key: cord-0872102-2piyywhp
authors: Bendahmane, Mohammed; Koo, Moses; Karrer, Erik; Beachy, Roger N
title: Display of epitopes on the surface of tobacco mosaic virus: impact of charge and isoelectric point of the epitope on virus-host interactions
date: 1999-07-02
journal: J Mol Biol
DOI: 10.1006/jmbi.1999.2860
sha: 216f0aafe55479012c2736fecd91b270aaeb2e0b
doc_id: 872102
cord_uid: 2piyywhp

The biophysical properties of the tobacco mosaic tobamovirus (TMV) coat protein (CP) make it possible to display foreign peptides on the surface of TMV. The immunogenic epitopes G5-24 from the rabies virus (RV) glycoprotein, and 5B19 from murine hepatitis virus (MHV) S-glycoprotein were successfully displayed on the surface of TMV, and viruses accumulated to high levels in infected leaves of Nicotiana tabacum Xanthi-nn. The peptide RB19, which contains an arginine residue plus the 5B19 epitope fused to the CP (TMV-RB19), resulted in the induction of necrotic local lesions on inoculated leaves of N. tabacum Xanthi-nn and cell death of infected BY2 protoplasts. RNA dot blot assays confirmed that expression of the acidic and basic pathogenesis-related PR2 genes were induced in infected Xanthi-nn leaf tissue. TMV that carried epitope 31D from the RV nucleoprotein did not accumulate in inoculated tobacco leaves. Analysis of hybrid CPs predicted that the isoelectric points (pI):charge value was 5.31: − 2 for wild-type CP, 5.64: − 1 for CP-RB19, and 9.14: + 2 for CP-31D. When acidic amino acids were inserted in CP-RB19 and CP-31D to bring their pI:charge to near that of wild-type CP, the resulting viruses TMV-RB19E and TMV-4D:31D infected N. tabacum Xanthi-nn plants and BY2 protoplasts without causing cell death. These data show the importance of the pI of the epitope and its effects on the hybrid CP pI:charge value for successful epitope display as well as the lack of tolerance to positively charged epitopes on the surface of TMV.

A variety of plant and animal viruses have been studied for their potential use as carrier and presenter of foreign peptides as linear or conformational antigens, and some have been tested as potential animal vaccines (Haynes et al., 1986; Hamamoto et al., 1993; Usha et al., 1993; Jagadish et al., 1993; Fitchen et al., 1995; Sugiyama et al., 1995; Turpen et al., 1995; Beachy et al., 1996; Lomonossoff & Johnson, 1996; Burke et al., 1988; Dedieu et al., 1992) . Plant viruses are of special interest as carriers of immunogenic peptides and for use as potential vaccines, in part because they are not infectious in humans and other animals. Compared to most animal viruses, the structure of most plant viruses is very simple, and are often formed with a single coat protein (CP) type. Most plant viruses are not enveloped, and many accumulate to very high levels in their host. For these reasons, viruses such as tobacco mosaic tobamovirus (TMV), cowpea mosaic comovirus (CPMV) and johnsongrass mosaic potyvirus (JGMV) have been investigated for their potential use as epitope presenters with different degrees of success (Haynes et al., 1986; Hamamoto et al., 1993; Usha et al., 1993; Jagadish et al., 1993; Fitchen et al., 1995; Sugiyama et al., 1995; Turpen et al., 1995; Beachy et al., 1996; Lomonossoff & Johnson, 1996) .

TMV is composed of a single type of CP of 17.5 kDa that encapsidates the 6.4 kb viral genome of positive-sense RNA into rod-shaped virus particles of about 300 nm  18 nm. TMV possesses characteristics which make it an attractive particle on which to display peptides including epitopes: (i) the structure of the TMV CP (Figure 1(d) ) is known at 2.9 A Ê resolution (Namba et al., 1989) , making it possible to apply strategies for rational design of foreign sequences on the surface of virus particles. (ii) TMV CP self assembles, and assembled CP and virions are stable at optimal conditions (Durham et al., 1971) . (iii) TMV accumulates to high levels in many host plants. Levels as high as 10 mg of virus per gram of fresh leaf material can be obtained by simple puri®cation procedures. Haynes et al. (1986) were the ®rst to take advantage of the self-assembly properties of the TMV CP and expressed a peptide from VP1 of poliovirus 3 as a fusion with TMV CP in Escherichia coli. Rats that were immunized with the hybrid CP carrying the poliovirus epitope produced neutralizing antibodies speci®c to the poliovirus VP1 peptide.

In some cases, epitopes expressed on the surface of plant viruses interfere with the normal virus infection cycle. Such results were observed with TMV and CPMV where insertion of certain foreign from the MHV Sglycoprotein. Numbers above the arrows represent amino acids in Sglycoprotein (Buchmeier et al., 1984; Luytjes et al., 1989) . (b) Peptides from rabies virus glycoprotein and nucleoprotein. The amino acid coordinates in the nucleoprotein (for the peptide 31D) and in the glycoprotein (for the peptide G5-24) are indicated (Dietzschold et al., 1990) . Amino acid residues inserted to neutralize the positive charge of the peptides are underlined. (c) Site of insertion of the peptides in TMV CP. The amino acid coordinates for w.t. CP are indicated (Goelet et al., 1982) . The map of the vector pCPs/p developed in this study is shown.

(d) Computer-assisted graphic representation of TMV CP structure. The position of the insertion of the different peptides at the C terminus of the CP is indicated. sequences resulted in interference with viral assembly, loss of systemic infection, and/or induction of a necrotic response in the systemic host (Takamatsu et al., 1990; Porta et al., 1994; Beachy et al., 1996) . Such effects limit the use of viruses for production of peptides, including their use as vaccines.

Here we used TMV as a model to investigate why some CP hybrids induce cell death or failed to accumulate in infected leaves. Our studies demonstrate the importance of the isoelectric point (pI) of the peptide and the pI:charge value of the hybrid CPs on cell death and virus infectivity. Such effects can be prevented by lowering the pI:charge value of the hybrid CP to a value close to that of wild-type (w.t.) TMV CP.

Chimeric CP molecules were produced by inserting nucleotide sequences encoding different peptides between TMV nucleotides 6176 and 6177 or between nucleotides 6188 and 6189 (Goelet et al., 1982) , which placed the peptides between amino acid residues Ser154 and Gly155 (as described by Fitchen et al., 1995) or after Thr158 (last amino acid of the CP). The chimeras were created using either PCR-based site-directed mutagenesis, or direct cloning in the vector pCPs/p (Figure 1(c) ). The vector pCPs/p was constructed by inserting the restriction sites SpeI and PpuMI in the CP gene sequence, making it possible to insert peptide sequences between amino acid residues Ser154 and Gly155 in the CP reading frame.

RNA transcripts produced from cDNA clone TMV-G5.24, which contains the coding sequence of the peptide G5-24 from the rabies virus (RV) glycoprotein (amino acid residues 253 to 275; Dietzschold et al., 1990) fused to the TMV CP, produced normal necrotic local lesions on Nicotiana tabacum Xanthi-NN plants three to four days postinoculation (d.p.i.). On inoculated N. tabacum Xanthi-nn, systemic infection with TMV-G5.24 was delayed by about 10 to 15 days compared with infection by w.t. TMV. Similar delays in systemic infection have been observed with certain assembly defective mutants of TMV CP Culver et al., 1995) . Symptoms produced by TMV-G5.24 were more mild than those produced by w.t. TMV, and consisted of large yellow patches rather than light green-dark green symptoms produced by w.t. TMV.

Virus particles could be puri®ed from the systemically infected leaves with a yield of 1 to 3 mg per gram of leaves. Total protein extracts from infected leaves and protein isolated from puri®ed virus were separated by SDS-PAGE. As shown in Figure 2 Expression of Epitopes on the Surface of TMV anti-G5-24 (MAb5) was successfully bound to puri®ed TMV-G5.24 in immunogold labeling experiments, as shown in Figure 2 (c). These results demonstrate that the G5-24 peptide is exposed on or near the surface of the virion.

In contrast, when the 31D epitope, comprising amino acid residues 404 to 418 from the RV nucleoprotein was similarly inserted into the CP, the resulting virus TMV-31D failed to produce local or systemic infections on N. tabacum Xanthinn plants. We were unable to detect CP-31D or virus particles in inoculated or non-inoculated leaves. However, Western blot analysis of total protein from BY2 protoplasts inoculated with TMV-31D revealed that the CP-31D accumulated in these cells (Figure 3(b) ). Virus particles could not be detected in extracts of the protoplasts. Northern blot analysis con®rmed TMV-31D replication with concomitant accumulation of viral genomic and CP subgenomic RNAs in infected BY2 cells (data not shown).

We addressed the basis for the lack of accumulation of the CP-31D in inoculated leaf tissues by looking at the composition of the peptide 31D. Peptide 31D contains a high content of basic amino acids (three Arg residues and one Lys residue; Figure 1 (b)), resulting in a peptide with a predicted pI of 12 (Table 1) . This resulted in a pI:charge value of the hybrid CP-31D molecule of 9.14: 2; the pI:charge value of w.t. CP is 5.31:-2. Using site-directed mutagenesis we introduced two Asp residues in the CP-31D between Ser154 of the CP and the 31D peptide. The resulting virus clone, referred to as TMV-2D:31D (Figure 1(b) ), encodes CP-2D:31D with a pI:charge value of 7.09:0. CP-2D:31D failed to accumulate in inoculated N. tabacum Xanthi-nn plants, and in infected BY2 protoplasts CP-2D:31D accumulated to slightly higher levels compared with CP-31D (Figure 3(b) ). Two additional Asp residues were inserted in CP-2D:31D to give rise to TMV-4D:31D (Figure 1(b) ), bringing the pI:charge value of the chimeric CP to 5.43: À 2 (Table 1) . TMV-4D:31D infected N. tabacum Xanthi-nn plants and the CP-4D:31D accumulated to high levels in inoculated leaves (Figure 3 (a)) and assembled to form viruslike particles. However, the infection did not spread to upper leaves.

The peptide 5B19 (amino acid residues 900 to 909; Buchmeier et al., 1984; Luytjes et al., 1989) from murine hepatitis virus (MHV) spike glycoprotein was inserted between Ser154 and Gly155 of the TMV CP (Koo et al., 1999) . Virus that carries the hybrid CP-5B19 produced a systemic infection on N. tabacum Xanthi-nn, and virus particles were puri®ed from infected plants (Koo et al., 1999) . However, most of virus particles formed by TMV-5B19 were insoluble and co-puri®ed with cell debris. The hybrid TMV-RB19, which contains an Arg residue between Ser154 of the CP and the 5B19 peptide ( Figure 1 (a); this virus was created to facilitate release of the MHV peptide by trypsin digestion) produced necrotic local lesions on N. tabacum Xanthi-NN similar to those produced by TMV-U1. On the systemic host N. tabacum Xanthi-nn, TMV-RB19 also produced a hypersensitive response, and the inoculated plants developed necrotic local lesions that limited the virus to the inoculated leaves (Figure 4(a) ).

To determine if the necrotic reaction was caused by the position at which the epitope was inserted or due to this peptide per se, the RB19 epitope was inserted in-frame after Thr158 (last amino acid residue of CP) to generate TMV-RB19-2. When TMV-RB19-2 was applied to Xanthi-nn plants, necrotic local lesions similar to those observed with TMV-RB19 developed on the inoculated leaves (data not shown). These results suggest that the inserted peptide is responsible for the observed phenotype and not its site of insertion in the CP. The size of the necrotic lesions that developed on Xanthi-nn leaves inoculated with TMV-RB19 was compared to those induced by w.t. TMV or by TMV-RB19 on Xanthi-NN plants (Figure 4 (b)). The local lesions on Xanthi-nn were about 1.5 to two times larger than those that developed on Xanthi-NN plants.

Cell death in BY2 protoplasts that were infected with TMV-RB19 was signi®cantly greater than in protoplasts infected by w.t. TMV or TMV-5B19 (Figure 4(c)). As shown in Figure 4 (d), infection with TMV-RB19 caused cell death starting between 24 and 36 hours post-inoculation (h.p.i.) regardless of whether the protoplasts were held at 24 C or 31 C. At 48 h.p.i. about 65 % to 80 % of cells were not viable. In comparison, protoplasts that were infected with w.t. TMV exhibited viability of about 60 % through 48 h.p.i.

Analysis by electron microscopy of extracts from BY2 cells infected with TMV-RB19, revealed the presence of virus particles similar in appearance to w.t. TMV; in contrast to w.t. TMV, TMV-RB19 was always attached to cell debris (data not shown).

The peptide RB19 has a predicted pI of 8.59, and when inserted in the CP it results in pI:charge value of 5.64: À 1 for CP-RB19. When a Glu residue was inserted between the RB19 peptide and Gly155 of the CP (Figure 1 ), the calculated pI:charge value of the hybrid CP-RB19E was 5.34:-2 (equivalent to the w.t. CP; Table 1 ). The resulting hybrid, referred to as TMV-RB19E, did not induce a hypersensitive response on the inoculated leaves of Xanthi-nn plants (Figure 4(a) ) and led to systemic infection. Likewise, the virus did not increase the rate of cell death in infected BY2 protoplasts (Figure 4 (c) and (d)). Typical mosaic symptoms were produced by TMV-RB19E infection on Xanthi-nn plants, and virus particles were puri®ed from systemically infected leaves.

Protein extracts prepared from infected plants and CP from puri®ed virus were separated on SDS-PAGE and analyzed by Coomassie blue staining and Western immunoblot assays, using anti-TMV antibody or the monoclonal anti-5B19 antibody (MAb5B19) (Figure 5(a) ). CP-RB19E hybrid protein of the predicted molecular mass (18.7 kDa) was present at high levels in infected plants and in assembled virus ( Figure 5(a) ). Electron microscopy following immunogold labeling using MAb5B19 showed speci®c labeling of TMV-RB19E, whereas no labeling was obtained with w.t. TMV ( Figure 5(b) ). In addition, treating puri®ed TMV-RB19E particles with trypsin released the RB19E peptide from the core of CP (data not shown). These results con®rm that the peptide is exposed on the surface of virus particles.

Analysis of induced cell death in protoplasts and hypersensitive response in Xanthi-nn Brederode et al. (1991) reported that infection by TMV of the necrotic local lesion host N. tabacum Xanthi-NN genotype activates high levels of expression of the acidic PR2 gene both locally and systemically, and low levels of expression of the basic PR2 gene locally, but not systemically. We investigated whether or not the PR2 genes are activated in BY2 protoplasts and in N. tabacum Xanthinn and Xanthi-NN plants following infection by w.t. TMV or TMV mutants. RNA dot blot analyses were performed using total RNA puri®ed from protoplasts (at 24 and 36 h.p.i.), Xanthi-nn (4 d.p.i.), and Xanthi-NN (4 d.p.i.) plants inoculated with TMV-RB19, TMV-RB19E or w.t. TMV-U1. 32 Plabeled DNA molecules speci®c to acidic or to basic PR2 gene were used as probes. There was little or no induction of acidic or basic PR2 mRNA in BY2 protoplasts infected with w.t. TMV or with hybrid viruses (data not shown). In contrast, TMV-RB19 induced accumulation of acidic and basic PR2 gene expression in inoculated leaves of N. tabacum Xanthi-nn and Xanthi-NN ( Figure 6 ). In Xanthi-NN plants inoculated with TMV-RB19, basic PR2 gene expression was about two times greater than that on TMV inoculated Xanthi-NN plants, whereas the accumulation of acidic PR2 mRNA was about the same in both plants ( Figure 6 ). On inoculated leaves (IL) of N. tabacum Xanthi-nn plants, high levels of mRNAs of acidic and basic PR2 gene expression was observed after inoculation with TMV-RB19, but not by w.t. TMV or TMV-RB19E ( Figure 6 and data not shown). Basic PR2 gene expression was induced in inoculated leaves, but not in non-inoculated upper leaves (SL; Figure 6 (a)). However, basic PR2 mRNA levels in Xanthi-nn inoculated with TMV-RB19 was about 20 times higher than in Xanthi-nn leaves inoculated with w.t. TMV, and about three to ®ve times higher than in Xanthi-NN leaves inoculated with TMV or TMV-RB19 (Figure 6(a) ). In contrast, the accumulation of acidic PR2 mRNA in Xanthi-nn inoculated with TMV-RB19 was similar to that observed on Xanthi-NN inoculated with TMV or TMV-RB19 (Figure 6(b) ). These studies demonstrate that the hypersensitive response on Xanthi-nn plants resulting from TMV-RB19 infection triggered the expression of the acidic and basic PR2 genes. 

Expression of immunogenic peptides as fusion proteins with plant virus capsid proteins can lead to production of large amounts of the peptide. TMV CP is a good carrier because it self-assembles in an ordered manner, and accumulates to large amounts in infected tissues. Furthermore, as the atomic structure of TMV is known, one can design proteins that display foreign peptides on the surface of the virion. CP molecules have presumably evolved a variety of functions (e.g. ef®cient assembly), and in some cases, play an important role in virus movement and/or replication (Matthews, 1991) . Mutations in viral CP can cause loss of certain CP functions (Bendahmane et al., 1997) . Similarly, display of peptides on the surface of TMV as fusion to the CP can reduce local and systemic virion accumulation in infected plants, and can induce the expression of host defense genes (Beachy et al., 1996; this study) .

In this study peptides from two different viruses, rabies virus (a lyssavirus) and murine hepatitis virus (a coronavirus) were produced as fusion proteins with TMV CP. The 23 amino acid residues of peptide G5-24 from the RV glycoprotein, and the ten amino acid residue peptide containing the 5B19 epitope from MHV spike protein, were successfully exposed on the surface of TMV. The respective viruses, TMV-G5.24 and TMV-5B19, systemically infected N. tabacum Xanthi-nn. However, insertion of the peptide 31D from the RV nucleoprotein in TMV CP (TMV-31D) reduced the infectivity of the hybrid virus TMV-31D in plants. In BY2 protoplasts TMV-31D replicated and CP-31D accumulated to high levels (Figure 3(b) ). The results indicate that the lack of virus accumulation in infected plant leaves may result from interference with local movement. This important ®nding remains unexplained, since it has been shown that local cell to cell spread of TMV does not require CP. A possible explanation for these results is that leaf cells that are infected are killed before local spread can occur.

TMV-RB19 contained an Arg plus the peptide RB19 (from the MHV S-glycoprotein) inserted in CP. Infection by this virus produced necrotic lesions on N. tabacum Xanthi-NN similar to those produced by w.t. TMV. However, in contrast to w.t. TMV, TMV-RB19 induced a hypersensitive response (HR) on the systemic host N. tabacum Xanthi-nn. In this case the addition of a single Comparison of the hybrid CPs that did not cause necrosis on Xanthi-nn plants with those that did, led us to propose that the overall charge of the CP is important in inducing necrosis. The peptides that, when inserted in the CP, resulted in a pI equal to or lower than the w.t. CP (e.g. G5-24, 5B19, and 5B19L; Figure 1 (a); Table 1 ), were successfully displayed on TMV particles, and the chimeric viruses caused systemic infections. In contrast, peptides that resulted in a pI of the CP that is higher than that of w.t. CP resulted in the induction of HR (Table 1) . Altering the pI:charge value in hybrid CP to bring it near to that of w.t. CP, 5.31: À 2 (by inserting Asp to Glu residues before or after the peptide) eliminated the necrotic reaction and resulted in local and systemic infection (Table 1) . These results con®rmed the importance of pI:charge value of the hybrid protein, and show that highly positively charged peptides can in¯uence the interaction of TMV with the host. In contrast, negatively charged epitopes are more tolerated than are positively charged epitopes (e.g. G5-24 and 5B19L peptides; Table 1 ). These principles were recently applied to successfully generate new hybrid TMV constructs that express the MG3 epitope from the lymphocyte choriomeningitis virus ( TMV induces local necrotic lesions on N. tabacum Xanthi-NN as a result of the N resistance gene (Whitham et al., 1994) and an uncharacterized interaction with TMV replicase (Padgett & Beachy, 1993; Padgett et al., 1997; Abbink et al., 1998) . Although N. tabacum Xanthi-nn lacks the N gene, the hybrid virus TMV-RB19 induced HR on these plants. The mechanism of induction of HR by TMV-RB19 on Xanthi-nn plants is not known. However, the signal transduction pathway that results in HR on Xanthi-NN and Xanthi-nn may, at least in a part, be different. This is supported by the observation that TMV-RB19 induces cell death in infected BY2 protoplasts whereas w.t. TMV does not. Furthermore, protoplasts derived from Xanthi-NN plants do not die when infected with TMV (Otsuki et al., 1972) . It is known that infection by TMV induces the expression of the basic PR2 gene in the inoculated leaves and the expression of the acidic PR2 gene both locally and systemically in Xanthi-NN (Brederode et al., 1991; Heitz et al., 1994) , but not in Xanthi-nn plants that lack the N gene (Linthorst, 1991) . In contrast to Xanthi-nn plants inoculated with w.t. TMV, Xanthi-nn plants that produce necrotic lesions following inoculation with TMV-RB19 resulted in the expression of the basic and the acidic PR2 genes. These results indicate that the HR observed in the Xanthi-nn plants inoculated with TMV-RB19 may share a portion of the HR signal transduction pathway with the HR in Xanthi-NN.

It is known that the N H resistance gene in Nicotiana sylvestris does not respond to infection by w.t. TMV. However, certain CP mutants of TMV that are structurally altered to affect its ordered quaternary structure elicit the expression of the N H gene (Knorr & Dawson, 1988; Culver et al., 1994) . It was suggested that structures created by the mutations interact with an uncharacterized host receptor to elicit the HR (Culver et al., 1994) . The HR that results from infection by TMV-RB19 on N. tabacum Xanthi-nn may be similar to, or different from, the N H elicited HR; additional studies are needed to address this question.

The question remains as to why BY2 protoplasts die as a result of TMV-RB19 infection. An answer Figure 6 . Accumulation of (a) basic and (b) acidic PR2 mRNA in N. tabacum Xanthi-nn and Xanthi-NN plants following inoculation with TMV-U1 or TMV-RB19 at 4 d.p.i. RNA dot blots were performed using similar amounts of total RNA and then hybridized with sequences to detect (a) basic or (b) acidic PR2 gene expression. The signal intensity of the isotope label was measured using phosphoimager followed by Multi-Analyst computer program analyses (Bio-Rad, Hercules, CA), and the relative values are displayed as graphs. may reside in the fact that TMV-RB19 and TMV-5B19 assemble to form virus particles which are very insoluble in plant cells. Indeed, electron microscopy experiments showed that virus particles were recovered from membrane-rich fractions rather than from a soluble supernatant fraction. This may indicate that insertion of the RB19 or 5B19 peptides at the C terminus of TMV CP creates a domain on the surface of assembled viruses that interacts with and destabilizes cellular membranes, thus causing BY2 cell death. Porta et al. (1994) studied the expression of an epitope from foot-andmouth disease virus (FMDV) on the surface of CPMV and found an unusual af®nity of the hybrid CPMV-FMDV with cellular membranes. Porta et al. (1994) proposed a direct involvement of three amino acid residues (Arg-Gly-Asp) of the FMDV epitope in af®nity with the membrane. The predicted hydrophobicity of the CP-5B19 and CP-RB19 compared to w.t. CP suggests that both proteins may create hydrophobic domains at the C terminus of the hybrid CP that is not present in w.t. CP. This domain may interact with and anchor the CP and/or the virus with cell membranes.

Our data support the conclusion that the pI of the peptide-CP fusion protein in¯uences the interaction of the virus with the host. The TMV CP tolerates negatively charged peptides, but is affected by positively charged epitopes. The data presented here also demonstrate that these effects are reversible, and that bringing the pI:charge value closer to that of w.t. CP reduces or eliminates the ability of the fusion protein to cause cell death in infected plants and protoplasts.

Construction of TMV-5B19, TMV-5B19L and TMV-RB19

TMV-5B19 and TMV-5B19L containing the nucleotide sequences coding for the peptides containing the 5B19 epitope ( 900 LLGCIGSTCA 909 ) or the epitope 5B19 plus ®ve neighboring amino acid residues (5B19L: 899 PLLGCIGSTCAEDGN 913 ) from murine hepatitis virus (MHV, coronavirus) S-glycoprotein are described by Koo et al., (1999) . TMV-RB19 harbors the peptide containing an arginine residue plus 5B19 fused between amino acid residues Ser154 and Gly155 of the TMV coat protein (Figure 1 ; Table 1 ), and was constructed using the following primers: 5 H tAGA TTA TTA GGA TGT ATA GGA TCT ACT TGT GC3 H and its complement 3 H gaTCT AAT AAT CCT ATA TAT CCT AGA TGA ACA C5 H . The two primers containing the nucleotide-coding sequence of the RB19 peptide were mixed, heated and allowed to anneal. The DNA fragment formed by the annealed primers contains AhdI compatible ends that was used for cloning into the TMV cDNA clone linearized with AhdI. The resulting clone TMV-RB19 contains an arginine residue followed by the nine amino acid residues of the 5B19 epitope between Ser154 and Gly155 of the TMV CP.

The primers 5 H CT TGT GCT GAA GGT CCT GCA ACT TGA GG3 H and its complement 5 H CC TTC AGC ACA AGT AGA TCC TAT AC3 H were used in PCR reactions to insert a glutamic acid residue between Ala165 and Gly167 of the hybrid CP:RB19 in the TMV-RB19 cDNA clone. The resulting construct is referred to as TMV-RB19E ( Figure 1 ; Table 1 ).

The nucleotide sequence coding for the peptide 31D (Dietzschold et al., 1990) of the rabies virus nucleocapsid protein ( 404 AVYTRIMMNGGRLKR 418 ) or for the peptide G5-24 (Dietzschold et al., 1990) from the RV glycoprotein ( 253 PPDQLVNLHDFRSDEIEHLVVEE 275 ) were fused between Ser154 and Gly155 of the TMV CP in the clone pKN2 (Bendahmane et al., 1997) via a PCR-based reaction using the following primers: 5 H CT CGA ATC ATG ATG AAT GGA GGT CGA CTA AAG AGA ggt cct gca act tga ggt agt c3 H and 5 H CG ACC TCC ATT CAT CAT GAT TCG AGT ATA AAC AGC aga ggt cca aac caa acc aga ag3 H , or 5 H CTG CAC GAC TTT CGC TCA GAC GAA ATT GAG CAC CTT GTT GTA GAG GAG ggt cct gca act tga ggt agt c3 H and 5 H AAT TTC GTC TGA GCG AAA GTC GTG CAG GTT CAC CAA CTG ATC GGG AGG aga ggt cca aac caa acc aga ag3 H , respectively. Sequences in upper case represent the epitope sequence, while those in lower case represent the sequence of the TMV CP. The resulting constructs pCP:31D and pCP:G5.24 contain the 31D or G5-24 epitopes, respectively, inserted between Ser154 and Gly155 in the TMV CP ( Figure 1 ; Table 1 ).

The hybrids pCP-2D:31D and pCP-4D:31D were constructed by inserting the dipeptide AspAsp or the pentapeptide AspAspGlyAspAsp between Ser154 and Ala155 of the 31D peptide in the pCP:31D construct via PCRbased reactions using the following primers: 5 H ggt ttg gtt tgg act agt GAC GAC GCT GTT TAT ACT CG 3 H or 5 H ct agt GAC GAC GGT GAC GAC GCT GTT TAT ACT CGA ATC3 H and 5 H GTC GTC ACC GTC GTC act agt cca aac caa acc aga aga g3 H , respectively. Sequences in upper case represent the epitope sequence, those underlined represent the AspAsp or AspAspGlyAspAsp, and sequences in lower case represent the sequence of the TMV CP.

The pCP-2D:31D construct was engineered to create a new vector with the unique restriction sites SpeI and PpuMI to facilitate the cloning of any nucleotide sequence between amino acid residues Ser154 and Gly155 of the TMV CP (Figure 1(c) ). The vector is termed pCPs/p.

The fragment NcoI-KpnI from the plasmids pCP:31D, pCP-2D:31D, pCP-4D:31D and pCP:G5.24 carrying hybrid CP sequences were used to replace the homologue in the TMV cDNA clone U3/12-4 (Holt & Beachy, 1991) linearized with NcoI plus KpnI to generate infectious full-length cDNA clones referred to as pTMV-31D, pTMV-2D:31D, pTMV-4D:31D and pTMV-G5.24, respectively.

In vitro transcription reactions were used to produce full-length viral RNA from 1 mg of plasmids pTMV-31D, pTMV-2D:31D, TMV-4D:31D, pTMV-G5.24, pTMV-5B19, Expression of Epitopes on the Surface of TMV pTMV-RB19, pTMV-RB19E or U3/12-4 (wild-type TMV clone) previously linearized with KpnI as described (Holt & Beachy, 1991) . Leaves of four-week-old N. tabacum Xanthi-NN and Xanthi-nn plants were ®rst dusted with carborundum (320 Grit, Fisher Scienti®c) and then inoculated with the in vitro transcription reaction products.

Wild-type TMV and the hybrid TMV-G5.24 were puri-®ed from systemically infected leaves of N. tabacum Xanthi-nn plants as described (Asselin & Zaitlin, 1978) . To purify hybrid TMV-RB19E, infected plant leaves were ground in liquid nitrogen and homogenized in 20 mM NaHPO 4 buffer (pH 7). Cell debris was removed by centrifugation at 3000 g. 5 % (w/w) diatomaceous earth (Sigma) was mixed with the supernatant and centrifuged at 10,000 g for 30 minutes. The supernatant was then applied to a cushion of 20 % sucrose in 20 mM Tris-HCl (pH 7). After centrifugation at 140,000 g for one hour 30 minutes, the pellet was dissolved in 20 mM NaHPO 4 buffer (pH 7), and sonicated for one minute. Triton X-100 (5 %, v/v) was then added to the virus solution and the suspension was applied again on a cushion of 20 % sucrose in 20 mM Tris-HCl (pH 7), as above, and subjected to centrifugation. The pellet containing the virus was dissolved in 20 mM NaHPO 4 (pH 7).

Puri®ed virus at concentrations of 10 to 100 mg/ml was applied to carbon-coated copper EM grids. The grids were then blocked with 0.1 % bovine serum albumin (BSA) solution in 1 Â PBS for 15 minutes. The monoclonal antibodies anti-5B19 (MAb5B19; Buchmeier et al., 1984) for TMV-RB19E or anti-G5.24 (MAb5; Dietzschold et al., 1990) for TMV-G5.24 (diluted 1/50 in 1 Â PBS containing 0.1 % BSA) were added for one hour, followed by ®ve washes with 50 ml of 1 Â PBS. Goat anti-mouse antibody conjugated with 5 nm gold particles (Ted Pella, Redding, CA) were added to the grid at room temperature for one hour. To remove non-speci®c binding, six washes with 50 ml of 1 Â PBS followed by four washes with water were applied. The grids were then negatively stained with 1 % phosphotungstic acid solution (pH 5.6), for two minutes. Grids were dried and examined on a Phillips CM100 electron microscope at 39,000Â magni®cation.

Tobacco protoplasts were prepared from BY2 suspension cell cultures; 2 Â 10 6 protoplasts were inoculated with RNA transcribed in vitro from above cDNA clones of w.t. or hybrid viruses via electroporation as described by Watanabe et al. (1987) . Protoplasts were harvested at different times post-inoculation and used for further analyses.

Suspensions of protoplasts were stained for live cells using¯uorescence diacetate (FDA) staining solution (0.1 ml of 0.05 % FDA stock solution plus 5 ml of protoplast culture medium). One volume FDA staining solution was mixed with one volume of suspended protoplasts in a hemocytometer. After ®ve minutes, live cells were visualized and counted with a Nikon Opti-phot2-UD microscope equipped with a Nikon B-2A ®lter set consisting of a 450-490 excitation ®lter as described by Kahn et al. (1998) .

Total protein extracts from leaves or from protoplasts were homogenized in 1 Â Laemmli loading buffer (Laemmli, 1970) , boiled for four minutes and applied to 12.5 % polyacrylamide gels containing SDS (SDS-PAGE). Proteins were either directly stained with Coomassie blue or analyzed by Western immunoblot using as ®rst antibody the polyclonal rabbit anti-TMV, the monoclonal MAb5B19 mouse anti-5B19 peptide or the monoclonal MAb5 anti-G5-24 peptide using standard methods (Sambrook et al., 1989) . Goat anti-rabbit or goat anti-mouse antibodies conjugated with horse radish peroxidase or alkaline phosphatase were used as secondary antibodies and the signal was revealed using the chemiluminescence ECL-kit (Amersham) according to the manufacturer's recommendation and as described by Sambrook et al. (1989) .

Total RNAs were puri®ed from BY2 protoplasts at 24 and 36 h.p.i. with TMV, TMV-RB19 or TMV-RB19E using the standard guanidine thiocyanate extraction method (Sambrook et al., 1989) . Total RNA was puri®ed from N. tabacum Xanthi-nn and Xanthi-NN plants four days post-inoculation with TMV-RB19 or TMV-U1. Leaf discs (1 cm in diameter) containing necrotic lesions were collected and total RNA was puri®ed as follows. Leaf discs were homogenized in the extraction buffer (0.2 M Tris-HCl (pH 7.5), 0.1 M LiCl, 5 mM EDTA, 1 % (w/v) SDS). The homogenate was extracted with phenol/ chloroform/isoamyl alcohol and RNA were then precipitated in the presence of 3 M LiCl. The pelletted RNA was resuspended in H 2 O and then precipitated in the presence of 2.5 volumes of ethanol. RNA dot blots were performed by applying similar amounts of total RNA onto a nylon membrane; blotted RNAs were hybridized with sequences to detect acidic or basic PR2 genes expression using speci®c PR2 DNA sequences as described by Karrer et al. (1998) . After stringent washes the intensity of the radioactive label was measured using a phosphoimager followed by Multi-Analyst computer program analyses according to the manufacturers recommendations (Bio-Rad, Hercules, CA).

Tobacco mosaic virus helicase domain induces necrosis in N gene-carrying tobacco in the absence of virus replication

Characterization of a second protein associated with virions of tobacco mosaic virus

Use of plant viruses for delivery of vaccine epitopes

Studies of coat protein mediated resistance to TMV: correlation between assembly of mutant coat proteins and resistance

Differential induction of acquired resistance and PR gene expression in tobacco by infection, ethephon treatment, UV light and wounding

Murine hepatitis virus-4 (starin JMH)-induced neurologic disease is modulated in vivo by monoclonal antibody

Antigen chimeras of poliovirus as a potential new vaccines

Structure-function relationship between tobacco mosaic virus coat protein and hypersensitivity in Nicotiana sylvestris

Site-directed mutagenesis con®rms the involvement of carboxylate groups in the disassembly of tobacco mosaic virus

Modi®cations of the tobacco mosaic virus coat protein gene affecting replication, movement, and symptomatology

Poliovirus chimeras expressing sequences from the principal neutralization domain of human immunode®ciency virus type 1

Structural and immunological characterization of a linear virus-neutralizing epitope of rabies virus glycoprotein and its possible use in a synthetic vaccine

States of aggregation of tobacco mosaic virus protein

Plant virus expressing hybrid coat protein with added murine epitope elicit autoantibody response

Nucleotide sequence of tobacco mosaic virus RNA

A new tobacco mosaic virus vector and its use for the systemic production of natigensin-I-converting enzyme inhibitor in transgenic tobacco and tomato

Development of a genetically-engineered, candidate polio vaccine employing the self-assembling properties of the tobacco mosaic virus coat protein

Local and systemic accumulation of pathogenesis-related proteins in tobacco plants infected with tobacco mosaic virus

In vivo complementation of infectious transcripts from mutant tobacco mosaic virus cDNAs in transgenic plants

High level of production of hybrid potyvirus-like particles carrying repetitive copies of foreign antigens in Escherichia coli

Domains of the TMV movement protein involved in subcellular localization

Cloning of tobacco genes that elicit the hypersensitive response

Protective immunity against murine hepatitis virus induced by intranasal or subcutaneous administration of hybrids of tobacco mosaic virus that carries an MHV epitope

A point mutation in the tobacco mosaic virus capsid protein gene induces hypersensitivity in Nicotiana sylvestris

Cleavage of structural proteins during the assembly of the head of bacteriophage T4

Pathogenesis-related proteins of plants

Use of macromolecular assemblies as expression systems for peptides and synthetic vaccines

Amino acid sequence of a conserved neutralizing epitope of murine coronaviruses

Plant Virology

Visualization of protein-nucleic acid interactions in a virus. Re®ned structure of intact tobacco mosaic virus at 2.9 A Ê resolution X-ray ®ber diffraction

Tobacco mosaic virus multiplication and expression of the N gene in necrotic responding tobacco varieties

Analysis of a tobacco mosaic virus strain capable of overcoming N gene-mediated resistance

Identi®cation of the TMV replicase sequence that activates the N gene-mediated hypersensitive response

Development of cowpea mosaic virus as a height-yielding system for the presentation of foreign peptides

Molecular Cloning -A Laboratory Manual

Systemic production of foreign peptides on the particle of tobacco mosaic virus

Production of enkephalin in tobacco protoplasts using tobacco mosaic virus RNA vector

Malarial epitopes expressed on the surface of recombinant tobacco mosaic virus

Expression of an animal virus antigenic site on the surface of a plant virus particle

Infection of tobacco protoplasts with in vitro transcribed tobacco mosaic virus RNA using an improved electroporation method

The product of the tobacco mosaic virus resistance gene N: similarity to Toll and the interleukin-1 receptor