AGRONOMIC PERFORMANCE AND WINE PHENOLIC COMPOSITION OF 'MERLOT' GROWN IN ALTITUDE REGION WITH DIFFERENT ROOTSTOCKS

Several studies have reported the influence of rootstocks on grape and wine composition, however, there is considerable variability in the results. The objective of this work was to evaluate the effects of three different rootstocks on vegetative-productive balance and the wine composition of 'Merlot'. The present work was conducted in the municipality of Painel, during the 2014/2015, 2015/2016 and 2016/2017 vintages. The vegetative, productive, and enological variables of ‘Merlot’ grafted onto ‘Paulsen 1103’, ‘3309 Couderc’ and ‘101-14 Mgt’ rootstocks were evaluated. The ANOVA results showed that‘3309C’ rootstock promoted an increase in the productive indices, and consequently in more adequate vegetative-productive balance for 'Merlot' variety compared to the rootstocks ‘Paulsen 1103’ and ‘101-14 Mgt’. The rootstocks ‘3309C’ and ‘101-14 Mgt’ provided more adequate maturation indices for 'Merlot'. The ‘101-14 Mgt’ rootstock was associated with the highest contents of flavonoid compounds, while the ‘3309C’ rootstock was related to higher contents of non-flavonoid compounds, especially resveratrol.


INTRODUCTION
The high-altitude region of Santa Catarina State/Brazil is characterized by presenting its vineyards at an altitude ranging from 900 to 1400 m above sea level, where grape phenological cycle is longer than in other Brazilian wine producing regions (Brighenti et al., 2013).Wine produced in this region has high-color quality and aroma capable of differentiating it from wine made in other regions (Marcon Filho et al., 2015).
The environmental conditions found in high altitude regions are challenging to obtain balanced vineyards.Among them, high water availability (Bem et al., 2016) and high organic matter in the soil (Mafra et al., 2011) stand out, which associated with the use of vigorous rootstock (Borghezan et al., 2011) promote excessive vegetative growth; this shows that it is necessary to improve vine balance, which allows grapes to be harvested at appropriate levels of maturity (Wurz et al., 2017).
The use of rootstocks in viticulture provides a platform for manipulating a wide range of vine traits that can improve the productive efficiency of the vineyard.Among them, we can report, the ability to absorb and transport nutrients (Stockert et al., 2013;Lecourt et al., 2015), resistance to water deficit (Pavlousek, 2011) and the transport of plant hormones (Nikolaou et al., 2000).Rootstocks can play an important role in controlling excessive vegetative growth and increasing fruit set in lowfertility varieties (Kidman et al., 2013).These skills depend on their adaptability to the environment and interaction with the canopy variety (Tandonnet et al., 2010).
Several studies have reported the influence of rootstocks on grape and wine composition, however, there is considerable variability in the results.According to Han et al. (2022), some research indicated that rootstocks with different characteristics confer different phenotypic traits to the scion.For example, rootstocks with high vigor increase yield and pruning weight, as well as the titratable acid content in the juice.Several studies, however, showed that rootstocks did not affect basic fruit composition parameters.Miele and Rizzon (2017) studied the composition of 'Cabernet Sauvignon' must on 15 rootstocks and found higher levels of sugars and pH in rootstocks that had 'V.riparia' in their genetics, such as '101-14 Mgt' and '3309C'.Ollat et al. (2003), found higher anthocyanin contents in grapes of plants grafted onto V. Riparia compared to '101-14 Mgt' and 'SO4'.The same study found differences in the concentration of sugars.Some authors argue that the rootstock has some significant effects on grape and wine composition, but that the determining factors are related to the scion variety and the environment factor (Harbertson and Keller, 2011).But in general, other studies have clearly demonstrated that the rootstock has considerable potential to determine important characteristics of grape and wine quality (Jogaiah et al., 2015).The mechanisms involved in this phenomenon may be related to the modification of the relationship between vegetative and reproductive development (Kidman et al., 2013;Miele and Rizzon, 2017), or even to the direct influence on fruit composition (Kodur et al., 2013;Jogaiah et al., 2015).
In the highlands of Santa Catarina State/Brazil, it is observed that 72% of the vines are grafted onto '1103 Paulsen' (Vianna et al., 2016), which is described in the literature as a vigorous genotype (Keller, 2015).Regarding the characteristics of '1103 Paulsen' with the edaphoclimatic conditions found in the highlands of Santa Catarina, an environment conducive to the vegetative growth of the canopy varieties exists, which leads to an imbalance in the vineyards (Brighenti et al., 2011;Zalamena et al., 2013).
In this context, the objective of this work was to evaluate the effects of different rootstocks on the vegetative-productive balance and the wine composition of 'Merlot'.For this purpose, in addition to '1103P', the rootstocks '3309C' and '101-14 Mgt', which are known to confer moderate to low vigor (Keller, 2015), were tested.

MATERIALS AND METHODS
The present work was carried out in the municipality of Painel (28°01'S, 50°08'W and 1,200 m above sea level) during the 2014/2015, 2015/2016 and 2016/2017 cycles.The vineyards were planted in 2004, spacing 3.0 m between rows and 1.5 m between plants.The plants are trained in Y trellis system with double spur cordon pruning.Soils in the region are considered Humic Cambisol, Litholic Neosol and Haplic Nitosol classes, developed from rhyodacite and basalt rocks (Santos et al., 2018).The climate of the region, according to the Koeppen climate classification system, is classified as humid mesothermal and mild summer (Cfb).
The experimental design was randomized blocks, with four blocks, each repetition consisting of 10 uniform plants in vigor.The vegetative, productive, and enological variables of the Merlot variety grafted onto 'Paulsen 1103', '3309 Couderc' and '101-14 Mgt' rootstocks were evaluated.
Harvesting took place on the same day for all three rootstocks, according to the winery's criteria.On the date of harvest, data on production and leaf area were recorded, and samples of clusters and berries were collected for further analysis and microvinification in the laboratory.
Yield, number of clusters and number of shoots were recorded for each plant of each treatment.Cluster mass was estimated by dividing production per plant by the number of clusters per plant.The number of clusters per shoot was obtained by dividing the number of clusters per plant by the number of shoots per plant.
To estimate the leaf area per plant, 10 shoots per plot were selected in the middle third of the cordon.In these shoots, the midvein length (cm) of all leaves was obtained with a ruler.For each leaf, the leaf area was calculated using the formula: y = 1.1265x 2.0445 , where y corresponds to the leaf area in cm² and x corresponds to the midvein length in cm (Borghezan et al., 2010).By adding the leaf area of the leaves for each shoot, the leaf area per shoot was obtained.From there, the average leaf area per shoot was calculated for each treatment, and by multiplying this value by the number of shoots per plant, an estimate of the leaf area per plant was obtained.The ratio between leaf area and fruit mass (LA/Fruit) was calculated by dividing the leaf area (m²) by the yield (kg), and the results were expressed in cm²/g.The Ravaz index was obtained from the ratio of fruit mass to mass of branches (Cus, 2004).
Samples of 100 berries per plot were collected for technological maturation analyzes and determination of berry diameter and mass.In the laboratory, the berries were weighed and individually crushed to obtain the must.The content of soluble solids was analyzed using a digital refractometer with temperature compensation (ATAGO®, PAL-1); the pH was analyzed with a potentiometer (Impac®); and the total acidity, through titration with 0.1N NaOH, using bromothymol blue indicator until the pH of the medium reached 8.2, with the results expressed in mEq/L (OIV, 2015).
For microvinification, 40 kg samples were collected per treatment, which were kept in a cold storage for 24 hours at 5 ºC.The microvinifications were carried out at the Enology Laboratory of Santa Catarina State University, in the city of Lages/SC and followed the protocol adapted from Pszczolkowski;Lecco (2011), Makhotkina et al. (2013), as described by Wurz et al. (2018).
The analysis of total polyphenol content was carried out in a UV-VIS spectrophotometer (Biospectro -SP220), according to the methodology described by Singleton and Rossi (1965), and the results were expressed in mg gallic acid/L.The anthocyanin content was determined by spectrophotometry, using the method described by Rizzon (2010).Wine color was measured in a spectrophotometer, through absorbance measurements at 420 nm, 520 nm and 620 nm.The color intensity and hue were calculated as follows: Intensity = 420 nm + 520 nm + 620 nm; Hue = 420 nm/520 nm.
The wines phenolic compounds were quantified by high performance liquid chromatography (HPLC) according to the methodology described by Cadahía et al. (2009) and adapted by Ferreira-Lima et al. (2013), according to the quantitative methodology described by Marcon Filho et al. (2022).A solution containing a mixture of all standards was prepared in a synthetic wine medium (5 g/L tartaric acid, 12% v/v ethanol and pH 3.2).Synthetic wine was used to avoid interference in chromatographic separation and detection response.The calibration solutions were also prepared in synthetic wine by diluting the stock solution containing the mixture of the standards.All the solutions used were previously filtered through a 0.45 μm membrane (PES-Kasvi membrane).
Approximately 2 mL of sample (wine or calibration solution) was filtered on a 0.45 μm membrane (PES-Kasvi membrane) with a syringe and placed in the vial for direct injection into the HPLC system.
Chromatographic analyzes were performed using Shimadzu High Performance Liquid Chromatography equipment (Kyoto, Japan).The column was C18 (5 μm, 250 mm x 4.6 mm, Restek).The software used to control the gradient system, the detector and for data acquisition was Shimadzu Class-VP.For each sample was analyzed in duplicate; when a variation of more than 10% was detected, a third analysis was performed.The phenolic compounds were detected at 280 nm.
The data were submitted to analysis of variance (ANOVA); when significant effects of the rootstocks were detected, the Fisher's least significant difference (LSD) test was performed (p<0.005).

RESULTS AND DISCUSSION
The yield of 'Merlot' was significantly affected by the rootstock, as shown in Table I.The highest yield was observed with '3309 C' (5.5 kg/plant), which had the highest number of clusters (81.6 clusters/plant) and the heaviest clusters (64.5 g/cluster).The rootstocks '1103 P' and '101-14 Mgt' were associated with the lowest yields (3.7 and 4.5 kg/plant, respectively), the lowest number of clusters (64.7 and 73.0 clusters/plant, respectively), and the lowest average cluster mass (55.5 and 57.8 g/cluster, respectively).
These observations are in line with the literature, especially with the work of Kidman et al. (2013), who report an increase in 'Merlot' bud fertility and fruit set when grafted onto rootstocks that have 'V.riparia' in their genetics.In another study, the authors also found the highest yields when 'Merlot' was grafted onto '3309 C', compared with other rootstocks during three years of studies (Keller et al., 2011).
The higher leaf area averages per plant were observed with '1103 P' (15.8 m 2 ) and '3309 C' (15.2 m 2 ), and the lowest was observed with '101-14 Mgt' (12.2 m 2 ), as shown in Table II.'Merlot' grafted onto '1103 P' and '101-14 Mgt' presented values of 1.8 and 1.6 kg of pruning material per plant, respectively.This difference is directly related to the variable shoot mass (g), since the rootstock '3309 C' had an average value of 28.1 g per shoot, while the rootstocks '1103 P' and '101-14 Mgt' had average values of 39.7 and 35.9 g per shoot, respectively.
Similar to previous reports, '1103 P' tended to increase scion pruning weights (Keller et al., 2012).According to Han et al. (2022), '1103 P' is a hybrid rootstock, with 'V.rupestris' and 'V.berlandieri' as parents, and absorb more water and nutrients to supply the vegetative growth of the canopy, resulting in a higher pruning weight, and greater vegetative growth.
Data related to vegetative-productive balance are described in Table III.It was observed that the '3309 C' and '101-14 Mgt' rootstocks resulted in a greater number of clusters per shoot (1.6 and 1.5 clusters per shoot, respectively), while the '1103 P' rootstock had 1.4 clusters per shoot.A similar behavior was observed for the leaf area/fruit ratio, with values of 33.4 and 34.7 cm 2 /g, for rootstocks '3309 C' and '101-14 Mgt', respectively, while rootstock '1103 P' showed a value of 52.4 cm 2 /g for the leaf area/fruit ratio.In high-altitude vineyards, ideal leaf area/yield ratios were established as 23 cm 2 /g for 'Merlot' (Borghezan et al., 2011).
Regarding the Ravaz Index, more adequate values were found for the rootstocks '3309 C' and '101-14 Mgt', with 4.1 and 3.2 kg/kg, respectively, while the '1103 P' rootstock presented a ratio of 2.2 kg/kg for this index.
Due to the increase of the production components, the rootstock '3309 C' improved the vegetative:productive balance variables.Kliewer;Dokoozlian (2005) reported that the Ravaz Index generally ranged from 5 to 10 between cultivars.Although the values found in '3309 C' are below the ideal minimum, they are still above the rates described in the high-altitude regions of Santa Catarina.In São Joaquim, for example, for 'Cabernet Sauvignon' and 'Merlot' varieties, values between 1.5 and 2.3 have already been described (Brighenti et al., 2011;Zalamena et al., 2013).Wooldridge et al. (2010) showed that the wine quality of 'Chardonnay' and 'Pinot Noir' onto different rootstocks was inversely proportional to pruning mass but positively correlated with the Ravaz Index.This shows the importance of control the vegetative canopy for sutiable ratio between source and sink (Panceri et al., 2018).This behavior affected the estimated yield, since plants with a more adequate vegetative-productive balance resulted in better productive indices.Furthermore, according to Callili et al. (2023), excessive vegetative vigor leads to humid conditions and low radiation, increasing shading in the cluster zone and impairing the accumulation of soluble solids in the berries.The different rootstocks influenced the maturation of 'Merlot', as shown in Table IV.Soluble solids were not influenced by the three rootstocks evaluated, however, regarding the pH, the higher values were found for the rootstocks '3309 C' (3.14) and '101-14 Mgt' (3.16) while the '1103 P' rootstock presented a value of 3.06.
Regarding total acidity, the highest value was observed for the rootstock '1103 P' (104.5 mEq/L), while the rootstocks '3309 C' and '101-14 Mgt' resulted in lower values, with averages of 96.7 and 96.9 mEq/L, respectively.Low vigor may also explain the lowest acid content for this combination, probably due to an increased berry exposure to sunlight, favoring acid degradation by rising in temperature (Zufferey et al., 2017).
Regarding the maturation of 'Merlot', similar data were observed by Han et al. (2022), which also reported the impact of rootstocks on total acidity and pH of the juice, without influencing soluble solids content.Nascimento et al. (2022), evaluated the effects of different rootstocks on 'Magna' variety and they observed similar behavior for these variables.Thought Azuma et al. (2019) suggested that under conditions of sufficient solar radiation, rootstocks may have less influence on fruit ripening.
These results agree with the work of Pouget (1986), who clearly demonstrated that vines with large canopies, induced by the rootstock, tend to prolong fruit maturation, in comparison with vines grafted onto low vigor genotypes, which in turn tend to accelerate the maturation process.This trend has also been reported in other studies (Brighenti et al., 2011;Keller et al., 2011;Neal et al., 2014;Jogaiah et al., 2015;Han et al., 2022).
The pH of 'Merlot' grapes was significantly higher in plants grafted onto '3309 C' and '101-14 Mgt'.These results are consistent with those found in other studies (Kodur et al., 2013;Neal et al., 2014;Jogaiah et al., 2015).In the case of grapes originated from plants grafted onto '1103 P', the low pH may be related to the lower capacity to absorb potassium in rootstocks with progeny of 'Vitis berlandieri' (Wolpert et al., 2005).The rootstock influenced the anthocyanin content, total polyphenols, and color intensity; only the color hue was not influenced by the rootstock, as shown in Table V.  Regarding the total anthocyanin content, the higher value was observed for the '101-14 Mgt' rootstock (259.7 mg/L), while the '1103 P' and '3309 C' rootstocks showed values of 203.7 and 202.8 mg/L, respectively.Regarding the total polyphenol content, '1103 P' and '101-14 Mgt' showed the highest values, 1279 and 1229 mg/L, respectively, while the rootstock '3309 C' showed the lowest value (1088 mg/L).
The color intensity presented a similar behavior to the total anthocyanin content, with the highest value observed for the rootstock '101-14 Mgt' (10.9) and the lowest ones observed for the rootstocks '1103 P' and '3309 C' (9.2 and 9.0, respectively).The color intensity of wines is mainly determined by grape anthocyanin content.One of the factors that contribute to increase color in red wines is linked to the decrease in grapes average size (Gil et al., 2015).
Relatively less vigorous plants reduce cluster shading, and incident solar radiation induces an increase in anthocyanin synthesis (Downey et al., 2006).In this study, the phenolic compounds analyzed are categorized into flavonoids and non-flavonoids.Flavonoids are the compounds mainly responsible for wines color and astringency, in addition to their great importance for human health (Santos-Buelga and Scalbert, 2000).As indicated in Table VI, rootstocks do not influence the catechin content in wines.However, for quercetin, rutin and kaempferol, a higher value was found for the rootstock '101-14 Mgt' (12.3, 6.0 and 1.03 mg/L, respectively).
In a study on the effect of vigor on the composition of flavonoids in 'Pinot Noir' grapes and wines, the authors concluded that vine vigor exerts a greater influence on the accumulation of flavonoids in the skins and less or no influence of these compounds on the seeds (Cortell et al., 2005).The concentration of some flavan-3-ols and flavonols in grape skin increased significantly with the decrease of vine vigor (Cortell et al., 2005).Flavonoid compounds concentrations in grapes are mediated by rootstocks, and these differences are most likely due to an imbalance in vegetative and reproductive growth between rootstock combinations (Rasool et al., 2020).Shading has been shown to cause significant reductions in the concentration of flavonols in grapes (Downey et al., 2006).Despite their low concentration, flavonols are important because they participate in co-pigmentation processes with anthocyanins, changing the color of wines and stabilizing pigments (Cheynier and Rigaud, 1986).Flavanols, mainly represented by catechin and epicatechin, are important because they provide astringency to wines (Downey et al., 2003).
The non-flavonoid compounds analyzed in this work are classified as: hydroxybenzoic acids (gallic and vanillic acids), hydroxycinnamic acids (p-coumaric) and stilbenes (resveratrol).For such compounds, a consistent rootstock effect was not observed.
The vanillic acid was not influenced by the different rootstocks.However, gallic acid was higher for rootstock '1103 P' (22.3 mg/L), followed by '3309 C' (21.0 mg/L), and '101-14 Mgt' (19.7 mg/L).For p-coumaric acid, the highest value was observed for rootstock '3309 C' (5.0 mg/L), followed by the rootstocks '101-14 Mgt' (4.3 mg/L) and '1103 P' (4.1 mg/L).Gallic acid is described as one of the most important phenolic compounds because it is a precursor of all hydrolysable tannins.This group of compounds is important in the wine composition, mainly due to their ability to react with anthocyanins, consequently stabilizing the wine color (Gris et al., 2007).
The resveratrol content showed higher values for the rootstocks '3309 C' and '101-14 Mgt', with values of 9.4 and 8.7 mg/L, respectively, while the '1103 P' rootstock showed a lower value of resveratrol, with 7.0 mg/L.
Hydroxycinnamic acids and their derivatives participate as pigmenting cofactors, therefore, a higher proportion of these compounds in wines made with smaller size grape berries could partially explain the intense color of these wines, because of intermolecular copigmentation (Gil et al., 2015).In the present work, the decrease in berry size, mediated by the rootstocks '3309 C' and '101-14 Mgt', seems to have been sufficient to increase the concentration of p-coumaric acid (in 'Merlot'/'3309 C') and resveratrol in 'Merlot'/'3309 C' and 'Merlot'/'101-14 Mgt'.
The stilbenes monomers, cis-and trans-resveratrol, are phytoalexins, what are components synthesized by the vine in response to a stress situation (Mattivi et al., 1993), and the ultraviolet (UV) light exposure and climatic conditions are examples of abiotic stress.The effects of sunshine duration and UV light exposure on the resveratrol concentrations of red wines were analyzed separately and were found to increase resveratrol concentrations to a considerable extent (Threlfall et al., 1999), with trans-resveratrol (trans-3,5,4-trihydroxystilbene) being the most studied due to its potential benefit to human health (Gris, 2010).The highest concentrations of these compounds were observed for the '101-14 Mgt' rootstock, which provided a better vegetativeproductive balance, which may have favored an increase in solar radiation in vegetative canopy and in the clusters, contributing to the increase of these constituents.
The present work demonstrated that the vegetative growth, yield, wine chemical and phenolic composition of 'Merlot' variety grown in altitude region of Santa Catarina is directly related to the rootstocks used.

CONCLUSIONS
The rootstocks '3309 Courdec' and '1014-14 Mgt' are recommended for the 'Merlot' grapevine in the soil and climate conditions of the Santa Catarina highlands.
The rootstock '3309 C' promoted an increase in the productive indices (yield, cluster per plant and cluster mass), and provided lower vegetative vigor, and consequently resulted in more adequate vegetative-productive balance for 'Merlot' plants.
The use of rootstock '101-14 Mgt' resulted in higher contents of total polyphenols, color intensity and anthocyanins, and showed the highest contents of flavonoid compounds.

CONFLICTS OF INTEREST:
The authors declare no conflict of interest.
same row followed by different letters differ significantly by Fisher's LSD test (p<0.05).Mean ± standard error.

Table I
Production components of 'Merlot' (Vitis vinifera L.) grafted onto different rootstocks, grown in altitude region, during the 2014/15,

Table III
Vegetative-productive balance of 'Merlot' (Vitis vinifera L.) grafted onto different rootstocks, grown in altitude region, during the 2014/15, 2015/16 and 2016/17 cycles Means within the same row followed by different letters differ significantly by Fisher's LSD test (p<0.05).Mean ± standard error.

Table V
Anthocyanins, total polyphenols and chromatic characteristics of 'Merlot' wines, originated from grapevines grafted onto different rootstocks, grown in altitude region, during the 2015/16 and 2016/17 cycles

Table VI
Flavonoid phenolic compounds of 'Merlot' wines, originated from grapevines grafted onto different rootstocks, grown in altitude region,