• Open Access

Engineering cherry rootstocks with resistance to Prunus necrotic ring spot virus through RNAi-mediated silencing


  • Guo-qing Song,

    Corresponding author
    1. Department of Horticulture, Michigan State University, East Lansing, MI, USA
    • Department of Horticulture, Plant Biotechnology Resource and Outreach Center, Michigan State University, East Lansing, MI, USA
    Search for more papers by this author
  • Kenneth C. Sink,

    1. Department of Horticulture, Plant Biotechnology Resource and Outreach Center, Michigan State University, East Lansing, MI, USA
    2. Department of Horticulture, Michigan State University, East Lansing, MI, USA
    Search for more papers by this author
  • Aaron E. Walworth,

    1. Department of Horticulture, Plant Biotechnology Resource and Outreach Center, Michigan State University, East Lansing, MI, USA
    2. Department of Horticulture, Michigan State University, East Lansing, MI, USA
    Search for more papers by this author
  • Meridith A. Cook,

    1. Department of Plant Biology, Michigan State University, East Lansing, MI, USA
    Current affiliation:
    1. Cargill Specialty Seeds & Oils, Fort Collins, CO, USA
    Search for more papers by this author
  • Richard F. Allison,

    1. Department of Plant Biology, Michigan State University, East Lansing, MI, USA
    Search for more papers by this author
  • Gregory A. Lang

    1. Department of Horticulture, Michigan State University, East Lansing, MI, USA
    Search for more papers by this author

Correspondence (Tel: 517 355 5191 x 1384; fax: 517 432 8853; email: songg@msu.edu)


Prunus necrotic ringspot virus (PNRSV) is a major pollen-disseminated ilarvirus that adversely affects many Prunus species. In this study, an RNA interference (RNAi) vector pART27–PNRSV containing an inverted repeat (IR) region of PNRSV was transformed into two hybrid (triploid) cherry rootstocks, ‘Gisela 6’ (GI 148-1) and ‘Gisela 7’(GI 148-8)’, which are tolerant and sensitive, respectively, to PNRSV infection. One year after inoculation with PNRSV plus Prune Dwarf Virus, nontransgenic ‘Gisela 6’ exhibited no symptoms but a significant PNRSV titre, while the transgenic ‘Gisela 6’ had no symptoms and minimal PNRSV titre. The nontransgenic ‘Gisela 7’ trees died, while the transgenic ‘Gisela 7’ trees survived. These results demonstrate the RNAi strategy is useful for developing viral resistance in fruit rootstocks, and such transgenic rootstocks may have potential to enhance production of standard, nongenetically modified fruit varieties while avoiding concerns about transgene flow and exogenous protein production that are inherent for transformed fruiting genotypes.


The most successful instances of creating viral resistance in fruit crops, to date, are limited to papaya ringspot virus (PRSV) in papaya (Carica papaya L.) and plum pox virus (PPV) in plum (Prunus domestica L.) (Scorza and Ravelonandro, 2006; Souza et al., 2005). In the former case, transformed genotypes have been introduced directly into commercial production following deregulation by APHIS in 1996, with acceptance into the Japanese market occurring as recently as 2011. In the latter case, transformed genotypes were used for conventional breeding of new potential varieties, a long and expensive process, which has resulted in APHIS deregulation in 2010 for the first such introduction, ‘Honeysweet’ plum. However, food and environmental safety of genetically modified (GM) fruit crops are two major public concerns. Commercial tree fruit production usually depends on a composite plant, comprised of a commercial fruiting genotype grafted onto a separate rootstock genotype. The potential use of a transgenic rootstock that can be grafted to standard nontransgenic commercial scion varieties to produce nongenetically modified fruits is one of many potential strategies available to address the concerns about transgene flow and exogenous protein production that are associated with the fruit of transformed scion cultivars.

Like most important Prunus crops, sweet (Prunus avium L.) and sour (P. cerasus L.) cherries are susceptible to a number of potentially pathogenic viruses, including Prunus necrotic ringspot virus (PNRSV), which can be transmitted by grafting, pruning, pollen, sucking insects like aphids and leaf hoppers, and even microscopic soilborne pests like nematodes (Barbara et al., 1978). Sweet cherry scions generally tolerate PNRSV infection with no or only mild symptoms, but certain viral isolates can cause rugose mosaic disease, which affects fruit quality, ripening time and tree health. Cherry rootstock genotypes range from hypersensitive to tolerant for some of these common viruses such as PNRSV (Lang et al., 1998). However, even if a cherry scion genotype tolerates infection by a mild strain of PNRSV, when the virus reaches the graft union with a sensitive rootstock genotype, a hypersensitive reaction will occur that can cause vascular necrosis at the union, resulting in tree decline and ultimately death.

Conventional breeding to achieve virus-resistant cherry cultivars is uncommon, difficult, and time-consuming due to heterozygosity, juvenility and a lack of natural sources of resistance. Genetic engineering can facilitate the incorporation of single or multiple genes into existing sweet cherry genotypes, but concerns about transgene flow and exogenous protein production in the fruit of transformed scion cultivars has hindered such approaches. Additionally, the genetic transformation of cherry, like many woody perennial fruit crops, currently is difficult and far from routine, and the challenge of one-by-one transformation of the dozens of commercially important varieties would be an extremely time-consuming and expensive undertaking (Cheong, 2012; Song et al., 2008). Of more than 30 cherry species, genetic transformation has been reported only for a few commercially important genotypes, including sour cherry, chokecherry (Prunus virginiana L.), black cherry (Prunus serotina Ehrh.) and the cherry rootstocks ‘Rosa’ (P. subhirtella autumno), ‘Gisela 6’ (P. cerasus × P. canescens), ‘Colt’ (P. avium × P. pseudocerasus), ‘Inmil’ (P. incisa × P. serrula) and ‘Damil’ (P. dawyckensis) (Da Cämara Machado et al., 1995; Dai et al., 2007; Dolgov and Firsov, 1999; Druart et al., 1998; Gutièrrez-Pesce and Rugini, 2004; Gutièrrez-Pesce et al., 1998; Liu and Pijut, 2010; Song and Sink, 2005, 2006, 2007).

The most promising avenue for improving plant resistance to viruses involves the strategy of post-transcriptional gene silencing, using RNA interference (RNAi) to prevent the virus from functioning normally in replication and pathogenicity (Mlotshwa et al., 2002; Baulcombe, 2004; Kusaba, 2004; Matzke and Birchler, 2005; Brodersen and Voinnet, 2006). This type of gene silencing has been reported to be transmitted systemically, across the graft union into nontransformed tobacco plants grafted onto transformed plants (Palauqui et al., 1997). Double-stranded RNA (dsRNA) is generally cleaved in plants by the cellular machinery into short interfering RNAs (siRNAs), which are efficient inducers of gene silencing (Fusaro et al., 2006; Kerschen et al., 2004; Li et al., 2008; Szankowski et al., 2009; Wagner et al., 2005). To generate artificial dsRNAs through transgenes, constructs have been designed to express small hairpin RNA (hpRNA) (Hily et al., 2007; Kusaba, 2004). The hpRNA-mediated RNAi in plants has been shown to engender long-distance silencing signals and acts through the viral defense pathway in plants (Fusaro et al., 2006; Hättasch et al., 2009; Ryabov et al., 2004; Tournier et al., 2006; Yaegashi et al., 2008). As the first stage of a proof-of-concept study to determine whether transformed rootstocks can be used to obtain viral resistance in scion cultivars, we report here on the transformation of the hybrid cherry rootstocks ‘Gisela 6’ (tolerant to PNRSV) and ‘Gisela 7’ (sensitive to PNRSV) with two RNAi vectors pTRAP6i and pART27–PNRSV, each containing inverted repeat (IR) regions of a PNRSV sequence (Figure 1a). ‘Gisela 6’ and ‘Gisela 7’ are triploids, thereby further minimizing any potential for inadvertent transgene flow.

Figure 1.

Transformation of cherry rootstocks using pART27-PNRSV. (a) Schematic representation of the T-DNA region of the binary vectors pART27–PNRSV and pTRAP6i. En35S = the cauliflower mosaic virus (CaMV) 35S promoter plus an enhancer; Ocs = Ocs terminator; nptII = neomycin phosphotransferase gene; 35S = CaMV 35S promoter; T = nos terminator; PNRSV, Prunus necrotic ringspot virus; TRSV, Tobacco ringspot virus; PDV, Prune dwarf virus; PPV, Plum pox virus; PMV, Peach mosaic virus; and APLPV, American plum line pattern virus; (b) Shoot regeneration from leaf explants of ‘Gisela 6’; (c) Proliferation of a putative transgenic ‘Gisela 7’ event on the SC medium containing 100 mg/L Km. (d) Rooting of transgenic plants. (e) Growing of transgenic plants. The white arrows are showing three procumbent plants derived from one transgenic event of ‘Gisela 6’.

Materials and methods

Plant materials

Shoot cultures of the rootstock cultivars ‘Gisela 6’ and ‘Gisela 7’ were maintained and subcultured at 6 weeks intervals on the stock culture medium [SCM: Murashige and Skoog (1962) medium (MS) +1.0 mg/L benzylaminopurine (BA) +0.1 mg/L indole–3–butyric acid (IBA) + 2% (w/v) sucrose]. The regeneration medium (RM) for both cvs. was that of Woody Plant Medium (WPM: Lloyd and McCown, 1980) supplemented with 2.0 mg/L BA, 1.0 mg/L IBA and 3% (w/v) sucrose. All solidified media used in this study contained 0.6% (w/v) Bacto–agar (Becton, Dickinson and Co., Sparks, MD). Media pH was adjusted to 5.2 before autoclaving (121 °C, 20 min at 105 kPa). Acetosyringone (AS) and all antibiotics were filter-sterilized (0.22 μm) and added to media cooled to 50–60 °C after autoclaving. All in vitro plant tissues were cultured at 25 °C under a 16-h photoperiod of 40 μmol/m2/s from cool white fluorescent tubes except as otherwise noted.

RNAi constructs and Agrobacterium strain

Two hairpin RNA (hpRNA) constructs (binary vectors), pTRAP6i and pART27–PNRSV, were used for plant transformation (Figure 1a). The pTRAP6i vector was from Dr. Zongrang Liu, USDA–ARS, Kearneysville, West Virginia. A 2.5-kb IR of the coding sequence targeted at six Prunus-related viruses was separated by a 1.0-kb intron. The pART27–PNRSV hairpin vector for PNRSV silencing contains two IR sequences and was constructed by amplifying 414-bp from the coat protein cDNA of Michigan PNRSV isolate CB7 (GenBank Accession number EF495168) using primers containing restriction enzyme sites for cloning. The primers for the sense IR, containing XhoI and KpnI sites, were Sense–F: 5′ – tttttctcgagaatggtttgccgaatttgc – 3′ and Sense–R: 5′ – ttttttggtacctagtcctccaccatccca – 3′. The primers for antisense IR, containing XbaI and ClaI sites, were Antisense–F: 5′ – tttttttctagaaatggtttgccgaatttgc – 3′ and Antisense–R: 5′ – ttttttatcgattagtcctccaccatccca – 3′. The PCR products of both sense IR and antisense IR primers were digested accordingly and cloned into pKANNIBAL (Wesley et al., 2001) (GenBank Accession number AJ311873) in an inverted repeat orientation interrupted by a 700-bp fragment. The PNRSV hpRNAi cassette driven by the CaMV 35S promoter in the pKANNIBAL was released by a digestion with NotI and cloned into binary vector pART27 (Gleave, 1992), resulting in the pART27–PNRSV. Both pTRAP6i and pART27–PNRSV hairpin constructs were transferred into Agrobacterium tumefaciens strain EHA105 (Hood et al., 1993) using the freeze-thaw method (An et al., 1988) and were used for transformation.


Transformation of ‘Gisela 6’ and ‘Gisela 7’ was carried out following our reported protocol (Song and Sink, 2006). Briefly, uniform size leaves with midribs, 1.5–2.0 cm in length, were harvested from 6-week-old stock cultures, each cut transversely and equidistant four times through the midrib and used as explants. Single colonies of EHA105 were cultured in 10 mL liquid YEB (Vervliet et al., 1975) +50 mg/L kanamycin (Km) at 28 °C in the dark for 48 h with rotary shaking (300 rpm). The cells were collected by centrifugation, 2 min, 2500 g, and resuspended to a final OD600 of 0.5 in liquid RM media with 100 μM AS. Fresh leaf explants were immersed in 10 mL bacterial suspension for 5 min at 28 °C, blotted dry on sterile filter paper and transferred onto a filter paper disc laid abaxial side down on solidified RM + 100 μM AS and incubated in the dark at 25 °C. After 4 days co-cultivation, leaf explants were rinsed three times (3 min each) in liquid RM, washed one time (3 min) in RM + 500 mg/L timentin to eliminate excess bacterial cells and then blotted dry on sterile filter paper. Subsequently, inoculated leaf explants, 10 per Petri dish, were placed abaxial side up on RM + 50 mg/L Km + 250 mg/L timentin. Selection was carried out in the dark at 25 °C for 2 weeks, after which explants were transferred onto fresh selection medium and cultured under a 16-h photoperiod of 40 μmol/m2/s. Subculture on fresh selection medium was performed at 3-week intervals. The number of explants producing Km-resistant calluses and/or shoots was recorded after 12 weeks of selection. Km-resistant shoots from an individual explant of each cv. were excised and cultured separately on SCM containing 50 mg/L Km + 250 mg/L timentin as 50 mL in each Magenta GA7 (Magenta Corp., Chicago, IL).

Km-resistant shoots, 2–3 cm in length, were excised and inserted into water soaked Suremix Perlite planting medium (Michigan Grower Products Inc., Galesburg, MI) in a 48-cell tray. The cell tray was covered with a clear plastic cover for 4 weeks and was then progressively opened and removed after 1 week. The young plants were grown in the greenhouse and they were repotted, watered, fertilized and cold-stored as needed.

PCR, Southern and Northern blot analyses

Total genomic DNA was isolated from young leaves of greenhouse-grown plants before viral inoculation, 0.5 g for each sample, using a cetyltrimethylammonium bromide (CTAB) method (Doyle and Doyle, 1987). RNA extraction from one gram of young leaf tissues was performed following a CTAB–based protocol (Zamboni et al., 2008). Two primer pairs, one for the 414-bp PNRSV fragment (CP–F: 5′ –aatggtttgccgaatttgc– 3′ and CP–R: 5′ –tagtcctccaccatccca– 3′) and another for a 600-bp fragment of the nptII coding region (nptII–F: 5′–GAGGCTATTCGGCTATGACTG–3′ and nptII–R: 5′–ATCGGGAGCGGCGATACCGTA–3′), were used for the primary screening of the transformants.

Southern analysis was performed on selected transgenic lines to confirm stable integration of the transgenes. Twenty micrograms of DNA was digested with XbaI (New England Biolabs, Ipswich, MA), separated on a 1% agarose gel and transferred to a Hybond N + nylon membrane (Amersham, Arlington Heights, IL). Northern blot (polyacrylamide gel) was carried out according to the protocol developed by Dr. Zongrang Liu (personal communication). Fifty micrograms of RNA was used for each sample. The digoxigenin (DIG)-labelled 414-bp PNRSV fragment was used as probe for both southern and northern hybridization. The DIG High Prime DNA Labelling and Detection Starter Kit II (Roche Applied Science, Indianapolis, IN) was used for probe synthesis, hybridization and detection.

Bark–graft for viral inoculation

A PNRSV strain Fulton G and a combined PNRSV/PDV (CH39) isolate were each stored separately in mature trees of ‘Bing’ sweet cherry on Prunus mahaleb (L.) seedling rootstocks. The infected 1-year-old budwood was obtained from the Clean Plant Center of the Northwest, Washington State University–IAREC, Prosser, Washington. Chips of bark were peeled off the virus-containing donor budwood and were grafted, one chip for each, onto receiver plants at a bark location about 25 cm above the soil in Spring of 2011. The grafted trees were grown in a secured greenhouse and evaluated periodically for visual symptoms of viral infection after successful grafting was confirmed.

Enzyme-linked immunosorbent assay (ELISA)

The presence of PNRSV was detected in Spring of 2012 by triple-antibody sandwich enzyme-linked immunosorbent assay (TAS–ELISA) using a commercial kit for the PNRSV detection according to the manufacturer's (Agdia Inc., Elkhart, IN) instructions. Samples were prepared for the assay by grinding 0.5 g of leaf tissue, collected from the lowest leaves above the infection graft, in 5 mL of general extract buffer (supplied with the kit). Results were evaluated visually or by measuring absorbance at 405 nm on a spectrophotometer (Shimadzu Corp., Kyoto, Japan).


Transformation and selection of regenerants

The length of the IR in the RNAi vectors affected transformation efficiency (Table 1). Our initial efforts focused on using the pTRAP6i construct with a long 2.5 kb IR, which was designed for silencing six Prunus-related viruses through siRNAs and has had limited success in producing virus-resistant Nicotiana benthamiana (Figure 1a). However, the first eight experiments using 1600 leaf explants of ‘Gisela 6’ did not yield any transgenic shoots under our previously reliable transformation conditions. We then constructed pART27–PNRSV that contains a shorter IR (414 bp). Three transformations for each of the pTRAP6i and the pART27–PNRSV constructs were subsequently conducted. For the pART27–PNRSV vector, transgenic shoots were obtained from leaf explants after 6 months; in contrast, no transgenic shoots were regenerated from leaf explants transformed with pTRAP6i (Figure 1; Table 1). These results suggest that pART27–PNRSV with a shorter IR is preferable to pTRAP6i containing a longer IR for the generation of transgenic cherry plants.

Table 1. Transformation of cherry rootstocks ‘Gisela 6’ and ‘Gisela 7’ using different RNAi constructs
VectorGenotypeNo. of leaf explants (Explants × No. of experiments)No. of Km-resistant shootsNo. of PCR-positive shoots
  1. PNRSV, Prunus necrotic ringspot virus.

pTRAP6i‘Gisela 6’200 × 11 = 220030
‘Gisela 7’100 × 3 = 30000
pART27-PNRSV‘Gisela 6’200 × 3 = 6002819
‘Gisela 7’100 × 3 = 30053

Growth of transgenic plants

In vitro shoots of all transgenic events were morphologically identical to nontransgenic shoots (Figure 1c). Over 90% of these shoots rooted directly in soil after 8 weeks and no obvious variation in morphology of the newly rooted plants was observed (Figure 1d). For plants grown in the greenhouse, 21 events (18 for ‘Gisela 6’ and three for ‘Gisela 7’) showed normal morphology. Interestingly, all transformants of one transgenic event of ‘Gisela 6’ were procumbent plants (Figure 1e). This morphological change did not appear until the plants were grown in the greenhouse.

Molecular analysis of putative transformants

Kanamycin (Km)-resistant shoots from separate explants were considered as individual transgenic events. Among proliferating shoots, 19 events for ‘Gisela 6’ and three for ‘Gisela 7’ were found to be PCR-positive using nptII and PNRSV primers (Table 1).

Southern analysis of selected pART27–PNRSV lines indicated single independent random insertion events (data not shown).

Northern blot analysis showed that the presence of expected siRNAs of PNRSV was detectable with very faint signals around the 19-nucleotide size marker in four of seven transgenic events of ‘Gisela 6’ and one of three transgenic events of ‘Gisela 7’; no detectable band appeared in any of the nontransgenic controls (data not shown). These results demonstrate that the hairpin RNAi strategy enabled production of expected siRNAs.

PNRSV resistance assay

Nontransgenic ‘Gisela 7’ was susceptible to both the PNRSV (Fulton G) and PNRSV/PDV (CH39). Of three nontransgenic plants inoculated with the CH39, all died within 1 year of inoculation (Figure 2a). In addition, the PNRSV symptoms were also observed on nontransgenic ‘Gisela 7’ one year after inoculation with the PNRSV (Fulton G) (Figure 2b). However, in spite of the small size of the transgenic ‘Gisela 7’ plant that was chosen (assuming it might die) for inoculation with CH39, all transgenic ‘Gisela 7’ were still healthy 1 year after inoculation. These plants showed no viral symptoms and were similar in foliage and health to both transgenic and nontransgenic ‘Gisela 7’ without viral inoculation (Figure 2a).

Figure 2.

(a) Responses of nontransgenic (wild type) ‘Gisela 7’ (G7–WT) and transgenic ‘Gisela 7’ with the pART27–PNRSV (G7–2) 1 year after inoculation with a combined PNRSV/PDV (CH39) isolate, and G7–2 not inoculated with CH39. Note: The starting plant in the middle is smaller than the other two because of the limited availability of equally sized G7-2 plants (so weaker plants were inoculated). (b) Responses of nontransgenic ‘Gisela 7’ (G7–WT) 1 year after inoculation with a PNRSV strain Fulton G. (c) ELISA assays for the detection of PNRSV in the plants infected with a combined isolate of PNRSV/PDV (CH39). Wild-type G7 = Non-transgenic ‘Gisela 7’; G7–2 = Transgenic ‘Gisela 7’ with the pART27–PNRSV; Positive Control = PNRSV-infected Chenopodium quinoa tissue supplied with ELISA kit (Agdia Inc). PNRSV, Prunus necrotic ringspot virus; PDV, Prune dwarf virus.

ELISA assays confirmed PNRSV resistance in transgenic ‘Gisela 7’ containing pART27–PNRSV. Yellow colour, indicative of the presence of the PNRSV, was observed in non-transgenic ‘Gisela 7’ inoculated with CH39 as well as in the positive PNRSV control (Figure 2c). In contrast, little to no yellow colour was detectable in the wells containing samples of either transgenic ‘Gisela 7’ inoculated with CH39 or noninoculated plants (Figure 2c). Therefore, after transformation with the RNAi vector pART27–PNRSV, ‘Gisela 7’ showed resistance to PNRSV.

‘Gisela 6’ is one of the most important new cherry rootstocks released in the past 20 years, due to its promotion of early and high productivity in commercial scion varieties. It has been characterized as being tolerant of infection by both PNRSV and Prune dwarf virus (PDV) (Lang et al., 1998). When we inoculated standard ‘Gisela 6’ plants with a combined source of PNRSV/PDV (CH39), our results were consistent with previous reports. Although few symptoms in some infected plants were observed relative to the noninoculated control, ELISA assays for PNRSV were positive with a dark yellow colour indicating a high viral titre and high PNRSV/PDV tolerance of ‘Gisela 6’ (Table 2).

Table 2. ELISA assays for detection of Prunus necrotic ringspot virus (PNRSV) in the plants infected with PNRSV Fulton or a combined PNRSV/Prune dwarf virus (PDV) (CH39)
Plant IDa Absorbance at 405 nmb
  1. a

    WT G6 = Nontransgenic ‘Gisela 6’; I5, II3, I10, III6, and III1 = Independent transgenic ‘Gisela 6’ with the pART27–PNRSV; Positive Control = PNRSV infected Chenopodium quinoa tissue supplied with ELISA kit (Agdia Inc.).

  2. b

    All reads are relative to WT G6 uninfected.

  3. c

    2.836 is the upper limit of the spec (100% absorbance at 405 nm).

Buffer only0   
Positive control2.836c   
WT G6 02.8362.836
I5 02.7110.957
I10 02.8361.288
II3 000
III1 02.8360
III6 000

ELISA assays confirmed reduced PNRSV replication compared with nontransgenic plants in transgenic pART27–PNRSV ‘Gisela 6’ events (Table 2). While uninfected controls of both transgenic and nontransgenic plants of ‘Gisela 6’, as well as the buffer controls, were negative for PNRSV in the ELISA assays, nontransgenic plants infected with either PNRSV (Fulton G) alone or a combined PNRSV/PDV (CH39) resulted in positive responses with dark yellow colour indicating high accumulation of viruses. In contrast, all plants of the five transgenic events tested, of which three had no visible yellow and the other two had lighter yellow colour compared with infected nontransgenic plants, showed resistance to Fulton; three of the five transgenic events showed resistance to the CH39. Overall, two of five transgenic events of ‘Gisela 6’, II3 and III6, showed high resistance to both Fulton and CH39. Thus, the tolerance (resistance) of the transgenic ‘Gisela 6’ plants appears to be due to inhibition of viral infection and/or replication, whereas the tolerance of standard ‘Gisela 6’ plants likely is not due to inhibition of PNRSV replication, because a high PNRSV titre was observed in the infected tissues even though the plants had no visible viral symptoms. Consequently, standard ‘Gisela 6’ plants can serve as potential symptomless carriers of PNRSV, while transgenic ‘Gisela 6’ plants would be unlikely to serve as sources of transmission.


In this study, we obtained transgenic ‘Gisela 6’ and ‘Gisela 7’ using a hairpin RNAi vector. This is the first report of transformed Prunus species containing an RNAi vector (Cheong, 2012).The size of IR region in RNAi vectors affected transformation efficacy. Using the pTRAP6i with a long 2.5 kb IR region targeting six major viruses for Prunus species, including PNRSV, Tobacco ringspot virus (TRSV), Prune dwarf virus (PDV), Plum pox virus (PPV), Peach mosaic virus (PMV) and American plum line pattern virus (APLPV) (Figure 1a), no transgenic plant was recovered in this study. The increased potential of formation of hairpin loops prior to integration of the T-DNA region might be responsible for the reduced transformation frequency of the long IR region RNAi vectors. This hypothesis was later supported by our transformation of petunia (Petunia hybrida ‘Mitchell’) using pTRAP6i, pART27–PNRSV, and a regular binary vector pGA643 (Song, G.-Q., unpublished data).

Internode sections of cherry have been reported to be attractive explants for regeneration and transformation, although they are not easily available from in vitro materials (Matt and Jehle, 2005). In our efforts to improve transformation frequency, we tried stem sections in this study (data not shown). Many shoots were produced in 4 months regardless of the RNAi vectors; however, none of these shoots survived when they were excised and subcultured separately on selection medium containing 100 mg/L Km for 6–8 weeks. So leaf explants were still preferable to internode sections for cherry transformation.

Theoretically, none of the transgene in pART27–PNRSV has a function to alter the morphology of recipient plants. Interestingly, all transformants of one transgenic event of ‘Gisela 6’ were procumbent plants (Figure 1e). This morphological change did not appear until the plants were grown in the greenhouse. Impact of insertion position of transgenes or tissue culture-induced somaclonal variation might be responsible for the mutation. Considering tree structure is important for fruit production, this mutant provides a unique material to study molecular mechanisms controlling plant structure through RNA-seq or identifying the insertion position(s).

Wild-type ‘Gisela 7’ was initially identified as a valuable rootstock for commercial release, but this action was rescinded when its sensitivity to PNRSV became apparent. Similar to the previous report (Lang et al., 1998), in this study nontransgenic ‘Gisela 7’ is susceptible to both PNRSV (Fulton G) and a combined PNRSV/PDV (CH39). After transformation with the RNAi vector pART27–PNRSV, transgenic ‘Gisela 7’ showed high resistance to PNRSV (Figure 2). With the resistance to PNRSV gained through transformation, ‘Gisela 7’ may become a commercially viable rootstock. In addition, our results provide solid proof-of-concept that RNAi-mediated gene silencing is an effective approach for engineering viral resistance in cherry. It is completely possible to engineer a ‘super rootstock’ with resistance to one or more viruses for woody plants using RNAi vectors such as pTRAP6i.

‘Gisela 6’ is one of the major cherry rootstocks. It was characterized as a rootstock with PDV and PNRSV-tolerance (Lang et al., 1998). Our results are consistent with the previous report. In addition, our ELISA assays for the presence of PNRSV in infected nontransgenic ‘Gisela 6’ suggested that the natural tolerance was not likely due to its inhibition of the growth of the PNRSV because high PNRSV densities were observed in the infected tissues at the time when the plants had no visible viral symptoms.

Both ‘Gisela 6’ and ‘Gisela 7’ were characterized as PDV tolerant (Lang et al., 1998). In this study, the CH39 infected plants gave more presence of PNRSV in both genotypes (Table 2, Figure 2), indicating that interaction of PDV and PNRSV could be responsible for the increased presence of the PNRSV.

Rootstocks are typically used for most tree fruit and nut crops to achieve higher yields, modulate tree vigour, or enhance tolerance to biotic or abiotic stresses. Genetically engineering rootstocks instead of scion cultivars may be a more efficient way to provide a specific production advantage across many standard commercial varieties as well as potentially being more acceptable with respect to the public's concern regarding GM fruits (Smolka et al., 2010).

Our results demonstrate the effectiveness of RNAi-mediated gene silencing for inducing viral resistance in cherry rootstocks. These resistant transgenic rootstocks provide excellent materials to explore the ideas of (i) using transgenic rootstocks to produce nongenetically modified commercial fruit cultivars, and (ii) using rootstock-derived siRNAs to prevent, reduce, or eliminate viral infection in scion cultivars grafted on the transgenic rootstock. Research is continuing to examine how sweet and sour scion cultivars grafted on these resistant rootstocks will respond to PNRSV infection.


As the first stage of a ‘proof-of-concept’ study to determine whether transformed rootstocks can be used to obtain viral-resistant scion cultivars, two hybrid cherry rootstocks ‘Gisela 6’ (tolerant to PNRSV) and ‘Gisela 7’ (sensitive to PNRSV) were transformed with two RNAi vectors, pTRAP6i from the USDA-Kearneysville containing fragments from six Prunus-related viruses and pART27–PNRSV. The smaller pART27–PNRSV vector yielded 19 independent transgenic events for ‘Gisela 6’ and three for ‘Gisela 7’. siRNA was detected by Northern blot. Subsequently, PNRSV-inoculation experiments confirmed the effectiveness of RNAi-mediated gene silencing in inducing viral resistance for ‘Gisela 6’ and ‘Gisela 7’.


We thank Dr. Zongrang Liu, USDA–ARS, Kearneysville, West Virginia, for kindly providing the pTRAP6i plasmid. We also thank Mrs. S. Qi for the assistance with Northern blot analysis and Michael Leasia for assisting with the data collection. This research was supported by MSU Project GREEEN (Generating Research and Extension to Meet Economic and Environmental Needs).