Analysis of transgenic grapevine (Vitis rupestris) and Nicotiana benthamiana plants expressing an Arabis mosaic virus coat protein gene

Plant Science 156 (2000) 235 – 244 www.elsevier.com/locate/plantsci

Analysis of transgenic grapevine (Vitis rupestris) and Nicotiana benthamiana plants expressing an Arabis mosaic 6irus coat protein gene Albert Spielmann a,*, Stoyanka Krastanova b, Ve´ronique Douet-Orhant a, Paul Gugerli c a

Laboratoire de biochimie, Uni6ersite´ de Neuchaˆtel, Emile-Argand 11, 2007 Neuchaˆtel, Switzerland b Profigen/Agri6itis, Columbia Crest Dri6e, PO Box 188, Paterson, WA 99345, USA c Station Fe´de´rale de Recherches en Production Ve´ge´tale de Changins, 1260 Nyon, Switzerland Received 13 September 1999; received in revised form 20 March 2000; accepted 20 March 2000

Abstract A disarmed LBA4404 strain of Agrobacterium tumefaciens harboring a binary vector which contained chimeric genes encoding the neomycin phosphotransferase (npt II) and the coat protein (CP) of Arabis mosaic nepo6irus (ArMV) was used in co-cultivation experiments with leaf discs of Nicotiana benthamiana and somatic embryos of the grapevine rootstock cultivar Vitis rupestris. Transgenic N. benthamiana expressing the ArMV CP gene were regenerated and six independent lines were characterized. Enzyme-linked immunosorbent assay (ELISA) performed on leaf tissue demonstrated the accumulation of the ArMV CP in five of the six lines analyzed. Immunosorbent electron microscopy (ISEM) studies revealed the presence of virion-like isometric particles (VLPs) reacting to a rabbit antiserum specific to ArMV virions. ArMV-CP expressing transgenic N. benthamiana lines showed protection against ArMV expressed as a delay in infection and a reduction of the percentage of infected plants. Four independent transgenic lines of V. rupestris transformed with the ArMV CP gene were regenerated and characterized. In contrast to N. benthamiana, transgenic V. rupestris did not accumulate the ArMV CP at levels detectable by ELISA and no VLPs could be observed by ISEM. Northern blot analysis showed that the ArMV CP mRNA was expressed at lower level in V. rupestris compared with N. benthamiana. The reason for this difference in transgene expression and/or mRNA stability between grapevine and N. benthamiana is unclear, but the genetic state of the transgene(s) (homozygous in N. benthamiana versus hemizygous in V. rupestris) may have an effect on gene expression. © 2000 Elsevier Science Ireland Ltd. All rights reserved. Keywords: Coat protein-mediated virus resistance; Nepovirus; Grapevine; Nicotiana benthamiana; Plant virus resistance

1. Introduction Nepoviruses are plant viruses with polyedric particles which are transmitted via soil-inhabiting longidorid nematodes (Xiphinema and Longidorus spp.). They are responsible for economically important diseases, especially of perennial plants such as grapevine and fruit trees. Their genome is composed of two separately encapsidated positive single-stranded RNAs. In grapevine, the most

* Corresponding author. Tel.: +41-32-7182222; fax: + 41-327182201. E-mail address: [email protected] (A. Spielmann).

damaging and widespread nepoviruses are Grape6ine fanleaf 6irus (GFLV) and Arabis mosaic 6irus (ArMV), both responsible for the disease known as infectious degeneration or ‘court-noue´’. The disease is controlled by soil disinfection using nematicides. This procedure is forbidden in Switzerland, Germany, Italy and some States in the USA, due to the high toxicity of the chemical (dichlorpropen). Virus-resistant transgenic grapevine plants would be an ecological alternative to control the disease. Since the first report on resistance to Tobacco mosaic 6irus (TMV) in transgenic tobacco plants expressing TMV coat protein (CP) [1], similar results have been obtained for a broad spectrum

0168-9452/00/$ - see front matter © 2000 Elsevier Science Ireland Ltd. All rights reserved. PII: S0168-9452(00)00259-4

236

A. Spielmann et al. / Plant Science 156 (2000) 235–244

of plant viruses [2]. CP-mediated protection against nepoviruses has been demonstrated in transgenic Nicotiana benthamiana expressing the ArMV CP gene [3], the Tomato ringspot 6irus CP gene [4], or the GFLV CP gene [5], and in transgenic N. tabacum expressing the Strawberry latent ringspot 6irus CP gene [6], or the Grape6ine chrome mosaic 6irus CP gene [7]. The formation of virus-like particles resulting from the accumulation of transgenic nepovirus CP in plant tissues was demonstrated for the ArMV CP in N. benthamiana [8]. Genetic transformation of various grapevine tissues and regeneration of transgenic plants have been reported for a number of grapevine rootstocks and cultivars [9]. Recents reports described the regeneration of transgenic grapevine plants expressing a chimeric GFLV CP gene after transformation of somatic embryos of the rootstock cvs 110 Richter [10,11], Gloire de Montpellier, 3309 Couderc, Millardet de Grasset 101-14 [11], 41B and SO4, and the Vitis 6inifera variety Chardonnay [12]. In these cases, expression of the GFLV CP transgene leads to the accumulation of the CP in transgenic tissue (detected by ELISA), but the presence of virion-like isometric particles (VLPs) was not reported. In this paper, we describe the expression of a chimeric ArMV CP gene in an annual herbaceous species, N. benthamiana, and a woody perennial species, V. rupestris. The transgenic N. benthamiana lines expressed the ArMV CP gene at high levels and accumulation of ArMV CP resulted in the formation of VLPs in plant tissue. In contrast, none of the transgenic V. rupestris lines accumulated the ArMV CP at a detectable level. Data from northern blot experiments showed that the apparent lack of ArMV CP protein accumulation in V. rupestris tissue was correlated to differences in transgene expression levels and/or RNA stability in grapevine tissues compared with N. benthamiana. Little or no transgene mRNA could be detected in transgenic grapevine tissues. Progeny of ArMV CP-expressing transgenic N. benthamiana lines showed resistance against ArMV. 2. Materials and methods

2.1. Multiplication and purification of Arabis mosaic 6irus ArMV isolate Triaca 782 was isolated in 1985 from degenerated grapevine (J.-J. Brugger, unpub-

lished results) and propagated on Chenopodium quinoa. The virus was purified [13] and viral RNA extracted as described [14].

2.2. Isolation of the ArMV CP gene The ArMV CP is encoded as part of a polyprotein by RNA-2 [15]. ArMV RNA-2 cDNA was synthesized according to standard procedures (Superscript, Life Technologies) and the cDNA was directly used in various PCR reactions. Based on the published ArMV-CP sequences [16–18], two primers were designed to engineer the ArMV CP gene by adding a methionine initiation codon: a sense upstream primer (ARMV-1), homologous to the first 18 nucleotides of the ArMV CP region, containing in addition an ATG initiation codon and a BclI site at the 5%-end (TGATCATCCATGGGACTTGCTGGTAGAGG), and a reverse dowstream primer (ARMV-2), complementary to the last 22 nucleotides of the ArMV CP gene, carrying in addition an XbaI site at the 5%-end (TCTAGAAACCTAAACTTTAAAACATGT). The amplified ArMV CP fragment was then cloned into the EcoRV site of the pUC-8 vector, creating plasmid pARMV-1 (Fig. 1). The ArMV CP gene was completely sequenced by the PCR cycle sequencing method (fmole kit, Promega).

2.3. Construction of a functional chimeric ArMV CP gene A BclI –XbaI fragment from pARMV-1 containing the complete ArMV CP gene was inserted into the BamHI and SstI sites of the binary vector pBI121.1, replacing the uidA gene [19]. The resulting plasmid (pCACP-1) contained the ArMV CP gene under the control of the CaMV 35S promoter and the nopaline synthase terminator (Fig. 1).

2.4. Plant transformation The plasmid pCACP-1 was transferred from E. coli to A. tumefaciens LBA4404 by triparental mating. In vitro-grown leaf explants of N. benthamiana were transformed with A. tumefaciens containing the pCACP-1 plasmid using standard methods [20]. Putative transformed plants were selected on media containing kanamycin (100 mg/l) and cefotaxime (100 mg/l). Grapevine transfor-

A. Spielmann et al. / Plant Science 156 (2000) 235–244

mation and regeneration were essentially performed as described [10].

2.5. Characterization of transgenic plants The ability to form roots on media containing kanamycin (50 mg/l for N. benthamiana and 10 mg/l for V. rupestris) was used as the first indicator of the transgenic status of the regenerated plantlets, since these amounts of kanamycin completely inhibited root formation of in vitro-grown wild type N. benthamiana or V. rupestris plants (data no shown). Integration of the transgenes was confirmed by Southern analysis using total DNA extracted from 0.5 to 1.0 g of fresh leaf tissue of in

237

vitro-grown N. benthamiana or V. rupestris plants following the procedure described [21]. For Southern analysis, 3 – 5 mg of plant DNA was digested with EcoRI or EcoRV, electrophoresed on a 0.7% agarose gel in 1×Tris–borate–EDTA buffer, and blotted onto a charged Nylon membrane (Boerhinger, Mannheim) by capillarity transfer under denaturing conditions (0.4 M NaOH). Transferred DNA were hybridized to 32P-labelled purified DNA fragment corresponding to the ArMV CP or the npt II coding region. Presence of the ArMV CP transcript sequences was assayed by northern blot analysis. Total plant RNA was extracted from in vitro-grown plants of N. benthamiana by the acid guanidium–thiocyanate–phenol–chloro-

Fig. 1. Construction of the vector pCACP-1. After cDNA synthesis using random hexamers as primers, the coding region of the ArMV CP was specifically amplified using an upstream primer (ARMV-1) complementary to the first 18 nucleotides of the CP coding region according to the published ArMV CP sequence [16 – 18] and carrying in addition a BclI site at its 5%-end and a methionine initiation codon, and a dowstream primer (ARMV-2) homologous to the last 22 nucleotides of the CP coding region and carrying an XbaI site at its 5%-end. The 1539 bp amplified fragment corresponding to the ArMV CP coding region was cloned into the SmaI site of pUC-8 in order to verify the integrity of the gene by sequencing. After digesting pARMV-1 with BclI and XbaI, the XbaI ends were filled-in and the fragment containing the ArMV CP gene was gel-purified and ligated to gel-purified plasmid pBI121.1 [19] cut with BamHI and Ecl136 I. The resulting plasmid (pCACP-1) contained the ArMV CP gene under the control of the CaMV 35S promoter (p35S) and the nopaline synthase terminator (tNOS) regions, replacing the uidA gene. The flags stand for the termini sequences of the T-DNA borders. RB, right border; LB, left border; pNOS, nopaline synthase promoter region; nptII, neomycine phosphotransferase II coding region.

238

A. Spielmann et al. / Plant Science 156 (2000) 235–244

2.8. Protection experiments

form method [22] following the recommendations of the supplier (reagent commercialized by Life Technologies as Trizol). For grapevine, total nucleic acids were isolated as described [21] and total RNA were subsequently purified by the acid guanidium – thiocyanate–phenol–chloroform method procedure. About 10 mg of RNA were separated by electrophoresis on a denaturing formaldehyde – agarose gel, transferred to a nylon membrane, and probed with a digoxigenin-labeled purified DNA fragment corresponding to the ArMV CP region. ArMV CP mRNAs were visualized by chemiluminescent detection (Boehringer Mannheim). The R1 progeny of transgenic N. benthamiana plants were screened for kanamycin resistance (kanR) phenotype by germinating seeds (at least 100 seeds for each individual lines) on agar plates containing MS medium supplemented with 400 mg/l kanamycin sulfate. Seeds were surface-sterilized in 10% commercial bleach for 20 min and rinsed twice in sterile distilled water. Lines producing 100% kanR seeds were used for further analysis. For grapevine, primary transformants were propagated in vitro from nodal cuttings according to standard procedures.

R2 seedling progeny from self-fertilized R1 transgenic N. benthamiana producing 100% kanR seeds were used to evaluate resistance to ArMV infection. Similar numbers (at least 15 plants) of transgenic and control (non-transformed) plants consisting of 6-week-old greenhouse grown seedlings were dusted with Carborundum and were mechanically inoculated with a 1:50 diluted crude sap from systemically ArMV-infected C. quinoa leaves. The inoculum was prepared in 0.01 M phosphate buffer pH 7.0 containing 0.01 M sodium diethylthiocarbamate. Because ArMV CP expressing transgenic N. benthamiana lines contained serologically detectable amount of ArMV CP, the extent of infection was checked 2, 4 and 8 weeks post-inoculation by back transmission of a 1:100 dilution of inoculated N. benthamiana upper leaf homogenate to healthy C. quinoa. ArMV infection in C. quinoa was monitored by ELISA 9, 11 and 14 days after back-transmission. The percentage of infected N. benthamiana plants was calculated as the percentage of infected C. quinoa measured 14 days post back inoculation.

2.6. Expression of the ArMV CP gene in transgenic plants

3. Results

Expression of the ArMV CP gene in transgenic plants was first determined by enzyme linked immunosorbent assay (ELISA), using a rabbit antiserum to ArMV virion essentially as described [13]. Leaf material (0.3–0.5 g) from in-vitro grown N. bentamiana or V. rupestris were ground in 3 ml of extraction buffer (0.5 M Tris–HCl, 2% PVP, 1% PEG 6000, 0.14 M NaCl, 0.05% Tween, 0.02% NaN3, pH 8.2). Optical densities were measured at 405 nm 2 h after addition of the substrate (p-nitrophenylphosphate at 1 mg/ml).

2.7. Electron microscopy Crude leaf sap from N. benthamiana plants grown under greenhouse conditions (4–6-weekold) or from 2-week-old seedlings were negatively stained with phosphotungstic acid and viewed on carbon-coated, Formvar-filmed grids using a Philips 300 electron microscope according to [23,24].

3.1. Construction of the pCACP-1 transformation 6ector The structure of the pCACP-1 transformation vector is illustrated in Fig. 1. The integrity of the ArMV CP and the correct addition of the ATG initiation were verified by sequencing. Comparison of the CP deduced amino acid sequence of isolate Triaca 782 with that of published ArMV sequences of [8,16] reveals a very high degree of homology (93 and 97% identity, respectively).

3.2. Characterization of regenerated plants Seventeen putative transgenic N. benthamiana lines transformed with the pCACP-1 plasmid were regenerated. Genomic DNA isolated from a subset of seven kanamycin resistant N. benthamiana plants was digested with the restriction enzymes

A. Spielmann et al. / Plant Science 156 (2000) 235–244

239

Fig. 2. Molecular characterization of seven transgenic N. benthamiana plants transformed with the ArMV CP gene. (A) Southern blot analysis of DNA isolated from N. benthamiana plants. Total DNA (3 – 5 mg) prepared from an untransformed control plant (wt), or transgenic lines 65-1, 65-2, 65-3, 65-5, 65-6, 65-8 and 65-10 was digested with EcoRI or EcoRV, electrophoresed, and blotted onto Nylon membrane. The probe was a 1.5-kb fragment corresponding to the entire ArMV CP region (solid bar in B). 32 P-labelled lambda DNA was also included in the hybridization mixture to reveal the position of the molecular weight marker fragments. M, lambda DNA cut with HindIII used as molecular weight marker; C, 40 pg of plasmid pCACP-1 cut with EcoRV. In the EcoRV digest, the expected 1.1 kb ArMV-CP internal fragment is indicated by an arrow. (B) The top line shows a schematic representation of the T-DNA from the plasmid pCACP-1 integrated in the plant genome. The symbols used are the same as in Fig. 1. The size of the expected internal EcoRV fragment and the minimum expected sizes for the EcoRI or EcoRV border fragments are shown below.

EcoRI or EcoRV, and characterized by Southern analysis using a 1.5 kb fragment corresponding to the entire ArMV CP as probe (Fig. 2B). In the EcoRI digest, one to three fragments hybridized to the probe in each plant (Fig. 2A). The bands represent right border fragments (Fig. 2), allowing the determination of T-DNA insertion loci and showing that, with the exception of plants 65-5 and 65-6, all the plants gave a specific hybridization pattern, demonstrating that they arose from independent transformation events. The use of EcoRV allowed the detection of both an internal T-DNA fragment of 1.1 kb and border fragments

corresponding to the junction of the T-DNA with the plant genomic DNA (Fig. 2A). Results from both EcoRI and EcoRV digests showed that three lines contained at least one apparently intact single T-DNA copy (lines 65-1, 65-5= 65-6, and 65-10), one line contained at least two T-DNA copies (line 65-3), one line contained at least three T-DNA copies (line 65-2) and one line contained a single incomplete (or rearranged) copy of the ArMV CP gene (line 65-8, in which the expected 1.1 kb internal fragment was missing). Inheritance of the kanamycin resistant trait by plating seeds on selective medium showed that the ratio of resistant to

240

A. Spielmann et al. / Plant Science 156 (2000) 235–244

sensitive R2 seedlings was  3:1 from a self for all the lines analyzed, except line 65-3 (ratio 15:1), demonstrating that T-DNA insertion at a single locus occured in most of the lines, except line 65-3 (2 loci). Only seven transgenic V. rupestris lines transformed with pCACP-1 could be regenerated. Hybridization results of five kanamycin resistant V. rupestris plants are shown in Fig. 3, in which a 1.5-kb fragment corresponding to the entire ArMV CP was used as probe. Data for EcoRV digests (Fig. 3) showed that, with the exception of lines 030502 and 030505, which appeared to be identical based on their hybridization patterns, all transgenic lines arose from independent transformation events. In addition, all have at least one intact copy of the ArMV CP gene. Line 030502 (= 030505) contained a very high number of T-

Fig. 4. ELISA detection of the ArMV CP in the R2 progeny of six transgenic N. benthamiana lines and in four transgenic V. rupestris lines. The OD450 nm was measured 2 h after addition of the substrate (p-nitrophenylphosphate). Ext. buf, extraction buffer; N.b. wt, untransformed N. benthamiana plant; 03wt, healthy V. rupestris plant; 03(ArMV), ArMV-infected V. rupestris plant.

DNA inserts (seven to ten copies), probably assembled as complex tandem and/or inverted repeat units, whereas lines 030501 and 030506 contained at least one insert and line 030503 probably two inserts. Results from Southern blots using other probe/enzyme combinations (data not shown) confirmed these findings and revealed that out of the seven regenerated transgenic V. rupestris lines, only four arose from independent transformation events (lines 030501, 030502, 030503 and 030506).

3.3. Expression of the ArMV CP gene

Fig. 3. Molecular characterization of five transgenic V. rupestris plants transformed with the ArMV CP gene. Southern blot analysis of DNA isolated from V. rupestris plants. Total DNA (3–5 mg) prepared from an untransformed control plant (wt), or transformants 030501, 030502, 030503, 030505 and 030506 digested with EcoRV, electrophoresed and blotted onto Nylon membrane. Hybridization conditions were as described in Fig. 2B. M, lambda DNA cut with HindIII used as molecular weight marker. For a schematic representation of the T-DNA from the plasmid pCACP-1 integrated in the plant genome, see Fig. 2B. The expected 1.1 kb ArMV CP internal fragment is indicated by an arrow.

Expression of the ArMV CP transgene in transgenic N. benthamina and V. rupestris was first tested by ELISA using a rabbit antiserum specific to ArMV virions (Fig. 4). Six of the seven transgenic N. benthamiana lines analyzed accumulated the ArMV CP at various levels, whereas no expression was detected in untransformed N. benthamiana plants. As expected, line 65-8 did not accumulate the ArMV CP at detectable levels, since molecular analysis showed that this line contained an incomplete ArMV CP gene. In contrast, no detectable amount of ArMV CP could be detected by ELISA in four V. rupestris lines containing the same ArMV CP gene (Fig. 4). An electron microscopy study was undertaken to test for the presence of VLPs in transgenic plants, as reported previously [10]. Virion-like structures consisting of empty shells could be detected in the three N. benthamiana lines analyzed (65-2, 65-3,

A. Spielmann et al. / Plant Science 156 (2000) 235–244

65-5) (Fig. 5a). Confirmation of the identity of the VLPs was performed by immunodetection using a rabbit polyclonal antiserum specific to ArMV virions (Fig. 5b). In contrast, VLPs were not detected in V. rupestris despite several attempts of enrichment by sucrose gradient. To determine whether the lack of ArMV CP accumulation in transgenic V. rupestris was due to

241

transgene expression level, we compared the ArMV CP gene expression by northern blot in transgenic V. rupestris and N. benthamiana plants using approximately the same amount of total RNA (10 mg) (Fig. 6). The expected mRNA transcript was detected in most of the transgenic N. benthamiana lines analyzed, with the exception of line 65-8 whereas none was detected in a control plant (wt). A band of the expected size was also visible in the four transgenic V. rupestris lines (030501, 030502, 030503, 030506), but the hybridization signal was weaker than that found in transgenic N. benthamiana plants, particularly in line 030506. No band corresponding to the ArMV CP mRNA was detected in an untransformed grapevine plant (wt), and in ArMV-infected V. rupestris (lane wt +ArMV). In the latter case, the expected 3760 nt band corresponding to the genomic ArMV RNA-2 was clearly visible.

3.4. Protection experiments Fig. 5. Electron microscopy of negatively stained VLPs in crude leaf sap of transgenic N. benthamiana: (a) without prior incubation with an ArMV specific antiserum; (b) with prior incubation with a rabbit antiserum specific to ArMV virions. Bar=100 nm.

Fig. 6. Northern blot analysis. Total plant RNA were extracted from in 6itro grown plants of N. benthamiana or V. rupestris. The probe was a 1.5 kb DNA fragment corresponding to the entire ArMV CP region labelled by PCR according to the recommendations of the supplier of the labeling kit (Boerhinger Mannheim). Two run-off transcripts corresponding to the first half (Tr1 =806 nt) and the entire ArMV CP (Tr2=1537 nt) were synthesized in vitro and 10 pg of each transcript were loaded as size markers. The position of the expected 1596 nt ArMV CP and the ArMV RNA-2 mRNA are shown with an arrow. Wt, untransformed plant; 65-1, 65-2, 65-5, 65-8, 65-10, independent transgenic N. benthamiana lines; 030501, 030502, 030503, 030506, independent transgenic V. rupestris lines; wt/ArMV, ArMV-infected untransformed V. rupestris plant.

Homozygous N. benthamiana R2 progeny were used for protection studies. Preliminary experiments using an inoculum consisting of a 1:10, 1:50 and 1:100 diluted crude leaf sap of ArMV-infected Chenopodium quinoa demonstrated that the 1:50 concentration was sufficient to obtain more than 90% infection of control plants. This dilution was therefore used in further studies. Two weeks postinoculation, 69% of the non-transformed plants were infected, while the percentage of infection of the six independent transgenic lines was significantly lower (from 11% for lines 65-1 and 65-3 up to 55% for lines 65-5 and 65-8) (Fig. 7). Four weeks post inoculation, almost 85% of the non-transformed plants were infected, whereas this percentage remained significantly lower for most of the transgenic lines. Two lines (65-1 and 65-2) showed a significant resistance phenotype, with only 13% (65-1) or 28% (65-2) of infection 4 weeks post inoculation. Three lines (65-3, 65-8 and 65-10) had a low level of resistance (55 to 60%). One line (65-5) showed a high percentage of infection not significantly different from the non-transformed plants. After 8 weeks, all the transgenic lines were infected at the same level as the control plants, demonstrating that the protection phenotype was overcome over time.

242

A. Spielmann et al. / Plant Science 156 (2000) 235–244

Fig. 7. Protection experiments in N. benthamiana plants homozygous for the ArMV CP gene. Each curve represents at least 15 plants. Six-week-old greenhouse grown seedlings were dusted with Carborundum and were mechanically inoculated with 1:50 diluted crude sap from systemically ArMV-infected Chenopodium quinoa leaves. Inoculated plants were observed daily for the development of systemic symptoms. Because ArMV CP expressing transgenic N. benthamiana lines contained serologically detectable amount of ArMV CP, the extent of infection was tested 2, 4 and 8 weeks post-inoculation by back transmission to C. quinoa. The percentage of infected N. benthamiana plants was deduced from the percentage of infected C. quinoa measured 14 days post back inoculation by ELISA.

4. Discussion In this paper, we compare the expression of a chimeric ArMV CP gene in an annual herbaceous species, N. benthamiana, and a woody perennial species, V. rupestris. Six transgenic N. benthamiana lines and four V. rupestris lines that expressed a chimeric gene encoding the CP of the ArMV nepovirus were characterized. Southern blot analysis using a probe specific to the ArMV CP region confirmed that the chimeric ArMV CP gene was integrated in the N. benthamiana and V. rupestris genomes. The number of T-DNA loci varied from one to three in transgenic N. benthamiana lines and from one to ten in transgenic V. rupestris lines. Expression of the ArMV CP gene was demonstrated by northern blot. A mRNA corresponding to the ArMV CP could be detected in all the transgenic lines of N. benthamiana and V. rupestris, except for N. benthamiana line 65-8, which do not contain an intact T-DNA copy. No apparent correlation between T-DNA insertion loci and RNA accumulation levels was observed. All five transgenic N. benthamiana plants expressing the ArMV CP mRNA accumulated the corresponding protein at high level, resulting in VLPs formation visible by electron microscopy. Surprisingly, transgenic V. rupestris lines transformed

with the same chimeric gene did not accumulate the ArMV CP product at detectable levels and no VLPs were produced in grapevine cells. The lack of ArMV CP accumulation in transgenic V. rupestris may have several reasons. Data from southern blotting experiments showed that all transgenic V. rupestris lines had at least one apparently intact copy of the transgene. However, transgene expression studies by northern blot analysis demonstrated that the ArMV CP transcript accumulated at lower level in V. rupestris compared to N. benthamiana. Several hypothesis could account for the apparent reduced level of ArMV CP in transgenic grapevine tissues. First, the rate of transgenes transcription driven by the CaMV 35S promoter might be significantly lower in grapevine compared to other dicot plants, but this can be ruled out because several reports have shown good expression of various transgenes in grapevine using this promoter [9,10]. Second, failure to extract RNA of good quality from grapevine may lead to partial or complete degradation of the transgene RNA. However, recent northern blot data obtained after successive hybridizations of the same blot to a control housekeeper gene (N. benthamiana actin gene), then to a specific ArMV CP sequence showed only a slight degradation of the grapevine RNA samples which could not account for the differences observed. Third, transgene RNA turn-over may be higher in grapevine cells, resulting in lower steady-state level and thus reduced translation rate. Fourth, the genetic state of the transgene(s) present in the grapevine or the N. benthamiana genome may have a significant influence on gene expression. Indeed, all the expression experiments were performed on R2 progeny of N. benthamiana which were homozygous for the transgene(s). In contrast, the V. rupestris lines were vegetatively propagated and were therefore hemizygous for the transgenes. This particular state might influence gene expression level. Finally, the involvement of gene silencing at the transcriptional or post-transcriptional level cannot be ruled out. Transgenic expression of nepovirus CP genes has been reported in herbaceous species for five nepoviruses (ArMV [8], TRSV [4], GFLV [5], GCMV [7] and SLRV [6]. In these studies, the transgenic CP could readily be detected by ELISA in crude plant extracts of most of the transgenic lines analyzed. Interestingly, VLPs formation was

A. Spielmann et al. / Plant Science 156 (2000) 235–244

only demonstrated for ArMV CP expressing plants. Accumulation of nepovirus CP in grapevine species was reported for the GCMV CP [25] and the GFLV CP [10–12]. Krastanova et al. [10] reported the transformation of the rootstocks V. rupestris and 110R (V. berlandieri x V. rupestris), but accumulation of the GFLV CP by ELISA was demonstrated for only two plants. Mauro et al. [12] succeeded in producing a high number of independent transgenic lines of the roostocks 41B (V. 6inifera x V. berlandieri ), SO4 (V. berlandieri x V. riparia) and the cultivar Chardonnay (V. 6inifera). Overall,  30% of the plants were GFLV CP negative by ELISA, whereas the presence of GFLV CP gene was demonstrated by PCR or Southern analyses. In these cases, the steady-state levels of GFLV CP mRNA accumulation was not verified experimentally, so it is difficult to explain the lack of GFLV CP accumulation. Taken together, these observations suggest that, although transgenic GFLV CP accumulation could be obtained in both herbaceous and grapevine plants, a relatively high number of grapevine plants do not produce the GFLV CP at detectable level by ELISA, which is in good agreement with the data described in this paper on ArMV CP accumulation in herbaceous versus grapevine tissues. In contrast, Le Gall et al. [25] reported the obtention of five lines of the grapevine rootstock 110R transformed with the GCMV CP and demonstrated the accumulation of the GCMV CP by ELISA and western blot analysis. In this case, accumulation levels of GCMV CP were in the same range in grapevine as in tobacco. When inoculated with ArMV, the six transgenic N. benthamiana showed a lower percentage of infected plants compared to control plants. The resistance phenotype appeared essentially as a delay in infection, since most of the plants became infected after prolonged incubation (Fig. 7). There was no direct correlation between the level of ArMV CP protein accumulation and the protection phenotype. Similar data were obtained. From the results obtained with the transgenic N. benthamiana lines, it is clear that the expression of an ArMV CP gene in the plant cell is sufficient to protect the plant against infection by ArMV, although this protection phenotype appeared as a delay in virus spread rather than a true resistant phenotype where the plant is immune to virus replication, as described for pathogen-derived pro-

243

tection based on homology-dependent gene silencing [26,27]. In this case, called RNA-mediated virus resistance, both the transgene mRNA and the viral RNA are specifically degraded by an unknown cellular mechanism. The partial protection observed in N. benthamiana may not be effective in long-term protection of a perennial crop such as grapevine, especially if a susceptible variety would be grafted onto a partially resistant transgenic rootstock. On the other hand, such protection may be sufficient to inhibit ArMV replication originating from an inoculum naturally delivered by the nematode vector Xiphinema di6ersicaudatum. The transgenic V. rupestris lines described here are currently being tested for their protection against ArMV infection by grafting onto ArMV-infected V. rupestris plants.

Acknowledgements The authors are grateful to Sophie-Marc-Martin and Marie-He´le`ne Prince Siegrist for their excellent technical support. We thank Professor E. Stutz for critical reading of the manuscript and helpful discussions. This work was supported by a grant from the Swiss National Foundation (Swiss Priority Program). A.S. was partly supported by a fellowship from the Ciba-Geigy Jubilee Foundation (Novartis).

References [1] P. Powell Abel, R.S. Nelson, B. De, N. Hoffmann, S.G. Rogers, R.T. Fraley, R.N. Beachy, Delay of disease development in transgenic plants that express the tobacco mosaic virus coat protein gene, Science 232 (1986) 738 – 743. [2] A.F. Hackland, E.P. Rybicki, J.A. Thomson, Coat protein-mediated resistance in transgenic plants, Arch. Virol. 139 (1994) 1 – 22. [3] D.J. Bertioli, J.I. Cooper, M.L. Edwards, W.S. Hawes, Arabis mosaic nepovirus coat protein in transgenic tobacco lessens disease severity and virus replication, Ann. Appl. Biol. 120 (1992) 47 – 54. [4] L.M. Yepes, M. Fuchs, J.L. Slightom, D. Gonsalves, Sense and antisense coat protein gene constructs confer high levels of resistance to tomato ringspot nepovirus in transgenic Nicotiana species, Phytopathology 86 (1996) 417 – 424. [5] N. Bardonnet, F. Hans, M.A. Serghini, L. Pinck, Protection against virus infection in tobacco plants expressing the coat protein of grapevine fanleaf nepovirus, Plant Cell Rep. 13 (1994) 357 – 360.

244

A. Spielmann et al. / Plant Science 156 (2000) 235–244

[6] S. Kreiah, M.L. Edwards, W.S. Hawes, A.T. Jones, D.J.F. Brown, W.J. Mcgavin, J.I. Cooper, Some coat protein constituents from strawberry latent ringspot virus expressed in transgenic tobacco protect plants against systematic invasion following root inoculation by nematode vectors, Eur. J. Plant Pathol. 102 (1996) 297–303. [7] V. Brault, T. Candresse, O. le Gall, R.P. Delbos, M. Lanneau, J. Dunez, Genetically engineered resistance against grapevine chrome mosaic nepovirus, Plant Mol. Biol. 21 (1993) 89–97. [8] D.J. Bertioli, R.D. Harris, M.L. Edwards, J.I. Cooper, W.S. Hawes, Transgenic plants and insect cells expressing the coat protein of arabis mosaic virus empty viruslike particles, J. Gen. Virol. 72 (1991) 1201–1209. [9] A. Perl, Y. Eshdat, DNA transfer and gene expression in transgenic grapes, Biotechnol. Genet. Eng. Rev. 15 (1998) 365–386. [10] S. Krastanova, M. Perrin, P. Barbier, G. Demangeat, P. Cornuet, N. Bardonnet, L. Otten, L. Pinck, B. Walter, Transformation of grapevine rootstocks with the coat protein gene of grapevine fanleaf nepovirus, Plant Cell Rep. 14 (1995) 550–554. [11] B. Xue, K.S. Ling, C.L. Reid, S. Krastanova, M. Sekiya, E.A. Momol, S. Jule, J. Mozsar, D. Gonsalves, T.J. Burr, Transformation of five grape rootstocks with plant virus genes and a virE2 gene from Agrobacterium tumefaciens, In Vitro Cell. Dev. Biol. Plant 35 (1999) 226 – 231. [12] M.C. Mauro, S. Toutain, B. Walter, L. Pinck, L. Otten, P. Coutos The´venot, A. Deloire, P. Barbier, High efficiency regeneration of grapevine plants transformed with the GFLV coat protein gene, Plant Sci. 112 (1995) 97–106. [13] P. Gugerli, Grapevine fanleaf virus, in: H.U. Bergmeyer (Ed.), Methods of Enzymatic Analysis, Antigens and Antibodies 2, vol. XI, Verlag Chemie, Weinheim, 1986, pp. 431–444. [14] M. Pinck, J. Reinbolt, A.M. Loudes, M. Leret, L. Pinck, Primary structure and location of the genome-linked protein (VPg) of grapevine fanleaf nepovirus, FEBS Lett. 284 (1991) 117–119. [15] G.P. Martelli, C.E. Taylor, Distribution of viruses and their nematode vectors, Adv. Dis. Vector Res. 6 (1989) 151–189. [16] H. Steinkellner, G. Himmler, D. Mattanovich, H. Katinger, Nucleotide sequence of AMV-capsid proteingene, Nucleic Acids Res. 18 (1990) 71–82.

.

[17] D.J. Bertioli, R.D. Harris, M.L. Edwards, J.I. Cooper, W.S. Hawes, Transgenic plants and insect cells expressing the coat protein of arabis mosaic virus produce empty virus-like particles, J. Gen. Virol. 72 (1991) 1801– 1809. [18] A.M. Loudes, C. Ritzenthaler, M. Pinck, M.A. Serghini, L. Pinck, The 119 kDa and 124 kDa polyproteins of arabis mosaic nepovirus (isolate S) are encoded by two distinct RNA 2 species, J. Gen. Virol. 76 (1995) 899– 906. [19] R.A. Jefferson, T.A. Kavanagh, M.W. Bevan, GUS fusions: b-glucuronidase as sensitive and versatile gene fusion marker in higher plants, EMBO J. 6 (1987) 3901 – 3907. [20] R.B. Horsch, J.E. Fry, N.L. Hoffmann, D. Eichholz, S.G. Rogers, R.T. Fraley, A simple and general method for transferring genes into plants, Science 227 (1985) 1229 – 1231. [21] M.A. Lodhi, G.N. Ye, N.F. Weeden, B.I. Reisch, A simple and efficient method for DNA extraction from grapevine cultivars and Vitis species, Plant Mol. Biol. Rep. 12 (1994) 6 – 13. [22] P. Chomczynski, N. Sacchi, Single-step method of RNA isolation by acid guanidium thiocyanate – phenol–chloroform extraction, Anal. Biochem. 162 (1987) 156–159. [23] R. Bovey, Plant virology, in: 7th Conference on Czechoslovak Plant Virologists, High Tatras. Publ. House Slovak Acad Sci, Bratislava, 1973, p. 91. [24] M.E. Ramel, P. Serrant, P. Kulling, P. Gugerli, Monoclonal and polyclonal antibodies for the detection of grapevine fleck associated virus, in: Extended Abstracts of the 11th Meeting ICVG, Montreux, Switzerland, 1993. [25] O. Le Gall, L. Torregrosa, Y. Danglot, T. Candresse, A. Bouquet, Agrobacterium-mediated genetic transformation of grapevine and regeneration of transgenic plants expressing the coat protein of grapevine chrome mosaic nepovirus (GCMC), Plant Sci. (1994) 161 – 170. [26] J.A. Lindbo, L. Silva-Rosales, W.M. Proebsting, W.G. Dougherty, Induction of a highly specific antiviral state in transgenic plants: implications for regulation of gene expression and viral resistance, Plant Cell 5 (1993) 1749 – 1759. [27] H.A. Smith, S.L. Swaney, T.D. Parks, E.A. Wernsman, W.G. Dougherty, Transgenic plant virus resistance mediated by untranslatable sense RNAs: expression, regulation and fate of non essential RNAs, Plant Cell 6 (1984) 1441 – 1453.