Effect of Cladosporium rot on the composition and aromatic compounds of red wine

High incidences of Cladosporium rot frequently occur (over 50% infected berries per clusters) in delayed harvests in Cabernet Sauvignon and other red cultivars of Vitis vinifera in Chile. The objective of this study was to determine the effect of Cladosporium rot (Cladosporium cladosporioides and C. herbarum) on composition and aroma of red wine. Cabernet Sauvignon and Carménère wines made with 50% of Cladosporium rot infected grapes (CRG) were characterized and compared with wines made identically with apparently healthy grapes (AHG). Wine composition, color, and aroma were determined and the wines were subjected to sensory evaluation. Differences in volatile acidity, residual sugar, anthocyanins, color hue, and tannins between CRG wines and AHG wines of both Cabernet Sauvignon and Carménère were significant (p<0.05). In Cabernet Sauvignon, differences in the aromatic profiles were obtained between CRG wines and AHG wines. Panelists negatively differentiated the aroma and astringency of the CRG wines. Cladosporium rot reduced color, aroma, and flavor in Cabernet Sauvignon and Carménère wines. To our knowledge, this is the first report demonstrating the effect of Cladosporium rot, a common late fall disease of grapevine in Chile, on red wines, and the results suggest the need to prevent it for obtaining high quality wines. Additional key words: aromas, Cladosporium spp., color, delayed harvest, fungal disease.


Introduction
Cladosporium rot occurs frequently in wine grapes (Vitis vinifera L.) in the Central Valley of Chile.It is particularly severe (over 50% infected berries per clusters) when growers delay harvest considerably in order to optimize phenolic composition to assure the best possible wine quality (Pszczólkowski et al., 2001;Briceño and Latorre, 2007).Therefore, it mainly occurs in overripe berries of red cultivars, e.g.Cabernet Sauvignon, Carménère, and it is characterized by an olive green mold that invades the grape skin superficially (Briceño andLatorre, 2007, 2008).In Chile, it has been associated to Cladosporium herbarum (Pers.:Fr.)Link and C. cladosporioides (Fresen.)De Vries (Mujica and Vergara, 1980).
The quality of the wines is highly dependent on the grape composition at harvest (Ribéreau-Gayon and Peynaud, 1960).Therefore, the ripeness of the pulp, optimal ratio between the total soluble solid content (TSS) and the titratable acidity (TA) (Barceló, 1997), is recognized as a critical factor for quality wines.Consequently, grapes for winemaking have been harvested in accordance with these parameters (TSS > 22.5% and TA=6-7 g L -1 of tartaric acid) in Chile (Pszczólkowski et al., 2001).However, this harvest criterion does not necessarily coincide with "phenolic ripeness" of seeds and skins, needed for obtaining quality red wines (Saint-Cricq et al., 1998).Therefore, growers postpone harvest waiting for a softening of the tannins from the seeds and skins.This has resulted in an excessive delay in harvest dates, with a negative impact on the quality and phytosanitary condition of the grapes.A slow or incomplete fermentation frequently occurs in musts made with late harvest grapes, and it has been postulated that a considerably loss of color, flavor, aroma, and other sensorial attributes occur associated to Cladosporium rot (Pszczólkowski et al., 2001).Therefore, the objective of this study was to determine the effect of Cladosporium rot on the physicochemical and volatile composition of these red wines.

Material and methods
Winemaking.Three and four grape lots, each of 15 kg per cultivar, were harvested in May when Cladosporium rot incidence was above 50% (CRG) in a commercial Cabernet Sauvignon and Carménère vineyards in 2006 and 2007, respectively.An equal number of lots of apparently healthy grapes (AHG) were harvested as controls in the same vineyards.Each grape lot was made into wine in separate 20 L polyvinyl chloride (PVC) containers.The Cladosporium rotted grapes used in this study exhibited an olive green superficial mold caused by C. herbarum and C. cladosporioides.
Cabernet Sauvignon and Carménère bunches were destemmed, manually crushed, and fermented in the presence of skins and seeds.The nitrogen content of the musts was corrected with 300 mg L -1 ammonium phosphate and then treated with 20 mg L -1 pectolytic enzymes (Rapidase® ex color, Gist-brocades, Beverages Ingredients Group, France) and 30 mg L -1 sulphur dioxide (SO 2 ).The musts were homogenized, and the temperature and density were determined before the addition of 200 mg L -1 of Saccharomyces cerevisiae (Anchor WE 372, Anchor Bio-technologies, Eppindust, South Africa), which was previously hydrated at 37ºC for 30 min.
The containers were sealed with water valves to maintain semi-anaerobic conditions, and fermentation took place in an isothermal fermentation chamber at 28-30ºC.Skins were pushed down daily and the temperature and density of each wine were determined.At the end of fermentation, the free run wine was transferred into 5 L glass bottles that were kept in the isothermal chamber (20-24ºC) until the end of malolactic fermentation (MLF), which was determined by paper chromatography (Ribéreau-Gayon and Peynaud, 1962).The wines obtained were decanted into clean containers and treated with 35 mg L -1 SO 2 .The wines were then stored at 0ºC for 30 days and then the free SO 2 was corrected to 35 mg L -1 (Buechsenstein and Ough, 1978).
Basic analysis of must and wine.The TSS content, pH, and TA of musts were determined.Analytical determinations of alcohol, TA, pH, volatile acidity and reducing sugars were assessed in the wines (Bordeu and Scarpa, 1998).
Analysis of the volatile fraction.The volatile fraction of Cabernet Sauvignon wines was determined and quantified using a gas chromatograph (Hewlett-Packard 6890 Series, USA) coupled with a Hewlett-Packard 5972 mass detector (GC-MS).Samples (100 mL) of bottled wines were subjected to a double liquid-liquid extraction with 25 mL of dichloromethane.They were mixed for 30 min at 0°C in a nitrogen atmosphere and centrifuged between each extraction (Sorvall®RC 285, Dupont, USA) at 5000 rpm for 20 min to separate the organic phase, which was dried over anhydrous Na 2 SO 4 .Prior to GC-MS analysis, the extracts obtained were concentrated to 350 µL in a Dufton column in a thermo-regulated bath (Julabo Labortechnik, GMBH, Germany) at 45ºC.As an internal standard, 4-nonanol (100 µL of a 3.5 mg L -1 solution) was used as a response factor equal to 1.A DB-WAX ETR column (J&W Scientific, Folson, CA, USA) of 60 m x 0.25 mm x 0.25 µm film thickness was used.The temperature of the injector was 180°C and the GC oven was initially regulated at 40°C for 5 min, increasing 3°C per minute and then 25 min at 240°C.Samples (2 µL) were injected to the chromatograph in splitless mode with an helium flux of 1.9 mL min -1 .The compounds obtained in each chromatogram were compared with the information from the NIST-EPA-NIH mass spectral database (Gaithersburg, MD, USA), which is composed of 130,000 spectra.Each wine sample was processed in triplicate.
Odor activity values (OAVs) were calculated by dividing the mean concentration of each aromatic compound by its odor threshold value, and it was used to estimate the sensory contribution of the aromatic compounds to the overall wine flavor.The OAVs in Cabernet Sauvignon wines were calculated on the basis of odor threshold values reported previously (Guth, 1997;Ferreira et al., 2000Ferreira et al., , 2002;;López et al., 2002;Culleré et al., 2004;Peinado et al., 2004;Selli et al., 2004;Gómez-Míguez et al., 2007).
Sensory evaluation.The sensory evaluation was performed in one session in individual booths.Each replicate was presented as pairs of wines with both treatments in random order for each of the cultivars.The samples were served at room temperature (18-20ºC) in glasses identified with a random 3-digit number and covered with a polyethylene cover.The sensorial tasting panel was composed of 11 panelists (oenologists) with no previous training, 4 women and 7 men, between 25 and 60 year old, and a set of parameters describing appearance, aroma and mouth-feel was evaluated on a graphical scale of 10 cm length, where 0 and 10 represented low and high perception of each parameter respectively.
Statistical design and analysis.The effect of vintage year and Cladosporium rot on the musts and wines, phenols and the volatile fraction of the wine were studied by two-way analysis of variance (ANOVA) with four replicates, considering vintage year as main factor and health grape status (with and without visible Cladosporium rot symptoms) as sub-factor.The SigmaStat® 3.1 software (Systat Software Inc., San José, CA, USA) was used.
The results of the sensory evaluation were analyzed separately for each of the parameters using a two way analysis where the type of wine (made with or without Cladosporium-rotted grapes) was the main factor and panelists the sub-factor.Data were analyzed with the aid of SigmaStat® 3.1.

Results
Basic composition of must and wine.Cladosporium rotted berries were significantly (p=0.05)lighter than apparently healthy berries with a mean weight reduction of 14.3% in Cabernet Sauvignon and 12.5 to 25% in Carménère.Significant differences in TA, and pH were obtained among must prepared with diseased berries and apparently healthy berries, but differences in TSS content were only significant in Carménère.In Carménère must, the interaction between vintage year and berry weight, TSS and pH were significant (p<0.041),while in Cabernet Sauvignon must, only the interaction between vintage year and TA was significant (p<0.026),meaning that differences in the above mentioned parameters were significant only one season (Table 1).
In both cultivars, Cabernet Sauvignon and Carménère, the volatile acidity and reducing sugars were significantly higher in CRG wines than AHG wines.Moreover, ethanol content was significantly different in Carménère wines but not in Cabernet Sauvignon.The significant interaction in residual sugars again reflects differences only one season.Wine acidity and pH results were not clear.
Color and phenolic composition of wine.Color hue and total tannins were always higher in CRG wines than in AHG wines.Total phenols were significantly higher in CRG wines than AHG wines prepared with grapes cv.Carménère.Except for Carménère in 2007, the CRG wines were characterized by significantly lower anthocyanin content than AHG wines (Table 2).
Analysis of the volatile fraction.A total of 93 and 130 volatile compounds including acids, alcohols, alde-hydes and ketones, esters, furans, lactones, norisoprenoids, phenols, sulphur derivatives, and vanillin derivatives were identified in wine samples in 2006 and 2007, respectively (Table 3).
Cladosporium rot significant reduced the concentration of decanoic acid in CRG wines in both vintage years.Among alcohols, isoamyl alcohol and 2phenylethanol had the highest concentrations but the effect of Cladosporium was not significant.Other alcohols were significantly (p<0.001)altered in CRG (Table 3).Among ten ester compounds, eight were significantly altered in both vintage years.For instance, isoamyl acetate, ethyl 3-hidroxybutirate and phenyl acetate were significantly higher in AHG wines than CRG wines.Higher concentration of furan compounds were found in CRG wines than AHG wines, but significant effect of Cladoporium rot was only determined for furfural in 2006.Norisoprenoid compounds were only detected in 2007 and their concentration significantly increases in CRG wines.Phenol compounds were always higher in CRG wines than AHG wines but, only phenol, and 3,4,5-trimethoxy-phenol were consistently altered in both vintage years.Benzaldehyde, ethyl vanillate and γbutyrolactone were significantly higher in CRG wines than AHG wines in both vintages years (Table 3).
OAVs were calculated for 49 of the 167 compounds, but only seven compounds had OAVs higher than one and significant differences between AHG wines and CRG wines in both vintage years (Table 4).With the exception of isoamyl acetate and 2,3-butanedione, the effect of CRG on the other compounds was dependant significantly (p<0.009) on the vintage year.The interaction between vintage year and Cladosporium rot was only significant (p=0.048) in relation to γ-butyrolactone (Table 4).
Sensory evaluation.In Cabernet Sauvignon wine, all parameters were significantly affected by Cladosporium rot (p<0.045),except for acidity (Table 5).In Carménère wine, the effect of Cladosporium rot was significant only in relation to aroma and tannins.In both cases a significant interaction with panellist was found in several parameters, showing some discrepancy in the panel.
In Cabernet Sauvignon, color, aroma and acceptability were found to be considerably lower in CRG wines than AHG wines, while sweetness, alcohol, acidity and astringency were higher in CRG wines than AHG wines.In Carménère aroma was lower in CRG than AHG wines and astringency was higher in CRG wines than AHG wines (data not shown).

Discussion
On the basis of the results obtained in this study, the delayed harvest of Cabernet Sauvignon and Carménère,  (Pszczólkowski et al., 2001;Briceño and Latorre, 2007).Confirming previous results (Pszczólkowski et al., 2001;Briceño and Latorre, 2007), the delayed harvest was associated with berry dehydration, 12.5 to 25.0% berry weight losses.The relatively high TSS contents (> 23.4%) increased the risk of delayed must fermentation.In this study, fermentation was delayed between 24 h in Carménère in 2007 and 96 h in Cabernet Sauvignon in 2006 in relation to AHG.These differences in the fermentation period can result from differences in TSS contents but it is possible that the high proportion of Cladosporium rotted berries could affect yeast performance.However, further research is needed to verify this subject.Independently of the grape cultivar, wines produced with 50% Cladosporium rotted grapes (CRG) consistently had higher residual sugars (9.2 to 61.7%) than AHG wines in both vintage years, reflecting incomplete fermentations.
Important changes in phenolic composition occur during grape ripening that improve red wine quality (Kennedy et al., 2000(Kennedy et al., , 2006;;Harberston et al., 2002;Adams, 2006;Kennedy, 2008).On the basis of this knowledge, Chilean growers considerably delay harvest of Cabernet Sauvignon and other red cultivars until late in autumn (April-May) waiting for the best possible phenolic composition of berries.However, this delay often implies that Cladosporium infected grape lots are used for winemaking.Total phenols were significantly higher in CRG wines only in Car-ménère and expected differences between seasons were found in Cabernet Sauvignon (Kennedy, 2008).
Berry dehydration and low pulp-seed ratio were possible responsible for the higher tannins and phenol contents found in wines elaborated with grape lots having 50% Cladosporium rotted berries and it may explain the high astringency detected by panelists in the sensory analysis.In spite of this lower pulp-skin ratio, the anthocyanin contents in CRG wines were lower than in AHG wines, confirming previous results on Cabernet Sauvignon elaborated with heavily mold infested grapes (Pszczólkowski et al., 2001).This relatively low anthocyanin contents and the high hue values found in the CRG wines, corresponding to an evolution towards a red brick color was also detected by panelists in Cabernet Sauvignon wines.
It is probable that this color change results from the physical and chemical skin damages caused by Cladosporium spp. that colonized the skin surface of the grapes.However, additional research is needed to probe this hypothesis.Nevertheless, similar results have been reported for wines made grape lots infected with powdery mildew (Erysiphe necator), where the loss of red color and low anthocyanin content was mainly due to a low concentration of delphinidin, cyanidin, petunidin, peonidin, and malvidin (Amati et al., 1996;Piermattei et al., 1999).
In both Cabernet Sauvignon and Carménère, the sensory panel detected a considerable loss of aromas in CRG wines.In Cabernet Sauvignon wine, it is possible that relatively low contents of compounds such as 2phenyl ethanol, isoamylacetate, and ethyl hexanonoate, found in the aroma profile, reduced the floral and/or fruity aromas.Additionally, the CRG wines had a larger number and higher concentration of aromatic phenolic compounds than AHG wines.Some of these phenolic compounds contributed aromas like smoke, phenol, and medicine (e.g.guaiacol, p-cresol) that could be considered negative because they may mask favorable aromas in Cabernet Sauvignon wines (Ferreira et al., 2000(Ferreira et al., , 2002;;López et al., 2002).
The concentration of esters like monoethyl succinate, isoamyl acetate and ethyl 3-hydroxybutyrate were in lower concentrations in CRG wines than AHG wines which may affect the fruity aroma.In both vintage seasons, γ-butyrolactone was significantly lower in CRG wines AHG wines.This could be an undesirable factor considering that γ-butyrolactone confers pleasant aroma with floral and sweet attributes (Selli et al., 2004).Even if compounds with pleasant aromas like ethyl vanillate, benzaldehyde and furans were relatively high in CRG, these compounds had OAV<1, therefore, they are not important in their olfactory properties.
With the exception of octanoic acid, most compounds with important olfactory properties (OAV>1) and significant differences between CRG wines and AHG wines have positive sensory properties.Therefore, the lost of fruity aroma associated to Cladosporium rotted grapes in Cabernet Sauvignon wines could be primarily due to a loss of esters (e.g.isoamyl acetate), lactones (e.g.γ-butyrolactone), ketones (e.g.2,3-butanedione), and alcohols (e.g.2-phenylethanol).Previous studies have demonstrated the loss of varietal aromas in wines elaborated with grapes infected with powdery mildew, however, the compounds affected were not determined (Calonnec et al., 2004).
Finally, it was interesting that some volatile phenol compounds (i.e.3,5 dimethoxy phenol, p-cresol) and furan compounds (2,5 dimethyl furan) were found only in wines made with 50% Cladosporium rotted grapes.It is unknown whether these are unique compounds associated to C. cladoporioides and/or C. herbarum.Additional investigations are needed in order to confirm this and to determine the effect of these compounds on the sensory qualities of red wines.
On the basis of the results obtained, the use of grape lots having 50% Cladosporium rotted grapes, caused by C. cladosporioides and C. herbarum, negatively affected Cabernet Sauvignon and Carménère wines, reducing color, aroma, and flavor significantly.An important reduction in fruity and floral aromas was obtained in Cabernet Sauvignon wines.Additionally, Cladosporium rot considerably reduced berry weight and so yields of Cabernet Sauvignon and Carménère grapevines.

Table 1 .
Effect of the vintage year and Cladosporium rot on berry weight and must composition of Vitis vinifera cvs.Cabernet

Table 2 .
Effect of vintage year and Cladosporium rot on the composition of Cabernet Sauvignon and Carménère wines in 2006 and 2007 vintages

Table 3 .
Effect of vintage year and Cladosporium rot in aroma profile obtained by gas chromatography-mass spectrometry in Cabernet Sauvignon wines in 2006 and 2007 vintages

Table 4 .
Aroma descriptor, odour threshold and effect of vintage year and Cladosporium rot on odour activity values (OAV) of compounds in Cabernet Sauvignon wines in 2006 and 2007 vintages.Mean of four infield replicates

Table 5 .
P values showed the effect of panellists and Cladosporium rot on sensory evaluation of wines Cabernet Sauvignon and Carménère.
Data analysed by two-way analysis of variance.