Tailoring tobacco hairy root metabolism for the production of stilbenes
Tobacco hairy root (HR) cultures, which have been widely used for the heterologous production of target compounds, have an innate capacity to bioconvert exogenous t-resveratrol (t-R) into t-piceatannol (t-Pn) and t-pterostilbene (t-Pt). We established genetically engineered HR carrying the gene encoding stilbene synthase (STS) from Vitis vinifera and/or the transcription factor (TF) AtMYB12 from Arabidopsis thaliana, in order to generate a holistic response in the phenylpropanoid pathway and coordinate the up-regulation of multiple metabolic steps. Additionally, an artificial microRNA for chalcone synthase (amiRNA CHS) was utilized to arrest the normal flux through the endogenous chalcone synthase (CHS) enzyme, which would otherwise compete for precursors with the STS enzyme imported for the flux deviation. The transgenic HR were able to biosynthesize the target stilbenes, achieving a production of 40 μg L−1 of t-R, which was partially metabolized into t-Pn and t-Pt (up to 2.2 μg L−1 and 86.4 μg L−1, respectively), as well as its glucoside piceid (up to 339.7 μg L−1). Major metabolic perturbations were caused by the TF AtMYB12, affecting both primary and secondary metabolism, which confirms the complexity of biotechnological systems based on seed plant in vitro cultures for the heterologous production of high-value molecules.
Due to its wide-ranging therapeutic potential, resveratrol (3,5,4′-trihydroxy-trans-stilbene) (t-R) is amongst the most studied stilbenes. Its antiviral, antioxidant, anti-inflammatory, and cardioprotective effects have been extensively reported, as well as platelet anti-aggregation and melanoma chemoprevention activities1. The hydroxylated t-R derivative t-piceatannol (t-Pn) has significant anti-cancer and cancer chemopreventive activity but with higher bioavailability than t-R2,3, while the methoxy derivative t-pterostilbene (t-Pt) has greater anti-proliferative effects against human colon cancer cells4. The scarce distribution of these derivatives in nature calls for the devel- opment of alternative sources for their sustainable supply. In this scenario, biotechnological factories based on plant cell cultures (the so-called “green cell factory” concept) constitute a promising new biosustainable system for plant secondary metabolite (PSM) production. Similarly, hairy root (HR) cultures hold great potential as sys- tems for the bioconversion and bioproduction of PSM5, due to their genetic stability6,7 and facile scale up to biore- actor level8. Additionally, HR do not need exogenous hormones and allow the large-scale harvesting of secondary products9, in contrast with plant cell cultures, which are frequently unable to produce significant levels of target compounds8. The use of HR as biofactories has been reviewed by Chandra and Chandra10 and Rischer et al.11.The heterologous biosynthesis of t-R has been reported in microorganisms and whole plants12,13. Recently, this strategy has been successfully used in metabolically engineered grapevine cell cultures to obtain the highly bio- active stilbenes, t-Pn and t-Pt14. The capacity of engineered tobacco HR cultures to bioconvert t-R into the target derivatives has also been demonstrated15. In addition, biotechnological studies have shown that Vitis vinifera cell cultures need elicitation treatments for an efficient t-R production, the best results being achieved by combining methyl jasmonate with the expensive cyclodextrins16. Silybum marianum transgenic cell suspensions carrying the V.vinifera stilbene synthase3 (VvSTS) gene treated only with cyclodextrins have also yielded t-R17, as have HR of peanut18,19 and V. rotundifolia20.
Metabolic engineering has been used to enhance biosynthetic pathways in plants21,22, extending existing routes or introducing new ones to obtain novel compounds10,23,24. Approaches based on the modification of single steps have certain drawbacks, such as the inability to address multiple bottlenecks or simultaneously target multiple metabolites. Consequently, more effective approaches have been developed involving the overexpression of tran- scription factors (TF) that can simultaneously regulate multiple genes in a metabolic pathway25. On the other hand, the use of omics technologies has led to significant advances in the elucidation of secondary metabolism and has shed light on the competition between primary and secondary pathways26,27. In most cases TF are pres- ent as gene families and can directly activate all or several of the enzymes involved in metabolite formation in response to biotic and abiotic stress28,29. Moreover, TF allow gaps or undefined stages in metabolic routes to be reduced or eliminated upon transfer to heterologous systems30. The MYB TF family is the most useful in enhanc- ing flavonoid biosynthesis31,32, when expressed in tobacco and tomato. Notably, AtMYB12 TF increased the levels of phenolic compounds and insect resistance33,34. Recent studies have reported strong modulation in transcrip- tome behavior by MYB TF, which above all affects genes of the phenylpropanoid pathway35. In strategies aimed at increasing the production of a specific metabolite, the insertion of a single gene to complete the desired metabolic pathway can be more effective when accompanied by a TF-mediated holistic modification34.
The blocking of competitive pathways is another successfully used approach in metabolic engineering to improve PSM production. In artificial micro RNA (amiRNA) technology, which involves endoge- nous primary-miRNAs (pri-miRNAs), the miRNA and miRNA* sequences are replaced with corresponding artificial-miRNA (amiRNA/amiRNA*) sequences for the specific target gene36. The application of amiRNA in plants has led to the elucidation of gene function in the phenylpropanoid pathway34,35, as well as the identifica- tion of genes involved in plant growth and development37–41. Recently, Carbonell et al.42 developed a relativelystraightforward protocol for the fast and effective construction of amiRNAs for plant species, which opens up new possibilities for research in plant biotechnology.In the present work, we report the engineering of tobacco HR carrying the VvSTS gene to tailor the HR metab- olism for the bioproduction of t-R and its derivatives t-Pn and t-Pt. Thus, the AtMYB12 TF was overexpressed to generate a holistic response in the phenylpropanoid pathway and coordinate the up-regulation of multiple steps. In addition, the use of artificial microRNA for chalcone synthase (amiRNA-CHS) disrupted the normal flux through the endogenous CHS enzyme, which competes for precursors with the imported stilbene synthase (STS) enzyme, thereby promoting the formation of t-R (Fig. 1).ResultsEstablishment and genetic characterization of transgenic tobacco root cultures. Transgenic tobacco HR cultures were obtained by co-culture of leaf segments from in vitro tobacco plantlets with a wild type Agrobacterium rhizogenes A4 (pRiA4) strain (control line) or the engineered bacteria A. rhizogenes con- taining the plasmids [pRiA4 + pJCV52-STS3], [PRiA4 + pJCV52-AtMYB12] and [pRiA4 + pDMDC32B-At- MiR390a-B/c amiRNACHS].
The HR appeared at 2–4 weeks, after which they were cultured individually in MSsolid medium, supplemented with cefotaxime (500 mg L−1) and kanamycin (100 mg L−1) and/or hygromycin (50 mg L−1), depending on the target gene. Most of the transgenic root lines isolated after the infection with A. rhizogenes A4-S or co-infection with A4-S and A4-TF grew well under kanamycin selection, whereas only 40% of the root lines that appeared after the co-infection with the Agrobacteria A4-S, A4-TF and A4-Inh survived under joint kanamycin and hygromycin selection. After successive 2-week subcultures, the antibiotic concentration was reduced. Once the antibiotic was removed, the HR clones generated were tested for the virD gene by PCR to confirm that they were free of agrobacterium infection (Fig. 2). In general, the presence of the transgenes did not affect root morphology, which was very similar to the wild type HR. Several root clones with a high growth capacity (growth index, GI > 3.5) and displaying the typical HR phenotype were selected, and the presence of the respective transgenes in their genome was further confirmed by PCR (Fig. 2). Remarkably, only 23% of roots generated after co-infection with 3 Agrobacterium reached the aforementioned GI, probably due to the presence of both antibiotics in the culture medium, in contrast with 37% of those cultivated only with kanamycin. The verified presence of the VvSTS gene, AtMYB12 TF and the amiRNA-CHS in the selected root lines allowed us to establish four tobacco HR populations (Table 1) for further experiments.Transcriptional analysis of the introduced transgenes and their effects on the expression of phenylalanine ammonia lyase and chalcone synthase and the production of stilbenes, total phenolics and total flavonoids. Ten selected HR clones of the different types (denoted as C, S-1, S-4, S-9; TF/S-6, TF/S-8, TF/S-24 and S/TF/Inh-2, TF/S/Inh-12, TF/S/Inh-16) were transferred to a liquid medium and subcultured every two weeks.
After several rounds of subculture, samples were taken in sextuplicate for each analysis. The qRT-PCR experiments revealed transgene expression driven by the CaMV 35 S promotor in all the selected HR lines, although at variable levels (Fig. 3A). The expression of the VvSTS gene was notably low in the S/TF/Inh transgenic roots, even though we took into account the identity between the nucleotide sequence of endogenous CHS and VvSTS (70%) and verified a non-homologous zone to select the amiRNA (see Supporting Information). In contrast, as expected, no interactions between AtMB12 TF and VvSTS expression were observed, because both genes were under the control of the 35S promotor (Fig. 3A).Among the metabolic changes induced by the heterologous expression of the different transgenes, stilbene biosynthesis was driven by the VvSTS gene, resulting in 57.4 μg L−1 of total stilbenes (t-R + t-Pt + t-Pn + Piceid), which were not detected in the wild type HR (C lines; Fig. 3B). These results demonstrate that the designed bio- technological platform carrying an active STS enzyme was able to produce t-R and then metabolize it to other stil- benes by the action of unspecific tobacco enzymes constitutively present in the HR cultures. As shown in Fig. 3B, Table S2, the VvSTS gene expression was significantly correlated with the total stilbene content in the HR clones. As expected, the co-expression of the AtMYB12 TF-boosted stilbene biosynthesis and the total stilbene con- tent in TF/S-type HR lines was more than 6.9-fold higher than in the S-type. A similar increase in total phenolic compounds (TPC) (3.9-fold) and total flavonoids (TFC) (2.3-fold) also occurred in the TF/S-type roots (Fig. 3B).These results show that the TF effectively enhanced phenolic metabolism in the tobacco HR cultures (Fig. 1).
Furthermore, there was a significant positive correlation between the transcript levels of the AtMYB12 TF and those of the phenylalanine ammonia-lyase (PAL) (ccP = 0.970) and CHS (ccP = 0.967) genes from tobacco (Figs 1 and 3C), indicating that the TF could activate early and late genes involved in phenolic metabolism in the tobacco HR cultures.On the other hand, decreased levels of TFC (Fig. 3B) were observed in the HR lines carrying the amiRNA-CHS (TF/S/Inh lines), confirming an interference with the natural CHS gene expression in the biotechnological sys- tem (Fig. 3C). This was also ratified by the negative correlation between the amiRNA and CHS gene expression (ccP = −0.665) (Fig. 1).The metabolomic perturbations induced by the AtMYB12 TF. In order to examine the metabolo- mic alterations caused by the expression of the AtMYB12 TF, excluding those induced by the VvSTS gene or the amiRNA-CHS, we obtained transgenic HR clones carrying the T-DNA of A. rhizogenes together with theAtMYB12 TF. The presence of the transgene was confirmed by PCR (see Fig. S2). 1H-NMR fingerprinting coupled with principal component analysis (PCA) allowed the identification of major perturbations in the metabolome of the transgenic roots harboring the AtMYB12 TF in comparison with wild type HR (control line). To examine in depth the metabolomic changes, the large data set from the 1H-NMR was subjected to PCA to reduce the numer- ous NMR signals of the extracts (Fig. 4). A total of 79.5% of model variance was described with the principal components PC1 and PC2 (Fig. 4A), revealing significant metabolomic differences. Both groups of samples were clearly divided by component two (Fig. 4A) and the amino acid (δ 0.5–3.0 ppm), carbohydrate (δ 3.0–5.5 ppm) and aromatic (δ 5.5–9.8 ppm) regions were highly clustered in the root line with the AtMYB12 TF (Fig. 4B). The signals of compounds detected by NMR are included in the Table S3.The TF significantly affected the organic acid profile, both qualitatively and quantitatively, especially that of phenolic acids. The accumulation of two acids from the tricarboxylic acid cycle (TAC), fumaric and malic acid, as well as GABA, was notably enhanced (Fig. 5A,B, Table S3). Also, together with higher levels of cinnamic and p-coumaric acid, which are involved in the biosynthesis of stilbenes and flavonoids, the transgenic roots accu- mulated increased levels of caffeic and sinapic acid, which are eventually involved in lignin/lignan biosynthesis (Fig. 5A,B, Table S3).
Similar glutamine, leucine or threonine levels were observed in both transgenic and wild type HR; in contrast, the contents of proline and valine were significantly affected by the expression of AtMYB12 TF (Fig. 5A,B, Table S3).Bioproduction of stilbenes in elicited hairy roots. In order to investigate the effects of elicitation on the stilbene production in the designed biotechnological system, 100 μM methyl jasmonate (MeJA) and 50 mM meth- ylated β-cyclodextrin (MBCD) were added to the tobacco HR cultures after 11 days of culture and samples were taken 3 days later. This elicitor treatment has been reported as optimum for inducing t-R production in V. vinifera cell cultures16. Under the non-elicited control conditions, total stilbene accumulation was highest in roots harbor- ing both the VvSTS gene and AtMYB12 TF (TF/S lines). Specifically, the root line TF/S-24, which displayed the highest expression of both transgenes (Fig. 3A), reached a stilbene production of 468.6 μg L−1 (Fig. 6A). In most lines, the stilbene accumulation pattern was piceid > t-R > t-Pt > t-Pn. In the TF/S-type lines, the contents were 142.4–339.7 μg L−1 for piceid, and 31.3–39.9 μg L−1 for t-R; the maximum yield of t-Pt was 86.4 μg L−1, whereas t-Pn levels were significantly lower at 2.2–3.4 μg L−1 (Fig. 6A). These results show that unspecific tobacco methyl- transferases effectively bioconverted t-R into t-Pt when the AtMYB12 TF was co-expressed with STS in the root cultures. On the other hand, the lowest stilbene production occurred in the TF/S/Inh-type root lines (Fig. 6A), which may be related with the concomitant low expression of the VvSTS and AtMYB12 TF genes (Fig. 3A).The combined elicitor treatment (MeJA + MBCD) did not increase total stilbene production in the studied transgenic HR (Fig. 6A), but it profitably changed the stilbene pattern to t-Pt > t-R ≥ piceid > t-Pn, with a notable increase in the two most active derivatives, t-Pt and t-Pn, whose yields were on average 200% and 80% higher, respectively. However, as shown in Fig. 6B, the elicitor treatment was extremely effective in releasing the accumu- lated stilbenes into the culture medium, with only piceid remaining in the root tissues.
Discussion
As mentioned, tobacco is a useful model plant system, which can be easily transformed by A. rhizogenes to gen- erate HR cultures and this trait may be harnessed for the heterologous expression of foreign genes harbored in engineered A. rhizogenes strains43. This fact, along with the growing interest in t-R and its derivatives t-Pn and t-Pt14,15, prompted us to set up a production platform based on transgenic tobacco HR. Thus, together with the A. rhizogenes T-DNA, the engineered tobacco HR lines harbored the following: the VvSTS gene for heterologous t-R production; AtMYB12 TF from Arabidopsis thaliana together with VvSTS to generate a general response in the phenylpropanoid pathway and coordinate the up-regulation of multiple steps35; and amiRNA-CHS together with STS and AtMYB12 TF to limit the phenolic flux through the endogenous CHS enzyme, which competes for precursors with the STS enzyme imported for the flux redirection.
Our results confirmed the potential of engineered HR cultures for t-R production, and the natural capacity of tobacco roots to bioconvert t-R into derivatives such as piceid, t-Pn and t-Pt. The bioproduction of t-R was apparently strongly correlated with the expression level of the VvSTS gene in the different transgenic HR clones obtained.
The genetic transformation of plants with STS genes has given interesting results, including plant disease resistance44,45, and is still being intensively studied8. In this research pipeline, the first experiments were per- formed by Jeandet et al.46, who expressed the STS gene from Arachis hypogea in tobacco plants, leading to resver- atrol accumulation after treatment with short-wavelength ultraviolet light. Heterologous expression of the STS gene has subsequently been reported in several edible plant species, including tomato, kiwifruit, apple, lettuce, barley and wheat, as a defense against plague attack, based on the phytoalexin activity of resveratrol8.Several metabolic engineering strategies have also been developed for t-R production in microorganisms, but unlike plant systems, these require the introduction of a complex set of genes. In this context, the entire resveratrol pathway has been induced in modified yeast, Yarrowia lipolytica (ATCC 20362 strain), including the genes encod- ing phenylalanine/tyrosine ammonia lyase (PAL/TAL), cinnamate-4-hydroxylase (4CH), p-coumaroyl-CoA ligase (4CL), and STS47. More recently, Trantas et al.48 constructed the complete resveratrol biosynthetic pathway in S. cerevisiae to produce resveratrol from the precursor phenylalanine (10 mM), obtaining a content of 290 μg L−1 t-R after 120 h of culture. Higher yields have been achieved in engineered Escherichia coli strains; for instance, Katsuyama et al.49 reported a production of 171 mg L−1 of t-R. However, the production of t-R in microorganisms always entails the addition of exogenous precursors50.The production of t-Pn and t-Pt was recently reported by Li et al.51 in engineered yeast after feeding the culture with phenylalanine, and Wang et al.52 obtained t-Pt from t-R and p-coumaric acid in both engineered yeast and E. coli. In seed plant systems, the production of t-Pn and t-Pt has been enabled by engineering either V. vinifera plant cell cultures14 or tobacco transgenic HR15.
In both cases the heterologous genes HsCYP1B1 and VvROMT were constitutively expressed, whilst t-R was induced by elicitation in the cell cultures or supplied by direct feeding in the HR.As mentioned in the introduction, TF overexpression can boost a complete metabolic pathway3. In this con- text, the TF of the MYB family have been described as the most useful in enhancing flavonoid biosynthesis31,32. In tobacco plants, AtMYB12 TF expression causes a strong transcriptome change by up-regulating genes encoding enzymes downstream in the phenolic pathway, which increases flavonoid content, as well as the expression of several genes encoding upstream enzymes involved in the biosynthesis of common precursors in the stilbene and flavonoid biosynthetic pathways34.In our study, in agreement with the results reported by Misra et al.34, AtMYB12 TF expression was tightly correlated with an enhanced flavonoid content in the transgenic HR clones, probably because of an up-regulation of the CHS gene. It has previously been demonstrated that ectopic expression of the AtMYB12 TF increases the flavonoid content of tomato leaves and fruits33,35. In our tobacco HR, the TF also enhanced the expression of the upstream PAL, resulting in a higher total phenolic accumulation compared with the control roots. Similar results have been reported in AtMYB12 TF transgenic tobacco plants, although PAL expression was not regulated by the TF in Arabidopsis thaliana32.A more detailed view of the metabolism modulation caused by ectopic expression of MYB12 TF in transgenic HR was provided by an NMR-based metabolomics approach. Major qualitative and quantitative metabolomic changes were analyzed in roots expressing only the AtMYB12 TF and compared with wild type HR (control line). As well as amino acids and sugars, the transgene expression affected organic acids, especially phenolic acids. The accumulation of fumaric and malic acids from the tricarboxylic acid cycle was notably enhanced, and high levels of GABA were found, but contrary to previous studies of transgenic plants, citrate contents were not affected34. These results reflect that the regulation of metabolic pathways varies according to the plant organs and tissues.
Thus, while the levels of several amino acids, including valine, alanine, and phenylalanine, increased in the MYB-12 transgenic tobacco plants, in the HR system we observed the up-regulation only of valine, and the down-regulation of proline biosynthesis.Among phenolic acids, the transgenic roots accumulated higher levels of cinnamic and p-coumaric acid, which are involved in the biosynthesis of stilbenes and flavonoids, as well as caffeic and sinapic acid, which maybe involved in lignin biosynthesis. These results are in agreement with those reported for AtMYB12 TF transgenic tobacco plants by Misra et al.34, who observed an increased expression of genes involved in lignin biosynthesis. Altogether, the results suggest that part of the increased precursor availability for stilbene biosynthesis could be redirected to the lignin pathway, undermining the effectiveness of our biotechnological system for the production of the target compounds, t-R and its derivatives t-Pn and t-Pt (Fig. 5).The heterologous expression of STS can alter flower morphology, and cause male sterility in tobacco and petunia53. The sterility may be linked to a competition for substrates between STS and endogenous CHS, since fertility can be restored in tobacco by adding exogenous flavonol53. As mentioned, STS and CHS are both involved in the polyphenol pathway, and it is reasonable to assume that the expression of exogenous STS may lead to a competition for substrates. Indeed, among the analyzed tobacco HR lines, those transformed with amiRNA-CHS achieved the lowest levels of flavonoids due to inhibition of the CHS gene expression. By silencing the branching point of the flavonoid biosynthetic pathway, a greater flow of precursors was channeled toward the production of t-R and its derivatives. Unfortunately, the transgenic root lines carrying the amiRNA-CHS also showed a low VvSTS gene expression and their stilbene contents were lower than in the TF/S root clones.
Differences in trans- gene expression among HR clones can be attributed to transgene rearrangements, which may reduce their expres- sion54. It is known that exogenous DNA sequences can undergo recombination, rearrangement and truncation before or during integration but not afterwards55, which could have a significant effect on transgene expression. Random transgene insertion into the plant genome is another possible reason for the differences in transgene expression among HR lines56,57.We have previously discussed that elicitation can improve the heterologous production of a target compound in plant cell cultures58,59. Elicitation of the engineered HR with 50 mM MBCD and 100 µM of MeJA significantly increased the release of t-R and its derivatives t-Pn and t-Pt to the liquid medium, whereas piceid remained mainly inside the roots. In contrast, the positive effects of the combined elicitor treatment (MBCD + MeJA) on the t-R production has been previously reported in non-transgenic and transgenic grapevine cell cultures14,60, and also in transgenic S. marianum cell cultures carrying the VvSTS gene and supplemented with MBCD17. Theineffectiveness of elicitation in the HR culture system to improve the total stilbenoid contents was perhaps due to the fast metabolization of stilbenoids in the culture medium, as reported previously15, and an increased biocon- version of t-R into the more bioactive derivatives, t-Pn and t-Pt.
Conclusions
Taken as a whole, our results show the suitability of the engineered HR cultures for the heterologous production of t-R and its derivatives t-Pn and t-Pt. The effectiveness of the AtMYB12 TF for activating the phenolic pathway and amiRNA CHS for blocking competitive pathways was also demonstrated. However, these results also show the extreme complexity of metabolically engineering biotechnological systems based on in vitro seed plants. A large part of the t-R biosynthesized by the root cultures was metabolized not only to the target compounds t-Pn and t-Pt, but also to piceid and probably other non-identified resveratrol derivatives. Moreover, the boosting of phenolic metabolism, reflected by the high levels of caffeic and sinapic acid in the HR lines harboring the AtMYB12 TF, could also provide precursors for competitive pathways, such as lignan and lignin biosynthesis. Therefore, new approaches are required in which competitive upstream pathways are blocked by specific amiR- NAs and the carbon flux is more effectively channeled to stilbene biosynthesis.Bacteria and plasmids. Four strains of Agrobacterium rhizogenes A4 were used: wild type and three engi- neered strains carrying the pRiA4. The specific engineered strains and binary plant expression vectors of each one are described in Table 2 and Fig. S1.
The VvSTS gene (Acc. XM_002264953) was cloned as described by Martinez-Marquez et al.14. The clone TOPO-U04-A02 that contains the AtMYB12 gene was purchased from the Arabidopsis Biological Resource Center (ABRC). An LR recombination reaction was performed to generate an expression clone using LR Clonase (Invitrogen, Carlsbad, CA). The construction of amiRNA-CHS (see Supp information) was as described by Carbonell et al.42 using the following sequences: forward oligo: 5′TGT ATT AAT CAT TGA TTT TTC ACA GAT GAT GAT CAC ATT CGT TAT CTA TTT TTT CTG TGA AAA AGC AAT GAT TAA 3′; reverse oligo: 5′ATG TTA ATC ATT GCT TTT TCA CAG AAA AAA TAG ATA ACG AAT GTG ATC ATC ATC TGT
GAA AAA TCA ATG ATT AA 3′. The Agrobacterium strains were transformed by electroporation as described by Shaw61.Stable transformation and hairy root culture induction. Leaf discs of Nicotiana tabacum cv Xhanti plantlets grown in vitro on Murashige and Skoog (MS) medium were co-infected as described by Gallois and Marinho62, using a wild type A. rhizogenes A4 or A4-S or A4-S plus A4-TF or A4-S plus A4-TF plus A4-Inh. After 2–4 weeks, HR appeared in leaf discs maintained on MS solid medium with 30 g L−1 of sucrose and 500 mg L−1 cefotaxime to eliminate the agrobacteria. Further, HR were excised and individually cultured on the same medium plus kanamycin (50 mg L−1) and/or hygromycin (50 mg L−1) for selection, depending on the target genes (i.e. HR obtained by co-infection with 3 Agrobacterium was cultivated for selection with kanamycin and hygro- mycin). The tobacco HR were kept in these conditions at 25 °C in the dark for several rounds of subculturing until the elimination of agrobacteria and confirmation by PCR analysis.
PCR analysis. The analysis was performed with DreamTaq Green PCR Master Mix (Thermo Fisher Scientific Inc). Previously, genomic DNA was isolated as described by Dellaporta et al.63. The specific primers and ampli- fication reactions used are described in Table S1. PCR products were analyzed by electrophoresis on 1% agarose gels.Gene expression by qPCR analysis. Expression of the VvSTS, AtMYB12, amiRNA-CHS, PAL and CHS genes was verified by qPCR in selected HR lines as described by Hidalgo et al.15. Briefly, total RNA was iso- lated with TRIzol reagent (Invitrogen, Carlsbad, CA) and treated with DNase I (Invitrogen, Carlsbad, CA). Then, cDNA was synthesized with SuperScript III reverse transcriptase (Invitrogen, Carlsbad, CA) and finally, qRT-PCR was performed with iTAqTM Universal SYBR Green Supermix (BioRad, Hercules, CA, EEUU) in a 384-well platform system (LightCycler 480 Instrument; Roche). Conditions and primers used are described in Table S1.Stilbene extraction and determination. The HR and culture medium were processed as described by Hidalgo et al.15. Briefly, two times one volume of ethyl acetate was added per four volumes of the medium, stirring vigorously, and the non-polar solvent recovered and evaporated. For the HR, 50 mg of freeze-dried powder was extracted twice with 2 volumes of 100% methanol, sonicated for 30 min, and the supernatant was collected and evaporated. Stilbenes were determined by a Linear Ion Trap Quadrupole LC/MS/MS Mass Spectrometer, 4000 Q TRAP of AB Sciex Instruments with MRM scan type in negative mode as described by Hidalgo et al.15.
Total phenolic content (TPC). The compounds were extracted as indicated in the previous paragraph and their quantity was estimated by the Folin-Ciocalteu colorimetric method as described by Thiruvengadam et al.64 with slight modifications. Briefly, 100 µL of HR extracts were mixed with 3.1 mL of distilled water, followed by addition of 0.2 mL Folin-Ciocalteu reagent. After 5 min, 0.6 mL of 20% sodium carbonate solution was added, and after 60 min of incubation the solution absorbance was measured at 760 nm. The concentration was calculated as mg of gallic acid equivalents using a calibration curve.Total flavonoid content (TFC). The flavonoid quantity in extracts was estimated using the aluminum chlo- ride colorimetry method described by Thiruvengadam et al.64 with slight modifications. Briefly, 200 µL of extract, 100 µL of 10% (w/v) aluminum chloride solution, 100 µL of 1 M potassium acetate solution, and 4.6 mL of distilled water were mixed. After 30 min of incubation the solution absorbance was measured at 415 nm. The concentra- tion was calculated as mg of rutin equivalents using a calibration curve.Sample preparation, NMR conditions and data analysis. Six biological replicates of HR were treated as described by Zahmanov et al.65. Briefly, in a 2 mL tube 50 mg freeze-dried HR powder, 750 µL of CD3OD and 750 µL of D2O (KH2PO4, buffer, pH 6.0, containing 0.01% w/v TSPA-d4) were mixed. The mixture was homog- enized, sonicated for 20 min, and centrifuged for 20 min at 12,000 rpm. Finally, 800 µL was placed into the NMR tube. The proton spectra (1H NMR) were recorded at 25 °C on an AVII+ 600 spectrometer (Bruker, Karlsruhe,Germany) at a proton frequency of 600.13 MHz with 4.07 s relaxation time and CD3OD as the internal lock66. The data analysis was processed as described by Marchev et al.67. Briefly, the spectra were phased, baseline corrected, set to TSPA at 0.0 ppm and binned to 0.04 ppm using MestReNova software (version 6.2.1, Mestrelab Research, Santiago de Compostela, Spain). The principal component analysis (PCA) was performed with SIMCA-P14.0 (Umetrics, Umea, Sweden), excluding the signals of water Piceatannol and methanol.