Yeast Strain-Derived Sensory Differences Retained in Sauvignon blanc Wines after Extended Bottle Storage

Results and Discussion

Of the 11 aroma attributes rated in the 2005 wines after three years of bottle aging, three attributes were significantly different among treatments at p < 0.05: aroma intensity, canned lychee, and fresh green. The sweaty attribute (p < 0.10) was also included in the analysis. For the chemical composition of the 2005 wines at three years, there were differences among the treatments for most of the volatile compounds measured. The principal component analysis (PCA) biplot (Figure 1A) shows the different aroma attributes and explains 92% of the variance. The different volatile compounds are also shown (Figure 1B), with 59% of variance explained.

• Download figure
• Open in new tab

Figure 1

Principal component analysis (PCA) biplots of the different aroma attributes (A) and volatile compounds (B) for three-year-old wines made in 2005 using seven different yeast treatments (closed circles). Volatile compounds in italic font indicate an odor activity value greater than 1.0.

The yeast treatment Vin7, situated on the far right side of the biplots, was rated highest in the attributes aroma intensity, canned lychee, and sweaty (Figure 1A) and had relatively high concentrations of the volatile thiols 4MMP, 3MH, and 3MHA and the fruity compounds 3-methylbutyl acetate, 2-methylpropyl acetate, ethyl butanoate, and ethyl-2-methylpropanoate, as well as most of the higher alcohols and volatile acids measured (Figure 1B). These results are similar to the Vin7 wines at six months, where the wines had high concentrations of all volatile thiol compounds and high ratings for the sweaty sensory attribute (Swiegers et al. 2009). The canned lychee attribute was closely associated with the Vin7 yeast treatment, which had the highest volatile thiol concentrations; therefore, we can speculate that the canned lychee attribute is related to the presence of volatile thiols in aged white wines, although it has not previously been described.

The X5 yeast treatment was rated highest in the fresh green attribute (Figure 1A). All other yeast treatments had similar sensory profiles, except the VL3 yeast treatment had significantly lower ratings for the attribute aroma intensity than the Vin13 treatment. There were significant differences in the chemical compositions of the yeast treatments, illustrated by their vertical spread along PC2 (Figure 1B). The Vin13 and NT116 yeast treatments were highest in ethyl propanoate, ethyl lactate, and butanol, whereas the QA23, L2056, and X5 treatments were high in propanoic acid, and VL3 yeast treatment was high in 2-methylbutyl acetate and hexanol.

From an ANOVA of the sensory descriptive data for the 2006 wines assessed three years postbottling, the attributes yellow (color), overall fruit flavor, prickly, yeasty/cheesy, sweetness, and bitter were significantly different among the wines at p < 0.05. Two other attributes were included in further analysis: cat urine/sweaty and solvent (p < 0.10). The sensory and chemical data for the 2006 wines at three years are shown. The PCA biplot of the aroma attributes explains 76% of the total variance (Figure 2A), while the PCA biplot of the volatile compounds explains 78% of the variance (Figure 2B).

• Download figure
• Open in new tab

Figure 2

Principal component analysis (PCA) biplots of the different aroma attributes (A) and volatile compounds (B) for three-year-old wines made in 2006 using seven different treatments (closed circles). Italicized volatile compounds have an odor activity value greater than 1.0. Yeast treatments with a slash (/) represent coinoculations and and those with a plus (+) represent blended wines of Vin7 and QA23, and Vin7 and Vin13 yeast strains.

The Vin7 yeast treatment and two blended wines Vin7+QA23 and Vin7+Vin13 were rated highly in the attributes solvent and overall fruit flavor (Figure 2A). The Vin7 yeast treatment and the blended wines had high concentrations of volatile acidity and ethyl acetate, as well as 4MMP, 2-methylpropyl acetate, 2-methylbutyl acetate, 2-methylpropanol, and 3-methylbutanol (Figure 2B). The aroma profiles of these wines reflect the high concentrations of volatile acidity and high ratings of white vinegar observed in the young wines (King et al. 2008). These treatments also retained relatively high concentrations of the volatile thiol 4MMP (King et al. 2008), the sensory contribution of which may be masked by the solvent attribute.

The Vin13 yeast treatment and the two coinoculated wines Vin7/Vin13 and Vin7/QA23 were situated in the left side of the biplot (Figure 2A), rated high in the cat urine/sweaty and prickly aromas. Vin13 was rated highly in the yeasty/cheesy attribute, along with QA23 yeast treatment, situated at the bottom of the biplot. Vin13 and Vin7/Vin13 yeast treatments had high concentrations of the volatile thiol 3MH, as well as 3-methylbutyl acetate, phenylethyl acetate, ethyl-2-methylpropanoate, ethyl-3-methylbutanoate, and butanol (Figure 2B). Vin7/QA23 coinoculation also had relatively high concentrations of 3MH and high concentrations of hexyl acetate, ethyl propanoate and hexanol, similar to QA23, hence its position near the bottom of the biplot (Figure 2B).

The wine aroma profiles of the coinoculated wines at three years were similar to those measured at six months, with high concentrations of volatile thiols and highest ratings for the volatile thiol-related sensory attributes box hedge and passionfruit (King et al. 2008). The cat urine/sweaty attribute in the three-year-old wines is synonymous with box hedge (King et al. 2011), which was used as a reference standard for this attribute (Table 1).

The wines in both sets of samples studied had large differences in sensory properties after three years of bottle storage, which could be characterized using sensory descriptive analysis methodology, rather than requiring more sensitive sensory tests, such as discrimination tests (Stone and Sidel 2004). As expected, the three-year-old Sauvignon blanc wines no longer exhibited most of the fresh, fruity aromas that distinguished the treatments at six months, such as estery, confectionary, and floral (King et al. 2008, Swiegers et al. 2009). This result was common to all treatments for both vintages and was associated with a decrease in most of the fruity acetate esters, as reported elsewhere (Rapp 1998). There was, nonetheless, a substantial degree of commonality evident among the sensory properties measured at six months and those perceived at three years for both vintages, partly because young wines with high initial concentrations of volatile thiols remain relatively high in these compounds after three years of bottle aging and continue to display volatile thiol-related sensory attributes. A similar theory was suggested whereby yeast strains that produce high concentrations of ester compounds, and thus, higher initial fruity aroma ratings soon after bottling, can result in a longer shelf life of the wine (Lilly et al. 2000).

There was some treatment variation in free and total sulfur dioxide concentrations. All treatments had relatively low free sulfur dioxide levels, ranging from 8 to 13 mg/L (data not shown), with the lowest concentrations in the wines made with the strains Vin13, Vin7, and NT116 in the 2005 wines, and all the treatments in the 2006 wines, except the coinoculated wines. There was no significant relationship between free sulfur dioxide levels and ratings for any of the sensory attributes in the two sample sets, indicating that the slight variation in sulfur dioxide was not the cause of the sensory differences observed.