Vineyard site, plant material and weather conditions

The research was conducted in 2015 and 2016 on cultivars Merlot, Syrah, and Cabernet Sauvignon. The vineyard is located 20 km north of Zadar (Baštica, Suhovare) in Dalmatia region, subregion Dalmatian hinterland (latitude 44°06´N; longitude 15°13´E) and is a part of the University of Zadar, Croatia. All three grapevine cultivars were grafted on Kober 5BB (Vitis berlandieri Planch. × Vitis riparia Michx.) rootstock, which was planted in 2007 on anthropogenic soil called regosol with a sandy clay texture. The vines were planted with a spacing of 90 cm within the row and 280 cm between rows (planting density of 4100 vines per ha). All three grapevine cultivars were trained to a vertical shoot position with single-cane-pruned Guyot, leaving about 12 to 14 buds per vine. The basal wire was placed at 100 cm above the ground, with two sets of catch wires positioned 50 and 90 cm above the cordon. The maximum canopy height was 200 cm. The experimental field had no irrigation system and the space between the rows was grassed. The same vineyard management practices were used for all treatments.

The beginning of the main grapevine phenophases was determined visually. Full flowering was estimated when 50 % of the flower caps had fallen off, corresponding to stage 23 according to the modified Eichorn and Lorenz (E-L) system (23), while vérasion was estimated when the berries started to brighten in colour, corresponding to stage 35 according to the same scale.

The harvest date was determined by measuring the total soluble solids (Brix), total acids (g/L), and pH. Harvesting began when the total soluble solids were above 19 °Brix. Grapes were harvested manually at different times depending on the grapevine variety and measured parameters. Grapes for each treatment were harvested separately.

Weather conditions, including average temperature and precipitation for both seasons from April to September, were measured by the Croatian Meteorological and Hydrological Service (Weather station Benkovac), 25 km from the experimental vineyard, and the data are shown in Table S1. The weather conditions were reflected at the beginning of flowering and vérasion, and also during the leaf removal treatments. The harvest time differed only by a few days in both years. Merlot was harvested three days earlier in 2016 (22 September) than in 2015 (25 September), probably due to the previously mentioned dry period in July, which affected the slightly earlier harvest. Syrah was harvested on the same day as Merlot in 2015 (25 September), but 8 days later than Merlot in 2016, on 30 September. Cabernet Sauvignon was harvested on 9 October in 2015, and on 13 October in 2016.

Experimental design

The experiment was a completely randomised block design with three treatments in three replications for each cultivar. Each replication consisted of 15 continuous plants, so there were 135 plants per cultivar and 405 plants in total. The cultivars were in the same vineyard, but one was next to the other, so the experiment was set up in the same way for different grapevine varieties. All treatments were repeated for two years in the same part of the vineyard.

The three treatments were: (i) leaf removal during flowering (full flowering, 50 % open flowers); (ii) leaf removal during vérasion (beginning of vérasion, 30 % of the berries are coloured), and (iii) control (C) - without leaf removal. In both leaf removal treatments, the basal leaves were removed up to the height of the last cluster on the shoot (4 to 6 leaves).

Vinification

Manually harvested grapes were destemmed and crushed separately for each variety and treatment and placed in an open plastic container (100 L) for maceration and fermentation. All vinifications were sulphited with 5 g K₂S₂O₅ per 100 L and after a few hours, Saccharomyces cerevisiae yeast (ICV D254; Lallemand, Montreal, Canada) was inoculated at a concentration of 25 g/100 L. The pomace was stirred manually twice a day and the temperature was between 25 and 28 °C. After seven days of maceration and fermentation, the wine was racked and fermentation continued in glass containers. At the end of fermentation, the wine was additionally sulphited with 5 g K₂S₂O₅ per 100 L, racked again and bottled in 0.75-litre bottles two months after the end of fermentation.

Must samples were collected immediately after primary processing, for analysis of total soluble solids, titratable acidity and pH. Total soluble solids in the must were measured using a handheld refractometer (RHB 32 ATC; PR China) (expressed in °Brix) and pH was determined with a pH meter (Lab 860; Schott Instruments; Mainz, Germany). The titratable acidity (g/L) was determined using the colouration pattern volumetric method according to the O.I.V. (24).

The wine was stored and matured under the cellar conditions for one year after bottling. Samples for analysis were taken at random in triplicate after bottling, i.e. after 0 months and after 6 and 12 months.

Analysis of phenolic content using HPLC-DAD

The concentration of anthocyanins and other phenolic compounds (phenolic acids, procyanidins, flavan-3-ols and flavonol glycosides) was determined in all wine samples using high-performance liquid chromatography (HPLC). The wine samples were filtered through 0.45-µm syringe filters (Macherey-Nagel GmbH & Co. KG, Duren, Germany) into glass vials and analysed using the HPLC Agilent Infinity 1260 system equipped with an Agilent 1260 photodiode array detector (PDA; Agilent, Santa Clara, CA, USA), an automatic injector and Chemstation software (v. C.01.03) for data processing and instrument control. Phenolic compounds were separated using Luna 100-5C18 column, 5 µm (250 mm×4.6 mm; Phenomenex, Aschaffenburg, Germany). The injection volume was 5 µL, and the solvent composition and gradient conditions were as previously described by Zorić et al. (25).

All anthocyanins were identified at λ=520 nm by comparing their retention times and absorption spectra with those of authentic standards. All identified compounds were quantified according to the calibration curves of standards. Standards of delphinidin-3-glucoside (Del-3-Glc), cyanidin 3-glucoside (Cy-3-Glc), petunidin-3-glucoside (Pet-3-Glc), peonidin-3-glucoside (Peo-3-Glc) and malvidin-3-glucoside (Mal-3-Glc) were prepared as stock solutions at a concentration of 100 mg/L in methanol acidified with φ(formic acid)=1 %. The stock solutions were diluted to obtain five concentrations between 20 and 100 mg/L.

Phenolic acids, procyanidins, flavan-3-ols and flavonol glycosides were identified by comparing the retention times and spectral data with those of authentic standards prepared in methanol, namely: chlorogenic acid, caffeic acid, p-coumaric acid, gallic acid, procyanidins B1 and B6, epigallocatechin gallate, catechin, quercetin-3-glucoside and kaempferol-3-rutinoside.

All results were expressed as mg/L in the form of mean value±standard deviation.

Statistical analysis

Statistica v. 14.0 software (26) was used for the statistical analysis. Descriptive statistics was used to assess the basic information about the experimental data set, and the data are presented as mean value±S.E. The normality and homoscedasticity of the data were analysed using the Shapiro-Wilk test and Levene’s test, respectively, and were evaluated accordingly by ANOVA coupled with the post hoc Tukey’s HSD test with multiple comparisons of mean ranks. A statistically significant difference at p≤0.05 was assigned for all tests.

Basic chemical parameters of the must

The basic chemical parameters measured in the must of three grape varieties (total soluble solids, titratable acidity and pH) were mainly influenced by the experimental year, while there was no significant difference among varieties and leaf removal treatments (Table 1).

The two experimental years differed in the average temperature and precipitation during the vegetation period, with 2015 being 0.7 °C warmer and having about 125 mm less precipitation. Furthermore, the ripening period was on average 1.3 °C warmer in August 2015 than in 2016. Higher temperatures have an effect on increased cell respiration, which leads to malic acid breakdown (27) and a lower acidity. This was observed in the 2015 samples, which had lower acidity and, consequently, higher pH than the 2016 samples. A similar observation that experimental year has a significant effect on basic chemical parameters, compared to leaf removal treatments, was made by Mosetti et al. (28) in Sauvignon blanc and Anić et al. (29) in Merlot, although in some cases, a mild influence of leaf removal treatments on basic chemical parameters was observed (30, 31). The time of leaf removal also had no influence on the basic chemical parameters of the must samples. There is no difference between the treatments in titratable acidity and pH, which is consistent with other research (32, 33).

Defoliation treatments did not affect the increase in total soluble solids, regardless of when they were applied. These observations are consistent with other studies (29, 32, 33).

Effect of leaf removal on anthocyanin content in wine

Leaf removal treatments positively influenced the accumulation of anthocyanin in all three grapevine varieties, which was expected and is consistent with other research on different varieties (29, 34, 35). Similar results were obtained by other authors. For example, in the research on the Italian cultivar Nebbiolo, the concentration of individual anthocyanins and polyphenols depended on the year and climatic conditions. Nevertheless, the total concentration was consistently higher in defoliated samples than in the control (35). The effect of the leaf removal on the composition of individual anthocyanins in Merlot, Syrah and Cabernet Sauvignon wine is shown in Table 2. In all three grapevine varieties, malvidin-3-O-glucoside (Mal-3-Glc) was the most abundant anthocyanin, with concentration depended on the year and leaf removal treatment, which is consistent with the studies of Shi et al. (36). The second most abundant anthocyanin in all three varieties was malvidin-3-O-acetyl-glucoside (Mal-3-Ac-Glc).

The influence of the time of defoliation on the concentration of total and individual anthocyanins depended on the variety and the year (Table S2, Table S3 and Table S4). Similar effects were observed on Merlot, Pinot noir and Gamay, where the experimental year had an important role in the success of the leaf removal treatments (32-37). In both years, there was a significant influence of defoliation treatment on the concentration of Pet-3-Glc, Malv-3-Glc, Peo-3-Coum-Glc, Mal-3-Ac-Glc and Mal-3-Coum-Glc in Merlot, while defoliation had no significant effect on the remaining individual anthocyanins. In contrast to Merlot, the defoliation treatments consistently increased the anthocyanin content of Syrah in both years, with defoliation during vérasion having the greatest effect. Unlike Merlot, the effect of defoliation did not vary significantly between years. In 2016, Pet-3-Coum-Glc was undetectable in all treatments. In Cabernet Sauvignon, the effect of leaf removal depended on the experimental year. Only defoliation during flowering in 2016 had a significant influence on the individual anthocyanin concentration in Cabernet Sauvignon wines, while the control wines had the highest anthocyanin content in 2015 (Table 2).

Regarding the time of leaf removal, different results have been reported. According to some studies, a higher concentration of anthocyanins was found after early leaf removal during flowering than after leaf removal during vérasion (11, 38), which is similar to our results in Merlot wines from 2015 and Cabernet Sauvignon wines from 2016 (Table 3 and Table 4). In contrast, Merlot from 2016 and Syrah from both years were found to have a significant influence on the increase in the concentration of total anthocyanins at defoliation during vérasion (Table 3 and Table 5). The highest concentration of total anthocyanins in Cabernet Sauvignon from 2015 was found in the control sample (Table 4).

The positive influence of early defoliation on anthocyanin concentration due to increased UV radiation was also recorded on the Merlot in the studies by Anić et al. (29). Due to the increasingly warmer years and the influence of high temperatures on anthocyanin degradation, late leaf removal at vérasion loses its advantages over early leaf removal during flowering. Comparing the effect of both defoliation treatments, Sternard Lemut et al. (39) measured a higher concentration of total anthocyanins in Pinot Noir when the leaves were removed early, in contrast to our results for Syrah from both years.

Anthocyanin content in wines during ageing

The wine ageing period had a significant effect on the reduction of the concentration of the individual anthocyanins in all three varieties analysed in both years (Table 6). Anthocyanin concentration decreased during wine ageing (Fig. 1), which is consistent with previous studies (40, 41). Although free anthocyanins are responsible for the red colour of young red wines, their concentration decreases significantly during wine ageing to as little as 0–50 mg/L, leading to a loss of colour in red wine (22).

Fig. 1

Effect of ageing and leaf removal treatment on anthocyanin content in: a) Merlot from 2015, b) Merlot from 2016, c) Syrah from 2015, d) Syrah from 2016, e) Cabernet Sauvignon from 2015 and f) Cabernet Sauvignon from 2016 wines. LRF=leaf removal during flowering, LRV=leaf removal during véraison

The decrease in the concentration of anthocyanins in wine is partly influenced by external factors (temperature, light and precipitation). Nevertheless, some of the anthocyanins decrease due to their instability and strong reactivity with other compounds. This refers primarily to reactions of anthocyanins with other anthocyanins and their co-pigmentation and to polymerisation reactions with flavan-3-ols and procyanidins, forming new pigments of proanthocyanins and polymeric anthocyanins that can stabilise the wine colour (19-21).

According to the available literature data, previous studies confirm a steady decrease in total anthocyanin content during bottle ageing for up to 42 months (42-45). Anthocyanin degradation during ageing was high in all three wine varieties, ranging from 36 to 90 %, depending on the year and treatment. In 2016, the degradation of anthocyanins ranged from 36 to 70 %, while in 2015, the degradation was 65 to as high as 90 %, depending on the variety (Fig. 1). The highest degradation of anthocyanins was observed in Merlot wine from 2015 in the treatment of defoliation during flowering, and increased to a high 90 % after 12 months of ageing.

Anthocyanin concentration in Merlot decreased during ageing, with Pet-3-Glc, Pet-3-Coum-Glc and peonidin-3-O-coumaroyl-glucoside (Peo-3-Coum-Glc) being the most unstable. Peo-3-Glc and Pet-3-Coum-Glc were no longer detectable in any treatment after 12 months. However, the wine from the 2016 obtained after leaf removal during vérasion retained the highest total anthocyanin concentration after ageing, even though some of the individual compounds could not be detected, i.e. were degraded in the wines after 12 months of storage or were present in very low concentrations (Table S2).

In Syrah, the leaf removal during vérasion had the most significant positive influence on the anthocyanin content in both years (Table 5). Pet-3-Coum-Glc was undetectable in all treatments in 2016. The anthocyanin stability in the wine varied depending on the year and treatment. The treatment with the most stable anthocyanins in the wine after 12 months of ageing seems to be the 2015 control. In 2016, the control and the samples with the leaves removed during vérasion had the same effect on the stability of anthocyanins in the wine (Fig. 1 and Table S3).

The effect of ageing on the anthocyanin concentration in Cabernet Sauvignon wine is shown in Table 6. In 2015, all varieties had a significant loss of anthocyanins during ageing, while in 2016, the stability of anthocyanins in the stored wines was similar in the control and the samples defoliated during flowering. As with Syrah, Pet-3-Coum-Glc was undetectable in the samples from 2016, and Pet-3-Glc was the most unstable anthocyanin in the wine and disappeared from all wines after 12 months (Fig. 1 and Table S4).

Only in Merlot from 2016 and Cabernet Sauvignon from 2015 did the removal of leaves had a positive effect on the stability of anthocyanins. The degradation of anthocyanins in the samples defoliated during vérasion and during flowering was lower than in the control. A lower percentage of degradation was observed in the samples defoliated during vérasion.

Although there are significant differences between the anthocyanin contents in the wines after 12 months of ageing (Table 6), they cannot be related to the influence of the leaf removal treatments, regardless of their effect on the increase in anthocyanin concentration in young wines. This can be explained by the fact that the stability of anthocyanins in wine is influenced by a number of factors, such as wine storage conditions and cultivars, but also by the different reactions that anthocyanins undergo during wine ageing (40).

Furthermore, the degradation of anthocyanins seems to have been lower in the colder year 2016 than in the warmer year 2015. This could be explained by the differences in the basic chemical parameters, i.e. pH differences, since the stability and colour of red wines are strongly influenced by pH and the amount of free sulphur dioxide (20). The red colour of the wine comes mainly from the anthocyanins, which are in the flavylium state and whose concentration depends on the pH and the free sulphur dioxide. At a low pH, the concentration of the flavylium state increases, the hydrolysis of anthocyanins slows down and the colour is more intense, while with an increase in pH, the colour intensity and the concentration of anthocyanins in the flavylium state decrease significantly (20). The grapevine variety significantly influenced the total anthocyanin content in the wine. Syrah wine had the highest anthocyanin content compared to Merlot and Cabernet Sauvignon wines (Table 7).

Similar results were obtained with Merlot, Syrah, Cabernet Sauvignon and Marselan in the study by Shi et al. (36). Differences in the anthocyanin content are cultivar-specific. However, the accumulation of anthocyanins in grapes is influenced by other factors, such as agroecological conditions, climate, soil conditions, canopy management and irrigation, agrotechnical practices and yield (2, 46). A significantly higher concentration of total anthocyanins was found in the drier and warmer year 2015 than in 2016 (Table 7), which is in contrast to previous results (7, 29), which confirm that increased solar radiation and temperature in the fruit zone reduce anthocyanin accumulation in the berry skin.

Effects of leaf removal on other phenolic compounds in wines

Considering that anthocyanins react with other phenolic compounds in polymerisation reactions during wine ageing, other groups of phenolic compounds were analysed in all wines. In both analysed years, both leaf removal treatments increased the concentrations of total phenolic acids (TPA), total procyanidins (TPro), total flavan-3-ols (TFL-3-ols) and total flavonol glycosides (TFG) in Merlot wines compared to the control (Table 3). At the same time, such an effect was not observed in Cabernet Sauvignon and Syrah wines (Table 4 and Table 5).

This result could be a consequence of cultivar characteristics and canopy porosity, which has already been suggested by Tardaguila et al. (14). Differences were observed in the influence of the time of the leaf removal treatment, so earlier leaf removal during flowering affected the increase of TFG in Merlot and Syrah in both years, which is consistent with other studies (29, 32). Defoliation increases sun exposure and UV radiation in the grape zone. Flavonols protect plants from excessive UV radiation, and their accumulation is strongly influenced by environmental conditions (6, 47). Together with anthocyanins in co-pigmentation processes, flavonols form more complex compounds that affect colour stability and wine quality (48). Leaf removal during vérasion led to an increase in TFL-3-ol also in Merlot and Syrah wines in 2015 and 2016, which is in contrast to the results reported by Osrečak et al. (33). The end of the synthesis of flavan-3-ols in berry skin is around vérasion, so it is considered that the practice of late defoliation cannot be reflected in their concentration.

Differences in the results also exist between the two years of study, which can be related to the different meteorological and microclimatic conditions in the two vegetation seasons. Other authors also confirmed this and stated that the vintage year effect plays an important role in the successful implementation of the treatment (37).

Other phenolic compounds during wine ageing

The concentration of TPA increased during wine ageing in all treatments in Syrah wines in both years (Table 5) and also in Merlot and Cabernet wines in 2015 (Table 3 and Table 4), which is consistent with previous results (40, 49). The highest concentration of phenolic acids was found in the Syrah wine from 2016 and the lowest in the Cabernet Sauvignon wine from 2015. The TPro concentration in all wines decreased during ageing, except for the Cabernet control wine from 2016, and their degradation ranged from 5 to 42 %, depending on variety, year and treatment (Table S5).

The concentration of TFL-3-ols and TFG increased in some varieties, while it decreased in others during ageing. The highest concentration of TFL-3-ols was measured in Syrah wine samples defoliated during vérasion in 2015, while the highest degradation of TFL-3-ols was 33 % in Syrah samples defoliated during vérasion in 2016 and Merlot samples defoliated during flowering in 2015 (Table S5). The concentration of TFG decreased in all wines except Merlot from 2016. The lowest percentage of degradation was found in Cabernet Sauvignon wine samples defoliated during flowering in 2016, and the highest was over 70 % in Syrah samples defoliated during vérasion in 2015. The percentage of degradation of TFL-3-ols during ageing was largely influenced by the grape variety (46).

The interaction between the leaf removal treatment and wine ageing had a significant effect on the concentration of phenolic compounds in all three wines (Table 3, Table 4 and Table 5).