Nucleotide Sequence-Homology-Independent Breakdown of Transgenic Resistance by More Virulent Virus Strains and a Potential Solution

Controlling plant viruses by genetic engineering, including the globally important Papaya ringspot virus (PRSV), mainly involves coat protein (CP) gene mediated resistance via post-transcriptional gene silencing (PTGS). However, the breakdown of single- or double-virus resistance in CP-gene-transgenic papaya by more virulent PRSV strains has been noted in repeated field trials. Recombination analysis revealed that the gene silencing suppressor HC-Pro or CP of the virulent PRSV strain 5-19 is responsible for overcoming CP-transgenic resistance in a sequence-homology-independent manner. Transient expression assays using agro-infiltration in Nicotiana benthamiana plants indicated that 5-19 HC-Pro exhibits stronger PTGS suppression than the transgene donor strain. To disarm the suppressor from the virulent strain, transgenic papaya lines were generated carrying untranslatable 5-19 HC-Pro, which conferred complete resistance to 5-19 and other geographic PRSV strains. Our study suggested the potential risk of the emergence of more virulent virus strains, spurred by the deployment of CP-gene-transgenic crops, and provides a strategy to combat such strains.

of this fruit 17 . The lack of naturally resistant sources for use in conventional breeding, as well as the ineffectiveness of control measures such as agricultural practices for repelling aphid vectors and crossprotection from mild PRSV strains, necessitate the generation of transgenic resistance 18 . In Hawaii, Gonsalves and co-workers initiated the development of PRSV CP-transgenic papaya in the late 1980s. The CP-transgenic resistance of the resulting SunUp cultivar 17 is triggered by PTGS and confers a high degree of resistance to PRSV Hawaii strain HA 19 . In 1998, the SunUp and Rainbow papaya cultivars were deregulated and approved for commercial application in the United States 17 , representing the first successful case of the cultivation of a transgenic fruit tree. However, the high genetic divergence of CP genes from other geographic PRSV strains 17,18 limits the application of these two CP-transgenic cultivars in regions outside of Hawaii.
We previously used the CP coding sequence and contiguous 3'untranslated region (CP-3'UTR) of a native PRSV strain, YK, to generate PTGS-based transgenic resistance using the papaya cultivar Tainung No. 2 in Taiwan 20 . The resultant YK CP-3'UTR transgenic lines confer a high degree of broad-spectrum resistance to different geographic strains from Taiwan (YK), Hawaii (HA), Thailand (TH) and Mexico (MX) 21 . However, the infection of these YK CP-3'UTR transgenic papaya lines by a different potyvirus Papaya leaf-distortion mosaic virus (PLDMV) during field trials 22 necessitated the development of double-virus-resistant transgenic papaya lines that carry a chimeric untranslatable transgene comprising parts of the PLDMV CP and PRSV YK CP-3'UTR sequences 23 . Nevertheless, during repeated field trials over 4 years, both PRSV single-virus- 24 and PRSV1PLDMV double-virus-resistant lines were found to be susceptible to a more virulent strain, PRSV 5-19 23 . The emergence of more virulent virus strains represents a serious threat to the application of CP-transgenic papaya cultivars resistant to either PRSV or PRSV1PLDMV.
The YK CP-3'UTR transgenic papaya lines exhibit broad-spectrum resistance to PRSV strains whose CPs share .89% nucleotide (nt) sequence identity 21,24 with the transgene, including the different geographic strains TH (Thailand), HA (Hawaii) and MX (Mexico) 21,24 . However, PRSV 5-19 is able to overcome the singleor double-virus transgenic resistance despite the significantly higher nucleotide identity of its CP/3'UTR compared with YK (96/98%) 24 , implying that the resistance-breaking ability of PRSV 5-19 is not correlated with the level of sequence divergence between its CP-3'UTR and that of the transgene. Previously, Mäki-Valkama 25 reported that although potato lines containing the P1 gene of PVY O (a common strain) are resistant to PVY O , they are not resistant to PVY N (a necrotic strain), despite complete sequence identity between the P1 genes of PVY O and PVY N . Such anomalous, homology-independent breakdown of PTGS-mediated transgenic resistance by a more virulent strain suggests the involvement of viral elements that suppress PTGS-mediated resistance and may shut off the RNA silencing mechanism.
In this study, the mechanism of the homology-independent breakdown of CP-transgenic resistance by virulent strain 5-19 was investigated by analysing the reactions of the PRSV single-virus-and PRSV1PLDMV double-virus-resistant lines to the transgene donor strain YK, the resistance-breaking strain 5-19, and their recombinants. The results of the recombinant analysis indicated that 5-19 HC-Pro, a PTGS suppressor 15,26 , is responsible for the resistancebreaking ability of [5][6][7][8][9][10][11][12][13][14][15][16][17][18][19]. Gene silencing suppression assays performed using transient expression in Nicotiana benthamiana verified that 5-19 HC-Pro exerts strong gene-silencing suppression than YK HC-Pro. Thus, the virulent strain 5-19 may have emerged in conventional plants and been inoculated into transgenic trees via aphid-mediated transmission. The stronger ability of this strain to suppress PTGS allows it to overcome CP-transgenic resistance in a sequence-homology-independent manner. To counter the threat of the 5-19 strain, transgenic papaya lines carrying an untranslatable 5-19 HC-Pro construct were developed. The resulting HC-Pro-transgenic papaya lines are resistant to the virulent strain 5-19, as well as to other geographic PRSV strains, suggesting that this approach may solve the problem of the emergence of more virulent strains from CP-transgenic lines.

Results
CP-transgenic resistance can be overcome in a nucleotide sequence-homology-independent manner. RNA viruses can be recombined at the cDNA level in the laboratory. Using an infectious, full-length in vitro YK cDNA clone as a backbone 27 , which can generate a full-length, infectious YK RNA genome following in vitro transcription from a bacteriophage T3 promoter, the recombinant viruses YK/519HC and YK/519CP3U were generated by replacing the HC-Pro or/and CP-3'UTR regions of YK with the corresponding sequences from 5-19 (Fig. 1A).
A total of 30 plants, each from the PRSV single-virus-resistant lines 16-0-1,17-0-1,18-2-4 and 18-0-9 21 , were mechanically inoculated with YK, 5-19 or the recombinant viruses under greenhouse conditions in four independent tests. Two months after inoculation, plants from lines 18-2-4 and 17-0-1 were completely resistant to YK infection, whereas 3 and 13% of the plants from lines 16-0-1 and 18-0-9 were infected with YK, respectively. The recombinant viruses YK/519HC and YK/519CP3U caused symptoms of severe leaf mosaic and distortion in non-transgenic (NT) papaya plants two weeks after inoculation. At two months post inoculation, YK/519HC and YK/519CP3U caused infection rates of 51% and 53%, respectively, in the four transgenic lines ( Fig. 1B; Table 1). Similar symptoms of comparable severity were also caused by YK/ 519HCCP3U in 56% of the transgenic plants, suggesting the absence of an additive effect in overcoming the resistance of CP-transgenic lines, even with the combined presence of both 5-19 HC-Pro and CP genes in the YK backbone ( Fig. 1B; Table 1).
All three recombinants were also able to overcome the resistance of PRSV-PLDMV CP transgenic lines 23 (Fig. 1C), with infection rates as high as 95-100% in the four tested lines 9-5, 10-4, 12-4, and 14-1 23 at eight weeks after inoculation (Table 2). Our results indicated that 5-19 HC-Pro enables recombinant YK/519HC, which possesses an identical CP-3'UTR to that of the transgene (100% nt identity), to overcome both single-and double-virus resistances in a sequence homology-independent manner. Because the resistance-breaking ability of 5-19 HC-Pro was not enhanced by the presence of the 5-19 CP-3'UTR in recombinant YK/519HCCP3U (Table 1 and 2), we concluded that 5-19 HC-Pro plays a vital role in the breakdown of nucleotide sequence-homology-independent resistance.
Intriguingly, YK/519CP3U also significantly increased the percentage of breakdown, implying a possible contribution from CP-3'UTR. Because the nt identity of 5-19CP/3'UTR to that of YK is much higher than to the exotic strains HA, TH, and MX, which can not overcome single-and double-virus-transgenic resistance 23,24 , the breakdown mediated by the recombinant YK/519CP3U is also probably not sequence-homology-dependent. However, we cannot exclude the possibility of homology-dependent breakdown mediated by the heterologous 5-19 CP and 3'UTR in recombinant YK/ 519CP3U because the two virus strains still show very little nucleotide divergence (95.9% identity for CP and 97.9% for 3'UTR).
Breakdown of CP-transgenic resistance is due to suppression of the host defence reaction. We investigated the ability of YK, 5-19 and the three recombinant viruses to supress PTGS in double-virus resistant lines (10-4 and 14-1) carrying the chimeric construct (PLDMV CP-PRSV YK CP-3'UTR) 23 . In northern blots, the PLDMV and PRSV CP probes did not cross-hybridise with preparations from nontransgenic plants infected with the heterologous virus ( Fig. 2A)    Because the PLDMV CP probe did not cross-hybridise with the siRNA from PRSV-infected non-transgenic plants ( Fig. 2A), the source of the detected siRNA must be the chimeric transgenic transcript. In these same infected transgenic plants, the corresponding 21-22 nt siRNAs were also observed to accumulate to higher levels than in YK-resistant transgenic plants, despite considerable accumulation of the chimeric transcript (Fig. 2C). Potyviral HC-Pro suppresses gene silencing by interfering with the RNA silencing machinery at multiple levels, including (i) reducing siRNA levels by incompletely inhibiting the processing of dsRNA by Dicer 28,29 ; (ii) preventing mRNA degradation by interfering with the function of the RNA-induced silencing complex (RISC) 29,30 , possibly by inhibiting the 3'-methylation of siRNAs, which leads to enhanced oligouridylation and subsequent degradation of siRNA 31,32 ; and (iii) binding and trapping siRNAs to compromise signaling 33,34 . Although the binding and trapping of siRNAs by HC-Pro in RDR6-dependent systems can inhibit the maintenance/amplification step and lower the total levels of signalling siRNAs 33 , the sequestered siRNA may accumulate along with the preserved target RNA 33 , as observed in the present study ( Fig. 2B&C) and also reported elsewhere 35 .
PRSV strain 5-19 has a stronger silencing suppressor HC-Pro than other strains. The gene-silencing-suppression capabilities of 5-19 in comparison with that of YK were compared via Agrobacterium tumefaciens-mediated transient expression of their respective HC-Pro constructs in the leaf tissues of Nicotiana benthamiana plants that were co-infiltrated with a translatable green fluorescent protein (GFP) construct (reporter) and a GFP hairpin construct (PTGS inducer), as described for Tobacco etch virus (TEV) P1/HC-Pro 36 . UV irradiation of the Agrobacterium-infiltrated leaves 5 days after infiltration showed that the fluorescence intensity of 5-19 HC-Proinfiltrated areas was 1.25-fold higher (mean of six repeats) than that of YK HC-Pro-infiltrated areas (Fig. 3A), indicating significantly stronger gene silencing suppression by 5-19 HC-Pro compared with YK HC-Pro (Fig. 3A). Since HC-Pro is the weapon of the virus to counteract the defensive RNA silencing from the host, if a virus mutant (strain) can resist RNA silencing, its titers would increase immediately and subsequently amplify the silencing suppression effect. This suggests that a slight difference would affect the consequence of the fight between the host and the invading virus dramatically. We strongly believe that a 25% increase in silencing suppression would significantly result in different biological phenotype as reflected by the results of this study, especially the virus recombination assays (Fig. 1). Similarly,    (Fig. S1A, B), with the latter being responsible for viral replication and PTGS suppression 37 (Fig. S1C). HC-Pro sequences from diverse strains to which the YK CP-3'UTR-transgenic plants were resistant share 95-97% aa identity with 5-19 HC-Pro (Table S1). Potyviral HC-Pro is a multifunctional protein with different domains that fulfil its multiple roles 38 , and slight changes in these different functional domains may have a concerted effect on silencing suppression. Our results revealed that there was no correlation between particular sequence differences and the ability to break CP-transgenic resistance.
An approach for solving the problem of CP sequence-homologyindependent transgenic resistance breakdown. Three transformation vectors, pBI-519HCF, pBI-519HCN and pBI-519HCC, carrying untranslatable full-length (nt 1-1371), N-terminal (nt 1-750) and C-terminal (nt 622-1371) 5-19 HC-Pro, respectively, were constructed. Each construct carries two engineered termination-codons, along with an inserted T nucleotide that causes frame shift, and is controlled by the CaMV 35S promoter and nos terminator (Fig. 4A). Somatic embryos derived from in vitro adventitious roots 39 of papaya cultivar Sunrise were transformed using A. tumefaciens harbouring the individual vectors, and 60 transgenic lines were established (Table S2). Micropropagated 5-19 HC-Pro transgenic plants were evaluated for their resistance to PRSV by mechanical inoculation with 5-19 or YK. Non-transgenic (NT) plants and plants from YK-CP-3'UTR transgenic line 18-2-4 21 were used as controls. Two weeks after inoculation in NT plants, 5-19 induced stunting and severe mosaic symptoms ( Fig. 4B-a), which subsequently deteriorated into lethal wilting; mosaic symptoms were also observed in 18-2-4 plants ( Fig. 4B-b). Thirty-one 5-19 HC-Pro transgenic lines were susceptible to 5-19 (Table S2; Fig. 4B-c), thirty of which were also susceptible to YK (Table S2), as defined by no delays in the development of symptoms. Of the twenty-nine 5-19-resistant transgenic lines, 10 were weakly resistant (WR) (Table S2), as defined by a delay in the development of symptoms, and 19 were highly resistant (HR) to 5-19 (Table S2; (Table S2; Fig. 4B-h).
A significantly higher percentage of HR lines (46.0%) was found among the full-length HC-Pro lines than among lines carrying the Nterminal 519HCN construct (29.0%) or the C-terminal 519HCC construct (21.4%) (Table S2). Southern hybridisation analysis of sixteen of lines that were highly resistant to both YK and 5-19 indicated the presence of 1-6 HC-Pro transgene inserts (Fig. S2). We also obtained 12 lines of 5-19 CP transgenic papaya, however, they did not show significant degrees of resistance to both PRSV 5-19 and YK strains (data not shown). Because high percentages of transgenic lines with different constructs of 5-19 HC-Pro (average 31.6%, Table S2) resistant to 5-19 were obtained, we did not proceed further experiments to obtain more 519-CP lines for comparison.
Four HR lines (F2-1-4, F2-7-1, F3-2-2, and N11-1) with single insertions of the untranslatable 5-19 HC-Pro construct were further analysed for their resistance to different geographical PRSV strains. Line F2-7-1 was susceptible to the exotic PRSV strains HA, TH and MX. Lines F2-1-4 and N11-1 showed delayed symptoms to HA and TH but were susceptible to MX (Table S3). However, line F3-2-2 showed complete resistance to all the exotic strains tested and was thus regarded as an extremely HR line.
Of the nine HR lines and three S lines (three HR and one S lines for each full-length, N-half, and C-half untranslatable HC-Pro) analysed by northern hybridisation, five HR lines (F2-7-1, N11-1, C9-4-1, C9-5-3, C10-13) accumulated significantly lower levels of transgenic transcripts than the S lines. However, transgenic transcripts were not detectable in the other four HR lines (Fig. 4C). In addition, siRNAs were detected in all the nine HR lines but not in any of the S lines (Fig. 4C). The highest level of siRNA was detected in the single-insert full-length HC-Pro transgenic line F3-2-2 that conferred broadspectrum resistance to 5-19 and all other PRSV strains (Table S3). Our results demonstrated that the transgenic resistance to the virulent virus 5-19 and the other PRSV strains is based on the targeting of the gene-silencing suppressor HC-Pro by PTGS. However, HC-Pro resistance is not merely dependent on sequence identity but also on its function as a suppressor.
Segregation analysis of the nptII gene in R 1 progeny. The segregation ratios of the nptII gene in the R 1 progeny from the 2 HR lines (F2-1-4 and F3-2-2) were 3:1, as calculated by the ratios of geminating seedling resistant or sentive on the selection medium containing kanamycin (Table S4), verifying that the nptII gene was integrated into the chromosome at a single locus, as revealed by Southern blotting (Fig. S2) . When these kanamycin-resistant plants were mechanically challenged with PRSV 5-19, all of them were resistant to viral infection. Thus, our results further indicated that the transgenic resistance to 5-19 is linked with kanamycin resistance and undergoes nuclear inheritance as a dominant trait.

Discussion
The severity of viruses can be increased by a few mutations in gene silencing suppressors 34 . As demonstrated by the present investigation, such mutant viruses can overcome PTGS-mediated transgenic resistance independent of transgene sequence homology. Although transgenic papaya lines grown in Hawaii are not resistant to PRSV strains from other geographic areas, no breakdown in the transgenic resistance to PRSV strains indigenous to Hawaii has been observed 3 . The geographical isolation of Hawaii restricts the diversity of PRSV strains and contributes to the stable and continuous effectiveness of PRSV CP-transgenic papaya for more than a decade 18 . However, the emergence of more virulent virus strains through the selection mechanism described in this study is expected to prevail in areas where more diverse PRSV strains exist.
Influenced by European trends, the government of Taiwan has yet to approve any of the transgenic crops cultivated on the island. Nevertheless, field tests of CP-transgenic papaya lines under isolated conditions have been strongly supported by the government over the past two decades. During more than 10 years of regulated field trials, the complexity of the viruses that infect CP-transgenic papaya lines has helped advance the development of single-virus 21 , doublevirus 22,23 , and virulent virus (this study) resistance to PRSV, PRSV/ PLDMV, and virulent PRSV strains, respectively. The lethal wilting type PRSV 5-19 with a stronger HC-Pro described herein infects the YK-CP-3'UTR-transgenic papaya plants as a non-lethal mosaic virus. Here, we demonstrate that the virulent strain 5-19 exhibits a strong ability to suppress gene silencing and overcome CP-transgenic resistance, which allows it to establish in CP-transgenic plants and cause non-lethal mosaic symptoms. The emergence and accumulation of such virulent strains in CP-transgenic plants poses a serious risk that may completely wipe out CP-transgenic and nontransgenic papaya plantations and also threatens other cucurbitaceous crops that can be infected by PRSV. Our results further indicate that the breakdown of CP-transgenic resistance by virulent strains, such as PRSV 5-19, is independent of the genetic diversity of PRSV CPs. Thus, regardless of investigations into the diversity of CPs to develop CP-transgenic resistance against different PRSV strains present in a particular region, CP-transgenic plants could still facilitate the emergence of PRSV mutants or variants with stronger RNA silencing suppression from amongst the natural virus population. This complexity emphasises the importance of monitoring the possible emergence of more virulent virus strains during the cultivation of CP-transgenic crops.
Our study also demonstrated that untranslatable constructs targeting the gene silencing suppressor HC-Pro can disarm viral protections against host defences and minimise the potential emergence of more virulent virus strains that can overcome CP-transgenic resistance. Currently, we are pyramiding single-virus, double-virus, and virulent virus transgenic resistances in papaya to develop a hybrid variety with broad-spectrum resistance for global application and reduced risk for the emergence of more virulent virus strains during cultivation.

Methods
Virus isolates. Papaya ringspot virus (PRSV) isolates YK 40 , 5-19 24 , HA 41 , TH and MX 21 and Papaya leaf-distortion mosaic virus (PLDMV) isolate P-TW-WF 22 were used in this study. PRSV YK, a typical mosaic type strain collected from southern Taiwan, was previously characterised and its genome sequenced 40 . PRSV 5-19 24 and PLDMV P-TW-WF 22 , both severely infecting transgenic papaya (Carica papaya L. cultivar Tainung No. 2) lines carrying the PRSV YK CP coding region along with the contiguous PRSV YK 3'-UTR 21 , were collected from central Taiwan during field trials 42 . The geographically distant PRSV strains HA 41 , TH and MX 21 , which originate from Hawaii, Thailand and Mexico, respectively, were used to evaluate the broadspectrum resistance of the transgenic papaya lines. All virus isolates were maintained in papaya (Carica papaya L. cv. Tainung No. 2) plants.

Confirmation of recombinant viruses in infected plants.
NotI-linearised constructs for the wild-type and recombinant viruses were transcribed in vitro, and the products were mechanically introduced into papaya seedlings (cv. Tainung No. 2) as described previously 44 . The inoculated plants were maintained in a temperature-controlled (25 6 3uC) greenhouse to observe symptom development.
Appropriate cDNAs amplified by RT-PCR using RNA extracted from virusinfected papaya plants were analysed by restriction-enzyme digestion and sequencing to confirm the identities of the recombinant viruses. cDNA containing the HC-Pro coding region was amplified using the PYK1646/MYK3139 primer pair and analysed by XhoI digestion to specifically release the fragment corresponding to the 5-19 HC-Pro gene. cDNA containing the CP coding region along with the 3'-UTR was amplified by RT-PCR using the YK905/YK10326 primer pair and then sequenced to verify its origin from PRSV 5-19.
All transgenic lines were evaluated against the recombinant viruses YK/519HC, YK/519CP3U and YK/519HCCP3U (Fig. 1A), which were derived from the pT3YK/ 519HC, pT3YK/519CP3U and pT3YK/519HCCP3U constructs, respectively, by mechanical inoculation, with YK and 519 serving as controls. Inocula of the challenge viruses were prepared by extracting leaf tissues from virus-infected papaya plants www.nature.com/scientificreports SCIENTIFIC REPORTS | 5 : 9804 | DOI: 10.1038/srep09804 (cv. Tainung No. 2) with 20 volumes (w/v) of 0.01 M phosphate buffer (pH 7.0) at 15 days after inoculation with the individual viruses. Plants from the YK CP-3'UTR transgenic lines were inoculated on two fully expanded upper leaves at the 10-cm stage. The infection rates from four independent tests, with a total of 30 plants/ line, were statistically analysed using Duncan's multiple range test. Plants from the doublevirus resistant PY16 lines were similarly inoculated at the 15-cm stage, and two experiments were performed, with a total of 20 plants/line. All inoculated plants were kept in a temperature-controlled (25 6 3uC) greenhouse for two months after inoculation to evaluate their resistance.
Indirect enzyme-linked immunosorbent assay. Indirect ELISA was performed as described previously 45 , with modifications. Leaf extracts diluted (1:100) in 50 mM sodium carbonate buffer (pH 9.6) with 0.01% sodium azide were used to coat the wells of a polystyrene microtiter plate. PRSV CP antiserum 46 diluted 2000-fold in conjugate buffer 46 was added (200 ml/well) and incubated at 37uC for 1 hr. After the wells were washed the three times, 200 ml of alkaline phosphatase-conjugated goat anti-rabbit immunoglobulin (KPL, Inc., Gaithersburg, MD, USA), diluted 5000-fold in conjugate buffer, was added to each well. After additional washing, 100 ml of 1 mg/ml p-nitrophenyl phosphate (Sigma-Aldrich Corporation, St. Louis, MO, USA) in substrate buffer (100 mM diethanolamine, pH 9.6) was added to each well. The absorbance was measured at 405 nm on a Rainbow microplate reader (SLT Lab Instruments, Winosski, VT, USA) 30 min after the addition of enzyme substrate. Transgenic plants that exhibited two-fold or higher absorbance compared with the negative control were classified as ELISA positive.
Preparation of 5-19 HC-Pro antiserum. Amorphous inclusions of PRSV 5-19 were purified from infected Cucumis metuliferus (Naud.) Mey. (Acc. 2459) using the previously described differential centrifugation method 47 . The HC-Pro protein was further dissociated and then purified by preparative gel electrophoresis as described 46,48 . BALB/c mice were immunised by intraperitoneal injection of 50 mg of purified HC-Pro protein once a week for three weeks. The antiserum was collected to detect HC-Pro from PRSV 5-19 and YK in the agro-infiltration assay.
Suspensions of Agrobacterium tumefaciens ABI carrying pBA-GFP (a GFP expressor, driven constitutively by a 35S promoter, prepared in our laboratory) and pBA-GFi (a 2/3 GFP ORF construct with an inverted repeat to induce silencing, provided by Dr. Shih-Shun Lin, National Taiwan University) were mixed (ratio 2: 1: 2) with the individual pBA-519HC, pBA-YKHC and pBA-NSs vectors containing 5-19 HC-Pro, YK HC-Pro and the NSs coding region (silencing suppressor positive control) of Watermelon silver mottle virus (WSMoV) 48 , respectively, and then injected into the lower side of leaves using a syringe without a needle, and the infiltrated plants were kept in a temperature controlled (25 6 3uC) greenhouse. The empty vector pBA 49 was used as a negative control. GFP fluorescence was monitored at 5 dpi using a hand-held UV light B-100AP (UVP, CA, Upland, USA) and photographed using a digital camera (D7000, Nikon, Tokyo, Japan) with a Cokin P series filter (Cokin, 84 mm, 524 nm, Rungis, French). To minimize the variation in leaf age, size and position, half leaf assay was performed (n56). The GFP fluorescence signals were quantified by Kodak 1D image analysis software (Eastman Kodak, Rochester, NY). The density of GFP fluorescence in YK HC-Pro infiltrated leaves was arbitrarily set as 1:00 to relatively quantify that in the same half leaves infiltrated with 5-19 HC-Pro. The transient expression of recombinant proteins was monitored at 5 dpi via chemiluminescent western blotting (Amersham, Bucks, U.K.) using 5-19 HC-Pro antiserum (prepared as described above), GFP antiserum 50 , or WSMoV NSs MAb 48 as the primary antibody and horseradish peroxidase-conjugated goat anti-rabbit immunoglobulin (Amersham, Bucks, U.K.) or horseradish peroxidase-conjugated goat anti-mouse immunoglobulin (Amersham, Bucks, U.K.) as the secondary antibody.
The protein signals were quantified by Kodak 1D image analysis software (Eastman Kodak, Rochester, NY).
Northern hybridisation for mRNA/siRNA detection. For the detection of particular transcripts, total RNA was extracted from young leaves of transgenic and nontransgenic papaya plants (cv. Tainung No. 2) using the ULTRASPEC TM RNA isolation system (Biotecx Laboratories Inc., Houston,TX,USA). Fifteen micrograms of total RNA was separated on a 1.2% agarose gel with formaldehyde, transferred to a Gene Screen Plus nylon membrane (Dupont Co., Boston, MA) and hybridised with a-32 P-dATP-labeled probes specific to the coding sequences from PLDMV CP, PRSV YK CP, GFP or PRSV 5-19 HC-Pro. The probes were prepared using the Primer-It II random primer labelling kit (Stratagene, La Jolla, CA) from 25 ng DNA amplified from the corresponding constructs using appropriate primers. Prehybridisation, hybridisation, washing, and autoradiography were carried out as described 23 .
For the detection of the siRNA derived from the transgenic constructs, thirty micrograms of total RNA was resolved on a 15% polyacrylamide/13TBE (8.9 mM Tris-HCl, 8.9 mM boric acid, 20 mM EDTA)/8 M urea gel and transblotted to a Hybond-N1 membrane (Amersham Phamacia Biotech, Bucks, UK). For hybridisation, a-32 P-labeled probes were prepared as described above.
Hybridisation was carried out in ULTRAHyb-Oligo solution (Ambion Inc., Austin, TX), and signals were detected by autoradiography. The DynamarkerH Prestain Marker (BioDynamics Laboratory Inc., Tokyo, Japan) for small RNA was used as the molecular weight standard. The mRNA/siRNA signals were quantified by Kodak 1D image analysis software (Eastman Kodak, Rochester, NY).
Cloning and sequencing of PRSV TH and MX HC-Pro coding sequences. Total RNA was isolated from papaya leaves infected with PRSV TH or PRSV MX 3 weeks after inoculation. The primer MYK3234 (5'-GTTAAAAGTACGCTTGGTGACATC-3'), which corresponds to nucleotide positions 3211-3235 in the PRSV YK RNA sequence, was used for first-strand cDNA synthesis by reverse transcription (RT). The forward primer PPYK1607 (5'-GCTGACAGAATAGGTCGATA-3'), which corresponds to nucleotide positions 1607-1626 in the PRSV YK RNA sequence, was used with the reverse primer MYK3234 to PCR amplify the HC-Pro gene from each virus . RT-PCR-amplified PRSV TH and MX HC-Pro fragments were cloned into pCR-TOPO (Invitrogen, San Diego, CA, USA) and sequenced.
The nucleotide and amino acid sequences of HC-Pro from PRSV TH and MX were compared with those from 5-19, YK 40 and HA 43 using PC/GENE 6.85 software (IntelliGenetics, Inc., Mountain View, CA, USA). Multiple sequence alignments were constructed using the PILEUP program of the GCG package, version 9.0 (Genetics Computer Group, Madison, WI, USA).
The primers 519HCStopB (5'-GACAGAAATGGGCTCTAGATGTGGGGT GAATGATGATATCATCGCCAAAAG-3'), containing two stop codons (bold), an XbaI site (underlined) and an inserted extra nucleotide (italicised) for frame shifting, and 519HCSacB were used to amplify the C-terminal fragment (750 bp) of the PRSV 5-19 HC-Pro gene. The amplified product was digested with XbaI/SacI and ligated into XbaI/SacI-digested pBI121 to generate the untranslatable construct pBI-519HCC. All three untranslatable constructs were transferred into A. tumefaciens LBA4404 by electroporation using a Gene Pulser (BioRad, Hercules, CA, USA) as previously described 51 .
Generation of HC-Pro transgenic papaya lines. Papaya transformation and regeneration were carried out as described previously 39 , with the required modifications. Embryogenic tissues derived from the adventitious roots of the PRSVsusceptible papaya cultivar Sunrise wounded in liquid-phase by carborundum were transformed using A. tumefaciens carrying pBI-519HCF, pBI-519HCN or pBI-519HCC. After selection on kanamycin (100 mg/l) medium for 3 months and subsequent growth without kanamycin for one month, the mature somatic embryos were germinated on MSNB shoot medium [MS medium with 0.02 mg/l a-naphthaleneacetic acid (NAA) and 0.2 mg/l BAP] containing 100 mg/l kanamycin for shoot development. Shoot regeneration, micropropogation and acclimatisation were carried out as described previously 20 . The micropropagated R 0 plants from individual putative transformed shoots were established in vermiculite soil under temperature-controlled (25 6 3uC) greenhouse conditions 20 and used for further evaluation. The presence of the transgene in the putative transgenic lines was checked by PCR using the primer pairs 519HCStopA/519HCSacB, 519HCStopA/519HCSacA or 519HCStopB/519HCSacB. The nptII-specific primers NA and NB 23 were also used for PCR analysis to confirm the presence of the selection marker.
Evaluation of HC-Pro transgenic lines for resistance to 5-19 and YK. The viral resistance of the putative HC-Pro transgenic lines was evaluated by mechanical inoculation with PRSV 5-19 24 or PRSV YK 40 using sixty micropropagated R 0 plants (5 plants/line) at approximately the 15-cm stage. The first two fully expanded leaves were dusted with 600-mesh carborundum and rubbed with 200 ml of inoculum that had been prepared from papaya leaf tissues 3 weeks after inoculation with PRSV 5-19 or PRSV YK (1:20 w/v in 0.01 M potassium phosphate buffer, pH 7.0). Non-transgenic papaya plants (cv. Sunrise) and YK-CP3'-UTR transgenic papaya line 18-2-4 21 were used as controls. Plants were kept in a temperature-controlled greenhouse (25 6 3uC), and symptom development was monitored for 7 weeks.
Southern hybridisation. The DNeasy Plant Mini kit (Qiagen Inc., Valencia, CA, USA) was used to extract genomic DNA from HC-Pro transgenic papaya plants. Thirty micrograms of genomic DNA was digested with AseI (two AseI sites are located at nt 4764 and 4823 in pBI121, prior to the 35S promoter) and the electrophoretically resolved profile (0.8% agarose gel) was transferred to a Gene Screen Plus nylon www.nature.com/scientificreports SCIENTIFIC REPORTS | 5 : 9804 | DOI: 10.1038/srep09804 membrane (Dupont Co., Boston, MA, USA). The Primer-It II random primer labelling kit (Stratagene, La Jolla, CA, USA) was used to prepare the a-32 P-dATPlabeled HC-Pro probe from 25 ng of DNA (1371 bp) amplified from pBI-519HCF using the primer pair 519HCStopA/519HCSacB. Southern hybridisation and autoradiography were performed as described previously 23 .
Resistance of HC-Pro lines to different geographic PRSV strains. Plants from four highly resistant transgenic lines that contained a single transgene insert, as identified by Southern blotting, were tested against geographically distinct PRSV strains from Hawaii (HA) 43 , Thailand (TH) 21 , and Mexico (MX) 21 by mechanical inoculation, as described above.
Analysis of transgene segregation in R 1 progeny. R 0 plants from two highly resistant 5-19 HC-Pro transgenic papaya lines, F2-1-4 and F3-2-2, were self-crossed under greenhouse conditions, and the R 1 premature seeds were collected at 90 days after pollination to analyse the inheritance of the transgene. To analyse the pattern of inheritance of the nptII gene, the R 1 premature seeds of F3-2-2 were cultured on MSNB medium containing 100 mg/l kanamycin for two weeks, until plantlets developed. The kanamycin-resistant R 1 progeny were further established and challenged by mechanical inoculation with PRSV 5-19, as described above, to ascertain the inheritance of the transgenic resistance.