Expression of self-complementary hairpin RNA under the control of the rolC promoter confers systemic disease resistance to plum pox virus without preventing local infection

Background Homology-dependent selective degradation of RNA, or post-transcriptional gene silencing (PTGS), is involved in several biological phenomena, including adaptative defense mechanisms against plant viruses. Small interfering RNAs mediate the selective degradation of target RNA by guiding a multicomponent RNAse. Expression of self-complementary hairpin RNAs within two complementary regions separated by an intron elicits PTGS with high efficiency. Plum pox virus (PPV) is the etiological agent of sharka disease in Drupaceae, although it can also be transmitted to herbaceous species (e.g. Nicotiana benthamiana). Once inside the plant, PPV is transmitted via plasmodesmata from cell to cell, and at longer distances, via phloem. The rolC promoter drives expression in phloem cells. RolC expression is absent in both epidermal and mesophyll cells. The aim of the present study was to confer systemic disease resistance without preventing local viral infection. Results In the ihprolC-PP197 gene (intron hair pin rolC PPV 197), a 197 bp sequence homologous to the PPV RNA genome (from base 134 to 330) was placed as two inverted repeats separated by the DNA sequence of the rolA intron. This hairpin construct is under the control of the rolC promoter.N. benthamiana plants transgenic for the ihprolC-PP197 gene contain siRNAs homologous to the 197 bp sequence. The transgenic progeny of ihprolC-PP197 plants are resistant to PPV systemic infection. Local infection is unaffected. Most (80%) transgenic plants are virus free and symptomless. Some plants (20%) contain virus in uninoculated apical leaves; however they show only mild symptoms of leaf mottling. PPV systemic resistance cosegregates with the ihprolC-PP197 transgene and was observed in progeny plants of all independent transgenic lines analyzed. SiRNAs of 23–25 nt homologous to the PPV sequence used in the ihprolC-PP197 construct were detected in transgenic plants before and after inoculation. Transitivity of siRNAs was observed in transgenic plants 6 weeks after viral inoculation. Conclusions The ihprolC-PP197 transgene confers systemic resistance to PPV disease in N. benthamiana. Local infection is unaffected. This transgene and/or similar constructs could be used to confer PPV resistance to fruit trees where systemic disease causes economic damage.


Background
Plants use post-transcriptional gene silencing (PTGS) as a mechanism of adaptive protection against viruses [1]. PTGS is closely related to RNA interference and, consequently, represents an ancient eukaryotic genetic phenomenon [2]. PTGS correlates with the presence of small interfering RNAs (siRNAs) [3]. These latter molecules are composed of 21-26 ribonucleotides that originate from double-stranded RNA (dsRNA) by the action of Dicer RNAses [4] [5]. dsRNA can be generated by several mechanisms that include host or virus-encoded RNA-dependent RNA polymerases, transcripts bearing inverted repeats, and transcription of complementary DNA strands. siRNAs are a component of the RNA-induced silencing complex (RISC) [6]. RISC ribonuclease activity degrades RNA homologous to siRNAs [7]. siRNAs homologous to the target RNA guide selective RNA degradation [7]. Thus, siR-NAs confer specificity to the ribonuclease activity of RISC [3][6] [7].
Post-transcriptional gene silencing acts intracellularly. However, the RNA silencing state can be transmitted from a plant cell to its neighbour via plasmodesmata and at longer distances via the vascular system [8]. In plants, it has been proposed that the 21-22 nt-long siRNAs are involved in the selective intracellular degradation of RNA, while the longer (24-26 nt) siRNAs are involved in systemic signaling of PTGS and/or other aspects of RNA silencing [9][10] [11].
PTGS has been achieved with high efficiency in transgenic plants expressing self-complementary hairpin RNA (hpRNA) [12] [13]. hpRNAs have two complementary regions that form a double-stranded region separated by a short loop. Hairpin constructs having a spliceosomal intron inserted in the loop region (ihpRNA) show an efficiency of up to 90% in eliciting RNA silencing [13].
Plum pox potyvirus (PPV) is the etiological agent of sharka, a severe plant disease that infects Drupaceae, such as peach, apricot, plum, cherry and almond [14]. Systemic infection by PPV impairs plant growth and drastically curtails fruit productivity and quality [14] [15]. PPV may be transmitted either by aphids in a non-persistent manner or by grafting [15]. The PPV genome consists of a positive single-stranded RNA of approximately 10 kb [16]. Viral replication begins in infected epidermal cells and then, via plasmodesmata, the virus reaches mesophyll cells, bundle sheaths, phloem parenchyma and companion cells [17]. Long distance movement occurs through the phloem sieve-tubes and causes rapid systemic spreading of the infection within the plant [17]. Resistance to PPV disease has been achieved by PTGS in one of 5 independent peach plants transgenic for the PPV coat protein expressed under the control of the constitutive 35S CaMV promoter [18].
Other approaches have been used to obtain transgenic plants transformed with PPV sequences in herbaceous species [15]. Fully resistant N. benthamiana plants have been obtained in one of 7 lines transgenic for a mutated CI protein [19]. N. benthamiana plants transgenic for mutated DNA sequences corresponding to the NIa-NIb and N-terminus of the capsid protein, showed virus resistance and, depending on the specific mutations, in all or in a fraction of the plants tested [20].
We hypothesised that if systemic movement of PPV could be impaired, the virus should remain confined to inoculated leaves and, consequently, would not cause systemic disease. In this regard, primary inoculated leaves never show the severe symptoms that characterise systemic infection by potyvirus. Furthermore, the number of primary infection foci determined by aphid transmission does not extensively affect plant growth and productivity. Moreover, PTGS is an adaptative mechanism of the defence where the virus genome is both target for degradation and substrate for amplification of the PTGS response. Thus, in our opinion, the enduring presence of viruses in locally infected leaves might boost PTGS.
Thus, we have tested whether systemic resistance to PPV disease can be conferred by an hpRNA gene construct that would be predicted to be processed by Dicer RNAse(s) to siRNAs. siRNAs homologous to the PPV genome would then trigger the degradation of PPV RNA. To drive expression of the hpRNA construct, we have used the rolC promoter from Agrobacterium rhizogenes [21] [22]. The rolC gene, and consequently the rolC promoter, is not expressed in either epidermal or mesophyll cells, but is expressed in cells of the vascular system. In Nicotiana tabacum, by using a rolC-uidA reporter gene, rolC driven expression was localised to cells of the phloem vascular system [21] [22]. In the same species, using anti-rolC affinity purified antibodies, rolC expression was localised in companion and protophloem cells [23]. In potato, the rolC promoter has been used to confer viral resistance to potato leafroll virus (PLRV) by expressing the PLRV coat protein under its control [24]. In this species, by using a rolC-uidA reporter gene in a strongly expressing transgenic plant, rolC driven expression was also observed in bundle sheath and vascular parenchyma cells, but not in epidermal or mesophyll cells [24].
HairpinRNA gene constructs in which the 2 complementary regions of the hpRNA separated by an intron are extremely efficient in triggering PTGS [13]. In the ihprolC-PP197 (intron-hairpin rolC PPV 197) gene construct used for the present experiments, the DNA sequence corresponding to the rolA intron was inserted between the two sequences derived from the PPV genome that build the double-stranded region of the hpRNA. The rolA intron of A. rhizogenes is 85 bases long and is spliced in both Arabidopsis [25] and Nicotiana tabacum [26]. Agrobacterium rhizogenes causes hairy-root disease in many plant species [27]. Consequently, the rol genes are likely to be widely expressed among plants. Thus, the rolC promoter and the rolA intron should provide the genetic information necessary for the construction of a hpRNA that would be expressed in most, if not all, plant species. Here, we show that the ihprolC-PP197 gene confers systemic PPV resistance to Nicotiana benthamiana plants. Consistent with the known expression pattern of the rolC promoter, PPV infection and replication are not prevented in inoculated leaves of ihprolC-PP197 transgenic plants. Local symptoms develop in inoculated leaves, although systemic PPV disease is either prevented or drastically curtailed.

Rationale and design of the ihprolC-PP197 transgene
Expression of the ihprolC-PP197 transgene (Fig. 1, panel A) was placed under the control of the rolC promoter from Agrobacterium rhizogenes. The rolC promoter is located in the intergenic region between the rolC and the rolB T-DNA genes of A. rhizogenes and spans 850 bp upstream of the transcriptional initiation site of rolC [28]. In the ihprolC-PP197 gene, an 1182 bp DNA fragment spanning the entire intergenic region was used to drive gene expression. However, shorter DNA fragments might be used as well (e.g. the 850 bp long HindIII/HpaI DNA fragment [24] [28]). The rolC promoter drives vascular-specific expression. RolC gene expression takes place mainly in companion cells [23] and is also present in other phloem cell types [24], but is absent in epidermal and mesophyll cells [23] [24]. The transcribed region of the ihprolC-PP197 transgene is composed of 2 DNA sequences placed in inverted orientation in order to produce a self-complementary dsRNA molecule. Each of the 2 sequences consists of 197 bases derived from the genome of PPV strain D. The 197 bp DNA sequence corresponds to bases 134 to 330 of the PPV genome (accession number X16415). Thus, its sequence begins 13 bases before the AUG initiation codon of the PPV polyprotein at nucleotide 147 and ends 181 bases after the translational start site. The 197 base-long region shows 99.5% homology to the corresponding sequence of PPV strain M (the so-called PPV epidemic strain). An 87 bp-long DNA sequence containing the rolA intron of A. rhizogenes was placed between the two 197 bp complementary sequences (Fig. 1, panel A). Transcriptional termination and polyadenylation signals are present in a 251 bp DNA sequence derived from the rolC gene. The DNA sequence of the ihprolC-PP197 is shown in Figure 1, panel B. Thus, in transgenic plants, after splicing the pre-mRNA transcribed from the ihprolC-PP197 gene produces a mRNA with 2 virus-derived self-complementary regions of 197 bases each, separated by 14 bases that are predicted to form a loop of 10 bases.

Transgenic state of ihprolC-PP197 Nicotiana benthamiana plants
The transgenic state of independent Nicotiana benthamiana transformation events was analyzed by Southern blotting. The 5 independent transgenic plants (T0 primary transformants) have either one (lines 13 and 29; Fig. 2, panel A, lanes 4 and 2, respectively), two (line 8; Fig. 2, panel B, lane 1), or several copies (i.e. at least 5 in line 9 and at least 4 in line 2; Fig. 2, panel A, lanes 3 and 5, respectively) of the ihprolC-PP197 transgene per haploid genome. It is worthwhile mentioning that plant line 13 has a ihprolC-PP197 transgene deleted of the region upstream of the HindIII site positioned 850 bp upstream the rolC initiation of transcription (Fig. 2, panel A, lane 4). Consequently, the deletion does not involve rolC promoter sequences.
The T0 plants were backcrossed with wild-type plants and the T1 progeny were analyzed for kanamycin resistance (data not shown). The transgenic state of the T1 progeny were either confirmed by PCR (data not shown) or characterised by Southern blot analysis (Fig. 2, panel B). In summary, on the basis of Southern blot analysis ( Fig. 2 and data not shown) for either the ihprolC-PP197 transgene or the nptII (kanamycin resistance) gene, in addition to segregation analysis for kanamycin resistance, the following conclusions can be drawn. Plant line 13 had one copy of the nptII gene and one copy of the ihprolC-PP197 transgene. Plant line 29 had one copy of the transgene and 2 copies of the nptII gene, with one of the 2 nptII genes unlinked to the ihprolC-PP197 transgene. In plant line 8, kanamycin resistance segregated as 2 loci. Southern blot analysis of the kanamycin resistant progeny of line 8 confirmed the segregation data of 2 unlinked T-DNA insertions, each one containing one ihprolC-PP197 and one nptII gene (Fig. 2,

RNA analysis of ihprolC-PP197 transgenic plants
The 5 independent T0 transformants were analyzed for the presence of ihprolC-PP197 mRNA (Fig. 3, panel A). In each plant, Northern blot analysis showed a mRNA of the expected size, i.e. approximately 700 bases (Fig. 3, panel A, lanes 2-6). Thus, all 5 independent transgenic plants (T0) expressed the ihprolC-PP197 transgene, albeit at rather low steady-state levels for rolC driven expression. In fact, the steady-state level of ihprolC-PP197 mRNA was 5 to 10 times lower than that of a transgene (rolC-PP2x197) composed of the same regulatory and transcribed regions of the ihprolC-PP197 construct, but bearing the 197 bp viral sequences repeated in a direct orientation ( . The blots were hybridised to a DNA probe corresponding to a 484 bp long sequence at the 3' end of the ihprolC-PP197 gene. The ihprolC-PP197 and rolC-PP2x197 transcripts were detected after 3 days and 12 h of autoradiography, respectively. The membranes were also hybridised to an actin probe and exposed for 12 h. Both probes were labeled with (α-32 P)-dCTP (3000 Ci/mmol) using "Ready to go DNA labeling beads (-dCTP)" (Amersham Biosciences); 10 6 cpm/ml of labeled probe were added to the hybridization buffer.
RNA gel blot analysis of siRNAs in ihprolC-PP197 transgenic plants immunocapture-rtPCR (IC-RT-PCR). Seven days after inoculation, samples from inoculated basal leaves (half leaves) and from non-inoculated apical leaves of all plants were analyzed. Inoculated basal half-leaves were positive by ELISA in both wild-type and transgenic plants (Table  1). Moreover, the ELISA values were rather similar in wildtype and transgenic plants (Table 1). Thus, neither local viral infection nor replication is prevented in ihprolC-PP197 transgenic plants. The non-inoculated half leaves from 25 transgenic plants were also analysed by DAS-ELISA. In 21 of 25 plants tested, the ELISA values indicated that they were free of virus (below detection limits) (data not shown).
The systemic spreading and replication of PPV was evaluated by ELISA performed on apical non-inoculated leaves ( Table 2). At 7 days, apical (non-inoculated) leaves from the progeny of plant lines 13 and 29 were negative in all transgenic plants. The absorbance values were never higher than the average value plus 3 times the standard deviation measured in leaves of mock-inoculated transgenic and healthy wild type plants. Seven days after inoculation, 10 out of 25 plants from the progeny of plant lines 2, 8 and 9, were positive, while the other 15 plants gave values below detection levels ( Table 2).
In order to evaluate whether PPV was able to overcome the systemic resistance conferred by the ihprolC-PP197 PPV inoculated ihprolC-PP197 transgenic and untransformed control plants  (Table 2). It is noteworthy to stress that all the other symptomless plants that were ELISA positive one week after inoculation, were negative 6 weeks after inoculation.
In order to validate the ELISA analysis, 30 samples from ELISA-negative, asymptomatic plants were randomly chosen and analysed by IC-RT-PCR, a technique that is at least 1000-fold more sensitive than ELISA in detecting PPV [29]. All samples were also negative by IC-RT-PCR, con-firming the absence of the virus and validating the ELISA assay.

Transitivity of siRNAs in PPV inoculated Nicotiana benthamiana plants transgenic for the ihprolC-PP197 gene
SiRNAs were analysed in PPV inoculated plants. Northern blot analysis was performed using probes homologous either to the 197 bp region used in the hairpin construct (Fig. 8, panel A) or to PPV sequences located between 1000 and 1500 bases at the 3' of the 197 bases used in the hairpin construct (Fig. 8, panel B). siRNAs homologous to the 197 bp region of the hairpin construct were detected in inoculated transgenic plants both 2 days after inoculation (Fig. 8, panel A, lanes 2 and 3) and 42 days after inoculation (Fig. 8, panel A, lane 4). Two classes of siRNAs were found: namely, 23 and 25 nt long species (Fig. 8, panel A, lanes 2-4). siRNAs were absent in inoculated wild type controls collected 2 days after inoculation (Fig.  8, panel A, lane1). When the same blot was hybridised to a probe derived from viral sequences located between 1000 and 1500 bases at the 3' of the 197 bases used in the hairpin construct, siRNAs were detected in inoculated transgenic plants 42 days after inoculation (Fig. 8, panel B, lane 4), but were undetectable in inoculated transgenic plants 2 days after inoculation (Fig. 8, panel B, lanes 2 and  3). However, as different plant lines were tested, it is possible that the absence of transitivity observed 2 days after viral inoculation is dependent from the particular line assayed.
Thus, after viral inoculation, transitivity of siRNAs takes place in ihprolC-PP197 genetically modified plants. It is noteworthy that the transgenic plant analysed 42 days after inoculation (Fig. 8,

Discussion
Plum pox potyvirus causes severe economic losses of fruit trees belonging to Drupaceae. Sharka, the PPV disease, has affected more than 100 million trees worldwide and is considered the most devastating disease of stone fruit trees in Europe [30]. PPV disease has also caused significant damage in South America, Africa and Asia [31]. Just a few years ago, sharka disease was also detected in North America, which is presently facing damage to selected but important stone fruit species [31].
Damage and economic losses are due to systemic PPV infection and disease [15]. To date, conventional breeding has had only limited success in producing commercial varieties of resistant plants [32]. In contrast, the expression of the gene coding for the PPV coat protein under the control of the 35S CAMV promoter conferred resistance to PPV infection in one of 5 peach plants tested [18]. It is interesting to note that in the PPV resistant transgenic peach plants, PPV was detected by IC-RT-PCR [18]. However, the presence of a very low level of virus genomes, insufficient to cause damage, might be relevant for sustaining an enduring PTGS-based virus resistance.

Symptoms in PPV inoculated transgenic plant
Recently, dsRNA expressed in bacteria has been shown to prevent PPV infection up to 7 days after spraying [33].
PPV can be transmitted either by aphids in a non-persistent manner or by grafting [15]. Aphids inject PPV into leaf epidermal cells. Subsequently, PPV moves to mesophyll cells to bundle sheath and vascular parenchyma cells. From these cells, PPV then moves to companion cells [17]. From the companion cells, PPV enters into the sieve tubes and then is spread by the phloem transport system [17]. Thus, within the plant, PPV spreading over a long distance takes place via the phloem. Consequently,  In this regard, it is worthwhile to mention that potyviruses encode proteins that act as suppressors of PTGS [34]. In particular, it has been shown that the HC-Pro curtails siR-NAs accumulation and inhibits PTGS intracellularly [35]. Current knowledge would suggest that without preventing HC-Pro activity, the viral resistance conferred by PTGS might not be stable. However, HC-Pro does not appear to inhibit systemic signaling of PTGS [36]. During our study, a loss of resistance was never observed. In fact, we have observed in few cases that ELISA values decreased below threshold levels during plant growth (see Table 2).
We have demonstrated that expression of an antiviral-hairpinRNA construct under the control of the rolC promoter results in resistance to systemic PPV infection, while local leaf infection is not prevented. This finding is consistent with the fact that the rolC promoter does not drive gene expression in non-vascular tissues [21][22][23] [24]. In fact, the rolC promoter drives vascular-specific expression and is present mainly in companion cells [23], but can also be found in protophloem and other vascular cell types [21][22][23] [24]. In this regard, it is worthwhile to stress that the systemic resistance conferred by the ihprolC-PP197 gene does not prevent virus infection and replication in epidermal and mesophyll cells. PTGS is an adaptative mechanism of defence. Thus, the restricted presence of the viral genome in the inoculated leaves is meant to amplify the PTGS response.
Our data confirm that ihpRNA gene constructs are efficient in eliciting PTGS [12] [13]. Moreover, considering that hpRNAs less than 100 nucleotides in length are usually efficient in eliciting RNA silencing [13] and that viral RNA genomes are usually several kb long, it would be RNA gel blot analysis of siRNAs in virus-inoculated ihprolC-PP197 transgenic plants feasible to design several different hpRNA from the sequence of any plant RNA viral genome. In the ihprolC-PP197construct, the region of homology to PPV RNA genome spans the AUG initiation codon of the PPV polyprotein. It includes 13 nucleotides of the 5'UTR before the AUG initiation codon and the first 181 bp of the PPV polyprotein coding region, including sequences that are necessary for efficient translation [37] and part of the 18 nucleotide long sequence from base 127 to 145 reported to be required for efficient PPV-host interaction [38]. IhprolC-PP197 transgenic plants contain siRNAs of 23 and 25 bases homologous to the 197 bases of the PPV genome used in the hairpin construct. It has been previously reported that PTGS correlates with the presence of siRNAs of 2 size classes [9]. Virus inoculated transgenic plants show siRNAs homologous to regions of the viral genome located at the 3' of the sequence used in the hairpin construct. Transitivity of RNA interference involves that the RNA targeted by PTGS functions as template for 5' to 3' synthesis of new dsRNA. Processing of these dsRNA by Dicer RNAse will produce secondary siRNAs. PTGS is an adaptative mechanism of defense. The viral genome is both a target for RNA degradation and a template for PTGS amplification. Thus, resistance results from the interplay of host mechanisms and virus inoculum and replication.
The ihprolC-PP197 gene, or similar constructs, could be used for engineering resistance to systemic PPV disease in plants belonging to the Drupaceae. Moreover, considering that the PTGS state can be transmitted by grafting from a silenced stock to a non-silenced scion [8], it will be interesting to test whether transgenic ihprolC-PP197 rootstocks will be able to confer resistance to a grafted non-transgenic scion. It is worthwhile to stress that the ihprolC-PP197 construct does not code for a protein. Moreover PTGS does not appear to alter the transcriptome except for the target RNA [39]. Although the rolC promoter is bound to drive expression also in the vascular tissue of the transgenic fruits, fruits should not be modified in their transcriptome, but only in their resistance to PPV.

Gene constructs, recombinant plasmid vectors, and bacterial strains
A DNA sequence homologous to 197 bases of the viral genome of PPV strain D was synthesised in vitro using Taq polymerase and appropriately designed oligonucleotides. The sequence spans the viral genome from base 134 to base 330 (accession number X16415) and, consequently, it includes 13 bases upstream and 181 bases downstream from the AUG initiation codon of the PPV polyprotein. Two 197 bp DNA fragments homologous to the PPV genome were placed in inverted orientation downstream from the 1182 bp DNA fragment containing the promoter of the RolC gene of Agrobacterium rhizogenes strain A4. An 87 bp DNA fragment containing the 85 base-long intron of the rolA gene of Agrobacterium rhizogenes was inserted between the two virus-derived sequences (Fig. 1). A 251 bp DNA fragment containing the termination and polyadenylation sequences of the RolC gene was placed downstream of the second virus-derived sequence. The resulting gene, termed ihprolC-PP197, was subcloned in the T-DNA region of a derivative of pBin19 binary vector [40]. The recombinant pBin19 vector was used to transform Agrobacterium tumefaciens strain LBA4404 as described by Shen and Forde [41].

Tobacco leaf disc infection and regeneration of transgenic plants
Nicotiana benthamiana plants were cultured on MS medium in Magenta boxes and grown in a growth chamber at 20°C under a 16 h photoperiod. Fully expanded leaves were removed and cut into 1.5 cm 2 discs. Leaf pieces were dipped for 20 min in a bacterial suspension. The leaf explants were then removed from the solution and placed upside down on MS medium supplemented with 30 g/l saccharose, 1.0 mg/l 6-benzylaminopurine, 0.1 mg/l α-naphtalene acetic acid, and 7 g/l plant agar. After two days, leaf explants were transferred to fresh MS medium containing 30 g/l saccharose, 1.0 mg/l 6-benzylaminopurine, 0.1 mg/l α-naphtalene acetic acid, 100 mg/l kanamycin, 500 mg/l cefotaxime, and 7 g/l agar. When elongated, putative transgenic shoots were excised from the calli and cultured in Magenta boxes on MS medium supplemented with kanamycin (100 mg/l) without growth regulators to induce rooting. Rooted plantlets were subsequently transferred to soil and grown to maturity in a greenhouse.

Plant culture conditions
Nicotiana benthamiana plants were grown in sterile potting mixture in a greenhouse at 24-21°C (day and night temperatures, respectively) with a photoperiod of 14 hours. Independent transformation events (T0) were characterised for their transgenic state and backcrossed with wildtype N. benthamiana plants. The T1 progeny of each primary trasformant were tested for kanamycin resistance and resistance to PPV.

Virus source and plant inoculation
The original PPV isolate, a non-aphid transmissible Dserotype, was supplied by DSMZ Plant Virus Collection (PV0001) and maintained by passage on N. benthamiana. The viral inoculum was prepared by grinding 0.5 g of symptomatic N. benthamiana leaves in a precooled mortar in presence of 10 ml of 0.05 M potassium phosphate buffer, pH 7.2. Transgenic T1 plants and wild-type controls were inoculated at the stage of 6-7 true leaves. The third, fourth, and fifth leaves were dusted with sterile Celite ® and mechanically inoculated on one half of each leaf. Plants were monitored daily by visual inspection.

DAS-ELISA and IC-RT-PCR
PPV-inoculated, transgenic progeny (T1) plants from 5 independent transformation events were tested for virus multiplication by double-antibody sandwich (DAS) ELISA [42] and by immunocapture-RT-PCR (IC-RT-PCR) [43]. Samples from inoculated basal half-leaves and apical non-inoculated leaves were collected from each plant 7 days after inoculation. Samples from apical leaves were also collected 42 days after inoculation. Leaf tissues were ground (1:5 w/v) in extraction buffer and clarified by centrifugation. The supernatant was then used in the tests. Reagents for ELISA (polyclonal antibody and alkaline phosphatase-conjugated polyclonal antibody) were supplied by LOEWE Biochemica GmbH (Germany). Absorbance values at 405 nm were determined using a Microplate Reader 550 (Biorad). Immunocapture-RT-PCR detection of PPV was carried out on the same clarified supernatant used for DAS-ELISA tests and with the same polyclonal antibody supplied in the ELISA kit.

Southern blot and PCR analysis
Genomic DNA was extracted from 0.5-1 g of frozen leaves using the Nucleon PhytoPure system (Amersham Biosciences) according to the manufacturer's instructions. DNA (15 µg) was digested with 70 Units of either HindIII or DraI, electrophoresed on a 0.7% agarose gel at 4.5 V cm -1 , and transferred to a nylon membrane (Hybond N, Amersham Biosciences). The membrane was hybridised with 100 ng of fluorescein-labeled probe prepared using the Random prime labeling module (Amersham Biosciences). Detection was performed with anti-fluorescein conjugated alkaline phosphatase and the chemiluminescent alkaline phosphatase CDP-Star substrate (Amersham Biosciences) according to the manufacturer's instructions. Membranes were exposed for 1 h using Kodak XAR-5 film.

Small RNA (siRNA) analysis
Small RNA analysis was performed following the protocol described by Hamilton and Baulcombe [3] with slight modifications. Total RNA was extracted from young leaves with TRIZOL (Invitrogen). The RNA pellet was dissolved in 0.5 ml of sterile DEPC-treated water, heated at 65°C for 5 min, and placed on ice. PEG-8000 (5% final concentration) and NaCl (0.5 M final concentration) were then added. The RNA preparation was then kept for 30 min on ice and high molecular weight nucleic acids were pelleted by centrifugation at 10,000 × g for 10 min. The supernatant containing small RNAs was recovered and further purified by using a RNA/DNA purification kit (Qiagen) to remove contaminants such as proteins, polysaccharides, and carbohydrates. The low molecular weight RNAs were precipitated by adding 3 volumes of ethanol and 1/10 volume of sodium-acetate pH 5.2.
Small RNAs (usually 40 µg per lane) were separated on a 15% polyacrylamide-7 M urea gels and transferred to a Hybond N + membrane (Amersham Biosciences) by electroblotting for 1 hour at 100 V in 0.5X TBE. The membrane was placed for 10 min on three layers of Whatman 3 MM paper dipped in 20X SSC and then UV cross-linked. Blots were probed with a single stranded RNA corresponding to the region from 134 to 330 of the viral genome obtained by in vitro transcription with the Riboprobe in vitro transcription system (Promega) using either T7 or T3 polymerase and (α-32 P)-CTP. The probe was hydrolyzed with 300 µl of alkaline buffer (80 mM sodium bicarbonate, 120 mM sodium carbonate) at 60°C for 2.5 hours. The time of hydrolysis was calculated according to the formula reported by Hamilton and Baulcombe [3]. The probe was then neutralised with 20 µl of 3 M sodium acetate, pH 5.2. The membrane was prehybridised at 40°C for 1 hour and, after the addition of the labeled probe, blots were hybridised overnight at the same temperature. The membrane was rinsed twice in 2X SSC/0.2% SDS at 50°C for 10 min and exposed to Kodak X-AR5 film. Oligoribonucleotides of 23 and 21 nt were used as molecular markers.

Northern blot analysis
Total RNA was isolated with TRIZOL (Invitrogen) and 10-12 µg of total RNA were separated on 1% agarose-formaldehyde denaturing gels. The gel was blotted on a Hybond N + membrane (Amersham Biosciences) in 10X SSC. RNA markers were stained with 0.1% toluidine blue in 20% ethanol. The DNA probe, corresponding to the 484 bplong sequence at the 3' end of the construct, was labeled with (α-32 P)-dCTP using "Ready to go DNA labeling beads (-dCTP)" (Amersham Biosciences). Unincorporated nucleotides were removed by using Probe G-50 micro columns (Amersham Biosciences). The membranes were hybridised overnight at 42°C in ULTRAhyb (Ambion). The membranes were washed 2 times in 2X SSC/0.1% SDS and 2 times in 0.1X SSC/0.1% SDS at 42°C. Autoradiography was then performed using Kodak X-AR5 film.