Resveratrol production potential of grape flowers and green berries to screen genotypes for gray mold and powdery mildew resistance

Abstract

The ability of grape cultivars to produce resveratrol in response to UV radiation is associated with their resistance to fungal pathogens. In this study, we evaluated the UV-induced resveratrol production potential of grape flowers and green berries of 72 grape genotypes. Their ability to produce resveratrol in response to UV radiation was used to establish a selection criterion for screening genotypes for resistance to gray mold and powdery mildew. We quantified resveratrol in grape flower extracts from pre-bloom and early bloom flowers after exposure to UV radiation. There was a strong negative correlation between UV-induced resveratrol production and susceptibility to Botrytis infection. The grape genotype was considered to be susceptible to gray mold when the resveratrol content of whole flower clusters was less than 10 μg g⁻¹ FW after the UV treatment. We determined resveratrol production in response to UV radiation in whole grape berries from 0 to 30 days after full bloom. The ability to produce resveratrol in response to UV increased by 8- to 20-times during this period, depending on the genotype. At 30 days after full bloom, Vitis vinifera genotypes generally had low levels of resveratrol (<50 μg g⁻¹ FW), while interspecific hybrids, especially the North American cultivars, had high levels (>50 μg g⁻¹ FW). There were clear varietal differences in resistance to Erysiphe (powdery mildew) infection. Susceptibility to Erysiphe infection was strongly negatively correlated with UV-induced resveratrol production.

Introduction

Gray mold (Botrytis cinerea Pers.) and powdery mildew (Erysiphe necator Schwein) are widespread fungal diseases of field- and greenhouse-grown grapevines in temperate areas (Pearson and Goheen 1994; Reisch and Pratt 1996). In viticulture, serious economic losses result from the effects of these fungal infections on the fruit, e.g., poor fruit set, premature drop of bunches, uneven maturation, cracking that leads to fruit rot, and decreased yields. In Japan, fungicides are applied at high levels to grapes from the bloom stage to the green berries stage to control these diseases, especially in greenhouse-grown grapevines. Breeding grape cultivars with innate resistance to these fungal diseases could reduce the costs of chemical control.

In grapevines, disease resistance involves both structural and biochemical defense mechanisms. One of these mechanisms, the production of the stilbene-type phytoalexin, trans-resveratrol (hereafter resveratrol), is particularly effective against fungal diseases, as it inhibits spore germination and growth of the fungal pathogen (Dercks et al. 1995). Resveratrol production of grape leaves in response to fungal attacks or short ultraviolet (UV)-treatments has been extensively studied (Langcake and Pryce 1977; Langcake and McCarthy 1979; Langcake 1981; Pool et al. 1981; Stein and Blaich 1985; Barlass et al. 1987; Bavaresco and Eibach 1987; Dercks and Creasy 1989; Adrian et al. 1997). The results of these studies indicated that the capacity for resveratrol production is positively correlated with disease resistance of grapevines.

Ultraviolet radiation is a useful induction technique, because it induces resveratrol synthesis in a large proportion of exposed cells. Thus, the potential for resveratrol production in response to UV radiation may be a useful marker in assessing resistance of flowers/green berries to diseases, as was the case for predicting disease resistance of leaf tissue (Sbaghi et al. 1995). However, there are few studies on resveratrol production potential of grape flowers and green berries (Creasy and Coffee 1988; Jeandet et al. 1991; Bavaresco et al. 1997; Keller et al. 2003). If there are varietal differences in the potential for resveratrol production in flowers and green berries, this could be used to screen grape genotypes for disease resistance. Screening procedures are a crucial part of any breeding program, and must be simple and rapid. The purpose of this study was (1) to evaluate UV-induced resveratrol production in grape flowers and green berries, and (2) to use these results to establish a selection criterion for screening grapevine genotypes (table and wine cultivars) for resistance to gray mold and powdery mildew.

Materials and methods

Plant material

We used 72 grape genotypes, as follows (Table 1): 48 Vitis labruscana × Vitis vinifera interspecific hybrids (Aki Queen, Akitsu No. 27, Alden 4X, Black Beat, Buffalo 4X, Campbell Early, Delaware, Festivee, Fujiminori, Fukuoka No. 12, Gorby, Honey Venus, Houman, Kyoho, Kai Noir, Muscat Bailey A, Neo Alicante, Niagara Rosada, Ohtama Portland, Olympia, Oriental Star, Pierce, Pione, Portland, Red Port, Red Port 4X, Rubired, Schuyler 4X, Shine Muscat, Steuben, Steuben 4X, Suiho, Sunny Rouge, TSA-4, Wayne 4X, Yoho, 3891, 5512, 95-13-2, 02-2-38, 02-3-24, 02-3-32, 02-7-2, 02-9-4, 02-9-13, 02-10-2, 02-12-44, and 03-6-1); 22 V. vinifera genotypes (Alicante Bouschet, Bijou Noir, Cabernet Sauvignon, French Colombard, Fukuoka No. 13, Fukuoka No. 14, Hakata White, Harmo Noir; Merlot, Muscat of Alexandria, Muscat Hamburg, Neo Muscat, Palestine, Red Ohanez, Rosario Bianco, Ruby Okuyama, Ruby Seedless, Sunny Dolce, Syrah, Yatomi Rosa, 6706, and 97-1-14); and two rootstocks (Kober 5BB and Riparia Gloire de Montpellier). All genotypes were grown with a vine-by-row spacing of 0.8 × 2.0 m in a North–South orientation in a greenhouse at the Fukuoka Agricultural Research Center, Fukuoka, Japan. We used one self-rooted vine per genotype. The vines were cane-pruned to 15–20 nodes and trained onto a horizontal trellis system. Fungicides were applied to prevent latent infection with gray mold and powdery mildew before flower or berry sampling. The spraying regime was as follows: at the five- to six-leaves unfolded stage, we applied famoxadone; at the beginning of flowering, iprodione; at late flowering, diethofencarb; at fruit set, azoxystrobin; and when berries were pea-sized, triflumizole. These fungicide applications were carried out for all experiments except for experiment 5, in which we evaluated natural infection of grapevines by E. necator.

Table 1 Origin and use of 72 grape genotypes used in this study

Sample preparation of flower clusters and green berries

Detached clusters were washed in running water for 1 h. After surface sterilization (70% ethanol, 30 s) and rinsing in sterile water, three clusters were placed on a Petri dish (ϕ 150 × 15 mm) containing moist filter paper. In the following experiments, we used three to five Petri dishes (replicates) of each genotype for UV irradiation and Botrytis inoculation.

Gray mold experiments

Experiment 1

In 2007, we sampled eight American hybrids and one vinifera genotype to determine resveratrol production and Botrytis resistance at the pre-bloom stage. Unopened flower clusters were sampled at stage 17 (fully developed; flowers separating) as defined by Eichhorn and Lorenz (1977). The detached clusters (three replicates for each cultivar) were irradiated for 15 min at a distance of 10 cm with a 254 nm wavelength UV lamp (SUV-16, 16 W, 2020 μW cm⁻² at 5 cm, AS ONE, Osaka, Japan) in a dark room at 25°C. After UV irradiation, the clusters were enclosed in a polyethylene bag (0.03 × 260 × 380 mm), kept in the dark at 25°C for 48 h (Pool et al. 1981), and then immediately frozen at −20°C until resveratrol analysis. For the Botrytis inoculation, clusters from five replicates were dipped in Botrytis conidial suspension (10⁵ spores ml⁻¹). After inoculation, the clusters were placed in a Petri dish and enclosed in a polyethylene bag as above, and kept in the dark at 20°C for 5 days.

Experiment 2

In 2007, we sampled four interspecific hybrids and one vinifera genotype to determine resveratrol production and Botrytis resistance at the early bloom stage. Opened flower clusters were sampled at stage 21 (25% of caps fallen) as defined by Eichhorn and Lorenz (1977). The detached clusters (five replicates of each genotype, including controls) were UV irradiated and incubated as described above. After incubation, clusters were immediately frozen at −20°C until resveratrol analysis. For the Botrytis inoculation, clusters (five replicates per cultivar) were inoculated and incubated as described in experiment 1.

Experiment 3

In 2008, we determined resveratrol production at early bloom in 69 genotypes (46 interspecific hybrids, 21 vinifera genotypes, and two rootstocks). In each genotype, five replicates of detached clusters were UV irradiated and incubated as described in experiment 1. After incubation, the clusters were immediately frozen at −20°C until resveratrol analysis.

B. cinerea cultures and preparation of conidial suspension

B. cinerea (strain BCT01, MAFF 665011) was supplied by Dr. R. Nakaune (Grape and Persimmon Research Station, NIFTS, Higashi-Hiroshima, Japan). Fungal colonies were cultured on potato dextrose agar at 20°C in the dark for 72 h. Sporulation was induced by BLB irradiation (365 nm, 4 W, 513 μW cm⁻² at 5 cm) at a distance of 10 cm for 48 h, then incubation for 48 h. A conidial suspension was prepared from these 7-day-old cultures by flooding the Petri dishes with sterile potato dextrose solution (0.2% potato starch, 1.0% dextrose, 0.01% Tween 20) and scraping the agar surface to dislodge conidia. The conidial suspension was obtained by filtering the mixture through two layers of cheese cloth.

Visual assessment of symptoms

Disease symptoms on the clusters were classified according to the following scale: 0 = no symptoms; 1 = one to two pedicels infected; 3 = three to four pedicels infected; 5 = more than five pedicels infected. The infection index for Botrytis (IIB, %) was calculated as follows:

$$ IIB = \left( {{{\left[ {\sum\limits_{{i = 0}}^{5} {n_{i} \times i} } \right]} \mathord{\left/ {\vphantom {{\left[ {\sum\nolimits_{{i = 0}}^{5} {n_{i} \times i} } \right]} {\left[ {\sum\nolimits_{{i = 0}}^{5} {n_{i} \times 5} } \right]}}} \right. \kern-\nulldelimiterspace} {\left[ {\sum\limits_{{i = 0}}^{5} {n_{i} \times 5} } \right]}}} \right) \times 100 $$

where i = the infection scale rating (0–5) and n = the number of clusters in each infection class.

Powdery mildew experiments

Experiment 4

In 2007, we selected three interspecific hybrid genotypes to monitor changes in UV-induced resveratrol production after full bloom. The clusters were sampled at 0, 15, and 30 days after full bloom, and were UV irradiated as described in experiment 1.

Experiment 5

In 2008, we examined resveratrol production in grape berries at 30 days after full bloom. For this analysis, we selected 68 genotypes (46 interspecific hybrids, 21 vinifera genotypes, and 1 rootstock). We selected 6–10 berries, depending on berry size, for UV irradiation using the procedures described in experiment 1. The remaining fresh berry samples were crushed in a hand press, and the percentage of soluble solids in the juice was measured with a refractometer (N1, ATAGO, Japan).

Visual assessment of symptoms

A total of 15 genotypes (Alden 4X, Buffalo 4X, Campbell Early, Delaware, Festivee, French Colombard, Muscat of Alexandria, Muscat Hamburg, Portland, Red Ohanez, Rubired, Ruby Okuyama, Ruby Seedless, Schuyler 4X, and Steuben 4X) were naturally infected by E. necator in the isolation greenhouse in 2008. No fungicides or pesticides were sprayed in this greenhouse. We used one self-rooted vine per genotype, and each vine was 5–7 years old. We monitored disease symptoms on 5–7 berry clusters per clone from 30 to 40 days after full bloom. The disease severity was classified as follows: 0 = no symptoms; 1 = up to 10% of berries infected; 2 = 11–30% of berries infected; 3 = 31–50% of berries infected; 4 = more than 51% of berries infected. The infection index for powder mildew (IIE, %) was calculated as follows:

$$ IIE = \left( {{{\left[ {\sum\limits_{{i = 0}}^{4} {n_{i} \times i} } \right]} \mathord{\left/ {\vphantom {{\left[ {\sum\limits_{{i = 0}}^{4} {n_{i} \times i} } \right]} {\left[ {\sum\limits_{{i = 0}}^{4} {n_{i} \times 4} } \right]}}} \right. \kern-\nulldelimiterspace} {\left[ {\sum\limits_{{i = 0}}^{4} {n_{i} \times 4} } \right]}}} \right) \times 100 $$

where i = scale rating (0–4) and n = number of clusters in each scale rating.

Extraction and analysis of resveratrol

The whole flower cluster (0.8–1.2 g) was weighed and homogenized with a homogenizer (POLYTRON PT10-35, Switzerland) in 40 ml methanol. The homogenate was vigorously shaken for 30 min in the dark at 25°C. The mixture was centrifuged at 5000×g for 10 min. The supernatant volume was topped up to 50 ml with methanol and filtered through a 0.45 μm membrane filter for high performance liquid chromatography (HPLC). For green berry extraction, the berries were detached from their clusters, cut in half longitudinally, and deseeded while still frozen. The berry samples (skin + pulp, 2.0–3.0 g) were weighed, homogenized, and prepared for resveratrol analysis using the same methods that were used to prepare flower clusters. HPLC analysis of resveratrol was performed as described by Sato et al. (1997) with some modifications. The HPLC system was a Shimadzu LC-10A gradient liquid chromatograph equipped with a SCL-10A system controller, LC-10AD pumps, a CTO-10A column oven, and a SPD-10A detector. We used a C18 reverse-phase column (CAPCELL PAK Phenyl UG120, 4.6 × 250 mm, 5 μm particle size, SHISEIDO, Tokyo, Japan). The column was operated at 35°C with a mobile phase flow rate of 1.0 ml min⁻¹. Solvent A was 0.4% phosphoric acid (v/v), and solvent B was 20% solvent A (v/v) and 80% acetonitrile (v/v). The resveratrol peak was detected at 303 nm with a SPD-10A detector (range: 1.0000AUFS; response: 4) using a trans-resveratrol standard (Sigma Chemical Co., St. Louis, MO, USA). The gradient elution was initially 0% solvent B, which increased linearly to 30.0% in 20 min. The injection volume was 30–50 μl.

Results

Gray mold experiments

Experiment 1

After UV irradiation, extracts were prepared from flowers and were analyzed by HPLC. The resveratrol peak in the chromatograms of extracts was positively identified by comparison of its retention time with that of the standard. Figure 1 shows typical HPLC profiles of the standard (a), UV-induced resveratrol in Muscat Bailey A (b) and Rosario Bianco (c), and photos of Botrytis inoculation in 2007. Muscat Bailey A synthesized more resveratrol and showed few disease symptoms, whereas the susceptible cultivar Rosario Bianco produced very little resveratrol and showed extensive disease symptoms, e.g., aerial mycelia.

Fig. 1

Typical HPLC profiles of UV-induced resveratrol in grape flower clusters at pre-bloom, and Botrytis infection in 2007. a standard (trans-resveratrol) b Botrytis-resistant cv. Muscat Bailey A (few disease symptoms) c Botrytis-susceptible cv. Rosario Bianco (extensive infection with aerial mycelia) Arrow indicates trans-resveratrol peak

Production of resveratrol in response to UV radiation (X) was significantly negatively correlated with the infection index of Botrytis (IIB, Y) in 2007 (Y = −3.0758X + 105.83, R² = 0.8367, Fig. 2). Muscat Bailey A (27.3 μg g⁻¹ FW and IIB = 2.9%) was considered to be highly resistant to Botrytis. Moderately Botrytis-resistant genotypes with 40–60% IIB produced lower levels of resveratrol, ranging from 13.6 μg g⁻¹ FW (5512) to 25.3 μg g⁻¹ FW (02-10-2). Susceptible genotypes with more than 80% IIB synthesized low levels of resveratrol ranging from 3.2 μg g⁻¹ FW (Aki Queen) to 6.2 μg g⁻¹ FW (Rosario Bianco).

Fig. 2

Relationship between UV-induced resveratrol production and infection index of Botrytis (IIB) in grape flower clusters at pre-bloom, 2007. 1 Muscat Bailey A, 2 Kyoho, 3 Yoho, 4 02-10-2 5:02-2-38 6:5512 7: Red Port, 8 Rosario Bianco, 9 Aki Queen

Experiment 2

In 2007, we evaluated resveratrol production in flowers at the early bloom stage in response to UV radiation, and their resistance to B. cinerea (Table 2). In all genotypes, UV radiation resulted in resveratrol levels 6- (5512) to 41-times (Yoho) greater than that of their respective controls. Similar to unopened flowers, in opened flowers the UV-induced resveratrol production was strongly negatively correlated with IIB. Delaware (resveratrol, 24.7 μg g⁻¹ FW; IIB = 18%) was highly Botrytis resistant. Yoho was moderately Botrytis-resistant (resveratrol, 20.6 μg g⁻¹ FW; 24% IIB). Susceptible genotypes with IIB values greater than 80% synthesized low levels of resveratrol (2.4 μg g⁻¹ FW in 5512; 9.5 μg g⁻¹ FW in Hakata White). The UV-induced resveratrol production decreased from the pre-bloom to the early bloom stage in some cultivars, e.g., Kyoho (24.7 at pre-bloom vs. 17.2 at early bloom), Yoho (23.7 vs. 20.6), and 5512 (13.6 vs. 2.4).

Table 2 Resveratrol production following UV-irradiation and resistance to Botrytis cinerea for grape flower clusters at early bloom in 2007

Experiment 3

Table 3 shows the UV-induced resveratrol production of 69 grape genotypes at early bloom in 2008. The genotype had a significant effect on UV-induced resveratrol production (ANOVA, P < 0.05) and the least significant difference (LSD) among the means of the genotypes was 3.0 μg g⁻¹ FW. In general, vinifera genotypes produced low levels of resveratrol (<10 μg g⁻¹ FW). In American hybrids, resveratrol production varied from 1.1 (Olympia) to 22.8 (Delaware) μg g⁻¹ FW. Compared with cultivars, rootstocks derived from American species produced high levels of resveratrol, e.g., Kober 5BB (22.1 μg g⁻¹ FW) and Riparia Gloire de Montpellier (26.3 μg g⁻¹ FW).

Table 3 Varietal difference in resveratrol production (in parentheses) of grape flowers following UV-irradiation at early bloom in 2008

From the results of experiments 1–3, we classified grape genotypes according to their resveratrol production as follows: 37 low-resveratrol (LRG: <10 μg g⁻¹ FW), 26 medium-resveratrol (MRG: 10–20 μg g⁻¹ FW), and 6 high-resveratrol (HRG: >20 μg g⁻¹ FW) genotypes. With respect to gray mold, MRG and HRG cultivars are considered to be field resistant under the general fungicide spray programs used in commercial Japanese viticulture.

Powdery mildew experiments

Experiment 4

In 2007, we determined the changes in resveratrol production in response to UV radiation in green berry extracts of three grape genotypes at 0, 15, and 30 days after full bloom (Table 4). The berries produced significantly (P < 0.05) more resveratrol in response to UV radiation over the 30-day period, reaching 20.3 μg g⁻¹ FW (5512), 146.0 μg g⁻¹ FW (Yoho), and 208.9 μg g⁻¹ FW (Campbell Early) by 30 days after full bloom. The amount of resveratrol produced in response to UV radiation differed among the genotypes, ranging from an 8- (Yoho) to 20-times (5512) increase from day 0 to day 30.

Table 4 Resveratrol production in green grape berries following UV-irradiation after full bloom in 2007

Experiment 5

At 30 days after full bloom, soluble solids content of 68 grape genotypes ranged from 2.6 (Red Port) to 4.4 (95-13-2)°Brix, averaging 3.1°Brix among all genotypes. For Riparia Gloire de Montpellier (rootstock), berry sampling was not possible, because this genotype has male flowers with nonfunctional pistils, and therefore very low fruit set. Table 5 shows varietal differences in the UV-induced resveratrol production of 68 grape genotypes at 30 days after full bloom in 2008. The genotype had a significant effect (ANOVA, P < 0.05) and the LSD among the means of the genotypes was 23.1 μg g⁻¹ FW. Portland produced the highest levels of resveratrol (273.1 μg g⁻¹ FW), and Muscat Hamburg the lowest (11.6 μg g⁻¹ FW). Among the vinifera genotypes, resveratrol production was generally low (<50 μg g⁻¹ FW), although several cultivars produced higher levels, e.g., Alicante Bouschet (51.3 μg g⁻¹ FW), Palestine (76.9 μg g⁻¹ FW), and Neo Muscat (85.7 μg g⁻¹ FW). Meanwhile, interspecific hybrids generally produced high levels of resveratrol (>50 μg g⁻¹ FW), especially North American cultivars, e.g., Delaware (163.9 μg g⁻¹ FW), Festivee (170.6 μg g⁻¹ FW), and Steuben (245.0 μg g⁻¹ FW). However, some American hybrid genotypes produced low levels of resveratrol, e.g., Olympia (17.5 μg g⁻¹ FW), Houman (36.1 μg g⁻¹ FW), and Kyoho (44.8 μg g⁻¹ FW).

Table 5 Varietal differences in resveratrol production (in parentheses) of grape green berries following UV-irradiation at 30 days after full bloom in 2008

Figure 3 shows typical symptoms of green berries naturally infected by E. necator in 2008. These vines were grown in an isolation greenhouse and were not sprayed with fungicides or pesticides. Delaware (IIE = 22.0%) was considered to be Erysiphe-resistant, Rubired was considered to be moderately resistant (IIE = 55.6%), and Ruby seedless (IIE = 100.0%) was considered to be susceptible. In 2008, UV-induced resveratrol production (X) was significantly (P < 0.05) negatively correlated with infection index for powdery mildew (IIE, Y) (Y = −33.995Ln(X) + 199.43, R² = 0.9242, Fig. 4). In general, genotypes with low levels of UV-induced resveratrol (<50 μg g⁻¹ FW) were susceptible to powdery mildew, especially vinifera genotypes, e.g., Muscat Hamburg, Ruby Okuyama, Red Ohanez, and Ruby Seedless (UV-induced resveratrol levels of 11.6, 17.8, 24.0, and 33.8 μg g⁻¹ FW, respectively). In contrast, genotypes producing higher levels of resveratrol (>100 μg g⁻¹ FW) in response to UV radiation were generally resistant to powdery mildew, especially interspecific hybrids such as Buffalo 4X (132.2 μg g⁻¹ FW), Steuben 4X (191.0 μg g⁻¹ FW), and Campbell Early (218.9 μg g⁻¹ FW).

Fig. 3

Typical symptoms of grapes naturally infected with Erysiphe necator (IIE) in isolation greenhouse (no fungicide or pesticide sprays) in 2008. a Erysiphe-resistant cv. Delaware (partial infection) b Moderately Erysiphe-resistant cv. Rubired c Erysiphe-susceptible cv. Ruby Seedless (total infection with mycelia) Open circles show areas of infection

Fig. 4

Relationship between UV-induced resveratrol production and infection index of Erysiphe (IIE) in grape green berries in 2008. 1 Portland, 2 Campbell Early, 3 Steuben 4X, 4 Festivee, 5 Delaware, 6 Schuyler 4X, 7 Buffalo 4X, 8 Alden 4X, 9 French Colombard, 10 Rubired, 11 Muscat of Alexandria, 12 Ruby Seedless, 13 Red Ohanez, 14 Ruby Okuyama, 15 Muscat Hamburg

As in the gray mold experiments, the grape genotypes were classified according to their resveratrol production, as follows: 25 low-resveratrol (LRP: <50 μg g⁻¹ FW), 23 medium-resveratrol (MRP: 50–100 μg g⁻¹ FW), and 20 high-resveratrol (HRP: >100 μg g⁻¹ FW) genotypes. LRP genotypes, especially vinifera cultivars, are likely to be susceptible to powdery mildew in commercial Japanese viticulture.

Discussion

We evaluated the potential of grape flowers and green berries at several developmental stages (pre-bloom until 30 days after full bloom) to produce resveratrol in response to UV treatment. A short UV treatment (15 min) is sufficient to induce resveratrol production in grape genotypes, as shown by other workers (Barlass et al. 1987; Creasy and Coffee 1988; Jeandet et al. 1991). The UV-induced resveratrol production in flowers and berry clusters was positively correlated with resistance to gray mold and powdery mildew. There was also a wide variation in the resveratrol production potential among the genotypes at the early bloom stage and at 30 days after full bloom. Our findings suggest that assaying resveratrol production in response to UV irradiation is effective method to screen for resistance to gray mold at the early bloom stage, and for resistance to powdery mildew at 30 days after full bloom. With respect to sampling of flower clusters, it is easier to select opened flowers (early bloom) than unopened ones (pre-bloom) for screening, because it is easy to judge the stage of flower development in open flowers.

At pre-bloom and early bloom, grape genotypes that produced low levels of resveratrol (<10 μg g⁻¹ FW; LRG) in the whole flower in response to UV radiation were susceptible to gray mold. In contrast, those that produced higher levels of resveratrol (>20 μg g⁻¹ FW; HRG) were resistant to gray mold. This was observed in both artificial inoculation experiments and in field-grown grapevines. There is little published data about resveratrol content of flower clusters. Creasy and Coffee (1988) analyzed unopened and opened flowers of the interspecific hybrid Chancellor, and found trace quantities of resveratrol ranging from 0.3 (0.02 μg cm⁻²) to 1.2 (0.07 μg cm⁻²) μg g⁻¹ FW. Keller et al. (2003) showed that the resveratrol levels in Botrytis-inoculated vinifera Gamay flowers ranged from 8.2 (full bloom) to 14.0 (pre-bloom) μg g⁻¹ FW. In addition, they reported that Botrytis-inoculated vinifera Chardonnay flowers accumulated low levels of resveratrol (1.0–3.0 μg g⁻¹ FW). Considering the results of Keller et al. (2003) and those of our study, it appears that vinifera cultivars have limited potential for resveratrol production. Adrian et al. (1997) tested the activity of solutions of resveratrol against gray mold, and found that 160 μg ml⁻¹ FW inhibited gray mold spore germination, and 60–140 μg ml⁻¹ FW inhibited mycelial growth. Compared with their results, however, the gray mold resistant genotypes (HRG) in this study produced much lower levels of resveratrol at early bloom. According to Keller et al. (2003), conidial germination does not guarantee successful penetration and establishment of Botrytis infection in grape flowers. Instead, they proposed that the high susceptibility of vinifera grape flowers to gray mold may be related not only to their poor capacity for resveratrol synthesis, but also to their constitutively low levels of soluble phenolic compounds (mainly derivatives of quercetin and hydroxyl-cinnamic acid). Gabler et al. (2003) also suggested that disease resistance is a result of the balance between the plant’s ability to produce stilbenes (resveratrol) and the ability of the fungus to degrade stilbenes via stilbene oxidase.

The mean soluble solids content among all the genotypes at 30 days after full bloom (the green berries stage) was approximately 3°Brix. This stage coincides with the period of maximum resveratrol production potential in grape berries (Jeandet et al. 1991; Bavaresco et al. 1997). Creasy and Coffee (1988) pointed out that resveratrol production potential decreased as the soluble solids content increased, therefore, berries could not produce potentially inhibitory concentrations of resveratrol when they reached 5% soluble solids. Gadoury et al. (2001, 2003) reported that ontogenic resistance to powdery mildew in grape berries increased markedly at approximately 3–4 weeks postbloom in an interspecific V. labruscana × vinifera hybrid and in vinifera cultivars. That is, fruit were almost immune to infection before °Brix levels reached 8%. The results of those studies demonstrated that UV-induced resveratrol production of grape berries at 30 days after full bloom is associated with their powdery mildew resistance. In general, American species and their interspecific hybrid genotypes synthesize more resveratrol than vinifera genotypes (Bavaresco and Fregoni 2001). In the present study, we determined resveratrol content on a whole berry basis. Therefore, we cannot compare the values obtained in our study with those reported previously, as those values represent resveratrol extracted from fresh skin, the primary site of resveratrol production. As expected, however, the interspecific hybrid genotypes generally accumulated more resveratrol in response to UV treatments than their counterparts. This resulted in the wide variation in UV-induced resveratrol production among grape genotypes in the present study. In a preliminary test, we determined UV-induced resveratrol content of 16 genotypes on the basis of both skin and berry FW. We determined that resveratrol of 1.0 μg g⁻¹ skin FW would be equivalent to approximately 16.0 μg g⁻¹ berry FW. Takayanagi et al. (2004) reported that the UV-induced resveratrol content of Muscat Bailey A (interspecific hybrid), Chardonnay (vinifera), and Koshu (vinifera) at 30 days after full bloom was 119.4 μg g⁻¹ berry FW (1910 μg g⁻¹ skin FW), 79.4 μg g⁻¹ berry FW (1270 μg g⁻¹ skin FW), and 78.8 μg g⁻¹ berry FW (1260 μg g⁻¹ skin FW), respectively. Our results are similar to their reports.

When grapevines were cultivated in a greenhouse with no spray program, those that produced low levels of resveratrol (<50 μg g⁻¹ FW) in response to UV radiation were more susceptible to powdery mildew infection, particularly the vinifera genotypes. In contrast, interspecific hybrid genotypes frequently showed higher resistance to powdery mildew and high UV-induced resveratrol production (>100 μg g⁻¹ FW). As is the case in Botrytis-resistance, the resistance of grape berries to E. necator represents a combination of many attributes, and cannot be attributed to a single resistance factor, i.e., resveratrol production. Previous studies have reported that Botrytis infection of grape berries also depends on the number and thickness of epidermal and hypodermal cell layers, the thickness of the cuticle, and the amount of cuticular wax (Dubos et al. 1993; Gabler et al. 2003). For example, highly resistant interspecific hybrid cultivars (e.g., Niagara, Niabell, Fredonia) had few or no pores on the berry surface. The varietal differences in Erysiphe-resistance observed in the present study may be attributed, as least in part, to such differences in the anatomy of grape berries. In addition, stilbene synthesis in grapes depends on various viticultural factors such as the cultivar, the environment, and cultural practices. Bavaresco (2003) reported that stilbene concentration in grapes was positively correlated with vineyard elevation.

Some interspecific hybrid genotypes consistently showed low levels of UV-induced resveratrol (<10 μg g⁻¹ FW at early bloom, and <50 μg g⁻¹ FW at 30 days after full bloom). This is possibly because of recurrent crosses between resistant American interspecific hybrids and susceptible European vinifera cultivars to improve fruit quality. Furthermore, resistances of grapevines to gray mold and powdery mildew are polygenic traits (Eibach et al. 1989; Sbaghi et al. 1995). If the potential for resveratrol production is genetically controlled and specific to each grape genotype at the early bloom stage and at 30 days after full bloom, we can establish a simple classification system according to resveratrol production at the two stages, as follows: low and low (type A); low and high (type B); high and low (type C); high and high (type D). For example, Muscat Bailey A is a useful breeding material with high resveratrol production potential (type D). Muscat Bailey A is highly resistant to gray mold infection (Goto and Aono 1981), and accumulates high levels of resveratrol because the genes associated with resveratrol biosynthesis, such as stilbene synthase and phenylalanine ammonia-lyase, are expressed at high levels in this cultivar (Takayanagi et al. 2004). For grape breeding, low levels of UV-induced resveratrol (<10 μg g⁻¹ FW) indicate that the genotype is susceptible to gray mold, and levels of less than 50 μg g⁻¹ FW indicate susceptibility to powdery mildew.