|Home | About | Journals | Submit | Contact Us | Français|
Plant viral vectors have great potential in rapid production of important pharmaceutical proteins. However, high-yield production of heterooligomeric proteins that require the expression and assembly of two or more protein subunits often suffers problems due to the “competing” nature of viral vectors derived from the same virus. Previously we reported that a bean yellow dwarf virus (BeYDV)-derived, three-component DNA replicon system allows rapid production of single recombinant proteins in plants (Huang et al. 2009). In this article, we report further development of this expression system for its application in high-yield production of oligomeric protein complexes including monoclonal antibodies (mAbs) in plants. We showed that the BeYDV replicon system permits simultaneous efficient replication of two DNA replicons and thus, high-level accumulation of two recombinant proteins in the same plant cell. We also demonstrated that a single vector that contains multiple replicon cassettes was as efficient as the three-component system in driving the expression of two distinct proteins. Using either the non-competing, three-vector system or the multi-replicon single vector, we produced both the heavy and light chain subunits of a protective IgG mAb 6D8 against Ebola virus GP1 (Wilson et al. 2000) at 0.5 mg of mAb per gram leaf fresh weight within 4 days post infiltration of Nicotiana benthamiana leaves. We further demonstrated that full-size tetrameric IgG complex containing two heavy and two light chains was efficiently assembled and readily purified, and retained its functionality in specific binding to inactivated Ebola virus. Thus, our single-vector replicon system provides high-yield production capacity for heterooligomeric proteins, yet eliminates the difficult task of identifying non-competing virus and the need for co-infection of multiple expression modules. The multi-replicon vector represents a significant advance in transient expression technology for antibody production in plants.
The use of plants as factories to produce important pharmaceutical proteins, including vaccine antigens and therapeutic monoclonal antibodies (mAbs), is attractive for several reasons. First, plants can produce large volumes of proteins efficiently and sustainably and, under certain conditions, can have significant advantages in manufacturing costs (Giddings 2001; Hood et al. 2002). Second, the growth of plants does not require animal- or human-derived nutrients and therefore has minimal risks of contamination with animal or human pathogens and toxins. Third, plants possess an endomembrane system and secretory pathway that are similar to mammalian cells (Vitale and Pedrazzini 2005), permitting appropriate post-translational modification of recombinant proteins, which are often critical to their proper functions.
Since the first demonstration of mAb expression in transgenic plants in 1989 (Hiatt et al. 1989), many different forms of mAbs have been produced in plant systems (reviewed in (Chen 2008; Ma 2003; Ma et al. 2005; Stoger et al. 2004)) using either transient, viral based expression systems (Canizares et al. 2005; Yusibov et al. 2006) or stably transgenic plants (Floss et al. 2007; Giddings et al. 2000; Twyman et al. 2005). The latter strategy suffers from the long time frame required to generate transgenic plants and the generally low protein yields (<40 µg/g of fresh biomass). In contrast, plant viral vectors have the potential to rapidly produce high-levels of foreign proteins owing to their efficient replication and the resulting high copy numbers of gene of interest (Lico et al. 2008). However, until recently it has been difficult to efficiently express multi-component mAbs with plant viral vectors, because co-delivery of viral vectors built on the same virus backbone always results in early segregation and subsequent preferential amplification of one of the vectors in one cell – a common scenario of “competing replicons” (Dietrich and Maiss 2003; Diveki et al. 2002; Hull and Plaskitt 1970). This problem has been recently overcome by utilizing two sets of vectors derived from non-competing tobacco mosaic virus (TMV) and potato virus X (PVX), to produce full-size IgG at levels as high as 0.5 mg of mAb per gram leaf fresh weight (Giritch et al. 2006). However, it will be a difficult task to identify additional viruses compatible with both TMV and PVX expression systems to allow efficient co-expression and assembly of three or more distinct subunit proteins such as secretory IgA and IgM in same cells. Thus, there is a need to develop an advanced transient expression system which consists of a minimum number of vectors but still permits high-level co-expression and assembly of multiple subunit proteins.
Previously we reported that a bean yellow dwarf virus (BeYDV)-derived, three-component DNA replicon system allows rapid high-yield production of single recombinant proteins in plants (Huang et al. 2009). In the present study, we showed that this three component replicon system permits simultaneous efficient replication of two separate DNA replicons and high-level accumulation of two proteins encoded by the replicons. Moreover, we constructed a single vector that contains multiple replicon cassettes and demonstrated that it was as efficient as the three-component system in directing expression of two distinct protein molecules. Using either the non-competing, three-vector system or the multi-replicon single vector, we demonstrated co-expression of green fluorescent protein (GFP) and Discosoma sp. red fluorescent protein (DsRed). Using proteins that fluoresce at different wavelengths, it was possible to show that two different foreign proteins were expressed in the same cells. Moreover, we produced both the heavy and light chain molecules of a protective IgG mAb 6D8 against Ebola virus GP1 (Wilson et al. 2000) at ~0.5 mg of mAb per gram leaf fresh weight (LFW) within 4 to 8 days post infiltration (dpi) of Nicotiana benthamiana leaves. We further demonstrated that full-size IgG complex containing two heavy chains and two light chains was efficiently assembled, readily purified, and functioned properly to bind inactivated Ebola virus. Thus, our single-vector replicon system provides high-yield production capacity, eliminates the difficult task of identifying non-competing virus, and obviates the need for co-infection of multiple expression modules. The multi-replicon vector represents a significant advance in transient expression technology for antibody production in plants.
For the construction of pBYDsRed (Fig. 1), the DsRed gene was amplified from pDsRed1-1 (Clontech cat# 6922-1) with primers 5’-ATCGTCTAGAACCATGGTGCGCTCCTCCAAG and 5’-ATTAGAGCTCCTACAGGAACAGGTGGTG, digested with XbaI and SacI, and ligated into pIBT210 to make pIBT-DsRed, from which the XhoI-SacI fragment was substituted into pBYGFP to make pBYDsRed (Fig. 1). Tandem dual replicon constructs (e.g. pBYGFPDsRed.R, Fig. 1) were designed to contain two replicons in tandem, linked by a LIR element. The first replicon contains LIR-35S cassette-SIR-LIR, and the second contains LIR-35S cassette-SIR-C2/C1-LIR, with the bold/underlined LIR element of each cassette overlapping to generate LIR-35S cassette-SIR-LIR-35S cassette-SIR-C2/C1-LIR. Each replicon is thus defined by the LIR borders. We used CaMV 35S promoters with a single enhancer element, obtained by amplification of the expression cassettes in pBYDsRed and pIBT210.3 (Judge et al. 2004) with primers 35S-Sda (5’-TGACCTGCAGgCATGGTGGAGCACGACA) and VSPHT (5’-TGAATAGTGCATATCAGCATACCTTA). The 35S promoter and 5’-UTR of TMV in the fragment amplified from pIBT210.3 (SdaI-NcoI), the GFP gene from pBYGFP (NcoI-KpnI), and the pea ribulose1,5-bisphosphate carboxylase small subunit (rbcS) terminator (Friedrichsen et al. 2000) on a KpnI-EcoRI fragment were ligated together into the PstI and EcoRI sites of pBY024 (Mor et al. 2003) to make pBYGFP210.3. The fragment containing the GFP partial replicon (AscI)-LIR-35S/TMV/GFP-SIR-(blunt), obtained by digestion of pBYGFP210.3 with BamHI and filling with Klenow enzyme, then digestion with AscI, was ligated with the 35S-Sda primer-amplified pBYDsRed partial replicon (blunt)-LIR-35S/TEV/DsRed-SIR-(SacI) digested with AscI and filled with Klenow enzyme, and then digested with SacI, into the vector pBYHBc.R (Huang et al. 2009) that had been digested with AscI and SacI, to make pBY-GFPDsRed.R (Fig. 1).
The gene sequences for heavy (H2) and light (K3) chains of mouse monoclonal antibody 6D8 (Wilson et al. 2000) were de-immunized for humans by substitution of human constant region sequences for gamma type 1 and kappa chains (Biovation, Edinburgh, Scotland). The resulting sequences were used to design plant codon-optimized genes, and synthesized commercially (Retrogen, San Diego, CA). The H2 gene in pCHF4-6D8-H2 (Mapp Biopharmaceutical, Inc.) was end-tailored to add a C-terminal ‘SEKDEL’ hexapeptide by PCR with the primer H2-SEKDEL-Kpn (5’-GCGGTACCTTAAAGCTCATCCTTCTCTGATTTACCCGGAGACAAGGAGAG), digested with NcoI and KpnI, and inserted into pPS1 (Huang and Mason 2004) to make p6D8-H2, from which the XhoI-EcoRI fragment containing TEV5’-UTR-H2-VSP3’ was substituted into pBYGFP to make pBYH(6D8) (Fig. 1). The K3 gene in pCHF4-6D8-K3 (Mapp Biopharmaceutical, Inc.) was obtained as an NcoI-KpnI fragment, ligated with the 35S promoter-TMV5’ and the rbcS3’ elements, and substituted into pBYGFP vector to make pBY-L(6D8) (Fig. 1). The replicon (LIR to SIR) from pBY-L(6D8) was substituted into pBYGFP.R to make pBYK3R. The H2-SEKDEL fragment was amplified from pBY-H(6D8) with primers H2-Xba (5’- ACGATCTAGAACAATGGGATGGTCTTGCATC) and VSPHT, digested with XbaI and KpnI, and substituted into the vector pBY027 (Mor et al. 2003) to make pBY-H2K210. The replicon from pBY-H2K210 was then inserted into pBYK3R to make the tandem dual replicon construct pBY-HL(6D8).R (Fig. 1).
Binary vectors were separately introduced into Agrobacterium tumefaciens LBA4404 by electroporation. The agroinfiltration procedure was performed as previously described (Huang et al. 2009).
Total DNA extraction and Southern blotting was performed as previously described (Huang et al. 2009). Digoxygenin (DIG)-labeled gene-specific probes were synthesized by PCR with primers (5’-GTCACCATGGTGAGCAAGGGCGAG -) and (5’-CTCAGGATCCTTACTTGTACAGCTCGTC -) for the GFP gene, and with primers (5’-ATCGTCTAGAACCATGGTGCGCTCCTCCAAG-) and (5’-ATTAGAGCTCCTACAGGAACAGGTGGTG-) for the DsRed gene, respectively, according to the manufacturer’s instructions (Roche Applied Science, Indianapolis, IN).
For expression time-course experiments, Agroinfiltrated Nicotiana benthamiana leaves were harvest on days 2, 4, 6, and 8 dpi. For other protein analysis, plant leaves were harvest 4 dpi.
Total protein extract was obtained by homogenizing leaves with extraction buffer (25mM sodium phosphate, pH 6.6, 100mM NaCl, 1mM EDTA, 0.1% TritonX-100, 10mg/ml sodium ascorbate, 10µg/ml leupeptin, 0.3mg/ml PMSF) using a FastPrep machine (Bio101, Vista, CA) following the manufacture’s instruction. The crude plant extract was clarified by centrifugation at 14,000g for 10 min at 4°C.
The clarified total protein extract was analyzed by ELISA designed to detect the assembled form of mAb (with both light and heavy chains) as described previously (Giritch et al. 2006). Briefly, plates were coated with a goat anti-human IgG specific to gamma heavy chain (Southern Biotech, AL). After incubation with plant protein extract, a HRP-conjugated anti-human-kappa chain antibody (Southern Biotech, AL) was used as the detection antibody, using human IgG reference standard (Southern Biotech, AL).
SDS-PAGE and western blotting were performed as previously described (Santi et al. 2008). Briefly, plant protein example or human IgG standard was mixed with sample buffer either under reducing or nonreducing condition and then subjected to SDS-PAGE. PAGE gels were either stained with Coomassie blue or used to transfer proteins onto PVDF membranes. The anti-human gamma and kappa specific antibodies described above (ELISA) were used for western blot analysis.
Polyvinyl chloride 96-well ELISA plates were coated with 50µl of 1:1000-diluted irradiated Ebola virus (a gift from Dr. William Pratt, USAMRIID) and incubated at 4°C for 12 hr. Plates were blocked with 5% skim milk in PBST (10.6mM Na2HPO4, 2.9mM KH2PO4, 104.7mM NaCl, 0.05% Tween-20) at room temperature for 2 hrs. Subsequently, the plates were incubated with different concentrations of the plant-derived 6D8 mAb or human IgG (Southern Biotech, AL) diluted in 1% skim milk in PBST for 1 hr at 37°C. HRP-conjugated goat anti-human IgG (Southern Biotech) was then added and incubated for 1 hr at 37°C. The plates were developed with TMB substrate (KPL inc., MD) and the OD was read at 450nm.
N. benthamiana leaves infiltrated with 6D8 mAb constructs were harvest on 4 dpi and were homogenized with the extraction buffer (25mM sodium phosphate, pH 6.6, 100mM NaCl). Crude extract was filtered through Miracloth and centrifuged at 17,700g for 15 min at 4°C to remove cell debris. Ammonium sulfate was slowly added to the clarified plant extracts to 35% saturation with thorough mixing at 4°C. The sample was centrifuged at 17,700g for 15 min at 4°C and the pellet was saved for analysis. The 35% ammonium sulfate supernatant was further processed by adding ammonium sulfate to 60% saturation. The sample was again centrifuged at 17,700g for 15 min and the supernatant was discarded. The 60% ammonium sulfate pellet was resuspended in the extraction buffer and was then applied to a Protein G column (Pierce, IL). After sample loading, the resin was washed with the extraction buffer and the column was eluted with 50mM sodium citrate, pH2.5. The elution fraction was neutralized immediately with 1M Tris-base to a final pH of 7.5. The purity of 6D8 mAb was determined by quantitating Coomassie blue–stained protein bands on SDS-PAGE using a densitometer.
Protoplasts were released from infiltrated leaf tissue by incubation for 10–16 hours at room temperature in a solution containing 0.4M mannitol, 8mM CaCl2, 0.4% cellulase and 0.4% Macerozyme R-10. Released protoplasts were filtrated through a nylon mesh with a 62-µm pore diameter.
Leaves co-expressing GFP and DsRed were viewed under a UVGL-25 lamp (UVP, Upland, CA). Protoplasts were viewed with a Nikon inverted microscope with GFP filter sets (Chroma #41028; excitation filter, 500/20 nm; emission filter, 535/30 nm) and DsRed filter sets (Chroma #42005; excitation filter, 540/40nm; emission filter, 600/50 nm).
We previously showed (Huang et al. 2009) that a BeYDV-derived three-vector (e.g., pBYGFP, pREP110 and pP19 in Fig. 1) transient expression system provides extremely efficient replication of a single DNA replicon and high-level expression in plant cells of protein encoded by the single replicon vector (e.g., pBYGFP). In the present study, we first tested whether two replicon vectors encoding different proteins can be co-expressed. We infiltrated N. benthamiana leaves with BYGFP and/or BYDsRed in combination with REP110 and P19 vectors (Fig. 1). BYGFP/REP110/P19 and BYDsRed/REP110/P19 infiltrated areas displayed green and red fluorescence, respectively, while the fluorescence of BYGFP/BYDsRed co-infiltrated area appeared yellow (Fig. 2A), probably due to overlapping of both green and red fluorescence. Mesophyll protoplasts were subsequently isolated from infiltrated areas and observed under fluorescence microscope. When two populations of protoplasts that individually expressed BYGFP or BYDsRed were mixed, no overlapping fluorescent protoplast was observed (Fig. 2B and 2C). In contrast, the majority (>80%) of the fluorescent protoplasts from co-infiltration of BYGFP/BYDsRed exhibited both green and red fluorescence (Fig. 2D and 2E), indicating the efficient co-expression of GFP and DsRed in the same cells.
To determine whether amplification of both replicons is efficient, we performed Southern blot of DNA extracted from co-infiltrated leaves. As shown in Fig. 3A, the ethidium bromide-stained gel showed the expected ~3kbp band in the BYGFP/REP110/P19, the BYDsRed/REP110/P19, and the BYGFP/BYDsRed/REP110/P19 samples, indicating efficient formation of DNA replicons. No discrete bands were observed in DNA samples from uninfiltrated, BYGFP alone, or BYDsRed alone leaves. Southern blotting with both GFP and DsRed probes showed that the BYGFP/REP110/P19 replicon reacted with only the GFP probe, while the BYDsRed/REP110/P19 bound only the DsRed probe (Fig. 3B and 3C). However, the BYGFP/BYDsRed/REP110/P19 sample produced positive signals for both GFP and DsRed probes, with the signal intensity similar to those representing the corresponding single replicons, suggesting that the BYGFP and BYDsRed replicons are near-equally co-expressed. Together, these results demonstrate that the BeYDV replicon system is non-competing, which will enable its application in producing multiple-subunit proteins of pharmaceutical importance, such as mAbs.
To demonstrate the utility of our non-competing expression system, we expressed a protective mAb against Ebola virus glycoprotein GP1 (6D8) by co-infiltration of N. benthamiana leaves with two replicon vectors, pBY-H(6D8) and pBY-L(6D8) (Fig. 1), encoding the heavy chain and light chain subunits of 6D8 mAb, respectively, along with vectors for REP110 and P19. ELISA showed that 6D8 mAb accumulated within 2 dpi and rapidly reached a peak in 4 dpi at 0.4–.5 mg/g LFW (Fig. 4). Western blot with SDS-PAGE under reducing condition confirmed the presence of both heavy and light chains with the expected molecular weights of 50 kDa and 25 kDa, respectively (Fig. 5A and 5B, lane 3). When the same samples were analyzed under non-reducing condition, a ~170 kDa band was observed, which represents the fully assembled hetero-tetrameric form of 6D8 mAb (Fig. 5C, lane 3). The slightly higher molecular weight (~170 kDa vs. the calculated 150 kDa) suggests that 6D8 is glycosylated as predicted for proteins with potential N-link glycosylation sites targeted to the endomembrane system of plant cells. The western blots also indicated that no proteolytic clipping of the light or heavy chains occurred, since only the expected size bands were observed (Fig. 5A–5C, lane 3).
We previously showed that the three-vector replicon system can be simplified without compromising yields (Huang et al. 2009). A single vector replicon containing a native REP expression cassette (C1/C2 coding region under the control of the viral LIR promoter, e.g. pBYGFP.R in Fig. 1) works as well as the three-vector system for the expression of GFP (Huang et al. 2009). To test if the single-vector replicon system can be used to simultaneously express multiple proteins, we constructed pBYGFPDsRed.R for co-expression of GFP and DsRed (Fig. 1). Examination of protoplasts prepared from pBYGFPDsRed.R infiltrated leaves showed that both GFP and DsRed fluorescence were simultaneously detected in ~95% of the fluorescent protoplasts (Fig. 2F and 2G), indicating high-efficiency co-expression of both fluorescent proteins within same cell. Analyses of DNA from infiltrated leaves reveals the BYGFPDsRed.R sample produced replicons of different sizes which also reacted differently with GFP or DsRed probes (Fig. 6), indicating the simultaneous presence of GFP and DsRed replicons. For the BYGFPDsRed.R sample, the strongest DsRed-probe-reacting band is ~3.5kb (Fig. 6), which is similar to the size of the BYGFP.R replicon, suggesting this band represents the DsRed replicon comprising the 35S/TEV 5’-DsRed-VSP 3’-SIR-C2/C1 sequence within the two LIRs (see Fig. 1); while the top GFP-probe-reacting band is ~2.6kb (Fig. 6), which is expected for a smaller replicon consisting of the 35S/TMV 5’-GFP-rbcS 3’-SIR region between two adjacent LIR elements (see Fig. 1). These results demonstrate that the two tandemly linked replicons can be released and amplified independently, leading to high-level proliferation of two distinct double-stranded DNA replicons and efficient expression of both proteins. We also observed that in the BYGFPDsRed.R lanes a ~1.6kb band reacted with the GFP probe and a slightly higher ~1.7kb band less strongly with the DsRed probe (Fig. 6), probably representing the single-strand DNA forms of the corresponding replicons. Overall, the data (Fig. 2F and 2G and Fig.6) indicate that the multi-replicon single vector system produces high-level co-expression of two target proteins.
To further demonstrate the effectiveness of the multi-replicon single-vector system, we used pBY-HL(6D8).R (Fig. 1), which harbors expression cassettes for the light and heavy chains of 6D8 mAb and REP (C1/C2). ELISA of pBY-HL(6D8).R leaf extracts showed that 6D8 mAb accumulated to a level comparable to that produced by the co-infiltration of four separate vectors (Fig. 4). The time course of the expression of 6D8 mAb with this single vector was also similar to that obtained with separate vectors (Fig 4). Furthermore, SDS-PAGE and western blot analyses demonstrate that the mAb produced by pBY-HL(6D8).R has the correct light and heavy chain components (Fig. 5A and 5B, Lane 4) and is fully assembled (Fig. 5C, Lane 4 ).
To further validate the multi-replicon single-vector expression system and to examine the structural and functional properties of the plant-made IgG, we used ammonium sulfate precipitation followed by protein G affinity chromatography to purify the single-vector-derived 6D8 mAb. SDS-PAGE Coomassie blue staining analysis of purification fractions showed that the combination of ammonium sulfate precipitation and Protein G affinity chromatography effectively removed most plant host proteins including the most abundant plant protein RuBisCo (Fig. 7, Lanes 2 and 4) and 6D8 mAb was purified to >90% purity with intact light and heavy chains (Fig. 7, Lane 5). A similar analysis under non-reducing condition indicated that the purified 6D8 antibody is in its fully assembled tetrameric form (data not shown). The purified mAb was further examined in a binding assay in which different concentrations of plant-derived 6D8 were incubated with irradiated Ebola virus coated on an ELISA plate. A generic human IgG was used as a negative control for the assay. Figure 8 shows the OD450 increased as more 6D8 mAb was applied in the reaction. In contrast, the OD450 for a generic human IgG remained at a basal level regardless the amount of this IgG used for the reaction (Fig. 8). This result indicates that plant produced 6D8 retains its specific affinity for Ebola virus GP1 protein. Analyses for 6D8 produced by co-infiltration of plant leaves with separate light chain and heavy chain replicons showed similar results in antibody purity, SDS-PAGE pattern, and its ability to specifically bind to Ebola virus (data not shown).
Plant production of mAbs employs two main approaches: plant virus-based transient expression and stable transgenic plants. For the latter strategy, the long time frame (months to years) required to create stable transgenic plants is impractical for rapid production of protein samples for pre-clinical studies. In addition, the lack of strong regulatory elements to drive high-level protein accumulation and the position effect associated with the randomness of transgene integration in the plant genome still presents challenges for stable transgenic technology (Chen 2008). In contrast, transient systems that are focused on production speed have greatly reduced the time needed to obtain material for testing the function of recombinant antibodies in preclinical trials (Chen et al. 2009; Lico et al. 2008). In addition, highly efficient viral replicon vectors have enabled high-level antibody expression in plants. For example, it has been recently reported that full-size mAbs can be rapidly produced at levels as high as 0.5 mg of mAb per gram leaf fresh weight (Giritch et al. 2006). However, the process requires two sets of noncompeting viral vectors derived from TMV and PVX (Giritch et al. 2006). For the production and assembly of more complex mAbs such as secretory IgA and IgM, it remains a difficult task to find additional viruses compatible with the existing TMV/PVX expression system to allow efficient co-expression of three or more distinct subunit proteins in same cells.
Unlike traditional or “deconstructed” viral expression systems (Giritch et al. 2006) which rely on in planta assembly and replication of near-full-length viral genome, the BeYDV-derived DNA replicon system requires only two viral elements, the LIR region and the Rep/RepA, for the initiation and amplification of episomal replicons (Mor et al. 2003). We hypothesized that this system could be used for non-competing co-expression of multiple protein subunits. We found that co-infiltration of replicons encoding two fluorescence proteins (GFP and DsRed) resulted in the exhibition of both fluorescences, with >80% of cells expressing both proteins in the same cell (Fig. 2), indicating our BeYDV replicon system is indeed noncompeting. Our subsequent effort focused on simplifying the system so that it can be feasible for practical production of multiple-subunit protein pharmaceutical candidates. Our group has previously showed that single replicon vectors containing a gene of interest and the Rep/RepA expression cassettes (such as pBYGFP.R in Fig. 1) were as efficient as their three-component counterparts in directing the expression of target proteins (Huang et al. 2009). In this study, we created single vectors containing two tandemly linked replicons expressing distinct genes of interest, with the hypothesis that the two replicons can efficiently replicate without interference. Transient leaf infiltration with one such vector, pBYGFPDsRed.R (Fig. 1), resulted in efficient formation of both GFP and DsRed replicons as evidenced by Southern blotting (Fig. 6) as well as co-expression of GFP and DsRed in ~95% cells (Fig. 2F and 2G), thus proving that our hypothesis is correct. Collectively, our results with model proteins GFP and DsRed indicate the feasibility of using our replicon system for high-level co-expression of two subunit proteins.
To evaluate the utility and capacity of our replicon systems for antibody production, we co-expressed both the heavy chain and light chain molecules of a mAb protective against Ebola virus GP1 using either the three-component noncompeting vectors or the single vector pBY-HL (6D8).R. Our results showed that both systems could rapidly (4 dpi) produce 6D8 mAb at ~0.5 mg of mAb per gram leaf fresh weight (Fig. 4), which is comparable to the highest level ever achieved with a plant-based expression system (Giritch et al. 2006). Furthermore, the mAb can correctly assemble into its tetrameric structure (Fig. 5C). We did not observe any non-assembled or partial assembled 6D8 mAb, suggesting that the assembly of the mAb using the DNA replicon system is highly efficient. We further demonstrated that 6D8 mAb produced with this single vector system not only has the correct assembled tetrameric structure, but also can be readily purified, and is functionally active in specific binding to inactivated Ebola virus. In addition to 6D8, we have also successfully expressed several other mAbs and mAb-fusion proteins twice as large as 6D8 with full assembly and functionality (Chen, unpublished data). Our success in producing the fully-assembled tetrameric functional IgG (two heterooligomeric subunits) with a two-replicon single vector strongly suggest that simultaneous expression of as many as four hetero-subunits can be easily achieved using two of such vectors, or by creating single vectors with three or more tandem linked replicons. To our best knowledge, this technology is the first and the only one that has such a potential, among all plant transient replicon expression systems.
While the BeYDV multi-replicon vector provided high expression levels, there may be additional ways to improve its performance. As suggested by others, molecular chaperones could be included in the expression cassette to accommodate the increasing protein folding and assembly demand by the high-level accumulation of oligomeric mAbs (Nuttall et al. 2002). Furthermore, the anti-silencing element P19 could be incorporated into the single vector system. We have not fully explored the effect of P19 with the multi-replicon single-vector. Finally, incorporation of translation enhancer elements from cowpea mosaic virus RNA-2 (Sainsbury and Lomonossoff 2008) have the potential to substantially increase expression of mRNA produced by the DNA replicon system.
In summary, the results presented in this paper collectively demonstrate that the BeYDV-derived DNA replicon expression system is capable of rapidly producing high levels of multi-subunit antibodies in plants. The ability of a simplified multi-replicon single vector to deliver multiple expression cassettes/replicons for production of mAbs and other hetero-oligomeric pharmaceuticals provides some key advantages. Once optimized, this expression system has the potential to provide a technology platform to produce a wide spectrum of pharmaceutical proteins with speed, efficiency, cost-effectiveness and safety.
This work was supported by NIH Grants # 5U01AI061253, 1U01AI075549, U01AI061270 and 5U19AI066332.