Phylogenetic Relationships of SPCSV Isolates

SPCSV was detected by TAS-ELISA from plants belonging to 12 wild species (Ipomoea acuminata, I. cairica, I. obscura, I. repens, I. rubens, I. sinensis, I. spathulata, I. stenobasis, I. tenuirostris, I. wightii, Hewittia sublobata and Lepistemon owariensis) growing in their natural habitats in 13 different geographical regions of Uganda. A total of 30 SPCSV isolates from different wild hosts and representing the 13 different regions (Table 1) were selected for further study along with 34 isolates of SPCSV detected in sweetpotatoes in the same 13 regions of Uganda.

Sequences of the 3’-proximal region of RNA1 in the SPCSV isolates were obtained by reverse transcription, amplification by PCR using a high-fidelity DNA polymerase and direct sequencing of the PCR products without cloning. The ORF for RNase3 was of identical length (684 nt; 228 aa) in all SPCSV isolates characterized in this study, including the 64 new isolates from Uganda, the two isolates detected in local sweetpotato cultivars in Central America (HN1 in Honduras and A1GT in Guatemala) [30] also characterized in this study, and the six previously characterized isolates from Tanzania, Uganda, Israel and Peru included for comparison (Table 1, Figure 1). The ORF for p7 was 171 nt (57 aa) in all EA strain isolates, but was shorter (165 nt; 55 aa) in the WA strain isolates. The ORF for p22 was of identical length (572 nt; 192 aa) in all isolates containing this gene. However, isolates KML33b, KTK40 (I. sinensis), SOR71 (I. obscura), HOM40, KTK39 and KTK41 (sweetpotato) characterized from Uganda and the two SPCSV isolates from Central America did not contain p22. Similarly, the previously characterized two EA strain isolates of SPCSV from Tanzania, one EA strain isolate from Kenya, the EA strain isolates from Peru and the WA strain isolates do not contain p22 [8,10]. In all SPCSV isolates lacking p22, the gap in RNA1 was identically positioned. It began immediately after the stop codon for p7 in the EA strain isolates and six nt downstream from the stop codon of p7 in the WA strain isolates.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 1. Alignment of the 43 different RNase3 protein amino acid (aa) sequences of Sweet potato chlorotic stunt virus.

Groups of isolates containing identical RNase3 aa sequences are represented each by a single isolate. Numbers on top of the alignment indicate the aa positions with reference to isolate Ug (AJ428554) whereas numbers on the right indicate the number of the last amino acid at that position for each isolate. Numbers at the bottom of the alignment indicate aa positions in the new unknown virus (KML33b) related to SPCSV and detected in this study. The class 1 RNase III signature motif at aa positions 36-44 in SPCSV isolates (aa 42-50 in KML33b) is boxed. Two aa sites predicted to be under positive selection are indicated with black shades. Names of isolates characterized from wild plants are in bold.

https://doi.org/10.1371/journal.pone.0081479.g001

Analysis of the RNA1 sequences using six different methods available in RDP v3.29 for detection of recombination breakpoints provided no evidence of recombination in any SPCSV isolate, as confirmed also with a PHI test (p = 0.999).

The nt sequences of SPCSV isolates characterized in this study were subjected to phylogenetic analysis, including the previously characterized WA strain isolate from Israel and five EA strain isolates from Uganda, Tanzania and Peru for comparison. Results on RNase3 sequences showed that all SPCSV isolates from Uganda belonged to the EA strain (Figure 2), no matter whether infecting wild species or sweetpotato. Only the two SPCSV isolates (HN1 and A1GT) from Central America belonged to the WA strain and formed a separate main clade with the WA isolate Is from Israel (Figure 2). Phylogenetic analysis of the p7 gene sequences resulted in a similar grouping of isolates to the two SPCSV strains (Figure S1). Only few significant subclusters of two or three isolates were supported by high bootstrap values, and the subclusters differed depending on whether RNase3 or p7 was analyzed. Phylogenetic analysis based on the p22 gene sequences revealed no significant subclusters (data not shown).

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 2. Phylogenetic analysis of genes coding for RNase3 of Sweet potato chlorotic stunt virus and the corresponding sequence of an unknown related virus (KML33b) detected in this study.

The branch of KML33b is not fully depicted. Names of isolates characterized from wild plants are indicated in bold, whereas the ten SPCSV isolates lacking the p22 gene are indicated with a black triangle (▲). Numbers at branches represent bootstrap values of 1000 replicates. Only bootstrap values of ≥ 50% are shown. Scale indicates Kimura units in nucleotide substitutions per site [55].

https://doi.org/10.1371/journal.pone.0081479.g002

Irrespective of the gene used for analysis, there was no phylogenetic congruence of isolates with their host species, geographical origins in Uganda, and whether or not they encoded the p22 gene.

Nucleotide Diversity and Selection Pressure

Nt sequence identities of EA strain isolates ranged from 92.5 % (p7), 96.0% (p22) and 97.7% (RNase3) to 100 % depending on the gene analyzed. The aa sequence identities ranged from 93.1% (p7), 94.8% (p22) and 98.7% (Rnase3) to 100%. The p22 protein of isolate MBL16 (from I. acuminata) was distinguished from other isolates by 12 unique aa substitutions (Supporting information, Figure S2). The sequences of WA strain isolates from Honduras and Guatemala were identical to each other and nearly identical to the isolate Is; the nt (aa) sequence identities were 99.4% (98.7%) for RNase3 and 99.4% (100%) for p7. The EA and WA strain isolates showed nt (aa) sequence identities of 83.8-84.3% (81.6-82.5%) for RNase3 and 76.2-79.8% (60.7-66.7%) for p7.

Irrespective of the host species from which the SPCSV isolates were characterized, nt diversity (π) values for each of the three protein coding regions studied were relatively low being 2.2%, 4.3% and 0.71% for the RNase3, p7 and p22 gene, respectively. The non-synonymous nt diversity (π_a) was 1.2%, 4.1% and 0.4% for the RNase3, p7 and p22 gene, respectively, whereas the synonymous nt diversity (π_s) was 5.6%, 5.5%, and 1.6%, respectively. The value of π_a was 4.7 or 4.0 fold lower than the value of π_s for the RNase3 and p22 gene, respectively. The value of π_a (4.1%) was slightly lower than π_s (5.5%) for the p7 protein. The selection pressures suggested by the aforementioned results were assessed by a maximum likelihood framework of codon substitution under model M0 which yielded ω values of 0.268, 0.562 and 0.324 for the RNase3, p7 and p22 genes, respectively, indicating purifying selection (Table 2). Heterogeneity of selective pressure was revealed by a test using an M3 vs. M0 LRT which showed that M3 fitted the data significantly better than M0 (Table 2) for all proteins. M3 for RNase3 suggested that 55.5% of sites were evolving under strong purifying selection (ω = 0.072), 43.9% of sites under weak purifying selection (ω = 0.513) and only 0.6% of sites under positive selection (ω = 7.781) (Table 2). The proportions of different site classes under M3 for the p7 and p22 genes are shown in Table 2.

LRTs of nested models M2a vs. M1a showed that M2a was a better fit than M1a for RNase3, p7 and p22 proteins (Table 2) which is consistent with positive selection on some aa sites. Parameter estimates under M2a for the RNase3 showed a large proportion of sites (86.4%) under purifying selection (ω = 0.170), 13.1% under neutral evolution (ω = 1.000) and only 0.5% under positive selection (ω = 9.690). These results were verified in an independent LRT involving M8 vs. M7 in which M8 fitted the data for each gene better than M7 (Table 2). The Näive Empirical Bayes (NEB) inference provided similar results as BEB inference under M3, M2a and M8 (Table 2).

Interactions Between SPCSV and SPFMV from Wild Plants in Sweetpotato

Twenty-four of those 64 isolates of SPCSV, which were characterized at the molecular level, had been initially transmitted with whiteflies from the original 12 wild plants and 12 sweetpotato plants sampled from the field to healthy plants of sweetpotato cv. Tanzania (Table 1). These wild plants belonged to seven species (I. hederifolia, I. obscura, I. rubens, I. setosa, I. sinensis, Hewittia sublobata and Lepistemon owariensis) which were amenable for establishment in the screenhouse using cuttings and were hence feasible for the experiment. The whitefly-inoculated plants of cv. Tanzania infected with SPCSV were multiplied by rooting stem cuttings and the new plants were observed for symptoms in the greenhouse. All isolates induced similar symptoms of mild chlorosis in younger leaves and purpling symptoms on older leaves, as shown for isolates HOM76 and SOR71 in Figures 3B and 3D, respectively.

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 3. Symptoms and virus accumulation in single or double infections with Sweet potato chlorotic stunt virus (SPCSV) and Sweet potato feathery mottle virus (SPFMV) in sweetpotato plants of cv.

Tanzania. (A) A plant infected with SPFMV (isolate RUK73) shows no obvious virus symptoms 6 weeks post-inoculation. (B) Chlorosis and purpling of older leaves induced by SPCSV isolate HOM76. (C) Typical symptoms of sweetpotato virus disease (SPVD) including retarded growth, severe leaf strapping and puckering induced by co-infection with SPCSV isolate HOM76 and SPFMV. (D) Chlorosis and purpling of older leaves induced by SPCSV isolate SOR71, and (E) similar symptoms in a plant co-infected with SOR71 and SPFMV. Young leaves develop normally in (D) SOR71-infected plants and (E) plants co-infected with SOR71 and SPFMV. In both cases the plants display only mild chlorosis typical of SPCSV infection, which indicates that SOR71 is not able to induce SPVD in co-infection with SPFMV. (F) SPCSV RNA detected by dot blot hybridization with a digoxigenin-labelled RNA probe specific to the RNase3 gene in (a) plants infected with SPCSV alone or (b) plants co-infected with the SPCSV and SPFMV. The amounts of total plant RNA dotted on the membrane are indicated. The SPCSV isolates tested were 1, KTK39; 2, KTK40; 3, KTK41; 4, SOR71; 5, MAS69; 6, SET5; 7, MSK7; 8, TOR16; and 9, HOM76. Isolates 1 to 4 are lacking the p22 gene. Note that SOR71 (4a and 4b) accumulates at very low concentrations in sweetpotato leaves and is barely detectable. H, non-inoculated healthy plant of cv. Tanzania. (G) SPFMV RNA detected by dot blot hybridization with a digoxigenin-labelled RNA probe specific to the CP-encoding region. Samples 1 to 9 are those co-infected with SPCSV and SPFMV and tested for SPCSV (i.e., samples 1b to 9b) in (F). Three additional samples co-infected with SPCSV (10, MAS46; 11, TOR14; and 12, MPG88) and SPFMV were included. Note that SOR71 synergises SPFMV, which is detected by the enhanced SPFMV concentrations (sample 4) as compared to the samples (FM) from cv. Tanzania infected with SPFMV only. H, non-inoculated healthy plant of cv. Tanzania.

https://doi.org/10.1371/journal.pone.0081479.g003

Subsequently, new cuttings were graft-inoculated with SPFMV isolate RUK73 (obtained from a plant of I. obscura in Uganda; [29]) in two experiments. All SPCSV isolates (except SOR71) induced the characteristic symptoms of SPVD including small, crinkled and strap-like leaves and severe retardation of plant growth in co-infected plants, as exemplified with coinfection of HOM76 and SPFMV in Figure 3C. Furthermore, three tested isolates lacking the p22 gene (KTK39, KTK40 and KTK41) induced similar symptoms of SPVD as those 20 isolates which contained p22.

Accumulation of SPCSV and SPFMV RNA was compared between the plants and treatments by dot blot hybridization using different amounts of total RNA extracted from the fully-grown upper leaves of cv. Tanzania. The viral RNA was detected with DIG-labelled RNA probes specific to RNase3 and the CP-encoding region of SPFMV. The titres of SPCSV RNA1 in the leaves of plants co-infected with SPCSV and SPFMV were similar or slightly higher than those in plants infected with SPCSV alone (Figure 3F). The low concentrations of SPFMV RNA were barely detectable in plants of cv. Tanzania infected with SPFMV alone, but were greatly enhanced following co-infection with SPCSV, irrespective of the presence or absence of p22 in the SPCSV isolate (Figure 3G). The relative differences in viral RNA concentrations were reproducible in the two experiments.

Suppression of RNA silencing is supposed to increase virus titers in infected tissues, but virus concentrations were enhanced also in absence of p22 that is a strong suppressor of RNA silencing in SPCSV isolate Ug [20]. The p22 proteins of two new SPCSV isolates (ARU59 from I. stenobasis and HOM89 from sweetpotato) were tested for their ability to suppress gene silencing, including p22 of isolate Ug for comparison. Silencing of the jellyfish (Aequorea victoria) green fluorescent protein (GFP) gene expression was induced by double-stranded (hairpin) RNA in leaves of Nicotiana benthamiana using an Agrobacterium tumefaciens infiltration assay, as described [31]. All three p22 proteins supported GFP expression, i.e., suppressed gfp silencing, in contrast to expression of ß-glucuronidae (GUS) included as a negative control (Figure 4). These results were consistent with previous studies indicating that suppression of RNA silencing by p22 is dispensable for development of SPVD or enhancement of the titers of co-infecting viruses unrelated to SPCSV [21].

Download:

• PPT
  PowerPoint slide
• PNG
  larger image
• TIFF
  original image

Figure 4. Suppression of RNA silencing by the p22 and RNase3 proteins of SPCSV isolates and the RNase3-like protein of the new virus KML33b.

Upper row of leaves: “Silencing on the spot” to induce “strong silencing” of the gfp gene for green fuorescent protein (GFP) was achieved by co-expressing gfp from one A. tumefaciens strain and double-stranded (hairpin) RNA homologous to gfp from another strain in coinfiltrated leaf tissue of Nicotiana benthamiana, and coinfiltration of a third strain expressing p22 protein to suppress gfp silencing. The p22 proteins of isolates ARU59 (I. sinensis) and HOM89 (sweetpotato) were compared with the previously characterized p22 protein of isolate Ug by expressing them at the opposite sites of the midrib in the same leaf. An Agrobacterium strain expressing ß-glucuronidae (GUS) was included as the negative control. Leaves were illuminated with UV light and photographed from the underside with a digital camera 3 days postinfiltration. Lower row of leaves: Cosuppression of gfp in transgenic N. benthamiana plants (line 16c) constitutively expressing gfp (note the green fluorescence in veins). The spots were co-infiltrated with a mixture of two Agrobacterium strains, one expressing gfp to achieve cosuppression (silencing) of gfp and another expressing RNase3 of isolate SOR71 (I. obscura), Ug, or the RNase3-like protein of the new virus KML33b (I. sinensis). Leaves were illuminated with UV light and photographed from the underside with a digital camera 6 days postinfiltration.

https://doi.org/10.1371/journal.pone.0081479.g004

SPCSV isolate SOR71 was an exception, because the titers of this isolate both in single infection and double infection with SPFMV were too low to generate detectable signals within the time of exposure used (Figure 3F, samples 4a and 4b). Nevertheless, SOR71 was readily detected in these plants by RT-PCR and sequencing of the PCR products (data not shown). The plants co-infected with SOR71 and SPFMV (Figure 3E) did not develop SPVD symptoms, but symptoms remained mild and similar to those caused by SOR71 alone (Figure 3D). However, SPFMV titres were enhanced to the same extent as with other SPCSV isolates (Figure 3G, sample 4).

Due to all these unexpected results on SOR71 (which lacks p22), RNase3 of SOR71 was tested for its ability to suppress RNA silencing using the aforementioned agroinfiltration assay. However, because RNase3 suppresses only sense-mediated gene silencing [21], i.e., caused cosuppression or “weak gfp silencing” [31], the gfp gene was overexpressed by agroinfiltration in leaves of transgenic N. benthamiana constitutively expressing gfp (line 16c) [32] and RNase3 of SOR71 was coexpressed in the same leaf tissue from another A. tumefaciens strain. RNase3 of the previously characterized SPCSV isolate Ug [20] and ß-glucuronidase (GUS) were included as a positive and a negative control, respectively. Results of the two experiments showed consistently that RNase3 of SOR71 and Ug supported expression of GFP fluorescence and suppressed sense-mediated silencing with similar efficiency (Figure 4).

A New Putative Crinivirus encodes an RNase3-like Suppressor of Silencing

One plant of I. sinensis was co-infected with two distinct virus isolates, which were designed as KML33a and KML33b (Table 1). KML33a was nearly identical to other EA strain isolates of SPCSV (Figures 1 and 2). In contrast, the nt (aa) sequence of RNase3 in KML33b was only 58.1-59.5% (53.0-55.6 %) identical to the EA strain isolates of SPCSV and 56.3% (50.9-51.3%) identical to the WA strain isolates (Figures 1 and 2). Furthermore, the RNase3 of KML33b was predicted to contain six additional aa residues at the N-terminus and two additional aa residues at the C-terminus, compared to that of SPCSV (Figure 1). The ORF for p7 was of identical length in KML33b and the SPCSV isolates, and the level of sequence similarity of p7 in KML33b and SPCSV isolates was comparable with the similarity of RNase3 nt and aa sequences. KML33b contained no ORF for p22. The corresponding gap in the genomic sequence of RNA1 started at an identical position but was 20 nt shorter as compared with SPCSV isolates. The low identity of KML33b to SPCSV and other viral sequences available in databases suggested that KML33b is a new virus related to viruses in genus Crinivirus.

The RNase3-like protein of KML33b was tested for its ability to suppress sense-mediated gene silencing (cosuppression), as described above. Leaves of the gfp-transgenic N. benthamiana line 16c were coinfiltrated with Agrobacterium strains for expression of gfp and KML33b RNase3 or GUS. Silencing was suppressed by KML33b RNase3, allowing continued GFP expression, whereas GFP fluorescence faded out by 6 days postinfiltration following infiltration with GUS (Figure 4). Hence, the KML33b RNase3 protein was able to suppress silencing.

Collection of Virus Isolates

Wild plants of family Convolvulacea were collected in 2007, identified and maintained as described by Tugume et al. [28,29,35]. Samples were collected from private land with the owner’s permission. For any other locations, no specific permissions were required for sampling. The field studies did not involve endangered or protected species.

Isolates of SPCSV from 13 wild plants and 12 plants of cultivated sweetpotato were whitefly-transmitted to healthy sweetpotato cv. Tanzania at MUARIK. Whiteflies (Bemisia tabaci Gennadius) were raised and maintained on cassava (Manihot esculenta Crantz) plants in insect-proof cages for three weeks. Single source plants infected with SPCSV were placed into insect-proof cages and 70-100 adult whiteflies were transferred to each. After a 36-h access period, a 3-week old healthy sweetpotato plant cv. Tanzania (raised from a cutting) was added into the cage for 72 h to be inoculated by the whiteflies and then treated with pymetrozine (Fulfill; Syngenta, Greensboro, NC) to kill the whiteflies. The inoculated sweetpotato plants were tested for SPCSV after 3 wk. Stem cuttings were taken and transferred to University of Helsinki, Finland, where they were grown in an insect-proof greenhouse (temperature 25-30 °C, relative humidity 70%) under natural daylight extended to 16h by illumination with high-pressure sodium halide lamps (light intensity of 150-200 μmol s^-1m^-2).

Serological and Molecular Detection of SPCSV

Triple antibody sandwich enzyme-linked immunosorbent assay (TAS-ELISA) was used for serological detection of SPCSV as described [46]. Polyclonal and monoclonal antibodies to SPCSV were provided by the International Potato Centre (Lima, Peru) and Julius Kühn Institute (Braunschweig, Germany), respectively. Testing was repeated on the plants established in the screenhouse. The SPCSV in wild plants was graft-transmitted to healthy sweetpotato plants of cv. Tanzania for ease of maintenance and study. The presence of SPCSV in plants grown from the tuberous roots obtained from Central America was tested using molecular methods.

Total RNA was extracted using TRIzol Reagent (InVitrogen Ltd, UK) according to the manufacturer’s instructions from 200 mg of leaf tissue. First-strand cDNA was synthesized on 3 µg of total RNA with random primers and Moloney murine leukemia virus (M-MuLV) reverse transcriptase RNase H^- (Finnzymes Oy, Finland) according to the manufacturer’s instructions. The 3’ genomic region of the RNA1 of SPCSV was amplified by polymerase chain reaction (PCR) using a forward CSVR3-F2 (5’-GTGTTTCATACATTGTTTGTGTGCT-3’) and a reverse CSVp22-R2 (5’-AGGTGTATGACTCTAGGGTATAAAC-3’) primers designed in this study and which were complementary to the genomic positions 7491-7515 and 9241-9265, respectively, of the SPCSV EA strain RNA1 (AJ428554; Kreuze et al., 2002). The primers PolymF4b (5’-CGAATACGTGGTTGTCAA-3’) and WA-R1 (5’-CTCTAGGATACAAACATTAATC-3’) complementary to genomic positions 7046-7063 and 1270-1291 of SPCSV RNA1 accessions AJ428554 and EU124491, respectively, were used to amplify the 3’ genomic region of two SPCSV isolates from Central America. The PCR mixture and cycling parameters were those recommended for Phusion High-Fidelity DNA Polymerase (Finnzymes Oy, Finland).

Sequence Analyses

The PCR products were purified using a combination of Exonuclease I and Calf Intestine Alkaline Phosphatase (Fermentas) as recommended [47]. The purified PCR products were sequenced directly with the Big Dye Terminator kit version 3.1 on an ABI automatic 3130 XL Genetic Analyzer at Haartman Institute, University of Helsinki, using the PCR primers. Sequencing was done in both directions for each of the two independent amplicons of each isolate. The sequences were compared by BLAST search with the existing sequences available in the NCBI database.

Nucleotide (nt) sequences were aligned using CLUSTALX version 1.83 [48], and examined visually. The ORFs for the RNase3, p7 and p22 genes of SPCSV were translated into amino acid (aa) sequences using the EMBOSS tool (http://www.ebi.ac.uk/emboss/transeq/index.html?). Percent nt and aa identities between sequences were computed using the CLUSTALW procedure [49] as implemented in MEGALIGN program of the DNASTAR software package (DNASTAR Inc., Madison WI, USA).

Average nucleotide diversity (π), the number of nonsynonymous nt substitutions per nonsynonymous site (d_N) and the mean of dN values in all pairwise comparisons of the sequences of the gene (i.e., the nonsynonymous nucleotide diversity, symbolized π_a), and the number of synonymous nt substitutions per synonymous sites (d_S) and the mean of dS values in all pairwise comparisons of the sequences of the gene (i.e., the synonymous nucleotide diversity, symbolized π_s) for each of the genes RNase3, p7, and p22 were calculated using DnaSP version 5 [50]. The sequence data were analyzed for recombination breakpoints using RDP, Maximum Chi square, Bootscan, Chimaera, and Sister Scan (Recombination Detection Program RDP3 Package version 3.29; [51]) and verified using the Pairwise homoplasy index (PHI) test for recombination [52].

Phylogenetic relationships based on the RNase3, p7 and p22 genes were analyzed using the Neighbor-joining algorithm [53] as implemented in MEGA4 [54] using the Kimura two-parameter nucleotide substitution model [55]. Statistical significance of tree branching was tested by performing 1,000 bootstrap replications.

Analysis of Selection Pressures

Non-synonymous to synonymous nt substitution rate ratio (ω) which gives a measure of selection pressure at the protein level was assessed using maximum likelihood codon substitution models implemented in CODEML programme of PAML4 package [56]. Six site models including M0 (one-ratio), M1a (nearly neutral), M2a (positive selection), M3 (discrete), M7 (beta), and M8 (beta&ω) were employed as described [57-59]. Likelihood ratio tests (LRTs) of M3 vs. M0 were used to test the heterogeneity of selective constraints among codons, while two comparisons of M2a vs. M1a, and M8 vs. M7 were used to test for positive selection as described [59]. Where the LRTs suggested positive selection, the Bayes empirical Bayes (BEB) approach [58] was used to identify amino acids under positive selection.

Analysis of Interactions Between SPCSV and SPFMV in Cultivated Sweetpotato and Accumulation of Viruses

Scions of sweetpotato cv. Tanzania singly infected with SPFMV EA strain isolate RUK73 (FJ795761) isolated from I. obscura in Uganda [29] were grafted onto 20 sweetpotato plants of cv. Tanzania raised from cuttings taken from the rootstocks that had been inoculated with different isolates of SPCSV via whitefly transmission. The grafted plants were maintained as above and observed for SPVD symptoms.

Concentrations of SPFMV and SPCSV (RNA1) were estimated by dot blot hybridization analysis, using digoxigenin (DIG) labeled RNA probes for the CP encoding region of SPFMV isolate RUK73 or the RNase3 gene of SPCSV isolate HOM76 (Table 1), respectively. The RNA probes were synthesized by in vitro transcription of the cloned DNA fragments encoding these genes as recommended in the DIG Application Manual for Filter Hybridization (Roche, Mannheim, Germany). Total RNA was extracted from the sixth leaf below the uppermost fully expanded leaf at 6 wk after planting the SPCSV-infected cuttings or 6 wk after top graft-inoculation of SPFMV onto SPCSV-infected rootstock. The quantity and quality of the RNA were measured using a spectrophotometer and the concentrations of RNA samples were equalized. A dilution series was prepared from total plant RNA starting from 2000 ng/μL and 5 µl from each RNA dilution was dotted on positively charged nylon membrane (Hybond-N+, Amersham Bioscience Ltd, UK) and detected using the digoxigenin-labelled RNA probes following the procedures described in the DIG Application Manual for Filter Hybridization (Roche).

Agroinfiltration Assays on Suppression of RNA Silencing

The experiments testing suppression of RNA silencing by p22 and RNase3 were carried out essentially as described by Kreuze et al. [20] and Cuéllar et al. [21]. Agrobacterium tumefaciens (strain C58c1; pGV3850) was transformed with binary vectors (pA35Shp200) each expressing one of the proteins under Cauliflower mosaic virus 35S promoter, and the 5’ untranslated region (5’UTR) of Potato virus A (PVA) was used for translation enhancement [20]. GUS was included as a negative control. Leaves of N. benthamiana were coinfiltrated with Agrobacterium strains for expression of the gfp gene for green fluorescent protein (GFP), gfp-specific double-stranded (hairpin) RNA to induce “strong gfp silencing” [31], and a third construct to express the putative silencing suppressor, as described previously [20]. Alternatively, leaves of transgenic N. benthamiana plants (line 16c) constitutively expressing gfp [32] were co-infiltrated with a mixture of two Agrobacterium strains, one expressing gfp to achieve cosuppression (silencing) of gfp and another expressing the putative silencing suppressor protein. Cultures of A. tumefaciens were diluted to OD₆₀₀=0.25 or OD₆₀₀=0.50 for the two types of experiments, respectively. Plants were grown from seeds in growth rooms under controlled conditions (16 h photoperiod at 18/22 °C, 70% relative humidity; 200 μE m^–2s^–1 light intensity) and given fertiliser (0.3 g/l) (16:9:22 N:P:K; Yara, Espoo, Finland) at every watering. The first 3-4 fully expanded leaves were agroinfiltrated.