Abstract
The potential for synergy between temperature and ethanol as a means to control infections by Brettanomyces bruxellensis in red wines was studied. Using a commercially prepared Merlot wine, we employed a 4 × 5 factorial experimental design with storage temperature (12, 15, 18, or 21°C) and ethanol (12, 13, 14, 15, or 16% [v/v]) as variables. Culturabilities of two B. bruxellensis strains (I1a and F3) isolated from Washington wines (I1a and F3) were monitored for 100 days before concentrations of volatile acidity, 4-ethylphenol (4-EP), and 4-ethylguaiacol (4-EG) were quantified. While growth of both strains was observed in 12 to 15% (v/v) ethanol, lag phase duration generally increased with a decrease in temperature. The two strains displayed similar growth patterns under the various temperature × ethanol conditions, except in the wines containing 15% (v/v) ethanol. At this concentration, F3 exhibited less growth and reduced concentrations of volatile acidity, 4-EP, and 4-EG at the higher temperatures (18 and 21°C) than did I1a. In fact, F3 grew better at 18°C than at 21°C at 15% ethanol. Culturabilities of both strains quickly declined in wines containing 16% v/v ethanol. Wines in which B. bruxellensis reached >106 cfu/mL frequently contained concentrations of 4-EP and 4-EG in excess of 1290 and 155 μg/L, respectively, and thus above the olfactory thresholds for the two compounds. Given the significant interactions between temperature and ethanol, wines containing >13% (v/v) ethanol should be stored at ≤12°C to help limit spoilage by this yeast.
Associated with spoilage of red wines, Brettanomyces bruxellensis is well known to synthesize volatile phenols, most notably 4-ethylphenol (4-EP) and 4-ethylguaiacol (4-EG) (Fugelsang and Edwards 2007). As reviewed by Licker et al. (1998), 4-EP has an aroma described as “smoky,” “medicinal,” or “animal,” whereas the 4-EG aroma has been described as “spicy” or “clove.” Methods evaluated for controlling growth of B. bruxellensis have included use of SO2, dimethyl dicarbonate, and filtration (Barata et al. 2008a, Agnolucci et al. 2010, Curtin et al. 2012, Umiker et al. 2013, Zuehlke and Edwards 2013, Zuehlke et al. 2015).
Two factors known to affect the growth of B. bruxellensis in synthetic media and wine are temperature and ethanol concentration. B. bruxellensis has a growth optimum temperature of 25 to 28°C in synthetic media (Smith 2011), and temperatures above 15°C therefore favor its growth, with growth cessation having been reported above 35°C (Couto et al. 2005, Barata et al. 2008b). However, growth of the yeast has also been observed at temperatures as low as 10°C in both synthetic media and wine (Conterno et al. 2006, Zuehlke and Edwards 2013). Incubation temperature also impacts production of acetic acid (Brandam et al. 2008) and volatile phenols (Dias et al. 2003, Benito et al. 2009). Regarding ethanol, B. bruxellensis can tolerate concentrations as high as 15% (v/v) depending on the strain and conditions (Childs et al. 2015). Although slower growth has been noted at concentrations ≥10% (v/v) (Rodrigues et al. 2001), Ibeas et al. (1997) isolated Brettanomyces from commercial wines containing 15% (v/v) ethanol. Of 30 strains examined, Barata et al. (2008a) reported that 23 exhibited growth up to 14.5% (v/v) ethanol in media, but only two tolerated 15% v/v.
To date, few studies have focused on interactions among control methods as a way to limit spoilage by B. bruxellensis. Interactions known to affect yeast viability include SO2 × dimethyl dicarbonate (Costa et al. 2008) and SO2 × temperature (Zuehlke and Edwards 2013). Dittrich (1977) suggested that a relationship between temperature and ethanol also existed that affected non-Saccharomyces yeasts, but this author did not examine Brettanomyces. Here, the author reported no growth of some yeasts in wines containing 10 to 12% (v/v) alcohol when stored at 8 to 12°C, whereas growth was observed in wines containing 14% (v/v) but held at warmer temperatures. More recently, Sturm et al. (2014) examined the interactive impacts of pH (3.4, 3.7, or 4.0), ethanol (10, 12.5, or 15% [v/v]), and free SO2 (0, 30, or 50 mg/L) on B. bruxellensis.
Thus, the objective of this study was to examine the interactive effects between temperature and ethanol on the growth and spoilage characteristics of B. bruxellensis in a red wine.
Materials and Methods
Yeast starter cultures
B. bruxellensis strains I1a and F3 (Jensen et al. 2009) were acquired from the Washington State University culture collection and grown on Wallerstein Laboratory Differential Medium (WLDM) agar (Difco). Single colonies were transferred into 10 mL of yeast mold broth (pH 3.85, Difco). After incubation at 27°C for four days, 100 μL of the culture was transferred into 50 mL of the same medium containing 5% (v/v) ethanol. Cells were harvested in the late exponential growth phase by centrifuging the samples at 2000 × g for 20 min, washing twice with 0.2 M Na2HPO4 (pH 7.0) buffer, and then resuspending in Merlot wine for starter cultures.
Wine preparation and treatment
A 2011 Merlot wine (pH 3.45, 13.2% [v/v] ethanol) was obtained from a commercial winery. Total SO2 was measured with the aeration-oxidation method (Buechsenstein and Ough 1978) and removed by addition of 30% (v/v) hydrogen peroxide.
A 4 × 5 factorial design with storage temperature (12, 15, 18, or 21°C) and ethanol (12, 13, 14, 15, or 16% [v/v]) as variables was conducted in three triplicates. Fixed volumes with variable proportions of ethanol:water mixtures were added to yield wines with differing ethanol concentrations. Ethanol concentrations were confirmed with an ebulliometer (Alla). After supplementation with 0.5 g/L glucose, 0.5 g/L fructose, and 0.1 g/L yeast extract (Difco), the pH was adjusted to 3.75 with 10 M NaOH. Wines were then sterile-filtered through 0.22 μm membranes (Millipore) into 100 mL sterile milk-dilution bottles. After incubation for 24 hr at 12, 15, 18, or 21°C, all wines were inoculated with 104 cfu/mL B. bruxellensis strain I1a or F3.
Analyses
Wines were sampled twice per week for the first four weeks, and then once per week thereafter. Culturability was determined by plating on WLDM incubated at 27°C using the spread-plate technique and an Autoplate® 4000 spiral plater (Spiral Biotech). At the end of 100 days, volatile acidities were determined with segmented flow analysis (Astoria-Pacific), while 4-EP and 4-EG were measured by gas chromatography-mass spectroscopy (GC-MS). Here, 4-EP and 4-EG were analyzed using a headspace solid-phase microextraction (HS SPME) method with an 85 μm polyacrylate fiber (Supelco). The fiber was thermally desorbed at 280°C for 3 min in the injection port of a GC-MS-MS (Varian model 4000). Separation was achieved with a DB-5MS capillary column (0.18 mm ID × 20 m) with 0.18 μm film thickness obtained from J&W/Agilent Technologies. The carrier gas, helium, was held at a constant flow of 0.8 mL/min. The temperature program consisted of 40°C held for 2.0 min, an increase of 20°C/min to 160°C and held for 0.0 min, followed by an increase of 50°C/min to 300°C and held for 0.2 min. The volatile phenols were identified by retention times, as well as by fragmentation patterns compared with chemical standards.
Analysis of variance and Tukey’s HSD were carried out for mean separation at probability p ≤ 0.05 with XLSTAT software (Addinsoft).
Results
Strains I1a and F3 both grew well in red wines that contained 12% (Figure 1), 13% (Figure 2), and 14% (v/v) (Figure 3) ethanol when incubation temperatures were ≥15°C. As shown in Figure 1, strain I1a required 11, 18, or 30 days to achieve peak populations close to 107 cfu/mL at temperatures of 21, 18, or 15°C, respectively, in wines containing 12% (v/v) ethanol. These trends were similar for strain F3 at the same ethanol concentration. As the ethanol concentration increased, both strains exhibited longer lag phases at lower temperatures. Here, strain I1a reached 106 cfu/mL within 20 days of inoculation into the 12% (v/v) ethanol wine maintained at 15°C, but required ~40 or 70 days to reach the same population when incubated in 13% (Figure 2) or 14% (v/v) ethanol (Figure 3). Culturabilities of both strains decreased and recovery was either slowed (12% and 13% [v/v] ethanol) or nonexistent (14% [v/v] ethanol) in wines incubated at 12°C.
Above 14% ethanol, growth of both I1a and F3 strains was increasingly retarded. On the one hand, strain I1a reached populations >106 cfu/mL 60 days after inoculation in wines containing 15% ethanol when incubated at 21 or 18°C. However, maximum populations were <105 cfu/mL at temperatures ≤15°C (Figure 4A). On the other hand, culturabilities of F3 in 15% ethanol wines declined from 104 to ~102 to 103 cfu/mL by day 20 regardless of incubation temperature and reached only 104 cfu/mL (21 and 18°C) or 103 cfu/mL (15°C) 100 days after inoculation (Figure 4B). In wines with 16% (v/v) ethanol, populations of both strains quickly declined and remained undetectable regardless of incubation temperature (Figure 5).
Concurrent with yeast growth, concentrations of 4-EP, 4-EG, and volatile acidity depended on the peak-culturable populations of B. bruxellensis. As examples, concentrations of 4-EP and 4-EG were ≥1810 and ≥264 μg/L (Table 1) or ≥1290 and ≥155 μg/L (Table 2), respectively, in wines in which populations achieved >106 cfu/mL. Conversely, wines incubated at 12°C, regardless of ethanol concentration or all those with 16% (v/v) ethanol, yielded low peak populations and contained ≤152 μg/L 4-EP or ≤37.7 μg/L 4-EG. At 15% (v/v) ethanol, high concentrations of 4-EP (≥2360 μg/L) and 4-EG (≥331 μg/L) were found in wines inoculated with I1a and incubated at 21 or 18°C, which was in contrast to F3, having much lower levels of 4-EP (≤32.7 μg/L) and 4-EG (≤13.0 μg/L). At this alcohol concentration, neither strain produced 4-EP concentrations of >11.3 μg/L when incubated at ≤15°C. Similarly, volatile acidities ranged from 0.53 to 0.55 g/L in wines that exhibited limited growth of B. bruxellensis, but did achieve >1 g/L in some wines in which populations were >106 cfu/mL. Furthermore, as ethanol concentrations increased from 12 to 14% (v/v), the volatile acidities produced by I1a and F3 increased as incubation temperatures increased from 12 to 21°C. Statistical analyses revealed significant main and interactive effects between temperature and ethanol, and these interactions affected synthesis of volatile acidity, 4-EP, and 4-EG by both strains (Table 3).
Discussion
Growth of B. bruxellensis, synthesis of 4-EP and 4-EG, and volatile acidity all depended on storage temperature and ethanol concentration of the Merlot wines. In general, B. bruxellensis exhibited longer lag phases as the ethanol concentrations of the wine increased from 12 to 15% (v/v) over a range of temperatures (i.e., 15 to 21°C), with a total loss of culturability noted for all wines at 16% (v/v) ethanol. In agreement, Chandra et al. (2014) reported increases in lag-phase duration of 3 to 10 days with increases in ethanol, although growth of their strain ceased at 15% (v/v), in contrast to the strains studied here. In fact, none of the wines stored at 12°C contained more than 152 μg/L 4-EP, suggesting that these wines did not possess sensorily detectable “Brett character” based on the definitions of Licker et al. (1998). Although reducing cellar temperatures is a common industry practice to help inhibit Brettanomyces, this factor limited concentrations of 4-EP, 4-EG, and volatile acidities present, but not necessarily growth, especially when ethanol concentrations were <14% (v/v).
Production of volatile phenols and volatile acidities have been generally observed in infected wines in which populations reached at least 106 cfu/mL, a level commonly associated with late-log or stationary phase (Dias et al. 2003, Harris et al. 2008, Romano et al. 2008). In these wines, the concentrations of 4-EP could be sensorily defined as having “medium” Brett character (Licker et al. 1998). However, some wines reached these populations, yet did not contain excessive concentrations of 4-EP, 4-EG, or volatile acidities (e.g., those containing 12% ethanol and stored at 12°C). While Barata et al. (2008b) reported that volatile phenol production was directly related to population, Fugelsang and Zoecklein (2003) found poor correlation between population (cfu) and 4-EP at any stage of the growth cycle. Rather, these authors noted that 4-EP production correlated with “cumulative cell exposure,” suggesting release of volatile phenols on yeast cell death. Like production of volatile phenols, synthesis of acetic acid depends on the extent of yeast growth but also on the presence of oxygen. In fact, Aguilar-Uscanga et al. (2003) noted that maximum amounts of acetic acid were produced by B. bruxellensis under “moderate” levels of aeration (60 L/hr).
Few studies have examined interactive effects of temperature and ethanol on yeast growth. Earlier work by Sá-Correia and van Uden (1986) suggested that cellular death of Saccharomyces cerevisiae due to high ethanol and/or low temperature is a function of increases in intracellular acidity and losses of certain metabolites (e.g., K+). Later work by Aldiguier et al. (2004) and Couto et al. (2005) also reported synergistic interactive effects between temperature and ethanol on yeast growth, but these authors studied fermentation temperatures of ≥27°C. More recently, Sturm et al. (2014) systematically examined the impact of multiple variables (pH, ethanol, and free SO2) on B. bruxellensis growth. In fact, the media used by these authors were incubated while placed in microtiter plates in which oxygen availability would affect yeast growth and synthesis of such compounds as acetic acid (Ciani and Ferraro 1997, Aguilar-Uscanga et al. 2003). While the conclusions of Sturm et al. (2014) were based on growth in a synthetic microbiological medium, Coronado et al. (2015) observed that B. bruxellensis grows better in wine than in synthetic media despite similar composition. The current results are the first to indicate significant main and interactive effects between temperature and ethanol that affected the metabolism of B. bruxellensis in wine.
Complicating interpretations and comparisons with previous research is the fact that B. bruxellensis exhibits a range of responses to temperature and/or ethanol depending on the strain(s). For example, Medawar et al. (2003) and Conterno et al. (2013) characterized strains that could only endure 13% (v/v), whereas others have described strains that could grow and produce volatile phenols in wines containing >14% (v/v) ethanol (Ibeas et al. 1997, Barata et al. 2008a, Benito et al. 2009). In fact, there does not appear to be a strong correlation between the groupings of B. bruxellensis (genetic clusters a through e) and physiological responses (Conterno et al. 2006). As an example, Conterno et al. (2006) noted that strains within genetic cluster b varied in growth at low temperature, SO2 tolerance, and synthesis of 4-EP. Consistent with these results, strains isolated from a commercial wine from Washington state have variably responded to SO2, dimethyl dicarbonate, and processing methods such as filtration (Umiker et al. 2013, Zuehlke and Edwards 2013, Zuehlke et al. 2015), even though all strains also belong to cluster b.
In general, mechanisms related to ethanol tolerance by yeasts have focused primarily on the roles of unsaturated fatty acids, phospholipids, and sterols such as ergosterol (Stanley et al. 2010). Several authors have noted that increases in unsaturation of fatty acids and/or ergosterol are correlated with increased ethanol tolerance (Alexandre et al. 1994, You et al. 2003, Aguilera et al. 2006, Vanegas et al. 2012). In contrast, Venter et al. (1989) observed that the concentration of saturated fatty acids correlated with high fermentation rates toward the end of fermentation conducted by S. cerevisiae. Although the ethanol tolerances of Hanseniaspora guilliermondii and Zygosaccharomyces bailii depend on the unsaturated fatty acid/sterol composition of membranes (Balterias Couto and Huis in’t Veld 1995, Pina et al. 2004), Pina et al. (2004) could not establish this relationship in Debaryomyces hansenii, H. uvarum, or Torulaspora delbrueckii. Of the few reports on B. bruxellensis, Huffer et al. (2011) noted that the ratio of unsaturated to saturated lipids increases only slightly (from 1.1 to 1.2) on ethanol exposure. Although Renouf et al. (2006) suggested that B. bruxellensis tolerates ethanol better than S. cerevisiae, the mechanisms remain unknown. Other suggested mechanisms of ethanol tolerance have included synthesis of specific proteins (Espinazo-Romeu et al. 2008) or the presence of certain sugars (Torre-Gonzalez et al. 2016).
Like ethanol tolerance, fermentation temperature can alter fatty acid composition of yeasts in order to maintain optimal fluidity of plasma membranes. Depending on the S. cerevisiae strain, Venter et al. (1989) noted that saturated fatty acid compositions generally decrease with a decrease in fermentation temperature. In fact, Torija et al. (2003) concluded that a higher degree of fatty acid unsaturation did not ensure acceptable fermentation performed under low temperatures. Rather, increases in the concentrations of medium-chain fatty acids at the lower temperatures were noted for S. bayanus, suggesting different mechanisms for altering membrane fluidity (Torija et al. 2003). For Z. bailii, Balterias Couto and Huis in’t Veld (1995) noted that a decrease in growth temperature results in a decrease in total unsaturated cellular fatty acids.
Conclusion
Storage temperature and the concentration of ethanol synergistically affected the ability of two strains of B. bruxellensis to grow and synthesize 4-EP, 4-EG, or acetic acid. In the current study utilizing a Merlot wine under vinification conditions, the combination of low storage temperature (≤12°C) and higher ethanol concentration (≥14%) resulted in declines of cell populations to levels below the detection limit of 30 cfu/mL for up to 100 days. However, as culturable cells tend not to produce ethylphenols until late-log or early stationary phase, these findings further suggest that microbial detection studies, rather than volatile phenol analyses, should be performed to assess potential spoilage risk by B. bruxellensis.
Acknowledgments
Sincere appreciation is expressed to Washington Wine Advisory Committee, Lallemand (Montréal, Quebec, Canada), and the School of Food Science for financial and material support.
- Received November 2016.
- Revision received December 2016.
- Accepted January 2017.
- Published online December 1969
- ©2017 by the American Society for Enology and Viticulture