Effect of vine-shoot and oak extract foliar grapevine applications on the oenological parameters, phenolic acids and glutathione content of white musts and wines

Aim: An oxidation process frequently occurs during white winemaking, affecting its quality. The aim was to study, for two years (2013 and 2014), the effects that foliar applications of vine-shoot (Airén (AVS) and Moscatel (MVS)) and oak wood (OW) extracts on Airén grapevines have on wine color, and must and wine glutathione, trans-GRP, trans-caftaric acid, and trans-p-coutaric acid content. Methods and results: These compounds were analyzed by HPLC. The results showed that, in general, foliar application of Airén vine-shoot extracts did not affect glutathione concentration, and the other treatments decreased it. AVS2013-50 and AVS2014 samples were characterized by a high content of glutathione and trans-GRP, while MVS2014 samples retain high levels of trans-caftaric acid. trans-p-Coutaric acid concentration decreased after AVS2013 treatment, meanwhile in the 2014 season all applications increased its content. The content of these compounds in the wines was similar to those obtained in the musts. OW2013 showed the lowest value of Abs 420, likely because this treatment decreased glutathione concentration in musts. Conclusions: The foliar applications of vine-shoots and oak extracts on Airén grapevines had a clear effect in reducing the wine alcohol degree and Baumé in case of grapes. This finding is interesting in the context of the current global warming change scenario. Although the two vintages studied were different, an improvement in the color quality was observed in both. The content of glutathione decreased from must to wine: the content in both matrices was lower in treated samples than in control samples, except for Airén extracts in both vintages. This means that glutathione is oxidized, avoiding the oxidation of other must and wine molecules. Significance and impact of the study: These findings are important in relation to revalorizing waste from the vineyard, and thus being able to improve the quality of the white wine in relation to the oxidation processes that take place in the winemaking process.


INTRODUCTION
It has become mandatory for the wine industry to apply waste products to grapevines, with the aim of reducing chemicals according to sustainable practices (European Parliament, 2008). In this sense, vine-shoots are one of the most important waste products and have been used as organic fertilizers and food additives, to produce paper pulp, or to create solid biofuels, etc. (Jiménez Gómez et al., 1993;Jiménez et al., 2006;Portilla et al., 2008;Mendívil et al., 2013). Sánchez- Gómez et al. (2014) studied the chemical composition of Airén vine-shoot waste aqueous extracts in terms of phenolic, volatile, and mineral compounds, and reported that the extracts showed an interesting composition that may be assimilated by plants. Thereby, Airén or Moscatel vineshoot extracts applied foliarly to the grapevines improved wine amino acids, phenolic acids, and volatile composition, increased grape yield and decreased the alcohol degree (Sánchez- Gómez et al., 2016a;Sánchez-Gómez et al., 2016b;Sánchez-Gómez et al., 2017).
Several residues are generated in the manufacturing process of oak barrels, which can be added in the form of chips or pellets as an alternative to wine aging. Pardo-García et al. (2014) observed that oak extracts, obtained from oak waste from the cooperages, are interesting applications for the sustainability of the grapevine and oak industry. Thereby, the foliar application of the oak extracts affected grape composition, producing less alcoholic and acidic red wines with higher color intensity, lower shade, a more stable color and a higher content of polyphenols such as gallic acid, hydroxycynnamoyltartaric acids, acylated anthocyanins, flavanols and stilbenes (Pardo-García et al., 2014). In addition, after these applications, several compounds from the oak extracts were assimilated by vines and stored in grapes as glycosidic forms, affecting the wine volatile composition, mainly months after bottle storage (Martínez-Gil et al., 2011;Martínez-Gil et al., 2012;Martínez-Gil et al., 2013).
Color is one of the most important properties of must and wines from white varieties. The oxidation process that occurs frequently in white wines is a well-known problem in the winemaking industry because the color of musts and wines can be modified (Cejudo-Bastante et al., 2010). Sulfur dioxide is usually used to prevent must oxidation and browning development during bottle storage. However, there are strict regulations about sulfur dioxide use in the food industry, because of its toxicity and allergenic effects on human health (Lester, 1995). Hydroxycinnamic acids are prone to enzymatic oxidation reactions during the winemaking process, forming o-quinones and leading to color darkening. Thereby, glutathione, is an antioxidant tripeptide formed by three amino acids -cysteine, glutamate and glycine -and present in musts and wines. This tripeptide can regenerate the o-diphenol group of enzymatically oxidized trans-caftaric acid, giving rise to 2-S-glutathionyltrans-caftaric acid (GRP) and thus inhibiting wine browning (Cheynier et al., 1989;Cejudo-Bastante et al., 2010;Kritzinger et al., 2013). It has been recently reported that bisulfite competes as a nucleophile with glutathione for the o-quinone of trans-caftaric acid (Hayasaka et al., 2017). Thus, the aim of this work was to study the effect that foliar applications of Airén and Moscatel vine-shoot (AVS and MVS) extracts and oak wood (OW) extracts from non-toasted and toasted wood have on Airén grapevines. The effect of these treatments on must and wine glutathione, trans-GRP, trans-caftaric acid, and trans-pcoutaric acid concentration, and on wine color was evaluated during two consecutive vintages.

Vine-shoot and oak samples
Vine-shoot waste from the Airén Vitis vinifera cultivar in 2013 and from Airén and Moscatel in 2014 were sampled in the Castilla-La Mancha Spanish region, four months after each harvest, by randomized selection. Each year, vine-shoots were dried at room temperature for three months until a final humidity of 6.5 % (g water /100g of sample ). The 2013 oak sample was obtained from a commercial aqueous French oak extract (103 C), hereinafter OW 2013 , supplied by Protea France S.A.S. (Gensac la Pallue, France). This extract was obtained by macerating French oak chips (Quercus sessiliflora Salisb) for natural seasoning with water at high temperatures, and this was later toasted at mediumintensity levels, as described by Martínez-Gil et al. (2011). In 2014, the French oak wood used was provided by Protea France S.A.S. in the form of non-toasted chips (Quercus sessiliflora Salisb), naturally seasoned for 24 months.

Extracts preparation
All non-toasted vine-shoot waste samples (2013 and 2014) and the oak chips from 2014, were ground in a hammer miller (RETSCH, SM100 Comfort, Haan, Germany), sieved in a 10-mesh to get a homogeneous sawdust, and kept under vacuum at room temperature (25-27ºC) until use. To obtain the toasted samples for the 2014 experiment, dried Airén and Moscatel vine-shoots and oak chip materials were heated at 180 °C for 45 min according to the methods in Sánchez- Gómez et al. (2016c) and Sánchez-Gómez et al. (2016d) and the extraction procedure was as described in Sánchez- Gómez et al. (2014). The resulting Airén vine-shoot wastes extracts were as follows: Airén aqueous extract from non-toasted vine-shoots in 2013 (AVS 2013 ), Airén aqueous extract from nontoasted vine-shoots in 2014 (AVS 2014 ), and Airén aqueous extract from toasted vine-shoots in 2014 (AVS Toasted ). For Moscatel vine-shoot waste, the resulting extracts were Moscatel aqueous extract from non-toasted vine-shoots in 2014 (MVS 2014 ) and Moscatel aqueous extract from toasted vineshoots in 2014 (MVS Toasted ). For oak wood, the extracts obtained were French oak wood extract from non-toasted chips in 2014 (OW 2014 ) and French oak wood extract from toasted chips in 2014 (OW Toasted ). For each extract, 1.8 L was prepared.

Vines treatments
The white grapevines used in this study were from Airén Vitis vinifera grown at O.D. Jumilla in southwest Spain (38º 35' 17" N and 1º 16' 20" W; 511 m above sea level) during the 2013 and 2014 vintages. The grapevines were trained on the traditional bush vines Gobelet system with a row distribution north to south.
The formulations from 2013 (AVS 2013 and OW 2013 ) and 2014 (AVS 2014 , AVS Toasted , MVS 2014 , MVS Toasted , OW 2014 and OW Toasted ) were prepared using 0.05 % (v/v) of the Agral (Syngenta, Switzerland) adjuvant, which is a superficial wetting agent typically used for treatments with foliar applications, constituted by an inert mixture of polymers. The chemical composition of the extracts from 2013 and 2014, used in these applications, has been discussed in detail in previous works (Sánchez -Gómez et al., 2014;Sánchez-Gómez et al., 2016c;Sánchez-Gómez et al., 2016d).
In 2013, the formulation effect of vine-shoot extracts was tested, and the previous dilutions carried out with oak extracts (Martínez-Gil et al., 2011 were used as a reference. The AVS 2013 extract was applied once on day 7 post-veraison (AVS 2013 -100) and the second application was carried out with AVS 2013 diluted with water at 50 % and applied twice on days 7 and 14 post-veraison (AVS 2013 -50). The commercial aqueous oak extract (OW 2013 ) was applied once on day 7 post-veraison for comparing these results the previous ones from the same oak extract (Martínez-Gil et al., 2011. In 2014, the toasted procedure effect and possible vine-shoot varieties effect were tested. Because of the results obtained in 2013, all treatments were applied once. Thus, Airén grapevines were treated with four different vine-shoot extracts (AVS 2014 , AVS Toasted , MVS 2014 , MVS Toasted ) and two oak extracts (OW 2014 , OW Toasted ), which were applied once on day 7 post-veraison. Figure 1 summarizes the experimental design for the two years.
Each year, the treatments were applied on plants arranged in rows and in an edaphologically homogeneous portion of the plot, avoiding its edges. Each treatment was carried out on six plants randomly selected in the same row, distributed on three replicates with two plants each (n = 3), leaving other rows with untreated plants between the different applications to avoid contamination. Moreover, six additional plants were selected each year as controls, which were treated with water and the adjuvant following the same treatment protocol. The treatments were carried out when the environmental temperature was below 20 °C, between approximately 7 a.m. and 9 a.m. The grapevines were treated with 300 mL of each formulation per plant by spraying locally over the leaves.

Winemaking
Grapes were manually harvested on 26 September 2013 and 9 September 2014 at the optimum technological maturation moment for the grapevines used as the control sample. In both years, the grapes (approximately 20 kg for each treatment) were separately processed. These grapes were destemmed and the whole grape mass was mixed and homogenized (must and solid parts combined). Then, the grape mass from each treatment was placed into 30 L stainless steel tanks and ammonium bisulfite (150 mg/mL) was added to the resulting musts. Maceration was carried out at 11 °C ± 1 °C for 17 h.
Skins and seeds from each treatment were removed with a traditional vertical hand-press and the free-run musts were put into stainless steel tanks to finish the alcoholic fermentation. Musts were inoculated with 20 g/hL of UCLM S377 active dry yeast strain (Springer Oenologie, Maisons-Alfort Cedex, France) of Saccharomyces bayanus, previously rehydrated and activated according to the supplier. The alcoholic fermentation took place at 16 °C ± 1 °C. The density was measured each day and alcoholic fermentation was considered complete after 16 days in 2013 and 17 days in 2014, when it was constant at 0.990 g/L and the residual sugars value were below 5 g/L. As no malolactic fermentation was carried out, at this final point three wines from each treatment (33 wines in total: 12 from 2013 and 21 from 2014) were obtained, bottled and stored at 22 °C ± 2 °C.

Musts and wines oenological parameters
Must parameters such as °Baumé (ºBé), pH, total acidity (g/L of tartaric acid) and malic acid (g/L), along with wines parameters such as alcoholic degree (°A), pH, total acidity (g/L of tartaric acid), volatile acidity (g/L), malic acid (g/L), lactic acid (g/L) and color intensity (Abs 420 nm) were analyzed using equipment based on Fourier Transform-infrared spectroscopy (FT-IR Multispec; TDI, Barcelona, Spain) using the methods of ECC (1990) as reference.

Determination of glutathione by HPLC
The glutathione analysis was performed using the method described by Garde-Cerdán et al. (2014). This compound was analyzed by reverse-phase HPLC using a liquid chromatograph Agilent 1100 Series (Palo Alto, USA). Each sample (5 mL) was mixed with 100 µL of norvaline (internal standard). The mixture was submitted to an automatic precolumn derivatization with o-phthaldialdehyde (OPA Reagent, Agilent). The injected amount from the derived sample was 10 mL, with a constant temperature of 40 ºC. All separations were performed on a Hypersil ODS (250 × 4.0 mm, I.D. 5 µm) column (Agilent).
Two eluents were used as mobile phases: eluent A was 75 mM sodium acetate, 0.018 % triethylamine (pH 6.9) + 0.3 % tetrahydrofuran; eluent B was water, methanol, and acetonitrile (10:45:45, v/v/v). Detection was performed using a fluorescence detector (FLD). The identification of glutathione was performed by comparison of its retention time with that of pure reference standard. The pure reference compound and internal standard were from Sigma-Aldrich (Steinheim, Germany). The results for glutathione correspond to the average of three analyses (n = 3).

Statistical analysis
Data statistical analysis was performed using the SPSS Version 22.0 statistical package for Windows (SPSS, Chicago, USA). Oenological parameters, glutathione, trans-GRP, trans-caftaric and trans-p-coutaric acids data concentration from musts and wines were processed using variance analysis (ANOVA). Differences between means were compared using the Tukey test at 99.95 % probability level. Canonical discriminant analyses were performed in both years for all treatments with the glutathione, trans-GRP, trans-caftaric and trans-p-coutaric acids concentration in musts and wines.

Must and wine oenological parameters
The oenological parameters for musts in both years of study are shown in Table 1. In 2013, control must presented significant differences in all parameters studied in relation to the treatments. Control musts showed the highest ºBaumé content and the lowest pH, although in this last case pH range values were very close among control and musts from treatments. On the contrary, OW 2013 presented the lowest value of ºBaumé and the highest total acidity and malic acid content followed in these two last parameters for control must. Musts from grapevines treated with Airén vine-shoot extracts (AVS 2013 -50 and AVS 2013 -100) showed intermediate content of alcohol and the lowest values of total acidity and malic acid, and therefore, the highest pH values.
In the 2014 vintage, ºBaumé in the control must showed a similar behavior than in the 2013 must, presenting the highest content and significant differences regarding the treatments. Thus, in general, the treatments tested seem to affect the ºBaumé content, providing in ºBaumé pH Total acidity (g/L tartaric acid)    (Table 1). However, in general, the differences were very low, so the effect was minimal. This grape ripening delay could be due to the application of these vegetal extracts, so this should be taken into account in the current global warming/climate change scenario (Jones et al., 2005;Hannah et al., 2013;Webb et al., 2012). Regarding pH, AVS 2014 must showed the highest value, with the same behavior in 2013 for AVS 2013 -100 treatment whereas AVS Toasted , MVS 2014 and OW Toasted musts had the highest total acidity content and OW 2014 the lowest. In relation to malic acid content, must from AVS 2014 treatment and both from oak extracts applications showed the highest content, whereas the MVS Toasted value was the lowest.
The oenological parameters of the wines at the end of alcoholic fermentation after different grapevine treatments for both vintages, 2013 and 2014, are shown in In 2013, no significant differences were observed among treatments and control wines in total acidity. In relation to other parameters, control and OW 2013 wines presented lower pH values than wines from vine-shoot treatments (AVS 2013 -50 and AVS 2013 -100), which may be due to the high malic acid content of OW 2013 wine and the high lactic acid content of control wine, as total acidity, quantified as tartaric acid, was similar in all wines. The values of malic acid found in all the wines revealed that no malolactic fermentation occurred. Wines from AVS treatments showed lower values of malic and lactic acids, exhibiting the highest value of volatile acidity with AVS 2013 -50. Notwithstanding, this was below 0.6 g/L, indicating that all wines were well conserved, which means that there were no microbiological problems (Ribéreau-Gayon et al., 2006). OW 2013 wine was the only one that presented significant differences in Abs 420 nm with respect to control and AVS treatments (AVS 2013 -50 and AVS 2013 -100), as the OW 2013 value the lowest. In 2014, the control wines had the highest pH, whereas in the case of total and volatile acidity, AVS 2014 presented the lowest value for both parameters, but only in the first one; these differences were statistically significant with respect to the control and the rest of the wines.

Alcohol degree (%, v/v) pH
Total acidity (g/L tartaric acid) Volatile acidity (g/L) Malic acid (g/L) Lactic acid (g/L) Abs 420 Control 2013 9.70 ± 0.13 c 3.25 ± 0.01 a 6.81 ± 0.11 a 0.15 ± 0.02 ab 3.65 ± 0.03 b 0.29 ± 0.03 c 0.52 ± 0.04 b AVS2013-50 9.95 ± 0.30 c 3.28 ± 0.01 b 6.83 ± 0.08 a 0.26 ± 0.03 c 3.40 ± 0.09 a 0.14 ± 0.02 ab 0.52 ± 0.03 b AVS2013-100 8.96 ± 0.14 b 3.33 ± 0.01 c 6.71 ± 0.09 a 0.11 ± 0.03 a 3.60 ± 0.09 b 0.10 ± 0.03 a 0.58 ± 0.04 b OW2013 8      For each year, letters in the same column indicate significant differences between treatments and Greek letters indicate, for each compound, differences between musts and wines according to the Tukey test ( As in 2013 vintage, malolactic fermentation was not carried out, which was confirmed by the high values of malic acid (Table 2): the wines with the MVS Toasted treatment had a significantly higher value of malic and lactic acid. Although the vintage variability is observed, foliar applications seem to have a clear effect on the alcohol degree, as in both vintages the treated wines were less alcoholic than control ones. Additionally, the formulation effect along with the wood toasting procedure or wood type, does not seem to have an influence on the alcohol degree.

Glutathione, trans-GRP and trans-caftaric and trans-p-coutaric acids content in musts and wines
Glutathione, trans-GRP and trans-caftaric and trans-p-coutaric acids content in musts and wines for the two years studied after different grapevines treatments are summarized in Table 3.
In both studied vintages, the application of Airén vine-shoot extracts (AVS and AVS Toasted ) did not affect glutathione concentration in musts with respect to the control vines. However, in both cases, the treatment carried out with the oak wood extracts (OW and OW Toasted ) significantly decreased the glutathione content in the musts compared to control samples. In relation to the Moscatel vineshoot extracts (MVS and MVS Toasted ), their effect was similar to those obtained by the OW extracts, as glutathione concentration in musts decreased compared to control samples. The concentration of all these extracts decreased considerably after alcoholic fermentation, even below 2 mg/L (Table 3). In the 2013 vintage, wines from grapevines treated with AVS 2013-100 showed the lowest glutathione content with respect to the control wines. However, in the 2014 vintage, wines elaborated from the grapevines treated with different treatments presented lower glutathione concentration than control samples, except for AVS 2014 wines. The different behavior in the glutathione concentration could be due to the climatic conditions, among other reasons, as those affect the vines development and by the pathways that modify or activate vine secondary mechanisms due to the different extract applications.
All treatments applied during the 2013 vintage increased significantly the trans-GRP content in musts, with respect to controls. In this regard, the most effective treatments were both vineshoot extracts (AVS 2013 -50 and AVS 2013 -100) compared to the oak wood extract application (OW 2013 ) (Table 3). In the 2014 vintage, the content of transGRP in musts was increased by the application of non-toasted Airén vineshoot extract (AVS 2014 ) and from both Moscatel vine-shoot extracts, non-toasted and toasted. The rest of the treatments, such as application of oak wood extracts (OW 2014 and OW Toasted ) and AVS Toasted produced a decrease in trans-GRP content respect to control musts. trans-GRP concentration also decreased after alcoholic fermentation, but to a lesser extent than glutathione (Table 3). In the 2013 vintage, wines from grapevines treated with AVS 2013 -50 and AVS 2013 -100 had higher trans-GRP content than the control and OW 2013 wines. In 2014, the highest trans-GRP concentration was found in the AVS 2014 wines, while the lowest content was found in OW Toasted wines.
In the first year of study, the concentration of trans-caftaric acid decreased in the musts from grapevines treated with vine-shoot and OW extracts (  (Table 3). The increase in the trans-p-coutaric acid content from must to wine was also reported by Portu et al. (2015) when jasmonate foliar application was carried out on Tempranillo grapevines. In 2013, all wines showed different trans-p-coutaric acid concentration, where OW 2013 wines had the highest content and AVS 2013 -100 the lowest. In the 2014 vintage, control and MVS 2014 wines presented the lowest and the highest trans-p-coutaric acid content, respectively. Except for trans-p-coutaric acid, whose content was similar during both vintages, the concentration of the rest of the compounds studied in the Airén musts was higher in 2013 than in 2014 vintage. trans-p-Coutaric acid is an ester formed from trans-p-coumaric acid and tartaric acid. Acid esters are formed in simple hydrogenion-catalyzed esterification (Margalit, 2004)  and it has been shown that their formation and the hydroxycinnamic acid production depends on pH (Edwards et al., 1985;Hall and De Luca, 2007). As for the glutathione concentration, the differences found in the phenolic acid composition could be attributed to the physiological impact of the use of the extracts on biosynthesis of the compounds analyzed in must and wines, related to the secondary mechanism.
In an attempt to differentiate the treatments of both vintages in musts and wines according to glutathione, trans-GRP and trans-caftaric and trans-p-coutaric acids compounds, complementary discriminants analyses were also carried out ( Figure 2). Figure 2a shows the graphical representation of the must samples in the 2013 vintage for each treatment. The two canonic discriminant functions explained 97.60 % of the total variance, reaching 84.40 % with the first. In case of function 1, trans-GRP was the compound that contributed more to the differentiation with a negative correlation. For function 2, glutathione had the highest weight, also with a negative correlation. Figure 2b shows the discriminant analysis for wines from 2013 vintage, which presented a wide distance between control, OW 2013 and AVS 2013 -100, with respect to AVS 2013 -50 treatment. Function 1 and 2 explained 90.50 % and 9.40 % of the total variance, respectively. trans-GRP and trans-caftaric acid were the compounds that contributed more to the differentiation with a positive correlation in function 1, whereas for function 2 trans-p-coutaric acid had a higher weight and a positive correlation. According to the information provided by the variables related to functions 1 and 2, the must from AVS 2013 -50, and AVS 2013 -100 keep a higher level of trans-GRP; along with this, AVS 2013 -50 was also describedas having a high glutathione content. In contrast, the control kept a lower level of trans-GRP, while OW 2013 had a lower content of glutathione. With respect to wines, as occurred in musts, AVS 2013 -50 was described by trans-GPR, while OW 2013 was described by glutathione and trans-p-coutaric acid, contrary to those that occurred in the musts. Figure 2c shows a graphical representation of the must samples in the 2014 vintage for each treatment: 98.30 % of the total variance was explained by the two canonic discriminant functions. In the case of function 1, which explained the 57.60 % of the total variance, trans-GRP was the compound that contributed more to the differentiation with a positive correlation. For function 2, which showed 40.70 % of the total variance, the trans-caftaric acid had the highest weight, also with a positive correlation. Figure 2d shows the graphical representation for wines from 2014: 73.30 % and 26.00 % of the total variance were explained by function 1 and 2, respectively. trans-GRP and trans-caftaric acid were the compounds that contributed more to the differentiation with a positive correlation in function 1, whereas for function 2 trans-GRP had a higher weight and a positive correlation. According to the information provided by the variables related to functions 1 and 2, it could be concluded that musts from AVS 2014 were described by trans-GRP, while musts from MVS 2014 retain higher levels of trans-caftaric acid. With respect to the wines, AVS 2014 and control samples were described by trans-caftaric acid and trans-GRP, while MVS 2014 samples were characterized by trans-GRP. AVS Toasted , MVS Toasted , OW 2014 and OW Toasted wine samples presented similar behavior.