Original papers
Factors Affecting Dimethyl Trisulfide Formation in Wine
2017 Volume 23 Issue 2 Pages 241-248
Details
2017 Volume 23 Issue 2 Pages 241-248
Dimethyl trisulfide (DMTS) develops in wine during storage, creating an unpleasant aroma. However, the mechanisms of DMTS formation during the wine-making process are poorly understood. We attempted to identify the factors that affect DMTS development during storage. We found that the death of yeast cells, followed by the leakage of their cell contents into the must, enhanced DMTS formation, but this effect varied between grape cultivars. Oxidation of the must and insufficient clarification of the grape juice can also lead to DMTS development during storage. However, DMTS-P1 and methanethiol, precursors of DMTS in other beverages, contributed little to DMTS formation in wine. When synthetic grape must was fermented instead of grape juice, less DMTS developed, despite most yeast cells having died. These findings indicate that unknown compounds contained in grape juice are involved in DMTS formation and that the key compounds involved in DMTS formation in wine originate from grapes.
Low molecular weight sulfur compounds, including hydrogen sulfide, methanethiol (MTL), dimethyl sulfide, dimethyl disulfide, and dimethyl trisulfide (DMTS), have an unpleasant smell with low perception thresholds and produce off-flavors in wine. However, some high molecular weight thiols, such as 3-mercaptohexal-1-ol, 3-mercaptohexyl acetate, and 4-mercapto-4-methyl-pentan-2-one, have a positive effect on the aroma profile of wine.
DMTS develops during wine storage and has an odor that is similar to pickled radish or cooked vegetables. Although DMTS has not been well-studied in wine because of its low concentration, it was reported to be an odor-active compound in wines that were analyzed using gas chromatography/olfactometry (Guth, 1997). Furthermore, although its perception threshold level in wine has not been documented, it is as low as 0.1 µg/L in beer and 10% (w/w) ethanol (Landaud et al., 2008; Leppänen et al., 1979) and 0.18 µg/L in sake (Utsunomiya et al., 2004). DMTS has been found in various kinds of wine, including sherry, madeira, and white and red wines. The concentrations of DMTS in sherry and madeira range from undetected (ND) to 0.5 µg/L (Leppänen et al., 1979), and the concentrations in white and red wines range from ND to 0.9 µg/L (Davis, 2012; Fang and Qian, 2005; Nguyen et al., 2012). Because the threshold level of this polysulfide is low, these concentrations are thought to be sufficient to affect the flavor of wine.
DMTS is formed from various precursors. In whisky and aged beer, the precursors are MTL and methional, which are generated from methionine through the Strecker degradation process (Gijis et al., 2000; Lee et al., 2001), while S-methylcysteine sulfoxide from hops is a precursor in beer (Peppard, 1978). In sake, DMTS-P1 and DMTS-P2 have also been reported as precursors of DMTS, which can develop after at least 6 months of storage at room temperature and negatively affects the odor (Isogai et al., 2009, 2010). The structure of DMTS-P1 was determined as 1,2-dihydroxy-5-(methylsulfinyl)pentan-3-one, which is structurally similar to 1,2-dihydroxy-5-(methylthio)-1-penten-3-one, an intermediate compound of the methionine salvage pathway (MTA cycle), and therefore, it has been assumed that DMTS-P1 originated from this cycle (Isogai et al., 2009). Supporting this, it was recently shown that disruption of the MTA cycle genes MDE1, which encodes a methythioribulose-1-phosphate dehydrogenase, and MRI1, which encodes a methylthioribose-1-phosphate isomerase, resulted in a drastic reduction of DMTS-P1 production (Wakabayashi et al., 2013).
It has previously been reported that several factors that affect DMTS formation are present in sake. For example, the concentration of sulfur amino acids in sake was shown to be significantly positively correlated with the production of DMTS after storage (Okuda et al., 2009; Sasaki et al., 2014); and the lysis of yeast cells and subsequent release of their cell contents during sake fermentation were found to accelerate DMTS formation after storage (Nishibori et al., 2014). Moreover, the enzymes that are released from yeast cells are likely involved in DMTS formation. However, the factors that influence DMTS formation and the precursors of DMTS have not been well studied in wine. Therefore, in this study, we investigated the factors that affect DMTS development during wine storage.
Yeast strains The following wine yeast strains were maintained in our laboratory: Saccharomyces cerevisiae Kyokai No. 4 (W-3), OC-2, sake yeast Kyokai No. 7 (K-7), MED1 deletion mutant of the K7 strain (Wakabayashi et al., 2013; which was kindly donated by Dr. Isogai, National Research Institute of Brewing, Japan), Kyokai No. 9 (K-9), and S. bayanus UCD530. In addition, the commercial wine yeasts Lalvin 71B, BM45, Rhone2323, EC1118, Uvaferm CS2 (Lallemand, QC, Canada), and Zymaflore X5 (Laffort, Bordeaux, France) were purchased as dried yeast and rehydrated according to the manufacturers' instructions. All yeast strains were maintained on YPD agar plates (1% yeast extract, 2% peptone, and 2% glucose).
Preparation of grape juice for fermentation Blush-type rose wine and white wine were made in small batches using the grapes listed in Table 1. All grapes used in this study were harvested from our experimental vineyard in 2014 and stored at −20°C. The grape berries were crushed and pressed by hand, and 100 mg/kg of potassium metabisulfite was added during crushing. The juice yield of each grape was 70 – 75%. The sugar concentration was adjusted to 22% (w/v) with sucrose, the pH was adjusted to 3.5 with malic acid, and the yeast available nitrogen (YAN) content was adjusted to 250 mg N/L with diammonium phosphate. To examine the effect of grape cultivars on DMTS formation, the YAN content was adjusted to 300 mg N/L.
| Degree Brix | pH | Total acid* | YAN | Total amino acid | Sulfur amino acid | |
|---|---|---|---|---|---|---|
| (%, w/v) | (mg/L) | (mg/L) | (mg/L) | |||
| Chardonnay | 18.0 | 3.6 | 0.59 | 288.4 | 1753.4 | 5.3 |
| Niagara | 12.6 | 3.2 | 0.58 | 299.6 | 2557.7 | 8.6 |
| Koshu | 16.4 | 3.3 | 0.68 | 198.8 | 1438.5 | 2.1 |
| Merlot | 19.6 | 3.7 | 0.37 | 193.2 | 1154.6 | 8.3 |
| Campbell Early | 16.4 | 3.5 | 0.43 | 319.2 | 2718.5 | 6.3 |
| Malbec | 17.8 | 3.7 | 0.63 | 232.4 | 1697.2 | 10.0 |
| SGM | 22.0 | 3.5 | 0.30 | 300.0 | 1807.0 | 52.0 |
YAN=yeast available nitrogen
Clarification of the grape juice Grape juice was prepared as outlined above and then allowed to settle for 16 h at 4°C. The clear layer was then carefully collected to obtain the clarified grape juice.
Preparation of synthetic grape must for fermentation Synthetic grape must (SGM) was also used in the fermentation tests, and was prepared as described by Beltran et al. (2004) with some modification. Briefly, glucose and fructose (110 g/L each) were used as carbon sources, the YAN content was adjusted to 250 mg N/L with 100 mg N/L of NH4Cl and 150 mg N/L of amino acid mixture, and the organic acids composition was 3 g/L of tartaric acid, and 0.3 g/L each of citric and malic acids. The pH was adjusted to 3.5 using NaOH.
Small-scale wine fermentation Fermentations were conducted in triplicate at 20°C in 200-mL glass bottles, which were loosely covered with caps, containing 150 mL of grape juice or SGM under static conditions. Each yeast strain was precultured in SGM at 25°C for 48 h, and then inoculated into each grape juice or SGM at a concentration of 3 × 106 cells/mL. The fermentation nutrient preparation Fermaid® K (Lallemand) was then supplied at a concentration of 125 mg/L. When fermentations were performed under anaerobic conditions, glass bottles were covered with airlocks instead of caps. Fermentation was monitored by measuring the weight loss of the must, which corresponded to the CO2 evolution due to alcohol fermentation.
At the end of fermentation, the must was centrifuged at 9000 g for 10 min without racking or the addition of potassium metabisulfite. The supernatants (i.e., the wine samples) were then stored at −20°C until analysis. However, total acid, pH, volatile sulfur compounds, and DMTS-producing potential were analyzed without freezing of wine samples.
Determination of the percentage of dead cells Prior to centrifugation of the must, the percentage of dead cells in the must was determined under a microscope using the methylene blue (MB) staining method (Murakami, 1974).
Measurement of alcohol concentration The concentration of ethanol was measured using a gas chromatography system (GC-17A; Shimadzu, Kyoto, Japan) equipped with a flame ionization detector and a capillary column (DB-624, 30 m × 0.53 mm, 3 µm film thickness; Agilent Technologies, CA, USA). The temperatures of the oven, detector, and injector were maintained at 50, 250, and 250°C, respectively, the flow rate of the helium carrier gas was 6 mL/min, and the split ratio was 40:1.
Measurement of total acid Total acid was determined by titration of 10 mL of wine samples with 0.1N sodium hydroxide to an end point of 8.2, and was expressed as tartaric acid equivalent.
Quantification of total sulfur and volatile sulfur compounds The total sulfur content was analyzed using inductively coupled plasma atomic emission spectrometry (ICP-AES, ICPE-9000; Shimadzu) under the following analytical conditions: radio frequency output, 1.2 kW; argon gas flow rate, 10.0 L/min (plasma gas), 0.6 L/min (auxiliary gas), and 0.7 L/min (carrier gas); axial observation; and analytical wavelength, 180.731 nm.
The volatile sulfur compounds were analyzed using a headspace gas chromatography system (6890N; Agilent Technologies) equipped with a pulsed flame photometric detector (OA Analytical, TX, USA) and a DB-1 capillary column (30 m × 0.53 mm, 3 µm film thickness; Agilent Technologies). The oven temperature was programmed as follows: initial hold at 35°C for 5 min; ramp at 10°C/min to 250°C; and hold for 5 min. The flow rate of the helium carrier gas was 1.5 mL/min, and the temperatures of the detector and injector were maintained at 220 and 250°C, respectively.
Quantification of amino acids and DMTS-P1 The amino acid contents were measured using an amino acid analyzer (JLC-500; JEOL Corporation, Tokyo, Japan). The concentration of DMTS-P1 was measured using liquid chromatography-mass spectrometry (LCMS-8040; Shimadzu), following the method described by Wakabayashi et al. (2013).
Measurement of the DMTS-producing potential DMTS usually develops during wine storage and, supporting this, we detected DMTS in aged wines but not in young wines (data not shown). Therefore, we mimicked wine aging in the laboratory by carrying out an accelerated aging treatment and measured the amount of DMTS that developed, which was defined as the DMTS-producing potential (DMTS-pp) (Isogai et al., 2009). A 9-mL sample of each wine was placed into a 10-mL glass vial, sealed with a PTFE/silicon septum, and incubated at 70°C for 1 week. The amount of DMTS was then analyzed using stir bar sorptive extraction coupled to a thermal desorption-GC-MS, following the method of Isogai et al. (2010), but using DMTS-d6 as an internal standard rather than 3-octanol.
Statistical analysis Correlations between multiple pairs of variables were calculated using JMP software ver. 12.01 (SAS Institute Inc., NC, USA).
Effect of yeast strain on the DMTS-producing potential Small-scale fermentations were conducted with Merlot juice (YAN 250 mg/L) using different yeast strains at 20°C. The fermentation process of each yeast strain is shown in Figure 1. All nine of the wine yeasts, including S. bayanus, showed similar fermentation profiles, whereas sake yeasts K7 and K9 demonstrated lower fermentation rates that took longer to complete. The final amount of evolved CO2 was similar among all yeasts.
Changes in total CO2 evolution during fermentation in wine yeasts (solid lines) and sake yeast (dotted lines). The fermentations were conducted at 20°C in 200-mL glass bottles containing 150-mL Merlot must (YAN 250 mg/L). These data represent the means of three separate assays with SDs within 5% of the mean.
The analytical wine production values and the correlation coefficients between DMTS-pp and the other analytical values are shown in Table 2. Total amino acids included all major proteinogenic amino acids except proline, which is neither assimilated by yeast under wine-making conditions nor affects DMTS-pp (data not shown), β-alanine, cystathionine, γ-aminobutyric acid, phosphoserine, and taurine. Sulfur amino acids included cystine, methionine, and cystathionine. Values for the MB staining ratio (defined as the percentage of dead yeast cells), and the concentrations of total amino acids and sulfur amino acids varied between the yeast strains. In addition, DMTS-pp differed greatly between the yeast strains and was significantly positively correlated with the MB staining ratio, total amino acid concentration, and sulfur amino acid concentration (p < 0.01). It is well known that the death of yeast cells in the lees followed by the leakage of their cell contents increases the amino acid concentrations in wine. Thus, it is plausible that the death of yeast cells would increase both the total amino acid and sulfur amino acid concentrations in wine. This hypothesis was supported by the high correlation coefficients between the MB staining ratio and both the total amino acid and the sulfur amino acid concentrations (0.90 and 0.87, respectively). However, the other analytical values, including the concentrations of DMTS-P1 and MTL, which are known precursors of DMTS, were not related to DMTS-pp. Therefore, it is likely that the differences in DMTS-pp between yeast strains resulted from differences in the death rates of these strains.
| Yeast strains | Fermentation time | Alcohol | Total acid* | pH | MB | Total amino acid | Sulfur amino acid | Cys | Met | Cysta | Total Sulfur | H2S | MTL | DMS | DMTS-P1 | DMTS-pp |
|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (mg/L) | (mg/L) | (mg/L) | (mg/L) | (mg/L) | mg/L | (µg/L) | (µg/L) | (µg/L) | (µg/L) | (µg/L) | ||
| 71B | 13 | 12.8±0.1 | 0.63±0.00 | 3.7±0.0 | 98.0±0.6 | 211.0±5.9 | 25.1±0.8 | 19.6±1.5 | 4.7±1.2 | 0.8±0.2 | 65.0±1.3 | 4.3±0.2 | 2.0±0.2 | 1.1±0.0 | 24.5±0.7 | 2.6±0.2 |
| X5 | 13 | 12.6±0.1 | 0.94±0.01 | 3.4±0.0 | 6.5±2.4 | 41.5±2.3 | 6.5±0.1 | 5.4±0.1 | 0.4±0.0 | 0.7±0.0 | 61.0±0.4 | 3.4±0.1 | 1.8±0.0 | 1.1±0.0 | 43.2±5.2 | 0.4±0.1 |
| Rhone2323 | 13 | 12.5±0.1 | 0.84±0.01 | 3.4±0.0 | 13.7±11.1 | 37.4±3.3 | 6.7±0.5 | 5.5±0.4 | 0.4±0.0 | 0.8±0.1 | 60.3±0.8 | 3.7±0.3 | 1.8±0.1 | 1.2±0.0 | 35.6±5.4 | 0.6±0.2 |
| BM45 | 13 | 12.4±0.0 | 0.89±0.01 | 3.4±0.0 | 14.9±1.9 | 66.1±11.5 | 8.6±0.7 | 6.9±0.6 | 0.8±0.1 | 1.0±0.0 | 80.3±1.5 | 2.8±0.0 | 1.7±0.1 | 1.2±0.0 | 28.0±2.2 | 0.8±0.3 |
| CS2 | 13 | 12.7±0.01 | 0.80±0.00 | 3.5±0.0 | 98.8±0.4 | 179.4±5.1 | 17.1±0.4 | 12.1±0.3 | 3.9±0.1 | 1.1±0.1 | 68.3±1.0 | 4.3±0.2 | 2.0±0.0 | 1.2±0.0 | 27.6±2.2 | 4.6±0.2 |
| W3 | 13 | 13.1±0.0 | 0.66±0.01 | 3.4±0.0 | 95.2±1.0 | 201.8±7.8 | 25.4±0.5 | 21.0±0.5 | 3.8±0.3 | 0.6±0.1 | 69.0±0.4 | 8.2±0.6 | 2.2±0.1 | 1.1±0.0 | 27.5±3.4 | 2.4±0.4 |
| OC2 | 13 | 12.8±0.1 | 0.70±0.01 | 3.4±0.0 | 95.6±1.9 | 205.4±3.2 | 25.0±0.9 | 20.7±0.7 | 3.5±0.1 | 0.8±0.2 | 64.9±1.8 | 7.9±0.2 | 2.8±0.1 | 1.1±0.0 | 63.1±2.2 | 4.7±0.6 |
| EC1118 | 13 | 12.8±0.5 | 0.63±0.01 | 3.4±0.0 | 89.4±0.4 | 132.6±4.8 | 15.5±0.2 | 12.2±0.1 | 2.0±0.2 | 1.2±0.1 | 85.3±1.9 | 5.0±0.2 | 2.3±0.1 | 1.1±0.0 | 28.5±3.3 | 4.2±0.5 |
| UCD530 | 13 | 12.8±0.3 | 0.66±0.01 | 3.5±0.0 | 90.8±1.1 | 233.2±11.0 | 28.3±1.0 | 23.1±0.9 | 4.1±0.1 | 1.1±0.0 | 66.0±0.4 | 7.5±0.3 | 2.6±0.1 | 1.1±0.0 | 30.2±1.8 | 3.6±0.3 |
| K7 | 19 | 12.3±0.0 | 0.87±0.01 | 3.5±0.0 | 96.2±0.7 | 262.1±5.8 | 32.2±1.7 | 25.8±1.7 | 5.4±0.0 | 1.1±0.1 | 69.1±1.6 | 2.7±0.1 | 1.5±0.1 | 1.1±0.0 | 45.5±8.9 | 7.7±2.3 |
| K9 | 19 | 12.5±0.1 | 0.81±0.00 | 3.5±0.0 | 98.1±0.4 | 313.4±3.2 | 33.5±0.4 | 27.6±0.6 | 4.5±0.1 | 1.3±0.1 | 70.6±1.2 | 2.7±0.0 | 1.5±0.1 | 1.1±0.0 | 55.6±4.6 | 10.5±0.3 |
| Correlation coefficients with DMTS-pp (p < 0.01**) | 0.41 | −0.03 | 0.12 | 0.72** | 0.86** | 0.81** | −0.02 | 0.00 | −0.38 | 0.20 | −0.19 | −0.23 | −0.45 | 0.41 | 1.00 | |
MB = the ratio of cells stained with methylene blue in must, Cys=cystine, Met=methionine, Cysta=cystationine, MTL= methanthiol, DMS=dimethylsulfide, DMTS-pp = the amount of DMTS after accelerated aging Values represent the mean of at least three separate assays and standard deviations were calculated.
Since total sulfur and the concentrations of Cys, Met, Cysta, H2S, MTL, DMS, and DMTS-P1 were not correlated with DMTS-pp, they have been omitted from the following discussion.
Effect of yeast cell death on the DMTS-producing potential To confirm the relationship between DMTS-pp and the death of yeast cells, we obtained wine samples with different MB staining ratios from small-scale fermentation. We found that the yeast cells did not die quickly when the carbon source was exhausted, but died gradually as the must was left longer (known as sur lie). Therefore, we took wine samples at different times to compare DMTS-pp in wine with different MB staining ratios. To kill the yeast cells, we added ethanol to the musts to create a final concentration of more than 17% and then incubated the samples at 20°C for 3 days. The MB staining ratios were more than 90% the day after ethanol addition and 100% by day 3. Thus, both a longer sur lie period and the addition of ethanol increased the MB staining ratio regardless of the yeast strain and grape variety used (Table 3). Similarly, the total amino acid and sulfur amino acid concentrations, and DMTS-pp were also increased by extending the sur lie period and by adding ethanol.
| Must | Yeast strain | Fermentation time | Alcohol | Total acid*2 | pH | MB | Total amino acid | Sulfur amino acid | DMTS-pp |
|---|---|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (mg/L) | (mg/L) | (µg/L) | |||
| Merlot | 71B | 9 | 12.8±0.0 | 0.66±0.13 | 3.6±0.0 | 24.1±1.0 | 50.3±1.6 | 11.1±0.1 | 0.4±0.2 |
| 9*1 | 18.0±0.1 | 0.62±0.03 | 3.7±0.0 | 100.0 | 162.2±3.6 | 25.4±0.7 | 1.3±0.1 | ||
| 16 | 12.8±0.1 | 0.64±0.00 | 3.6±0.0 | 77.4±3.1 | 158.6±3.2 | 22.2±0.6 | 1.7±0.1 | ||
| CS2 | 6 | 12.7±0.1 | 0.76±0.01 | 3.5±0.0 | 13.5±1.9 | 34.2±1.5 | 7.3±0.1 | 0.4±0.1 | |
| 6*1 | 17.8±0.1 | 0.68±0.01 | 3.7±0.0 | 100.0 | 136.2±0.8 | 17.8±0.3 | 3.3±0.1 | ||
| 12 | 12.6±0.2 | 0.70±0.00 | 3.6±0.0 | 99.2±0.1 | 195.9±3.8 | 20.8±0.2 | 4.1±0.8 | ||
| K7 | 15 | 12.3±0.1 | 0.80±0.01 | 3.5±0.0 | 19.3±3.5 | 47.3±5.7 | 7.6±0.6 | 0.9±0.7 | |
| 19 | 12.3±0.0 | 0.87±0.01 | 3.6±0.0 | 96.2±0.7 | 262.1±5.8 | 32.2±1.7 | 7.7±2.3 | ||
| Campbell Early | 71B | 5 | 12.6±0.1 | 0.74±0.01 | 3.5±0.0 | 29.5±4.7 | 133.9±6.9 | 3.5±0.5 | 1.1±0.2 |
| 5*1 | 18.1±0.2 | 0.61±0.01 | 3.7±0.0 | 100.0 | 307.3±6.9 | 18.5±1.9 | 2.5±0.2 | ||
| 11 | 12.7±0.1 | 0.70±0.01 | 3.6±0.0 | 98.3±0.3 | 426.7±10.3 | 16.9±0.7 | 4.5±0.4 | ||
| CS2 | 5 | 12.8±0.1 | 0.72±0.00 | 3.4±0.0 | 18.1±2.7 | 143.7±4.7 | 6.3±0.6 | 2.5±0.1 | |
| 5*1 | 17.7±0.4 | 0.60±0.01 | 3.6±0.0 | 100.0 | 319.4±4.2 | 22.3±0.1 | 3.6±0.1 | ||
| 11 | 12.7±0.0 | 0.68±0.01 | 3.5±0.0 | 98.6±0.5 | 402.4±31.6 | 25.7±8.2 | 5.2±0.3 |
Values represent the mean of at least three separate assays and standard deviations were calculated.
These results indicate that the increasing DMTS-pp was due to the death of the yeast cells followed by the leakage of their cell contents into the must. In sake, it has previously been reported that there is a positive correlation between the production of DMTS after storage and the concentration of sulfur amino acids in samples (Okuda et al., 2009). Thus, the increasing sulfur amino acid concentrations in the wine samples are thought to contribute to the increasing DMTS-pp. Moreover, we previously reported that the lysis of yeast cells and subsequent release of their cell contents into sake mash accelerated DMTS development, and that the enzymes released from yeast cells are involved in DMTS formation (Nishibori et al., 2014). Therefore, similar phenomena likely occur in wine.
Contribution of DMTS-P1 to the DMTS-producing potential DMTS-P1 is the major precursor of DMTS in sake (Isogai et al., 2009). However, in this study, DMTS-pp was not positively correlated with DMTS-P1 in wine (Table 2). These findings suggest that DMTS-P1 makes only a limited contribution to the formation of DMTS in wine, and so other precursors of DMTS must be present. To confirm this hypothesis, we used the K7 and K7Δmde1 strains for small-scale fermentations. It has previously been shown that the K7Δmde1 strain produced only a low concentration of DMTS-P1 in sake and resulted in a much lower DMTS-pp than the K7 strain (Wakabayashi et al., 2013). In the present study, both strains produced very similar MB staining ratios and general properties in wine (Table 4). Furthermore, although the DMTS-P1 concentrations were much lower in wine made from the K7Δmde1 strain than in wine made from the K7 strain, the DMTS-pp in the samples were similar between the two strains. These results demonstrated that there are several precursors of DMTS in wine.
| Must | Yeast strain | Fermentation time | Alcohol | Total acid* | pH | MB | DMTS-P1 | DMTS-pp |
|---|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (µg/L) | (µg/L) | |||
| Merlot | K7 | 21 | 12.3±0.1 | 0.79±0.01 | 3.6±0.0 | 96.8±0.3 | 89.4±19.5 | 4.4±0.6 |
| Δmde1 | 21 | 12.4±0.3 | 0.82±0.01 | 3.6±0.0 | 98.4±0.7 | 9.9±0.8 | 6.8±0.7 | |
| Campbell Early | K7 | 9 | 12.2±0.1 | 0.62±0.01 | 3.6±0.0 | 95.4±2.6 | 171.7±20.1 | 7.4±0.7 |
| Δmde1 | 9 | 12.4±0.3 | 0.62±0.01 | 3.6±0.0 | 97.4±2.2 | 3.2±0.3 | 7.6±1.6 |
Values represent the mean of at least three separate assays and standard deviations were calculated.
DMTS-producing potential of wine from different grape varieties One of the precursors of DMTS in beer originates from plant material (Peppard, 1978). Therefore, to examine the possibility that the precursors of DMTS in wine originate from the grapes, we used several grape varieties for small-scale fermentations with 71B wine yeast. The properties of the grape juices are shown in Table 1.
As shown in Table 5, there was no difference in the general properties of the wines or any of the other analytical values except DMTS-pp between the grape musts and SGM. However, DMTS-pp did vary between the grape musts, but was not correlated with any other analytical values, including the MB staining ratio. As shown in Tables 2 and 3, the death of yeast cells increased DMTS-pp. However, although nearly all of the yeast cells had died in all of the musts, there was a more than 10-fold difference in DMTS-pp across the grape musts (Table 5), and the DMTS-pp of SGM was low (0.2 µg/L). These results indicate that certain compounds that originate from the grapes are involved in DMTS formation in wine, which may be further enhanced by the death of yeast cells if these compounds are present in the must. However, it is unclear what these compounds are and whether certain precursors are present in grape must or become precursors through yeast metabolic processes. Moreover, it was not possible to clarify whether the observed differences in DMTS-pp resulted from differences between the grape varieties or some other factor(s).
| Fermentation time | Alcohol | Total acid* | pH | MB | Total amino acid | Sulfur amino acid | DMTS-pp | |
|---|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (mg/L) | (mg/L) | (µg/L) | ||
| Chardonnay | 13 | 12.5±0.1 | 0.63±0.00 | 3.8±0.0 | 97.2±0.6 | 260.6±16.4 | 28.8±1.0 | 0.6±0.3 |
| Niagara | 13 | 12.6±0.2 | 0.63±0.01 | 3.7±0.0 | 98.6±0.7 | 279.8±3.8 | 22.6±0.4 | 1.7±0.3 |
| Koshu | 13 | 12.5±0.2 | 0.67±0.01 | 3.7±0.0 | 99.0±0.3 | 191.1±18.0 | 24.2±0.9 | 1.2±0.2 |
| Merlot | 13 | 12.6±0.1 | 0.62±0.00 | 3.7±0.0 | 98.0±1.0 | 185.3±9.0 | 22.0±1.2 | 4.3±0.6 |
| Campbell Early | 13 | 12.6±0.1 | 0.61±0.01 | 3.8±0.0 | 99.1±0.5 | 384.9±15.9 | 30.1±0.7 | 7.1±0.8 |
| Malbec | 13 | 12.3±0.0 | 0.53±0.01 | 3.8±0.0 | 98.9±0.3 | 324.7±8.6 | 28.1±3.5 | 1.6±0.3 |
| SGM | 13 | 12.6±0.1 | 0.61±0.00 | 3.4±0.0 | 89.0±1.1 | 243.9±11.6 | 12.8±0.2 | 0.2±0.0 |
The YAN content was adjusted to 300 mg N/L
All the wine was fermented by 71B wine yeast.
Values represent the mean of at least three separate assays and standard deviations were calculated.
Effects of clarifying grape juice and air conditioning during fermentation on the DMTS-producing potential In the standard method for making white wine, the grape juice is clarified by allowing it to settle overnight at a low temperature and is then fermented under anaerobic conditions. The clarification of grape juice improves the wine quality and reduces the amount of undesirable sulfur compounds in the final product (Ribereau-Gayon et al., 2005). In the present study, our small-scale fermentations were performed under slightly aerobic conditions, particularly when most of the sugars had been exhausted and the fermentation speed had slowed down. Thus, we examined the effect of clarifying grape juice and air conditioning during fermentation on the DMTS-pp of wine samples.
As shown in Table 6, the clarification of Merlot juice resulted in a reduction in DMTS-pp, whereas the clarification of Campbell juice did not. The reason for this difference was not clear, but it does show that clarifying grape juice can contribute to the reduction in DMTS-pp. Fermentation under anaerobic conditions also reduced DMTS-pp, despite most yeast cells being dead (Table 7). Under standard wine-making conditions, the musts are rarely exposed to oxygen during fermentation, and so commercial wines usually contain very small amounts of DMTS–and clarifying the grape juice may further suppress DMTS development. However, oxidation of the must and insufficient clarification of the grape juice may cause DMTS development during storage.
| Must | Fermentation time | Alcohol | Total acid* | pH | MB | Total amino acid | Sulfur amino acid | DMTS-pp | |
|---|---|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (mg/L) | (mg/L) | (µg/L) | |||
| Merlot | Non-clarified | 13 | 12.7±0.1 | 0.60±0.00 | 3.7±0.0 | 98.0±0.5 | 181.4±5.4 | 23.4±0.8 | 2.0±0.3 |
| Clasrified | 13 | 13.0±0.1 | 0.60±0.01 | 3.7±0.0 | 94.7±0.5 | 205.6±2.6 | 22.5±0.1 | 0.7±0.1 | |
| Campbell Early | Non-clarified | 11 | 12.7±0.1 | 0.68±0.00 | 3.6±0.0 | 98.6±0.5 | 402.4±10.3 | 18.9±0.7 | 4.5±0.4 |
| Clasrified | 10 | 12.7±0.1 | 0.64±0.01 | 3.6±0.0 | 98.9±0.2 | 665.8±70.2 | 18.1±4.5 | 3.6±0.2 |
All the wine was fermented by 71B wine yeast.
Values represent the mean of at least three separate assays and standard deviations were calculated.
| Must | air condition | Fermentation time | Alcohol | Total acid*2 | pH | MB | DMTS-pp |
|---|---|---|---|---|---|---|---|
| (days) | (%, v/v) | (%, w/v) | (%) | (µg/L) | |||
| Merlot | normal | 13 | 13.1±0.2 | 0.58±0.01 | 3.7±0.0 | 97.6±0.6 | 4.0±0.6 |
| anaerobic*1 | 13 | 13.1±0.2 | 0.58±0.02 | 3.7±0.0 | 96.8±1.4 | 1.0±0.2 | |
| Campbell Early | normal | 11 | 12.7±0.1 | 0.70±0.01 | 3.6±0.0 | 98.3±0.3 | 4.5±0.4 |
| anaerobic*1 | 12 | 12.7±0.3 | 0.74±0.03 | 3.6±0.0 | 98.9±0.2 | 1.8±0.2 |
All the wine was fermented by 71B wine yeast.
Values represent the mean of at least three separate assays and standard deviations were calculated.
The findings of this study suggest that certain compounds that originate from grapes are involved in DMTS formation in wine, and that the lysis of yeast cells and subsequent release of cell contents into the must enhance DMTS development. They also indicate that DMTS formation is accelerated when the must is exposed to oxygen. Although the specific mechanisms of DMTS formation during the wine-making process require further investigation, the findings of this study will contribute to reducing DMTS development during storage.
