See discussions, stats, and author profiles for this publication at: https://www.researchgate.

net/publication/348846648

Early Leaf Removal Increases Berry and Wine Phenolics in Cabernet

Sauvignon Grown in Eastern Serbia

Article in Agronomy · January 2021

DOI: 10.3390/agronomy11020238

CITATIONS READS

0 52

5 authors, including:

Ljiljana Kostic Miroslav Nikolic

Institute for Multidisciplinary Research University of Belgrade
25 PUBLICATIONS 286 CITATIONS 90 PUBLICATIONS 2,584 CITATIONS

SEE PROFILE SEE PROFILE

Some of the authors of this publication are also working on these related projects:

Plant Adaptations to Mineral Stress on Marginal Agricultural Soils View project

All content following this page was uploaded by Nina Nikolic on 10 February 2021.

The user has requested enhancement of the downloaded file.

 agronomy
Article
Early Leaf Removal Increases Berry and Wine Phenolics in
Cabernet Sauvignon Grown in Eastern Serbia
Dejan Stefanovic 1,2,† , Nina Nikolic 3,† , Ljiljana Kostic 3 , Slavica Todic 1 and Miroslav Nikolic 3, *

1 Faculty of Agriculture, University of Belgrade, Nemanjina 6, 11080 Belgrade, Serbia;

[email protected] (D.S.); [email protected] (S.T.)
2 Agricultural Extension Service (PSSS), Bukovski Put bb, 19300 Negotin, Serbia
3 Institute for Multidisciplinary Research, University of Belgrade, Kneza Viseslava 1, 11030 Belgrade, Serbia;
[email protected] (N.N.); [email protected] (L.K.)
* Correspondence: [email protected]; Tel.: +381-11-3058-956
† Both authors contributed equally to this work.

Abstract: Cluster zone leaf removal is a well-established viticulture practice for improving cluster mi-
croclimate and wine quality in cooler climates, while its efficacy in warmer conditions is less is known.
Here we compared the effect of early (ELR, after fruit set; diameter of berries 3–5 mm) and late (LLR,
beginning of veraison) leaf removal on berry composition and wine phenolic profile of grapevine
(Vitis vinifera L.) variety Cabernet Sauvignon grown in a temperate, warmer region of Eastern Serbia.
Compared to the control (no leaf removal), both leaf removal treatments increased the sugar content
in fresh juice and alcohol concentration in wine. Over three consecutive years (2011–2013) markedly
different in temperature and rainfall, ELR was clearly most effective in decreasing weights of cluster





 and of one berry, and in increasing of skin share in a berry. The content of total phenols, tartaric
acid esters, anthocyanins, and flavanols in berry skin and wine was the highest in ELR treatment.
Citation: Stefanovic, D.; Nikolic, N.;
ELR prominently modified the phenolic profile: Increasing flavanols, myricetin and quercetine in
Kostic, L.; Todic, S.; Nikolic, M. Early
Leaf Removal Increases Berry and
skin and wine, and anthocyanins, peonidin-3-glucoside in skin and delphinidin-3-glucoside in wine.
Wine Phenolics in Cabernet This work demonstrated that early leaf removal positively influenced the chemical composition of
Sauvignon Grown in Eastern Serbia. berries and wine of Cabernet Sauvignon and might be recommended for practice in the temperate
Agronomy 2021, 11, 238. https:// warm conditions.
doi.org/10.3390/agronomy11020238
Keywords: anthocyanin; berry skin; canopy management; flavanol; grape yield; wine quality
Academic Editors: Luciano Gristina,
Ettore Barone, Rosario Di Lorenzo
and Antonino Pisciotta
Received: 17 December 2020 1. Introduction
Accepted: 22 January 2021
Cluster zone leaf removal (also termed as defoliation) is one of the most effective
Published: 28 January 2021
canopy management practices for altering microclimatic conditions surrounding the grape
cluster (e.g., better aeration and sun light exposure of clusters, decreased humidity) and
Publisher’s Note: MDPI stays neutral
has shown to improve berry and wine quality [1,2]. The canopy leaf area greatly influences
with regard to jurisdictional claims in
published maps and institutional affil-
the changes in the properties of clusters and berries, such as the contents of sugars, total
iations.
acids, and colored and aromatic compounds. Accordingly, partial removal of leaves in the
cluster zone has been recognized as one of the most important activities in the vineyard.
The modification of canopy microclimate and light exposure of clusters by leaf removal
in the cluster zone showed the influence on the chemical and organoleptic properties of
grapes and wines [3–8]. This effect mainly depends on the degree of leaf biomass removal,
Copyright: © 2021 by the authors.
and on the timing [8,9]. In most vine regions worldwide, leaf removal is usually performed
Licensee MDPI, Basel, Switzerland.
in the period after flowering to the beginning of veraison. Leaf removal from the cluster
This article is an open access article
distributed under the terms and
zone is usually practiced in cooler to moderate climates to increase temperature and light
conditions of the Creative Commons
quality in the cluster zone [8,10]. However, in the warm climates a removal of leaves in
Attribution (CC BY) license (https:// cluster zone, in particular before or during veraison, may cause sunburns and overheating
creativecommons.org/licenses/by/ of berries, which lowers the wine quality [11]. In a cooler climate, early leaf removal usually
4.0/). decreases grape yield due to decreased berry size, but in parallel increases components

Agronomy 2021, 11, 238. https://doi.org/10.3390/agronomy11020238 https://www.mdpi.com/journal/agronomy

Agronomy 2021, 11, 238 2 of 17

of wine quality [12]. However, leaf removal in warmer climates has less been studied. In
such climates, early leaf removal and subsequently increased cluster exposure to sun light,
although less affecting grape yield, increased berry sugar and phenolics, and maintained
must acidity [8], but occasionally led to berry discoloration [13].
Phenolic compounds are very important for the quality of red wines, affecting both
organoleptic characteristics and benefits for human health [14]. The content and composi-
tion of the phenolic compounds depends on the variety, ecological conditions, and vineyard
management. Hunter et al. [5] demonstrated that the concentration of anthocyanins in-
creased with partial defoliation at veraison, while the wine quality was considerably
improved by this measure regardless of the timing. Late defoliation (post-veraison and
pre-harvest) on Sangiovese increased the total flavanols in berry skin; however, it did
affect the total anthocyanins [15]. Moreover, in the same study [15], no significant differ-
ences between late defoliation and control (no defoliation) treatments were found for total
polyphenols, alcohol, titratable acidity, and pH in wine.
Most of phenolic compounds, especially anthocyanins, are concentrated in the berry
skin. Early leaf removal (after flowering to the berry set stages), for instance, reduced the
size and number of berries and increased the skin/flesh ratio, resulting in an increased
content of phenolic compounds in wine [16]. It has further been shown that the smaller
size of berries was a result of the reduction of the assimilation surface during the flowering
and fruit set stages [17]. Leaf removal also resulted in the higher sugar content in must,
increased phenolic content in berry skin and more of stable anthocyanins in wine [18,19].
Despite the extensive research in the past decade, the effect of the timing of cluster zone
leaf removal for the specific region and rather continental climatic conditions of Eastern
Serbia characterized by warm summertime (July–September), is not well understood. It has
been hypothesized that the cluster zone leaf removal will significantly influence the quality
of both grapes and wine of Cabernet Sauvignon grown in the specific climate conditions of
Eastern Serbia. Therefore, the aim of this study is to evaluate the effects of cluster zone leaf
removal timing (early and late, compared to no leaf removal) on the quality of Cabernet
Sauvignon grapes and wine with special emphasis on the phenolic profile.

2. Materials and Methods

2.1. Vineyard Description
This research was conducted over three consecutive seasons (2011–2013) on the Vi-
tis vinifera L. cv. Cabernet Sauvignon. The vineyard was planted in 2001 on rootstock
5BB, with the 3333 plants per hectare (3 × 1 m). The location of vineyard is in Negotinska
Krajina wine region, Serbia, South-eastern Europe (44◦ 040 0500 N and 22◦ 330 2600 W; altitude
of 140 m). Bud break for Cabernet Sauvignon in this area usually occurs by mid April,
flowering by mid June; veraison by early August, with harvest by late September, and leaf
fall in November. The climate of the experimental area is temperate mesothermal with
hot and humid summers (Köppen class Cfa), characterized by both sub-continental and
sub-Mediterranean influences [20]. With average annual precipitation about 570 mm (of
which 54% is delivered during vegetation period), Lang’s rain factor about 45 and de Mar-
tonne’s aridity index about 24, the research locality is under semi-arid conditions. Mean
annual temperature is 11.8 ◦ C and average daily temperature from April to September of
19.2 ◦ C [21].
The vineyard was on a slight slope with the row orientation approximately north-west
to south-east. The soil was a Vertisol; pH (in 1 M KCl) 6.07, 1.87% of soil organic matter,
9.7 mg kg−1 soil of ammonium lactate-extractable P and 80.5 mg kg−1 of exchangeable K;
no fertilization/irrigation during the entire experiment were applied. Vine training system
was a modified two-arm cordon with a trunk height of 1.4 m. The number of buds retained
by pruning was 24, distributed on two canes with ten buds and two replacement spurs
with two buds each. The locally common soil management and fungicides were applied to
all the treatments.
Agronomy 2021, 11, 238 3 of 17

2.2. Experimental Design

Leaf removal from the cluster zone was done manually in two period of times: Early
leaf removal (ELR) after berry set (diameter of berries 3–5 mm; BBCH 73 growth stage,
Lorenz et al., [22] and late leaf removal (LLR) at the beginning of veraison (BBCH 81 growth
stage), with the same intensity (removing 5 fully developed basal leaves per shoot), and no
leaf removal in the control. The lateral shoots were retained. The experiment was carried
out in a random block design with three blocks and three treatments with 10 plants per
treatment in each block.

2.3. Grape Harvest, Analyses of Grape Juice and Estimation of Vegetative Growth
At commercial harvest time all clusters per vine were harvested (in the stage of
“technological maturity” reached at different dates: 25 September 2011, 15 September
2012 and 28 September 2013). Ten vines were harvested in each replicate (block) per
treatment. The clusters were weighted to determine the yield and mean cluster weight.
The berries were counted from 30 randomly taken clusters per each replicate. In addition,
50 berries were randomly taken from each sample for weighting and then crushed by
fingers in order to remove the skins. The skins were rinsed with tap water to completely
separate the pulp and, after which they were blotted with filter paper and weighed. The
relative skin share (%) was calculated following equation: skin weight × 100/berry weight.
The composite grape samples from 10 vines (each containing 12 kg) were divided into two
parts; one part (about 2 kg) was frozen (−20 ◦ C) for further chemical analyses, and the
other (about 10 kg) was crushed by hand and then subjected to microvinification.
At harvest, total sugar content in the fresh grape juice was determined by portable
refractometer (MR200ATC Milwaukee Meters, Neutron Pl Rowville, VIC, Australia) and
the titratable acidity (total acid concentration) was determined by titration with 0.1 N
NaOH to a pH 8.2 end point.
The vine vegetative growth was estimated by measuring pruning weight per vine
during the winter period following the growing season.

2.4. Microvinification
The must was treated with potassium metabisulfite (0.1 g kg−1 ) and inoculated with
the commercial wine yeast (BDX, Lallemand Oenology, Castel dAzzano, Italy) (0.25 g kg−1 ).
After 10 days of maceration and primary fermentation in 15 L plastic pots, the fermenting
juice was removed and the skins are pressed by hand to extract the remaining juice. The
secondary fermentation was continued in 5 L glass demijohn for the next 10 days and then
treated with 0.05 g of potassium metabisulfite. After one month wine was transferred to
another 5 L glass demijohn for and after 3 months wine was bottled (0.75 L) and aged for
6 months prior to analyses.

2.5. Chemical Analyses of Berry Skin and Wine

Thirty berries were randomly taken from each frozen sample (2 kg collected from
10 vines), the skin was removed by forceps, homogenized in a mortar using liquid nitrogen,
and 1 g of the skin homogenate was taken and mixed with 10 mL 50% methanol in the plas-
tic tubes. Tubes were closed, shaken and left for 60 min at room temperature, with shaking
every 10 min. The supernatants were subjected to further analysis of phenolic compounds.
Total phenols, total tartaric acid esters and total flavanols were determined following
the method of Mazza et al. [23]. The samples (skin extract as described above or 6-month
bottle aging wine) were mixed with acidified ethanol (2% HCl in 95% ethanol) in the ratio
of 1:10 (v/v), incubated for 15 min at room temperature, filtered through a membrane filter
(0.45 µm pore size) and measured spectrophotometrically (Agilent 8453 UV-VIS Spectropho-
tometer, Santa Clara, CA, USA) at 280 nm (total phenols; gallic acid standard), 320 nm
(tartaric acid esters; caffeic acid standard) and 360 nm (total flavanols; quercetin standard).
The concentrations were expressed in mg standard equivalent per g of berry skin.
Agronomy 2021, 11, 238 4 of 17

Total anthocyanins were determined by the modified method of Di Stefano [24].

After 15-min incubation of 0.5 mL samples (skin extract as described above or 6-month
bottle aging wine) in 10 mL solution containing mixture ethanol:d-water:hydrochloric acid
(70:29:1 v/v/v), the absorbance was measured spectrophotometrically (Agilent 8453 UV-
VIS Spectrophotometer, Santa Clara, CA, USA) at 520 nm, and the total anthocyanins were
expressed as malvidin-3-glucoside per g of berry skin.
In addition, the individual phenolic compounds (flavanols and anthocyanins) were
determined in the berry skin and wine samples from the year 2012, which happened to be
rather typical for the Eastern Serbia terroir (see Table 1 and Supplementary Table S7). The
profile of flavanols (e.g., malvidin-3-glucoside, peonidin-3-glucoside, petunid-3-glucoside,
cyanidin-3-glucoside and delphidin-3-glucoside) and anthocyanins (e.g., quercetin-glucoside,
quercetin, rutin, morin, myricetin, and kaempferol) was determined by the high-performance
liquid chromatography (HPLC, Agilent Technologies 1200, Santa Clara, CA, USA) with
diode array detector (DAD), whereas the catechin was determined by HPLC with fluores-
cence detector (275–322 nm). Prior to the HPLC analysis, the extracts were further filtered
through membrane filters with a pore size of 0.45 µm. Separation was carried out on the
column Agilent-Eclipse XDB C-18 4.6 × 150 mm; the column was calibrated at 30 ◦ C; the
solvents were: Formic acid:water (5:95 v/v) − solvent A, and acetonitrile:formic acid:water
(80:5:15 v/v) − solvent B. The linear elution gradient was used as follows: From 0 to 10 min,
0.0% B, 10 to 28 min, 10.0% B, 28 to 35 min, 25% B, 35 to 40 min, 50% B, 40 to 45 min, 80%
B and for the last 10 min again 0% B. The injection volume was 5 µL and the flow of the
mobile phase was 0.9 mL min−1 . For detection wavelengths, 280, 320, 360, and 520 nm
for DAD were selected, and 275/322 nm (λEx /λEm ) for fluorescent detection. Different
phenolic compounds were identified by comparing their retention times and spectral char-
acteristics with the data of original standard components. Quantification was performed
using external calibration with the standards of phenolic compounds (HPLC-grade, ≥98%,
Sigma-Aldrich, St. Louis, MO, USA). The calibration curves were linear, with the coefficient
of determination R2 = 0.99.

Table 1. The selected weather parameters of the experimental years (2011–2013). Growing degree
days (GDD; base 10 ◦ C), rainfall registered during the bud burst to harvest period, annual rainfall,
insolation and average daily maximum air temperature (AMDT) during the veraison to harvest
period for Cabernet Sauvignon (August–September).

Parameter 1 2011 2012 2013

Growing degree days (GDD) 1784 1976 1686
Average daily maximum air temperature (AMDT) (◦ C) 30.3 32.1 28.9
Insolation June–September (h) 1951 2009 1872
Rainfall April–September (mm) 148 239 280
Annual rainfall (mm) 352 533 700
1 Data from the Republic Hydrometeorological Service of Serbia (http://www.hidmet.gov.rs).

Alcohol in the wine samples was determined using pycnometer method. The titratable
acidity (total acid concentration) in wine was determined by titration with 0.1 N NaOH
to a phenolphthalein end point. Volatile acidity in wines was determined by distillation
using a Cazenave-Ferré followed by titration with phenolphthalein.

2.6. Statistical Analysis

The data were statistically analyzed by the 2-way ANOVA model with interaction
term, after normality of residuals and homoscedasticity tests. The two treatments were leaf
removal (control–no leaf removal, late leaf removal, and early leaf removal), and growth
season (i.e., experimental year: 2011, 2012, and 2013). The significance of the differences
between treatment means were analyzed by Tukey’s Honestly significant test, using a
conventional level of significance of 5% (α = 0.05). Analyses of variance were performed
by the STATISTICA 6 software (StatSoft Inc., Tulsa, OK, USA).
Agronomy 2021, 11, 238 5 of 17

The multivariate structure of the collected data was visualized by Principal Com-
ponent Analysis (PCA) by PC-ORD 6 software (MjM Software Design, Gleneden Beach,
OR, USA), using correlation cross-product matrix, Euclidean distance and no rotation.
Correlation cross-product matrix was centered and standardized (relativized by standard
deviate), what brought all the analyzed variables to an equal footing. The differences
among the chemical profiles of skin and wine, induced by the leaf removal treatment, were
analyzed by Multi-Response Permutation Procedure (MRPP; Mielke and Berry [25]). This is
a nonparametric procedure for testing the hypothesis of no difference between pre-defined
groups; data were relativized by standard deviate to bring all the variables to the same
footing prior to analyses and be in accordance with the PCA.

3. Results and Discussion

3.1. Vine Growth and Grape Yield
Vine vegetative growth, grape yield and yield components were differently affected by
leaf removal treatment and experimental year (Table 2, Supplementary Table S1). Figure 1
shows the main effect of a factor with the strongest influence on each measured parameter
(after ANOVA analysis, Supplementary Table S1). Overall, pruning weight (Figure 1a),
grape yield (Figure 1b), and number of clusters per vine (Figure 1c) were strongly modified
by experimental year and the highest in 2013, which was incidentally unusually moist
year (Table 1; see also Supplementary Table S7). Vegetative growth (expressed as pruning
weight per plant) and number of clusters per vine were even not affected by leaf removal
treatment (Table 2, Supplementary Table S1). Compared to other years, pruning weight per
vine was on average by 37% higher in the year with the highest amount of precipitation
(2013—see Table 1), Figure 1a). Grape yield ranged from 1.99 kg (ELR treatment in 2012) to
5.08 kg (LLR treatment in 2013) (Table 2). Across all the canopy management strategies
evaluated (no leaf removal, early, and late leaf removal), the grape yield was in 2012 by 43%
lower compared to the year 2013, and by 21% lower compared to the year 2011. However,
across the three growth seasons, the average grape yield was the lowest in ELR treatment
(3.34 ± 1.14 kg per vine); it decreased by 9% in comparison with the LLR treatment, and
by 17% compared to the control. The number of clusters per vine (Figure 1c) followed the
same pattern as grape yield; it was the highest in 2013, and decreased by 73% in 2012.

Table 2. Effects of leaf removal time and experimental year on the vegetative growth and yield components of Cabernet
Sauvignon. LLR, late leaf removal; ELR, early leaf removal. Data are means of three replicates ± sd. Different letters in a
column denote significant differences among treatments according to Tukey’s test following a two-way ANOVA at α = 0.05.

Experimental Pruning Weight Grape Yield Clusters Cluster Berries Weight One Berry
Treatment
Year (kg Vine−1 ) (kg Vine−1 ) per Vine Weight (g) (g Cluster−1 ) Weight (g)
Control 1230 ± 27 a 4.0 ± 0.4 cd 36 ± 2 b 114 ± 3 bc 110 ± 3 d 1.02 ± 0.01 b
2011 LLR 1089 ± 46 a 2.9 ± 0.1 b 27 ± 2 a 106 ± 6 bc 101 ± 6 bcd 1.00 ± 0.03 b
ELR 1177 ± 65 a 3.5 ± 0.4 bc 35 ± 2 b 100 ± 13 bc 96 ± 13 bc 0.81 ± 0.04 a
Control 1212 ± 15 a 3.2 ± 0.1 b 26 ± 2 a 117 ± 7 d 113 ± 7 bcd 1.02 ± 0.02 b
2012 LLR 1141 ± 52 a 3.1 ± 0.2 b 27 ± 2 a 115 ± 6 bc 111 ± 6 cd 1.0 ± 0.1 b
ELR 1173 ± 72 a 2.0 ± 0.1 a 27 ± 2 a 76 ± 1 a 73 ± 0.9 a 0.80 ± 0.02 a
Control 1576 ± 73 bc 4.9 ± 0.3 e 46 ± 3 c 107 ± 4 bc 103 ± 4 bcd 1.07 ± 0.02 b
2013 LLR 1698 ± 31 c 5.1 ± 0.3 e 47 ± 4 c 108 ± 4 bc 104 ± 4 bcd 1.04 ± 0.02 b
ELR 1526 ± 89 b 4.5 ± 0.1 de 46 ± 3 c 97 ± 10 b 93 ± 10 b 0.9 ± 0.1 a
Agronomy 2021, 11, 238 6 of 17

Figure 1. Effect of leaf removal and experimental year on the on the vegetative growth and yield components. (a) pruning
weight (full model R2 adj. 0.94, F = 55.7, p = 0.000000); (b) grape yield (full model R2 adj. 0.96, F = 29.4, p = 0.000000);
(c) number of clusters per vine (full model R2 adj. 0.92, F = 40.2, p = 0.000000); (d) cluster weight (full model R2 adj. 0.78,
F = 12.5, p = 0.000006); (e) weight of berries per cluster (full model R2 adj.0.77, F = 12.2, p = 0.000007); (f) weight of one berry
(full model R2 adj. 0.93, F = 45.0, p = 0.000000). Mean values marked with the same letter are not significantly different at
α = 0.05. Vertical bars denote 95% confidence intervals.

The weight of clusters (Figure 1d), weight of berries in a cluster (Figure 1e), and weight
of one berry (Figure 1e) were decisively influenced by canopy management practices, i.e.,
by leaf removal (Supplementary Table S1). For the weight of clusters and weight of berries
per cluster, main effect of growth season could not even be detected. These three parameters
were the lowest in ELR treatments (see also Table 2). In particular, the weight of a single
berry in vines subjected to early leaf removal was consistently the lowest (compared to
control and LLR treatments) in each experimental year (Table 2). Across three years, these
three parameters did not differ between control and LLR. The ELR treatment however
1
Agronomy 2021, 11, 238 7 of 17

reduced cluster weight by 18%, weight of berries by 23%, and weight of one berry by 20%
in comparison to the averages in control and late removal treatments (Figure 2).

Figure 2. Effect of leaf removal and experimental year on the quality parameters of grape juice. (a) Concentration of sugars
(total soluble solids (TSS)) (full model R2 adj. 0.99, F = 104.3, p = 0.000000); (b) concentration of titratable acidity (full model
R2 adj. 0.96, F = 29.4, p = 0.000000). Mean values marked with the same letter are not significantly different at α = 0.05.
Vertical bars denote 95% confidence intervals.

With an exception for the year 2011 (which was, incidentally, a very dry one; see
Table 1), the results for the other two years suggest that the ELR treatment might be an
efficient practice in controlling the yield. Since the yield depends on several factors, such
as number of berries and weight and number of clusters, the lower ELR yield in the
experiment is probably due to the lower average weight of clusters and berries compared
to control and LLR treatments. Present results are in agreement with the results of various
studies which reported that early leaf removal (flowering and berry set) is an effective
measure in controlling the yield in different climate conditions [9,26–28]. A lower grape
yield recorded in ELR treatment was mainly caused by a decreased berry weigh, and
thereby weight of clusters (Table 2; Figure 1), which is in accordance with the results of
Moreno et al. [26] in Tempranillo grown in a temperate warm vineyard similar to ours (see
Table 1 in this study and Table 1 in the study of Moreno et al.). However, Bledsoe et al. [29]
reported that defoliation did not affect grape yield components and composition (with
exception of an increased TSS in the early leaf removal treatment) of Sauvignon Blanc
grown in the climate conditions of Napa (California), with more robust training system
and lower vine density than in the present study.

3.2. General Quality Parameters of Gape Juice and Wine

Experimental year exhibited the strongest effect on both total sugar content (Figure 2a)
and titratable acidity (Figure 2b) in fresh grape juice (see Supplementary Table S2). The
tested canopy management practices, furthermore, did not show any statistically significant
effect on the total acidity of grape juice (Supplementary Table S2). On average, across
three leaf removal treatments, sugar content (expressed as TSS) in 2012 was by 12% and
15% higher than in 2011 and 2013, respectively. This increase in sugar content in the
year 2012 was most prominent in ELR treatment (Figure 2a), which is in agreement with
findings of Moreno et al. [26]. Likewise, total juice acidity across all the leaf removal
treatments in 2012 (on average 8.2 g L−1 ) was by 9% and by 18% higher than in 2011 and
2013, respectively. Furthermore, across three different growth seasons, TSS in juice was the
lowest in the control, and did not significantly differ among the two leaf removal timings.
Incidentally, the year 2012 happened to be very favorable for grapevine, with the highest
temperatures and insolation, and no drought stress (Table 1). Also, in Tempranillo grown in
La Rioja (Spain), basal leaf removal at fruit setting stage led to more ripened grape berries
Agronomy 2021, 11, 238 8 of 17

in terms of higher TSS and decreased acidity [30]. However, very recent study on Cabernet
Sauvignon grown in the North Serbia terroir showed that early leaf removal decreased
total acidity of grape juice, but did not affect TSS [31].
Consistently with the sugars and acids in fresh berry juice (Figure 2), the experimen-
tal year exhibited the strongest influence also on quality parameters in wine (Figure 3;
Supplementary Table S3). Nevertheless, it has been shown that wine quality might not
always be simply a function of TSS and berry acidity [8]. Alcohol concentration in wine was
the highest in the year 2012, and in particular in the in ELR treatment (15.7%, Figure 3a).
In an apparently most favorable year for wine (2012), alcohol concentration in wine was
by 24% higher than in other two experimental years. On average, across three years of
the experiment, the concentration of alcohol in wine was the lowest in the control treat-
ments (no leaf removal, Figure 3a). The years 2011 and 2013 conspicuously differed in the
rainfall (2011 was rather dry, and 2013 rather humid; Table 1, Supplementary Table S7).
and what coincided with a prominent difference in pruning weight, grape yield and num-
ber of clusters (Figure 1a–c). Yet no differences between these non-optimal years was
observed for the TSS in fresh juice (Figure 2a) nor for the alcohol in wine (Figure 3a).
Furthermore, LLR treatment caused, on average across the experimental years, a small, but
statistically significant increase of titratable acidity in wine by 4% compared to ELR, and
by 7% compared to the control treatments (Figure 3b). In contrast to our findings, wine
parameters such as alcohol content and titratable acidity were unaffected by defoliation of
Grenache vines grown in La Rioja terroir, although early leaf removal positively affected
wine color and sensory parameters [27]. On the other hand, the experimental year did
not markedly change the pattern of how the volatile acidity in wine were affected by leaf
removal (Figure 3c). Volatile acids in wine were, across the three experimental years, by up
to 30% lower in control, then in the treatments involving leaf removal (Figure 3c). These
results are in agreement with the findings that wine from shaded clusters of Shiraz showed
significantly lower alcohol content, total and titratable acidity [32].

3.3. Polyphenolic Composition of Grape Berry Skin and Wine

Interestingly, across three consecutive experimental years markedly differing in pre-
cipitation and temperature (Table 1), ELR treatment had a clear, consistently favorable
effect on polyphenolic composition of berry and wine. For instance, leaf removal treatment
was the strongest modifier of a proportion of skin in a fresh berry (Supplementary Table S4).
Share of skin was in fact the only parameter measured in this study not affected by leaf
removal x experimental year interaction. In this study, ELR treatment resulted in the
highest skin proportion in grape berry in all the experimental years, while there was no
difference between LLR and control treatments (Figure 4). Across the three experimental
years, average skin share of 6.71% in ELR treatment was 1.14-fold and 1.20-fold higher than
in LLR and control treatments, respectively. Early leaf removal (before or after flowering)
may also induce a so-called “photosynthetic shock” i.e., reduced overall photosynthetic
activity of grapevines, which in turn may cause a reduced influx of assimilates from
leaves to blossoms or newly formed clusters and consequently formation of the smaller
berries with an increased skin-to-flesh ratio [30,31]. Excessive sunlight exposure of clusters
due to an intensive leaf removal, together with high cluster temperature also resulted in
smaller berries comparing with those in the shade [13,32]. Increased percentage of the
skin (Figure 4) in smaller berries (Figure 1f) was associated with the increased content
of phenolic compounds in berry skin (Table 3), especially anthocyanins whose synthesis
occurs in the grape berry skin. Reduced mesocarp (flesh) and increased proportion of the
skin is crucial for the quality of the wine, since the phenolics from the skin are extracted
during the maceration process [17,31]. The highest proportion of the skin in the berry
mass (ELR treatment, Figure 4) consequently reflected the highest content of the phenolic
compounds in the ELR wine, extracted during the maceration process in each of the three
experimental years (Table 4).
Agronomy 2021, 11, 238 9 of 17

Figure 3. Effect of leaf removal end experimental 3

year on the quality parameters of wine (after
6 months of aging). (a) Concentration of alcohol (full model R2 adj. 0.998, F = 612.9, p = 0.000);
(b) concentration of total acids (full model R2 adj. 0.98, F = 46.6, p = 0.000000); (c) concentration of
volatile acids (full model R2 adj. 0.98, F = 198.0, p = 0.000000). Mean values marked with the same
letter are not significantly different at α = 0.05. Vertical bars denote 95% confidence intervals.
Agronomy 2021, 11, 238 10 of 17

Figure 4. Effect of leaf removal and experimental year on the relative proportion of skin in a grape
berry. Full model R2 adj. 0.75, F = 10.7, p = 0.000019. Mean values marked with the same letter are not
significantly different at α = 0.05. Vertical bars denote 95% confidence intervals.

Table 3. Effects of leaf removal time and experimental year on the concentrations of total phenols, tartaric acid esters, and
anthocyanins and flavanols in berry skin determined by HPLC. LLR, late leaf removal; ELR, early leaf removal. Data are
means of three replicates ± sd. Different letters in a column denote significant differences among treatments according to
Tukey’s test following a two-way ANOVA at α = 0.05.

Experimental Total Phenols Total Tartaric Acid Esters Total Anthocyanins Total Flavonols
Treatment
Year (mg g–1 FW) (mg g–1 FW) (mg g–1 FW) (mg g–1 FW)
Control 12.67 ± 0.05 c 1.371 ± 0.008 a 0.172 ± 0.007 a 1.469 ± 0.006 b
2011 LLR 16.69 ± 0.04 g 3.91 ± 0.04 e 0.173 ± 0.004 a 3.3 ± 0.03 f
ELR 18.79 ± 0.07 i 4.30 ± 0.04 d 0.199 ± 0.009 b 4.14 ± 0.03 h
Control 17.728 ± 0.04 h 4.09 ± 0.02 cd 0.286 ± 0.009 f 1.466 ± 0.003 b
2012 LLR 9.438 ± 0.008 a 1.34 ± 0.01 a 0.206 ± 0.003 bc 3.185 ± 0.005 e
ELR 15.468 ± 0.008 f 3.721 ± 0.007 c 0.252 ± 0.003 e 3.97 ± 0.03 g
Control 11.654 ± 0.001 b 1.363 ± 0.001 a 0.197 ± 0.003 b 1.09 ± 0.01 a
2013 LLR 14.43 ± 0.03 d 2.88 ± 0.01 b 0.217 ± 0.004 cd 2.717 ± 0.004 c
ELR 14.8 ± 0.2 e 2.9 ± 0.6 b 0.22 ± 0.01 d 3.01 ± 0.02 d

Table 4. Effects of leaf removal time and experimental year on the concentrations of total phenols, tartaric acid esters, and anthocyanins
and flavanols in wine (after 6 months of aging) determined by HPLC. LLR, late leaf removal; ELR, early leaf removal. Data are means of
three replicates ± sd. Different letters in a column denote significant differences among treatments according to Tukey’s test following
a two-way ANOVA at α = 0.05.

Experimental Total Phenols Total Tartaric Acid Esters Total Anthocyanins Total Flavonols
Treatment
Year (mg L–1 ) (mg L–1 ) (mg L–1 ) (mg L–1 )
Control 1541 ± 1 c 322.4 ± 0.7 c 11.0 ± 0.2 a 178.6 ± 0.4 b
2011 LLR 1984 ± 1 g 422.8 ± 0.8 h 11.2 ± 0.8 a 291 ± 1 g
ELR 2113 ± 115 h 522.8 ± 0.9 i 17.4 ± 0.4 b 417.1 ± 0.4 i
Control 1147.9 ± 0.2 a 272.7 ± 0.3 a 16.4 ± 0.4 b 147.7 ± 0.4 a
2012 LLR 1648.2 ± 0.9 d 338.7 ± 0.5 e 24.7 ± 0.4 d 278.7 ± 0.5 e
ELR 1802 ± 1 e 381.8 ± 0.9 g 33.2 ± 0.3 e 306 ± 0.4 h
Control 1366 ± 1 b 328.1 ± 0.3 d 21.6 ± 0.5 c 238.4 ± 0.7 c
2013 LLR 1363.8 ± 0.4 b 317.5 ± 0.5 b 21.2 ± 0.4 c 257.1 ± 0.4 d
ELR 1855.4 ± 0.8 f 368.6 ± 0.4 f 21.8 ± 0.2 c 283.2 ± 0.4 f

4
Agronomy 2021, 11, 238 11 of 17

In the present study on Cabernet Sauvignon, both leaf removal times changed the
phenolic composition of the berry skin (Table 3 and Figure 5a) and wine (Table 4 and
Figure 5b) in comparison to the control (no leaf removal). In berry skin, the influence
of statistical interaction of leaf removal x experimental year (Supplementary Table S5)
strongly modified the four analyzed parameters (Table 3). This interaction had the strongest
influence on tartaric acid esters, and the weakest on anthocyanins. The simple main effect
for year (across three leaf removal treatments) resulted in the highest content of skin total
phenols, tartaric acid esters and flavanols in 2011 and the lowest in 2013, while anthocyanins
Agronomy 2021, 11, x FOR PEER REVIEW 12 of 18
were the highest in 2012. The simple main effect of defoliation clearly showed a superior
effect of ELR on the increase of these parameters, while the differences between LLR
and control were inconsistent. On the contrary, leaf removal treatment was the strongest
effect of leaf removal treatment). Early leaf removal further caused an increase of the con-
modifier of the concentrations of total phenols and flavanols, and tartaric acid esters in
tent of total tartaric acid esters in wine by 18% and by 38% relative to the LLR and control
wine (Supplementary
treatment, Table S6).
respectively. Finally, Onlythe
although thecontent
concentrations of total anthocyanins
of total anthocyanins both in berry
in wine largely
skin and on
depended in wine were decisively
the experimental under the strongest
year (Supplementary influence
Table S6), of the characteristics
in the representative 2012 of the
growth
year, season
it also clearly(Supplementary Tables
had the same pattern S5>and
(ELR LLRS6).
> Control, Table 4).

Figure5.5.
Figure Major
Major gradients
gradients (PCA(PCA ordination)
ordination) in the concentrations
in the concentrations of four
of four major typesmajor types of chemical
of chemical
compounds
compounds determined
determined by HPLC
by HPLC in theinberry skin (a)skin
the berry and(a)
theand
winethe(after 6 months
wine (after 6ofmonths
aging) (b)of aging) (b) of
of CabernetSauvignon
Cabernet Sauvignon across
acrossthe three
the treatments:
three treatments:Control (no leaf
Control (noremoval); LLR (late
leaf removal); leaf(late
LLR re- leaf removal)
moval) and ELR (early leaf removal). The values in parenthesis denote the proportion of variance
and ELR (early
represented leaf removal).
by a statistically The values
significant in parenthesis
PCA axis. The angles denote the proportion
and lengths of variance
of the radiating lines represented
by a statistically significant PCA axis. The angles and lengths of the radiating
indicate the direction and strength of relationships of the variables with the ordination scores. lines indicate the
Data matrix: 27 samples collected over three consecutive years (2011–2013), 4 measured
direction and strength of relationships of the variables with the ordination scores. Data matrix: variables,
correlation
27 samples cross-product
collected overmatrix, Euclidean
three distance,
consecutive no rotation.
years Group centroids
(2011–2013), 4 measured (crosses) show correlation
variables,
the average position of a sample within each treatment. Convex hulls (polygons) encompass all
cross-product
the samples in each matrix, Euclidean
treatment. distance,
Different no rotation.
letters denote Group centroids
that biochemical fingerprints(crosses) show the average
of the leaf
position of a sample within each treatment. Convex hulls (polygons) encompass
removal treatments, as characterized by the 4 measured variables, are statistically different accord- all the samples in
ing to the
each MRPP test,
treatment. p < 0.05.letters
Different All thedenote
variables were
that correlated with
biochemical the ordination
fingerprints of thescores by
leaf removal treatments,
>35%.
as characterized by the 4 measured variables, are statistically different according to the MRPP test,
p < 0.05. All the variables were correlated with the ordination scores by >35%.
Light stimulates expression and activity of a number enzymes involved in the syn-
thesis of phenolic compounds, and shading reduces these processes [6]. Although light is
essential for the production of anthocyanins, direct exposure of clusters to sun light is
much more important for the production of flavanols then for the accumulation of antho-
cyanins; therefore, various authors proposed flavanols as some kind of markers for light
exposure and microclimate of clusters [12,28,33–38]. For instance, sun-exposed clusters
had greater berry skin concentrations of total and individual (e.g., myricetin, quercetin
Agronomy 2021, 11, 238 12 of 17

Despite a strong modifying influence of the experimental year on the effect of leaf
removal on skin concentrations of total phenols, tartaric acid esters, flavanols and in
particular on anthocyanins and (Table 3, Supplementary Table S5), a general pattern of
their increase (across all experimental years) is indicated along the gradient from control,
over LLR, to ELR (Figure 5a). Moreover, if these four chemical compounds are used as
“fingerprints” to characterize the effect of treatment, treatments involving defoliation are
significantly different (by higher concentrations) from the control treatment, with exception
of anthocyanins which were more influenced by year and uncorrelated with the main
gradient (along the PCA Axis 1, Figure 5). This is in agreement with Moreno et al. [26] in
Tempranillo grown in semiarid conditions of Western Spain. These authors reported that
early leaf removal did not significantly change the concentration of total anthocyanidins,
but increased the concentrations of flavanols. In the same variety (Tempranillo, two-year
trial), Diago et al. [30] reported that leaf removal also might significantly increase the
accumulation of anthocyanins in berries, but not in every year what is again consistent
with our findings (see Supplementary Tables S5 and S6).
In wine, the concentration of total phenols, total tartaric acid esters and total flavanols
very regularly increased from no leaf removal (control), over late leaf removal, and were the
highest in ELR treatments, in each experimental year (Table 4, Figure 5b). In all treatments
the profiles of these four compounds were statistically different (based on MRPP analysis).
The total phenols content in ELR wine (average across three years 1924 ± 144 mg L−1 ) was
by 15% higher than in LLR wine and about 42% higher than in the control wine. Likewise,
total flavanols in ELR wine (average for three years 335 ± 62 mg L−1 ) were 1.22-fold
higher than in LLR, and 1.79-fold higher compared to the control wine (main effect of leaf
removal treatment). Early leaf removal further caused an increase of the content of total
tartaric acid esters in wine by 18% and by 38% relative to the LLR and control treatment,
respectively. Finally, although the content of total anthocyanins in wine largely depended
on the experimental year (Supplementary Table S6), in the representative 2012 year, it also
clearly had the same pattern (ELR > LLR > Control, Table 4).
Light stimulates expression and activity of a number enzymes involved in the syn-
thesis of phenolic compounds, and shading reduces these processes [6]. Although light is
essential for the production of anthocyanins, direct exposure of clusters to sun light is much
more important for the production of flavanols then for the accumulation of anthocyanins;
therefore, various authors proposed flavanols as some kind of markers for light exposure
and microclimate of clusters [12,28,33–38]. For instance, sun-exposed clusters had greater
berry skin concentrations of total and individual (e.g., myricetin, quercetin and kaempferol)
flavanols than in shaded clusters of Merlot [36]. These findings are in agreement with
our results (Figures 5 and 6), where anthocyanins were not correlated to the statistically
significant PCA axis, while flavanols strongly corresponded to the gradient induced by leaf
removal along the PCA axis 1. According to various authors, the shaded clusters showed
significantly lower anthocyanins concentration than those exposed to sun [7,28,32,39].
Furthermore, the detailed profile of flavanols and antocyanins in response to leaf
removal treatments in berry skin (Table 5, Figure 6a) and in wine (Table 6, Figure 6b) are
shown for the year 2012. This year incidentally happened to be rather typical for the
Eastern Serbia terroir (see Table 1 and Supplemental Table S7). In fact, 2011 was rather dry
and 2013 was cooler and more humid than the average weather conditions of this region.
All the treatments (characterized by 12 phenolic compounds) were statistically significantly
different (MRPP test for 2012 data, skin and wine, not shown). The skin concentrations of
total flavanols measured over three-year experiment were the highest in ELR treatments
(Table 3, Figure 5a), and, consistently, all the six detected flavanol compounds (except
rutin) had the highest concentration in the ELR treatment of 2012 (Table 5, Figure 6a). Like
with total flavanols (Table 3, Figure 5a), the differences in 7 phenolic compounds detected
in skin were not consistent between LLR and control. The accumulation of kaempferol,
quercetin and morin in particular was 2-fold higher in ELR than in the control in 2012
(Table 5). The prominent accumulation of flavanols in the berry skin (catechin, quercetin,
Agronomy 2021, 11, 238 13 of 17

quercetin-glucoside, morine, myricetine, and kaempferol) in the ELR treatment (Table 5,

Figure 6a) is in agreement with findings of Downey et al. [6] in Shiraz, Ćirković et al. [28]
in Prokupac, and Spayd et al. [36] in Merlot. The skin concentrations of total anthocyanins
were the highest in ELR treatments in each experimental year (Table 3, Figure 5a), but their
values highly varied between years (as discussed previously). In the samples from 2012
(Table 5, Figure 6a), five out of six detected anthocyanins had the highest concentration
Agronomy 2021, 11, x FOR PEER REVIEW 13 of 18
in the ELR treatment. Malvidin-3-glucoside was the dominant anthocyanin in the berry
skin of Cabernet Sauvignon in 2012 (Table 5), which is in accordance with the findings
of Ristic et al. [32] in the berry skin of Shiraz. Malvidin-3-glucoside showed the highest
and kaempferol) flavanols than in shaded clusters of Merlot [36]. These findings are in
concentration in ELR treatment which differed significantly from the other two treatments
agreement with our results (Figures 5 and 6), where anthocyanins were not correlated to
(Table 5, Figure 6a). On the other hand, the concentrations of the three anthocyanins
the statistically significant PCA axis, while flavanols strongly corresponded to the gradi-
(peonidin-3-glucoside, petunid-3-glucoside, delphidin-3-glucoside) in berry skin were
ent induced by leaf removal along the PCA axis 1. According to various authors, the
significantly lower in the LLR treatment than in ELR and control treatments. Relative to
shaded clusters showed significantly lower anthocyanins concentration than those ex-
the control, the ELR treatment increased the concentration of anthocyanin compounds by
posed to sun
up to 40% in [7,28,32,39].
2012 (Table 5) and Spayd et al. [36].

Figure 6.
Figure Majorgradients
6. Major gradients(PCA(PCAordination)
ordination)inin
thethe concentrations
concentrations of of
12 12 chemical
chemical compounds
compounds de-
deter-
mined
terminedby HPLC
by HPLCin the
in berry skin skin
the berry (a) and
(a)the
andwine
the (after
wine 6(after
months of aging)
6 months of (b) of Cabernet
aging) Sauvi-
(b) of Cabernet
gnon acrossacross
Sauvignon the three
the treatments: Control
three treatments: (no leaf
Control removal);
(no LLR (late
leaf removal); LLRleaf removal)
(late and ELR
leaf removal) and(early
ELR
leaf removal). The values in parenthesis denote the proportion of variance represented
(early leaf removal). The values in parenthesis denote the proportion of variance represented by by each
statistically significant PCA axis. The angles and lengths of the radiating lines indicate the direc-
each statistically significant PCA axis. The angles and lengths of the radiating lines indicate the
tion and strength of relationships of the variables with the ordination scores. Data matrix: 9 sam-
direction and strength of relationships of the variables with the ordination scores. Data matrix:
ples collected in one representative year (2012), 12 variables, correlation cross-product matrix,
9 samples distance,
Euclidean collectednoin rotation.
one representative
Samples (3year (2012),
per leaf 12 variables,
removal treatment)correlation
are shown, cross-product
together withma-
trix, Euclidean distance, no rotation. Samples (3 per leaf removal treatment) are
group centroids (crosses) which point the average position of a sample within each treatment. shown, together
with group
Chemical centroids (crosses)
compounds which point
are abbreviated as in the average
Tables 5 andposition ofvariables
6. All the a samplewerewithin each treatment.
correlated with
Chemical
the compounds
ordination scores byare>35%.
abbreviated as in Tables 5 and 6. All the variables were correlated with the
ordination scores by >35%.
Furthermore, the detailed profile of flavanols and antocyanins in response to leaf re-
moval treatments in berry skin (Table 5, Figure 6a) and in wine (Table 6, Figure 6b) are
shown for the year 2012. This year incidentally happened to be rather typical for the East-
ern Serbia terroir (see Table 1 and Supplemental Table S7). In fact, 2011 was rather dry and
2013 was cooler and more humid than the average weather conditions of this region. All
Agronomy 2021, 11, 238 14 of 17

Table 5. Effects of leaf removal timing on the concentrations (mg g−1 FW) of different phenolic
compounds (flavanols and anthocyanins) in the grape berry skin determined by HPLC. LLR, late
leaf removal; ELR, early leaf removal. The results from representative year (2012) are shown. Data
are means of three replicates ± SD. Different letters in a row denote significant difference between
parameter means after Tukey’s test following one-way ANOVA (p < 0.05).

Compounds Abbreviation Control LLR ELR

1 20.119 ± 0.002 a b 22.306 ± 0.00 c
Malvidin-3-glucoside MalGluc 20.136 ± 0.004
Peonidin-3-glucoside 1 PeoGluc 2.110 ± 0.00 b 1.563 ± 0.004 a 2.953 ± 0.002 c
Petunidin-3-glucoside 1 PetGluc 2.675 ± 0.003 b 2.1830 ± 0.0008 a 2.842 ± 0.001 c
Cyanidin-3-glucoside 1 CyGluc 0.434 ± 0.002 b 0.235 ± 0.00 a 0.479 ± 0.000 c
Delphinidin-3-glucoside 1 DelGluc 3.411 ± 0.0014 b 2.534 ± 0.00 a 3.507 ± 0.003 c
Quercetin-glucoside 2 QuerGluc 0.326 ± 0.004 b 0.225 ± 0.00 a 0.357 ± 0.004 c
Quercetin 2 QuerTot 0.177 ± 0.004 b 0.156 ± 0.00 a 0.324 ± 0.003 c
Rutin 2 Rutin 0.527 ± 0.002 c 0.266 ± 0.00 a 0.519 ± 0.00 b
Morin 2 Morin 0.08 ± 0.0009 b 0.066 ± 0.00 a 0.116 ± 0.004 c
Myricetin 2 Myric 0.039 ± 0.0005 a 0.034 ± 0.00 a 0.051 ± 0.00 b
Kaempferol 2 Kaemph 0.029 ± 0.0007 a 0.045 ± 0.00 b 0.065 ± 0.003 c
Catechin 2 Cath 5.282 ± 0.008 a 5.438 ± 0.00 b 5.886 ± 0.004 c
1 Anthocyanins; 2 Flavonols.

Table 6. Effects of leaf removal on the concentrations (mg L−1 ) of different phenolic compounds
(flavanols and anthocyanins) in wine (after 6 months of aging) determined by HPLC. LLR, late leaf
removal; ELR, early leaf removal. The results from representative year (2012) are shown. Data
are means of three replicates ± SD. Different letters in a row denote significant difference between
parameter means after Tukey’s test following one-way ANOVA (p < 0.05).

Compounds Abbreviation Control LLR ELR

1 b 787.21 ± 0.04 a 921.40 ± 0.03 c
Malvidin-3-glucoside MalGluc 851.4 ± 0.02
Peonidin-3-glucoside 1 PeoGluc 11.1 ± 0.01 a 18.2 ± 0.05 c 16.4 ± 0.03 b
Petunidin-3-glucoside 1 PetGluc 60.4 ± 0.02 a 67.1 ± 0.02 b 81.5 ± 0.03 c
Cyanidin-3-glucoside 1 CyGluc 10.2 ± 0.02 a 10.9 ± 0.03 b 11.7 ± 0.7 c
Delphinidin-3-glucoside 1 DelGluc 28.4 ± 0.03 a 41.0 ± 0.02 b 52.1 ± 0.02 c
Quercetin-glucoside 2 QuerGluc 19.8 ± 0.04 b 18.2 ± 0.02 a 26.80 ± 0.0 c
Quercetin 2 QuerTot 4.9 ± 0.02 a 11.2 ± 0.02 b 27.1 ± 0.04 c
Rutin 2 Rutin 9.5 ± 0.03 a 15.0 ± 0.03 c 12.6 ± 0.02 b
Morin 2 Morin 3.4 ± 0.04 a 5.7 ± 0.01 b 7.6 ± 0.02 c
Myricetin 2 Myric 1.9 ± 0.03 a 4.5 ± 0.02 b 7.6 ± 0.04 c
Kaempferol 2 Kaemph 3.5 ± 0.02 a 3.6 ± 0.04 a 3.9 ± 0.01 b
Catechin 2 Cath 38.4 ± 0.04 a 42.1 ± 0.03 b 46.2 ± 0.04 c
1 Anthocyanins; 2 Flavonols.

In wine, the concentrations of total flavanols and total anthocyanins measured over
three-year experiment were also the highest in ELR treatments (Table 4, Figure 5b), and,
consistently, all the 11 detected anthocyanin and flavanol compounds (except rutin) were
the highest in the ELR treatment of 2012 (Table 6, Figure 5b). The significant gradient in
increase of total flavanols (Control < LLR < ELR, PCA Axis 1, Figure 6b) was best reflected
in the gradual increase of morin and myricetin in year 2012 (PCA Axis 1 in Figure 6b). This
trend was in fact observed with all the detected flavanols except rutin (Table 6). Across
the three-year period, the ELR treatment increased total phenols by 22% and by 80% in
comparison with the LLR and the control, respectively. In the year 2012, specifically, the
increase of total phenols in ELR were by 10% and by about 200% relative to the LLR and
the control, respectively (Table 4). This increase in 2012 was not uniform for all the detected
flavanols: In was most prominent in quercetin (up to 5-fold) and myricetin (up to 4-fold),
and least in kaempferol (Table 6). Malvidin 3-glucoside was the dominant anthocyanin in
wine, like in berries in 2012 (Table 5). Total anthocyanins in wine were the highest in year
2012 (Table 4), and decisively modified by experimental year (Supplementary Table S6).
So the better effect of ELR on concentration of total anthocyanins in wine (Table 4) was
found for years 2011 and 2012, but not for 2013 (the season with less insolation, Table 1).
This suggests that three year trials might not be sufficient to firmly establish the effect
Agronomy 2021, 11, 238 15 of 17

of leaf removal timing on total anthocyanins in wine. In samples from 2012, four out of
five detected anthocyanins in wine were the highest in the ELR treatments (Table 6). The
strongest increase in ELR treatment was found for delphinidin-3-glucoside (27% relative
to the LLR, and 83% relative to the control. The changes in phenolic profiles shown here
are only illustrative and have no predicative power. For instance, while in one year we
found that ELR increases delfphidine-, malvidine-, cyanindin-, and petunidin-3 glucosides
(compared to ELR and control, Table 6), Ivanisevic et al. reported, on the same variety
Cabernet, in Serbia, also for one year, a respective effect for delfphidin- and cyanindin-, no
effect on petunidin-, and a decrease of malvidine-3 glucosides.
Overall, the general ELR phenolic profile (in the year 2012) was distinguished from the
other treatments primarily by the highest concentrations of total quercetin and myricetin
in both skin and wine, followed by the highest morin in the skin and the highest petunidin-
glucoside in the wine (Figure 6). Interestingly, late defoliation prominently differed from
the other two treatments by the lowest concentrations of rutin, peonidin-, cyanidin-, and
delphinidine-glucosides in skin tissue (Figure 6a), while in the wine this treatment was set
apart by the highest concentrations of rutin and peonidin-glucoside (Figure 6b).

4. Conclusions
Both early and late leaf removal treatments increased the content of sugars in fresh
juice and alcohol in wine, compared to the control. In a three-year field experiment with
pronounced interannual variability of temperature and precipitation, the early timing of
leaf removal (ELR) showed a very clear effect on a decrease of cluster weight, weight
of berries in a cluster and weight of a single berry, and on an increase of skin share in
a berry (compared to LLR and control). Also, ELR resulted in the lowest grape yield.
Early leaf removal furthermore increased the content of total phenols, tartaric acid esters,
anthocyanins and flavanols in both berry skin and wine, with the exception of the total
anthocyanins content in 2013, which was atypically humid. Indeed, ELR treatment showed
irrespectively of very different growth seasons. In the representative year (2012), the ELR
most prominently increased the content of the flavanols, myricetin, and quercetine, in
both berry skin and wine, while with respect to anthocyanins the ELR caused the highest
increase of peonidin-3-glucoside in skin and the highest increase of delphinidin-3-glucoside
in wine.
This work clearly demonstrated the positive impact of early leaf removal on the chemi-
cal composition of berries and wine in a rather warm and arid temperature during veraison
to harvest period (July–September) of continental climate in Eastern Serbia. Therefore,
this measure can be recommended to increase wine quality of Cabernet Sauvignon, but
has to be carefully considered together with the other factors that affect wine quality (e.g.,
climatic and soil conditions, training system, and winemaking technology).

Supplementary Materials: The following are available online at https://www.mdpi.com/2073-439

5/11/2/238/s1, Table S1: Variance partitioning based on a two-way ANOVA with the interaction
term for the vegetative growth and yield components, attributable to treatment factors; Table S2:
Variance partitioning based on a two-way ANOVA with the interaction term for the parameters (total
soluble solids (TSS)) and total acids in fresh grape juice, attributable to treatment factors; Table S3:
Variance partitioning based on a two-way ANOVA with the interaction term for the content of
alcohol, total acids and volatile acids in wine, attributable to treatment factors; Table S4: Variance
partitioning based on a two-way ANOVA with the interaction term for the share of skin (%) in a
fresh berry mass attributable to treatment factors; Table S5: Variance partitioning based on a two-way
ANOVA with the interaction term for the concentration of four major groups of chemical compounds
in berry skin, attributable to treatment factors; Table S6: Variance partitioning based on a two-way
ANOVA with the interaction term for the concentration of four major groups of chemical compounds
in wine, attributable to treatment factors; Table S7: Average temperature (◦ C), precipitation (mm)
and insolation (h) during the bud burst to harvest period (2011–2013).
Agronomy 2021, 11, 238 16 of 17

Author Contributions: Conceptualization, D.S., S.T., and M.N.; methodology, D.S. and N.N.; soft-
ware, N.N.; validation, D.S., M.N., N.N., and L.K.; formal analysis, D.S. and N.N.; investigation, D.S.;
writing—original draft preparation, D.S. and N.N.; writing—review and editing, M.N. and N.N.;
visualization, N.N. and L.K.; supervision, S.T. and M.N.; funding acquisition, M.N. All authors have
read and agreed to the published version of the manuscript.
Funding: This research was supported by the Serbian Ministry of Education, Science and Technologi-
cal Development (Contract No. 451–03-68/2020–14/200053).
Institutional Review Board Statement: Not relevant.
Informed Consent Statement: Not applicable.
Data Availability Statement: No new data were created or analyzed in this study. Data sharing is
not applicable to this article.
Conflicts of Interest: The authors declare no conflict of interest.

References
1. Poni, S.; Gatti, M.; Bernizzoni, F.; Civardi, S.; Bobeica, N.; Magnanini, E.; Palliotti, A. Late leaf removal aimed at delaying ripening
in cv. Sangiovese: Physiological assessment and vine performance. Aust. J. Grape Wine Res. 2013, 19, 378–387. [CrossRef]
2. Bondada, B.; Covarrubias, J.I.; Tessarin, P.; Boliani, A.C.; Marodin, G.; Rombolà, A.D. Post-veraison shoot trimming reduces
cluster compactness without compromising fruit quality attributes in organically-grown Sangiovese grapevines. Am. J. Enol.
Vitic. 2016, 67, 206–211. [CrossRef]
3. Smart, R.E.; Smith, S.M.; Winchester, R.V. Light quality and quantity effects on fruit ripening for Cabernet Sauvignon. Am. J. Enol.
Vitic. 1988, 39, 250–258.
4. Rojas-Lara, B.A.; Morrison, J.C. Differential effects of shading fruit or foliage on the development and composition of grape
berries. Vitis 1989, 28, 199–208.
5. Hunter, J.J.; De Villiers, O.T.; Watts, J.E. The effect of partial defoliation on quality characteristics of Vitis vinifera L. cv. Cabernet
Sauvignon grapes. Skin color, skin sugar and wine quality. Am. J. Enol. Vitic. 1991, 42, 13–18.
6. Downey, M.O.; Harvey, J.S.; Robinson, S.P. The effect of bunch shading on berry development and flavonoid accumulation in
Shiraz grapes. Aust. J. Grape Wine Res. 2004, 10, 55–73. [CrossRef]
7. Chorti, E.; Gudoni, S.; Ferrandino, A.; Novello, V. Effect of different cluster sunlight exposure levels on ripening and anthocyanin
accumulation in Nebbiolo grapes. Am. J. Enol. Vitic. 2010, 61, 23–30.
8. Risco, D.; Perez, D.; Yeves, A.; Castel, J.R.; Intrigliolo, D.S. Early defoliation in a temperate warm and semiarid Tempranillo
vineyard: Vine performance and grape composition. Aust. J. Grape Wine Res. 2014, 20, 111–122. [CrossRef]
9. Sternad Lemut, M.; Sivilotti, P.; Franceschi, P.; Wehrens, R.; Vrhovsek, U. Use of metabolic profiling to study grape skin polyphenol
behavior as a result of canopy microclimate manipulation in a Pinot Noir vineyard. J. Agric. Food Chem. 2013, 61, 8976–8986.
[CrossRef]
10. Dokoozlian, N.K.; Kliewer, W.M. The light environment within grapevine canopies. I. Description and seasonal changes during
fruit development. Am. J. Enol. Vitic. 1995, 46, 209–217.
11. Tarara, J.M.; Lee, J.; Spayd, S.E.; Scagel, C.F. Berry temperature and solar radiation alter acylation, proportion, and concentration
of anthocyanin in Merlot grapes. Am. J. Enol. Vitic. 2008, 59, 235–247.
12. Diago, M.P.; Vilanova, M.; Tardaguila, J. Effects of timing of manual and mechanical early defoliation on the aroma of Vitis vinifera L.
Tempranillo wine. Am. J. Enol. Vitic. 2010, 61, 382–391.
13. Bergqvist, J.; Dokoozlian, N.; Ebisuda, N. Sunlight exposure and temperature effects on berry growth and composition of
Cabernet Sauvignon and Grenache in the central San Joaquin Valley of California. Am. J. Enol. Vitic. 2001, 52, 1–7.
14. German, J.B.; Walzem, R.L. The health benefits of wine. Annu. Rev. Nutr. 2000, 20, 561–593. [CrossRef]
15. Tessarin, P. Effects and Modes of Action of Canopy Management Practices on Vine Physiology and Berry Composition in
Organically-Cultivated cv. Sangiovese (Vitis vinifera L.). Ph.D. Thesis, University of Bologna, Bologna, Italy, 2016.
16. Poni, S.; Bernizzoni, F.; Briola, G. Effects of early leaf removal on cluster morphology, shoot efficiency and grape quality in two
Vitis Vinifera cultivars. Acta Hort. 2005, 689, 217–226. [CrossRef]
17. Intrieri, C.; Filippetti, I.; Allegro, G.; Centinari, M.; Poni, S. Early defoliation (hand vs mechanical) for improved crop control and
grape composition in Sangiovese (Vitis vinifera L.). Aust. J. Grape Wine Res. 2008, 14, 25–32. [CrossRef]
18. Baiano, A.; De Gianni, A.; Previtali, M.A.; Del Nobile, M.A.; Novello, V.; de Palma, L. Effects of defoliation on quality attributes of
Nero di Troia (Vitis vinifera L.) grape and wine. Food Res. Int. 2015, 75, 260–269. [CrossRef]
19. Mijowska, K.; Ochmian, I.; Oszmianski, J. Impact of cluster zone leaf removal on grapes cv. Regent on polyphenol content by the
UPLC-PDA/MS method. Molecules 2016, 21, 1688. [CrossRef]
20. Beck, H.E.; Zimmermann, N.E.; McVicar, T.R.; Vergopolan, N.; Berg, A.; Wood, E.F. Present and future Köppen-Geiger climate
classification maps at 1-km resolution. Sci. Data 2018, 5, 180214. [CrossRef]
Agronomy 2021, 11, 238 17 of 17

21. And̄elković, G.; Živković, N. Precipitation as an adverse climatic phenomenon in Negotin. Bull. Serb. Geogr. Soc. 2007, 87, 51–62.
(In Serbian)
22. Lorenz, D.H.; Eichhorn, K.W.; Bleiholder, H.; Klose, R.; Meier, U.; Weber, E. Growth stages of the grapevine: Phenological growth
stages of the grapevine (Vitis vinifera L. ssp. vinifera). Codes and descriptions according to the extended BBCH scale. Aust. J. Grape Wine
Res. 1995, 1, 100–103. [CrossRef]
23. Mazza, G.; Fukumoto, L.; Delaquis, P.; Girard, B.; Ewert, B. Anthocyanins, phenolics, and color of Cabernet Franc, Merlot and
Pinot Noir wines from British Columbia. J. Agric. Food Chem. 1999, 47, 4009–4017. [CrossRef] [PubMed]
24. Di Stefano, R. Metodi chimici nella caratterizzione varietale. Riv. Vitic. Enol. 1996, 1, 51–56.
25. Mielke, P.W.J.; Berry, K.J. Permutation Methods: A Distance Function Approach; Springer: Berlin, Germany, 2001.
26. Moreno, D.; Vilanova, M.; Gamero, E.; Intrigliolo, D.; Talaverano, I.; Uriarte, D.; Valdes, E. Effects of preflowering leaf removal on
phenolic composition of Tempranillo in the semiarid terroir of Western Spain. Am. J. Enol. Vitic. 2015, 66, 204–211. [CrossRef]
27. Tardaguila, J.; Diago, M.P.; de Toda, F.M.; Poni, S.; Vilanova, M. Effects of timing of leaf removal on yield, berry maturity, wine
composition and sensory properties of cv. Grenache grown under non irrigated conditions. J. Int. Sci. Vigne Vin 2008, 42, 221–229.
[CrossRef]
28. Ćirković, D.; Matijašević, S.; Deletić, N.; Ćirković, B.; Gašić, U.; Sredojević, M.; Jovanović, Z.; Djurić, V.; Tešić, Ž. The effect of
early and late defoliation on phenolic composition and antioxidant properties of Prokupac variety grape berries (Vitis vinifera L.).
Agronomy 2019, 9, 822. [CrossRef]
29. Bledsoe, A.M.; Kliewer, W.M.; Marois, J.J. Effects of timing and severity of leaf removal on yield and fruit composition of
Sauvignon Blanc grapevines. Am. J. Enol. Vitic. 1988, 39, 49–54.
30. Diago, M.P.; Ayestarán, B.; Guadalupe, Z.; Poni, S.; Tardáguila, J. Impact of prebloom and fruit set basal leaf removal on the
flavonol and anthocyanin composition of Tempranillo grapes. Am. J. Enol. Vitic. 2012, 63, 367–376. [CrossRef]
31. Ivanišević, D.; Kalajdžić, M.; Drenjančević, M.; Puškaš, V.; Korać, N. The impact of cluster thinning and leaf removal timing on
the grape quality and concentration of monomeric anthocyanins in Cabernet-Sauvignon and Probus (Vitis vinifera L.) wines.
OENO One 2020, 1, 63–74. [CrossRef]
32. Ristic, R.; Downey, M.O.; Iland, P.G.; Bindon, K.; Francis, I.L.; Herderich, M.J.; Robinson, S.P. Exclusion on sunlight from Shiraz
grapes alters wine colour, tanninand sensory properties. Aust. J. Grape Wine Res. 2007, 13, 53–65. [CrossRef]
33. Coombe, B.G. Research on development and ripening of the grape berry. Am. J. Enol. Vitic. 1992, 43, 101–110.
34. Poni, S.; Casalini, L.; Bernizzoni, F.; Civardi, S.; Intrieri, C. Effect of early defoliation on shoot photosynthesis, yield components
and grape composition. Am. J. Enol. Vitic. 2006, 57, 397–407.
35. Crippen, D.D.; Morrison, J.C. The effects of sun exposure on the compositional development of Cabernet Sauvignon berries. Am.
J. Enol. Vitic. 1986, 37, 235–242.
36. Spayd, S.E.; Tarara, J.M.; Mee, D.L.; Ferguson, J.C. Separation of sunlight and temperature effects on the composition of
Vitis vinifera cv. Merlot berries. Am. J. Enol. Vitic. 2002, 53, 171–182.
37. Keller, M.; Hrazdina, G. Interaction of nitrogen availability during bloom and light intensity during veraison: Effects on
anthocyanin and phenolic development during grape ripening. Am. J. Enol. Vitic. 1998, 49, 341–349.
38. Pereira, G.E.; Gaudillere, J.P.; Pieri, P.; Hilbert, G.; Maucourt, M.; Deborde, C.; Moing, A.; Rolin, D.l. Microclimate influence on
mineral and metabolic profiles of grape berries. J. Agric. Food Chem. 2006, 54, 6765–6775. [CrossRef]
39. Morrison, J.C.; Noble, A.C. The effects of leaf and cluster shading on the composition of Cabernet Sauvignon grapes and on fruit
and wine sensory properties. Am. J. Enol. Vitic. 1990, 41, 193–200.

View publication stats