The helicase domain of the TMV replicase proteins induces the N-mediated defence response in tobacco

Authors

  • F. Les Erickson,

    1. Plant Gene Expression Center, University of California, Berkeley and United States Department of Agriculture, Agriculture Research Service, Albany, California 94710, USA,
    2. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author
  • Steve Holzberg,

    1. Department of Environmental Science, Policy, and Management, University of California, Berkeley, California 94720, USA, and
    Search for more papers by this author
  • Alejandro Calderon-Urrea,

    1. Plant Gene Expression Center, University of California, Berkeley and United States Department of Agriculture, Agriculture Research Service, Albany, California 94710, USA,
    2. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author
    • Present address: Department of Biology, California State University, Fresno, California 93740, USA.

  • Vanessa Handley,

    1. Plant Gene Expression Center, University of California, Berkeley and United States Department of Agriculture, Agriculture Research Service, Albany, California 94710, USA,
    2. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author
  • Michael Axtell,

    1. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author
  • Catherine Corr,

    1. Plant Gene Expression Center, University of California, Berkeley and United States Department of Agriculture, Agriculture Research Service, Albany, California 94710, USA,
    2. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author
  • Barbara Baker

    1. Plant Gene Expression Center, University of California, Berkeley and United States Department of Agriculture, Agriculture Research Service, Albany, California 94710, USA,
    2. Department of Plant and Microbial Biology, University of California, Berkeley, California 94720, USA
    Search for more papers by this author

*For correspondence (fax +510 559 5678;
e-mail bbaker@socrates.berkeley.edu).

Summary

Tobacco mosaic virus (TMV) induces the hypersensitive response (HR) in tobacco plants containing the N gene. This defence response is characterized by cell death at the site of virus infection and inhibition of viral replication and movement. A previous study indicated that a portion of the TMV replicase containing a putative helicase domain is involved in HR induction. Here, this observation is confirmed and extended by showing that non-viral expression of a 50 kDa TMV helicase fragment (p50) is sufficient to induce the N-mediated HR in tobacco. Like the HR elicited by TMV infection, transgenic expression of p50 induces a temperature-sensitive defence response. We demonstrate that recombinant p50 protein has ATPase activity, as suggested by the presence of conserved sequence motifs found in ATPase/helicase enzymes. A point mutation that alters one of these motifs abolishes ATPase activity in vitro but does not affect HR induction. These results suggest that features of the TMV helicase domain, independent of its enzymatic activity, are recognized by N-containing tobacco to induce TMV resistance.

Introduction

The interaction between tobacco mosaic virus (TMV) and tobacco plants bearing the N gene is a classical system for the study of plant disease resistance ( Holmes 1938). The N gene from Nicotiana glutinosa (tobacco) confers resistance to TMV and all other known tobamoviruses except strain Ob ( Culver et al. 1991; Dawson 1992; Tobias et al. 1982). In tobacco cultivars lacking the N gene, TMV replicates and moves systemically, causing reduced growth and mosaic disease symptoms, characterized by intermingled areas of light and dark green leaf tissue. In contrast, TMV infection of N-containing tobacco produces cell death at the site of virus infection, and virus particles are restricted to the region immediately surrounding the necrotic lesions. The formation of necrotic lesions upon pathogen infection is known as the hypersensitive response (HR). The HR apparently occurs only when a specific resistance (R) gene is present in the plant and a corresponding avirulence (Avr) gene is present in the pathogen. Similar ‘gene-for–gene’ interactions ( Flor 1947; Flor 1971) have been shown to operate in resistance to viral, bacterial, fungal, nematode, and insect pathogens of plants ( Baker et al. 1997; Rossi et al. 1998).

How TMV infection induces the tobacco defence response is not clear. TMV has a positive-sense, RNA genome that encodes at least four proteins: a coat protein, a movement protein, and two replicase proteins. A 183 kDa replicase protein is produced by read-through of the stop codon that terminates a 126 kDa replicase protein ( Dawson 1992; Goelet et al. 1982). It has been shown that neither the TMV coat protein nor the movement protein is responsible for induction of the N-mediated HR ( Deom et al. 1991; Padgett et al. 1997). Two different studies have implicated the tobamovirus replicase proteins in HR induction. The first study identified HR-inducing mutants of the normally virulent Ob virus. These Ob mutants were found to harbour mutations altering amino acids in the helicase region of the replicase proteins ( Padgett & Beachy 1993). The second study demonstrated that a TMV replicase fragment containing the helicase region was required for HR induction by chimeric Ob/TMV viruses ( Padgett et al. 1997). Both of these studies implicated a region of the replicase proteins that may function as a helicase, as suggested by the presence of six sequence motifs found in other ATPase/helicase enzymes ( Gorbalenya & Koonin 1993). Helicases bind and hydrolyze nucleoside triphosphates to promote unwinding of duplex nucleic acids. The role of this proposed domain in HR induction is unknown; furthermore, the use of replicating tobamoviruses in these studies argues that the participation of additional viral components or processes in HR induction cannot be excluded.

To gain a better understanding of HR induction by TMV, we employed non-viral expression methods to determine the minimal TMV component required for inducing the N-mediated defence response in tobacco. Here, we report that either transient or stable transgenic expression of a 50 kDa replicase fragment (p50) that contains the putative helicase domain is sufficient to induce the HR. Biochemical and mutagenesis analyses indicated that this fragment has enzymatic activity consistent with helicase function, but this activity is not required for induction of the defence response.

Results

Transient expression of the p50 replicase fragment induces the N-mediated HR

To determine the minimal TMV component required for inducing HR in tobacco, we employed a transient expression method based on T-DNA transfer by Agrobacterium tumefaciens ( Van den Ackerveken et al. 1996 ). TMV cDNA sequences were inserted into the binary vector pMB under the control of a modified cauliflower mosaic virus 35S promoter and the 5′ untranslated region of TMV-U1. Viral coding sequences were derived from a TMV-U1 cDNA clone ( Dawson et al. 1986 ) and were modified, when needed, to contain translational start and stop codons. Agrobacterium strain C58C1::pGV2260 harbouring these expression plasmids or control plasmids were infiltrated into SR1 tobacco leaves that either lacked (SR1) or contained (SR1::NN) the N gene, allowing for vector transfer and transient expression. The infiltrated tissues were examined over a 2-week period and scored for the presence or absence of necrotic lesions, which indicate HR.

As a positive control for HR, a binary vector that contains AvrPto from Pseudomonas syringae ( Ronald et al. 1992 ) was employed because AvrPto induces an HR when transiently expressed in tobacco (Brian J. Staskawicz, personal communication). As expected, transient expression of AvrPto produced necrotic lesions in both SR1 and SR1::NN tobacco leaves ( Fig. 1c) within 2 days after infiltration. Transient expression of GUS, which encodes β-glucuronidase, failed to produce necrotic lesions in either tobacco line, indicating that neither Agrobacterium nor the infiltration process causes cell death in this assay ( Fig. 1c,d).

Figure 1.

Transient expression of TMV genes in tobacco.

(a) Schematic diagram depicting the 126 kDa and 183 kDa forms of the TMV replicase. Asterisks denote location of conserved helicase motifs.

(b) Control genes and TMV-U1 genes or gene fragments transiently expressed in transgenic SR1 tobacco that either contains (SR1::NN) or lacks (SR1) the N gene. TMV genes or gene fragments were inserted into the pMB binary vector under the control of a double CaMV 35S promotor and the TMV-U1 5′ untranslated region; refer to Experimental procedures for specific TMV nucleotides present in each replicase fragment. Results from Agrobacterium-mediated transient expression are shown to the right of each gene construct. (+) indicates HR was produced in over 95% of infiltrated tobacco leaves within 2 weeks after infiltration; (–) indicates no HR was observed in the same time period. At least 25 infiltrations were performed for each construct.

(c) and (d) Representative examples of Agrobacterium-mediated expression of HR in SR1::NN (c) and SR1 (d) tobacco leaves. Agrobacterium harbouring a binary vector containing the indicated gene construct was pressure-infiltrated into a 2–4 cm2 area of leaf tissue. Photographs were taken 10 days after infiltration.

(e) Protein gel blot analysis of soluble proteins extracted from SR1 tissue transiently expressing the indicated gene constructs. Proteins were separated by SDS–PAGE (10%), transferred to a nitrocellulose membrane, probed with polyclonal anti-FLAG antibodies and developed using SuperSignal ULTRA (Peirce). A 50-kDa protein was detected in lanes 3 and 6. No proteins were detected in lanes 2, 4 and 5. Similar results were obtained in three independent experiments. In lane 1, 2 ng of a FLAG-tagged, 56 kDa replicase fragment produced in and purified from insect cells was used as a size marker and a positive control for immunoblotting.

Transient expression of the replicase gene (RP, Fig. 1a,b) produced necrotic lesions in SR1::NN ( Fig. 1c), but not SR1 tobacco ( Fig. 1d), within 7–12 days after Agrobacterium infiltration, indicating an N-dependent HR. The TMV movement-protein gene (MP) did not produce necrotic lesions when transiently expressed in either SR1 or SR1::NN tobacco leaves ( Fig. 1c,d). Introduction of a frameshifting mutation 190 nucleotides downstream from the replicase start-codon in RP abolished the ability to produce necrotic lesions (data not shown), indicating that the encoded replicase proteins, but not their RNA, are necessary to induce HR. To further define the sequences within RP responsible for HR induction, a gene fragment encoding the 126 kDa replicase was divided into fragments that encode an 80 kDa N-portion terminal (p80, Fig. 1b) and a 50 kDa C-terminal portion (p50, Fig. 1b). p50 is essentially the same TMV gene fragment identified previously as required in chimeric Ob/TMV viruses for induction of HR ( Padgett et al. 1997 ). Transient expression of p50 produced necrotic lesions in SR1:NN ( Fig. 1c) but not SR1 leaves ( Fig. 1d) 2–3 days after infiltration, while transient expression of p80 did not produce lesions in leaves from either tobacco line ( Fig. 1c,d). No necrotic lesions were produced by either RP or p50 when Agrobacterium strain C58C1, which lacks the vir genes required for T-DNA transfer into plant cells, was used in the infiltration protocol (data not shown), indicating HR is induced by TMV protein produced in the plant cell. Taken together, these experiments showed that transient, ectopic expression of a 50 kDa TMV replicase fragment is sufficient to elicit an N-mediated HR in tobacco.

To further delimit the TMV sequences required for HR induction, truncations of p50 were constructed. p43 and p46 encode replicase fragments representing an N-terminal deletion of 61 amino acids (7 kDa) and a C-terminal deletion of 34 amino acids (4 kDa) of the 50 kDa replicase fragment, respectively ( Fig. 1b). Transient expression of either p43 or p46 failed to produce necrotic lesions in SR1 and SR1::NN plants (infiltrated leaves not shown). To ascertain if p50, p43, and p46 proteins accumulated in the plant, we performed immunoblot analysis of soluble proteins extracted from SR1 tissue transiently expressing FLAG-epitope tagged versions of these alleles (p50-FLAG, p43-FLAG, and p46-FLAG). The C-terminal tag did not alter the phenotype conferred by non-tagged versions of these alleles (data not shown). Figure 1(e) depicts a protein gel blot showing a 50 kDa protein in proteins extracted from tissue expressing p50-FLAG, while no proteins were observed in extracts derived from tissue expressing either p43-FLAG or p46-FLAG. These results imply p43-FLAG and p46-FLAG mRNA transcripts or their encoded proteins are unstable, which prevents making any conclusions regarding their ability to elicit HR.

Transgenic expression of the p50 replicase gene fragment induces a thermosensitive HR

To corroborate the results obtained using Agrobacterium-mediated transient expression, we determined whether stable, transgenic expression of p50 could induce HR in transgenic plants expressing the N gene. The pMB vector containing p50-FLAG was transformed into SR1 tobacco, and four independent transformants (SR1::p50) were isolated. Immunoblot analysis established that a 50 kDa FLAG-tagged protein was present at comparable levels in each of these lines ( Fig. 2a and data not shown). These SR1::p50 lines were crossed with transgenic SR1 plants homozygous for the N gene (SR1::NN), and F1 seedlings containing both p50 and N (SR1::p50/N) were selected and analyzed for the presence of necrotic lesions. F1 SR1::p50/N progeny from all crosses exhibited systemic necrotic lesions within the first week after germination at 22°C. The necrosis spread and killed the seedlings within 2 weeks after germination. An example of this phenotype is shown in Fig. 2(b). This lethality was N-dependent, because control seedlings, derived from SR1::p50 self-crosses, lacked necrosis and grew normally ( Fig. 2b). Thus, stable, transgenic expression of p50 is sufficient to elicit the N-mediated HR.

Figure 2.

Stable expression of p50 in tobacco.

(a) Protein gel blot analysis showing the accumulation of a 50 kDa protein in tobacco stably expressing p50-FLAG. Soluble proteins extracted from transgenic SR1::p50-FLAG (line 34–3) or SR1 leaf tissue were separated by SDS–PAGE (10%), transferred to nitrocellulose membrane, probed with polyclonal anti-FLAG antibodies and developed using SuperSignal ULTRA (Peirce). In lane 1, 2 ng of a FLAG-tagged, 56 kDa replicase fragment produced in and purified from insect cells was used as a size marker and a positive control for immunoblotting.

(b) Phenotype of transgenic p50-FLAG tobacco (line 34–3) either containing (SR1::p50/N) or lacking (SR1::p50) the N gene. Seed were germinated and grown on agar at the indicated temperatures. Plants involved in the temperature shift experiment (labelled 32°C→22°C) were germinated and grown at 32°C for 5 weeks prior to shifting to 22°C. Photographs were taken at 2 (22°C) and 7 weeks (32°C and 32°C→22°C) post germination.

We next determined if the HR induced in transgenic p50/N tobacco is similar to that induced by TMV. The HR elicited by TMV infection has been shown to be temperature sensitive, occurring only at temperatures below 28°C ( Samuel 1931). At temperatures above 28°C, HR is suppressed, and TMV spreads systemically. The HR is restored when the temperature is reduced below 28°C, and N-mediated cell death occurs throughout the plant, presumably due to systemic TMV infection. To determine if the HR induced by transgenic expression of p50 is temperature dependent, we monitored SR1::p50 seedlings that either contained or lacked N for necrotic lesions at 32°C. Unlike the lethality observed at 22°C ( Fig. 2b), SR1::p50/N seedlings lacked necrotic lesions when germinated at 32°C ( Fig. 2b) and grew normally for at least 2 months (plants not shown). To determine if this thermosensitivity is reversible, SR1::p50/N plants germinated and grown at 32°C for 5 weeks were shifted to 22°C and examined for the formation of HR. By 1 week after the temperature shift, all SR1:p50/N plants had developed necrotic and chlorotic lesions ( Fig. 2b) that spread and collapsed the plants within 2 weeks. Control SR1::p50 seedlings, which lack N, were unaffected by the temperature shift ( Fig. 2b). These observations indicated that the reversible thermosensitivity of the N-mediated response to TMV is also maintained in transgenic plants expressing both N and p50.

The p50 replicase fragment hydrolyzes ATP in vitro, but ATPase activity is not required for HR induction

p50 encodes a portion of the replicase proteins that contains conserved sequence motifs indicative of ATPase/helicase function ( Fig. 1a,b) ( Habili & Symons 1989). We sought to determine whether p50 protein has ATPase activity in vitro, and, if so, is this activity involved in HR induction. To perform in vitro analyses, we produced replicase fragments in E. coli fused to maltose-binding protein (MBP) using the pMAL expression vector (New England Biolabs). cDNA fragments corresponding to p80, p50, p46, and p43 were inserted into this vector to generate p80MBP, p50MBP, p46MBP and p43MBP. These genes were expressed in E. coli, and the encoded MBP-fusion proteins were purified from soluble protein extracts by affinity chromatography using amylose resin. The purified proteins are shown in Fig. 3(a). To test for their ability to release phosphate from adenosine triphosphate (ATPase activity), recombinant proteins were incubated with [γ-32P]-labelled adenosine triphosphate (ATP), followed by thin-layer chromatography to separate the reaction products and visualization of these products using a phosphorimager. p50MBP hydrolyzed ATP to release monophosphate (PO4), as did a commercially available ATPase ( Fig. 3b). p80MBP, p46MBP, and p43MBP lacked ATPase activity ( Fig. 3b). The presence of EDTA inhibited the ATPase activity of p50MBP, which indicated a requirement for Mg2 + , as seen for other viral ATPase/helicase enzymes ( Wonderling et al. 1995 ). It is not clear why p43MBP and p46MBP proteins lacked ATPase activity, since these truncated fragments contain all six ATPase/helicase motifs (see Fig. 1b).

Figure 3.

ATPase activity of TMV replicase fragments.

(a) SDS–PAGE of purified replicase fragments produced in E. coli. The indicated replicase fragments fused to maltose-binding-protein (MBP) were expressed in E. coli and purified from soluble protein extracts using amylose affinity chromatography. Proteins bound to amylose-beads were eluted using SDS sample buffer and boiling. Eluted proteins were separated by 10% SDS–PAGE and visualized using Coomassie brilliant blue

(b) Thin-layer-chromatograph showing ATPase activity of p50. 2–4 μg of purified replicase fragments bound to amylose beads were incubated with [γ-32P]-labelled adenosine triphosphate (ATP) and a portion of each reaction was spotted onto a polyethienine thin-layer chromatograph, which was developed in 0.5 m LiCl/0.5 m phosphoric acid pH 3.0, and visualized using a phosphorimager. Lane 6 represents the activity of a dog liver ATPase obtained from Sigma.

To evaluate if this ATPase activity of p50 is required for inducing HR, site-directed mutagenesis was used to change the codon for threonine 840 to an alanine codon in the p50-FLAG fragment. This conserved threonine is located in a P-loop motif (denoted by red asterisk in p50T/A in Fig. 1b) that functions in ATPase/helicase enzymes to bind the terminal phosphates of the nucleoside triphosphate cofactor ( Saraste et al. 1990 ; Subramanya et al. 1996 ). The mutated p50-FLAG fragment was inserted into pMB and pMAL-c2 to generate p50(T/A)-FLAG for transient expression in tobacco and p50(T/A)MBP for production of fusion protein in E. coli, respectively. Purified p50(T/A)MBP protein ( Fig. 3a) failed to hydrolyze ATP ( Fig. 3b), indicating that this conserved threonine is required for ATPase activity in vitro. Transient expression of p50(T/A)-FLAG in tobacco produced N-dependent necrotic lesions within 2–3 days after infiltration ( Fig. 1c,d), which is comparable in time to the necrosis produced by wild-type p50. Immunoblot analysis of proteins extracted from infiltrated SR1 tissue showed that p50T/A-FLAG protein accumulated in the plant like p50-FLAG protein ( Fig. 1d). These observations suggested that the ATPase activity of the p50 protein is not required for induction of the N-mediated defence response.

Discussion

Ectopic expression of the p50 replicase gene fragment is sufficient to elicit the thermosensitive, N-mediated HR

We have demonstrated that transient expression of a 50 kDa TMV replicase fragment (p50) is sufficient to induce the N-mediated defence response ( Fig. 1). This portion of the TMV replicase was shown previously to be required for HR induction in experiments involving chimeric tobamoviruses ( Padgett et al. 1997 ). Our results extend this observation by showing that other viral proteins or processes, such as viral RNA replication, viral RNA movement or any cellular alterations caused by TMV infection ( Reichel & Beachy 1998) are not required for HR induction by this replicase fragment.

Transient expression of p50 produced necrosis that was clearly visible by 2–3 days after infiltration, which is similar in time that HR lesions appear after inoculation of N-containing tobacco with TMV ( Whitham et al. 1996 ). Transient expression of the complete replicase gene (RP) also induced HR, although the response was much slower, with necrosis first appearing 7–12 days after infiltration. This weaker response may be due to low expression of RP, because the addition of the TMV 3′ UTR sequences downstream of RP in the pMB vector reduced the time for lesion formation to 3–5 days after infiltration (A. Calderon-Urrea and B. Baker, unpublished observations). The presence of the 3′ UTR may increase expression by stabilizing the RP mRNA and/or allowing for replication of the RP mRNA by its encoded replicase proteins.

We demonstrated that stable, transgenic expression of p50 induced the N-mediated HR ( Fig. 2). Transgenic SR1 tobacco stably expressing both p50 and N died of systemic necrosis shortly after germination at 22°C. Germination at 32°C, however, suppressed this lethality, indicating that the temperature sensitive interaction seen between TMV and N also occurs between p50 and N. The reason for this HR inactivation at high temperatures is not known. Previous studies suggested that high temperatures inhibit an early event in the initiation of HR induced by TMV ( Gulyas & Farkas 1978; Samuel 1931; Takahashi 1975). More recent studies have suggested that the temperature sensitivity of the N-mediated response is manifested at the level of interaction between the virus and the defence response mechanism ( Padgett et al. 1997 ). This latter hypothesis was based on the observation that the temperature at which the HR is inactivated differed among TMV and the HR-inducing Ob mutants, which suggests the thermosensitivity reflects a property of the virus and not the defence response mechanism. If this thermosensitivity is in fact a viral property, then our experiments indicate that this feature must reside entirely within the p50 replicase fragment.

ATPase activity not is required for HR induction

The p50 replicase fragment contains a putative helicase domain. The role of this domain in the TMV life cycle has not been characterized, although it presumably functions during viral RNA replication and/or cell-to-cell movement of the viral RNA, as suggested for other viral helicases ( Carrington et al. 1998 ; Hayes & Buck 1990). Structural studies of helicase enzymes have indicated that ATP binding and hydrolysis causes large conformational changes within the enzyme that serve to melt duplex RNA ( Korolev et al. 1997 ; Subramanya et al. 1996 ). Consistent with helicase function, we demonstrated that recombinant p50 protein could hydrolyze ATP in vitro ( Fig. 3b). A point mutation in one of the helicase motifs predicted to bind ATP abolished this activity but did not alter the ability to induce HR ( Figs 1 and 3), suggesting that HR induction depends on features of the p50 protein that are independent from its ATPase/helicase activity.

Tools for understanding the mechanisms of TMV recognition by N-containing tobacco

Many different pathogen products have been identified as specific inducers of R gene-mediated defence responses in plants ( Bonas & Van den Ackervaken 1997; Keen & Dawson 1992). These elicitors include extracellular proteins, polysaccharides produced by fungi and bacteria, and intracellular proteins produced by bacteria and viruses. The known viral inducers of plant defences are functionally diverse and include coat, movement, and replicase proteins ( Culver & Dawson 1991; Padgett et al. 1997 ; Pfitzner & Pfitzner 1992; Weber & Pfitzner 1998). In plants, more than a dozen different R genes have been isolated from numerous species ( Hammond-Kosack & Jones 1997). It has been hypothesized that plant R genes encode receptors that bind to pathogen elicitor molecules and signal for the defence responses ( Keen 1990). In fact, a physical interaction between R and Avr gene products has been demonstrated for the tomato Pto protein and its cognate avirulence protein, AvrPto from Pseudomonas syringae ( Scofield et al. 1996 ; Tang et al. 1996 ).

The tobacco N gene has been cloned and shown to be a member of the nucleotide-binding-site/leucine-rich-repeat class of plant disease resistance genes ( Whitham et al. 1994 ). Sequence similarities between N and Drosophila and mammalian Toll receptors are consistent with N functioning as a receptor that recognizes TMV infection and activates a signalling pathway leading to defence responses ( Dinesh-Kumar et al. 1995 ; Lamaitre et al. 1996 ; Whitham et al. 1994 ; Yang et al. 1998 ). Consistent with this idea and with the observations we have presented here, one reasonable model suggests that, upon TMV infection, N would interact with the helicase domain of the TMV replicase proteins or a protein complex dependent on the helicase domain. Because both components of this gene-for–gene interaction have now been identified and molecularly characterized, we are in a good position to address directly questions regarding the mechanisms involved in the recognition of TMV infection by N-containing tobacco.

Experimental procedures

Plant and bacterial materials

SR1 tobacco (Nicotiana. tabacum cv. Petite Havana SR1) was provided by H. Loerz, Universität Hamburg. Transgenic SR1 lines homozygous for the N gene, line 96–6 or 117–18 (referred to here as SR1::NN), were provided by S. P. Dinesh-Kumar, University of California, Berkeley, CA ( Whitham et al. 1996 ). SR1::NN contains a transformation marker for kanamycin resistance linked to the N gene. Plants were grown under greenhouse conditions (18–24°C) for experiments involving transient expression, or in a Conviron incubator (16 h light/8 h dark) at the indicated temperatures for experiments involving stable transgenic expression. The A. tumefaciens strain LBA4404 was used for stable tobacco transformations. For Agrobacterium-mediated transient expression in tobacco, A. tumefaciens strain C58C1::pGV2260 (vir +) and the C58C1 (vir-) strain, lacking plasmid pGV2260 containing the vir genes necessary for T-DNA transfer were used. All Agrobacterium strains were provided by P. Zambryski, University of California, Berkeley, CA. E. coli strain BL21 DE3 pLysS (Promega) was used for production of recombinant proteins. All cloning work used E. coli DH5α.

Construction of TMV expression plasmids

TMV gene constructs were made by cloning TMV-U1 cDNA sequences from plasmid pTMV004 (provided by William Dawson, University of Florida, Lake Alfred, FL, USA) into the plant binary vector pMB. pMB is a derivative of the binary vector pMD1 (provided by Chris Lamb, Salk Institute, La Jolla, CA), which is a derivative of pBIN 19 ( Bevan 1984). To construct pMB, a 0.9-kb Hind III/Sal I fragment containing most of the neomycin phospho-transferase (NPT II) gene was removed from pMD1. This fragment was replaced with a 1.5 kb HindIII/XhoI cassette containing CaMV 35S::BAR::gene 7 terminator, which encodes resistance to the herbicide phosphinothricin ammonium (PPT) ( De Block et al. 1987 ). A double cauliflower mosaic virus (CaMV) 35S promoter ( Carrington et al. 1991 ) and the TMV-U1 5′ untranslated region ( Carrington et al. 1991 ) derived from plasmid pBSG630 (provided by Tom Turpen, Biosource Genetics Corporation, Vacaville, CA, USA) was used for expression of TMV coding sequences in pMB. The transcription start site is predicted to be nucleotide 1 of the TMV 5′ untranslated sequence.

A multistep cloning strategy was used to construct the TMV expression plasmids into pMB. First, a fragment containing the double CaMV 35S promoter and TMV 5′ untranslated sequences were cloned into a pGEM (Promega) E. coli vector, and an NcoI restriction site was added to the 3′ end of the TMV untranslated sequence using PCR and oligonucleotides primers. TMV coding sequences were engineered using PCR to contain an NcoI site at the translational start codon (or an AUG codon was created by the NcoI site) and a XhoI restriction site following the translational stop codon (or a stop codon was also added where needed). These NcoI/XhoI fragments where inserted into the pGEM plasmid using the NcoI site and a XhoI site downstream of the NcoI site in the vector backbone. HindIII/XhoI-restricted DNA fragments containing the promoter, TMV 5′ untranslated, and the TMV coding sequences were mobilized from these intermediate constructs into HindIII/XhoI cut pMB, just upstream of the Nos transcriptional termination sequences present in pMB. This cloning step removes the single CaMV 35S promotor present in pMB.

p80, p50, p43 and p46 contain TMV-U1 nucleotides 69 to 2147, 2082 to 3418, 2265 to 3418, 2082 to 3314, respectively. To add a FLAG-epitope tag, sequences encoding the amino acids DYKDDDDK (single letter code) followed by a stop codon were added to the 3′ end of indicated gene fragments using PCR and oligonucleotide primers. To construct p50T/A, the antisense, PCR primer that has the sequence 5′-aattagatcttcatcaaaattaaccctggaaa gaatttctttggcttttccacagcccggaactcc-3′ was employed to change the codon for threonine 840 (residue number in replicase proteins; nucleotides 2586–2588 in TMV-U1) from ACC to GCC, which codes for alanine. The changed nucleotide is bolded in the above sequence. Sequences of additional oligonucleotide PCR primers used in these constructions are available upon request.

pAvrPto is a binary plasmid bearing the P. syringae AvrPto gene driven by the CaMV 35S promotor ( Scofield et al. 1996 ). Plasmid pGUS is pCLN65, a binary plasmid that contains a CaMV 35S::intGUS gene construct (provided by B. Staskawicz, University of California, Berkeley).

Agrobacterium-mediated transient expression

Agrobacterium-mediated transient expression assay was performed as described by ( Scofield et al. 1996). A. tumefaciens strains bearing the expression plasmids were grown to saturation in 2xYT media ( Sambrook et al. 1989). Following centrifugation, the cells were resuspended in infiltration buffer (60 m m sucrose, 55.5 m m glucose, 0.2 m m acetosyringone, 20 m m MES, pH 5.4) at a concentration of 1.0 O.D.600. This suspension was pressure-infiltrated into leaves using a syringe, soaking several square centimeters of leaf mesophyl tissue. Young, prebolting tobacco plants were infiltrated at up to 10 sites per leaf. Only the 2nd to 5th leaves, counting from the bottom of the plant, were used for infiltration. HR formation was scored over 2 weeks post-inoculation as either positive or negative for necrotic lesions. Each construction was tested a minimum of 25 times in numerous tobacco plants.

Tobacco transformation and screening

SR1 tobacco plants were transformed with TMV expression constructions using Agrobacterium-mediated transformation of leaf discs ( Horsch & Klee 1986). Transformants were selected in the presence of 100 mg l–1 phosphothricin ammonium (Basta; Hoescht Canada Inc.). Four independent primary transformants (T0 plants; lines 34–2, 34–3, 34–6 and 34–7) were chosen for analysis and crossed to homozygous SR1::NN tobacco. Only the results from line 34–3 are shown in Fig. 2, but comparable results were observed for all four lines. The F1 seed of the cross between SR1::p50 lines and SR1::NN and the self seed (T1) for each primary transformant were plated on solid growth media containing 200 mg l–1 kanamycin for selection of the N gene, and/or 100 mg l–1 Basta for selection of the integrated TMV construct. For each transgenic line tested, a minimum of 200 T1 and F1 seeds were plated and scored in two separate experiments. Seeds were germinated at 22°C with a 16-h light/8-h dark cycle. Seedlings that survived herbicide selection produced green cotyledons by 1 week post-germination and were scored for necrotic lesions over a period of 10 weeks. Crosses scored positive showed a necrotic phenotype in 100% of the seedlings examined.

Protein analysis

For analysis of transiently expressed proteins, leaves of SR1 tobacco plants were infiltrated with Agrobacterium cultures and the inoculated tissues were harvested 48 h later, ground to a fine powder in liquid nitrogen and extracted with 2 ml of buffer A (50 m m Tris–Cl pH 8.0, 300 m m NaCl, 10 m m EDTA, 10 m m 2-mercaptoethanol, 1 m m PMSF, 1 mg l–1 aprotinin, and 1 mg l–1 leupeptin) per gram of tissue. Soluble proteins were recovered by centrifugation at 2500 g for 5 min at 4°C, and the resulting supernatant was spun at 14 000 g for 15 min at 4°C. For immunoblotting, 20 μg of soluble protein, as determined by Bradford assay (BioRad, Hercules, CA), was separated using 10% SDS–PAGE and transferred to nitrocellulose membranes, which was probed with polyclonal anti-FLAG epitope antibodies (Santa Cruz Biotechnology, Inc.), and goat anti-rabbit, HRP-conjugated secondary antibody (Pierce, Rockford, IL). Bound antibodies were detected using SuperSignal ULTRA (Pierce, Rockford, IL) as described by the manufacturer.

Production of recombinant proteins

TMV cDNA fragments were cloned into the E. coli expression plasmid pMAL-c2 (New England BioLabs, Inc.) and transformed into E. coli strain BL21 DE3 pLysS (Promega). E. coli cultures were grown in LB media containing 100 mg l–1 ampicillin at 28°C, and MBP protein expression was induced at OD260 = 0.6 with 0.1 m m IPTG, followed by continued growth for 3 h. Cells were harvested by centrifugation at 5000 g and 4°C, and the cell pellets were stored at – 80°C until further processing. To purify MBP fusion proteins, frozen cell pellets were thawed in buffer A (50 m m Tris–Cl pH 7.5, 200 m m NaCl, 0.1% triton X-100, 1 m m DTT, 5% glycerol, 1 m m PMSF, 2 (g ml–1 each aprotinin and leupeptin) on ice and sonicated 3 times each for 30 sec. All steps were performed at 4°C or below. Soluble protein fractions were obtained by centrifugation at 10 000 g for 20 min. To the soluble protein fraction, 1 : 50 volume of amylose resin (1 : 1 in buffer A; New England Biolabs) was added, followed by incubation with gentle mixing for 1–2 h. The resin and bound proteins were washed three times with buffer A, followed by two washes with Assay buffer (25 m m HEPES-OH pH 7.5, 100 m m NaCl, 10 m m MgCl2, 1 m m DTT, 0.1 m m PMSF). The resin with bound proteins were suspended approximately 1 : 1 (v/v) in Assay buffer and used directly in the ATPase assay (see below) and SDS–PAGE analysis. Bound proteins were analyzed by boiling a portion of the resin (equal to amount used in ATPase assay) in SDS sample buffer (50 m m Tris–Cl pH 6.8, 1% SDS, 10 m m DTT, 20% glycerol), followed by separation by 10% SDS–PAGE and staining with Coomassie brilliant blue R-250 ( Fig. 3a). Protein quantities were estimated by comparison to serially diluted BSA standards separated and stained in the same gel. Protein identities were verified by immunoblotting using anti-FLAG antibodies, since each replicase fragment was tagged at the C-terminus with a FLAG epitope tag.

ATPase assay

ATPase activity was detected by the release of 32P-phosphate from ATP[γ32-P]. Reactions were in 50 μl of assay buffer containing 25 μl (2–4 μg) of MBP fusion proteins bound to amylose resin, 0.1 m m ATP and 10 (Ci ATP[γ32-P] (3000 Ci mmol–1), and incubated at 22°C for 30 min with occasional, gentle mixing. The reactions were stopped by the addition of 0.5 m EDTA to 50 m m and transferring the reactions to ice. A commercial dog-kidney ATPase (Sigma) was used as a positive ATPase control. 5 μg of BSA served as a negative control. In ATPase reactions testing for Mg2 + dependence, EDTA was added to 50 m m. Two μl of each reaction was spotted onto a polyethylenenimine (PEI) thin-layer chromatography plate, which was developed with 0.5 m LiCl/0.5 m phosphoric acid pH 3.0. Developed plates were analyzed using a phosphorimager (Molecular Dynamics).

Acknowledgements

We thank Brian Staskawicz, Patty Zambryski, Tom Turpen, S. P. Dinesh-Kumar, and William Dawson for providing materials. We also thank Patti Taranto and members of the Baker lab for comments on the manuscript. This work was supported by NIH postdoctoral fellowship F32 AI10135–01 awarded to F.L.E., US Environmental Protection Agency STAR Fellowship grant U914823 awarded to S.H., US Department of Agriculture grant 5335–22000–004–00D, and Monsanto Cooperative Research and Development Agreement 5335–22000–004–02T.

Ancillary