Early Leaf Removal has a Larger Effect than Cluster Thinning on Grape Phenolic Composition in cv. Teran

Transcription

1 Early Leaf Removal has a Larger Effect than Cluster Thinning on Grape Phenolic Composition in cv. Teran Marijan Bubola, 1 Paolo Sivilotti, 2 Danijela Janjanin, 1 and Stefano Poni 3 * Abstract: This study aimed to assess the effects of early leaf removal (ELR) and cluster thinning (CT) on vegetative growth, yield components, and berry composition of Teran (Vitis vinifera L.) grown in the Istria region of Croatia. Both treatments were compared with an untreated control (UC) during the 2012 and 2013 growing seasons. ELR involved the removal of six basal leaves at the preflowering stage, while in CT, 35% of clusters were removed at the onset of veraison. Both ELR and CT resulted in lower yield per vine compared to UC (22 and 37% less, respectively), and in greater Brix at harvest (+1.3 and +1.0, respectively). The concentration of total phenolics in grapes increased by 19% in ELR and by 6% in CT compared to UC, while total anthocyanin concentrations increased only in ELR (+20% versus UC). Despite greater yield and a lower leaf area-to-yield ratio compared to CT, ELR achieved higher total anthocyanin and phenolic concentration, suggesting that ELR is more suitable than CT for the production of high-quality Teran grapes. Key words: anthocyanins, canopy management, fruit exposure, leaf area, phenolic compounds, yield Canopy and crop management practices are of special concern for grapevine varieties such as Vitis vinifera L. cv. Teran that have high vigor and a tendency to overcrop. Teran, a variety from the northern Adriatic area, is mostly grown in the Istria region in Croatia (Maletić et al. 2015), the Carso region in Italy, where it is called Terrano, and the Primorska region in Slovenia, where it is called Refošk (Rusjan et al. 2015). Among the canopy management practices, traditional leaf removal, applied between fruit set and preharvest, is widely used to improve the fruit zone microclimate and grape composition in heavily shaded canopies (Kemp et al. 2011, Bubola et al. 2012, Lee and Skinkis 2013). Shoot and/or cluster thinning are primarily used to reduce yield (Keller et al. 2005, Silvestroni et al. 2016), and shoot trimming is usually performed to remove excess foliage and constrain canopy size (Šuklje et al. 2013, Herrera et al. 2015). Early leaf removal, performed around flowering, has been widely investigated over the last decade as a potential way 1 Institute of Agriculture and Tourism, Karla Huguesa 8, Poreč, Croatia; 2 Dipartimento di Scienze Agroalimentari, Ambientali e Animali, Università degli Studi di Udine, via delle Scienze 206, Udine, Italy; and 3 Dipartimento di Scienze delle Produzioni Vegetali Sostenibili, Area Frutticoltura e Viticoltura, Università Cattolica del Sacro Cuore, via Emilia Parmense 84, Piacenza, Italy. *Corresponding author (stefano.poni@unicatt.it; tel: ; fax: ) Acknowledgments: This research was partially supported by the Ministry of Science, Education and Sports, Republic of Croatia, under the project Valorization of grapevine resources and gene bank, and by the Croatian Council for Agricultural Research under the project Application of early defoliation in order to increase the quality of grapes and wine. Moreno Coronica is acknowledged for the use of his commercial vineyard and for technical assistance during the execution of this trial. Manuscript submitted July 2016, revised Oct 2016, accepted Oct 2016 Copyright 2017 by the American Society for Enology and Viticulture. All rights reserved. doi: /ajev to regulate crop levels and the fruit zone microclimate (Poni et al. 2006, Tardaguila et al. 2010, Bubola et al. 2012, Lee and Skinkis 2013, Cook et al. 2015, Komm and Moyer 2015, Sternad Lemut et al. 2015, Silvestroni et al. 2016, Sivilotti et al. 2016). Since carbohydrate supply at flowering is a primary determinant of fruit set (Coombe 1962), the removal of basal and fully functional leaves at this stage results in reduced fruit set and/or final berry size (Tardaguila et al. 2010, Gatti et al. 2012, Palliotti et al. 2012, Risco et al. 2014). Early leaf removal is especially appropriate for high-yielding vineyards and vines with large, compact clusters (Palliotti et al. 2011), and it reduces the sensitivity of clusters to Botrytis cinerea rot (Sternad Lemut et al. 2015). Early leaf removal also affects grape composition and usually leads to higher Brix and greater concentrations of total anthocyanins and phenolics (Diago et al. 2012a, Palliotti et al. 2012, Lee and Skinkis 2013, Cook et al. 2015, Silvestroni et al. 2016). Major factors leading to improved grape composition in vines subjected to early leaf removal include better exposure of grapes to sunlight, lower yield, adequate final leaf area-to-yield ratio, better photosynthetic performance of the remaining leaves, and increased skin-to-berry or flesh ratio in berries (Poni et al. 2006, Palliotti et al. 2011, Diago et al. 2012a). Cluster thinning is a crop management practice that is largely used to improve grape and wine quality in varieties that tend to overcrop, by reducing yield and increasing the leaf area-to-yield ratio. Several studies have confirmed positive effects of cluster thinning on grape and wine quality (Guidoni et al. 2002, Bubola et al. 2011, Gatti et al. 2012, Avizcuri-Inac et al. 2013); however, others have reported that cluster thinning had little or no effect on grape composition (Keller et al. 2005, Nuzzo and Matthews 2006). It is generally assumed that cluster thinning improves grape quality only in cases of true overcropping as gauged by low leaf area-to fruit ratios (Keller et al. 2005, Nuzzo and Matthews 2006). However, some growers prefer to reduce crop size even when the leaf area-to-yield ratio is above the threshold required 234

2 Leaf Removal and Cluster Thinning in Teran 235 for maximum Brix, berry weight, and berry color at harvest (Kliewer and Dokoozlian 2005). This unnecessary crop reduction might result in either no substantial change in grape composition or, when absolute crop size is severely reduced, in a faster ripening, which is considered undesirable under global warming scenarios (Herrera et al. 2015). Although cluster thinning may have a positive impact on grape quality, it does not significantly change the fruit-zone microclimate and it may be less effective than early leaf removal in high-vigor varieties. The goal of this study was to assess, over two seasons, the effects of preflowering leaf removal and cluster thinning applied at veraison on vegetative growth, yield components, and grape composition of fieldgrown Teran grapevines. Materials and Methods Vineyard site and experimental design. The experiment was conducted in 2012 and 2013 in a commercial, nonirrigated vineyard located near Koreniki (45 24 N; E; 60 m asl), in the Istria winegrowing region of Croatia. Teran grapevines (clone ISV-F2) grafted on 420A rootstock were planted in 2004 in a luvic, anthropogenized, deep terra rossa soil, with 2% north-west oriented slope. Vines were planted with a spacing of 0.8 m within rows and 2.0 m between rows (plant density of 6250 vines/ha), with rows oriented north-south. Vines were trained to a vertical shoot-positioned, single-canepruned Guyot trellis with a budload of ~10 nodes per vine. The cane was horizontally tied to a basal wire positioned 0.7 m from the ground. Three sets of catch wires were positioned at 25, 60, and 95 cm above the basal wire, for a canopy wall extending ~1.3 m. Shoot thinning was performed at grapevine growth stage 14 according to the modified E-L system (Coombe 1995), to leave one shoot per node and to remove stunted shoots. Because of strong shoot growth, mechanical shoot trimming was performed twice in both years: at berry set (on 9 June 2012 and 11 June 2013), and three weeks thereafter (on 1 July 2012 and 2 July 2013). At the first shoottrimming event, the top 50 to 60 cm of canopy was removed, and both sides of the row were hedged to maintain maximum canopy width within ~40 cm. The second trimming was performed to remove excessively vigorous laterals from the main apical nodes. Meteorological data were recorded by a Spectrum WD 1650 weather station (Spectrum Technologies) located in the vineyard. Vine phenological stages were recorded at weekly intervals according to the modified E-L system (Coombe 1995). Three adjacent rows were selected to build a randomized complete block design, with each row as a block. Within each row, three sections of three post spaces (21 vines per plot, 16.8-m row length) were tagged and randomly assigned to the following treatments: early (preflowering) leaf removal (ELR), cluster thinning (CT), and a nondefoliated, nonthinned control (UC). Three post spaces at the beginning of each row were excluded from the experiment and used as a buffer. ELR was carried out before flowering at E-L stage 18 (flower caps still in place but cap color fading from green), by manually removing from all shoots the primary leaves developed from nodes 1 to 6. Laterals that had burst at the time of leaf removal were maintained. CT was applied at stage 35 (berries beginning to color and enlarge), and ~35% of clusters were removed. Distal clusters and those developed on weak shoots were selected for removal first. ELR treatments were performed on 26 May 2012 and 29 May 2013, and CT was performed on 31 July 2012 and 2 Aug The treatments were applied on the same vines in both seasons. Leaf area measurement and yield components. Eight representative shoots per replicate were collected at harvest and brought to the laboratory in plastic bags. The primary and lateral leaf area of each sample was assessed using a method based on the fresh weight of leaf blades and the disc technique (Smart and Robinson 1991). All primary and lateral leaves of each sample were separately removed (without petioles) and weighed, and 100 discs (18-mm diam) were randomly cut from these leaves and weighed. The weight of the discs was compared with the weight of primary and lateral leaves to calculate primary and lateral leaf area per shoot. Shoot leaf area was multiplied by the number of shoots per vine to calculate total vine leaf area. To estimate total pretreatment leaf area and leaf area removed in the ELR treatment, eight shoots were taken to the laboratory when defoliation was performed, and the method described above was used. Point quadrat analysis (PQA; Smart and Robinson 1991) was performed at harvest as described for the Croatian sites in Diago et al. (2016) in order to determine canopy gaps and leaf layer number (LLN) in the fruit zone. All vines were harvested by hand on 18 Sept 2012 and 30 Sept Yield and number of clusters per vine were recorded, and average cluster weight was calculated. The number of clusters per shoot was determined as the clusters-to-shoot ratio (vine basis). Likewise, berry number was calculated from cluster and single berry weights, while cluster weight was reduced by 4.4% to take into account average rachis incidence as previously determined for the Teran variety (Vivoda 1996). Berry sampling and fruit composition. At harvest, portions of grape clusters (either wings or cluster tips, including ~10 berries) were randomly sampled with scissors to represent different canopy positions and within single clusters, until ~1 kg per treatment replicate was obtained. A sample from each replicate consisted of ~50 portions of clusters, taken from an equivalent number of clusters. Samples were transported to the laboratory within 2 hr of harvest. The weight of each sample was determined and added to the yield of the respective plot. A subsample of 200 berries was cut at the pedicel for determination of average berry weight and total anthocyanins and phenolics. A subset of 30 berries was collected for the determination of single berry components, and another subset of 20 berries was collected for the analysis of individual anthocyanins. All berry samples were weighed and immediately stored at -20 C. The remaining berries (~250) were manually pressed at room temperature, and the juice was used to measure Brix, ph, and titratable acidity (TA). Brix was determined using an HR200 digital refractometer

3 236 Bubola et al. (APT Instruments), ph was determined using an MP220 phmeter (Mettler Toledo), and TA (expressed as g/l tartaric acid equivalents) was measured by titration with NaOH 0.1N as recommended by the International Organization of Vine and Wine (OIV 2012). phenolic substances in berries were determined according to Iland et al. (2004) and expressed as mg/g berry fresh weight and mg/berry. To determine berry components, thawed berries were sliced in half with a razor blade, seeds and flesh were carefully removed from each berry half with a small metal spatula (without rupturing pigmented hypodermal cells), and the seeds were carefully separated by hand from the flesh. Skins and seeds were rinsed in deionized water, blotted dry, and weighed. Extraction procedure and determination of anthocyanin profile. The anthocyanin composition of the grape berry skin was determined according to Mattivi et al. (2006). For each sample, a set of 20 frozen berries was selected, peeled, and extracted for 24 hr in 100 ml of methanol (Sigma). The liquid phase was then separated from the skins and 50 ml of methanol was added for a further 2-hr extraction. Both methanolic extracts were then combined and stored at -20 C until analysis. The extracts were filtered through a 0.45-µm syringe filter (Chromafil Xtra, Macherey-Nagel), diluted with 1% trifluoroacetic acid (TFA; Sigma) in water (1:9 ratio), and transferred directly into high-performance liquid chromatography (HPLC) vials. The following anthocyanins were separated and quantified: delphinidin-3-glucoside (del-3- glu), cyanidin-3-glucoside (cya-3-glu), petunidin-3-glucoside (pet-3-glu), peonidin-3-glucoside (peo-3-glu), and malvidin- 3-glucoside (mal-3-glu), using gradient HPLC with UV-vis detection at 520 nm. The analysis was carried out with a Waters chromatographic system consisting of two Waters 510 pumps, a Waters 717+ autosampler, and a Waters 2487 UV-vis dual wavelength detector. Individual anthocyanins were separated using a Phenomenex Luna C18, 4.6 mm 150 mm, 5-µm column under the following chromatographic conditions. Mobile phase A consisted of methanol (Sigma) with 0.2% TFA, and mobile phase B contained water with 0.2% TFA (Sigma). The gradient of mobile phase B increased from 40 to 67% in 20 min. The percentage of mobile phase B was increased to 75% in the next 8 min, held for another 12 min, increased to 100% in 1 min, and held for 3 min; the initial gradient condition was then reached in 1 min and the equilibration lasted for 12 min. A flow rate of 0.9 ml/min was used throughout the gradient, and the injection volume was 10 µl. All analyses were carried out in duplicate. A commercially available standard of mal-3-glu (Extrasynthase, Genay Cedex) was separately dissolved in methanol and used as standard stock solution, diluted with 1% TFA (Sigma) in water, to generate calibration curves. Disubstituted (3 4 -OH; cya-3-glu + peo-3-glu), trisubstituted ( OH; del-3-glu + pet-3-glu + mal-3-glu), hydroxylated (OH; cya-3-glu + del-3-glu) and methoxylated anthocyanins (OCH3; pet-3-glu + peo-3-glu + mal-3-glu) were grouped together. The anthocyanin indices F3 5 /F3, 3 -OMT, and 5 -OMT were calculated according to Mattivi et al. (2006). In detail, F3 5 /F3 = (del-3-glu + pet-3- glu + mal-3-glu)/(cya-3-glu + peo-3-glu), 3 -OMT = peo-3-glu/ cya-3-glu, and 5 -OMT = mal-3-glu/del-3-glu. Statistical analysis. Data were processed in SAS (SAS Institute, Inc.) using a two-way mixed-model ANOVA, with year as a random factor and treatments as a fixed factor. When differences among treatments were significant, Fisher s least significant difference test (p 0.05) was used to separate the means. Results Meteorological conditions and grapevine phenological stages. Growing degree days, calculated on a base temperature of 10 C from 1 April to 30 Sept, were 1925 C in 2012 and 1819 C in 2013 (Figure 1). The higher value in 2012 was a result of higher temperatures from June to August, when rainfall was also very low. rainfall from April to September was 311 mm in 2012 and 350 mm in Based on the differences in meteorological conditions between the two seasons, budburst was recorded on 30 March 2012 and 10 April 2013, and anthesis (50% cap fall) occurred on 2 June 2012 and 4 June The onset of veraison was recorded on 29 July 2012 and 1 Aug 2013; grapes reached maturity and were harvested on 18 Sept 2012 and 30 Sept Leaf area and yield components. At the time of preflowering leaf removal, the average leaf area on ELR vines over the 2-yr period was 2129 cm 2 per shoot, and pulling the first six main leaves reduced actual leaf area by 84% (Table 1). At Figure 1 Monthly values for (A) growing degree days (GDD) and (B) rainfall recorded at the experimental site in 2012 (empty circles and bars) and 2013 (filled circles and bars).

4 Leaf Removal and Cluster Thinning in Teran 237 harvest, primary leaf area per shoot was lowest in ELR as a direct consequence of leaf removal. Lateral leaf area per shoot was slightly enhanced in ELR compared to UC (16%), yet final total leaf area was still lower in ELR than in the UC or CT treatments. ELR showed a greater percentage of canopy gaps in the fruit zone compared to the other two treatments (Table 1), and conversely, ELR had the lowest leaf layer number in the fruit zone. Berry weight was lower in ELR than in the UC or CT treatments (-10% and -15% as compared to UC and CT, respectively), whereas differences in berries per cluster were not significant (Table 2). Cluster weight was lowest in ELR (-23% compared to UC and CT), while cluster number per shoot and vine was, as expected, lowest in CT. Both ELR and CT showed lower yield per vine compared to UC (22% and 37% less, respectively). On a per-hectare basis, yield was reduced by 2.6 tons in ELR and by 4.3 tons in CT. The total leaf area-to-fruit ratio was higher in CT than in UC or ELR. The ratio was similar in the UC and ELR treatments, since yield and final leaf area were reduced quite proportionally. Berry tissue components. Relative skin weight on a perberry basis was significantly higher in ELR than in UC and CT, and relative flesh weight was correspondingly lower in ELR (Table 3). Relative seed weight on a per-berry basis and number of seeds per berry did not differ among the treatments. Table 1 Vegetative growth and canopy characteristics in Teran grapevines subjected to early leaf removal and cluster thinning (mean values for 2012 and 2013). LA, leaf area; LLN, leaf layer number; UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. before leaf removal (cm 2 ) Removed (cm 2 ) Fraction of removed LA (%) Primary (cm 2 ) Lateral (cm 2 ) (cm 2 ) LA/vine (m 2 ) Canopy gaps in fruit zone (%) LLN in fruit zone UC 2385 a a a 2.40 a 4 c 2.26 a ELR c c 1.97 b 16 a 1.29 c CT 1922 b b 2.08 b 9 b 1.86 b Signif. F b ** ns ** * ** ** T Y c ns ns ns ns ns ns a Different letters within columns identify significantly different means. b Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant c T Y = treatment year interaction. Table 2 Yield components and leaf-to-fruit ratio (vine basis) in Teran grapevines subjected to early leaf removal and cluster thinning (mean values for 2012 and 2013). LA, leaf area; UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. Berry wt Berries/ cluster Cluster wt Clusters/ shoot Yield/shoot Shoots/ vine Clusters/ vine Yield/vine (kg) LA/yield (m 2 /kg) UC 1.96 a a a 1.40 a 308 a a 1.85 a 1.29 b ELR 1.76 b b 1.37 a 235 b a 1.44 b 1.38 b CT 2.06 a a 0.91 b 201 c b 1.16 c 1.82 a Signif. F b * ns * ** ** ns ** ** ** T Y c ns ns ns ns ns ns ns ns ns a Different letters within columns identify significantly different means. b Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant c T Y = treatment year interaction. Table 3 Weight of berry components (absolute and relative basis) recorded in Teran grapevines subjected to early leaf removal and cluster thinning (mean values for 2012 and 2013). UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. Flesh wt Skin wt seed wt Seeds/ berry Skin-to-berry ratio (%) Seed-to-berry ratio (%) Flesh-to-berry ratio (%) Skin-to-flesh ratio (%) UC a a a b a 10.2 b ELR b c a b 11.1 a CT a b b a 10.3 b Signif. F b * ns * ns ** ns * ** T Y c ns ns ns ns ns ns ns ns a Different letters within columns identify significantly different means. b Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant c T Y = treatment year interaction.

5 238 Bubola et al. Berry composition. At harvest, Brix was higher in ELR and CT than in UC (+1.3 and +1.0 Brix, respectively, over the two-year period); TA was similar among treatments, and ph was significantly higher in CT than in UC and ELR (Table 4). The concentration of total phenolics was affected by treatments showing a 6% increase in CT up to 19% in ELR compared to UC. However, such differences disappeared when total phenolics were given on per-berry basis (Table 4). anthocyanin concentrations (mg/g of berry or skin mass) showed an increase in ELR (+20%), while values nearly comparable to UC were measured in CT (Table 5). As seen for total phenolics, the content of anthocyanins on a per-berry basis was not significantly affected by treatments. Most differences in anthocyanins occurred in the ELR treatment (Table 5). The concentration of all anthocyanins was higher in ELR berries, and the differences were significant for peo-3-glu, mal-3-glu, and p-coumaroyl/caffeoyl derivatives. Most anthocyanin concentrations were similar in CT and UC, with the exception of higher values for peo-3-glu (similar to ELR), and lower values for p-coumaroyl/caffeoyl derivatives in CT. When anthocyanins were given on a per-berry basis, the differences among treatments in most cases were not significant, with the exception of higher peo-3-glu values in CT. No significant differences among treatments were found for the indices F3 5 /F3, 3 -OMT, and 5 -OMT, indicating that even if the concentration of anthocyanins was increased in ELR, the relative percentages of di- and tri-substituted anthocyanins, and of methoxylation of OH-derivatives, did not change among treatments. Summed by type, ELR showed significantly higher values of 3 4 -OH, OH, and OCH 3 anthocyanins compared to UC (Table 6), and similar values of 3 4 -OH anthocyanins occurred in ELR and CT. Even if not significant, a higher F3 5 /F3 ratio and a lower methoxylation (both 3 -OMT and 5 -OMT) occurred in ELR. Similar methoxylation was observed in CT and UC, with a higher value of 3 4 -OH anthocyanins in CT. Discussion Canopy management practices are especially important in vigorous varieties to adequately manage fruit zone microclimate if high quality grapes are to be produced. In addition to high vigor, which can lead to excessive shoot development Table 4 Berry composition in Teran grapevines subjected to early leaf removal and cluster thinning (mean values for 2012 and 2013). UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. soluble solids (Brix) Titratable acidity (g/l) ph anthocyanins (mg/g berry) anthocyanins (mg/berry) anthocyanins (mg/g skin) phenolics (mg/g) phenolics (mg/berry) UC 21.5 b a b 1.45 b b 2.71 c 5.31 ELR 22.8 a b 1.74 a a 3.23 a 5.71 CT 22.5 a a 1.48 b b 2.88 b 5.97 Signif. F b * ns ** * ns * ** ns T Y c ns ns ns ns ns ns ns ns a Different letters within columns identify significantly different means. b Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant c T Y = treatment year interaction. Table 5 Anthocyanin profile at harvest as concentration (mg/kg) and content (μg/berry) in Teran grapes subjected to early leaf removal and cluster thinning (mean values for 2012 and 2013). UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. Del-3-glu a Cya-3-glu a Pet-3-glu a Peo-3-glu a Mal-3-glu a Acetyl p-coum + caff a Concentration (mg/kg) UC b b 659 b b ELR a 763 a a CT a 664 b c Signif. F c ns ns ns * * ns ** T Y d ns ns ns ns ns ns ns Content (μg/berry) UC b ELR b CT a Signif. F ns ns ns * ns ns ns T Y ns ns ns ns ns ns ns a Del-3-glu, delphinidin-3-glucoside; Cya-3-glu, cyanidin-3-glucoside; Pet-3-glu, petunidin-3-glucoside; Peo-3-glu, peonidin-3-glucoside; Mal- 3-glu, malvidin-3-glucoside; p-coum + caff, p-coumaroyl/caffeoyl. b Different letters within columns identify significantly different means. c Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant d T Y = treatment year interaction.

6 Leaf Removal and Cluster Thinning in Teran 239 and canopy shading if shoot trimming and leaf removal are not performed on time, Teran also has large leaves (Rusjan et al. 2015), which increases the tendency to develop a heavily shaded canopy. Because of the large leaves, the leaf area removed in this 2-yr study (1782 cm 2 per shoot, or 84% of the present leaf area) was greater than that in similar studies performed on other varieties such as Sangiovese (Poni et al. 2006), Barbera (Poni et al. 2009), and Tempranillo (Diago et al. 2012a). When ELR is applied, the regrowth of laterals is usually stimulated to compensate for the reduction in primary leaf area, as seen in Sangiovese and Trebbiano (Poni et al. 2006) and Tempranillo (Diago et al. 2012a). In our study, lateral regrowth in ELR was only slightly higher than in UC and did not lead to significant compensation for removed leaf area of primary leaves, resulting in a lower total leaf area per shoot in the ELR treatment (Table 1). Canopy gaps and LLN in the fruit zone, assessed at harvest by PQA, indeed revealed that ELR vines had a more open canopy than the other two treatments. However, the average value of LLN recorded in ELR (i.e., 1.29) indicated that some leaf cover was present around the clusters of this treatment, confirming that the ELR technique does allow seasonal compensation in terms of moderate leaf shading around the fruiting area. In this study, the prebloom removal of more than 80% of shoot leaf area induced severe stress in ELR vines, due to a reduction of available assimilates (Coombe 1962). Similar to previous studies (Intrieri et al. 2008, Tardaguila et al. 2010, Palliotti et al. 2011, Risco et al. 2014, Silvestroni et al. 2016), the lack of carbohydrate supply resulted in a lower cluster weight which, however, was mostly contributed by smaller berries. Although we did not determine berry set (i.e., the ratio of final berry to flower numbers), the lack of statistical differences in berries per cluster suggests that berry set is not affected by ELR in Teran, despite severe source limitation, as shown in other studies (Gatti et al. 2015, Komm and Moyer 2015). To explain such variability, a parameter that should be considered is also the amount of reserve storage that can be mobilized at leaf removal to buffer the abrupt source limitation (i.e., different depending on the presence of a renewed cane or permanent cordon, or as a function of vine age). Some studies of ELR found that defoliated treatments had significantly higher final leaf area-to-yield ratios than control treatments (Palliotti et al. 2011, Gatti et al. 2015); this was not the case with Teran, which showed an unchanged final leaf area-to-yield ratio versus UC. This response was due to a combination of an overall mild yield limitation (-23% versus UC), and to the moderate post-leaf removal lateral regrowth. Notably, in CT, the intensity of thinning (-35%, from cluster/shoot values in Table 2) mirrored the actual per-vine yield reduction at harvest (-37% versus UC), indicating that cluster thinning did not trigger significant yield compensation through higher berry weight in the remaining clusters. This behavior was shown in previous studies in which CT was performed late in the season to leave little time for compensation, and therefore to induce a significant and permanent increase in the leaf area-to-fruit ratio (Guidoni et al. 2002, Keller et al. 2005, Nuzzo and Matthews 2006, Gatti et al. 2012, Avizcuri-Inac et al. 2013). Higher skin-to-berry ratios are quite common reactions of berries to early defoliation treatments (Diago et al. 2012a, Palliotti et al. 2012, Silvestroni et al. 2016). In this study, the skin-to-berry ratio was higher in ELR than in UC and CT (+8% and +7%, respectively; Table 3). However, the magnitude of this effect was smaller than that observed by others (Poni et al. 2009, Gatti et al. 2015), who found a skin-to-berry ratio that was up to 20% greater in early defoliated vines compared to nondefoliated controls. In the present experiment, Brix, total phenolics, and total anthocyanins increased in both ELR and CT, but to a greater extent in ELR (Tables 4 and 5). These findings are consistent with previous studies that showed improved grape ripening that varied in magnitude according to the leaf area-to-yield ratio, growing conditions, and variety (Diago et al. 2012a, Palliotti et al. 2012, Lee and Skinkis 2013, Cook et al. 2015, Silvestroni et al. 2016). Poni et al. (2006) showed a significant increase in anthocyanins and total phenolics in the case of ELR in Sangiovese, and speculated that these increases could be related to an improved leaf area-to-yield ratio. Tardaguila et al. (2010) also showed similar increases in anthocyanins and phenolics in Graciano and Carignan grapes, although in Carignan, the leaf area-to-yield ratio did not increase in the preflowering leaf removal treatment. ELR also has differential effects on various classes of phenolic compounds. Risco et al. (2014) showed that anthocyanins increased but that tannins were mostly unaffected in Tempranillo vines subjected to ELR treatments. Moreover, Sternad Lemut et al. (2013) Table 6 Indices of anthocyanins in berries of Teran grapevines subjected to early leaf removal and cluster thinning. Data are mean values over two years (2012 and 2013). UC, control (undefoliated and unthinned); ELR, early leaf removal; CT, cluster thinning. F3 5 /F3 a 3 -OMT a 5 -OMT a (mg/kg) 3 4 -OH OH (mg/kg) OH (mg/kg) OCH 3 (mg/kg) UC b b 948 b b ELR a 1167 a a CT a 957 b b Signif. F c ns ns ns * * ns * T Y d ns ns ns ns ns ns ns a Indices F3 5 /F3, 3 -OMT, and 5 -OMT are unitless. b Different letters within columns identify significantly different means. c Data were analyzed using two-way mixed model ANOVA; when differences were significant, means were separated using Fisher s least significant d T Y = treatment year interaction.

7 240 Bubola et al. reported that ELR primarily affected flavonols and anthocyanins, and to a much lesser extent, all other classes of phenolic compounds. Meteorological conditions can sometimes also be responsible for masking the effects of preflowering leaf removal. Diago et al. (2012b) found that increases in phenolics and anthocyanins in Tempranillo grapes that underwent ELR were significant only in some seasons. Conversely, Sivilotti et al. (2016) showed no effects of preflowering leaf removal on anthocyanins or tannins in Merlot grapes during maturation and at harvest; though, the extraction of tannins during winemaking was higher in the case of ELR, as also reported for Pinot noir by Kemp et al. (2011). Sivilotti et al. (2016) reported that the leaf area-to-yield ratio in Merlot in all treatments was well above the optimal range as proposed by Kliewer and Dokoozlian (2005). However, Pastore et al. (2013), Šuklje et al. (2013), and Herrera et al. (2015) argued that even under a situation of source-sink equilibrium, uncoupling of primary and secondary metabolic processes may occur, meaning that while sugar accumulation remains unaffected, biosynthesis of flavonoids or other secondary metabolites could be significantly affected by canopy management. In this study, CT accounted for a slight increase in the concentration of phenolics but had negligible effects on anthocyanin concentrations. Other studies have shown a positive effect of CT on the increase of anthocyanins and phenolics (Guidoni et al. 2002, Bubola et al. 2011, Pastore et al. 2011, Gatti et al. 2012, Avizcuri-Inac et al. 2013). However, in other trials, differences between thinned plants and unthinned controls were inconsistent (Keller et al. 2005, Nuzzo and Matthews 2006). Clearly, in different grapevine varieties and growing conditions, the effects of CT are variable and mainly related to the leaf area-to-yield ratio. When vines are in balance in terms of leaf area-to-yield ratio (Kliewer and Dokoozlian 2005), further increases in leaf area-to-yield ratio caused by canopy management may not lead to increased Brix but still may trigger reprogramming of secondary metabolism (Pastore et al. 2011, Šuklje et al. 2013, Herrera et al. 2015). The anthocyanin profile in this study showed a high increase of mal-3-glu, peo-3-glu, and p-coumaroyl + caffeoyl derivatives in ELR, a significant increase of peo-3-glu in CT, and a decrease of p-coumaroyl + caffeoyl derivatives in CT (Table 5). Canopy management can indeed affect the relative proportions of monomeric anthocyanins or acylated and p- coumarated forms, and results among grapevine varieties and crop levels are conflicting. Matus et al. (2009) found that leaf removal at veraison promoted a higher proportion of acylated and p-coumarated anthocyanins in Cabernet Sauvignon, and among monomers, a higher increase in cya-3-glu, del-3-glu, and pet-3-glu. Although there were no significant differences in the anthocyanin indices among treatments, di-, tri- and OCH 3 -substituted anthocyanins increased significantly in ELR, while mild effects were assessed in CT (Table 6). Matsuyama et al. (2014) reported an increase in tri-substituted anthocyanins in the hybrid Muscat Bailey A subjected to leaf removal, showing that F3 5 H the gene was 4.0-fold upregulated compared to the control. Chorti et al. (2010) applied leaf removal at berry set on Nebbiolo vines and found that the concentration of total anthocyanins was higher in the case of leaf removal in the 2006 season, and not significantly different in 2007, while the F3 5 /F3 ratio was high and low with leaf removal in the same two seasons, respectively. In addition, the concentration of p-coumaroyl derivatives increased in the first season in case of leaf removal, similar to our results. While some studies have examined the effect of leaf removal on the anthocyanin profile, fewer have considered the effects of CT. Bubola et al. (2011) showed a positive effect of CT on the increase of concentrations of most anthocyanin monomers and p-coumaroyl + caffeoyl derivatives in Merlot, but no effects on mal-3-glu and acetylated forms. In contrast, Gil et al. (2013) found that CT slightly affected the anthocyanin profile of Syrah grapes, with p-coumaroyl derivatives being more affected than other anthocyanins. The most novel and useful result derived from this study is that the effects of ELR are superior to those of CT for regulating yield without compromising ripening. Despite 24% higher yield in ELR than in CT, grape composition is also improved. Notably, CT, despite a significant reduction in yield, caused only minor improvements in grape quality, essentially 1 Brix higher than UC. Moreover, the economic sustainability of the ELR approach could be increased as it can be quite easily performed by machine, therefore dramatically diminishing intervention cost (Intrieri et al. 2008). The relative differences between treatments observed here suggest caution against the assumption that final leaf areato-yield ratio is a good indicator of optimal berry ripening. Comparing UC versus ELR, the two treatments had similar leaf area-to-yield ratios, but grape quality was considerably improved by ELR. The comparison of ELR versus CT is even more striking, where despite a higher leaf area-to-yield ratio in CT, Brix was similar, and paradoxically, anthocyanin concentrations were either similar or lower (Tables 5 and 6). The effects on anthocyanins might appear to be due to the significant differences in berry size and relative skin weight; however, total anthocyanin concentrations on a skin-weight basis were very similar for UC and CT, and were higher in ELR (Table 4). This suggests that cluster exposure, rather than the final leaf area-to-yield ratio, affected the anthocyanin concentrations in ELR berries of Teran, similar to results reported by Sternad Lemut et al. (2013) in Pinot noir. Thus, the factors controlling color accumulation appear to be the percentage of canopy gaps and LLN in the fruit zone. For comparison, Mabrouk and Sinoquet (1998) used a 3-D digitizing method in the field to characterize cluster microclimate in different training systems, and found that among several indices of canopy structure, the fraction of canopy gaps had the highest correlation (r = 0.87) with total anthocyanins concentration at harvest. Conclusions Our findings for Teran grapevines grown in Croatian Istria revealed that both ELR and CT are effective practices for regulating yield while increasing Brix in grape berries. However,

8 Leaf Removal and Cluster Thinning in Teran 241 ELR, despite leading to higher yield than CT, led to greater concentrations of phenolics and anthocyanins in berries. This was not likely to be attributable to variation in the leaf area-toyield ratio, which was lower in ELR than in CT vines. Factors accounting for the greater efficiency of ELR likely include: i) good responsiveness of cv. Teran to a more exposed fruit zone from early in the season; ii) greater late-season canopy efficiency in ELR due to younger mean leaf age; and iii) yield reduction achieved by adjustments in berry size and number rather than by reduced cluster numbers per vine. Literature Cited Avizcuri-Inac JM, Gonzalo-Diago A, Sanz-Asensio J, Martínez-Soria MT, López-Alonso M, Dizy-Soto M, Echávarri-Granado JF, Vaquero- Fernández L and Fernández-Zurbano P Effect of cluster thinning and prohexadione calcium applications on phenolic composition and sensory properties of red wines. J Agric Food Chem 61: Bubola M, Peršurić D and Ganić K Impact of cluster thinning on productive characteristics and wine phenolic composition of cv. Merlot. J Food Agric Environ 9: Bubola M, Peršurić D, Kovačević Ganić K, Karoglan M and Kozina B Effects of fruit zone leaf removal on the concentrations of phenolic and organic acids in Istrian Malvasia grape juice and wine. Food Technol Biotech 50: Chorti E, Guidoni S, Ferrandino A and Novello V Effect of different cluster sunlight exposure levels on ripening and anthocyanin accumulation in Nebbiolo grapes. Am J Enol Vitic 61: Cook MG, Zhang Y, Nelson CJ, Gambetta G, Kennedy JA and Kurtural SK Anthocyanin composition of Merlot is ameliorated by light microclimate and irrigation in central California. Am J Enol Vitic 66: Coombe BG The effect of removing leaves, flowers and shoot tips on fruit-set in Vitis vinifera L. J Hortic Sci 37:1-15. Coombe BG Growth stages of the grapevine: Adoption of a system for identifying grapevine growth stages. Aust J Grape Wine Res 1: Diago MP, Ayestarán B, Guadalupe Z, Poni S and Tardáguila J. 2012a. Impact of prebloom and fruit set basal leaf removal on the flavonol and anthocyanin composition of Tempranillo grapes. Am J Enol Vitic 63: Diago MP, Ayestarán B, Guadalupe Z, Garrido Á and Tardaguila J. 2012b. Phenolic composition of Tempranillo wines following early defoliation of the vines. J Sci Food Agric 92: Diago MP, Krasnow M, Bubola M, Millan B and Tardáguila J Assessment of vineyard canopy porosity using machine vision. Am J Enol Vitic 67: Gatti M, Bernizzoni F, Civardi S and Poni S Effects of cluster thinning and preflowering leaf removal on growth and grape composition in cv. Sangiovese. Am J Enol Vitic 63: Gatti M, Garavani A, Cantatore A, Parisi MG, Bobeica N, Merli MC, Vercesi A and Poni S Interactions of summer pruning techniques and vine performance in the white Vitis vinifera cv. Ortrugo. Aust J Grape Wine Res 21: Gil M, Esteruelas M, González E, Kontoudakis N, Jiménez J, Fort F, Canals JM, Hermosín-Gutiérrez I and Zamora F Effect of two different treatments for reducing grape yield in Vitis vinifera cv. Syrah on wine composition and quality: Berry thinning versus cluster thinning. J Agric Food Chem 61: Guidoni S, Allara P and Schubert A Effect of cluster thinning on berry skin anthocyanin composition of Vitis vinifera cv. Nebbiolo. Am J Enol Vitic 53: Herrera JC, Bucchetti B, Sabbatini P, Comuzzo P, Zulini L, Vecchione A, Peterlunger E and Castellarin SD Effect of water deficit and severe shoot trimming on the composition of Vitis vinifera L. Merlot grapes and wines. Aust J Grape Wine Res 21: Iland P, Bruer N, Ewart A, Markides A and Sitters J Monitoring the Winemaking Process from Grapes to Wine: Techniques and Concepts. Patrick Iland Wine Promotions Pty Ltd, Adelaide, Australia. Intrieri C, Filippetti I, Allegro G, Centinari M and Poni S Early defoliation (hand vs mechanical) for improved crop control and grape composition in Sangiovese (Vitis vinifera L.). Aust J Grape Wine Res 14: Keller M, Mills LJ, Wample RL and Spayd SE Cluster thinning effects on three deficit-irrigated Vitis vinifera cultivars. Am J Enol Vitic 56: Kemp BS, Harrison R and Creasy GL Effect of mechanical leaf removal and its timing on flavan-3-ol composition and concentrations in Vitis vinifera L. cv. Pinot Noir wine. Aust J Grape Wine Res 17: Kliewer WM and Dokoozlian NK Leaf area/crop weight ratios of grapevines: Influence on fruit composition and wine quality. Am J Enol Vitic 56: Komm BL and Moyer MM Effect of early fruit-zone leaf removal on canopy development and fruit quality in Riesling and Sauvignon blanc. Am J Enol Vitic 66: Lee J and Skinkis PA Oregon Pinot noir grape anthocyanin enhancement by early leaf removal. Food Chem 139: Mabrouk H and Sinoquet H Indices of light microclimate and canopy structure of grapevines determined by 3D digitising and image analysis, and their relationship to grape quality. Aust J Grape Wine Res 4:2-13. Maletić E et al Ampelographic and genetic characterization of Croatian grapevine varieties. Vitis 54: Matsuyama S, Tanzawa F, Kobayashi H, Suzuki S, Takata R and Saito H Leaf removal accelerated accumulation of delphinidin-based anthocyanins in Muscat Bailey A [Vitis labruscana (Bailey) and Vitis vinifera (Muscat Hamburg)] grape skin. J Japan Soc Hort Sci 83: Mattivi F, Guzzon R, Vrhovsek U, Stefanini M and Velasco R Metabolite profiling of grape: Flavonols and anthocyanins. J Agric Food Chem 54: Matus JT, Loyola R, Vega A, Peña-Neira A, Bordeu E, Arce-Johnson P and Alcalde JA Post-veraison sunlight exposure induces MYB-mediated transcriptional regulation of anthocyanin and flavonol synthesis in berry skins of Vitis vinifera. J Exp Bot 60: Nuzzo V and Matthews MA Response of fruit growth and ripening to crop level in dry-farmed Cabernet Sauvignon on four rootstocks. Am J Enol Vitic 57: OIV Compendium of international methods of wine and musts analysis. Office International de la Vigne et du Vin, Paris, France. Palliotti A, Gatti M and Poni S Early leaf removal to improve vineyard efficiency: Gas exchange, source-to-sink balance, and reserve storage responses. Am J Enol Vitic 62: Palliotti A, Gardi T, Berrios JG, Civardi S and Poni S Early source limitation as a tool for yield control and wine quality improvement in a high-yielding red Vitis vinifera L. cultivar. Sci Hortic 145: Pastore C, Zenoni S, Tornielli GB, Allegro G, Dal Santo S, Valentini G, Intrieri C, Pezzotti M and Filippetti I Increasing the source/sink ratio in Vitis vinifera (cv Sangiovese) induces extensive transcriptome reprogramming and modifies berry ripening. BMC Genomics 12:631.

9 242 Bubola et al. Pastore C, Zenoni S, Fasoli M, Pezzotti M, Tornielli GB and Filippetti I Selective defoliation affects plant growth, fruit transcriptional ripening program and flavonoid metabolism in grapevine. BMC Plant Biol 13:30. Poni S, Casalini L, Bernizzoni F, Civardi S and Intrieri C Effects of early defoliation on shoot photosynthesis, yield components, and grape composition. Am J Enol Vitic 57: Poni S, Bernizzoni F, Civardi S and Libelli N Effects of prebloom leaf removal on growth of berry tissues and must composition in two red Vitis vinifera L. cultivars. Aust J Grape Wine Res 15: Risco D, Pèrez D, Yeves A, Castel JR and Intrigliolo DS Early defoliation in a temperate warm and semi-arid Tempranillo vineyard: Vine performance and grape composition. Aust J Grape Wine Res 20: Rusjan D, Bubola M, Janjanin D, Užila Z, Radeka S, Poljuha D, Pelengić R, Javornik B and Štajner N Ampelographic characterisation of grapevine accessions denominated Refošk, Refosco, Teran and Terrano (Vitis vinifera L.) from Slovenia, Croatia and Italy. Vitis 54: Silvestroni O, Lanari V, Lattanzi T, Palliotti A and Sabbatini P Impact of crop control strategies on performance of high-yielding Sangiovese grapevines. Am J Enol Vitic 67: Sivilotti P, Herrera JC, Lisjak K, Baša Česnik H, Sabbatini P, Peterlunger E and Castellarin SD Impact of leaf removal, applied before and after flowering, on anthocyanin, tannin, and methoxypyrazine concentrations in Merlot (Vitis vinifera L.) grapes and wines. J Agric Food Chem 64: Smart RE and Robinson M Sunlight into Wine. A Handbook for Winegrape Canopy Management. Winetitles, Adelaide, Australia. Sternad Lemut M, Sivilotti P, Franceschi P, Wehrens R and 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 61: Sternad Lemut M, Sivilotti P, Butinar L, Laganis J and Vrhovsek U Pre-flowering leaf removal alters grape microbial population and offers good potential for a more sustainable and cost-effective management of a Pinot Noir vineyard. Aust J Grape Wine Res 21: Šuklje K, Baša Česnik H, Janeš L, Kmecl V, Vanzo A, Deloire A, Sivilotti P and Lisjak K The effect of leaf area to yield ratio on secondary metabolites in grapes and wines of Vitis vinifera L. cv. Sauvignon blanc. J Int Sci Vigne Vin 47: Tardaguila J, de Toda FM, Poni S and Diago MP Impact of early leaf removal on yield and fruit and wine composition of Vitis vinifera L. Graciano and Carignan. Am J Enol Vitic 61: Vivoda V Teran i Refošk u Istri. Hrvatsko agronomsko društvo, Zagreb, Croatia.