Food Microbiology 26 (2009) 915–921
Contents lists available at ScienceDirect
Food Microbiology journal homepage: www.elsevier.com/locate/fm
Identification and characterization of Dekkera bruxellensis, Candida pararugosa, and Pichia guilliermondii isolated from commercial red wines Susanne L. Jensen a, Nicole L. Umiker b, Nils Arneborg a, Charles G. Edwards b, * a b
Department of Dairy and Food Science, University of Copenhagen, DK-1958 Frederiksberg C, Denmark School of Food Science, Washington State University, Pullman, WA 99164-6376, USA
a r t i c l e i n f o
a b s t r a c t
Article history: Received 11 February 2009 Received in revised form 16 June 2009 Accepted 22 June 2009 Available online 27 June 2009
Yeast isolates from commercial red wines were characterized with regards to tolerances to molecular SO2, ethanol, and temperature as well as synthesis of 4-ethyl-phenol/4-ethyl-guaiacol in grape juice or wine. Based on rDNA sequencing, nine of the 11 isolates belonged to Dekkera bruxellensis (B1a, B1b, B2a, E1, F1a, F3, I1a, N2, and P2) while the other two were Candida pararugosa (Q2) and Pichia guilliermondii (Q3). Strains B1b, Q2, and Q3 were much more resistant to molecular SO2 in comparison to the other strains of Dekkera. These strains were inoculated (103–104 cfu/ml) along with lower populations of Saccharomyces (105 cfu/ml. In wine, Q3 never entered logarithmic growth and quickly died in contrast to Q2 which survived >40 days after inoculation. B1b grew well in wine incubated at 21 C while slower growth was observed at 15 C. Neither Q2 nor Q3 produced 4-ethyl-phenol or 4-ethylguaiacol, unlike B1b. However, lower concentrations of volatile phenols were present in wine incubated at 15 C compared to 21 C. Ó 2009 Elsevier Ltd. All rights reserved.
Keywords: Non-Saccharomyces yeast Wine SO2 4-Ethyl-phenol
1. Introduction Yeasts naturally present in grape musts and wines include Candida, Dekkera, Hansenula, Kloeckera, Kluyveromyces, Pichia, Torulaspora, and several others (Jolly et al., 2006). In general, the activities of many of these yeasts (but not all) are restricted to prefermentation and early stages of alcoholic fermentation, in part due to lower ethanol tolerances compared to Saccharomyces (Jolly et al., 2006). Besides ethanol, other factors such as SO2 (Ough, 1993) temperature (Heard and Fleet, 1988; Gao and Fleet, 1988; Erten, 2002), a lack of oxygen (Hansen et al., 2001), or even cellto-cell contact with high populations of Saccharomyces (Nissen and Arneborg, 2003) can influence the growth of various yeasts. In fact, many of these factors interactively impact yeast growth as evidenced by Gao and Fleet (1988) who reported that the ethanol tolerance for Kloeckera apiculata and Candida stellata was enhanced at lower temperatures. Sulfites also impact yeast growth but efficacy depends on microbial species as well as the concentration of the antimicrobial form, molecular SO2 (mSO2) or SO2$H2O (Ough, 1993). In general, concentrations >1 mg/l mSO2
* Corresponding author. Tel.: þ1 509 335 6612; fax: þ1 509 335 4815. E-mail address:
[email protected] (C.G. Edwards). 0740-0020/$ – see front matter Ó 2009 Elsevier Ltd. All rights reserved. doi:10.1016/j.fm.2009.06.010
are required to inhibit yeasts found in grape musts or wine (Warth, 1985; Barata et al., 2008a). Amongst the various impacts yeasts have on wine quality, some are known to synthesize the volatile phenols, 4-ethyl-phenol and 4-ethyl-guaiacol, from the hydroxycinnamic acids, p-coumaric acid and ferulic acid, respectively (Jolly et al., 2006; Fugelsang and Edwards, 2007). The biochemical transformation is a two step process with an initial decarboxylation of the hydroxycinnamic acids catalyzed by cinnamate decarboxylase and the reduction of the vinyl-phenol intermediates by vinyl-phenol reductase (Sua´rez et al., 2007). While several wine microorganisms can synthesize vinyl-phenols, only Dekkera was thought to possess the ability to produce volatile phenols, making detection a chemical indicator for infections (Licker et al., 1999). However, it is now clear that yeasts other than Dekkera can synthesize these compounds including strains of Pichia guilliermondii and some species of Candida such as Candida helophila, Candida mannitofaciens, and Candida versatilis (Chatonnet et al., 1992, 1995; Dias et al., 2003a,b; Barata et al., 2006; Couto et al., 2006; Suezawa and Suzuki, 2007). These volatile phenols impart sensory characteristics that range from ‘clove’ or ‘spicy’ to ‘mousy,’ ‘wet wool,’ ‘barnyard,’ or even ‘sewage.’ During a recent study in which the selective medium DBDM (Rodrigues et al., 2001) was used to isolate Dekkera from commercial red wines (Umiker, in preparation), two additional
916
S.L. Jensen et al. / Food Microbiology 26 (2009) 915–921
isolates were found that microscopically resembled this yeast but did not belong to the genus. The objectives of this study were therefore to characterize some of the isolates found by Umiker (in preparation) with regards to growth under vinification conditions, in particular the impacts of mSO2, temperature, and ethanol. In addition, growth and the ability to synthesize 4-ethyl-phenol and 4-ethyl-guaiacol in grape juice and wine at two different temperatures was ascertained. 2. Materials and methods 2.1. Yeast strains As part of the study by Umiker (in preparation), a total of 48 red wine samples were obtained from 17 wineries located in Washington State. Wine samples were either directly plated or first centrifuged at 3000g for 30 min prior to transfer to the selective cultivation medium, DBDM which contained 60 ml/l ethanol, 6.7 g/l yeast nitrogen base, 100 mg/l p-coumaric acid, 22 mg/l bromocresol green, 10 mg/l cycloheximide, 20 g/l agar, and pH adjusted to 5.4 (Rodrigues et al., 2001). Once isolated through repeated streaking on DBDM and YM (Difco, Detroit, MI), the strains were identified by PCRs involving NL-1 and NL-4 primers as described by Conterno et al. (2006). PCR products were purified before sequencing following the standard methods used by the Cornell University BioResource Center Sequencing Facility (www.brc.cornell.edu). DNA consensus sequences of the portion of the large subunit of the 26S rDNA gene were submitted to a BLAST search on the National Center for Biotechnology Information web site (http://www.ncbi. nlm.nih.gov/blast). A phylogenetic tree was constructed in MegAlign sequence analysis software using the CLUSTAL W algorithm and default settings. Additional strains used in this research were D. bruxellensis ATCC 52905 obtained from the American Type Culture Collection (Manassas, VA), D. bruxellensis U45738, Dekkera anomala U84244 and Brettanomyces custersianus U76199 from Cornell University (NYSAES, Geneva, NY), and Saccharomyces cerevisiae Syrah from Lallemand (Montre´al, Quebec, Canada). All yeasts were maintained on YM agar slants (pH 6, Difco, Detroit, MI) prepared with 15 g/l agar and kept at 4 C. To prepare inoculums, one colony was transferred to 10 ml YM broth (pH 6) and incubated for 72 h. Cells were harvested by centrifugation at 3000g for 30 min and resuspended in various amounts of YM broth, red grade juice, or red wine depending on the experiment to reach the desired initial populations. 2.2. Inhibition by SO2 YM medium (6 ml, pH 3.5) was aseptically transferred to sterile 13 100 mm screw cap test tubes and varying amounts of potassium metabisulfite was added to yield 0, 0.2, 0.4, 0.6, and 0.8 mg/l mSO2 (in duplicate). The concentrations of mSO2 present were calculated based on medium pH and measurement of free SO2 by aeration/oxidation (Buechsenstein and Ough, 1978) just prior to inoculation. Yeast growth was monitored by Klett–Summerson photoelectric colorimeter (Klett Manufacturing Co., New York, NY) with a green filter. Infinite inhibitory concentrations (IIC) were calculated based on Marwan and Nagel (1986) using 70 Klett units as the cutoff value.
0.5 mm nominal filter pads (Gusmer Enterprises, Fresno, CA) before sterile filtration through a 0.45 mm absolute membrane (Millipore, Billerica, MA). The red wine was a blend of Cabernet Sauvignon, Merlot, and Syrah wines (pH 3.88, 13.0% v/v alcohol) and was sequentially passed through a 0.4 mm filter pad (CeluporeÒ 1935SD, Gusmer Enterprises, Fresno, CA) and then a 0.45 mm absolute membrane before inoculations. Fermentations (180 ml) were carried out in triplicate 250 ml Erlenmeyer flasks fitted with airlocks filled with sterile distilled water. Enough inocula of S. cerevisiae and/or non-Saccharomyces yeasts were added to yield populations of 100–500 cfu/ml and 103– 104 cfu/ml, respectively. The flasks were incubated at 15 C and 21 C with fermentations periodically sampled using sterile pipettes. Yeast populations were determined using an Autoplate 4000 spiral plater (Sprial Biotech Inc., Norwood, MA) with both YM and WLC media, the latter being WL agar (pH 5.5, Difco) with 10 mg/l cycloheximide added. All plates were incubated at 27 C. Difference counting between WLC and YM was used to enumerate fermentations co-inoculated with Saccharomyces and Dekkera, Candida, or Pichia. However, Candida and Dekkera were easily distinguished on YM through formation of small, pinpoint colonies in contrast to the large colonies of Saccharomyces observed after two days. Once populations reached late stationary phase, juices and wines were filtered through 0.45 mm absolute membranes and stored at 30 C prior to analysis of volatile phenols. 4-Ethylphenol and 4-ethyl-guaiacol were analyzed using a headspacesolid phase microextraction method with an 85 mm polyacrylate fiber (Supelco, Bellefonte, PA). The fiber was thermally desorbed at 280 C for 3 min by the injection port of a GC–MS/MS (Varian model 4000, Walnut Creek, CA). Separation was achieved using a DB-5MS capillary column (0.18 mm ID 20 m) with 0.18 mm film thickness obtained from J&W/Agilent Technologies (Wilmington, DE). 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, increased 20 C/min to 160 C and held for 0.0 min, and then increased 50 C/min to 300 C and held for 0.2 min. The volatile phenols were identified by retention times as well as fragmentation patterns compared to chemical standards. Fisher’s LSD was used for mean separation. 2.4. Impact of ethanol temperature on growth YM medium (6 ml, pH 5.8) was aseptically transferred to sterile 13 100 mm screw cap test tubes and varying amounts of ethanol was added to yield 0, 4, 8, 12, or 14% v/v. Strains were inoculated at 104 cfu/ml and media incubated at 10 C, 15 C, 18 C or 22 C. Turbidity was monitored by Klett–Summerson photoelectric colorimeter and defined as no growth (maximum turbidity achieved 60 Klett units). All growth conditions were conducted in triplicate. 2.5. Statistical analyses All statistics utilized analysis of variance (ANOVA) followed by Fisher’s least significant difference procedure to evaluate significant sources of variation amongst means. All tests of significance were conducted at a probability level of p 0.05. 3. Results
2.3. Growth and metabolism in grape juice and wine 3.1. Yeast strains Red grape juice was prepared from Syrah grapes (pH 3.59, 23.1 Brix) and kept frozen until use (70 C). Once thawed, the juice was sequentially filtered through 1 mm, 0.8 mm, and then
Of the 48 samples received, most were either red wine blends (18) or Cabernet Sauvignon (15) but additional samples of Syrah (8),
S.L. Jensen et al. / Food Microbiology 26 (2009) 915–921 Table 1 Yeast strains obtained from commercial red wines initially suspected of belonging to genus Brettanomyces. Strain number
Winery
B1a, B1b B2a E1 F1a F3 I1a N2 P2 Q2 Q3
Winery Winery Winery Winery Winery Winery Winery Winery Winery Winery
Wine lot B B E F F I N P Q Q
Syrah Syrah Cabernet Cabernet Merlot Merlot Cabernet Other Cabernet Syrah
Sauvignon Sauvignon
Sauvignon Sauvignon
Merlot (5), Cabernet franc (1) and Zinfandel (1) were also analyzed. Based on growth on DBDM agar with a possible ‘‘phenolic’’ odor and general microscopic appearance, 11 isolates were tentatively identified as being Dekkera (Table 1). To confirm identification, DNA sequences were aligned for comparison with NCBI 26S rDNA sequences for D. bruxellensis, D. anomala, and B. custersianus (Fig. 1). Of the 11 isolates, nine were found to be D. bruxellensis (B1a, B1b, B2a, E1, F1a, F3, I1a, N2, and P2). Isolates Q2 and Q3 were found to be strains of C. pararugosa and P. guilliermondii, respectively. 3.2. Inhibition by SO2 As illustrated in Fig. 2, strains B1a, B2a, E1, F1a, F3, I1a, N2, and P2 exhibited infinite inhibitory concentrations (IIC) for mSO2 that ranged between approximately 0.3 and 0.5 mg/l. Within this group, strain N2 was more sensitive (IIC ¼ 0.29 mg/l) while B2a less sensitive (IIC ¼ 0.46 mg/l). However, strains B1b, Q2, and Q3 were more resistant to mSO2 as demonstrated by relatively high IIC values of 1.78, 1.48, and 1.34 mg/l, respectively. 3.3. Growth and metabolism in grape juice and wine In general, Saccharomyces achieved populations in excess of 107 cfu/ml without or with co-inoculation of D. bruxellensis B1b, C. pararugosa Q2, or P. guilliermondii Q3 in grape juice (Fig. 3). Incubation at 15 C slightly slowed Saccharomyces from reaching maximum populations in comparison to fermentations conducted at 21 C. Both D. bruxellensis and C. pararugosa grew well in juice, achieving viability approaching or in excess of 106 cfu/ml with slightly better growth at 21 C. Although P. guilliermondii quickly reached >106 cfu/ml at 15 C, the strain died within days after inoculation in the 21 C fermentation. D. bruxellensis was the only yeast to produce volatile phenols in grape juice at either 15 or
917
21 C (Table 2), with concentrations
