• Open Access

pEAQ: versatile expression vectors for easy and quick transient expression of heterologous proteins in plants

Authors


  • Sequences submitted to Genbank: pM81-FSC2-POW, pEAQ, pEAQexpress, pEAQselectK, pEAQspecialK, pEAQ-HT, pEAQ-HT-DEST1, pEAQ-HT-DEST2, pEAQ-HT-DEST3.

*(fax +44 1603 450018; e-mail: george.lomonossoff@bbsrc.ac.uk)

Summary

Agro-infiltration of leaf tissue with binary vectors harbouring a sequence of interest is a rapid method of expressing proteins in plants. It has recently been shown that flanking the sequence to be expressed with a modified 5′-untranslated region (UTR) and the 3′-UTR from Cowpea mosaic virus (CPMV) RNA-2 (CPMV-HT) within the binary vector pBINPLUS greatly enhances the level of expression that can be achieved [Sainsbury, F. and Lomonossoff, G.P. (2008)Plant Physiol. 148, 1212–1218]. To exploit this finding, a series of small binary vectors tailored for transient expression (termed the pEAQ vectors) has been created. In these, more than 7 kb of non-essential sequence was removed from the pBINPLUS backbone and T-DNA region, and unique restriction sites were introduced to allow for accommodation of multiple expression cassettes, including that for a suppressor of silencing, on the same plasmid. These vectors allow the high-level simultaneous expression of multiple polypeptides from a single plasmid within a few days. Furthermore, vectors have been developed which allow the direct cloning of genes into the binary plasmid by both restriction enzyme-based cloning and GATEWAY recombination. In both cases, N- or C-terminal histidine tags may be fused to the target sequence as required. These vectors provide an easy and quick tool for the production of milligram quantities of recombinant proteins from plants with standard plant research techniques at a bench-top scale.

Introduction

Transient expression is a common method for assessing the capacity of plant cells to produce a particular protein and/or to test various expression cassette arrangements. This is often performed as a pilot study preceding the stable integration of foreign genes into the plant genome. More recently, transient expression by agro-infiltration (Kapila et al., 1997; Marillonnet et al., 2005; D’Aoust et al., 2009) has emerged as a bona fide production platform. As with stable transgenesis, this technique takes advantage of the ability of Agrobacterium tumefaciens to transfer a portion of its genome, the T-DNA, into the plant host cell (Gelvin, 2003; Tzfira and Citovsky, 2006). In general, transient expression is limited to the tissue which has been infiltrated with the appropriate Agrobacterium suspension(s). This inevitably limits the amount of tissue available and thus puts a premium on maximizing protein expression levels. One way of achieving this involves the use of replicating virus-based vectors (Cañizares et al., 2005; Giritch et al., 2006; Lindbo, 2007). However, these have disadvantages in terms of the size and complexity of the proteins that can be expressed, the genetic stability of constructs during replication of the viral genome, and concerns regarding biocontainment.

As an alternative, we have recently developed a transient expression system based on a deleted version of Cowpea mosaic virus (CPMV) RNA-2, CPMV-HT, which permits the extremely high-level and rapid production of proteins without viral replication (Sainsbury and Lomonossoff, 2008). This system involves inserting the gene to be expressed between a modified 5′-untranslated region (UTR) and the 3′-UTR from CPMV RNA-2. The elimination of the AUG codon at position 161 (by a U to C transition at position 162), upstream of the main translation initiation site at position 512, was found to greatly enhance the level of translation of the inserted gene compared with that obtained using the unmodified 5′-UTR (Sainsbury and Lomonossoff, 2008). The CPMV-HT system has several advantages over other transient expression systems: it is possible to express multiple foreign genes in the same tissue, there is no issue regarding genetic drift during replication of the sequence and there are few, if any, issues of biocontainment. However, cloning into the CPMV-HT system as originally described by Sainsbury and Lomonossoff (2008) required a two-step procedure: insertion of the gene of interest between the modified 5′- and 3′-UTR of CPMV RNA-2 in a plant expression cassette followed by transfer of the expression cassette, including the Cauliflower mosaic virus (CaMV) 35S promoter and nopaline synthase (nos) terminator, to the binary vector pBINPLUS (van Engelen et al., 1995). This was a rather time-consuming process which limited the number of constructs which can be produced. Furthermore, high-level expression required the co-infiltration of a separate construct expressing the suppressor of silencing P19 and co-expression of multiple proteins required the co-infiltration of separate Agrobacterium cultures containing each gene. In addition, to ensure that each cell received all the necessary T-DNA, the infiltration solutions had to have a relatively high cell density. Taken together, these factors constitute a substantial bottleneck for the use of CPMV-HT for medium- or high-throughput protein expression.

To overcome the above limitations, we have created a series of binary vectors that make cloning and expression with the CPMV-HT system much easier and quicker. The plasmid backbone has been reduced to less than half its original size to facilitate plasmid propagation and to eliminate unwanted restriction sites with no compromise in its performance in transient expression assays. CPMV-HT expression cassettes and the sequence encoding P19 have been incorporated into the T-DNA region of the resulting vector allowing the high-level expression of multiple polypeptides to be obtained by infiltration of a single construct. Finally, we have engineered an extremely high-level expression vector that allows efficient single-step insertion of foreign genes into the CPMV-HT cassette within the binary vector. This may be achieved either by restriction enzyme-based cloning or by use of the GATEWAY recombination system, which has the potential for high-throughput applications.

Results

pEAQ series construction

To monitor any effects on expression driven by the CPMV-HT cassette which might result from modifying the pBINPLUS vector backbone, we chose to start with the Green Fluorescent Protein (GFP)-containing plasmid pBD-FSC2-GFP-HT (Figure 1a) so that expression in plants after infiltration could be readily monitored. This plasmid contains an expression cassette consisting of a CaMV 35S promoter, the CPMV RNA-2 5′-UTR carrying the U162C (HT) mutation which was previously shown to enhance expression levels (Sainsbury and Lomonossoff, 2008), the sequence of GFP, the CPMV RNA-2 3′-UTR, and the nos terminator, inserted between the PacI and AscI sites of pBINPLUS and has a total size of 14 416 bp. The pBINPLUS backbone was subjected to computational analysis and three regions were determined to be essential for the plasmid to function as a binary vector: (i) the T-DNA region including right border (RB) and left border (LB); (ii) the RK2 (OriV) replication origin; and (iii) a segment containing the ColEI replication origin for higher copy number in Escherichia coli, the neomycin phosphotransferase (NPT) III gene conferring resistance to kanamycin, and TrfA from RK2, which promotes replication (Frisch et al., 1995). It was reasoned that much of the remaining backbone sequence was non-essential for bacterial growth and plant transformation and, therefore, could be removed as previously shown in other cases where binary vectors of reduced size have been constructed (Hajdukiewicz et al., 1994Xiang et al., 1999; Hellens et al., 2000). The three essential regions were amplified by the polymerase chain reaction (PCR) from pBD-FSC2-GFP-HT, and ligated to give plasmid pEAQbeta (Figure S1), which is 4584 bp smaller than its parent plasmid. A further round of PCR amplification of pEAQbeta removed an additional 2639 bp of non-essential sequence from the T-DNA region (including the NPTII gene) and inserted three unique restriction sites, AsiSI, MluI and FseI, adjacent to the original AscI site of pBINPLUS. The incorporation of these additional unique sites facilitates the insertion of additional expression cassettes into the T-DNA of the vector. The resulting vector, pEAQ-GFP-HT was fully sequenced and, at 4616 bp, the backbone of this plasmid, including the T-DNA borders, makes it one of smallest available binary vectors (Figure 1b).

Figure 1.

 Diagrammatic representation of pEAQ and its parent pBINPLUS including the GFP-HT expression cassette. (a) pBD-FSC2-GFP-HT and (b) pEAQ-GFP-HT with major features labelled and total sizes indicated. Light grey arrows, regions non-essential for plasmid growth or transient transformation; dark grey arrows, regions essential for plasmid growth and replication; black boxes, T-DNA left (LB) and right (RB) borders; white arrows, promoter sequences (CaMV, Cauliflower mosaic virus); solid black lines, CPMV RNA-2 UTRs; bordered grey arrows, GFP coding sequences; and black arrows, terminator sequences (nos, nopaline synthase). OriV, pRK2 replication origin; ColEI, pBR322 replication origin; NPT, neomycin phosphotransferase; TrfA, replication-essential locus.

Plasmid pEAQ-GFP-HT was used as the starting point for the development of a series of vectors by insertion of various features into the T-DNA (Figure 2). Plasmids pEAQexpress-GFP-HT and pEAQselectK-GFP-HT were created by the insertion of a 35S-P19 or an NPTII cassette, respectively, into the FseI site. pEAQspecialK-GFP-HT includes both the NPTII cassette (inserted in the FseI site) and the 35S-P19 cassette (inserted between the AsiSI and MluI sites) on the same plasmid. The presence of the P19 cassette was intended to obviate the need for co-infiltration with a separate P19-expressing construct, while the NTPII cassette was incorporated to potentially allow pEAQ-based vectors to be used for stable transformation. Unexpectedly, the presence of the NPTII gene in pEAQselectK-GFP-HT and pEAQspecialK-GFP-HT increased the yield of these plasmids in E. coli by approximately 50%–100% when compared with the pEAQ vectors not containing the cassette.

Figure 2.

 Schematic representation of the T-DNA regions of major derivatives of pEAQ. pEAQexpress-GFP-HT, pEAQselectK-GFP-HT, and pEAQspecialK-GFP-HT T-DNAs are shown with unique restriction sites indicated. Black boxes, T-DNA borders; white arrows, promoter sequences; solid black lines, CPMV RNA-2 UTRs; bordered grey arrows, coding sequences; and black arrows, terminator sequences.

Reduction in size does not compromise transient expression from pEAQ

To examine the performance of the pEAQ-based vectors for transient expression, each GFP-containing construct was transformed in A. tumefaciens and bacterial suspensions were used to agro-infiltrate Nicotiana benthamiana leaves. The original pBINPLUS-based construct, pBD-FSC2-GFP-HT, was infiltrated as a control. The constructs which did not contain the P19 cassette (pBD-FSC2-GFP-HT, pEAQ-GFP-HT and pEAQselectK-GFP-HT) were co-infiltrated with an Agrobacterium suspension containing pBIN61-P19 (Voinnet et al., 2003). The levels of GFP expression were determined by inspection of the infiltrated leaves under UV illumination, analysis of soluble protein extracts by sodium dodecyl–sulfate polyacrylamide gel electrophoresis (SDS-PAGE), and measurements of GFP fluorescence in leaf extracts by spectrofluorometry (Figure 3a–c respectively). In all cases, the pEAQ series of vectors gave levels of GFP expression at least as high as that obtained with pBD-FSC2-GFP-HT, showing that the large reduction in size of the vector backbone does not compromise expression levels. The results represent six replicates from two individual experiments and show that expression from pBD-FSC2-GFP-HT was more variable than from pEAQ vectors.

Figure 3.

 Expression levels from pEAQ and its derivatives compared with the parent plasmid pBINPLUS. 1 : 2 = constructs infiltrated at half the standard OD600; and * = constructs co-infiltrated with P19. (a) Leaves visualized under UV light, (b) Coomassie-stained 12% SDS-PAGE with molecular weight size markers indicated, and (c) spectrofluorometric analysis of GFP expression where values represent 6 samples from 2 separate experiments ± SE.

The GFP expression levels obtained by co-infiltration with pEAQ-GFP-HT + pBIN61-P19 or infiltration with pEAQexpress-GFP-HT (which contains the sequence of P19) alone, were very similar. This shows that expression of P19 from the same vector as the target gene is at least as effective as expressing it from a separate plasmid. The fact that only a single Agrobacterium suspension is used when both the GFP-HT and P19 expression cassettes are present on a single plasmid means that it should be possible to halve the total number of Agrobacteria used for infiltration. This was verified by diluting the Agrobacterium culture harbouring pEAQexpress-GFP-HT twofold, such that the final optical density (OD600) was that of each individual culture used for the co-infiltration of pEAQ-GFP-HT + pBIN61-P19, and demonstrating that the expression levels were similar (Figure 3, compare levels obtained with pEAQ and pEAQexpress 1 : 2).

As found with pEAQ-GFP-HT, co-infiltration of pEAQselectK-GFP-HT with pBIN61-P19 resulted in expression levels higher than those obtained by co-infiltration of pBD-FSC2-GFP-HT and pBIN61-P19. In addition, as observed for pEAQexpress-GFP-HT, inclusion of P19 in the T-DNA region of pEAQspecialK-GFP-HT gave high levels of GFP expression using Agrobacterium suspensions at half the total OD600 used for co-infiltrations (Figure 3). The inclusion of the NPTII resistance marker resulted in consistently higher levels of GFP expression in transient assays than was found with the corresponding plasmids which did not contain the marker (Figure 3). The reason for this is unclear but may be as a result of an increase in plasmid copy number in Agrobacteria harbouring the additional kanamycin-resistance cassette. Alternatively, it may be as a result of extensive degradation of the T-DNA from the LB during transfer, which would have less impact on the HT and P19 cassettes, both of which are required for high level expression, in cases where the NPTII cassette resides between it and the T-DNA border.

Full-size antibody expression from a single plasmid

To take advantage of the modular nature of the pEAQ series, CPMV-HT expression cassettes containing the endoplasmic reticulum-retained heavy chain (HE) and light chain (L) of the human monoclonal anti-human immunodeficiency virus antibody 2G12 (Buchacher et al., 1994), were inserted, in both orders, into the PacI/AscI and AsiSI/MluI sites of pEAQexpress-GFP-HT. This was performed to demonstrate accumulation of a multi-polypeptide chain protein and to test the effects of gene order in the plasmid. The resulting constructs provided the CPMV RNA-2 UTRs for the genes of both chains and were designated pEAQex-2G12HEL and pEAQex-2G12LHE depending on the order in which the sequences encoding the antibody chains were inserted (Figure 4a). Infiltration of N. benthamiana leaves with single Agrobacterium cultures containing the above plasmids resulted in the formation of fully assembled 2G12 antibodies identical in size to 2G12 produced by mixing three Agrobacterium cultures, each of which expressed the individual components, L, HE and P19 (Figure 4b; Sainsbury and Lomonossoff, 2008). The identity of this protein was confirmed by Western blot analysis using anti-Fc and anti-Fab antibodies (data not shown). Furthermore, analysis of the protein purified from leaves co-infiltrated with three constructs (Figure 4b), showed it had the properties of authentic 2G12; the results of this analysis will be presented elsewhere. No protein of the size expected of assembled IgG was seen in the sample from leaves infiltrated with the empty pBINPLUS vector with P19. The protein loaded in each lane represents the amount extracted from 3 mg of infiltrated tissue or 1/333 of the protein potentially obtainable from 1 g of tissue. The maximum amount of assembled IgG produced from the 3-culture mixture corresponds to 1 μg of CHO-produced 2G12 on the Coomassie-stained non-reducing SDS-PAGE gel. This suggests an expression level of 2G12 in excess of 325 mg/kg of fresh weight tissue, which is in agreement with the antibody concentrations measured previously using surface plasmon resonance (Sainsbury and Lomonossoff, 2008). The use of pEAQex-2G12HEL appears to surpass this already high-level of antibody accumulation.

Figure 4.

 Expression of the full size IgG, 2G12, with a single pEAQ plasmid compared with the infiltration of 3 separate cultures. (a) Schematic representation of the two pEAQexpress-derived plasmids constructed to express 2G12. Black boxes, T-DNA borders; white arrows, promoter sequences; solid black lines, CPMV RNA-2 UTRs; bordered grey arrows, coding sequences; and black arrows, terminator sequences. (b) Coomassie blue-stained 4-12% polyacrylamide-SDS gel with protein from the equivalent of 3 mg of infiltrated tissue: M, marker with sizes indicated; C, control extract from leaves infiltrated with the empty pBINPLUS vector; Std, 1 μg of CHO-produced 2G12. (c) Infiltration scheme indicating dilutions and their respective ODs for each plasmid combination, and the concentration of protein extracts made after infiltrations at each OD ± SD.

An advantage of pEAQ-derived vectors is that each component of a multi-chain protein such as an IgG will be present in each infected cell. Therefore, high expression levels should be maintained at higher dilutions of Agrobacteria suspensions than if multiple cultures have to be used. To test if this is the case in practice, cultures that were initially resuspended to OD600 1.2, and mixed where necessary, were subjected to two serial threefold dilutions (Figure 4b–c). This resulted in final OD600 of each individual culture in the three-culture mix being 0.4, 0.13 and 0.04. Single cultures harbouring the pEAQexpress constructs were infiltrated at OD600 of 1.2, 0.4 and 0.13. When three separate cultures were used, the level of assembled 2G12 decreased markedly upon serial dilution. In contrast, 2G12 expression from pEAQex-2G12HEL and pEAQex-2G12LHE, was maintained at a consistently higher level, with the reduction on dilution being relatively modest (Figure 4b). Interestingly, the amount of total protein extracted from the infiltrated tissue was almost halved when the OD600 of the infiltrate was reduced from 1.2 to 0.4 (Figure 4c). This suggests that a significant fraction of the protein in extracts from tissue in which the higher OD600 suspension had been infiltrated consisted of Agrobacteria-derived protein or plant proteins produced in response to the higher concentrations of Agrobacteria.

pEAQ-HT for direct cloning and protein tagging

The reduced size of pEAQ vectors improves their handling in cloning steps and combining elements of the system on to a single plasmid improves infiltration efficiency. However, the new vectors still require a two-step cloning procedure to introduce the sequence to be expressed: insertion of the sequence between the 5′- and 3′-UTR in the CPMV-HT cassette followed by transfer of the cassette to the appropriate binary vector. It would clearly be advantageous to construct a binary plasmid which already contains the CPMV-HT cassette into which a gene of interest could be directly inserted. To achieve this, a construct, pM81-FSC-POW, was created which contained a polylinker between a modified CPMV RNA-2 5′-UTR and the 3′-UTR. The polylinker was designed so that it not only allows the insertion of a foreign sequence into the CPMV-HT cassette, but also permits the fusion of a C- or N-terminal Histidine (His) tag if desired (Figure 5a). The modified 5′-UTR contained an additional mutation (A115G) as this has been shown to promote even higher expression levels than are obtained with the modified 5′-UTR containing only the U162C mutation (Sainsbury and Lomonossoff, 2008). The CPMV-HT cassette was then transferred into pEAQspecialK via the PacI/AscI sites to give pEAQ-HT (Figure 4a). The sequence of GFP was inserted in such a way within the polylinker of pEAQ-HT to give un-tagged GFP and 5′- (HisGFP) and 3′- (GFPHis) His-tag fusions from constructs pEAQ-HT-GFP, pEAQ-HT-GFPHis and pEAQ-HT-HisGFP respectively.

Figure 5.

 pEAQ-HT cloning and expression of GFP as native and His-tagged variants. (a) Diagrammatic representation of the T-DNA region of pEAQ-HT showing the polylinker in detail. Black boxes, T-DNA borders; green arrows, promoter sequences; and red arrows, terminator sequences. (b) Spectrofluorometric analysis of GFP expression. spK = pEAQspecialK-GFP-HT and GFP, HisGFP, and GFPHis are the pEAQ-HT clones. (c) 12% SDS-PAGE and Western blot analysis of GFP variant expression; −ve, control extract.

The levels of GFP expression obtained from pEAQ-HT-GFP, pEAQ-HT-GFPHis and pEAQ-HT-HisGFP were assessed by agro-infiltrating the constructs into N. benthamiana leaves using pEAQspecialK-GFP-HT (Figure 2) as a control. Untagged GFP was expressed from pEAQ-HT-GFP to a level (in excess of 1.5 g/kg fresh weight of tissue) even higher than that obtained with pEAQspecialK-GFP-HT as determined by spectrofluorometry (Figure 5b). This finding is consistent with the use of a 5′-UTR with the additional A115G mutation in the former. The incorporation of a His tag reduced the level of GFP detected by spectrofluorometry, although the levels were still in excess of 0.5 g/kg fresh weight of tissue (Figure 5b). The reduction was more pronounced for N-terminal compared with C-terminal His tag. SDS-PAGE of protein extracts revealed relative levels of GFP, as detected by Coomassie Blue staining, which were consistent with the amounts detected by fluorometry; as expected the presence of the His tag at either the N- or C-teminus reduced the mobility of the GFP band. Unfused and His-tagged GFP variants were all detectable in Western blots using anti-GFP antibodies whereas only HisGFP and GFPHis were detectable using anti-His antibodies, confirming that the His tags were correctly fused to GFP (Figure 5c).

GATEWAY compatibility

Expression from pEAQ-HT provides an extremely quick and easy method for protein production in plant tissue requiring only 2–3 weeks for cloning and expression. However, traditional restriction enzyme-based cloning is required, which in most cases represents a bottleneck for high-throughput expression platforms. To alleviate this, we have created a series of three GATEWAY-compatible pEAQ-HT destination vectors, which contain the double-mutated (A115G/U162C) 5′-UTR of pEAQ-HT to maximize expression and were assembled to allow the expression of an unfused protein (pEAQ-HT-DEST1) or a protein with either an N- (pEAQ-HT-DEST2) or C-terminal (pEAQ-HT-DEST3) His tag (Figure 6a). A GFP entry clone possessing both start and stop codons was used in GATEWAY recombination (LR) reactions with pEAQ-HT-DEST1 and pEAQ-HT-DEST2 destination vectors, resulting in constructs for expressing un-tagged GFP and N-terminally His-tagged GFP (HisGFP) respectively. pEAQ-HT-DEST3 was used in LR reactions with GFP entry vectors with and without stop codon thus yielding another un-tagged construct (GFP-stop) and a C-terminally His-tagged construct (GFPHis) respectively.

Figure 6.

 GATEWAY-compatible pEAQ-HT cloning and expression of GFP as native and His-tagged variants. (a) Diagrammatic representation of the HT expression cassettes of pEAQ-HT destination vectors. Black boxes, attR GATEWAY recombination sites; green arrows, promoter sequences; and red arrows, terminator sequences. (b) Spectrofluorometric analysis of GFP expression (c) 12% SDS-PAGE and Western blot analysis of GFP variant expression. −ve = control extract. CmR, chloramphenicol resistance gene; ccdB, E.coli lethal gene.

Infiltration of N. benthamiana leaves with the expression constructs created by GATEWAY recombination resulted levels of GFP expression comparable with those obtained from other pEAQ vectors. Expression levels measured by spectrofluorometry ranged from 0.7 to 1.25 g/kg fresh weight tissue (Figure 6b). A Coomassie-stained SDS-PAGE gel showed that the His-tagged GFP polypeptides had reduced mobility compared with the unfused version because of the presence not only of the His tag itself but also resulting from the addition of a further 16 and 17 amino acids for the C- and N-terminally tagged variants respectively. These are derived from the translation of the attB recombination sites and parts of both the entry clone and reading frame (Rf) cassettes used in the construction of the destination vectors. The correct fusions of His tags at both the N- and C-terminus of GFP were again confirmed by Western blotting (Figure 6c); all GFP variants were detectable with anti-GFP antibodies, whereas only HisGFP and GFPHis were detectable with anti-His antibodies (Figure 6c). However, the C-terminal tag exhibited some degree of instability as this construct yielded two bands corresponding to tagged and untagged GFP, as seen by SDS-PAGE and immunodetection of GFP and the His tag. This is most likely because of proteolysis between the tag and the GFP. Similar to the use of pEAQ-HT, utilizing pEAQ-HT-DEST vectors to generate His-tagged GFP results in variation in GFP levels. However, in this case, the presence of a His tag enhanced rather than reduced expression especially if the sequence encoding the His tag was present at the 3′ side of the GFP sequence. The precise reason for this enhancement is unclear. It is not directly because of the presence of the His residues on the GFP as incorporating a stop codon between the C-terminus of GFP and the six C-terminal His residues encoded by (pEAQ-HT-DEST3) resulted in a construct (GFP-stop) which also gave enhanced levels of GFP expression. Whatever the cause of the variation, the high levels of GFP expression from pEAQ-HT-DEST plasmids indicate that they can provide an efficient high-throughput expression platform in plants.

Discussion

To improve the ease of use and performance of the CPMV-HT expression system a modular set of vectors (the pEAQ series) has been created for the Easy And Quick transient expression of foreign proteins in plants. The binary vector, pBINPLUS, which served as the backbone of the original version of the CPMV-HT system (Sainsbury and Lomonossoff, 2008) is derived from the ground-breaking early generation binary vector pBIN19 (Bevan, 1984). It proved possible to remove more than half of the plasmid backbone and some of the T-DNA region without compromising transient expression levels. A similar proportion of the backbone had previously been removed from pBIN19 without a loss of performance (Xiang et al., 1999). However, pBINPLUS possesses two significant improvements over pBIN19 (van Engelen et al., 1995): an increased copy number in E. coli owing to the addition of the ColEI origin of replication, and a reoriented T-DNA ensuring the gene of interest is further from the LB that can suffer extensive deletions in planta (Rossi et al., 1996). The reduction in size resulted in greatly improved yields during cloning procedures using commercial plasmid extraction kits, as these are most efficient for plasmids below 10 kb (data not shown). The resulting construct, pEAQ-GFP-HT, could then be used as the basis for the development of efficient transient expression vectors.

Insertion of the P19 sequence into the T-DNA of pEAQ-GFP-HT enables both the gene of interest and the suppressor of silencing to be expressed from a single plasmid and obviates the need for co-infiltration with a separate suppressor-expressing construct to achieve high level transgene expression. Although the ability of P19 to enhance expression levels is well characterized (Voinnet et al., 2003), this study presents the first demonstration of its effectiveness when co-delivered to each cell on the same T-DNA. A previous study had reported the co-delivery of P19 in a transient assay from a separate T-DNA within the same Agrobacterium as the transgene-containing T-DNA (Hellens et al., 2005). However, there was no effect of P19 until 6 days after infiltration suggesting inefficient transfer of T-DNA. Unexpectedly, reinsertion of the NPTII gene into the T-DNA region of pEAQ-GFP-HT to potentially permit the pEAQ plasmids to be used for stable transformation, enhanced expression levels in transient expression assays.

The modular nature of the pEAQ binary vector allows any silencing suppressor and/or marker gene, if required, to be co-expressed with one or two CPMV-HT cassettes. For example, insertion of a second HT cassette into the AsiSI/MluI sites of pEAQexpress-GFP-HT would allow the user to track expression with GFP fluorescence. The presence of two cloning sites for accepting HT cassettes from our cloning vectors also allows efficient expression of multi-subunit proteins such as full-size antibodies. The expression of the monoclonal IgG 2G12 from a single plasmid represents the highest reported yield of recombinant antibody from plant tissue infiltrated with a single Agrobacterium culture. The only way of achieving similar levels with another system involves the infiltration of six separate cultures (Giritch et al., 2006). The use of a single plasmid reduces the number of bacteria needed to ensure co-delivery of multiple expression cassettes, which could provide a significant cost saving at industrial production levels. The infiltration process is also physically easier to carry out with more dilute cultures because of less clogging of the intercellular spaces of leaf tissue. A further advantage in the reduction of the Agrobacterium titres needed for efficient expression is that extracts from the infiltrated tissue contain lower levels of protein contaminants, arising from the infiltration process. This provides a very useful and unexpected advantage for downstream processing.

To make high-level expression with pEAQ vectors readily useable by research groups with no previous experience with CPMV-based expression, we have created versions of pEAQspecialK that allow the single-step cloning of foreign genes by either restriction enzyme- or GATEWAY-based methods (pEAQ-HT and the pEAQ-HT-DEST series of vectors; Figures 5 and 6). These vectors permit the expression of both unfused proteins and proteins with 6 × histidine residues at either the N- or C-terminus to facilitate protein purification. The resulting constructs also benefit from the second mutation in the 5′-UTR which enhances expression relative to HT. When tested with GFP, a degree of variability of expression was found; nevertheless, the levels of expression were still extremely high for a plant-based expression system. Furthermore, as expression takes only a matter of days, these plasmids could form the basis for a plant-based method for high-throughput screening of constructs with a minimum of resources and expertise.

Overall, the pEAQ vectors combined with CPMV-HT enable the production of milligram quantities of recombinant proteins within 2–3 weeks of sequence identification. This approach presents the opportunity to use plants as a eukaryotic alternative for protein production that does not require extensive technical knowledge or expensive facilities. Applications range from expression of proteins recalcitrant to expression in prokaryotic systems for biochemical and structural characterization, to medium- or high-throughput screening of vaccine candidates, especially where time and resources may be limiting. In addition to the co-expression of proteins for complex molecules such as antibodies and virus-like particles (unpublished data), the use of multiple expression cassettes in pEAQ plasmids also provides a means for applications such as metabolic engineering. Therefore, we anticipate that this system will provide an extremely valuable tool in both academic and industrial settings.

Experimental procedures

Vector series construction

Three essential segments of pBD-FSC2-GFP-HT, which is based on pBINPLUS (van Engelen et al., 1995), were amplified with the high fidelity polymerase PHUSION (New England Biolabs; http://www.neb.com) using oligonucleotides encoding unique restriction enzyme sites for re-ligation (Table 1). The T-DNA region was amplified with a sense primer complementary to sequences upstream of a unique AhdI site (pBD-LB-F), and an antisense primer that included an ApaI site (pBD-RB-ApaI-R). A region encompassing the ColEI origin of replication, the NPTIII gene, and the TrfA locus was amplified with a sense primer that included an ApaI site (pBD-ColEI-ApaI-F), and an antisense primer that included a SpeI site (pBD-TrfA-SpeI-R). The RK2 origin of replication (OriV) was amplified with a sense primer that included a SpeI site (pBD-oriV-SpeI-F) and an antisense primer that included an AhdI site (pBD-oriV-AhdI-R). Following purification, the products were digested according to the unique restriction sites encoded at their termini and assembled in a three-part ligation. This resulted in the plasmid pEAQbeta (Figure S1), for which the ligation junctions were verified by sequencing. We found a deletion of approximately 1.2 kb from the T-DNA which had removed a portion of the nos terminator of the CPMV-GFP-HT cassette. Therefore, we re-amplified a portion of the terminator including the RB from pBD-FSC2-GFP-HT with pMini>pMicroBIN-F2 and pBD-RB-ApaI-R, and amplified the pEAQbeta backbone including the RB with the primers pBD-ColEI-ApaI-F and pMini>pMicroBIN-R (Table 1). The purified products were digested with ApaI and FseI and ligated to give pEAQ-GFP-HT.

Table 1.   Oligos for polymerase chain reaction (PCR) amplifications and mutagenesis with restriction sites, or part thereof, in lower case, mutations in bold, and recombination sites underlined
OligoSequenceFunction
pBD-LB-FGCCACTCAGCTTCCTCAGCGGCTTTSense primer for amplification of the region 6338-12085 of pBD-FSC2-GFP-HT
pBD-RB-ApaI-RTATTAgggcccCCGGCGCCAGATCTGGGGAACCCTGTGGAntisense primer for amplification of the region 6338-12085 of pBD-FSC2-GFP-HT with ApaI site
pBD-ColEI-ApaI-FGACTTAgggcccGTCCATTTCCGCGCAGACGATGACGTCACTSense primer for amplification of the region 1704-5155 of pBD-FSC2-GFP-HT with ApaI site
pBD-TrfA-SpeI-RGCATTAactagtCGCTGGCTGCTGAACCCCCAGCCGGAACTGACCAntisense primer for amplification of the region 1704-5155 of pBD-FSC2-GFP-HT with SpeI site
pBD-oriV-SpeI-FGTAGCactagtGTACATCACCGACGAGCAAGGCSense primer for amplification of the region 14373-670 of pBD-FSC2-GFP-HT with SpeI site
pBD-oriV-AhdI-RCAGTAgacaggctgtcTCGCGGCCGAGGGGCGCAGCCCAntisense primer for amplification of the region 14373-670 of pBD-FSC2-GFP-HT with AhdI site
pMini>pMicroBIN-F2ggccggccacgcgtTATCTGCAGAgcgatcgcGAATTGTGAGCGGATAACAATTTCACACAGGAAACAGCTATGACCSense primer for amplification of the region 2969-85 of pEAQbeta with FesI-MluI-AsiSI sites
pMini>pMicroBIN-RgcgatcgcTCTGCAGATAacgcgtggccggccCTCACTGGTGAAAAGAAAAACCACCCCAGTACATTAAAAACGTCCAntisense primer for amplification of the region 2969-85 of pEAQbeta with AsiSI-MluI-FesIsites
35SP19-PacI-FttaattaaGAATTCGAGCTCGGTACCCCCCTACTCCSense primer for amplification of the 35S-P19 cassette with PacI site
35SP19-AscI-RggcgcgccATCTTTTATCTTTAGAGTTAAGAACTCTTTCGAntisense primer for amplification of the 35S-P19 cassette with AscI site
35SP19-FseI-FggccggccGAATTCGAGCTCGGTACCCCCSense primer for amplification of the 35S-P19 cassette with FseI site
35SP19-FseI-RggccggccATCTTTTATCTTTAGAGTTAAGAntisense primer for amplification of the 35S-P19 cassette with FseI site
pBD-NPTII-FseI-FggccggccTACAGTATGAGCGGAGAATTAAGGGAGTCACGSense primer for amplification of the NPTII cassette from pBD-FSC2-GFP-HT with FseI site
pBD-NPTII-FseI-RggccggccTACAGTCCCGATCTAGTAACATAGATGACACCGCGCAntisense primer for amplification of the NPTII cassette from pBD-FSC2-GFP-HT with FseI site
P19-ΔNruI-FCCGTTTCTGGAGGGTCTCGAACTCTTCAGCATCSense primer for the silent mutagenesis of the NruI restriction site within P19
P19-ΔNruI-RGATGCTGAAGAGTTCGAGACCCTCCAGAAACGGAntisense primer for the silent mutagenesis of the NruI restriction site within P19
POW-FcgaccggtATGCATCACCATCACCATCATcccgggCATCACCATCACCATCACTAGcSense oligo for polylinker, POW
POW-RtcgagCTAGTGATGGTGATGGTGATGcccgggATGATGGTGATGGTGATGCATaccggttcgSense oligo for polylinker, POW
GFP-AgeI-FATCGGaccggtATGACTAGCAAAGGAGAAGAACSense oligo for amplification of GFP with AgeI site
GFP-XmaI-FATCCGAcccgggACTAGCAAAGGAGAAGAACTTTTCACSense oligo for amplification of GFP with XmaI site and no start codon
GFP-XmaI-RATCCGAcccgggTTTGTATAGTTCATCCATGCCAntisense oligo for amplification of GFP with XmaI site and no termination codon
GFP-XhoI-RCGATCctcgagTTATTTGTATAGTTCATCCATGCCAntisense oligo for amplification of GFP with XhoI site
GFP/D-TOPO-FCACCATGACTAGCAAAGGAGAAGAACSense oligo for amplification of GFP for directional insertion into pENTR/D-TOPO
GFP-RTTATTTGTATAGTTCATCCATGCCAntisense oligo for amplification of GFP
GFP-nostop-RTTTGTATAGTTCATCCATGCCAntisense oligo for amplification of GFP with no termination codon

The P19 gene, flanked by CaMV 35S promoter and CaMV 35S terminator, was amplified from pBIN61-P19 (Voinnet et al., 2003) using either 35SP19-PacI-F and 35SP19-AscI-R, or 35SP19-FseI-F and 35S-P19-FseI-R (Table 1). The NPTII gene, flanked by the nos promoter and terminator, was amplified from pBD-FSC2-GFP-HT using pBD-NPTII-FseI-F and pBD-NPTII-FseI-R (Table 1). Following A-tailing, the amplified cassettes were ligated into pGEM-T EASY (Promega; http://www.promega.com). The P19 cassette excised from pGEM-T EASY with FseI was ligated into FseI-digested pEAQ-GFP-HT to give pEAQexpress-GFP-HT. The NPTII cassette excised with FseI was ligated into FseI-digested pEAQ-GFP-HT to make pEAQselectK-GFP-HT. The P19 cassette was also excised with PacI/AscI and ligated into the AsiSI/MluI sites of pEAQselectK-GFP-HT to give pEAQspecialK-GFP-HT.

To make the plasmids available for use with a greater range of expression cassettes, a polylinker incorporating an additional 12 unique restriction sites was inserted into the pEAQ vectors (Figure S2). This replaced the GFP-HT cassette in pEAQ, pEAQexpress, pEAQselectK and pEAQspecialK thereby bringing the total of unique sites within the T-DNA to 17, which will enable facile reconfiguration of the T-DNA as required.

pEAQ-HT for direct cloning

Oligonucleotides encoding the sense and antisense strands of a short polylinker (POW-F and POW-R; Table 1) were annealed leaving the downstream half of an NruI site at the 5′-end and an overhang matching that of XhoI at the 3′-end. The annealed oligos were ligated with NruI/XhoI digested pM81-FSC2-GFP(A115G) (U162C) (Sainsbury and Lomonossoff, 2008) to give pM81-FSC2-POW. The NruI site was removed from the P19 gene in pGEM-T EASY by site-directed mutagenesis (QUICKCHANGE; Stratagene; http://www.stratagene.com) with the primers P19-ΔNruI-F and P19-ΔNruI-R (Table 1), and was re-inserted into the AsiSI/MluI sites of pEAQselectK-GFP-HT to give pEAQspecialKΔNruI-GFP-HT which showed no reduction in expression compared with pEAQspecialK-GFP-HT (data not shown). The PacI/AscI fragment from pM81-FSC2-POW was then released and inserted into similarly digested pEAQspecialKΔNruI-GFP-HT thereby replacing the GFP-HT expression cassette and yielding pEAQ-HT (Figure 5a). GFP was amplified from pBD-FSC2-GFP-HT with a set of four primers (Table 1) in three combinations for insertion into pEAQ-HT: GFP-AgeI-F and GFP-XhoI-R; GFP-AgeI-F and GFP-XmaI-R; and GFP-XmaI-F and GFP-XhoI-R. Purified PCR products were digested with the enzymes specified in their primers and inserted into appropriately digested pEAQ-HT to give pEAQ-HT-GFP, pEAQ-HT-GFPHis, and pEAQ-HT-HisGFP.

GATEWAY compatibility

The vector pEAQ-HT was used to make three GATEWAY (Invitrogen; http://www.invitrogen.com) destination vectors. In each case, the vector was digested within the multiple cloning site to yield a linear blunt-ended backbone. A GATEWAY vector conversion kit (Invitrogen) was used to insert the appropriate Rf cassettes containing a chloramphenicol resistance gene and ccdB gene for selection, flanked by attR recombination sites. In this way, ligation of RfC.1 into the StuI/NruI digested vector yielded pEAQ-HT-DEST1; ligation of RfA into the SmaI/StuI digested vector yielded pEAQ-HT-DEST2; and ligation of RfB into the SmaI/NruI digested vector yielded pEAQ-HT-DEST3. ccdB SURVIVAL 2T1R cells (Invitrogen) were used to clone and amplify the plasmids. GFP was amplified with GFP/D-TOPO-F and either GFP-R or GFP-nostop-R (Table 1) and inserted into pENTR/D-TOPO (Invitrogen) according to the manufacturer’s instructions. To prepare entry clones for recombination reactions, the kanamycin-resistance gene present in the pENTR/D-TOPO plasmid backbone was disabled by restriction digest with BspHI. This was to prevent the generation of false positive colonies that would result from the transformation of the entry clone. The resulting fragment was gel purified and used in LR reactions with the appropriate pEAQ-HT-DEST vector performed according to the manufacturer’s instructions (Invitrogen).

Plant growth and infiltrations

Plants were grown in greenhouses maintained at 23–25 °C. Binary plasmids were maintained in Agrobacterium tumefaciens strain LBA4404, which was transformed by electroporation. Cultures were grown to stable phase in Luria-Bertani media supplemented with the appropriate antibiotics and pelleted by centrifugation at 2000 g. Following resuspension in MMA (10 mm MES (2-[N-morpholino]ethanesulfonic acid) pH 5.6, 10 mm MgCl2, 100 μm Acetosyringone) to an OD600 of 1.2 and 2–4 h incubation at ambient temperature, suspensions were pressure infiltrated into Nicotiana benthamiana leaves. Cultures harbouring binary plasmids for co-infiltrations were mixed at equal densities and/or diluted as indicated. Control infiltrations comprised a mixture of pBINPLUS and pBIN61-P19 or pEAQ-HT as appropriate, and tissue was harvested 6 days after infiltration.

Whole leaf imaging

To detect GFP, infiltrated leaves were photographed with a Nikon D1× digital camera (http://www.nikon.com) under UV illumination from a Blak-Ray B-100AP UV lamp (Blak-Ray, Upland, CA, USA).

Protein extractions, electrophoresis, and immunological detection

Protein extractions were performed on 5.3 cm2 of leaf tissue to control for differences in leaf tissue mass because of variation in the osmotic state of infiltrated tissue and the potential of some constructs to result in eventual necrosis, which can affect the mass of a given infiltrated area. This corresponds to 90 mg of healthy un-infiltrated leaf tissue and this value was used to calculate yields.

To extract GFP, infiltrated leaf tissue was homogenized in 270 μL of extraction buffer [50 mm Tris–HCl pH 7.25, 150 mm NaCl, 2 mm ethylenediaminetetraacetic acid (EDTA), 0.1% (v/v) Triton X-100]. To extract 2G12, infiltrated leaf tissue was homogenized in 270 μL of phosphate-buffered saline (PBS) with 5 mm EDTA, 3 mmβ-mercaptoethanol and 0.05% Triton X-100. Lysates were clarified by centrifugation and protein concentrations determined by Bradford assay. Approximately 20 μg of GFP-containing extracts were separated on 12% NuPage gels (Invitrogen) under reducing conditions and approximately 12.5 μg of 2G12-containing protein extract was separated by Tris–glycine SDS-PAGE under non-reducing conditions. For Western blotting, 4 μg of GFP containing extracts were separated on 12% NuPage gels and electro-blotted on to nitrocellulose using transfer buffer [20 mm Tris–HCI, 152 mm glycine, 20% (v/v) methanol]. Blocking of nonspecific binding sites was achieved with blocking solution [5% skim-milk powder in PBS plus 0.025% (v/v) Tween-20], which also served as the antibody dilution buffer. GFP was detected with Living Colors A.v. peptide antibody (Clontech; http://www.clontech.com) and an anti-rabbit horseradish peroxidase-conjugated secondary (Amersham Biosciences; http://www.gelifesciences.com). His tags were detected with the His-Tag monoclonal antibody (Novagen; http://www.merckbiosciences.co.uk) and an anti-mouse horseradish peroxidase conjugate (Amersham Biosciences). Signals generated by ECL were captured on Hyperfilm (Amersham Biosciences).

GFP assay

GFP fluorescence measurements were made using a protocol modified from Richards et al. (2003). Soluble protein extracts were diluted in 0.1 m Na2CO3 and loaded in triplicate onto a fluorescently neutral black 96-well plate. Recombinant GFP from Clontech is the same variant of GFP as was used in this study and was, therefore, used to generate standard curves in a control plant extract at the same dilution as samples. Excitation (395 nm) and emission (509 nm) maxima were matched to Clontech’s GFP and read using a SPECTRAmax spectrofluorometer (Molecular Devices; http://www.moleculardevices.com).

Acknowledgements

The authors would like to thank Andrew Davis for photography. F.S acknowledges funding from a Marie Curie Early Stage Training Fellowship MEST-CT-2004-504273 and the Trustees of the John Innes Foundation. E.C.T. acknowledges funding from a Marie Curie Early Stage Training Fellowship MEST-CT-2005-019727 and the Trustees of the John Innes Foundation. The John Innes Centre is grant-aided by the Biotechnology and Biological Science Research Council (BBSRC) UK.

Ancillary