|
A Technical Guide for Wine Producers
|
|
RECENT ARTICLES | WYNBOER HOME
A new yeast selection process for the expression of the varietal aroma of Sauvignon Blanc
Philippe Marullo1, Stéphane Bonhomme1, Marie-Laure Murat1, Isabelle Masneuf-Pomarède2, Charlotte Augustin3
Introduction
Over the last few years, the varietal aroma of Sauvignon Blanc has been extensively studied. Three volatile thiol compounds have been found to contribute largely to the varietal aroma of Sauvignon blanc: 4-mercapto-4-methylpentan-2-one (4MMP), 3-mercaptohexan-1-ol (3MH) and 3-mercaptohexyl acetate (3MHA), which are responsible for the hints of boxwood (4MMP, 3MHA), broom (4MMP), grapefruit (3MH) and passion fruit (3MH, 3MHA) in wines produced from this grape variety. These compounds are also found in wines made from other white grape varieties such as Gewürztraminer, Riesling, Colombard, Petit Manseng (Tominaga a.o., 2000a) and rosé wines made from the red varieties Cabernet Sauvignon and Merlot (Murat a.o., 2001a), as well as Grenache and Syrah (Cacho, 2004; Murat, 2005). Two sulphur compounds (4MMP and 3MH) are found in musts as non-volatile precursors, S-conjugated to cysteine (Tominaga a.o., 1995, 1998a; Murat a.o., 2001b). The volatile thiols are released through the action of yeasts during alcoholic fermentation from the corresponding cysteine S-conjugates (Tominaga a.o., 1998a). However, the transformation mechanism of the cysteinylated precursor into the final aroma compound is still unknown. In this respect, not all Saccharomyces cerevisiae strains have the same potential for the expression of thiols (Murat a.o., 2001c; Howell a.o., 2004). In addition, the amount of fermentation aromas in wines (esters and higher alcohols produced during alcoholic fermentation) varies depending on the yeast strain used (Rankine, 1967; Usseglio and Tomasset, 1967; Suomalainen, 1971).
A better understanding of Sauvignon blanc varietal aromas and the utilization of a traditional cross-breeding method have allowed the development of new, more precise and efficient methods for the selection and improvement of oenological yeast strains suited to this variety and others with similar characteristics and aroma precursors. Accordingly, SARCO laboratories and the Oenology Faculty of Bordeaux have used the breeding method: this technique consists in cross-breeding two yeasts in order to obtain a "daughter" possessing the performance qualities of both "parents" (Figure 1). Consequently, cross-breeding a yeast with proven fermentative capacities (fermentation kinetics, alcohol resistance, low production of volatile acidity and sulphur compounds) with a yeast known to produce typical and aromatic Sauvignon blanc wines led to the development and the commercialisation of the new Zymaflore X5 strain (Laffort ?nologie).
The objectives of this work were to study the advantages of Zymaflore X5 and to compare it with other strains with regards to the fermentation potential, the capacities to express thiolated aroma compounds and the final wine quality.
Figure 1 (below): Development of new oenological yeast strains using the crossing method or "breeding". The parental strains may belong to the same species (e.g. either S. cerevisiae or S. uvarum) or two different species (e.g. S. cerevisiae crossed with S. uvarum).
Figure 2 (below): 4MMP (ng/L) levels in wines obtained from different yeast strains
Figure 3 (below): Relation between the sum of the aroma indices for 4MMP and 3MHA of the wines and their sensory ranking.
Figure 4 (below): Descriptive sensory analysis of the three wines presenting the highest levels of 4MMP and 3MHA (15 tasters).
Materials and methods
Yeast strains used. Seven active dry yeast (ADY) strains were used for the trial: Zymaflore VL3 (Laffort Oenologie), Zymaflore X5 (Laffort Oenologie), and five of the commercial strains generally recommended for the production of typical and aromatic Sauvignon blanc, noted from A to E.
Experimental protocol. The trial was carried out in 2005 at the Laffort ?nologie experimental centre. The Sauvignon Blanc must was obtained from the Bordeaux area. Upon receipt, the grapes were pressed, pectolytic enzyme (Lafazym CL, 2 g/hL) and sulphite (as SO2, 5 g/hL) were added, and the must was cold settled (48 h at 12°C). Must turbidity was adjusted to 50 NTU. 200 litre tanks were used; the seven treatments were carried out in duplicate. The entire pre-fermentation process was performed in an oxygen-free environment (covered with CO2 until onset of alcoholic fermentation). Yeast inoculation (20 g/hL) was performed the day of tank filling and fermentations were carried out at 16°C. The wines were sulphated one week after the end of the fermentation, then aged on fine lees for two months with two stirrings per week.
Must characteristics. The initial must characteristics are presented in Table 1. The assimilable nitrogen deficiency observed was corrected by adding Thiazote© in order to reach 180 mg/L.
Alcoholic fermentation monitoring. The specific gravity and temperature were measured daily during the entire fermentation process.
Yeast implantation was controlled at a specific gravity of approximately 1,040. For this, 100 µL of must were taken from the tank and grown on solid media. The total biomass DNA was extracted and genetically analysed by PCR (MiniCyclerTM, MJ Research) at the same time as the DNA of the inoculated ADY. Electrophoresis allowed a comparison of the profiles obtained from the total biomass with the ones of the ADY, and thus control of the implantation of the latter.
Chemical and sensory analyses of the wines. The standard analyses and aroma measurements were performed with finished wines after sulphating. The levels of volatile thiols in the wines were assessed following the method described by Tominaga a.o. (1998b) and modified by the same author (2000b). The levels of isoamyl acetate (IA), phenylethyl acetate (PEA) and 2-phenylethanol (PE) were determined according to the method described by Bertrand a.o. (1978).
A panel composed of 15 experts familiar with Sauvignon Blanc carried out a blind tasting. An initial assessment (appearance, nose, palate) was performed with the seven wines, after which the replicates of every yeast treatment were combined. In addition, three of these samples were tasted again, separately, and rated from 1 to 5 (5 being the maximum) for 6 descriptors: aroma intensity, Sauvignon blanc character, hints of boxwood (4MMP and 3MHA) and grapefruit on the nose (3MH descriptor), roundness and length. The three samples selected for this descriptive trial were the ones that presented the highest levels of 4MMP and 3MHA after analysis; a preference test with these three wines was also conducted.
Statistical analysis of the results. The statistical software Stata 9.0 was used in order to carry out the variance analyses and Scheffer tests (assessment of the volatile thiols and fermentation aromas). Kramer and Friedmann's tests were performed for the sensory preference tests.
Results
Yeast implantation controls. The yeast implantation controls were positive for all treatments (data not shown), hence the differences found among finished wines can be ascribed entirely to the yeast strains used, since all other conditions were identical.
Fermentation kinetics. The fermentations were completed within 10 days for all strains (results not shown).
Standard analyses of the wines. There was no significant difference among yeast strain treatments with regards to the parameters analysed. Table 2 presents the average values obtained for each wine.
Volatile thiol analysis. Statistically, the levels of 4MMP in wines were significantly different (Figure 2 and Table 3).
Statistical analysis of the duplicate treatments show that the strain Zymaflore X5 expressed significantly more 4MMP than all other yeast strains tested (Figure 2). Interestingly, this crossbred strain expressed 4MMP better than the strains previously selected for this purpose.
In decreasing order, the strains A, Zymaflore VL3 and B displayed smaller expression potentials.
Concerning the expression of 3MH and its transformation into acetate, no significant difference could be found between the strains tested (Table 3). In fact, the levels of these two compounds were rather low in this vintage. However, wines produced with the strains Zymaflore X5, VL3 and A had the highest 3MHA (passion fruit) concentrations. It is interesting to compare these results with those obtained by a similar trial undertaken by the Australian Wine Research Institute in 2005 (Swiegers et al), where small fermentations (20 L) indicated significantly different levels of 4MMP production by the Laffort Oenologie strains X5 and VL3.
The authors noted that the results of small-scale experiments do not necessarily correlate with those of commercial fermentations (volatile thiol levels in both studies were measured by Laffort Oenologie). Indeed, the authors noted that for the same juice, when fermented in a commercial winery at a commercial scale (5000 L), they obtained different results. Since DNA fingerprinting did not confirm the specific fermenting yeast strains in the study by Swiegers et al, some conjecture over the origin of the observed differences between data sets remains.
Levels of fermentation aromas. Under the given experimental conditions, no differences could be found among the levels of IA, PEA and PE in the wines obtained with the tested strains (Table 4). These were absolutely comparable in terms of ester production.
Generally, a strong ester formation is not necessarily desired for the production of a fruity and typical Sauvignon Blanc. However, a moderate formation (such as in these trials) can confer a certain complexity.
Sensory analysis. The preference test performed with the seven samples was significant at the 0.1% level (Friedmann's test). Thus, the wine obtained using Zymaflore X5 was clearly preferred (significant at 1%, Kramer's test). In decreasing order, the next most preferred were those fermented with the strains A and Zymaflore VL3, which were very similar, followed by the wines produced with strains B, C, D and E. This wine ranking was very reproducible: the tasters often ranked the wines in the same order as found in the overall ranking. For example, the Zymaflore X5 wine was ranked first 10 times out of 15, and the wine obtained with strain C was ranked fifth, 11 times, which corresponded with their overall ranking.
There is a statistically significant correlation between the sum of the aroma indices for 4MMP and 3MHA and the preference ranking obtained after tasting (R2 = 0.7673 significant at 2%, cp. Figure 3).
The descriptive test was performed with the three wines showing the highest values of 4MMP and 3MHA, obtained using the strains Zymaflore X5, VL3 and strain A. In fact, these had also been ranked as the best wines in the preference test. Figure 4 shows the results of this descriptive analysis. The wine fermented with the X5 strain was considered to be the most typical and olfactively intense, strongly characterised by notes of boxwood and grapefruit. The wines produced with strain A and Zymaflore VL3 appeared to be rather similar according to this descriptive test. However, the ranking of this triangular test between strain A and Zymaflore VL3 was clear (significant at 1%, Kramer's test): the wine produced with the strain Zymaflore VL3 was preferred to the one made with strain A. Frequently, tasting comments referred to the fruitiness and elegance of the wine made with Zymaflore VL3, as well as the roundness of the wine produced with strain A. Once again, the wine made with the strain Zymaflore X5 was ranked first and was described as being well balanced, lively and having a long finish.
Conclusion
While all the strains tested within this comparative study displayed a satisfying fermentation potential (fermentation completed, limited volatile acidity, wines without defect), the results show that they differed with regards to the organoleptic characteristics of the wines obtained. The tasting not only considered the aroma criteria but also the mouth-feel and finish. The differences established during the analysis and assessment of these wines made from the same must, once again underline the role the yeast strain plays in the expression of volatile thiols.
The strain Zymaflore X5 distinguished itself very clearly during the tastings and wine analyses (typical character, complexity, long finish and vivacity). This strain produced wines with notable levels of 4MMP and 3MHA, compounds that were well correlated to the overall wine preference during tastings. This potential to express 4MMP confirms the results of previous trials carried out in 2004 in the experimental centre. The wine made with the strain Zymaflore VL3 was also distinguished wine preference testing, as well as the wine from strain A, but the latter to a lesser extent. These strains also presented a superior ability to express volatile thiols but strain A produced more fermentation esters. It should be noted that the differences among typical character between the wines made from each yeast could be mediated during blending.
Applied as a new technique of yeast selection, the method of breeding showed its advantage: Zymaflore X5 is already highly appreciated by numerous winemakers for the fermentation of Sauvignon blanc musts and the trials presented here offer some explanations. Breeding also led to the selection of a new strain for the production of fruity red wines: the strain Zymaflore RX60. Results obtained with this strain will be published in a future publication.
Bibliographic references
Bertrand A., Marly-Brugerolle C., Sarre C. (1978) Influence du débourbage des moûts et du sulfitage sur les teneurs en substances volatiles des vins et eaux-de-vie. Com. Vigne Vin, 12, 35-48.
|
|
Wynboer is incorporated in WineLand, magazine of the SA wine producers.
Visit our sister sites:
|
| UP |
COPYRIGHT (C) 2000 WineLand |