191

Molecular Breeding 9: 191–199, 2002. © 2002 Kluwer Academic Publishers. Printed in the Netherlands.

Expressing multiple genes in a single open reading frame with the 2A region of foot-and-mouth disease virus as a linker Chonglie Ma and Amitava Mitra* Center for Biotechnology and Department of Plant pathology, University of Nebraska-Lincoln, 406 Plant Sciences Hall, Lincoln, NE 68583-0722, USA; *Author for correspondence (e-mail: [email protected]; fax: 1- 402- 472-2853) Received 1 November 2001; accepted in revised form 4 February 2002

Key words: Expression, Gene, Plant, Polyprotein, Processing Abstract The Food-and-mouth disease virus (FMDV) 2A protein is only 16–20 amino acid long. It is responsible for the ‘cleavage’ of the FMDV polyprotein at its own carboxyl-terminus. We used the ‘cleavage’ property of the 2A protein to process artificial polyproteins produced in transgenic plants. In our system, single or multiple copies of the reporter CAT and GUS genes were fused into a single open reading frame (ORF) with a copy of the FMDV 2A protein gene placed between the reporter genes. Expression of various constructs in transgenic tobacco resulted in consistent detection of freed CAT and/or GUS proteins, suggesting that FMDV 2A protein functioned properly in plant cells. ‘Cleavage’ efficiency ranged from 80% to 100% depending on the constructs. The variability in ‘cleavage’ efficiency suggested that the contexts flanking a 2A protein might modulate its activity. We further expressed constructs where multiple copies of the 2A and reporter genes were fused into one ORF. The presence of freed GUS protein together with partially processed polyprotein intermediates in the transgenic plants indicated that multiple copies of the 2A protein in a single ORF function independently. Our data demonstrate that using the FMDV 2A protease as a linker, multiple genes could be easily expressed in a single ORF. Abbreviations: CAT – chloramphenicol acetyl-transferase, GUS – ␤-glucuronidase Introduction There are several approaches that are currently available for co-expressing multiple transgenes in planta. An early strategy was to put genes into separate expression cassettes and transfer each cassette into separate plants by plant transformation. Subsequently, these genes were brought together by progressively crossing transgenic plants carrying different genes. This approach was successfully used to introduce a polyhydroxybutyrate biosynthetic pathway, comprising three genes, into Arabidopsis thaliana (Nawrath et al. 1994). Similarly, 4 genes of a secretory immunoglobulin A were co-expressed in a tobacco plant, resulting in generation and assembly of functional antibodies in plants (Ma et al. 1995). However, intensive time and labor requirements make this approach

inefficient. The entire process requires construction of multiple expression cassettes, separate transformation of each gene, screening and characterization of plants for each transgene, multiple crosses and cumbersome verification operations following every cross. Another drawback of this approach is that equi-molar expression of each transgene is virtually impossible due to variation in the level of expression of transgenes. The variation could be the consequence of use of promoters with different strengths, gene silencing induced by use of homologous promoters or multiple insertions of T-DNA [for reviews see Flavell (1994) and Meyer and Saedler (1996)], and position effects, variable expression levels caused by integration of T-DNA into different chromosomal positions (Elkind et al. 1995 Iglesias et al. 1997 Leisy et al. 1990 Liu and Tabe 1998 Peach and Velten 1991).

192 To circumvent these drawbacks, alternative approaches, based on strategies exploited by several animal Picornavirus and plant Potyviruses to process their polyproteins, have recently been developed (Ceriani et al. 1998 Halpin et al. 1999 Marcos and Beachy 1997 Urwin et al. 1988). In these viruses, the genome often encodes a single, long, open reading frame. Expression of the genome results in production of a polyprotein, which is co-translationally and post-translationally processed to mature forms by virus-encoded proteases (Riechmann et al. 1992). These proteases could cleave the polyprotein in cis (intramolecular cleavage) or in trans (cleavage in a different polyprotein) orientations. Two such proteases, 2A protein from the Foot-and-mouth disease virus (Robertson et al. 1985 Ryan et al. 1991) and NIa protein from Tobacco etch poytyvirus (TEV) (Carrington and Dougherty 1987 Carrington et al. 1988) have been well characterized. The nuclear inclusion (NIa) protein of TEV is a 49 kDa protease that frees itself from the polyprotein in an autoproteolytic reaction (Carrington and Dougherty 1987), cleaving the polyprotein at 5 positions, each defined by a consensus heptapeptide, -Glu-Xaa-Xaa-Tyr-Xaa-Gln-Ser/ Gly- (Carrington and Dougherty 1987 Dougherty et al. (1988, 1989)). Using the proteolytic activity and cleavage site of NIa, multiple foreign genes have been expressed both in vitro and in vivo. In such a system, the NIa and other genes were fused into a single open reading frame driven by a single promoter with the cleavage heptapeptide inserted between the genes for later separation (Ceriani et al. 1998 Marcos and Beachy (1994, 1997) Rorrer et al. 1992). The utility of the NIa protease is limited due to the presence of a nuclear-localizing signal within the protease and the amount of metabolic energy necessary to express the 49 kDa protease. Recently, a novel strategy, similar to the NIa system, has been tested in Arabidopsis (Urwin et al. 1988). In this approach part of the spacer region of a metallothionein-like protein was used to link two protease inhibitors to form a polyprotein. When expressed in planta, the protease inhibitors were released from the polyprotein via cleavage at the spacer region by an unknown endogenous protease. Since the availability and expression level of this mysterious endogenous protease is completely unknown for a specific plant, usage of this approach for introducing multiple gene products in plants could be very limited. The 2A protein of Foot-and-mouth disease virus has several commendable characteristics including its

small size, only 16–20 amino acid (Robertson et al. 1985) and its co-translational ‘cleavage’ activity to ‘cleave’ the 2A from the FMDV polyprotein at its own carboxyl terminus (Donnelly et al. (2001a, 2001b) Ryan et al. 1991). Although the mechanism of its ‘self-cleaving’ remains largely unknown, recent studies demonstrated that the ‘cleavage’ is not a proteolytic process but a discrete translation of a 2A protein and its downstream product through a ribosomal ‘skip’ from one codon to the next with the formation of a peptide bond (Donnelly et al. (2001a, 2001b)). Nevertheless, 2A-mediated ‘self-cleaving’ was observed in various artificial polyproteins in both mammalian (Mattion et al. 1996 Ryan and Drew 1994) and plant (Halpin et al. 1999 Santa Cruz et al. 1996) cells. In this report we evaluated the cleavage of a series of artificial polyproteins in which reporter genes, chloramphenicol acetyltransferase (CAT) and ␤-glucuronidase (GUS), were linked by the 2A protein in various combinations. We also examined the possibility of expressing up to 4 genes in a single open reading frame (ORF) using multiple copies of 2A protein gene as linkers. Expression of these constructs in transgenic tobacco plants demonstrated that the FMDV 2A protein functioned properly in all these polyproteins. However, variation in ‘cleavage’ efficiency was observed in some polyproteins, suggesting that the ‘cleavage’ could be modulated by the contexts flanking the 2A protein. Our results demonstrate the possibility of expressing pathway transgenes in a single step.

Materials and methods Plasmid constructions To construct an expression cassette where the CAT and GUS reporter genes were fused with FMDV 2A protease in a single open reading frame, we designed two long primers. Primer CAT1 was designed as a 5⬘ primer containing the translation initiation codon ATG and an XhoI site. Primers used in this study are shown in Table 1. Primer CAT2 was the complementary sequence of the sense strand that encoded for 11 amino acids from the C-terminus of the CAT gene, 2 residues V (valine) and D (aspartic acid) (SalI site) and 7 residues from the N-terminus of 2A protein. In order to fuse 2A and CAT genes into one frame, the stop codon at the CAT 3⬘ end was removed. Primer GUS1 (Table 1) was a sense strand primer containing

193 Table 1. Primers used in this study. Name

Sequence

Descritption

CAT1

5⬘-GT CTC GAG ATG GAG AAA AAA ATC ACT GGA TAT ACC ACC

Xho I, 5⬘ CAT

CAT2

GTT-3⬘ 5⬘GGAAGCTTGAGAAGATCGAAGTTGTC-

Hind III, Sal I, 5⬘ 2A, 3⬘CAT with

CAT3 CAT4 CAT5 GUS1

GACTGCTCCACCTTGCCATTCAT CGCAGTACTGTTG-3⬘ 5⬘-TACCCGGGTGCTCCACCTTGCCATTCATCGCAGTACTGTTG-3⬘ 5⬘-GTCCCGGGATGGAGAAAAAAATCACTGAATATACCACCGTT-3⬘ 5⬘-GTGGATCCTTATGCTCCACCTTGCCATTCATCGCAGTA-3⬘ 5⬘-GTAAGCTTGCTGGTGATGT-

stop codon removed Sma I, 3⬘ CAT, no stop codon Sma I, 5⬘ CAT Bam HI, 3⬘CAT with stop codon Hind III, Sma I, 3⬘ 2A, 5⬘ GUS

TGAATCTAACCCTGGACCCGGGATGTTA GUS2 GUS3 GUS4 GUS5 2aSalI 2aKpnI CATSalI CATKpnI *

CGTCCTGTAGAAACCCCAA-3⬘ 5⬘-GTGGATCCTCATTGTTTGCCTCCCTGCTGCGGTTTT-3⬘ 5⬘-GTCTCGAGTCCCTTATGTTACGTCCTGTA-3⬘. 5⬘-GTGTCGACTTGTTTGCCTCCCTGCTGCGGTTTT-3⬘. 5⬘-GTCCCGGGTTGTTTGCCTCCCTGCTGCGGTTTT-3⬘. 5⬘ GCCGTCGACAACTTCGATCTTCTC-3⬘. 5⬘-GTAGGTACCAACTTGGATCTTCTCAAATTG-3⬘. 5⬘-GAAGTCGACTGCTCCACCTTGCCATTCATC-3⬘. 5⬘-GAAGGTACCTGCTCCACCTTGCCATTCATC-3⬘.

Bam HI, 3⬘GUS with stop codon Xho I, 5⬘ GUS Sal I, 3⬘ GUS, no stop codon Sma I, 3⬘ GUS, no stop codon Sal I, 5⬘ 2A Kpn I, 5⬘ 2A Sal I, 3⬘ CAT, no stop codon Kpn I, 3⬘ CAT, no stop codon

Underline nucleotides represent recognization sequences for restriction enzymes.

codons encoding for 11 residues from the C-terminus of 2A protease, 8 residues from the N-terminus of the GUS gene and 2 residues each for P (Proline) and G (Glycine) (SmaI site). To facilitate cloning a HindIII site (coding for K and L residues) was added to the 3⬘ end of primer CAT2 and the 5⬘ end of primer GUS1. Using primers CAT1 and CAT2, a fragment containing the whole CAT gene and a partial 2A gene was amplified from pUC119-35SCAT (unpublished) by polymerase chain reaction (PCR). The PCR product was digested with XhoI and HindIII and cloned into the XhoI and HindIII sites of pGEM-7z (Promega) to form pGEM-CAT2A⌬. Using primers GUS 1 and GUS2 (3⬘ primer with original stop codon and an extra BamHI site, Table 1), a GUS fragment (carrying partial 2A gene at its 5⬘) was amplified from the plasmid pBI121.2 (Jefferson 1987) by PCR. The GUS fragment was digested with HindIII and BamHI and cloned into the HindIII and BamHI digested pGEMCAT2A⌬ to form pGEM-CAT/2A/GUS (pGEMCAG). The CAT2AGUS DNA fragment was recovered by cutting pGEM-CAG with XhoI (filled in with Klenow) and BamHI and cloned into the HpaI and BglII sites of the binary vector pAM696 (unpublished) to form pCAG. This construct was used to for making all other constructs. A constitutive CaMV 35S promoter was used to express each protein or polyprotein diagramed in Figure 1.

Enzyme assays All GUS staining procedures and CAT enzyme assays were carried out as described by Jefferson (1987) and Gorman et al. (1982), respectively. Northern blots Total RNA was extracted from plant leaves using the TRIZOL Reagent (GIBCO BRL, Rockville, MD) in accordance with the manufacturer’s instructions. About 20 ␮g total RNA from each sample was fractionated on 1.2% agarose gels. Subsequently, RNA was transferred onto a Zeta-Probe membrane (Bio Rad Lab, Hercules, CA) and hybridized to probes generated from CAT or GUS gene DNA following the manufacture’s instructions. Western blots Total protein was extracted following the method described by Mitra and An (1989). Equal amounts of total protein (about 20–40 ␮g) from each sample were separated on a 7.5%, 12% or 16.5% polyacrylamide BioRad Ready Gel (BioRad Lab, Hercules, CA). Proteins were blotted onto PVDF membrane using MiniTrans-Blot Electrophoretic Transfer (BioRad). The membrane was first treated with a diluted (1:1000)

194 rabbit antibodies to GUS or CAT (5⬘–3⬘, Inc., Boulder, CO). The immune blot assay was carried out using an Amplified Alkaline Phosphatase Goat AntiRabit Immuno-Blot Assay Kit (BioRad) following the manufacturer’s instructions.

Results Plasmid construction and plant transformation To examine the ‘cleavage’ ability of FMDV 2A in engineered environments, we designed a series of constructs that would express various artificial polyproteins and in which the 2A was flanked by different contexts (Figure 1). In construct pCAG (expressing polyprotein CAT2AGUS), 2A was preceded by a small CAT gene (657 base pairs) and followed by a large GUS gene (1809 base pairs), while in construct pGAC (expressing polyprotein GUS2ACAT), the positions of CAT and GUS genes were reversed. In another construct pGAG, a copy of the 2A gene was flanked by 2 GUS genes. We also built 4 constructs (pCGC, pC2G, pGC2 and pC3G) that contained multiple copies of 2A. In each of these 4 constructs a copy of 2A was used to link each adjacent pair of reporter CAT and/or GUS genes (Figure 1). In addition, we built 4 control plasmids, pCAT (containing a single CAT gene), pGUS (containing a single GUS gene), pCG and pGC (identical to pCAG and pGAC except that they did not contain a 2A gene). All these constructs were transferred into Nicotiana tabacum var. Xanthi-nc via Agrobacterium- mediated transformation (Horsch et al. 1985). Transgenic plants were screened for CAT and/or GUS activity as described in Methods. For transgenic plants expressing CAT, GUS or polyproteins CAT2AGUS, GUS2ACAT, CATGUS and GUSCAT, high enzyme activities of corresponding genes were detected in more than 70% (37 out of 48) of transgenic plants. Northern blot analyses were performed on representative plants that showed high levels of CAT and/or GUS activity. Results demonstrated that transcripts encoding CAT or GUS protein or polyprotein CAT2AGUS, CATGUS, GUS2ACAT or GUSCAT were correctly produced. (Figures 2A and 2B).

The FMDV 2A region functions properly in plants but with different ‘cleavage’ effıciencies in different contexts To examine various constructs expressing GUS proteins in plants, Western blot analyses were performed using polyclonal antibodies against GUS and results are shown in Figure 3. Originally, independent transgenic plants with the same construct were tested by Western blot analyses (2 plants each for constructs pC2G, pGC2 and pC3G, and a minimum of 5 plants each for all other constructs). Since the ‘cleavage’ efficiency of the FDMV 2A did not show obvious variation among replicates, result of one representative from each construct is shown (Figure 3). We also carried out Western blot analyses to detect CAT proteins. The commercially available anti-CAT antibodies could only detect freed CAT protein and low molecular weight CAT2A polyprotein, but not high molecular weight polyproteins such as CAT2AGUS, CATGUS, CATCAT. Thus, these results could not be used in the estimation of 2A ‘cleavage’ efficiency. Two specific bands were detected by polyclonal anti-GUS antibodies in a transgenic plant expressing polyprotein CAT2AGUS (Figure 3A, lane 4). Only one band was detected in transgenic plants expressing a single GUS protein or polyprotein CATGUS (Figure 3A, lanes 1 and 2). The low molecular weight band detected in the CAT2AGUS plant had an identical migration rate as GUS, representing release of the GUS protein from the CAT2AGUS polyprotein via 2A-mediated ‘cleavage’. The high molecular weight band detected in the CAT2AGUS plant had a similar migration rate as the polyprotein CATGUS, representing ‘uncleaved’ CAT2AGUS polyprotein. Expression of CATGUS polyprotein resulted in detection of a unique band (Figure 3A, lane 3), indicating that the FMDV 2A region was the sole factor to mediate the release of GUS from the CAT2AGUS polyprotein. Comparing the density of the high and low molecular weight bands, we estimated that the ‘cleavage’ efficiency of the 2A was about 80%, similar to that reported previously (Halpin et al. 1999 Ryan and Drew 1994). However, higher ‘cleavage’ efficiency was observed in transgenic plants expressing either polyprotein GUS2ACAT or GUS2AGUS (Figure 3B, lane 4 vs 5 and Figure 3C, lane 3 vs 4). In the GUS2ACAT plant, a strong band of the freed GUS2A polyprotein was detected, whereas the band of unprocessed GUS2ACAT polyprotein was barely detectable, indicating a nearly 100% ‘cleavage’ effi-

195

Figure 1. Schematic diagram of various artificial polyproteins. The entire amino acid sequence of the FMDV 2A (underlined) and partial flanking sequences of GUS (bold) and/or CAT (italic) proteins are indicated on the top of each construct. The ‘cleavage’ site of 2A protein is marked with a solid triangle. Plasmid names are mentioned on the left of each construct.

ciency (Figure 3B, land 4). Similarly, a ⬃ 95% ‘cleavage’ efficiency was obtained in the GUS2AGUS plant (Figure 3C, lane 4). Together, these data demonstrate that the FMDV 2A region functions in all 3 polyproteins but with different efficiency.

Multiple copies of 2A in a single open reading frame function properly and independently In previous studies, the ‘cleavage’ activity of the 2A was tested only in two-gene systems (CAT2AGUS) (Halpin et al. 1999 Ryan and Drew 1994). If we want to express more than two genes in such a system,

196

Figure 2. Northern blot analyses of transgenic tobacco containing various constructs. pGUS (lane 1 in A and B), pCAT (lane 2 in A and B), pCG (lane 3 in A and B). For internal control an 18S rDNA probe was used. The approximate size of each band is indicated on the right.

multiple copies of 2A protease must be used and evaluated. For example, will multiple copies of 2A protease function properly, or will the copies interact with each other leading to enhanced or decreased proteolytic activity? To answer these questions, we made 4 constructs, pCGC, pC2G, pGC2 and pC3G (Figure 1) which were used to transform tobacco. Immunodetections of GUS protein and polyproteins were performed on plants of these constructs and results are shown in Figure 3D–3E. High level of GUS2A poplyprotein was detected in the plants with construct pCGC, suggesting that both the copies of 2A in polyprotein CAT2AGUS2ACAT function properly. In addition, partially ‘cleaved’ products, polyproteins CAT2AGUS2A and GUS2ACAT, were clearly detected, whereas the unprocessed CAT2AGUS2ACAT polyprotein was detected only as a faint band (Figure 3D, lane 4). However, unprocessed polyproteins were clearly detected in plants of both constructs pC2G and pC3G (lane 4 in Figure 3D vs lanes 4 and 6 in Figure 3E). An explanation for this difference is that the second 2A copy in polyprotein CAT2AGUS2ACAT processed the polyprotein in a nearly 100% efficiency as

observed in polyprotein GUS2ACAT (Figure 3B, lane 4) since it shares the same context as the 2A copy in polyprotein GUS2ACAT (Figure 1). Immunodetection of CAT protein in the plant with construct pCGC showed that the accumulation of CAT protein released from the ‘cleavage’ of the second 2A copy was higher than that of CAT2A polyprotein released from the ‘cleavage’ of the first 2A copy (data not shown). This further confirms that the second 2A copy had higher ‘cleavage’ efficiency than the first one. This also suggests that the two copies of 2A in polyprotein CAT2AGUS2ACAT function independently and that their ‘cleavage’ efficiencies are determined only by the 2A protein and its flanking sequences. All possible products derived from the polyproteins, including completely-, partially-, and unprocessed polyproteins, were clearly detected in both plants of pC2G and pC3G (Figure 3E, lanes 4 and 6), indicating that all the 2A copies function properly. The protein band pattern detected in the plant of pGC2 was very similar to that in the plant of pGAC (lane 5 in Figure 3E vs lane 4 in Figure 3B), indicating that the ‘cleavage’ efficiency of the first 2A copy is the same (nearly 100%) as in polyprotein GUS2ACAT.

197

Figure 3. Immunodetection of GUS proteins and polyproteins in transgenic plants expressing various constructs. Protein or polyprotein expressed in the corresponding plant is indicated on the top of each lane. Positions of proteins and polyproteins detected by anti-GUS antibodies are indicated on the right of each panel.

Discussion The Previous studies demonstrated that both GUS and CAT proteins were efficiently released from an artificial polyprotein, CAT2AGUS, via the ‘cleavage’ mediated by the FMDV 2A region in both mammalian and plant cells (Halpin et al. 1999 Ryan and Drew 1994). Since the two reporter genes, CAT and GUS, and 2A protease were fused into a single open reading frame that was driven by a single promoter, coordinated expression of CAT and GUS was achieved (Halpin et al. 1999). In addition, the FMDV 2A is a very small protein of only 16–20 amino acids (com-

pared to the 49 kDa NIa of TEV). Because of its small size the 2A protein can be an ideal system for co-expressing multiple genes in a single open reading frame. However, these researchers only examined a CAT2AGUS polyprotein. In another study, a TMV CP-GFP fusion using a single 2A was examined in the TMV viral genome and efficient ‘cleavage’ was observed (Santa Cruz et al. 1996). Here we present a series of data to demonstrate that (i) the FMDV 2A region functions properly within different contexts, but the cleavage efficiency varies as flanking context changes and; (ii) multiple copies of 2A protease can

198 be used for the co-ordinate expression of multiple genes. We first investigated the ‘cleavage’ activity of 2A protein in 3 different contexts CAT2AGUS, GUS2ACAT and GUS2AGUS. Similar to previous reports (Halpin et al. 1999 Ryan and Drew 1994 Santa Cruz et al. 1996), high processing efficiency (> 80%) was achieved in all 3 polyproteins, suggesting that the FMDV 2A protein functions in diverse foreign contexts. However, variation (ranging from ⬃ 80 to ⬃ 100%) in ‘cleavage’ efficiency was observed among these polyproteins (Figure 3A–3C). Variation was even found between the two 2A copies of polyprotein CAT2AGUS2ACAT. Examining the ‘cleavage’ efficiency in different plants with the same construct revealed that the 2A ‘cleavage’ efficiency of a particular polyprotein was consistent among individuals. However, variation was observed among different constructs. Collectively, our data suggest that although the ‘cleavage’ property of 2A protein is independent of context, its ‘cleavage’ efficiency might be modulated by the flanking context. This may not be a problem for a coordinate expression of two genes since a single ‘cleavage’ within a polyprotein will always result in equal amount of both freed products. However, variable ‘cleavage’ efficiency, as observed in the expression of polyprotein CAT2AGUS2ACAT where the amount of CAT protein detected was higher than that of CAT2A polyprotein, may complicate a system in which two or more 2A copies are used. We also examined the possibility of expressing more than two genes in this system. Up to 4 gene fragments were efficiently expressed (Figure 3D–3E). Importantly, the multiple 2A copies within a single polyprotein function independently. Furthermore, our results demonstrated that the ‘cleavage’ efficiency of a 2A protein within a certain context (for example GUS2ACAT) is consistent regardless whether the context is alone (polyprotein GUS2ACAT) or linked with other polyproteins (polyproteins GUS2ACAT2ACAT and CAT2AGUS2ACAT). Such a property will greatly facilitate construct designs for multiple gene expression systems. The ‘cleavage’ efficiency of each 2A copy can be easily predicated using an equivalent two gene version. However, there remains another potential problem in this system. Attachment of a 2A protein to the Cterminus of the upstream protein and a proline residue to the N-terminus of the downstream protein could have adverse effects on the enzymatic activity and/or biological function of certain proteins. Data

obtained in this study show that CAT enzymatic activity in plants expressing construct pGAC and pGC2 was decreased presumably due to the extra proline residue at the N-terminus of CAT protein (data not shown). Despite of these drawbacks, this system is still a very powerful tool to express multiple genes. With proper design and testing it allows introduction of multiple genes in a single transformation event without having to provide individual promoters and terminators.

Acknowledgements We thank Dr Allan Zipf for critical review of the manuscript and Dan Higgins for expert technical assistance. This manuscript has been assigned Journal Series No. 13176, Agricultural Research Division, University of Nebraska.

References Carrington J. and Dougherty W. 1987. A viral cleavage site cassette: Identification of amino acid sequences required for tobacco etch virus polyprotein processing. Proc. Natl. Acad. Sci. USA 85: 3391–3395. Carrington J.C., Cary S.M. and Dougherty W.G. 1988. Mutational analysis of tobacco etch virus polyprotein processing: cis and trans proteolytic activities of polyproteins containing the 49-kilodalton proteinase. J. Virol. 62: 2313–2320. Ceriani M.F., Marcos J.F., Hopp H.E. and Beachy R.N. 1998. Simultaneous accumulation of multiple viral coat proteins from a TEV-Nia based expression vector. Plant Mol. Biol. 36: 239– 248. Donnelly M.L.L., Luke G., Mehrotra A., Li X., Hughes L.E., Gani D. et al. 2001a. Analysis of the aphthovirus 2A/2B protein ‘cleavage’ mechanism indicates not a proteolytic reaction, but a novel translational effect: a putative ribosomal ‘skip’. J. Gen. Virol. 82: 1013–1025. Donnelly M.L.L., Hughes L.E., Luke G., Mendoza H., ten Dam E., Gani D. et al. 2001b. The ‘cleavage’ activity of foot-and-mouth disease virus 2A site-directed mutants and naturally occurring ‘2A-like’ sequences. J. Gen. Virol. 82: 1027–1041. Dougherty W., Carrington J., Cary S. and Parks T. 1988. Biochemical and mutational analysis of a virus polyprotein cleavage site. EMBO J. 7: 1281–1287. Dougherty W.G., Cary S.M. and Parks T.D. 1989. Molecular genetic analysis of a plant virus polyprotein cleavage site: A model. Virology 171: 356–364. Elkind Y., Nir B. and Nadler H.T. 1995. Quantitative analysis of the transgene variability among primary tobacco transformants. Transgenic Research 4: 30–38.

199 Flavell R.B. 1994. Inactivation of gene expression in plants as consequence of specific sequence duplication. Proc. Natl. Acad. Sci. USA 91: 3490–3496. Gorman C.M., Moffat L.F. and Howard B.H. 1982. Recombinant genomes which express chloramphenicol acetyltransferase in mammalian cell. Mol. Cell Biol. 2: 1044–1051. Halpin G., Cooke S.E., Barakate A., El Emarani A. and Ryan M.D. 1999. Self-processing 2A-polyproteins-a system for co-ordinate expression of multiple proteins in transgenic plants. Plant J. 17: 453–459. Horsch R.B., Fry J.E., Hoffmann N.L., Eichholtz D., Rogers S.G. and Fraley R.T. 1985. A simple and general method for transferring genes into plants. Science 227: 1229–1231. Iglesias V.A., Moscone E.A., Papp I., Neuhuber F., Michalowski S., Phelan T. et al. 1997. Molecular and cytogenetic analyses of stably and unstably expressed transgene loci in tobacco. Plant Cell 9: 1251–1264. Jefferson R.A. 1987. Assaying chimeric genes in plants: the GUS fusion system. Plant Mol. Biol. Reptr. 5: 387–405. Leisy D.J., Hnilo J., Zhao Y. and Okita T.W. 1990. Expression of a rice glutelin promoter in transgenic tobacco. Plant Mol. Biol. 14: 41–45. Liu J.-W. and Tabe L.M. 1998. The influences of two plant nuclear matrix attachment regions (MARs) on gene expression in transgenic plants. Plant Cell Physiol 39: 115–123. Ma K.C.J., Hiatt A., Hein M., Vine N.D., Wang F., Stabila P. et al. 1995. Generation and assembly of secretory antibodies in plants. Science 268: 716–719. Marcos J.F. and Beachy R.N. 1994. In vitro characterization of a cassette to accumulate multiple proteins through synthesis of a self-processing polypeptide. Plant Mol. Biol. 24: 459–503. Marcos J.F. and Beachy R.N. 1997. Transgenic accumulation of two plant virus coat proteins on a single self-processing polypeptide. J. Gen. Virol. 78: 1771–1778. Mattion N.M., Harnish E.C., Crowley J.C. and Reilly P.A. 1996. Foot-and-mouth disease virus 2A protease mediated cleavage in attenuated sabin 3 poliovirus vectors engineered for delivery of foreign antigenes. J. Virology 70: 8124–8127.

Meyer P. and Saedler H. 1996. Homology-dependent gene silencing in plants. Ann. Rev. Plant Physiol. Plant Mol. Biol. 47: 23– 48. Mitra A. and An G. 1989. Three distinct regulatory elements comprises the upstream promoter region of the nopaline synthase gene. Mol. Gen. Genet. 215: 294–299. Nawrath C., Poirier Y. and Somerville C. 1994. Targeting of the polyhydroxybutyrate biosynthetic pathway to the plastids of Arabidopsis thaliana results in high levels of polymer accumulation. Proc. Natl. Acad. Sci. USA 91: 1276–12764. Peach C. and Velten J. 1991. Transgene expression variability (position effect) of CAT and GUS reporter genes driven by linked divergent T-DNA promoters. Plant Mol. Biol. 17: 49–60. Riechmann J., Lain S. and Garcia J. 1992. Highlights and prospects of potyvirus molecular biology. J. Gen. Virol. 73: 1–16. Robertson B.H., Grumbman M.J., Weddell G.N., Moore D.M., Welsh J.D., Fishcer T. et al. 1985. Nucleotide and amino acid sequence coding for polypeptides of foot-and-mouth disease virus type A12. J. Virol. 54: 651–660. Rorrer K., Parks T.D., Scheffler B., Bevan M. and Dougherty W.G. 1992. Autocatalytic activity of the tobacco etch virus NIa proteinase in viral and foreign protein sequences. J. Gen. Virol. 73: 775–783. Ryan M.D., King A.M.Q. and Thomas G.P. 1991. Cleavage of foot-and-mouth disease virus polyprotein is mediated by residues located within a 19 amino acid sequence. J. Gen. Virol. 72: 2727–2732. Ryan M.D. and Drew J. 1994. Foot-and-mouth disease virus 2A oligopeptide-mediated cleavage of an artificial polyprotein. EMBO J. 13: 928–933. Santa Cruz S., Chapman S., Roberts I.M., Prior D.A.M. and Oparka K.J. 1996. Assembly and movement of a plant virus carrying a green fluorescent protein overcoat. Proc. Natl. Acad. Sci. USA 93: 6286–6290. Urwin P.E., McPherson M.J. and Atkinson H.J. 1988. Enhanced transgenic plant resistance to nematodes by dual proteinase inhibitor constructs. Planta 204: 472–479.