Enhanced Stilbene Production and Excretion in Vitis vinifera cv Pinot Noir Hairy Root Cultures

Stilbenes are defense molecules produced by grapevine in response to stresses including various elicitors and signal molecules. Together with their prominent role in planta, stilbenes have been the center of much attention in recent decades due to their pharmaceutical properties. With the aim of setting up a cost-effective and high purity production of resveratrol derivatives, hairy root lines were established from Vitis vinifera cv Pinot Noir 40024 to study the organ-specific production of various stilbenes. Biomass increase and stilbene production by roots were monitored during flask experiments. Although there was a constitutive production of stilbenes in roots, an induction of stilbene synthesis by methyl jasmonate (MeJA) after 18 days of growth led to further accumulation of ε-viniferin, δ-viniferin, resveratrol and piceid. The use of 100 µM MeJA after 18 days of culture in the presence of methyl-β-cyclodextrins (MCDs) improved production levels, which reached 1034 µg/g fresh weight (FW) in roots and 165 mg/L in the extracellular medium, corresponding to five-and 570-fold increase in comparison to control. Whereas a low level of stilbene excretion was measured in controls, addition of MeJA induced excretion of up to 37% of total stilbenes. The use of MCDs increased the excretion phenomenon even more, reaching up to 98%. Our results demonstrate the ability of grapevine hairy roots to produce various stilbenes. This production was significantly improved in response to elicitation by methyl jasmonate and/or MCDs. This supports the interest of using hairy roots as a potentially valuable system for producing resveratrol derivatives.


Introduction
Stilbenes have been the center of much attention over recent decades due to their valuable biological properties. The most studied stilbene, resveratrol, has been described as an important plant defense molecule, with antifungal and antibacterial activities [1][2][3][4]. Although stilbenes are found in grapevines at levels considerably lower than flavonoid-type compounds, they possess outstanding pharmaceutical properties [5] as antioxidants [6] as well as anticancer agents [7,8]. Resveratrol derivatives such as viniferins also show an interesting therapeutic potential, mainly in cancer treatment [9,10]. The main natural sources of resveratrol available on the market result from its extraction from Japanese knotweed (Falliopa japonica), which produces large quantities of piceid (a β-D-resveratrol glucoside), which is then de-glycosylated to obtain resveratrol [11,12]. Consumers' (Figure 1a). These roots were probably adventitious roots due to the very well-known rooting ability of grapevine and as a result of the addition of 2,4-dichlorophenoxyacetic acid in the bacterial suspension used to infect the explants. A second series of roots appeared on the same Pinot Noir explants fourweeks after infection with R. rhizogenes ATCC 15834 using the two previous methods. Once they were 1-2 cm long, roots were isolated and cultivated on solid B5 medium supplemented with 3% sucrose (B5-30). These roots showed a completely different phenotype from the first ones, with a diameter ranging from 1 to 3 mm and a higher branching rate (Figure 1a,b). This phenotype showed some similarity to that described on Muscadine grape, with thick roots, very short root hairs, and a higher branching rate (10-15 branches per 5 cm) compared to the 4-5 branches per 5 cm for Muscadine HRs, as reported in a previous work [24]. In the second series of roots that appeared, 80 different lines were isolated and cultivated on B5-30 solid medium. The 60 lines displaying the best growth rates were then transferred to a B5-30 liquid medium (Figure 1c) to be further selected on growth capacity. Nine lines were thus conserved and their transformation was verified by confirmation of the presence of transferred rolC gene and the absence of virD2 bacterial virulence gene ( Figure S1). As few differences in terms of phenotype, growth rate and phytochemical profile were observed between these lines (data not shown), the root line 19A was chosen as representative and used further.

Effects of Culture Medium and Sucrose Concentration on HR Growth
The composition of the medium used is known to have an important impact on the growth capacity and secondary metabolite production of in vitrocultures. The observed slow growth of the selected lines strongly suggested that the medium composition could be further improved. To search for better culture conditions, fourcommonly used culture media, Gamborg

Effects of Culture Medium and Sucrose Concentration on HR Growth
The composition of the medium used is known to have an important impact on the growth capacity and secondary metabolite production of in vitro cultures. The observed slow growth of the selected lines strongly suggested that the medium composition could be further improved. To search compared. In addition, six different sucrose concentrations ranging from 1% to 6% were tested ( Figure 2). Average growth rate µ and its standard error were calculated from weight measurements made over 35 days with renewal of culture media every three or four days. Each point is composed of threebiological replicates. Statistical differences were tested using ANOVA (α = 5%). Letters above dataset point out statistically different groups.
A preference for 1/2 media seems to occur regardless of the medium tested ( Figure 2a). Only the B5 and MC media enabled growth when non-diluted. Amongst the 1/2 media, it would seem that the SH medium yielded the best growth rate (µ), reaching µ = 0.063 day −1 (doubling time, Td = 10.9 days), along with B5. Determination of the growth rate using different sucrose concentrations in the B5 medium ( Figure 2b) showed that the range of 1% to 3% sucrose concentrations in the medium seems to bethe most effective, reaching a growth rate of 0.046 day −1 (doubling time Td = 14.9 days) for 2% sucrose. As well as showing a low growth rate, the roots grown on medium with 5% and 6% sucrose started browning after threeweeks, suggesting significant stress under these conditions. This effect could be caused by an osmotic stress due to high sugar concentrations [33]. Thus, 1/2 SH medium with 2% sucrose was used for further experiments.

Biomass Production
The growth kinetics were evaluated using line 19A in the improved growth conditions of 1/2 SH medium with 2% sucrose (Figure 3).  Average growth rate µ and its standard error were calculated from weight measurements made over 35 days with renewal of culture media every three or four days. Each point is composed of three biological replicates. Statistical differences were tested using ANOVA (α = 5%). Letters above dataset point out statistically different groups.
A preference for 1/2 media seems to occur regardless of the medium tested ( Figure 2a). Only the B5 and MC media enabled growth when non-diluted. Amongst the 1/2 media, it would seem that the SH medium yielded the best growth rate (µ), reaching µ = 0.063 day −1 (doubling time, Td = 10.9 days), along with B5. Determination of the growth rate using different sucrose concentrations in the B5 medium ( Figure 2b) showed that the range of 1% to 3% sucrose concentrations in the medium seems to be the most effective, reaching a growth rate of 0.046 day −1 (doubling time Td = 14.9 days) for 2% sucrose. As well as showing a low growth rate, the roots grown on medium with 5% and 6% sucrose started browning after three weeks, suggesting significant stress under these conditions. This effect could be caused by an osmotic stress due to high sugar concentrations [33]. Thus, 1/2 SH medium with 2% sucrose was used for further experiments.

Biomass Production
The growth kinetics were evaluated using line 19A in the improved growth conditions of 1/2 SH medium with 2% sucrose (Figure 3). compared. In addition, six different sucrose concentrations ranging from 1% to 6% were tested ( Figure 2). Effects of media and sucrose concentration on HR growth rate: (a) effect of various standard media and 1/2 dilutions, with 3% sucrose on the growth rate of line 16C; and (b) effect of sucrose concentrations in B5 medium on the growth of line 19A. Average growth rate µ and its standard error were calculated from weight measurements made over 35 days with renewal of culture media every three or four days. Each point is composed of threebiological replicates. Statistical differences were tested using ANOVA (α = 5%). Letters above dataset point out statistically different groups.
A preference for 1/2 media seems to occur regardless of the medium tested ( Figure 2a). Only the B5 and MC media enabled growth when non-diluted. Amongst the 1/2 media, it would seem that the SH medium yielded the best growth rate (µ), reaching µ = 0.063 day −1 (doubling time, Td = 10.9 days), along with B5. Determination of the growth rate using different sucrose concentrations in the B5 medium ( Figure 2b) showed that the range of 1% to 3% sucrose concentrations in the medium seems to bethe most effective, reaching a growth rate of 0.046 day −1 (doubling time Td = 14.9 days) for 2% sucrose. As well as showing a low growth rate, the roots grown on medium with 5% and 6% sucrose started browning after threeweeks, suggesting significant stress under these conditions. This effect could be caused by an osmotic stress due to high sugar concentrations [33]. Thus, 1/2 SH medium with 2% sucrose was used for further experiments.

Biomass Production
The growth kinetics were evaluated using line 19A in the improved growth conditions of 1/2 SH medium with 2% sucrose (Figure 3).  The semi-log representation of the growth curve ( Figure 3a) of grapevine HRs shows no significant lag phase, starting directly with an exponential growth phase up to Day 11, followed by a slowing phase until Day 31 and a stationary phase up to Day 35. The rapid diminution in growth could be due to the relatively low amounts of sucrose in the culture medium. Figure 3b displays a rapid fall in the pH during the first days of culture up to 4-7 days. The pH then rises to 5 at day 20 before stabilizing, in the phases corresponding to both the slower growth and the stationary phases. This suggests that pH can be used as a rapid, non-destructive tool to monitor growth in grapevine HR cultures.
Growth rate was calculated between Day 0 and Day 11, giving a µ value of 0.11 day −1 (Td = 6.2 days). In comparison, peanut and Muscadine HRs have been reported to have a doubling time of 10.2 and 10.7 days, respectively, without any addition of growth regulators [23,24].

Constitutive Stilbene Production in Roots
The analysis of stilbenes during the time course experiments shows that a basal production takes place in roots (Figure 4), where they probably play a role as phytoanticipins as already mentioned in natural grapevine roots as well as in the perennial parts of the plant [34]. The semi-log representation of the growth curve ( Figure 3a) of grapevine HRs shows no significant lag phase, starting directly with an exponential growth phase up to Day 11, followed by a slowing phase until Day 31 and a stationary phase up to Day 35. The rapid diminution in growth could be due to the relatively low amounts of sucrose in the culture medium. Figure 3b displays a rapid fall in the pH during the first days of culture up to 4-7 days. The pH then rises to 5 at day 20 before stabilizing, in the phases corresponding to both the slower growth and the stationary phases. This suggests that pH can be used as a rapid, non-destructive tool to monitor growth in grapevine HR cultures.
Growth rate was calculated between Day 0 and Day 11, giving a µ value of 0.11 day −1 (Td = 6.2 days). In comparison, peanut and Muscadine HRs have been reported to have a doubling time of 10.2 and 10.7 days, respectively, without any addition of growth regulators [23,24].

Constitutive Stilbene Production in Roots
The analysis of stilbenes during the time course experiments shows that a basal production takes place in roots (Figure 4), where they probably play a role as phytoanticipins as already mentioned in natural grapevine roots as well as in the perennial parts of the plant [34].  Here, only the resveratrol glucoside, piceid, resveratrol, and the two dehydrodimers, ε-viniferin and δ-viniferin, were quantified using external standards. Further use of the term "total stilbenes" refers to the sum of all these molecules. A close relationship between the growth phase and the production of secondary metabolites in roots can be observed, especially when the metabolism tends to switch from primary to secondary at the end of the exponential growth phase. As shown in Figure  4a, the basal production of stilbenes per flask rises slowly until the end of the exponential phase at Day 21, before rapidly increasing up to 175 µg at Day 25 and 217 µg after 35 days. It is interesting to note the same trend in the proportion of excreted stilbenes (Figure 4b). The low proportion of 1%-5% excreted stilbenes vs. total stilbenes is consistent with those measured in similar systems without elicitation [23,24].

Elicitation with Methyl Jasmonate
Stilbene production, and particularly that of resveratrol, is known to be strongly induced by various environmental stresses, signaling molecules and drug transporters [19,35]. Such compounds have been used in various plant cell systems to mimic a stress and thus induce the production of numerous metabolites of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate (MeJA), has been shown to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. Here, only the resveratrol glucoside, piceid, resveratrol, and the two dehydrodimers, ε-viniferin and δ-viniferin, were quantified using external standards. Further use of the term "total stilbenes" refers to the sum of all these molecules. A close relationship between the growth phase and the production of secondary metabolites in roots can be observed, especially when the metabolism tends to switch from primary to secondary at the end of the exponential growth phase. As shown in Figure 4a, the basal production of stilbenes per flask rises slowly until the end of the exponential phase at Day 21, before rapidly increasing up to 175 µg at Day 25 and 217 µg after 35 days. It is interesting to note the same trend in the proportion of excreted stilbenes (Figure 4b). The low proportion of 1%-5% excreted stilbenes vs. total stilbenes is consistent with those measured in similar systems without elicitation [23,24].

Elicitation with Methyl Jasmonate
Stilbene production, and particularly that of resveratrol, is known to be strongly induced by various environmental stresses, signaling molecules and drug transporters [19,35]. Such compounds have been used in various plant cell systems to mimic a stress and thus induce the production of numerous metabolites of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate (MeJA), has been shown to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. Figure 5 shows the response in terms of stilbene production in roots and in the culture medium following treatment with 100 µM MeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) (corresponding to the alcoholic content upon addition of MeJA) to the control cultures shows no significant effect of the solvent on root or medium stilbene contents. All stilbene production values in roots were expressed here as micrograms per gram of fresh weight. In order to relate the obtained data with root dry weight, we determined during replicate experiments the ratio dry root weight over fresh root weight which was in the order of 16% ± 7%. As a result, stilbene production in µg/g DW can be obtained by multiplying stilbene production in µg/g fresh weight (FW) by a factor of around 7. 2016, 21, 12 6 of 16 Figure 5 shows the response in terms of stilbene production in roots and in the culture medium following treatment with 100 µMMeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) (corresponding to the alcoholic content upon addition of MeJA) to the control cultures shows no significant effect of the solvent on root or medium stilbene contents. All stilbene production values in roots were expressed here as micrograms per gram of fresh weight. In order to relate the obtained data with root dry weight, we determined during replicate experiments the ratio dry root weight over fresh root weight which was in the order of 16% ± 7%. As a result, stilbene production in µg/g DW can be obtained by multiplying stilbene production in µg/g fresh weight (FW) by a factor of around 7. Figure 5a displays a basal stilbene production in roots of around 200 µg/g FW, reaching up to 390 µg/g FW 10 days after elicitation. The low stilbene concentrations measured in the extracellular medium, reaching a maximum of 2.3 mg/L at Day 10, are consistent with those previously described [23,24]. However, application of MeJA at doses of 100 µM and 200 µM has a clear inducing effect on the stilbene contents of both the roots and the extracellular medium. Unlike described in peanut hairy roots treated with MeJA, the stilbene concentration in the culture medium didnot reach its maximum after 12-48 h and no fall in concentration was observed during the tested period [22]. One hundred micromoles ofMeJA (Figure 5c) gives rise to a quick increase in the stilbene root content between twoand four days after induction, reaching 757 µg/g FW and followed by a steady increase up to 1126 µg/g FW after 10 days, corresponding to 2.2-and 2.9-foldincreases respectively in comparison to control. have been used in various plant cell systems to mimic a stress and thus induce the production of numerous metabolites of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate (MeJA), has been shown to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. Figure 5 shows the response in terms of stilbene production in roots and in the culture medium following treatment with 100 μMMeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) (corresponding to the alcoholic content upon addition of MeJA) to the control cultures shows no significant effect of the solvent on root or medium stilbene contents. All stilbene production values in roots were expressed here as micrograms per gram of fresh weight. In order to relate the obtained data with root dry weight, we determined during replicate experiments the ratio dry root weight over fresh root weight which was in the order of 16% ± 7%. As a result, stilbene production in μg/g DW can be obtained by multiplying stilbene production in μg/g fresh weight (FW) by a factor of around 7. c d e f Figure 5a displays a basal stilbene production in roots of around 200 µ g/g FW, reaching up to 390 µ g/g FW 10 days after elicitation. The low stilbene concentrations measured in the extracellular medium, reaching a maximum of 2.3 mg/L at Day 10, are consistent with those previously described ); ε-vin: ε-viniferin ( have been used in various plant cell systems to mimic a stress and thus induce the production of numerous metabolites of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate (MeJA), has been shown to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. Figure 5 shows the response in terms of stilbene production in roots and in the culture medium following treatment with 100 μMMeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) (corresponding to the alcoholic content upon addition of MeJA) to the control cultures shows no significant effect of the solvent on root or medium stilbene contents. All stilbene production values in roots were expressed here as micrograms per gram of fresh weight. In order to relate the obtained data with root dry weight, we determined during replicate experiments the ratio dry root weight over fresh root weight which was in the order of 16% ± 7%. As a result, stilbene production in μg/g DW can be obtained by multiplying stilbene production in μg/g fresh weight (FW) by a factor of around 7. c d e f Figure 5a displays a basal stilbene production in roots of around 200 µ g/g FW, reaching up to 390 µ g/g FW 10 days after elicitation. The low stilbene concentrations measured in the extracellular medium, reaching a maximum of 2.3 mg/L at Day 10, are consistent with those previously described ); δ-vin: δ-viniferin ( have been used in various plant cell systems to mimic a stress and thus induce the production of numerous metabolites of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate (MeJA), has been shown to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. Figure 5 shows the response in terms of stilbene production in roots and in the culture medium following treatment with 100 μMMeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) (corresponding to the alcoholic content upon addition of MeJA) to the control cultures shows no significant effect of the solvent on root or medium stilbene contents. All stilbene production values in roots were expressed here as micrograms per gram of fresh weight. In order to relate the obtained data with root dry weight, we determined during replicate experiments the ratio dry root weight over fresh root weight which was in the order of 16% ± 7%. As a result, stilbene production in μg/g DW can be obtained by multiplying stilbene production in μg/g fresh weight (FW) by a factor of around 7. c d e f Figure 5a displays a basal stilbene production in roots of around 200 µ g/g FW, reaching up to 390 µ g/g FW 10 days after elicitation. The low stilbene concentrations measured in the extracellular medium, reaching a maximum of 2.3 mg/L at Day 10, are consistent with those previously described ); Total ( ) (sum of previous stilbenes). Each point represents the mean with standard deviation of three biological replicates. Figure 5a displays a basal stilbene production in roots of around 200 µg/g FW, reaching up to 390 µg/g FW 10 days after elicitation. The low stilbene concentrations measured in the extracellular medium, reaching a maximum of 2.3 mg/L at Day 10, are consistent with those previously described [23,24]. However, application of MeJA at doses of 100 µM and 200 µM has a clear inducing effect on the stilbene contents of both the roots and the extracellular medium. Unlike described in peanut hairy roots treated with MeJA, the stilbene concentration in the culture medium did not reach its maximum after 12-48 h and no fall in concentration was observed during the tested period [22]. One hundred micromoles of MeJA (Figure 5c) gives rise to a quick increase in the stilbene root content between two and four days after induction, reaching 757 µg/g FW and followed by a steady increase up to 1126 µg/g FW after 10 days, corresponding to 2.2-and 2.9-fold increases respectively in comparison to control.
In the extracellular medium (Figure 5d), 100 µM MeJA causes a similar effect, reaching stilbene contents of 5 and 10 mg/L after four and 10 days of treatment, corresponding to 12.6-and 4.3-fold increases, respectively, in comparison to control. Application of 200 µM MeJA shows a comparable effect on the total stilbene root content up to Day 4, reaching 650 µg/g FW (Figure 5e), although stabilizing afterwards. The browning of the roots observed upon treatment with 200 µM MeJA suggests a possible effect of this compound on root viability and biomass, as previously reported in Brassica hairy root cultures [48]. However, only the treatment with 100 µM MeJA displays statistically significant differences in biomass in comparison to the control 10 days after elicitation. Quantification of the corresponding excreted stilbenes again shows a response similar to the use of 100 µM MeJA until Day 4, reaching 3.5 mg/L. The stilbene concentration then rises to 19 mg/L at Day 10 (Figure 5f), which corresponds to 8.8-and 8.1-fold increases, respectively, in comparison to control. This observation is probably due to a release of intracellular stilbenes into the medium. All the data indicate that four days following MeJA treatment is the best timing, i.e. when induction is at a maximum and to minimize a possible detrimental effect of MeJA on root viability. For further experiments, stilbene measurements were thus carried out four days after MeJA induction. Excretion rates of stilbenes are also significantly affected by MeJA treatments (Figure 6).
While the control excretion rate rises slowly from 1% to 11% of total stilbenes in 10 days, the cultures with MeJA display excretion rates of 31% and 37% for 100 µM and 200 µM MeJA, respectively, after 10 days. No piceid was found in the medium of MeJA-treated roots and less than 10% of δ-viniferin was excreted while almost 30% ε-viniferin and 60% of resveratrol produced were found in the culture medium ( Figure S2).

2016, 21, 12 7 of 16
In the extracellular medium (Figure 5d), 100 µM MeJA causes a similar effect, reaching stilbene contents of 5 and 10 mg/L after fourand 10 days of treatment, corresponding to 12.6-and 4.3foldincreases, respectively, in comparison to control. Application of 200 µM MeJA shows a comparable effect on the total stilbene root content up to Day 4, reaching 650 µg/g FW (Figure 5e), although stabilizing afterwards. The browning of the roots observed upon treatment with 200µM MeJA suggests a possible effect of this compound on root viability and biomass, as previously reported in Brassica hairy root cultures [48]. However, only the treatment with 100 µM MeJA displays statistically significant differences in biomass in comparison to the control 10 days after elicitation. Quantification of the corresponding excreted stilbenes again shows a response similar to the use of 100 µM MeJA until Day 4, reaching 3.5 mg/L. The stilbene concentration then rises to 19 mg/L at Day 10 (Figure 5f), which corresponds to 8.8-and 8.1-fold increases, respectively, in comparison to control. This observation is probably due to a release of intracellular stilbenes into the medium. All the data indicate that four days following MeJA treatment is the best timing, i.e. when induction is at a maximum and to minimize a possible detrimental effect of MeJA on root viability. For further experiments, stilbene measurements were thus carried out four days after MeJA induction. Excretion rates of stilbenes are also significantly affected by MeJA treatments (Figure 6).
While the control excretion rate rises slowly from 1% to 11% of total stilbenes in 10 days, the cultures with MeJA display excretion rates of 31% and 37% for 100 µM and 200 µM MeJA, respectively, after 10 days. No piceid was found in the medium of MeJA-treated roots and less than 10% of δ-viniferin was excreted while almost 30% ε-viniferin and 60% of resveratrol produced were found in the culture medium ( Figure S2).

Elicitation Experiments with Cyclodextrins and/or Methyl Jasmonate
Cyclodextrins (CDs) and particularly methyl-β-CDs (MCDs) have also been described as strong inducers of stilbene production in grapevine cell suspensions [49]. They have been reported to display a synergistic effect when used in combination with MeJA [35,41,[50][51][52][53]. However, their effect on the production of stilbenes in grapevine HRs is unknown as yet. The effect of CAVASOL ® W-7 MCDs alone or in combination with MeJA was evaluated on both root growth ( Figure 7) and stilbene content (Figure 8). UPLC chromatograms are displayed in Figure S3.   23,24]. However, application of MeJA at doses of 100 µ M and 200 µ M has a clear inducing effect on the stilbene contents of both the roots and the extracellular medium. Unlike described in peanut hairy roots treated with MeJA, the stilbene concentration in the culture medium didnot reach its maximum after 12-48 h and no fall in concentration was observed during the tested period [22]. One hundred micromoles ofMeJA (Figure 5c) gives rise to a quick increase in the stilbene root content between twoand four days after induction, reaching 757 µ g/g FW and followed by a steady increase up to 1126 µ g/g FW after 10 days, corresponding to 2.2-and 2.9-foldincreases respectively in comparison to control.
In the extracellular medium (Figure 5d), 100 µ M MeJA causes a similar effect, reaching stilbene contents of 5 and 10 mg/L after fourand 10 days of treatment, corresponding to 12.6-and 4.3foldincreases, respectively, in comparison to control. Application of 200 µ M MeJA shows a comparable effect on the total stilbene root content up to Day 4, reaching 650 µ g/g FW (Figure 5e), although stabilizing afterwards. The browning of the roots observed upon treatment with 200µ M MeJA suggests a possible effect of this compound on root viability and biomass, as previously reported in Brassica hairy root cultures [48]. However, only the treatment with 100 µ M MeJA displays statistically significant differences in biomass in comparison to the control 10 days after elicitation. Quantification of the corresponding excreted stilbenes again shows a response similar to the use of 100 µ M MeJA until Day 4, reaching 3.5 mg/L. The stilbene concentration then rises to 19 mg/L at Day 10 (Figure 5f), which corresponds to 8.8-and 8.1-fold increases, respectively, in comparison to control. This observation is probably due to a release of intracellular stilbenes into the medium. All the data indicate that four days following MeJA treatment is the best timing, i.e. when induction is at a maximum and to minimize a possible detrimental effect of MeJA on root viability. For further experiments, stilbene measurements were thus carried out four days after MeJA induction. Excretion rates of stilbenes are also significantly affected by MeJA treatments (Figure 6).
While the control excretion rate rises slowly from 1% to 11% of total stilbenes in 10 days, the cultures with MeJA display excretion rates of 31% and 37% for 100 µ M and 200 µ M MeJA, respectively, after 10 days. No piceid was found in the medium of MeJA-treated roots and less than 10% of δ-viniferin was excreted while almost 30% ε-viniferin and 60% of resveratrol produced were found in the culture medium ( Figure S2).

Elicitation Experiments with Cyclodextrins and/or Methyl Jasmonate
Cyclodextrins (CDs) and particularly methyl-β-CDs (MCDs) have also been described as strong inducers of stilbene production in grapevine cell suspensions [49]. They have been reported to display a synergistic effect when used in combination with MeJA [35,41,[50][51][52][53]. However, their effect on the production of stilbenes in grapevine HRs is unknown as yet. The effect of CAVASOL ® W-7 MCDs alone or in combination with MeJA was evaluated on both root growth ( Figure 7) and stilbene content (Figure 8). UPLC chromatograms are displayed in Figure S3.
) or 200 µM MeJA ( us plant cell systems to mimic a stress and thus induce the production of interest [14,36,37]. Amongst them, the jasmonic acid ester, methyl jasmonate to induce stilbene production mainly in grapevine cell suspensions [38][39][40][41][42][43][44][45][46][47]. esponse in terms of stilbene production in roots and in the culture medium 100 μMMeJA up to 10 days after induction. Addition of 0.1% EtOH (v/v) coholic content upon addition of MeJA) to the control cultures shows no olvent on root or medium stilbene contents. All stilbene production values ere as micrograms per gram of fresh weight. In order to relate the obtained ht, we determined during replicate experiments the ratio dry root weight hich was in the order of 16% ± 7%. As a result, stilbene production in μg/g multiplying stilbene production in μg/g fresh weight (FW) by a factor of

Elicitation Experiments with Cyclodextrins and/or Methyl Jasmonate
Cyclodextrins (CDs) and particularly methyl-β-CDs (MCDs) have also been described as strong inducers of stilbene production in grapevine cell suspensions [49]. They have been reported to display a synergistic effect when used in combination with MeJA [35,41,[50][51][52][53]. However, their effect on the production of stilbenes in grapevine HRs is unknown as yet. The effect of CAVASOL ® W-7 MCDs alone or in combination with MeJA was evaluated on both root growth ( Figure 7) and stilbene content (Figure 8). UPLC chromatograms are displayed in Figure S3. CD concentrations used were chosen according to previous reports showing high stilbene production levels using grapevine cell suspensions [51,54,55]. Due to the high MCD concentrations applied, these were prepared directly in the culture medium before autoclaving. The HRs were thus cultivated with 0, 30, 50 or 70 mM MCD for 18 days before treatment with MeJA. Measurements were made four days after the application of MeJA. Figure 7 shows a strong effect of MCDs on HR growth at concentrations of 30, 50 and 70 mM, with almost no growth during the first culture phase of 18 days. MeJA treatment inhibited the growth of roots without MCDs but had no additional effect when applied on cultures containing MCDs. Previous results obtained for grapevine cell suspensions treated with 50 mM MCDs described a detrimental effect of the latter alone on cell growth [51]. MeJA treatment is known to have an even stronger inhibitory effect on the growth of grapevine cell suspension cultures [18,19,39,42,56]. Similar observations were made on HR cultures of Scutellaria lateriflora [57]. The use of MCDs on HRs has been reported to enhance the production of other phenolic compounds in S. lateriflora and to induce the production of stilbenes including resveratrol and derivatives such as prenylated arachidin-1 and arachidin-3 in peanut hairy root cultures [22]. Different elicitors were tested on those cultures, including sodium acetate, H2O2 as well as MeJA, CD or the combination of both. The best stilbene production in this system was achieved with 100 µM MeJA in combination with 9 g/L MCD, reaching 249 mg/L stilbenes excreted in the culture medium with a majority of arachidin-1 and arachidin-3. The use of MCDs on grapevine HR cultures displays a large induction effect on stilbenes accumulated in roots and excreted into the medium (Figure 8). After 18 days of culture, a clear increase in all quantified stilbenes in roots treated with MCDs was observed: control roots showed intracellular concentrations of 175 µg/g FW. In contrast, MCD-treated roots contained 999, 762 and 580 µg stilbenes/g FW before MeJA treatment for 30, 50 and 70 mM MCDs, respectively. The concentrations of stilbenes measured in the culture medium were 0.1 mg/L for the control and 71, 106, and 98 mg/L for the same CD-treated cultures. These concentrations represent an increase of stilbene production in comparison to control of 5.7, 4.4 and 3.3 folds in the roots and of 711, 1060 and 980 folds in the culture medium ( Figure S4). Each value was compared to its corresponding control using ANOVA test (* p < 0.5, *** p < 0.001). Biomasses of different controls were not statistically different.
CD concentrations used were chosen according to previous reports showing high stilbene production levels using grapevine cell suspensions [51,54,55]. Due to the high MCD concentrations applied, these were prepared directly in the culture medium before autoclaving. The HRs were thus cultivated with 0, 30, 50 or 70 mM MCD for 18 days before treatment with MeJA. Measurements were made four days after the application of MeJA. Figure 7 shows a strong effect of MCDs on HR growth at concentrations of 30, 50 and 70 mM, with almost no growth during the first culture phase of 18 days. MeJA treatment inhibited the growth of roots without MCDs but had no additional effect when applied on cultures containing MCDs. Previous results obtained for grapevine cell suspensions treated with 50 mM MCDs described a detrimental effect of the latter alone on cell growth [51]. MeJA treatment is known to have an even stronger inhibitory effect on the growth of grapevine cell suspension cultures [18,19,39,42,56]. Similar observations were made on HR cultures of Scutellaria lateriflora [57]. The use of MCDs on HRs has been reported to enhance the production of other phenolic compounds in S. lateriflora and to induce the production of stilbenes including resveratrol and derivatives such as prenylated arachidin-1 and arachidin-3 in peanut hairy root cultures [22]. Different elicitors were tested on those cultures, including sodium acetate, H 2 O 2 as well as MeJA, CD or the combination of both. The best stilbene production in this system was achieved with 100 µM MeJA in combination with 9 g/L MCD, reaching 249 mg/L stilbenes excreted in the culture medium with a majority of arachidin-1 and arachidin-3. The use of MCDs on grapevine HR cultures displays a large induction effect on stilbenes accumulated in roots and excreted into the medium (Figure 8). After 18 days of culture, a clear increase in all quantified stilbenes in roots treated with MCDs was observed: control roots showed intracellular concentrations of 175 µg/g FW. In contrast, MCD-treated roots contained 999, 762 and 580 µg stilbenes/g FW before MeJA treatment for 30, 50 and 70 mM MCDs, respectively. The concentrations of stilbenes measured in the culture medium were 0.1 mg/L for the control and 71, 106, and 98 mg/L for the same CD-treated cultures. These concentrations represent an increase of stilbene production in comparison to control of 5.7, 4.4 and 3.3 folds in the roots and of 711, 1060 and 980 folds in the culture medium ( Figure S4). and (e,f) four days after applying MeJA at 100 µM. Each point represents the mean with standard deviation of three biological replicates. Pic: piceid; Res: resveratrol;ε-vin: ε-viniferin;δ-vin: δ-viniferin. Each value was compared to its corresponding control using ANOVA test (* p < 0.5, **p < 0.01, *** p < 0.001) The MeJA treatment was applied after 18 days of culture and its effect was measured four days afterwards. Addition of 100 µM MeJA induced further accumulation of stilbenes in the roots and the culture medium (Figure 8), as previously reported in grapevine cell suspensions [18,19]. Control roots and medium contained 183 µg/g FW and 0.3 mg/L stilbenes, respectively, after four days. A maximum production of stilbenes in roots was observed for roots treated with 30 mM MCDs and 100 µM MeJA, reaching up to 1034 µg/g FW total stilbenes after four days, which represents a 5.6-fold increase in production. In the medium, the highest stilbene concentrations were observed in the cultures treated with 50 mM MCDs and 100 µM MeJA, which yielded up to 165 mg/L total stilbenes after four days, corresponding to an increase of 571 folds in excreted stilbenes. The measured rates of excreted stilbenes show extremely high values of more than 90% of total stilbenes for all three tested MCD concentrations (Figure 9). δ-viniferin was slightly less excreted, with 80% found in the extracellular medium and piceid which remains mostly intracellular ( Figure S3b). and (e,f) four days after applying MeJA at 100 µM. Each point represents the mean with standard deviation of three biological replicates. Pic: piceid; Res: resveratrol; ε-vin: ε-viniferin; δ-vin: δ-viniferin. Each value was compared to its corresponding control using ANOVA test (* p < 0.5, ** p < 0.01, *** p < 0.001).
The MeJA treatment was applied after 18 days of culture and its effect was measured four days afterwards. Addition of 100 µM MeJA induced further accumulation of stilbenes in the roots and the culture medium (Figure 8), as previously reported in grapevine cell suspensions [18,19]. Control roots and medium contained 183 µg/g FW and 0.3 mg/L stilbenes, respectively, after four days. A maximum production of stilbenes in roots was observed for roots treated with 30 mM MCDs and 100 µM MeJA, reaching up to 1034 µg/g FW total stilbenes after four days, which represents a 5.6-fold increase in production. In the medium, the highest stilbene concentrations were observed in the cultures treated with 50 mM MCDs and 100 µM MeJA, which yielded up to 165 mg/L total stilbenes after four days, corresponding to an increase of 571 folds in excreted stilbenes. The measured rates of excreted stilbenes show extremely high values of more than 90% of total stilbenes for all three tested MCD concentrations ( Figure 9). δ-viniferin was slightly less excreted, with 80% found in the extracellular medium and piceid which remains mostly intracellular ( Figure S3b). These data support the fact that, besides their well-known eliciting effects on stilbene production, MCDs facilitate the outflow of stilbenes by increasing their solubility in the medium [58]. Although 30 mM MCDs leads to the highest stilbene content in roots, a dose of 50 mM is more suitable due the high quantity of excreted stilbenes obtained. In fact, high yields of excreted stilbenes facilitate their extraction from the cultures. Production of resveratrol reaching up to 5 g/L has been reported using MCDs in grapevine cell suspensions [59]. Although the use of HRs presents unique advantages, this system still needs further optimization in terms of growth conditions to reach similar stilbene concentrations. For example, the biomass reported here with HRs was 8 g FW/L after 18 days of culture with MCDs. In comparison, grapevine cell suspensions can reach up to 500gFW/L [19]. Various approaches could be used to improve growth and thus biomass production before elicitation to reduce this limitation. A thorough optimization of the culture medium, by identifying limiting nutrients and potential growth factors, could lead to an improvement of growth. In the same way, growth conditions can be modified to optimize gas transfers, by shaking or changing the volume of the cultures. Comparing specific stilbene production in grapevine cell suspensions shows a total stilbene production of 13.4 mg/g FW [53], whereas it reached 18.1 mg/g FW in grapevine HRs, demonstrating that this system could represent a promising tool for stilbene production.

Plant and Bacterial Materials
Six-week-old in vitro plantlets from Vitis vinifera cv Pinot Noir 40024 were used for the transformation experiments. They were cultivated in a 16 h/8 h day/night photoperiod on McCown Woody Plant Medium (Dushefa, Haarlem, The Netherlands). All plant cultures were kept at 23 °C.
Two bacterial strains were tested: Rhizobium rhizogenes strain ATCC 15834 and Rhizobium rhizogenes ICPB TR7. The cultures were initiated in 10 mL Yeast Mannitol Broth YMB medium [60] for 24 h at 28 °C before being pelleted and resuspended in 20 mL in the corresponding plant medium supplemented with 1 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D). 2,4-D was used to induce a short dedifferentiation to facilitate transformation [31,32]. The explants tested consisted of entire leaves with petioles or 1-cm-long sections of the stem. For plant infection, three different methods were compared: (1) wounding explants with a scalpel previously infected with the bacterial suspension; (2) immersing wounded explants in the bacterial suspension for 5 min; and (3) directly pricking 24 in vitro plantlets with an infected needle on 3 spots of the stem.
After infection, explants were transferred onto solid B5 medium [61] and co-cultivated in darkness. After 48 h, the explants were rinsed in B5 +600 mg/L cefotaxime and cultured on solid B5 medium 3% sucrose supplemented with 300 mg/L cefotaxime. When the first roots appeared on the explants, and measured 1 cm or more, they were isolated and cultured on B5 medium +300 mg/L cefotaxime. They were grown for 3 months on solid B5 medium +300 mg/L cefotaxime and These data support the fact that, besides their well-known eliciting effects on stilbene production, MCDs facilitate the outflow of stilbenes by increasing their solubility in the medium [58]. Although 30 mM MCDs leads to the highest stilbene content in roots, a dose of 50 mM is more suitable due the high quantity of excreted stilbenes obtained. In fact, high yields of excreted stilbenes facilitate their extraction from the cultures. Production of resveratrol reaching up to 5 g/L has been reported using MCDs in grapevine cell suspensions [59]. Although the use of HRs presents unique advantages, this system still needs further optimization in terms of growth conditions to reach similar stilbene concentrations. For example, the biomass reported here with HRs was 8 g FW/L after 18 days of culture with MCDs. In comparison, grapevine cell suspensions can reach up to 500 g FW/L [19]. Various approaches could be used to improve growth and thus biomass production before elicitation to reduce this limitation. A thorough optimization of the culture medium, by identifying limiting nutrients and potential growth factors, could lead to an improvement of growth. In the same way, growth conditions can be modified to optimize gas transfers, by shaking or changing the volume of the cultures. Comparing specific stilbene production in grapevine cell suspensions shows a total stilbene production of 13.4 mg/g FW [53], whereas it reached 18.1 mg/g FW in grapevine HRs, demonstrating that this system could represent a promising tool for stilbene production.

Plant and Bacterial Materials
Six-week-old in vitro plantlets from Vitis vinifera cv Pinot Noir 40024 were used for the transformation experiments. They were cultivated in a 16 h/8 h day/night photoperiod on McCown Woody Plant Medium (Dushefa, Haarlem, The Netherlands). All plant cultures were kept at 23 • C.
Two bacterial strains were tested: Rhizobium rhizogenes strain ATCC 15834 and Rhizobium rhizogenes ICPB TR7. The cultures were initiated in 10 mL Yeast Mannitol Broth YMB medium [60] for 24 h at 28 • C before being pelleted and resuspended in 20 mL in the corresponding plant medium supplemented with 1 mg/L 2,4-dichlorophenoxyacetic acid (2,4-D). 2,4-D was used to induce a short dedifferentiation to facilitate transformation [31,32]. The explants tested consisted of entire leaves with petioles or 1-cm-long sections of the stem. For plant infection, three different methods were compared: (1) wounding explants with a scalpel previously infected with the bacterial suspension; (2) immersing wounded explants in the bacterial suspension for 5 min; and (3) directly pricking 24 in vitro plantlets with an infected needle on 3 spots of the stem.
After infection, explants were transferred onto solid B5 medium [61] and co-cultivated in darkness. After 48 h, the explants were rinsed in B5 + 600 mg/L cefotaxime and cultured on solid B5 medium 3% sucrose supplemented with 300 mg/L cefotaxime. When the first roots appeared on the explants, and measured 1 cm or more, they were isolated and cultured on B5 medium + 300 mg/L cefotaxime. They were grown for 3 months on solid B5 medium + 300 mg/L cefotaxime and subcultured every 3 weeks, then transferred to liquid B5 medium + 300 mg/L cefotaxime. Each line was routinely subcultured every 40 days and the medium was changed every 10 days. The antibiotic pressure was removed 4 months after isolation, once the absence of Rhizobacteria was confirmed by PCR.

Confirmation of the Genetic Transformation
The status of genetically transformed roots was verified by PCR, by searching for the rolC transgene and the absence of bacterial virD2. Primer characteristics are described in Table 1 and sequences were taken from data available in the literature [62,63]. Specific amplification was achieved using Firepol ® polymerase (Solis Biodyne) with a thermal cycler (Progene, Thechne): 12 min activation at 95 • C, followed by 35 cycles of 30 s at 95 • C, 30 s at 56 • C and 45 s at 72 • C, and a final 2 min extension time at 72 • C. The bacterial strain was used as a double positive control, a wild type Escherichia coli strain was used as a double negative control and Brassica rapa hairy roots were used as a transformation-positive control [64].

Improvement of Growth Conditions
The best medium and sucrose concentrations tested were determined in non-limiting conditions. Each culture was started in 100-mL flasks, containing 20 mL medium and one primary root tip of approximately 2 cm. They were cultivated at 110 rpm in darkness for 35 days. Weight measurements and medium renewal, to avoid excessive medium loss over time, were carried out every 3 or 4 days. All tests were done in triplicate.
The growth rate µ was determined from the exponential trend line of the mean growth curve, with biomass equation X = X 0 e µ(t−T0) . The doubling time Td was calculated using Td = ln (2) u [68].

Growth Kinetics
The growth kinetic parameters were calculated from cultures in 100-mL flasks, containing 20 mL 1/2 SH medium with 2% sucrose, inoculated with two primary root tips of approximately 2 cm. Three flasks were used for each point to determine the fresh weight. Roots and media were stored at −20 • C for stilbene quantification. The growth rate µ was calculated using the slope of the curve given by mean values of triplicates.

Stilbene Extraction
Stilbenes were extracted from the medium as described in [69,70]. Liquid-liquid extractions using 10 mL ethyl acetate for 10 mL medium were carried out. Ethyl acetate phases were then collected, evaporated in a rotavapor (Laborota 4000-Efficient, Heidolph with PC3001 VARIO vacuum pump, Vacuubrand, Schwabach, Germany) and the extracts were resuspended in 1 mL methanol (MeOH) for UPLC quantification.
Stilbene extraction from roots was adapted from [69,70] using roots ground in liquid nitrogen. One hundred micrograms was weighed and extracted with 1mL 85% MeOH for 2 h in a thermoshaker at 850 rpm, then centrifuged. The supernatant was collected in a glass tube and the pellet was resuspended in 1 mL 100% MeOH for 1 h with a thermomixer (Eppendorf) at 850 rpm, and then centrifuged. The supernatants were pooled, evaporated in a Vacuum Concentrator 5301 (Eppendorf) and resuspended in 1 mL MeOH for UPLC quantification.

UPLC Analysis
UPLC analyses were done using an Acquity UPLC (Waters, Guyancourt, France) system, with a C18 Cortecs ® (Waters) column (2.1 × 100 mm, particles: 1.6 µm, pore size: 90 Å), maintained at 30 • C. The separation conditions were taken from [68]. Briefly, a gradient at 0.5 mL/min starting from 90% H 2 O +0.1% formic acid (A) and 10% acetonitrile +0.1% formic acid(B) to 40% A and 60% B, followed by 5 min rinsing was used. Stilbene fluorescence was measured with an Aqcuity fluorometer (Waters) with an excitation wavelength of 330 nm and an emission wavelength at 375 nm. Piceid (3-O-β-D-resveratrol glucoside), resveratrol (3,5,4 -trihydroxystilbene), δand ε-viniferins (resveratrol dehydrodimers), in both their cis and trans forms, were identified and quantified as described previously and used as external standards [71,72]. Resveratrol, piceid and ε-viniferin were purchased from Sigma at the time of the experiments. Supplementary data concerning stilbene identification are available in [73]. Quantification of piceid and resveratrol was expressed as trans-resveratrol equivalents and that of δ and ε-viniferins as trans-ε-viniferin equivalents. All extractions were done in subdue light to limit isomerization of trans stilbenes to the cis forms. The concentrations shown are the sum of cis and trans forms measured. Total stilbenes represent the sum of these stilbenes.

Induction Treatments
Methyl jasmonate (MeJA) (Sigma Aldrich, St. Quentin Fallavier, France) was added to the culture medium (1/2 SH supplemented with 2% sucrose) from an ethanol (EtOH) solution after the exponential growth phase, corresponding to 18 days of culture. Final concentrations at 100 µM and 200 µM were chosen, with a maximum concentration of 0.1% EtOH in the medium. Controls were obtained by addition of EtOH to a final concentration of 0.1%. All experiments were conducted in triplicate.
For the MeJA and CD experiments, CAVASOL ® W-7 MCDs (WackerChemie AG, Burghausen, Germany) were dissolved in the culture medium (1/2 SH supplemented with 2% sucrose) to final concentrations of 30 mM, 50 mM and 70 mM before autoclaving and MeJA was added after 18 days at a final concentration of 100 µM.

Conclusions
Hairy root systems are particularly attractive as study and bioproduction tools due to their specific properties such as high genetic and phytochemical stability, good growth rates and their highly controlled environment.
In the present work, the growth and phenotype of grapevine cv Pinot Noir HRs has been reported. Moreover, their ability to respond to elicitation treatments that lead to the significant production and excretion of stilbenes into the culture medium has been demonstrated. The results show that grapevine hairy root cultures are able to provide high yields of secreted stilbenes, particularly ε-viniferin, when treated with methyl jasmonate and MCDs. This makes them a good candidate for the production of highly pure stilbenes. The up-scaling of such systems has already been proven feasible [74] and achieved by different companies (10,000 L Panax bioreactor) [75].