of ELP fusion proteins could be transferred to other fields, such as vaccine and enzyme production, with comparable accumulation rates. Such proteins may also be less sensitive than scFvs to higher temperatures, such that inverse transition cycling could be used for purification (Meyer and Chilkoti, 1999).
Construction of binary vectors for anti-Ox-scFv and anti-Kres-scFv under the control of USP and LeB4 promoters for the transformation of tobacco plants
The anti-Ox-ScFv gene was amplified by polymerase chain reaction (PCR) using the primers 5′ScFvoxa (5′ CGCGGATCCGAGGTCAAGCTGCAGGAGTCTG 3′) and 3′ScFvoxa (5′ TCCCCCGGGCCGTTTCAGCTCCA 3′), with an anti-Ox expression cassette (Fiedler and Conrad, 1995) as template. The purified product was digested with BamHI and SmaI. The anti-Kres-ScFv gene was amplified by PCR using the primers 5′ScFvKres-B8 (5′ TCCCCCGGGAAAGATGTTTTGATGACCCAGAC 3′) and 3′ScFvKres-B8 (5′ TCCCCCGGGTGAGGAGACGGTGACTGAG 3′), with an scFv coding fragment (scFv-B8) (Leps, 2002) as template. The purified product was digested with SmaI. The PCR products were cloned into pRTRA7/3-USP-SO1 and pRTRA7/3-USP-SO1-100×ELP (unpublished results), digested with BamHI and NaeI, or SmaI and NaeI, respectively. The generation of the 100×ELP fusion protein has been described previously (Scheller et al., 2004). The resulting plasmids pRTRA7/3-USP-anti-oxa-ScFv, pRTRA7/3-USP-anti-oxa-ScFv-100×ELP, pRTRA7/3-USP-anti-kres-ScFv and pRTRA7/3-USP-anti-kres-ScFv-100×ELP were digested with HindIII, and the fragments containing the cassette USP promoter/legumin signal/ScFv/c-myc-tag/KDEL/CaMV 35S terminator or USP promoter/legumin signal/ScFv/c-myc-tag/100×ELP/KDEL/CaMV 35S terminator were cloned into pCB301-Kan. The vector pCB301-Kan is based on the vector pCB301 (Xiang et al., 1999), and was produced by the transfer of a BglII-BamHI-T-DNA-fragment of the vector pBIN19 (Bevan, 1984).
The LeB4 promoter was amplified by PCR using the primers M13 (5′ CGCCAGGGTTTTCCCAGTCACGAC 3′) and 3′LeB4 (5′ CATGCCATGGTGACTGTGATAGTAAACAAC 3′), with an LeB4 promoter-containing construct (Fiedler and Conrad, 1995) as template. The PCR product was digested with HincII and NcoI and cloned into the corresponding sites of the plasmids pRTRA7/3-USP-anti-oxa-ScFv, pRTRA7/3-USP-anti-oxa-ScFv-100×ELP, pRTRA7/3-USP-anti-kres-ScFv and pRTRA7/3-USP-anti-kres-ScFv-100×ELP. The resulting plasmids pRTRA7/3-LeB4-anti-oxa-ScFv, pRTRA7/3-LeB4-anti-oxa-ScFv-100×ELP, pRTRA7/3-LeB4-anti-kres-ScFv and pRTRA7/3-LeB4-anti-kres-ScFv-100×ELP were digested with HindIII and the fragments containing the cassette LeB4 promoter/legumin signal/ScFv/c-myc-tag/KDEL/CaMV 35S terminator or LeB4 promoter/legumin signal/ScFv/c-myc-tag/100×ELP/KDEL/CaMV 35S terminator were cloned into pCB301-Kan.
Transformation of tobacco
The constructs were transferred into Agrobacterium tumefaciens C58C1 (pGV2260; Deblaere et al., 1985) by electroporation. Tobacco (Nicotiana tabacum cv. SNN) leaf discs were transformed as described elsewhere (Zambrinski et al., 1983). Regenerated transgenic plants were grown in vitro on Murashige–Skoog medium containing 100 mg/L kanamycin. Plants containing the transgene were grown to maturity in the glasshouse, and ripe seeds were selected for further investigation after Western blot analysis.
Western blot analysis and silver staining
TSP was extracted from seeds of transgenic tobacco by homogenization under liquid nitrogen in 50 mm Tris-HCl, 200 mm NaCl, 5 mm EDTA, 0.1% Tween 20, pH 8.0. The homogenate was centrifuged for 5 min at 4 °C and 16 000 g. Protein concentrations in the supernatants were determined by Bradford assay (Bio-Rad, Hercules, CA, USA). Western blot analysis with anti-c-myc monoclonal antibody (Munro and Pelham, 1987) was carried out according to the method described elsewhere (Conrad et al., 1997). For blotting, Protran Nitrocellulose Transfer Membrane, BA 85 0.45 µm (Schleicher and Schuell; Whatman, Dassel, Germany) was used. The electro-transfer was performed in 25 mm Tris, 250 mm glycine, 0.1% SDS, 20% methanol, pH 8.3 at 18 V, 200 mA for 12 h. Blocking was performed for at least 2 h with 5% Marvel fat-free dried skimmed milk in 20 mm Tris-HCl, pH 8.0, 180 mm NaCl. Immunological detection was performed by incubation with anti-c-myc (9E10) supernatant at a suitable dilution, followed by anti-mouse Ig horseradish peroxidase linked whole antibody (from sheep) (Amersham Bioscience, Piscataway, NJ) (1 : 2000). The reaction was detected with the ECL Western blotting analysis system (Amersham). IL-6-c-myc-his-tag protein was expressed from pet23a-IL-6-c-myc-his-tag as inclusion bodies in E. coli. The inclusion bodies were dissolved and the protein was refolded as described previously (Arcone et al., 1991). Purification was completed using Ni2+-nitrilotriacetate (Ni-NTA) chromatography as shown by silver staining (data not shown). The extinction coefficient of IL-6-c-myc-his-tag was calculated to be 10 100/cm/m. The concentration of the pure IL-6 protein fraction was quantified by UV spectroscopy and used as a loading control in Western blot experiments. Western blots were quantified by comparing the intensity of the bands with this defined standard. Apparent molecular weights of proteins were estimated by comparison with the Benchmark pre-stained protein marker (Gibco BRL, Rockville, MD, USA). Silver staining was performed according to the manufacturer's instructions (http://www.Carl-Roth.de).
ELISAs were performed as described elsewhere (Conrad et al., 1997). Briefly, Ox-BSA was absorbed to ELISA plates at alkaline pH, blocked with 3% BSA in phosphate-buffered saline containing Tween 20, and dilutions of scFv- and scFv-ELP-containing seed extracts were added. The binding was detected using anti-c-myc-tag antibody 9E10 and anti-mouse IgG (whole molecule, developed in rabbit) alkaline phosphatase conjugate (Sigma-Aldrich, Munich, Germany). The cleavage of p-nitrophenol phosphate (p-nitrophenyl phosphate tablets, Sigma) was measured at 495 nm. For competitive ELISA, antigen (Ox-BSA) and scFv were pre-incubated for 1 h at room temperature with varying antigen concentrations and applied to solid-phase fixed Ox-BSA. Thereafter, the binding was detected using anti-c-myc-tag antibody 9E10 and anti-mouse IgG (whole molecule, developed in rabbit) alkaline phosphatase conjugate (Sigma), as for the indirect ELISA. For all ELISA experiments, five parallel measurements were made for each variant, and standard deviations were calculated.
We thank Christine Helmold, Elisabeth Nagel and Isolde Tillack for technical help.
We also acknowledge Jeremy Timmis for help with language and style. Parts of the experiments described here were performed within the framework of the Pharma Planta Project supported by the European Community.