Rice plants infected with RSV were collected from Jiangsu province in China. Young instar nymphs of SBPHs were fed on the RSV-infected rice plants for 2 days to acquire the virus and were maintained on “Wuyujing No. 3” rice plants grown in an insect-rearing room at a temperature of 25 ± 3°C, 55 ± 5% relative humidity and under a light intensity of 200 μmol m^-2 s^-1 (16 h photoperiod). The proportions of viruliferous SBPHs were confirmed by dot-ELISA (Sun et al., 2011).

Arabidopsis thaliana (ecotype Columbia-0, Col-0) seeds were grown in potting soil in a growth chamber at 24°C under 200 μmol m^-2 s^-1 illumination. Rice plants used in this study were grown inside a growth chamber at 26°C, 80% relative humidity. Additionally, all plants were kept under 16-hlight/8-h dark photoperiod conditions.

Rice plants (Oryza sativa L. spp. japonica. cv. Nipponbare) were inoculated with three viruliferous SBPHs per plant and were kept in a growth chamber. After inoculation for 3 days, planthoppers were removed back to rice field. And the inoculation does not hurt SBPHs at all, we just let them feed on the tested rice plants. Plants were maintained in a growth chamber for symptom development, and RSV-free SBPHs were used for mock inoculation.

Total RNA was extracted from RSV infected rice. cDNA synthesis and PCR were conducted as described previously (Sun et al., 2016) to get the full length of NS3. The full-length NS3 DNA was integrated into the pENTR/D-TOPO vector (Invitrogen, Carlsbad, CA, United States) and was then cloned into the pBA-Flag-6Myc-DC binary vector (Zhang et al., 2005) using the LR clonase reaction mixture to induce constitutive expression of the NS3 gene in Arabidopsis. The resultant 35S:NS3 construct was introduced into Agrobacterium tumefaciens strain GV3101. Agrobacteria-mediated transformation of A. thaliana ecotype Col-0 with the NS3 gene was conducted using the floral dip method (Clough and Bent, 1998; Zhang et al., 2006). The transgenic seeds were selected on standard MS medium (Murashige and Skoog, 1962) containing the appropriate antibiotics: 100 mg/L carbenicillin (Sangon Biotech, Shanghai, China).

Total RNA was extracted from Oryza sativa L. spp. japonica. cv. Nipponbare and the full-length OsCIPK30 cDNA was cloned into the pUN1301 vector (Biovector, Co., Ltd, Beijing, China) to generate the Ubi-OsCIPK30 plasmid. The construct was introduced into A. tumefaciens strain GV3101. Agrobacteria-mediated transformation protocols for rice were published previously (Wu et al., 2007). Regenerated transgenic plants (T₀) were transplanted into soil; and the T₁ progeny germinated from seeds were grown in soil inside a greenhouse. The T₁ generations were confirmed by RT-PCR for testing hygromycin gene. Empty pUN1301 vector-transformed plants were used as the wild type (WT) controls.

Sequence alignment of 33 members of the OsCIPK family in rice was performed using ClustalW. Phylogenetic analysis of the 33 OsCIPKs in rice based on amino acid sequences carried out using a neighbor-joining (NJ) method with MEGA version 4.0.

To construct the bait vector for the assay, the coding sequence of NS3 was cloned into the yeast GAL4 binding domain vector pGBKT7, as instructed (Clontech, Mountain View, CA, United States). Construction and screening of a rice cDNA library (Oryza sativa L. spp. japonica. cv. Nipponbare) were performed as described in the BD Matchmaker Library Construction and Screening Kits User Manual (Clontech, Mountain View, CA, United States). Briefly, the rice cDNA library was screened using pGBKT7-NS3 as the bait in the Saccharomyces cerevisiae strain Gold (Clontech, Mountain View, CA, United States). The positive clones were selected on a histidine-deficient medium and then confirmed by β-gal assays. The coding sequence of OsCIPK5 (Os01g0206700), OsCIPK25 (Os06g0543400), OsCIPK26 (Os02g0161000), OsCIPK29 (Os07g0678300), OsCIPK30 (Os01g0759200), and OsCIPK32 (Os12g0132200) was PCR-amplified from rice leaf cDNA and then inserted into the yeast GAL4 activation domain vector, pGADT7 (Clontech, Mountain View, CA, United States).

Combinations of the plasmids BD-NS3 and AD-OsCIPK5, BD-NS3 and AD-OsCIPK25, BD-NS3 and AD-OsCIPK26, BD-NS3 and AD-OsCIPK29, BD-NS3 and AD-OsCIPK30, or BD-NS3 and AD-OsCIPK32 were co-transformed into the S. cerevisiae strain Gold. BD-53 and AD-T were also co-transformed into S. cerevisiae strain Gold to serve as a positive control. S. cerevisiae co-transformed with BD-Lam and AD-T, BD-NS3 and AD, or BD and AD-OsCIPK30 were as used as negative controls. All transformants were grown at 30°C for 72 h on synthetic medium lacking Leu and Trp and then transferred to medium lacking adenine, His, Leu, and Trp and supplemented with 0.5 mg ml^-1 Aureobasidin A (AbA).

The LCI was performed on 4-week-old Nicotiana benthamiana (N. benthamiana) leaves infiltrated with various combinations of A. tumefaciens GV3101 harboring pCambia-Myc-NS3-nLUC and pCambia-cLUC-3HA-candidate proteins. The agrobacteria containing the plasmids were incubated with 25 μM beta-estradiol for 3 h prior to infiltration, and all cultures were adjusted to OD₆₀₀ = 0.8. The transfected leaves were assayed 2 days after agroinfiltration by adding the substrate (10 mM luciferin). The sprayed leaves were incubated in total darkness for 5 min and photographed using the Tanon 5200 Luminescent Imaging Workstation with an EMCCD camera. The images were processed with TanonImage software (Tanon Science & Technology, Shanghai, China).

Extraction of total protein from leaf tissues (100 mg per sample) for western blotting was done as described previously (Sun et al., 2016). The membranes were probed with anti-Myc antibodies (Sigma-Aldrich, St. Louis, MO, United States). The secondary antibody used in the study was a goat anti-rabbit IgG conjugated with alkaline phosphatase (Sigma-Aldrich, St. Louis, MO, United States).

Equal amounts of the RNA samples from five independent plants of the same treatment were pooled. Total RNA was extracted from each treatment using RNAiso Plus reagent (Takara, Dalian, China), and cDNA synthesis and PCR were conducted as described previously (Sun et al., 2016). Actin1 (Os03g0718100) was used as an internal standard. qRT-PCR was performed using the Bio-Rad iQ5 Real-time PCR System and SYBR Premix Ex Taq and the primers used in this assay are listed in Supplementary Data Sheet 1 (Takara, Dalian, China). The comparative cycle threshold method was used to calculate the relative mRNA levels (Applied Biosystems, 1997; Johnson et al., 2000). All qRT-PCR experiments were performed at least three times.

Data were obtained from three replicates and analyzed by unpaired two-tailed Student’s t-test with GraphPad Prism software (GraphPad Software, Inc., La Jolla, CA, United States). Values of P < 0.05 or P < 0.01 were considered to indicate different levels of statistical significance.

To study the function of NS3, we generated transgenic Arabidopsis lines overexpressing full-length RSV NS3 with Flag-6Myc epitopes. These transgenic plants were confirmed by western blot assays (Figure 1B). Importantly, the majority of the transgenic lines exhibited developmental abnormalities with serrate true leaves (Figure 1A). The phenotypes of transgenic plants indicated that NS3 might impact growth and development in Arabidopsis and might also influence rice.

To identify plant host factor(s) that interact with NS3 during RSV infection, we screened a rice cDNA library using RSV NS3 as the bait through a yeast two-hybrid assay. Twenty-one positive cDNA clones were identified. One of them was a fragment of the OsCIPK30 gene. To test the interaction between the full-length OsCIPK30 and NS3, the ORFs of OsCIPK30 and NS3 were amplified and integrated into pGADT7 and pGBKT7, respectively. The recombinant plasmid (AD-30) was co-transformed into yeast cells with BD-NS3. Figures 2A,B show the positive interaction between the full-length OsCIPK30 and NS3 through yeast two-hybrid assays.

To demonstrate whether NS3 and OsCIPK30 also interacted in plant cells, a luciferase complementation imaging (LCI) assay was used. In the LCI experiments, the N- and C-terminal regions of firefly luciferase (NLuc and CLuc) were fused to NS3 and OsCIPK30, respectively, and were transiently expressed in N. benthamiana. When NLuc and CLuc are close to each other through the interaction of the two proteins, catalytic activity is restored and recorded through the CCD camera. In our LCI screening, NS3 displayed LUC complementation with OsCIPK30, suggesting that NS3 is physically close to OsCIPK30 in vivo (Figure 3C). Together, these assays clearly indicated that NS3 and OsCIPK30 interact in vivo.

As mentioned above, OsCIPK30 belongs to the CIPK family in rice. According to the Institute for Genomic Research [TIGR] database and NCBI database, we collected all 33 CIPK proteins in rice named from OsCIPK1 to OsCIPK33. Phylogenetic analysis of OsCIPKs showed that the proteins could be classified into several groups (Figure 3A); we chose the other five OsCIPK proteins from different groups and performed Y2H and LCI assays to identify whether they also interacted with NS3. The results showed that only OsCIPK29 was able to interact with NS3, whereas OsCIPK5, 25, 26, and 32 could not (Figures 3B,C).

To investigate whether RSV can induce the expression of OsCIPKs (OsCIPK5, OsCIPK25, OsCIPK26, OsCIPK29, OsCIPK30, and OsCIPK32), we examined the relative transcript level of these six OsCIPK genes in RSV-infected rice at 7 days post-inoculation (dpi). Real-time PCR showed that only three OsCIPKs were up-regulated, and two of them exhibited highly significant differences compared to the mock rice, suggesting that OsCIPK29 and OsCIPK30 might be involved in the host response to RSV infection (Figure 4).

Given that the transcription level of OsCIPK30 is the highest in six OsCIPKs when RSV infects the host, we generated the overexpression transgenic rice of OsCIPK30 to further investigate its function. Under greenhouse conditions, the Ubi-OsCIPK30 rice plants showed normal growth compared with WT (Figure 5A). We then inoculated WT and Ubi-OsCIPK30 transgenic rice plants with RSV and observed the symptom development of the plants over a time course.

Fewer transgenic rice plants showed symptoms than WT rice at 7 and 14 dpi; moreover, the Ubi-OsCIPK30 rice showed milder RSV symptoms than those in the WT rice at 21 dpi (Figures 5B,C). These results suggested that Ubi-OsCIPK30 rice could delay the RSV symptoms and might influence the RSV infection. The real-time PCR results demonstrated that the expression levels of OsCIPK30 in transgenic rice plants were significantly increased, while the transcriptional level of the RSV NP gene was decreased (Figures 5D,E), indicating that the enhanced tolerance was associated with increased accumulation of OsCIPK30 transcripts.

A previous study showed several pathogenesis-related genes involved in RSV infection (Satoh et al., 2010), and our results also showed that OsCIPK30 might influence RSV infection. To determine whether OsCIPK30 is involved in pathogenesis-related defense systems, we analyzed the expression patterns of several pathogenesis-related genes. The second youngest leaves were collected from five independent plants of Ubi-OsCIPK30 rice and WT rice at 7 dpi. The real-time PCR showed that RSV infection strongly induced the expression of PR1, PR10, PRB1-2, and PRB1-3 in OsCIPK30-overexpressing plants, showing significant differences between transgenic plants and WT plants, while the expression levels of the PR genes were not changed in un-inoculated OsCIPK30-overexpressing plants compared with WT rice (Figures 6A–D). We also took an experiment showing the effect of the PR1, PR10, PRB1-2, and PRB1-3 expression in WT plants by RSV infection (Figure 7). The results showed that the expression levels of PRB1-3 increased after RSV infection in WT rice, but PR1, PR10, PRB1-2 not. Together with the results showing in Figure 6, we conclude that the pathogenesis-related genes are triggered in OsCIPK30-overexpressing plants after RSV infection. Taken together, these results suggest that OsCIPK30 positively regulates the pathogenesis-related genes to enhance tolerance to RSV.