The presence of stilbenes in the Vitaceae was first reported by Langcake and Pryce in 1976. They observed that during Botrytis cinerea infection of detached V. vinifera leaves of different genotypes (Cabemet Sauvignon, Gordo, Sultana, Clare, Mtiller-Thurgau, Riesling x Sylvaner), the zone of apparently healthy cells adjacent to the advancing mar- gin of the lesion showed a bright blue fluorescence, when examined under low wave- length radiation (366 nm).
A similar blue fluorescence was evident within 24 hours of brief exposure (ca. 10 min) of healthy leaves to short wavelength (254 nm) radiation. The authors considered the phenomenon of potential interest, since it had been found previously that fluorescent compounds produced by plants, in response to infections, frequently had a direct or indi- rect relationship with host defence mechanisms (Clarke, 1973; Mansfield and Deverall,
1974). The major component responsible for the fluorescence was identified as trans- resveratrol (3 ,5,4 '-trihydroxystilbene). This was not detectable in non-irradiated healthy leaves, while in UV -irradiated ones it was present at concentrations between 50 and 100 I1g gfWo\ (up to 396 I1g gfWo\ in members of the Vitaceae other than V. vinifera). The trans-iso-
156 L. BA V ARESCO and C. FREGONI
mer alone was obtained when leaf extracts were protected from light, but in the absence of this protection light-induced isomerization of trans-resveratrol to the less fluorescent cis-form occurred during the isolation. The presence of resveratrol was observed also in grapevine leaves infected with downy mildew (Plasmopara viticola) and powdery mil- dew (Uncinula necator) and seemed to be an apparently normal constituent of lignified stem tissues (at concentrations up to 700 /lg gfw-1) in Miiller-Thurgau, so it was con- cluded that this stilbene was constitutive in stems, but induced in leaves. In vitro bioas- says for trans-resveratrol antifungal activity evaluation showed that it had generally low fungitoxicity, when compared to the commercially available compound p- hydroxystilbene, as it did not inhibit spore germination of different fungi at concentra- tions up to 200 /lg mL-1 (Table 6.1).
Table 6.1- Antifungal activities ofresveratrol, its derivatives and pterostilbeoe.
Substances Botrytis cinerea Plasmopara viticola**
___ ---::--___ S .... p_o_r_e ",-,g_er-::m-::i-:-o_at_io_n_t_e_st_* ___ Z_o_o-,sp-::ore release Zoospore motility
resveratrol > 200 > 200 > 200
a-viniferin 97 35 11
y-viniferin > 200
f;-viniferin 100
pterostilbeoe 18 - 24
> 100 19 4.5
Source: Langcake and Pryce, 1977; Langcake el al., 1979; Pezet and Pont, 1988.
• Concentrations (rg mIãl ) causing 50% inhibition of spore genoination .
> 100 12.5
2.3
• - Concentrations (fig mrl) causing 50% inhibition of release of zoospores from sporangia or of motility of zoospores after their release.
Further work, work by the same authors (Langcake and Pryce, I 977a; Pryce and Langcake, 1977) reported the presence of other compounds of the same chemical nature of resveratrol in grapevine, for which they proposed the trivial generic name viniferins.
They were synthesized, once again, in leaves infected by Botrytis cinerea or irradiated with UV light and not in healthy or non-irradiated ones. These compounds were found to be oligomers of resveratrol and were named as follows: f:-viniferin (dehydro dimer of resveratrol), a-viniferin (cyclic trimer), p-viniferin (cyclic tetramer) and y-viniferin (a more highly polymerized oligomer) (Fig. 6.1). In leaves infected with Botrytis cinerea, a-viniferin was the major antifungal compound detected (more than 50 /lg gfw-\ fol- lowed by f:-viniferin (ca. 10 /lg gfw -I). In addition to a-viniferin (20 /lg gfw -I) and £-
viniferin (15 /lg gfw -I), UV irradiated leaves contained p-viniferin (9 /lg gfw -1) and y- viniferin (80 /lg gfw-1); f:-viniferin was also detected in lignified stem tissues (ca. 500 /lg gfw-\
In subsequent experiments (Langcake et al., 1979) focused on phytoalexin production during the infection of Cabemet Sauvignon leaves by Plasmopara viticola, a previously unidentified stilbenic compound, trans-pterostilbene (3,5-dimethoxy-4' hydroxystilbene), was detected in extracts of infected leaves (ca. 22 /lg gfw-1) and UV irradiated ones (ca. 9 /lg gfw -1). in vitro tests showed that it was more fungitoxic than the viniferins (Table 6.1).
GRAPEVINE STlLBENIC COMPOUNDS 157
Aesveratrol (trans-4, 3', 5' -trihydroxy stilbene)
HO
OH HO
OH
[-Viniterin
(dimer of resveratroQ QH
HO OH
OH a-Viniferin (cyclic trimer of resveratrol)
HO ~~ A 'l: ~ - A OCH'
Pterostilbene 0 CH3 (trans-3.5~dimethoxy-4ã -hydroxy stilbene)
Figure 6.1. Stilbenic compounds first detected in Vitis species.
Like resveratrol and viniferins, pterostilbene was absent in healthy leaves. Moreover, it was not detectable either in the blue fluorescent zone surrounding lesions caused by Botrytis cinerea on grapevine leaves nor in the rotten area or in normal lignified stem tissues. Pterostilbene synthesis appeared therefore under metabolic control independent of viniferin production.
4.2. Biotic elicitors
Grapevine stilbenes behave as phytoalexins if the plant is attacked by fungi, such as grey mould Botrytis cinerea Pers.; downy mildew Plasmopara viticola (Berk. and Curt,) Berl.
and de Toni; excoriose Phomopsis viticola (Sacc.) Sacc.; berry rot Rhizopus stolonifer (Ehrenb. Fr.) Lind. The detectable stilbenes are, in this case, trans-resveratrol, £-
viniferin, a-viniferin and pterostilbene, which are produced in leaves and berry skins (Table 6.2). Unelicited leaves and berries do not synthesize stilbenes. Disease resistant genotypes (American species and interspecific hybrids) show a quick and high leaf and berry accumulation of phytoalexins after elicitation, while in susceptible plants (V. vinif- era), phytoalexin synthesis is slow and does not reach high' concentrations. An indirect evidence of elicitor activity by Oidium tuckeri in berry level was given by Piermattei et
158 L. BA V ARESCO and C. FREGONI
Table 6.2. Stilbenic compounds induced in Vitis spp. by biotic elicitors (from literature cited).
Organ Elicitor Stilbenic Content
comeound (Il~ gfw-I )
Leaves Botrytis cinerea trans-resveratrol 4.79-9.40
a-viniferin 26.70-72.40
E-viniferin 9.59-15.83
Plasmopara viticola trans-resveratrol 10
E-viniferin 100
trans-pterostilbene 22 Bacillus sp. trans-resveratrol 31.06-78.30
Berry skins Botrytis cinerea trans-resveratrol 4-7
Fruit flesh Botrytis cinerea trans-resveratrol < 0.10 Undeseeded berries Botrytis cinerea trans-resveratrol 1.32-6.56
E-viniferin 2.26-10.30
trans-pterostilbene 0.10-0.24 Whole berries Rhizopus stolonifer trans-resveratrol 18
trans-eterostilbene 15
af. (1999), who detected more trans-resveratrol in Sangiovese wine from infected than from healthy berries.
4.2.1. Botrytis cinerea
After the first detection of resveratrol in Botrytis cinerea infected grapevine leaves, the distribution of the stilbenic compound in leaf lesions caused by the fungus was investi- gated (Langcake and McCarthy, 1979). Starting from the observations by Langcake and Pryce (1976), who had found that fluorescence caused by resveratrol was mainly concentrated in lesion surrounding areas, the lesion itself was subdivided into three concentric zones and their respective resveratrol contents were measured.
Actually, resveratrol was present predominantly in the apparently healthy tissue sur- rounding the lesion, the greatest concentration being in the first 5 nun outside the rotten area. In some of the leaves examined, the presence of resveratrol was also detected in the apparently healthy tissue portions further away from the lesion, corresponding to the observation that UV irradiation frequently showed streaks of blue fluorescence radiating from the lesion, predominantly along the major veins. The relationship between the amount of resveratrol produced and the resistance of leaves to fungal invasion was also investigated, taking lesion diameter as a measure of susceptibility. Leaves of different ages (different positions along the shoot of greenhouse-grown plants of V. vinifera culti- vars) were infected with Botrytis cinerea and incubated in order to allow lesions to de- velop. The fluorescent tissue surrounding the lesions was then excised and processed for resveratrol extraction. Susceptibility to B. cinerea (lesion diameter) was found to de- crease with leaf age, while resveratrol content of the excised tissue portions (all of the same area) increased markedly with leaf age. Similar relationships were obtained when resveratrol concentrations were expressed on a fresh weight basis. Resveratrol contents
GRAPEVINE STILBENIC COMPOUNDS 159 in infected leaves of different cultivars were compared and a clear negative correlation between resveratrol synthesis around the lesions and susceptibility to Botrytis cinerea was demonstrated. As resveratrol had been shown to have a low fungitoxicity, the au- thors concluded that the most likely explanation of the relationship between resistance and resveratrol concentration was that resveratrol functions as a precursor of the more fungitoxic viniferins.
The above evidence was confirmed by Langcake (1981), who found the following values for a 10 mm wide zone surrounding the rotted area caused by Botrytis cinerea on V vinifera leaves (Ilg gfw-1, from the youngest to the oldest tested leaf): resveratrol 9.40- 4.79; E-viniferin 9.59-15.83; a-viniferin 26.7-72.4. Pterostilbene was not detected. When the concentration of a-viniferin (Ilg gfw-I ), by far the most abundant compound, was plotted against susceptibility (Ilg of fungal glucosamine per lesion), an inverse but non linear relationship was evident. However, the results fitted well for an inverse relation- ship between 10glO a-viniferin concentration and susceptibility:
10glO a-viniferin concentration in the surrounding zone = 1.911 - 0.038 (susceptibility) where,
correlation coefficient, r = 0.814
The same experiments conducted on the B. cinerea resistant genotype V riparia demonstrated that qualitatively the response of V riparia was similar to that of V vinif- era, but quantitatively there were differences in the concentrations of stress metabolites (which were up to 10 times greater in V riparia) and in the relative amounts of the stress metabolites. Whereas non-spreading lesions on V vinifera were characterized by rela- tively large amounts of a-viniferin, with only small amounts of resveratrol and E- viniferin, the non-spreading lesions on V riparia produced predominantly E-viniferin, although substantial amounts of resveratrol and a-viniferin were also produced. Blaich et al. (1982) confirmed that American grapevines and interspecific hybrids generally pro- duce larger amounts ofresveratrol than the European varieties (V vinifera) as a response to Botrytis cinerea attacks or to an abiotic elicitor (such as mucic acid) and that, in addi- tion, they synthesize it much more rapidly, thus indicating the possibility that all the Vi- taceae are able to synthesize stilbene phytoalexins, but with differences concerning the induction and the regulation of synthesis.
To assess this general potential, leaves of different Vitis spp. were induced by UV ir- radiation (Dercks and Creasy, 1989a) and B. cinerea (Luczka, 1982) and consequent resveratrol synthesis measured. A similar induction experiment was conducted on differ- ent cultivar grape berries elicited by UV light (Creasy and Coffee, 1988). Although not comparable, the independent data obtained have a fundamental feature in common (Dercks et al., 1995): whereas Vitis spp. with "intermediate" stilbene production poten- tial show variable results, which are not easy to interpret, the order of species with high or low potential is never reversed. For instance, "Castor" or V. rupestris B-38 always exhibit a high potential whereas' V. vinifera cvs always show a low capacity for stilbene synthesis.
160 L. BAVARESCO and C. FREGONI
All the first experiments involving stilbene induction were conducted on leaves from greenhouse plants, which had been grown under the same conditions, in order to obtain the most homogeneous leaf material. Jeandet et at. (1995a), starting from the evidence that laboratory conditions are highly favourable to B. cinerea development, but not in- dicative of grapevine reaction to the fungus attack under natural conditions, tried to evaluate the plant defence mechanisms against B. cinerea colonization in the vineyard.
Grapes of Pinot noir, Gamay and Chardonnay were examined for resveratrol content just before (apparently healthy fruits) and after B. cinerea infection in the vineyard (ap- proximately 10% infected grapes). In all cases, resveratrol was found to be synthesized particularly in the skin cells (ca. 4-7 f.lg gfw-J skins), whereas only traces of resveratrol could be detected in the fruit flesh (less than 0.1 f.lg gfw-J), as already pointed out by Stein (1984). However, resveratrol constantly reached 1 f.lg gfw-' in the seeds_ Also, ap- parently healthy grapes had resveratrol present in the fruit skins. Similarly to what had been previously observed in leaves (see above), the major concentration in berry skins was detected in the non-infected fruits close to the necrotic area of each cluster. Never- theless, in the fruits far from the periphery of the lesions there was almost as much res- veratrol as in the ones close to the necrotic area_ The authors proposed three hypotheses:
~ Cracks in the cuticle are normally distributed on the fruit surface. B. cinerea hyphae can penetrate through these micro lesions and cause a phytoalexin response prior to any detectable lesion (Bess is, 1972)_
~ Phytoalexin synthesis in ripe grapes may be the consequence of quiescent infections caused by B. cinerea on young fruits and remained in state of latency, as reported by Pezet and Pont (1986).
~ Phytoalexin synthesis in parts of the host, distant from the infection site, may be induced by systemic signals synthesized by the plant in response to pathogen attack_
At the fruit level, an experiment was set up by Bavaresco et al. (l997a) to study the time course of phytoalexin (resveratrol, e-viniferin, pterostilbene) accumulation in ber- ries of two grapevine genotypes, Castor and V. vinifera cv Huxelrebe, infected by Botry- tis cinerea. Castor is considered to be resistant to B. cinerea, while Huxelrebe is suscep- tible_ Berries were sampled and infected with Botrytis cinerea at three different stages of growth (25 d after fruit set, at veraison, and during ripening)_ Stilbene phytoalexins elic- ited decreased during berry development, showing very low levels at ripening (expressed in f.lg gfw-' ofundeseeded berries: from 6.56 to 1.32 for trans-resveratrol, from 10.30 to 2.26 for e-viniferin, from 0.24 to 0.10 for pterostilbene). On average, berries of the Bo- trytis-resistant genotype Castor synthesized more trans-resveratrol (5 f.lg gfw-I ) and e- viniferin (6.6 f.lg gfw-') and at a higher rate as compared with the susceptible variety Huxelrebe (2.1 and 3.3 f.lg gfw-', respectively), while the amount ofpterostilbene showed no significant difference between the two cultivars. The decreasing stilbene concentra- tion from veraison to maturity could be explained by substrate competition between chalcone synthase and stilbene synthase, the former being involved in the flavonoid syn- thesis and the latter in the stilbene biochemical pathway (Jeandet et at., 1995b) or to a detoxification ofphytoalexins and/or a laccase-like stilbene oxidase activity (see below).
GRAPEVINE STILBENIC COMPOUNDS 161 Creasy and Pool (1999) also found decreasing resveratrol contents in Pinot noir berries sampled in the vineyard at different times during ripening.
The mode of action of resveratrol and pterostilbene on Botrytis cinerea cells was studied by Pezet and Pont (1995). Light microscope observation of conidia placed on media containing different concentrations of hydroxystilbenes showed a structural trans- formation and, sometimes, an outflow of cytoplasmic matter. Ultrastructural observa- tions revealed a complete cell disorganization. Inhibition of conidial respiration took place. From a chemical point of view, it must be stressed that in phenolic compounds an electron movement inside the molecules results in alternating single and double bonds.
This delocalization of electrons enhances the polarization of the molecules, when the substituents are electron-attracting groups. The more electron-attracting are the stilbenes, the more toxic are their effects, due to their affinity with proteins of fungal cell mem- branes. The low fungitoxicity of resveratrol, as compared with that of pterostilbene, is probably due to its hydrophilic character incompatible with the lipophilicity of biological membranes.
Botrytis cinerea produces polyphenol oxidases (laccase-like enzymes), whose in vitro function in the oxidative detoxification of resveratrol and pterostilbene produced by de- fence reactions of the host plant has been demonstrated (Hoos and Blaich, 1988 and 1990; Pezet et aI., 1991; Jeandet et aI., 1993; Adrian et aI., 1998; Breuil et aI., 1999), though laccase synthesis is a necessary but insufficient requirement for infection by Bo- try tis cinerea (Bar Nun et aI., 1988). Sbaghi et al. (1996) found that the ability of eight Botrytis isolates to degrade resveratrol and pterostilbene was positively correlated to their pathogenicity to in vitro cultures of V rupestris. A successful defence strategy of the plant may thus depend both on speed and intensity of stilbene production. This is probably the reason why vintages, when Botrytis infections are favoured by climatic conditions, do not always give grapes with high resveratrol concentrations (Jeandet et aI., 1995c; Celotti et aI., 1998). Consequent resveratrol concentrations in musts and wines (derived from contact with grape skins) are determined by the balance between plant production and pathogen degradation.
4.2.2. Plasmopara viticola
According to Langcake (1981), the differences between V riparia and V vinifera (re- spectively resistant and susceptible to downy mildew) in the accumulation of stilbenic stress metabolites after inoculation with Plasmopara viticola were very pronounced. In V vinifera, resveratrol accumulated mostly but reached only approximately 10 flg gfw-!
in leaves. In V riparia, however, the stress metabolites accumulated much more rapidly and to a greater extent. The predominant component in V riparia (as in the case of B.
cinerea infection) was E-viniferin, which reached a concentration of ca. 100 flg gfw-!.
Concentrations of resveratrol were only slightly lower. Alpha-viniferin was produced in notably smaller amounts although its concentration greatly exceeded that in V vinifera.
Only traces of pterostilbene were found.
By screening 17 Vitis spp. exhibiting different Plasmopara viticola resistance (the
162 L. SA V ARESCO and C. FREGONI
visible symptom of the resistance reaction is necrosis of plant tissue indicating the ability to suppress the fungus reproduction), Dercks and Creasy (l989a) found that four groups of plants could be described, with respect to the correlation between their potential for phytoalexin production in response to UV irradiation and their resistance:
y Group 1: High phytoalexin production and high resistance (cv Castor, cv Pollux, V riparia, among others).
y Group 2: Intermediate phytoalexin production and moderate resistance (V rupestris, V andersonii, among others).
y Group 3: Intermediate to low phytoalexin production and susceptibility (cvs Mliller- Thurgau, Chardonnay, Riesling, V treleasei, among others)
y Group 4: Low phytoalexin production + resistance (V cinerea, V champini).
The existence of genotypes belonging to the fourth group suggests that resistance may, in some cases, be associated with factors other than stilbene phytoalexins. Never- theless, high phytoalexin accumulation was never associated with susceptibility, support~
ing the involvement of stilbenes in Plasmopara viticola resistance (similar conclusions were drawed by Sbaghi et al. (1995) when testing 13 Vitis spp. for their Botrytis cinerea resistance). The resistant genotypes (i.e. cv Castor), accwnulated five times the maxi- mum resveratrol concentration of the susceptible cv Riesling and maintained these high levels during pathogenesis, whereas the resveratrol concentration in cv Riesling de- creased. The reproduction rate of Plasmopara viticola was directly related to inoculwn density. By increasing the number of sporangia applied, a loss of resistance in cv Castor was caused. This shows the importance of taking into account the severity of fungal challenge when assessing the comparative degree of phytoalexin response and corre- sponding level of disease resistance.
4.2.3. Phomopsis viticola
Hoos and Blaich (1989) found that growth on agar medium of Phomopsis viticola colo- nies starting from pieces of myceliwn was slowed down by resveratrol concentrations
<50 ppm and stopped by higher concentrations. Thus, it is possible that plants produce resveratrol, as a defence reaction, when attacked by this fungus, even if at present there is no in vivo evidence ofthis fact.
4.2.4. Rhizopus stolonifer
The fungus Rhizopus stulunifer is the main cause of post-harvest decay of table grapes in Israel (Sarig et at., 1997). The high levels of inoculum present in Israeli vineyards and their environment and the severe necrotrofic nature of this fungus make it more damag- ing than Botrytis cinerea. A negative correlation was found between the potential for resveratrol accumulation (estimated after elicitation by UV irradiation) in different table grape cultivars and their susceptibility to decay caused R. stolonifer. Phytoalexin accu- mulation declines with advancing maturity of the berries, concomitant with their increas- ing susceptibility to decay. The maximum resveratrol and pterostilbene levels detected in cv Perlette berries, following inoculation with R. stolonifer, were ca. 18 f.lg gfw-1 and 15