Abstract
A four-year trial (1999–2002) was carried out in commercial vineyards in order to test the effect of manual leaf removal in the cluster zone at veraison on grape yield, berry composition, and stilbene concentration at harvest. Two Vitis vinifera L. red wine cultivars (Barbera and Croatina) and one white wine cultivar (Malvasia di Candia aromatica) were considered. The vines were Guyot trained, with 10 shoots per meter of row, and leaf removal accounted for approximately 22% of removed canopy surface. Meteorological data were recorded, as well as grape production, fruit composition, and stilbenes (trans-resveratrol, trans-piceid, cis-piceid) at harvest. Grape yield was not affected by leaf removal in any of the tested cultivars. Grape sugars and acidity were affected by leaf removal differently depending on meteorological conditions and cultivar. Leaf removal increased trans-piceid grape concentration in Barbera and decreased trans-resveratrol and cis-piceid in Croatina and Malvasia di Candia aromatica under cool meteorological conditions. Leaf removal had no effect on stilbene grape concentrations under warmer and drier climatic conditions.
Leaf removal is a cultural practice performed in the fruit zone during the vegetative season between fruit set and ripening in order to improve spray treatment efficacy, to modify grape quality, and to reduce cluster density if performed at fruit set (Poni et al. 2006). At veraison, leaf removal affects primary and secondary metabolite synthesis and its effect is related to leaf layer number, photosynthesis rate, and canopy surface area. Leaf removal, at veraison, of very dense canopies increased sugars, flavors, total and phenol-free glycosides, and flavonoids and decreased acidity and gray mold attacks, as compared with untreated vines (Percival et al. 1994, Poni et al. 2006, Reynolds et al. 1994, Zoecklein et al. 1992, 1998). These effects are related to the modification of source/sink ratio, to the activation of light-induced enzymes such as the phenylalanine ammonia-lyase (PAL), and to better microclimate conditions (more air circulation and lower relative humidity) (Dokoozlian and Kliewer 1995, Haselgrove et al. 2000).
Leaf removal at veraison on plants with a low canopy density does not significantly affect grape sugar, acidity, and color (Reynolds et al. 1994, 1996). Under these conditions, basal leaf removal may cause a reduction in photosynthesis responsible in some cases for a decrease in must sugar concentration (Reynolds et al. 1994). According to one study, anthocyanin synthesis is reduced and quercetin-3-glycoside is increased under sunlight grape exposure (Haselgrove et al. 2000).
Stilbenic phytoalexins are synthesized if grapevine leaves or berries are challenged by fungi such as gray mold agent Botrytis cinerea Pers. (Langcake and Pryce 1976), downy mildew agent Plasmopara viticola (Berk. and Curt.) Berl. and de Toni (Langcake 1981), powdery mildew agent Oidium tuckeri Berk. (Langcake and Pryce 1976), berry rot agent Rhizopus stolonifer (Ehrenb. Fr.) Lind. (Sarig et al. 1997), and ochratoxin A-producing fungi Aspergillus carbonarius and Aspergillus ochraceus (Bavaresco et al. 2003, Vezzulli et al. 2007a).
In addition to biotic stresses, resveratrol is elicited by abiotic factors such as UV irradiation (Langcake and Pryce 1977), aluminum chloride (Adrian et al. 1996), ozone (Sarig et al. 1996), injury (Blaich and Bachman 1980), fosetyl-Al (Dercks and Creasy 1989), benzothiadia-zole (BTH) (Iriti et al. 2004), abscisic acid (ABA) (Ban et al. 2000), and methyl jasmonate (Vezzulli et al. 2007b).
Stilbenes (especially resveratrol) are claimed to have health functional properties in humans, primarily because of antioxidant activity. Resveratrol is present in wine as it is extracted from berry skins during alcoholic fermentation and it is found as free and bound forms (3-β-glucoside of resveratrol) (Mattivi et al. 1995).
Resveratrol concentrations of grapes (and wines) analyzed worldwide vary widely, depending on viticultural and enological conditions. Viticultural factors include grape variety (Bavaresco et al. 2007), climate, soil (Bavaresco et al. 2005, 2008, de Andrés-de Prado et al. 2007), and cultural practices such as training system (Bertamini and Mattivi 1999, Ban et al. 2000) and fertilizer (Bavaresco et al. 2001).
Flavonoid and stilbene pathways are branching pathways, the first catalyzed by chalcone synthase (CHS) enzyme and the latter by stilbene synthase (StSy) enzyme. Evidence suggests that leaf removal affects f lavonoid berry synthesis, and it is also likely that stilbene berry concentration may be modified.
The aim of this experiment was to evaluate the effect of veraison leaf removal around the cluster zone on grape production and composition and on stilbene concentration of cultivars Barbera, Croatina, and Malvasia di Candia aromatica.
Materials and Methods
Experimental design and plant material.
A four-year-trial (1999–2002) was carried out in Tidone Valley, Piacenza province, Italy (lat. 44°54′N). Climatic parameters (46 years of data, from 1962 to 2007) and the main parameters of the tested years, recorded by a meteorological station placed nearby the tested vineyards, are reported in Table 1⇓. The viticultural area, some 3,500 ha wide, belongs to DOC (Denominazione di Origine Controllata: denomination of wines by origin) Colli Piacentini. The three most commonly cultivated Vitis vinifera L. varieties of this region were tested: two red wine cultivars, Barbera and Croatina, and one white wine cultivar, Malvasia di Candia aromatica (Malvasia C.a.). Malvasia C.a. is typically vinified alone, producing a sweet aromatic wine, while Barbera and Croatina are vinified together, giving a wine suitable for aging called Gutturnio. The grape cultivars were grown in commercial vineyards, located at 240 m asl and close together inside an area of ~0.75 ha. The soil was a fertile clay–calcareous flysch, which was tilled to control weed growth. The vineyards were located on an eastern exposure slope, vines were about 20 years old and Guyot trained (cane pruning with vertical shoot-positioning and 10 shoots per meter of row) with an average of 20 buds/plant (including a 3-node-spur and a 17-node-cane), and plant density was 2450 vines/ha (2.4 m between rows and 1.7 m between vines in a row). Complete manual leaf removal around the cluster zone of each shoot was performed at veraison, for each grape variety, to increase cluster exposure, while untreated vines were used as a control. At veraison the average shoot length was about 1.6 m, while the basal shoot zone where leaves were removed covered about 35 cm; therefore, the percentage of canopy surface area removed per each fruiting shoot was ~22%.
Three blocks of 10 plants each were considered per each variety and treatment, and four plants per block were sampled at random, resulting in 12 plants per treatment. The vines did not show any disease symptom along the vegetative season and during harvest, and disease control methods consisted of about 10 spray treatments per year, every 10 days (average value) beginning from May (shoot 10 cm long) until the first 10 days of August (berry touch), by using sulfur (wettable powder) against powdery mildew and mancozeb against downy mildew; copper hydroxide was used against downy mildew in the last two spray treatments each year.
Yield components and fruit composition.
At harvest 12 vines per treatment were chosen, all the clusters of each plant were counted and weighed, and the basal clusters of two medial shoots were sampled (2 clusters per vine) for fruit composition analyses. The following yield and fruit parameters were measured: grape yield (kg/plant), mean cluster weight (g), juice total soluble solids (Brix) by a temperature-compensating refractometer, juice titratable acidity (g/L) by titration with NaOH 0.1 N, juice pH by a pH meter; and berry trans-resveratrol (trans-3,4′,5-trihydroxystilbene) and trans- and cis-piceid (trans- and cis-resveratrol-3-O-β-d-glucopyranoside) were extracted by a previous method (Bavaresco et al. 2001) and analyzed by HPLC (Bavaresco et al. 2007).
Statistical analysis.
A two-way-ANOVA with interactions was used per each grape variety, and the means were compared by LSD at the 5% level. The sources of variation were vintage year and leaf removal.
Results
Weather.
During the four experimental years the annual mean temperature and rainfall were average as compared with the 46-year data, except for 1999 which was cooler and less rainy (Table 1⇑). Comparison of growing season indices found that degree days were much lower in 1999 than in the subsequent years and were highest in 2000. Rainfall was very high during 2002, as well as the relative humidity. Meteorological parameters of the time between veraison (when leaf removal was performed) and harvest, which was some 45 days for Barbera and Croatina and 35 days for Malvasia C.a., are shown in Table 2⇓.
Yield components and fruit composition.
Barbera.
Leaf removal did not significantly affect cluster number per vine, crop yield, or cluster weight in any of the tested years (Figure 1⇓). Cluster number per vine and crop yield were significantly higher (p < 0.05) in 1999 than in the following years. Soluble solids were significantly ( p < 0.05) increased by leaf removal only in 1999, while no effect was observed during the other years (Figure 2⇓). The highest soluble solids were recorded in 2000, and the lowest in 1999. No significant effects of leaf removal on titratable acidity and pH were observed. The highest acidity and the lowest pH were detected in 2002.
Croatina.
Leaf removal did not significantly affect cluster number per vine and crop yield in any of the tested years (Figure 1⇑). However, cluster weight was significantly higher ( p < 0.05) in the untreated vines during 1999 and 2000. The crop yield recorded in 2000 was higher than that of the other vintages. In 2000 leaf removal led to a significant (p < 0.05) increase in soluble solids, while in 2001 there was a significant (p < 0.05) reduction (Figure 2⇑). Concerning the year effect, 2002 induced the lowest values of soluble solids. Leaf removal significantly ( p < 0.05) decreased titratable acidity in 2000, while increasing it in 2001 and 2002, and no significant effect was observed in 1999 (Figure 2⇑). Concerning the vintage year, the highest titratable acidity was recorded in 2002 and the lowest in 2000. Leaf removal significantly (p < 0.05) reduced pH in 2001, whereas no effect was observed in other years (Figure 2⇑).
Malvasia di Candia aromatica.
Cluster number per vine and crop yield were not significantly affected by leaf removal in any of the tested years, but it did significantly (p < 0.05) reduce cluster weight in 2000 (Figure 1⇑). During 1999 the productive parameters were not recorded. Leaf removal significantly ( p < 0.05) reduced soluble solids in 1999, while increasing it in 2002. No effects of leaf removal were observed in the other years (Figure 2⇑). Vintage year 2001 had the lowest sugar accumulation. Titratable acidity was significantly ( p < 0.05) reduced by leaf removal in 1999 and 2002, while no effects were observed in the other years. Year 2002 had the highest acidity, while year 2000 had the lowest (Figure 2⇑). Leaf removal had no significant effects on pH, which decreased from 1999 to 2002 (Figure 2⇑).
Stilbene composition.
Barbera.
Leaf removal did not significantly affect trans-resveratrol and cis-piceid berry concentration in any of the tested years, while it significantly ( p < 0.05) increased trans-piceid in 1999 (Figure 3⇓). The highest trans-resveratrol and trans-piceid concentrations of the year 1999 (0.139 and 0.845 mg/kg, respectively, average values of control and leaf removal treatments) were related to the lowest degree days. During 2001 trans-resveratrol was not synthesized, while the lowest values of trans- and cis-piceid were recorded in 2000, the warmest and least rainy ripening time (Table 2⇑). The highest cis-piceid concentration (0.295 mg/kg) of 2002 seemed to be positively related to rainfall and relative humidity of ripening time (Figure 3⇓, Table 2⇑).
Croatina.
A significant reduction (p < 0.05) of trans-resveratrol due to leaf removal was evident only in 1999 (Figure 3⇑). No effect of leaf removal on trans-piceid occurred in 1999, 2000, and 2002, while in 2001 leaf removal significantly ( p < 0.05) decreased trans-piceid concentration (Figure 3⇑). Leaf removal led to a significant (p < 0.05) reduction of cis-piceid in 1999 and 2001, while no significant effect was observed in the other years (Figure 3⇑).
Malvasia di Candia aromatica.
The high trans-resveratrol berry concentration of 1999 (0.136 mg/kg) and 2002 (0.103 mg/kg) seemed to be related to degree days, rainfall, and relative humidity of ripening time (Figure 3⇑, Table 2⇑), with degree days being the lowest and the other two parameters the highest. The trans-resveratrol reduction as a consequence of leaf removal was significant ( p < 0.05) in 1999 and 2002, while leaf removal significantly (p < 0.05) reduced cis-piceid only in 1999 (Figure 3⇑).
Discussion
The cumulative effect (on yield reduction) of repeating leaf removal on the same vines season after season reported elsewhere (Bennett et al. 2005) was not observed in this experiment. After the first year, particularly in Barbera, cluster number per plant, crop yield, and cluster weight were reduced at the same extent in both control and defoliated vines. The recorded data does allow any speculation about the cause. Cluster weight seemed to be under a genetic control, because leaf removal increased bunch size in Barbera while reducing it in Croatina and Malvasia C.a.
Genotype also affected the qualitative traits of control and defoliated vines. Genotype environment interactions showed that under cool meteorological conditions (1999) leaf removal improved grape ripeness in Barbera, while no effects were observed under warmer conditions (2000). In Croatina, leaf removal under cool meteorological conditions (1999 and 2001) had the opposite effect as compared with Barbera, reducing grape ripening, while under warmer and drier conditions (2000) leaf removal had a positive effect on grape maturity. In Malvasia C.a. leaf removal under cool meteorological conditions (2002) improved grape ripeness, while under warmer and drier conditions (2001) leaf removal had no effect. The data on the interactions between stilbenes on one hand and grape variety, leaf removal, and climate on the other hand are interesting. Meteorological data of the time between veraison and harvest (ripening period) were recorded (Table 2⇑) because, according to the literature (Jeandet et al. 1995, Bertamini and Mattivi 1999, Bavaresco et al. 2007), they might be related to grape stilbene occurrence, and particularly they might be useful to better understand the effect of canopy management practices under different meteorological conditions. The role of genotype was evident since the varieties had very different stilbene concentrations, especially trans- and cis-piceid, with Barbera the highest and Malvasia C.a. the lowest. Moreover, each variety reacted differently to leaf removal. Stilbene concentration in Barbera grapes was poorly affected by leaf removal over the years, except for a trans-piceid increase in 1999, which was the coolest year (considering both heat summation from April to September and from veraison to harvest time). Still in 1999, soluble solids of leaf removed clusters were higher and acidity lower than control bunches (see also Reynolds et al. 2007), indicating that, from a practical point of view, Barbera benefits from leaf removal in terms of better ripening and more trans-piceid in cool years (like 1999), while in warmer years the practice is not effective and therefore useless. Croatina reacted to leaf removal more significantly than Barbera. Under the cool temperatures (veraison-harvest period) of 1999 and 2001, stilbenes were on the average depressed by leaf removal, as well as soluble solids, while acidity increased. In the warmest and least rainy ripening time of 2000, no significant effect of leaf removal occurred on stilbenes, while soluble solids increased and acidity and cluster size were depressed. In other words, leaf removal in Croatina variety does not have positive effects on grape stilbenes, irrespective of the climate. From a practical point of view, leaf removal in this variety can be useful in warm and very dry years (like 2000) to improve grape ripening and have a less compact cluster. Malvasia C.a. reacted to leaf removal more than Barbera, but less than Croatina. trans-Resveratrol was reduced by leaf removal under the cool temperatures, high rainfall, and relative humidity (veraison-harvest period) of 1999 and 2002. No significant effect of leaf removal on trans-resveratrol occurred in the warmer and drier (veraison-harvest period) years 2000 and 2001. No effects of leaf removal on trans-piceid were detected, while cis-piceid was reduced by leaf removal only during 1999. Even with Malvasia C.a., leaf removal had no positive effect on grape stilbenes, irrespective of the climate. From a practical point of view, leaf removal could be useful in Malvasia C.a. in mid-warm and dry years (like 2000) because it favors a less compact cluster without a reduction of stilbenes, sugars, and acidity: this is important to reduce gray mold attacks.
Further investigations are necessary to better understand stilbene physiology in the vine, and particularly in the cluster, under direct sunlight bunch exposure, and the interaction with the temperatures.
Conclusion
Leaf removal did not affect crop yield of the tested grape varieties irrespective of the yearly meteorological conditions. Cluster weight, berry composition, and stilbenes were affected differently depending on the meteorological conditions of the ripening time and the cultivar. Under cool meteorological conditions, leaf removal had a positive effect on cluster size, grape soluble solids, and trans-piceid concentrations of Barbera and had negative effects on Croatina soluble solids and stilbenes. In Malvasia C.a., leaf removal decreased grape acidity and trans-resveratrol.
Under warmer and drier meteorological conditions Barbera did not benefit from leaf removal in terms of grape sugar and stilbene concentration; Croatina did benefit from leaf removal in terms of sugar content and less compact cluster, while stilbene concentration remained unaffected; and Malvasia C.a. did benefit from leaf removal in terms of less compact cluster, without a reduction of sugars, acidity, and stilbene concentration.
Footnotes
Acknowledgments: The research was supported by Centro Ricerche Produzioni Vegetali, Faenza RA, and Ministero Istruzione Università e Ricerca.
The authors thank Giuseppe Bruzzi (Università Cattolica lab crew) for his contribution to the project and the Servizio Agrometeorologico of Piacenza provincial administration and Osservatorio Sismico Meteorologico Alberoni for providing meteorological data.
- Received November 2007.
- Revision received March 2008.
- Copyright © 2008 by the American Society for Enology and Viticulture