Influence of Enological Practices on Ochratoxin A Concentration in Wine

Wines.

Seven 100-kg lots of Falanghina (Vitis vinifera L.) grapes from different vineyards located in the Taburno DOC area (Campania region, Italy) were used for the production of white wines. Each lot was processed separately in a small-scale experimental winery. Grapes were de-stemmed, crushed, and 10 g/hL of potassium metabisulfite was added. The crushed grapes were pressed into musts, 2.0 g/hL of pectic enzyme (Lallzyme HC, Lallemand Inc., Castel d’Azzano, Italy) was added, and the mixture allowed to settle for 24 hr at 14°C. After clarification, musts were racked and inoculated with 30 g/hL of Saccharomyces cerevisiae strain 71 B (Lallemand). Fermentations took place in 140-L stainless steel tanks at 14°C. At the end of alcoholic fermentation (residual sugars <2.0 g/L), wines were racked with 8.0 g/hL of potassium metabisulfite, cold-stabilized for one month at 10°C, and bottled. Four white wines were obtained from Falanghina grapes left to dry for two months in an aerated room, adopting the winemaking procedure previously described.

Seven 100-kg lots of Aglianico (Vitis vinifera L.) grapes from the Taburno DOC area were used for the production of red wines. Grapes were destemmed and crushed. The must was treated with 6.0 g/hL of potassium metabisulfite. Fermentation started with indigenous yeast and the cap was immersed twice a day. Fermentation temperature was 26°C and maceration of the pomace lasted 10 days. Successively, the pomace was pressed (~8 Atm), giving ~65 L of wine. The effect of pressing on the concentration of OTA in wine was evaluated on four Aglianico red wines produced with the same red winemaking procedure previously described, except that must pressing was carried out at ~80 Atm.

In order to evaluate the influence of filtration and heating, three samples of Aglianico wine (4 L each) were divided into four aliquots. The first aliquot was filtered through a 0.45-μm membrane (Beco-Membran’PF Begerow, Milano, Italy), the second was filtered through a 10-μm membrane (SEITZ-PREcart PPII, Scott Labs, Petaluma, CA), and the third was heated on a hot plate at 55°C before being filtered through a 10-μm membrane (SEITZ-PREcart PPII). The fourth aliquot was the control.

In order to evaluate the influence of hygienic conditions of the winery on wine OTA contamination, seven Aglianico red wines and seven Falanghina white wines were provided by local farmers producing wine for their own consumption. These wineries have the following features: the presence of molds in containers, high storage temperature (up to 28–30°C in summer), and generally poor sanitation (premises were not perfectly clean).

OTA extraction and HPLC analysis.

Ochratoxin A was extracted from 10 mL of wine and purified following an immunoaffinity clean-up procedure (Visconti et al. 1999). Quantification of OTA was performed by HPLC, as described by Visconti et al. (1999). HPLC analysis was carried out with a Shimadzu apparatus (Shimadzu Italy, Milan) consisting of a LC-10 ADVP, a SCL-10A system controller, two LC-A pumps, a fluorescence detector, an injection system full-loop Rheodyne model 7725 (Rheodyne, Cotati, CA) and equipped with a Bio-Sil C18 column (150 x 4.6 mm, 5.0-mm particle diameter) (Bio-Rad, Milan, Italy) with an injection loop of 100 μL.

Quantification of OTA was carried out by comparison to an external standard curve constructed by injecting six standard solutions containing OTA at concentrations from 0.50 to 10 μg/L. The calibration curve equation had a slope of 45730.53 ± 343.19 and r² = 0.9997. The limit of detection (signal-to-noise ratio 3:1) was 0.05 μg/L and was obtained by injecting standard solutions at increasing OTA concentrations (from 0.025 to 0.5 μg/L). The limit of quantitation (signal-to-noise ratio 10:1) was 0.5 μg/L. This amount did not take into account the 20-fold concentration of OTA in the wine extract as a consequence of immunoaffinity chromatography. The extraction procedure and HPLC analysis of each sample was carried out in triplicate.

The precision of the method used was tested by 10 replicate analyses on a white wine sample containing OTA at a concentration of 0.006 μg/L and a red wine sample containing OTA at 0.066 μg/L. The coefficient of variation was less than 6.7%, demonstrating good repeatability of the HPLC analysis. To validate the method, wine containing OTA at a concentration of 0.006 μg/L was fortified with OTA at concentrations of 5.0, 3.0, and 2.0 μg/L. Each fortified sample was divided into four aliquots and submitted to the extraction procedure and HPLC analysis. The recoveries were 77.75 ± 2.59%, 82.25 ± 1.71%, and 83.75 ± 1.89% at 2.0, 3.0, and 5.0 μg/L, respectively.

Fining agents and clarification.

Eleven fining agents were used at a dosage of 50 g/hL: natural polysaccharide, cellulose ester, PVPP (N-vinyl-2-pyrrolidone homopolimer), cellulose fiber, rind yeast, silica gel, bentonite, deodorant carbon, decolorizing carbon, high mesoporosity carbon, and high decolorizing carbon (all provided by Perdomini, Verona, Italy). Clarification experiments were performed as follows. Red wine (500 mL) was fortified with OTA at a concentration of 10 μg/L and a fining agent was added to fortified wine. The solution was mixed for 3 hr at room temperature and filtered through a 12-μm membrane (Schleicher & Schuell AE 100, Keene, NH). The experiment conducted with bentonite was performed as follows: bentonite was dispersed in water (1:20) overnight, then added to the wine sample (500 mL), mixed for 5 min at room temperature, and left to flocculate until complete sedimentation. The supernatant was filtered through a 12-μm membrane (Schleicher & Schuell AE 100).

For the experiments with different concentrations of enological carbon with decolorizing activity, samples of red wine containing pure OTA at concentrations of 5.0 μg/L were added to enological carbon at 5.0, 10, 20, and 30 g/hL, mixed, and, after 30 min, filtered on a 12-μm membrane. Each fining agent and clarification experiment was carried out in triplicate.

Polyphenol and volatile fraction analyses.

Total polyphenol content was carried out by Folin-Ciocalteu colorimetric analysis as described by Singleton and Rossi (1965), catechins as described by Margheri and Falcesi (1972), color intensity and tone as described by Glories (1984), and anthocyanins and tannin content as described by Ribereau-Gayon and Stonestreet (1965). Extraction and analysis of volatile compounds was performed as described by Moio et al. (2004).

Statistical analysis.

Analysis of variance and Tukey’s test were used to interpret differences in means, if any, at the 95% confidence level. Regression analysis was carried out to evaluate the effect of increasing dosages of carbon on wine OTA concentration. Elaborations were carried out by means of JMP 4 (SAS Institute Inc., Cary, NC).

Effects of enological practices on OTA.

Ochratoxin A levels in wines and the effect of some enological factors are shown in Table 1⇓. Concentration ranges for OTA were 0.006 to 0.022 μg/L and 0.017 to 0.050 μg/L for white and red wines, respectively. White wines obtained from grapes left to dry for two months had OTA levels more than two times higher than white wines obtained from grapes at full maturity. Intensive pressing of the pomace increased OTA concentrations in red wines, resulting in wines with a concentration of toxin approximately four times higher than red wines traditionally obtained. Analysis of homemade wines revealed that winery sanitation was the main factor influencing OTA contamination of wine. Wines from contaminated wineries showed higher OTA concentrations, ranging between 0.130 and 0.240 μg/L for white wines and 0.200 and 1.351 μg/L for red wines.

Following filtration, a decrease in the toxin content of about 80% was observed in red wine samples filtered through the 0.45-μm membrane (p < 0.05). Treatments filtered through the 10-μm membrane did not show a significant decrease in the toxin (p < 0.05). Regarding the pasteurization of wine, heating on hot plates at 55°C followed by microfiltration through 10-μm membranes did not affect OTA concentration (p > 0.05).

Adsorptive capacity of fining agents.

Among products evaluated for the clarifying treatment of contaminated wines, only enological carbon reduced OTA concentration (Figure 1⇓). Four different types of carbon were used: (1) deodorant; (2) decolorizing with a medium activity and activated by H₃PO₄; (3) decolorizing with a medium-high activity; and (4) decolorizing with a high activity and activated by H₃PO₄. The addition of deodorant carbon to wine at 50 g/hL, used in enology to remove abnormal odors, decreased the initial concentration of OTA by 68%. The other three types of enological decolorizing carbon showed a 96 and 98% decrease of initial concentration under similar conditions.

• Download figure
• Open in new tab

Figure 1

Effect of different fining agents on OTA concentration in red wine. Error bars represent the standard deviation of replicate measurements.

The effect of increasing dosage of decolorizing carbon on OTA content is reported in Table 2⇓. At a carbon concentration of 10 g/hL, OTA decreased to 46% of the initial level; for samples treated at 30 g/hL, OTA decreased to 72% of initial level. As expected, OTA concentration decreased as the carbon concentration increased. Polyphenol content and other color parameters of the samples treated with activated carbon at concentrations of 5.0 to 30 g/hL were not significant compared with the controls (Table 2⇓).

The effect of carbon with high decolorizing activity on volatile compounds in the experimental wines was also studied (Table 3⇓). Fifty-four compounds belonging to the chemical classes of alcohols (20), esters (11), acids (8), aldehydes (2), phenols (3), sulfur compounds (3), ketones (1), and furanes (1) were identified and measured. Addition of carbon at 10 g/hL caused a significant decrease in the concentrations of some principal wine volatiles such as 3-methyl-1-butanol, geraniol, 2-phenylethanol, 3-methylbutyl acetate, ethyl hexanoate, ethyl octanoate, and octanoic and decanoic acids. For many of these compounds, such as ethyl esters, acetates, and geraniol, further concentration decreases were observed at 30 g/hL of carbon.

Effects of enological practices on OTA.

Ochratoxin A values found in wines were consistent with data reported by Zimmerli and Dick (1996) and Soleas et al. (2001). The concentrations of OTA detected in the present study, however, were below the average values reported by Visconti et al. (1999) on 55 wines from southern Italy (Puglia region). This decrease may be due to the fact that in the last few years, the wide application of HACCP (Hazard Analysis and Critical Control Point) imposed processing practices, as well as the obligation to keep buildings and grounds in an orderly, hygenic, neat, and tidy condition, and the use of containers, equipment, and tanks clean and made of a suitable material, which guarantee hygenic conditions in Italian wineries (Italian Legislative Decree 155/97).

The higher OTA concentration detected in red wines with respect to white wines has been previously reported by several authors (Zimmerli and Dick 1996, Visconti et al. 1999). This phenomenon is likely linked to maceration, which may promote the release of OTA from grape skins. The role of grape skins as a carrier of OTA in wine was confirmed by the observation that intensive pressing of pomace increases OTA concentration in wine (Table 1⇑). Drying of grapes also increased the OTA concentration in the corresponding wines, resulting from the presence of OTA-producing fungi during the drying process. The data agree with those reported in the literature on Noble late-harvest South African wines (Stander and Steyn 2002) and Malaga wines from Spain (Zimmerli and Dick 1996). However, the concentration of berry constituents occurring during drying may have contributed to the increased OTA concentration observed in the final product.

Among the enological practices able to reduce OTA contamination in wine, microfiltration through a 0.45-μm membrane decreased contamination by about 80%. This reduction was likely a result of retaining the toxin by the filtration bed formed on a 0.45-μm membrane by wine macromolecules during treatment. Because 0.45-μm filtration is largely used in the production of white wines, this process may improve toxicological safety for winemaking.

Adorptive capacity of fining agents.

Consistent with Castellari et al. (2001), we observed that treatment with decolorizing carbon significantly decreased OTA concentration. Regression analysis indicates a decrease in OTA concentration with increased carbon dosage. In order to understand OTA adsorption to decolorizing carbon, the Freundlich adsorption isotherm was determined (Treybal 1955). The Freundlich model (x/m = K_FC^1/n) well fitted the experimental data (r² = 0.9886,) and the specific absorption capacity value, K_F, was 3.172 (mg/g)/(mg/L)^0.8474 (Figure 2⇓). The value of K_F was less than values reported by Castellari et al. (2001) when two different kinds of carbons were used to adsorb OTA from the wine; differences may be easily explained by the fact that the value of K_F varies greatly with the nature of carbon used (Chen et al. 1997). This isotherm describes adsorption where the adsorbate has a heterogeneous surface with adsorption sites that have different energies of adsorption. Sites having the same adsorption energy are grouped together into one patch (Duong 1998). Thus the observed OTA behavior at increasing carbon concentration was due to the presence of specific binding sites and their gradual saturation. The decrease in OTA sorption capacity of carbon may be also due to direct competition for adsorption sites between OTA and other wine components. Competitive adsorption may be one of the more important factors influencing the binding efficiency of activated carbon toward target organic contaminant (Newcombe et al. 2002).

• Download figure
• Open in new tab

Figure 2

Freundlich isotherm for the adsorption of OTA from wine by carbon. Dependent variable x/m is the amount of OTA that is adsorbed per mass of carbon (mg/g), and C is the concentration of the OTA in the solution that is in equilibrium with the carbon (mg/L).

In our study, no significant reduction in the concentration of polyphenols in the experimental wines was observed, indicating no interference of these compounds on OTA adsorption (Table 2⇑). However, a decrease in some important aroma compounds of wine was observed (Table 3⇑). Among these, the decrease of alcohols such as 3-methyl-1-butanol is interesting, because the concentration of this compound is negatively correlated to the aroma quality of wine (Etiévant 1991). However, the 3-methyl-1-butanol concentrations detected after treatment were higher than the threshold value (30,000 μg/L) (Ferreira et al. 2000). Treatment with carbon (30 and 50 g/hL) for wine decontamination resulted in losses of other volatiles, such as ethyl fatty acids and acetates, volatile fatty acids, and geraniol. Among the esters with concentrations decreased by carbon treatment, 3-methylbutyl acetate, ethyl hexanoate, and ethyl octanoate are of primary importance for the aroma of white wines. These compounds are characterized by fruity aromas that are frequently used to describe wine flavor and are correlated with enhanced aroma quality (van der Merve and van Wyk 1981). The concentrations detected after treatment were higher than the odor threshold values of each compound: 30 μg/L for 3-methylbutyl acetate, 14 μg/L for ethyl hexanoate, and 5.0 μg/L for ethyl octanoate (Ferreira et al. 2000). Octanoic and decanoic acids (threshold values 500 and 1000 μg/L, respectively) were also decreased by carbon treatment, although the sensory contribution of these compounds is generally considered less significant than that of esters (Ferreira et al. 2000, Etiévant 1991). Finally, the concentration of geraniol was significantly reduced by all treatments with carbon, although its decrease was not linked to the amount of carbon used. This terpene alcohol with an odor threshold value of 30 μg/L (Guth 1997) and linalool are the main volatiles responsible for the characteristic aroma of Muscat wines and the flowery aroma of Muscat-type wines. The results from the current study as well as those of Peòa et al. (2001) suggest that these wine volatiles compete with the carbon adsorption sites more readily than polyphenols. Although the use of carbon to remove OTA from wine is effective, it may also decrease overall wine aroma quality, which may limit the usefulness of this treatment during the production of high-quality wines.