Fully expanded leaves from 2-year old Chinese wild V. quinquangularis cv. “Shang-24” were infected with PM through gentle contact with leaves exhibiting disease symptoms. Infected samples were subsequently collected at 0, 6, 12, 24, 36, 48, 72, and 96 hours post-inoculation (hpi). Control leaves were sprayed with sterile distilled water and were harvested at the same time points. Samples were immediately frozen in liquid nitrogen, and then stored at -80°C until subsequent RNA extraction.

Total RNA was extracted from V. quinquangularis cv. “Shang-24” leaf samples inoculated with PM 0, 6, 12, 24, 36, 48, 72, and 96 hpi using the E.Z.N.A. ^®Plant RNA Kit (Omega Bio-tek, Norcross, GA, USA). Subsequent first-strand cDNA synthesis was carried out using PrimerScript^TMRTase according to the manufacturer’s instructions (TaKaRa Bio Inc., Dalian, China). Amplification of cDNA was conducted using 2× Taq PCR MaterMix (BioSci Biotech, Hangzhou, China) and primers specific to each of the 31 VqSTS transcripts, respectively (primer sequences are provided in Supplementary Table S1). In each case, reactions generated amplicons ranging in size from 77 to 670 bp. The grape Actin1 gene (GenBank Acc. No. AY680701) was utilized as an internal reference. Thermal parameters for PCR amplification were as follows: 94°C for 2 min, followed by 30–40 cycles of 92°C for 30 s, 60 ± 5°C for 30 s, 72°C for 30 s, and final extension at 72°C for 2 min. PCR products were separated on a 1.5% (w/v) agarose gel and imaged under UV light. GeneSnap and HemI 1.0 (Deng et al., 2014) programs were utilized for assessing relative semi-quantitative expression levels. All reactions were carried out in triplicate.

The full-length 1170 bp VqSTS21 coding sequence was amplified using gene-specific primers and 2× Taq PCR MasterMix (BioSci Biotech, Hangzhou, China). The resulting PCR product was then cloned into the pGEM^®-T Easy vector (Promega, Madion, WI, USA). Restriction sites were added to the ends of the VqSTS21 coding sequence in order to facilitate downstream cloning using gene-specific primers with BamH I and Sma I sites added to their 5′ termini, respectively. Following cloning, the VqSTS21 coding sequence was inserted downstream of the CaMV 35S promoter in the plant expression vector pCambia 2300 (Cambia, Brisbane, QLD, Australia).

The resulting plant transformation vector was introduced into A. thaliana (Col-0) using the floral dip method (Clough and Bent, 1998). To identify transgenic lines, T₁ seeds were harvested and sown on MS medium (PhytoTechnology Laboratories, Overland Park, KS, USA) containing 10 g/L sucrose, 8 g/L agar at pH 5.8 and supplemented with 75 mg/L kanamycin. Fifteen independent lines were found by segregation analysis of T₂ seeds to bear a single copy of the transgene and were confirmed as transgenic using PCR (data not shown). Subsequently, these 15 lines were inoculated with PM and the three lines (L1, L2, and L3) exhibiting the highest level of disease resistance were selected for all further experiments. T₃ homozygous L1, L2, and L3 lines were utilized for the remainder of this study and wild-type (Col-0) plants were used as the untransformed control. All Arabidopsis plants were grown at 21∼23°C with a 16 h/8 h photoperiod (100 μmol m^-2 s^-1 photon flux density) at ∼60% relative humidity (RH) on soil.

Powdery mildew (10–14 days post-inoculation, dpi) cultivated on 4-week old pad4 (phytoalexin deficient 4) mutant Arabidopsis plants, which are very susceptible to this pathogen (Reuber et al., 1998), was utilized as inoculum to inoculate 24 four-week old Col-0 and transgenic plants, respectively. Treated plants were incubated at 22°C with 16 h light and ∼80% RH for 3 days following inoculation and were then transferred to an ambient environment at 22°C and 30∼40% RH. The response of VqSTS21 transgenic plants and untransformed controls was monitored between 0 and 7 dpi.

Botrytis cinerea was obtained from tomato and cultured on potato dextrose agar medium at 25°C in the dark. Conidial spore suspensions (2 × 10⁷ conidia/ml) were prepared with 14-day old cultures using sterile, distilled water as described previously (Guo et al., 2016). Fifty leaves from 24 four-week old plants, including the three transgenic lines and untransformed controls, respectively, were rinsed with distilled water and 10 μl conidial suspension was applied, after which time the leaves were placed on 1% agarose overlaid with wet filter paper and the glass Petri dishes were then sealed. Inoculated leaves were incubated at 22°C with a 16/8 h light cycle and ∼95% RH. Disease incidence and lesion diameter were recorded daily until 4 dpi.

Pst DC3000 was grown at 28°C in King’s B medium (Tornero and Dangl, 2001) supplemented with 50 μg/ml rifampicin in a shaker at 280 rpm until it reached an OD₆₀₀ of 0.8∼1.0. The resulting cell suspension was centrifuged at 5000 g for 10 min, and was then diluted with 10 mM MgCl₂ to an OD₆₀₀ of 0.02. Four-week old transgenic and Col-0 plants were dipped into the cell suspension containing 0.02% Silwet L-77 for 10 min as described previously (Tornero and Dangl, 2001), and were then covered with a plastic lid to maintain a high level of humidity for 24 h. Disease symptoms were assessed 3 dpi. For bacterial population assays, leaf disks (0.5 cm × 0.5 cm) were collected from 8 independent L1, L2, L3 lines and untransformed control plants, respectively, at 3 dpi. Disks were rinsed with sterile water three times, homogenized in 100 μl 10 mM MgCl₂, and the solution was gradually diluted. One hundred microliters of the resulting diluted solution were then plated onto King’s B agar plates supplemented with 50 μg/ml rifampicin and 50 μg/ml kanamycin for 48 h at 28°C (Guo et al., 2016).

The accumulation of resveratrol and its stilbenoid derivates were analyzed in triplicate in PM-inoculated and mock-inoculated leaf samples (0.5 g each, 7 dpi) harvested from 4-week old VqSTS21 transgenic lines and untransformed controls. The resulting samples were ground to a fine powder in liquid nitrogen using a mortar and pestle. Ground leaves were extracted with 5 ml 80% methanol, and supernatant fractions were collected following centrifugation at 4500 g for 5 min. The resulting extracts were evaporated using a vacuum rotary evaporator (CS110-4, LaboGene, Denmark), and were immediately re-dissolved in 0.2 ml of pure methanol. These extracts (30 μl) were then filtered through a 0.45 μm sterile Durapore^® PVDF filter (Millipore, USA). Samples were run on an Agilent 1200 HPLC system (Agilent, Waldbronn, Germany) with an Agilent ZORBAX SB-C18 column (5 μm, 4.6 × 250 mm), H₂O-methanol as eluent (H₂O:methanol [60:40], flow rate 0.8 ml/min), and a wavelength of 306 nm for detection. The column temperature was maintained at room temperature. Stilbenoids in transgenic lines were identified by comparing the retention time with those of standards.

Superoxide anions (O2–) accumulation and cell death were monitored in 30 detached leaves from 12 transgenic and untransformed control plants, respectively, following pathogen inoculation using nitro blue tetrazolium (NBT) and trypan blue staining. Experiments involving PM were conducted 5 dpi while those involving B. cinerea and Pst DC3000 were carried out 3 dpi. In the case of NBT staining, inoculated leaves were incubated in HEPES buffer (pH 7.5) containing 6 mM NBT for 2∼3 h (Tu et al., 2016). Trypan blue staining was carried out as described previously (Guo et al., 2016). Briefly, infected leaves were soaked in boiled trypan blue solution (a 1:1:1:1:1 ratio of trypan blue, phenol, glycerol, lactic acid, and water) for 2–3 min, and were subsequently depigmented in 2.5 g/ml chloral hydrate for 1–2 days.

T₃ seeds from transgenic lines and Col-0 plants were vernalized at 4°C for 3 days, disinfected in 75% ethanol for 30 s, washed with sterilized distilled water three times, incubated in 10% NaClO for 5 min, and finally washed with sterilized distilled water five times. To determine the inhibitory concentrations of NaCl and mannitol for cotyledon greening rates, both transgenic and Col-0 seeds were sown on MS medium containing different concentrations of NaCl and mannitol, respectively. Subsequently, 120 sterilized seeds from each line were sown on MS medium supplemented with 130 mM NaCl or 250 mM mannitol to stimulate salt and drought stress, respectively (Guo et al., 2016; Tu et al., 2016), and were grown at 21∼23°C with a 16 h/8 h photoperiod (100 μmol m^-2 s^-1 photon flux density). Cotyledon greening rates were recorded daily. Each experiment was conducted in triplicate.

Fifteen 4-day old transgenic and Col-0 seedlings, respectively, that had been cultivated on MS medium were transferred to either fresh MS medium or MS medium supplemented with 130 mM NaCl, 250 mM mannitol or 0.75 μM ABA. Root lengths were measured 6 days following osmotic stress treatment.

To assess the response of mature plants to osmotic stress, 4-week old transgenic lines and Col-0 plants were treated with salt or drought. In the case of salt treatment, soil was allowed to dry somewhat prior to watering with 130 mM NaCl to prevent dilution, and watering with salt solution was carried out once every 3 or 4 days for 7 days. For drought treatment, watering was halted for 7 days. Performance of the plants was assessed and survival rates recorded 7 days following both types of osmotic treatment. In each instance, experiments were carried out in triplicate.

VqSTS21 transgenic lines and Col-0 untransformed controls were sown on MS medium plates, and 1-week old seedlings were subsequently transferred to flasks containing MS liquid medium supplemented with 130 mM NaCl or 250 mM mannitol. Seven days following commencement of osmotic stress treatment, seedlings were collected for physiological assessments. To measure chlorophyll content, 0.1 g seedlings with their roots removed were submerged in 5 ml 96% ethanol and incubated at 4°C until the seedlings turned white (Guo et al., 2015).

For assessment of relative electrolyte leakage, 0.1 g seedlings were incubated in ultrapure water under vacuum for 20 min and were then left at room temperature for 2 h. Conductivity (C1) of the incubation solution was subsequently measured using a conductivity detector (FE30, METTLER-TOLEDO, China). Seedlings were then submerged in boiling ultrapure water for 20 min, cooled to room temperature, and the conductivities of the resulting solutions (C2) were once again determined. The values of relative electrolyte leakage were calculated using the ratio of C1–C2 (Bajji et al., 2002).

To determine malondialdehyde (MDA) content, 0.5 g seedlings were ground with 5 ml trichloroacetic acid in a mortar and then transferred to 10 ml centrifuge tubes for centrifugation at 4,000 rpm for 10 min. Subsequently, 2 ml of the supernatant was combined with 2 ml 0.67% (w/v) thilbarbituric acid, heated to 100°C for 15 min, immediately cooled on ice, and centrifuged at 4,000 rpm for 5 min. Finally, the absorbance of the supernatant was measured at 532 nm (A532), 600 nm (A600), and 450 nm (A450) using a spectrophotometer (UV-1700, Shimadzu, Japan). MDA content was calculated as follows: MDA content = 6.45 (A532- A600)-0.56 × A450 (Tu et al., 2016).

Five 4-week old VqSTS21 transgenic L1, L2, and L3 lines, as well as Col-0 untransformed controls, respectively, were removed from soil and weighed immediately to determine the initial fresh weight of the plants. The plants were subsequently weighed every 15 min at room temperature in ambient conditions of approximately 40% RH. The experiment was carried out in triplicate.

Leaves were harvested from L1, L2, L3 and Col-0 plants at 0, 24, 72, 120, and 168 hpi (PM infection), 0, 24, 48, 72, and 96 hpi (B. cinerea infection), 0, 24, and 48 hpi (Pst DC3000 infection), or at 0 and 7 dpi (osmotic stress treatment). Three biological replicates were collected from each line at each time point. Total RNA was extracted using the Ultrapure RNA kit (ComWin Biotech, Beijing, China), and first-strand cDNA synthesis was carried out using TransScript^® (Transgene Biotech, Beijing, China) in a reaction volume of 20 μl, which included 200 ng total RNA, 1 μl gDNA Remover and 1 μl Anchored Oligo (dT)₁₈ Primer (0.5 μg/μl).

Subsequent quantitative real-time PCR assays were performed in triplicate in a final reaction volume of 20 μl, which included 1 μl of a sixfold cDNA dilution as template and 2× TransStart Tip Green qPCR Supermix (Transgene Biotech, Beijing, China). Gene-specific primers used in the reactions are listed in Supplementary Table S2. Assays were carried out using a CFX96 real-time PCR detection system (Bio-Rad, CFX96, USA) with thermal parameters of 94°C for 30 s, followed by 40 cycles of 94°C for 5 s and 60°C for 30 s. The A. thaliana Actin2 (TAIR: AT3G18780) gene was used as an internal reference gene (Tu et al., 2016).

To obtain insight into the potential roles of all 31 VqSTS genes in the grape genome in terms of disease resistance, we inoculated 2-year old seedlings of V. quinquangularis cv. “Shang-24” with PM and assessed STS expression via semi-quantitative real-time PCR. Ten of the VqSTS genes analyzed demonstrated significant alterations in their expression levels following challenge with this pathogen (Figure 1), including VqSTS19, VqSTS20, VqSTS28, VqSTS30, VqSTS15, VqSTS21, VqSTS36, VqSTS46, and VqSTS47. Among these genes, the expression levels of VqSTS15 and VqSTS21 increased until reaching a peak at 12 hpi, which was earlier than other genes, and then declined at subsequent time points (Figure 1). As such, we selected the VqSTS21 gene for all further functional analyses in this study.

Transgenic Arabidopsis lines constitutively expressing the VqSTS21 gene, along with untransformed plants, were inoculated with PM and assessed 7 dpi. Leaf surfaces of transgenic lines showed fewer disease symptoms than untransformed controls at 7 dpi (Figure 2A). Staining with trypan blue (Figure 2B) and NBT (Figure 2C) revealed that transgenic lines exhibited higher levels of cell death and accumulated more superoxide anions (O2–), respectively, than untransformed plants as evidenced by increased staining.

To obtain deeper insight into the pathway(s) driving these alterations in VqSTS21 transgenic plants in response to PM, we assessed the relative expression levels of several SA- (Figure 2D) and JA-responsive genes (Figure 2E) in 4-week old plants at 0, 24, 72, 120, 168 hpi, respectively. qPCR assays indicated that pathogenesis-related gene 1 (AtPR1) and isochorismate synthase 1 (AtICS1), which play major roles in the SA-dependent disease resistance response, were significantly up-regulated following inoculation with PM in untransformed control plants (Figure 2D). In transgenic plants subjected to PM inoculation, expression of AtPR1 was significantly up-regulated 12-fold compared to untransformed controls from 72 hpi onward. Similarly, while AtICS1 expression was initially down-regulated at 72 hpi compared to untransformed controls, 2.4- and 5.5-fold increases in expression compared to untransformed plants were observed at 120 and 168 hpi, respectively (Figure 2D).

Conversely, although expression of plant defensin1.2 (AtPDF1.2) and lipoxygenase-3 (AtLOX3), which is key components of the JA-mediated signaling pathway, were also enhanced following PM inoculation in both untransformed and transgenic plants, their expression levels in transgenic plants were significantly reduced compared to untransformed controls. In particular, the expression of AtPDF1.2 was decreased by 45.8-fold in transgenic lines compared to untransformed plants (Figure 2E).

Four-week old transgenic STS and untransformed plants were inoculated with PM, leaves were collected 7 dpi, and stilbenoid content was measured using HPLC. Untransformed control plants were not found to produce stilbenoids following infection, which agrees with previous studies (Yu et al., 2006; Liu et al., 2011). However, since Arabidopsis contains all of the substrates and enzymes required for stilbenoid production, with the exception of STS, transgenic lines expressing VqSTS21 were found to produce an abundance of stilbenoids. Interestingly, the vast majority of the stilbenoids produced in these plants was in the form of trans-piceid rather than resveratrol (Supplementary Figure S1). In addition, trans-piceid content in transgenic lines infected with PM was found to be 1.3-fold higher than un-inoculated transgenic lines (Table 1).

To determine whether VqSTS21 induces differential responses to particular pathogens, we assessed transgenic lines for their susceptibility to the necrotrophic fungal pathogen, B. cinerea. Three days following inoculation with the pathogen, transgenic plants exhibited more severe disease lesions than untransformed controls (Figure 3A), with a significant increase in lesion diameter (Figure 3D). Furthermore, histochemical assays indicated that transgenic lines displayed enhanced cell death (Figure 3B) and superoxide anion (O2–) production (Figure 3C) compared to untransformed plants.

Expression levels of defense-related genes were also assessed in response to B. cinerea inoculation using qPCR 0, 24, 48, 72, and 96 hpi, respectively (Figure 3E). Transcript levels of all four genes tested, including AtPR1 and AtICS1, which are key components of SA-mediated defense signaling, as well as AtPDF1.2 and AtLOX3, which are components of JA-mediated defense signaling, increased by 48 hpi in untransformed controls inoculated with B. cinerea (Figure 3E). From 48 hpi onward, transgenic lines expressing VqSTS21 displayed significant up-regulation of AtPR1 and AtICS1, with up to 87.9-fold and 14.4-fold increases in transcript levels, respectively, compared to untransformed controls. Conversely, AtPDF1.2 and AtLOX3 were significantly down-regulated from 72 and 24 hpi onward, respectively, in transgenic plants compared to untransformed controls (Figure 3E).

Transgenic lines and untransformed controls were inoculated with Pst DC3000 to investigate whether VqSTS21 was able to enhance resistance to the pathogen. Three days following inoculation, disease symptoms, as evidenced by chlorosis, were found to be more severe in untransformed control plants than transgenic plants (Figure 4A). Indeed, bacterial population assays (Figure 4D) suggested that VqSTS21 inhibited the development of Pst DC3000 in transgenic lines, resulting in significant reductions in bacterial numbers in transgenic lines compared to untransformed controls. In addition, as was the case with PM and B. cinerea infection, trypan blue (Figure 4B) and NBT staining (O2–) (Figure 4C) indicated reduced levels of cell death and accumulation of superoxide anions in response to Pst DC3000 infection in transgenic lines compared to untransformed controls.

We have previously ascertained that a distinction can be made in gene expression levels within 72 h of inoculation with this pathogen (Guo et al., 2016). Therefore, expression levels of defense-related genes were assessed in response to Pst DC3000 inoculation using qPCR 0, 24, 48, and 72 hpi, respectively (Figure 3E). In a similar fashion to PM and B. cinerea infection, expression levels of AtPR1, AtICS1, AtPDF1.2 and AtLOX3 (Figures 4E,F) were all found to be up-regulated following inoculation with Pst DC3000 by 48 hpi in untransformed controls. In transgenic lines, the JA-mediated signaling-related genes, AtPDF1.2 and AtLOX3, exhibited significant up-regulation compared to untransformed controls following inoculation from 48 hpi onward. Similarly, the SA-mediated signaling gene AtICS1 demonstrated a significant increase in transcript levels compared to untransformed controls by 48 hpi. Conversely, the remaining SA-mediated signaling gene tested, AtPR1, exhibited a significant reduction in transcript levels compared to untransformed controls at both 24 and 48 hpi, but not 72 hpi.

To investigate the response of VqSTS21 transgenic lines to osmotic stress, transgenic and Col-0 plants were exposed to salt and drought stress at three stages of development, including seeds, seedlings and mature plants. In the case of seeds, transgenic lines and Col-0 plants were grown on MS medium supplemented with 130 mM NaCl and 250 mM mannitol to simulate salt and drought stress, respectively, and rates of cotyledon greening in transgenic lines were found to be significantly higher than those of Col-0 seeds following both treatments (Figures 5A,B). This suggests that transgenic seeds acquired less damage from salt and drought stress than untransformed controls, and that VqSTS21 expression enhanced the tolerance of Arabidopsis seeds to osmotic stress.

In the case of seedlings, we assessed various physiological parameters to determine whether transgenic plants were better able to tolerate osmotic stress than untransformed controls. Since both root development and nutrient uptake are known to be negatively impacted by osmotic stress and high concentrations of ABA (de Dorlodot et al., 2007; Rowe et al., 2016), we first sought to evaluate the effect of VqSTS21 overexpression in Arabidopsis on root development. Obvious differences were noted between VqSTS21 transgenic seedlings and Col-0 seedlings following osmotic stress treatment (Figures 6A–D). Indeed, root lengths of transgenic seedlings subjected to 130 mM NaCl, 250 mM mannitol or 0.75 μM ABA were found to be significantly longer than those of Col-0 (Figure 6E), which indicates that they are capable of better tolerating osmotic stress than wild-type plants (Tardieu, 2012). To investigate the degree to which leaves were damaged by osmotic stress and whether the enhanced osmotic stress tolerance seen in transgenic lines is related to membrane permeability, chlorophyll content and relative electrolyte leakage assays were conducted, respectively (Bajji et al., 2002; Zhang et al., 2012; Tu et al., 2016). The chlorophyll content of transgenic seedlings was found to be significantly higher following salt (0.46 mg/g FW) and drought (0.56 mg/g FW) treatment than Col-0 (0.3 and 0.4 mg/g FW, respectively) plants. Electrolyte leakage was also significantly reduced in transgenic lines compared to untransformed controls following both types of osmotic stress (Figure 6F), which suggests that transgenic lines possess enhanced cell membrane integrity that could contribute to their ability to better withstand osmotic stress. Water deficit and high salinity are often associated with the production of MDA in plants, which causes membrane damage and cell death (Levine et al., 1994). In this study, we found MDA content to be significantly reduced in transgenic lines compared to untransformed controls following osmotic stress treatment (Figure 6F); a factor that could also play a role in osmotic stress resistance in these lines.

In the case of mature plants, we found that the leaves of untransformed control plants displayed substantial amounts of chlorosis and withering 7 days following the initiation of both salt and drought treatments. Conversely, in transgenic lines subjected to the same treatment, these morphological changes were greatly reduced compared to Col-0 plants (Figures 7A–D). ROS accumulation was also altered in transgenic leaves following osmotic stress treatment compared to untransformed controls, whereby transgenic lines were found to accumulate significantly lower levels of O2– than Col-0 (Figure 7E). Furthermore, the water loss rate of detached rosette leaves from 4-week old transgenic plants were significantly reduced compared to Col-0 controls (Figure 7F), which indicates that transgenic lines were less affected than Col-0 by osmotic stress.

To investigate the defense response pathways behind VqSTS21-mediated osmotic stress tolerance in Arabidopsis, we analyzed the expression of various genes in 5-week old transgenic and Col-0 plants 7 days after the initiation of salt and drought treatments (Figures 8A,B). Since ABA-mediated signaling is known to be one of the most important aspects of abiotic stress tolerance in plants (Zhang et al., 2012), we focused on the expression of genes known to be involved in this pathway. Genes included AtRD29A and AtRD29B, which are key downstream genes in the ABA-mediated response pathway, AtRD22, which is responsive to dehydration and is also involved in the ABA signaling pathway in response to abiotic stress (Huang et al., 2012), and 9-cis-epoxycarotenoid dioxygenase-3 (AtNCED3), which is an indicator of ABA biosynthesis (Iuchi et al., 2001). Following salt treatment, expression levels of AtNCED3, AtRD29A and AtRD29B were significantly higher in VqSTS21 transgenic plants than in Col-0 (Figure 8A). Furthermore, the transcript level of salt overly sensitive-2 (AtSOS2), which is a component of the salt overly sensitive (SOS) pathway (Zhu, 2002), was also significantly enhanced in transgenic lines compared to Col-0 plants (Figure 8A). This suggests that along with ABA-mediated signaling, the SOS pathway may also play a role in the enhanced salt stress tolerance seen in our VqSTS21 transgenic lines. In the case of drought stress, while the expression of AtNCED3 and AtRD29A remained unaltered, AtRD29B and AtRD22 expression was significantly increased by 4.5-fold and 3.4-fold in transgenic plants compared to control plants, respectively (Figure 8B).