Abstract
Yeasts isolated from a prefermentation cold maceration (at 8 to 9°C) of Vitis vinifera L. cv. Pinot noir grapes were evaluated for β-glucosidase activity. Among the 13 distinct yeast isolates investigated, five had β-glucosidase activity. Sequence analysis of the D1/D2 domain of the 26S rDNA gene confirmed the identity of these yeasts as Metschnikowia pulcherrima, Hanseniaspora uvarum, Lachancea thermotolerans, and two isolates of Saccharomyces cerevisiae. The β-glucosidase activity of these yeast isolates was further investigated at different sugar concentrations and temperatures before the isolates were used in the production of Pinot noir wines. Pinot noir grapes sterilized by high hydrostatic pressure were inoculated with the individual yeast species or a mixture of all yeast species. A seven day prefermentation cold maceration (9°C) was then performed, followed by alcoholic fermentation conducted by S. cerevisiae RC212 at 27°C. Pinot noir grapes were also cold-macerated without the addition of any yeast, while fermentations were also undertaken with no cold maceration. Solid-phase microextraction-gas chromatography-mass spectrometry analysis of the wines indicated that the presence of the different yeast species during the prefermentation cold-maceration altered the volatile aroma composition of the wines. These yeasts altered the concentrations of ethyl esters, branch-chained esters, higher alcohols, and some monoterpenes in the wines.
- β-glucosidase
- non-Saccharomyces yeast
- Pinot noir
- prefermentation cold maceration
- SPME-GC-MS
- wine volatile aromas
Wine aroma is one of the most important components of wine quality (Ebeler 2001, Swiegers et al. 2005), but it is also one of the most complex, as more than 800 aroma compounds have been identified that contribute to it (Ebeler 2001, Cordente et al. 2012). Because of wine aroma’s importance, much research has been undertaken to understand the factors that impact it during the winemaking process. These include grape varieties, growing conditions, viticultural cultural practices, and winemaking procedures (Ebeler 2001, Mendes Ferreira et al. 2001, Swiegers et al. 2005, Miller et al. 2007). A winemaking procedure that can modify wine aroma is prefermentation cold maceration, commonly known as cold soak (CS) (Jackson 2000). During this process, the grape must is generally held at <10°C for 1 to 14 days to prevent the growth of Saccharomyces cerevisiae and delay the beginning of alcoholic fermentation. Winemakers employ this process for two main reasons: to improve the color of the wine, and/or to modify the flavor and aroma of the wine (Moreno-Perez et al. 2013, Casassa et al. 2015, Aleixandre-Tudo et al. 2016).
Changes in wine aroma due to the CS process have been proposed to result from two different mechanisms. First, volatile compounds or volatile precursor compounds are extracted during the maceration process (Moreno-Perez et al. 2013). Second, cold-tolerant yeast present during this process may affect the concentrations of volatile aroma compounds through their metabolic activity (Zott et al. 2008, Maturano et al. 2015). Microbial flora on the grapes at harvest have been repeatedly shown to alter wine aroma if they persist and grow during alcoholic fermentation (Lema et al. 1996, Garde-Cerdán and Ancín-Azpilicueta 2006, Viana et al. 2008, Zott et al. 2008, Ocón et al. 2010, Comitini et al. 2011). Although Saccharomyces species typically dominate the alcoholic fermentation, the growth of non-Saccharomyces yeasts that are often present in high numbers on the grapes has also been reported to impact wine aroma (Ciani and Maccarelli 1997, Zott et al. 2008, Comitini et al. 2011, Medina et al. 2013). For example, Medina et al. (2013) reported increased aroma and flavor diversity in Chardonnay wines produced by sequential inoculations with Hanseniaspora vineae and S. cerevisiae compared with S. cerevisiae alone. Given the cold tolerance of many of the non-Saccharomyces yeast species (Egli et al. 1998, Zott et al. 2008), it is possible that the growth of these yeast during a CS could significantly modify final wine aroma and flavor.
Aside from the production of primary and secondary metabolites such as esters, alcohols, and volatile acids, yeasts present during a CS may alter wine aroma through production of β-glucosidase enzymes that release glycosidically bound aroma compounds in grape (McMahon et al. 1999, Mendes Ferreira et al. 2001, Swangkeaw et al. 2011). Several studies have reported that non-Saccharomyces yeasts, such as Pichia anomala, Candida pelliculosa, H. uvarum, and several Debaryomyces spp., can have significantly higher β-glucosidase activities than S. cerevisiae (Rosi et al. 1994, Charoenchai et al. 1997). Although many studies of β-glucosidase activities in yeasts have been performed with either model systems or white wines, such as Riesling (Egli et al. 1998), Chardonnay (Egli et al. 1998, Soden et al. 2000), and Albariño (Lema et al. 1996), few have investigated red wines. In addition, most studies using red grapes do not use sterilized grapes or must (Morata et al. 2003, Caridi et al. 2004), but rely on the inoculation of a large population of a selected yeast strain. This practice makes it difficult to draw any conclusions about the specific contribution of each yeast strain or species (Egli et al. 1998), as background microorganisms may contribute to the final wine aroma composition. In this study, Pinot noir grapes were processed by high hydrostatic pressure (HHP) to eliminate background microorganisms from the grapes (Takush and Osborne 2011). These grapes were then used in a series of prefermentation cold-maceration experiments to investigate the impact of yeasts typically present during CS on volatile aroma compounds in Pinot noir wine.
Materials and Methods
Yeasts isolation and identification from prefermentation cold maceration of Oregon Pinot noir grape must
Grapes
Vitis vinifera L. cv. Pinot noir grapes were harvested from a commercial vineyard in Dundee, OR. After harvest, the grapes were stored overnight at 4°C before destemming in a Velo DPC 40 crusher/destemmer. The destemmed grapes were distributed into two 100 L stainless steel tanks, which each contained ~70 to 80 L of grape must. 50 mg/L SO2 was added to the grape must, argon gas was blanketed on top of the grapes, and bladder-equipped tank lids were placed on top of the tanks and sealed.
Prefermentation cold maceration and fermentation
Tanks were placed into a temperature-controlled room and maintained at between 8 and 10°C during eight days of CS. After eight days, the tanks were moved to a temperature-controlled room set at 25°C, and alcoholic fermentation proceeded without inoculation. Brix and temperature were monitored with an Anton-Paar DMA 35N density meter.
Yeast enumeration, isolation, and identification
Samples were aseptically taken daily from each tank with a sampling device prerinsed with 70% (v/v) ethanol. All samples were plated on Wallerstein Laboratory (WL) medium (Difco) supplemented with 0.15 g/L biphenyl (Sigma) and lysine (Difco) media after appropriate dilution in 0.1% (w/v) peptone. All plates were incubated aerobically at 25°C for 48 hrs before yeast colonies were counted and examined for morphological differences. For the WL medium, colonies were described in detail based on color, shape, consistency, and size. Unique colony types were restreaked on WL medium for isolation. Purified colonies were maintained on potato dextrose agar (Difco) slants and stored at 4°C. Glycerol cultures (15% [v/v]) were prepared for long-term storage at −80°C.
DNA sequence analysis
Cultures stored frozen in glycerol were streaked onto yeast extract peptone dextrose (YPD) (10 g/L yeast extract, 20 g/L peptone, 20 g/L dextrose, and 20 g/L agar, pH 6.5) plates and incubated at 25°C for 48 hrs. Single colonies were suspended in 50 μL nuclease-free purified water. The D1/D2 domain of the 5′ end of the large subunit 26S rDNA gene was amplified by direct colony polymerase chain reaction (PCR) in a Thermo Hybaid PCR Express thermocycler with the primers NL1 (5′-GCA TAT CAA TAA GCG GAG GAA AAG-3′) and NL4 (5′-GGT CCG TGT TTC AAG ACG G-3′) as described by Swangkeaw et al. (2011). The PCR was performed with an initial denaturation at 98°C for 30 sec, followed by 30 cycles of denaturation at 98°C for 10 sec, annealing at 66°C for 30 sec, and extension at 72°C for 15 sec. A final extension was performed at 72°C for 10 min. The PCR products were purified with the QIAGEN QIAquick PCR Purification Kit, and Sanger sequencing was performed by the Oregon State University Center for Genome Research & Biocomputing Core Laboratory (Corvallis, OR). Sequences were analyzed in NCBI BLASTN 2.2.26+ (Zhang et al. 2000).
Microsatellite multilocus variable number tandem repeat analysis
Microsatellite multilocus variable number tandem repeat (VNTR) analysis was conducted by ETS Laboratories (Sonoma, CA). Five loci were used to discriminate among yeast isolates. S. cerevisiae isolates from Pinot noir prefermentation cold macerations were compared with those from 80 commercial S. cerevisiae strains.
β-Glucosidase activity assay
Yeast isolates were streaked onto 4-methylumbelliferyl-β-D-glucopyranoside (4-MUG) media (40 mg/L 4-MUG [Sigma], 1.7 g/L yeast nitrogen base (YNB) [Difco], 5 g/L glucose, and 20 g/L agar, pH 5.0). Hydrolysis of the substrate (4-MUG) results in the release of the fluorescent compound 4-methylumbelliferone, causing a blue fluorescent zone surrounding yeast colonies when observed under long-wave UV light.
Yeast isolates that gave positive results on the 4-MUG plates were further assayed for β-glucosidase activity according to the method by Charoenchai et al. (1997). Isolates were streaked from glycerol cultures onto YPD agar plates and incubated at 25°C for 48 hrs. Single colonies were inoculated into 10 mL Wickerman’s malt extract yeast extract glucose peptone medium (3 g/L malt extract, 3 g/L yeast extract, 10 g/L glucose, and 5 g/L peptone, pH 5.5) and incubated for 24 hrs at 25°C. Yeast cells were harvested by centrifugation at 4650g for 10 min and washed with sterile saline buffer twice. The cells were resuspended in sterile saline and inoculated in triplicate into 10 mL of filter-sterilized ρ-nitrophenyl-β-D-glucopyranoside (NPG) medium, containing 6.7 g/L YNB, 5 g/L glucose, 0.9 g/L tartaric acid, 1.0 g/L dibasic potassium phosphate, and 1 mM ρ-NPG (Sigma), buffered at pH 3.5. The cultures were incubated at 8 or 25°C for 48 hrs. These assays were also conducted with media containing 100 g/L glucose and 100 g/L fructose.
After incubation, cultures were centrifuged at 4650g for 10 min. The supernatant (1.0 mL) was mixed with 2.0 mL sodium carbonate (0.2 M, pH 10.2) and the absorbance of the solution was measured at 400 nm on a Thermo Scientific Genesys 10 UV spectrophotometer. To determine dry cell mass, 1.0 mL of the culture was removed prior to enzyme analysis and transferred to preweighed, dry microcentrifuge tubes. The cells were harvested by centrifugation (4650g for 10 min) and washed twice with sterile saline. Cells were then dried for 24 hrs in a 60°C oven and weighed after cooling. Liberated ρ-nitrophenyl was determined with the extinction coefficient of 18,300/M cm. β-Glucosidase activity is reported as nmole ρ-nitrophenyl released per g of dry cells per mL of supernatant.
Effects of isolated yeasts on Pinot noir wines
Grapes
V. vinifera L. cv. Pinot noir grapes were harvested at Oregon State University’s Woodhall Vineyard (Alpine, OR) on 5 Oct 2011, and stored overnight at 4°C. Grapes were destemmed and crushed as described above and then pooled and divided into 3 kg portions. Each portion was placed into a Food Saver bag (Jarden Corp.), and 30 mg/L SO2 was added before the bags were sealed. The grape must was processed by HHP for 5 min at 551 MPa as described by Takush and Osborne (2011). The HHP unit was from National Forge Company and has a 22 L maximum capacity and 689 MPa maximum pressure. The high-pressure intensifier pump was a model 7XS-6000 (Flow International Corporation) and has a maximum pressure of 620 MPa. Post-HHP treatment juice samples were plated on WL (Difco) and modified apple Rogosa (MRS) media (20 g/L tryptone, 5 g/L glucose, 5 g/L yeast extract, 5 g/L peptone, and 20% [v/v] preservative-free apple juice, pH 4.5) containing 100 mg/L cycloheximide (Sigma). Plates were incubated at 25°C for three (WL plates) or seven (MRS medium) days to determine effectiveness of the HHP processing.
Grape must analysis was performed following HHP processing. Titratable acidity (TA) was determined by titration with 0.1 N NaOH and recorded as grams tartaric acid per 100 mL. Brix was determined with an Anton-Paar DMA 35N density meter, and the pH was measured with a Mettler-Toledo Delta 320 pH meter. Yeast assimilable nitrogen (YAN) status was determined by measuring primary amino acids according to Dukes and Butzke (1998) and by measuring ammonia with an enzymatic test kit (r-Biopharm, Darmstadt, Germany). The grape must parameters were: TA of 7.98 g tartaric acid/100 mL, pH 3.53, 22.0 Brix, and 147.6 mg/L YAN.
Microorganisms
Yeast isolates previously isolated from Pinot noir grape must undergoing CS and screened for β-glucosidase activity (Metschnikowia pulcherrima, H. uvarum, Lachancea thermotolerans, S. cerevisiae isolate 1, and S. cerevisiae isolate 2) were streaked from glycerol cultures onto YPD agar plates and incubated at 25°C for 48 hrs. Single colonies were inoculated into 250 mL acidic grape juice broth (25% [v/v] grape juice, 5 g/L yeast extract, 0.125 g/L magnesium sulfate, 0.00275 g/L manganese sulfate, and 0.5% [v/v] Tween, pH 4.5) and incubated at 25°C for 48 hrs. Cells were harvested by centrifugation at 4650g for 10 min and washed once with sterile phosphate buffer (27.8 g/L NaH2SO4•H2O and 28.38 g/L Na2HPO4, pH 7.0). The cells were resuspended in sterile 0.2 M phosphate buffer (pH 7.0) prior to inoculation.
Prefermentation cold maceration
Microscale fermentations (4 L) described by Takush and Osborne (2011) were used in this study. Fermentors were autoclaved at 121°C for 30 min and cooled before use. HHP-treated grape must (3 kg) was aseptically transferred to the fermentors in a laminar flow hood. The must was inoculated with M. pulcherrima, H. uvarum, L. thermotolerans, S. cerevisiae isolate 1, or S. cerevisiae isolate 2 at ~104 cfu/mL. A treatment containing all of the isolates was also prepared with each isolate being inoculated at the same rate as in the individual fermentations (104 cfu/mL). In addition, two sets of three fermentors were not inoculated with any microorganisms. One set of fermentors represented an uninoculated control that would undergo a CS, and the other set was a fermentation treatment that would not undergo a CS. This allowed for a comparison of wines produced from grapes that underwent a CS with or without the presence of yeasts as well as of wines produced from grapes that did not undergo a CS. All treatments were conducted in triplicate. Treatments that underwent a CS were held in a temperature-controlled room at 9°C for seven days. Samples were aseptically taken before and after inoculation and then every 48 hrs. After appropriate dilutions in 0.1% (w/v) peptone, samples were plated on WL medium (Difco) supplemented with 0.15 g/L biphenyl (Sigma).
Microscale Fermentation
After seven days at 9°C, the fermentors were transferred to a temperature-controlled room at 27°C, warmed to room temperature, and inoculated with S. cerevisiae RC212 (Lallemand) at ~105 cfu/mL. RC212 was prepared by streaking from glycerol cultures on YPD media and incubating at 25°C for 48 hrs. A single colony was then inoculated into 250 mL acidic grape juice broth and incubated at 25°C for 48 hrs. Cells were harvested by centrifugation at 4650g for 10 min and washed once with sterile phosphate buffer. The cells were resuspended in sterile 0.2 M phosphate buffer (pH 7.0) prior to inoculation. The treatment that did not undergo a prefermentation cold maceration had previously been inoculated following the same protocol. Prior to inoculation with S. cerevisiae RC212, Fermaid K (Lallemand) (0.45 μm filter-sterilized) was added to all fermentations at 0.25 g/L. Samples were taken before and after inoculation with RC212 and every 48 hrs until alcoholic fermentation was complete.
Viable cell populations were monitored by plating on WL (with addition of biphenyl) and lysine (Difco) media after appropriate dilutions in 0.1% (w/v) peptone. Brix and temperature were monitored with the Anton-Paar DMA 35N density meter. Fermentations were pressed at dryness (<0.5 g/L reducing sugar as measured by Clinitest) with a modified basket press with an applied constant pressure of 0.1 MPa for 5 min. No malolactic fermentation was conducted. 30 mg/L SO2 was added to all wines before settling at 4°C for five days. Samples were taken from each replicate and stored at −80°C for further analysis.
Wine volatile analysis by SPME-GC-MS
A 2 mL aliquot of the wine samples was diluted with 8 mL saturated sodium chloride solution in a clean 20 mL autosampler glass vial to which 20 μL of internal standard solution (containing 43.9 mg/L 4-octanol and 41.7 mg/L octyl propionate) was added. The vials were tightly capped with Teflon-faced silicone septa. A solid-phase microextraction (SPME) fiber coated with divinylbenzene-carboxen-polydimethylsiloxane (DVB/CAR/PDMS) phase (2 cm, 50/30 μm, Supleco) was used for the extraction of wine volatile compounds. The vials were placed into an automatic headspace-sampling system, and samples were preincubated at 50°C for 30 min, then extracted with the SPME fiber for 30 min before desorption.
The SPME fiber–extracted volatile compounds were analyzed with an Agilent 6890 gas chromatograph (GC) equipped with a 5973 mass spectrometry (MS) detector (Agilent Technologies). Compounds separation was achieved by using a ZB-Wax capillary column (30 m × 0.25 mm × 0.5 μm, Phenomenex). Helium was used as the carrier gas at a constant flow rate of 1.5 mL/min, and the inlet temperature was 250°C. The desorption was performed in splitless mode. The oven temperature cycle consisted of an initial 35°C step for 4 min, followed by an increase of 5°C/min to 230°C; the final temperature was held for 10 min. The electron impact energy was 70 eV, and the MS transfer line and ion source temperatures were 230°C. Electron impact mass spectrometric data from m/z 35 to 350 were collected in a scan mode.
Statistical analysis
Statistical analysis of β-glucosidase activity was performed in Microsoft Excel 2008 (ver. 12.3.3) with two-tailed Student’s t tests. Statistical analysis of volatile aroma compounds was performed with SPSS 15 (SPSS, Inc.) with Tukey’s HSD test for mean separation.
Results
Yeast populations were monitored during the CS (days 0 to 8) and alcoholic fermentation (days 9 to 14) of Pinot noir grapes from a commercial vineyard (Figure 1). The initial population density of non-Saccharomyces yeast in the grape must was ~1.7 × 103 cfu/mL (Figure 1). This population increased to ~1 × 105 cfu/mL by the end of the CS and reached a maximum of ~1 × 108 cfu/mL on the third day of alcoholic fermentation (Figure 1). The population of non-Saccharomyces yeasts was dominated by H. uvarum and L. thermotolerans during both CS and alcoholic fermentation (data not shown). Alcoholic fermentation commenced two days after the must was warmed (day 9) and was completed by day 13 (Figure 1).
Grape must samples plated on WL medium were assessed and colonies described in detail based on color, shape, consistency, and size as described by Pallman et al. (2001) and Edwards (2005). Many of the colony descriptions correlated with those reported by Pallman et al. (2001) or Edwards (2005) and allowed tentative identification of M. pulcherrima, H. uvarum, Issatchenkia orientalis, C. oleophila, C. vini, P. membranifaciens, L. thermotolerans, and what was initially identified as Hansenula anomala. Isolates were then screened for β-glucosidase activity on 4-MUG plates and five of these isolates screened positive for activity (Table 1). M. pulcherrima, H. uvarum, and L. thermotolerans had weak activity, and H. anomala and unknown isolate 10 both had strong activity.
The identity of yeast isolates that displayed β-glucosidase activity was assessed by DNA sequencing, which confirmed the results of the WL medium platings for L. thermotolerans, M. pulcherrima, and H. uvarum (Table 1). However, DNA sequencing indicated that the isolate identified as H. anomala on WL medium was in fact S. cerevisiae and that the unknown isolate 10 was also a S. cerevisiae isolate (Table 2). Because of these results, the H. anomala isolate was renamed S. cerevisiae isolate 1 and unknown isolate 10 was renamed S. cerevisiae isolate 2. VNTR analysis of these two isolates showed that they were two distinct S. cerevisiae isolates and that they did not match any of the 80 commercial yeast strains against which they were compared (data not shown).
The β-glucosidase activity of isolates that screened positive for activity on 4-MUG plates was quantified with the ρ-NPG assay (Table 2). All isolates had β-glucosidase activity under the assay conditions of 5 g/L glucose, pH 3.5, and incubation at 25°C. The highest activity under these conditions was observed for H. uvarum, and the lowest was seen for L. thermotolerans and the two S. cerevisiae isolates. β-Glucosidase activity decreased in all isolates when the sugar concentration of the assay media was increased from 5 g/L glucose to 100 g/L glucose and 100 g/L fructose. The greatest reduction was seen for the H. uvarum isolate (99.8%) and the lowest for S. cerevisiae isolate 1 (11%). When the assay temperature was lowered to 8°C, while glucose and fructose concentration and pH were maintained, the β-glucosidase activity of M. pulcherrima was further reduced (Table 2), but the activities of L. thermotolerans and S. cerevisiae isolates 1 and 2 increased (Table 2).
Yeast isolates that demonstrated β-glucosidase activity were subsequently used in microscale fermentations of HHP-treated Pinot noir grapes. Prior to the inoculation and after HHP treatment of the grape must and aseptic transfer to microscale fermenters, HHP-treated grape samples were plated on WL and MRS media, and no microorganisms were detected in the must (data not shown). In addition, no microorganisms were detected in the uninoculated treatments prior to inoculation with S. cerevisiae RC212. In the other treatments, all yeasts grew well during the prefermentation maceration after inoculation with yeast populations for all isolates reaching greater than 5 × 106 cfu/mL after seven days incubation at 9°C (Figure 2). When all the yeast isolates were inoculated together, the yeasts grew well and reached concentrations comparable to those in cultures inoculated with individual isolates (Figure 2). The length of alcoholic fermentation varied depending on the yeast isolate present, although all fermentations were complete within 15 days (Figure 2).
Wines produced from these fermentations were subsequently analyzed for volatile compounds. A selection of key aroma-active compounds, including esters, alcohols, acids, terpene alcohols, and C13-norisoprenoids, were quantified. In general, wines made from grapes that did not undergo a CS had higher levels of ethyl esters than treatments that underwent CS (Table 3). For example, wines made without a CS contained 413.0 μg/L ethyl hexanoate, and wines made with a CS containing no microorganisms had significantly lower concentrations of this ester. There were also significant differences in ethyl esters among wines made with a CS that contained microorganisms versus wines made with a CS containing no microorganisms. For example, wines made with a CS with no microorganisms had significantly lower ethyl hexanoate than many of the wines made with a CS containing various yeast species (Table 3). The concentration of ethyl acetate was the same in all wines except those made with a CS with H. uvarum, which contained significantly higher amounts of this ester. Among the wines made with a single yeast isolate during the CS, wines inoculated with S. cerevisiae isolate 2 had the highest levels of ethyl butanoate, ethyl hexanoate, and ethyl octanoate (Table 3).
We also noted significant differences in concentrations of branch-chained esters, alcohols, and organic acids. Wines made with a prefermentation maceration containing M. pulcherrima had the highest levels of isoamyl acetate and 2-methylbutyl acetate and the second highest level of isobutyl acetate. In contrast, wines made with a CS containing S. cerevisiae isolate 1 or 2 consistently had the lowest amounts of the branch-chained esters. While there were differences between wines made with CS and containing different yeast species, there were also differences among wines that did or did not undergo a CS. For example, all wines made with a CS contained significantly higher concentrations of phenethyl acetate (Table 3) than the wine made without a CS. For the alcohols, the highest concentrations were often measured in wines made without a CS or with a CS in which no microorganisms were present (Table 3).
All treatments had similar concentrations of β-damas cenone and β-ionone. Wines made without a CS generally contained the highest amounts of monoterpenes, whereas those made with a CS with S. cerevisiae isolate 1 or 2 generally contained the lowest amounts of monoterpenes, such as linalool, α-terpineol, and trans-geraniol (Table 3). The exception was β-citronellol because concentrations of this terpene in CS-treated wines with S. cerevisiae isolate 1 or 2 were as much as twice as high as its concentrations in the other wines (Table 3).
Discussion
There is increased interest in the contribution of various yeast species and strains to wine aroma, in particular, in the contribution of non-Saccharomyces yeast that may produce different levels of volatile aroma compounds than does S. cerevisiae. In this study, a total of 13 unique yeast colony types were identified during a CS of Pinot noir grapes. Although other authors have reported a larger number of different yeast species sometimes being present on grapes at harvest (Renouf et al. 2007), the present study focused on those that grew during the CS period, as these yeast species would be most likely to contribute to changes in the composition of the wine. The genera identified were consistent with those observed in winegrape musts by other authors (Jolly et al. 2006, Zott et al. 2008, Ocón et al. 2010), as were their initial population levels.
Despite the addition of 50 mg/L SO2, high population levels of non-Saccharomyces yeasts were present at the beginning of the CS. This was consistent with recent reports of SO2 concentrations of as high as 150 mg/L not completely eliminating the presence of non-Saccharomyces yeasts in Chardonnay grape juice (Bokulich et al. 2015). Aside from the non-Saccharomyces yeasts, two S. cerevisiae isolates that had distinct colony morphology on WL medium were also isolated during the CS. Because of their slow growth on lysine media, they were initially characterized as non-Saccharomyces isolates. However, D1/D2 analysis of their 26S RNA revealed that they were S. cerevisiae. Additional analysis (VNTR) revealed that they were distinct strains that were likely not commercial S. cerevisiae strains.
Total yeast populations increased during prefermentation maceration despite the cold temperatures (8 to 9°C). Growth of non-Saccharomyces yeasts during CS has been reported in previous studies (Jolly et al. 2006, Zott et al. 2008), although a number of these studies were performed at higher temperatures than those used in the present study. The growth and persistence of yeast species throughout the CS suggest that these yeasts could be a source of differences in wine flavor and aroma often noted by winemakers who perform CS. Furthermore, several of the yeasts isolated, including both non-Saccharomyces and Saccharomyces species, displayed β-glucosidase activity. As reported by others (Charoenchai et al. 1997, Swangkeaw et al. 2011), β-glucosidase activity was inhibited by some grape must conditions, such as high sugar concentrations. In the present study, the degree of β-glucosidase inhibition was species dependent, with some species experiencing severe inhibition and others only slight inhibition. In particular, the β-glucosidase activity of H. uvarum was strongly inhibited by a high sugar concentration, while M. pulcherrima, L. thermotolerans, and the two S. cerevisiae isolates were less inhibited and still demonstrated high β-glucosidase activity in the presence of 100 g/L glucose and 100 g/L fructose media. Charoenchai et al. (1997) and McMahon et al. (1999) have previously reported that the β-glucosidase activities of Saccharomyces species were two to three times lower than those of non-Saccharomyces species. While this finding was supported in the present study, this only occurred under conditions of low sugar (5 g/L glucose). When the sugar content was increased, the β-glucosidase activity of the Saccharomyces isolates was among the highest, with only one of the three non-Saccharomyces species (M. pulcherrima) having similar activity.
Aside from high sugar concentration, the other condition present during a prefermentation cold soak that may impact β-glucosidase activity is low temperature. However, in this study, when β-glucosidase assays were conducted at 8°C, an increase in β-glucosidase activity was observed for both Saccharomyces isolates and for L. thermotolerans compared with their activity at 25°C. The increased β-glucosidase activity was significant with an increase of 70% or more. To our knowledge, this is the first report of an increase in yeast β-glucosidase activity due to lower temperatures. Furthermore, this increase was not due to differences in yeast growth, as the assay accounts for differences in yeast growth by expressing activity on a dry cell weight basis.
When yeast isolates with β-glucosidase activity were inoculated into Pinot noir grape must and held at 8 to 9°C, all isolates grew well, with similar increases in population observed. Despite similarities in yeast growth, the concentrations of volatile compounds in the wines produced from each treatment were significantly different. In general, wines that underwent a CS had lower concentrations of ethyl esters but higher concentrations of branch-chained esters. We also noted differences among wines made with a CS with or without microorganisms. These results demonstrate that both the physical process of the CS and the presence and growth of yeast during this process can contribute to differences in volatile aroma among the wines. These differences may have been due to either the production of esters by the yeast species or compositional changes caused by the growth of the yeasts present during the CS.
The high population of yeast may have reduced the nutrients available for the synthesis of ethyl esters by S. cerevisiae RC212. Ethyl esters are derived from medium-chain fatty acids formed by yeast during fatty acid biosynthesis (Malcorps and Dufour 1992), and ethyl ester production is reported to vary with grape nitrogen composition (Miller et al. 2007). In addition, branch-chained esters are derived from the products (alcohols) of the degradation of amino acids, carbohydrates, and lipids, and amino acid composition of the must can impact branch-chained esters production in wine (Miller et al. 2007). Future studies investigating the impact of prefermentation maceration on wine quality should include analysis of the amino acid composition of the grape must at the end of the cold maceration to determine whether the growth of yeast significantly affected the concentration of those amino acids that may impact volatile aroma production by S. cerevisiae.
In addition to the overall trends observed in ester production, the growth of individual yeast isolates during CS resulted in wines with higher concentrations of particular esters. For instance, the treatment inoculated with H. uvarum had the highest concentration of ethyl acetate among all treatments. This finding was not surprising given that previous studies have reported high production of ethyl acetate by H. uvarum (Ciani and Maccarelli 1997, Viana et al. 2008). However, the ethyl acetate concentration in the H. uvarum treatment (64 mg/L) was below the 150 mg/L level considered to produce an off-odor in wine (Garde-Cerdán and Ancín-Azpilicueta 2006). In fact, at lower concentrations, ethyl acetate may be considered a desirable wine aroma (Maarse 1991). The treatment inoculated with H. uvarum also had high concentrations of the aforementioned branch-chained esters and also of isoamyl acetate and isobutyl acetate, along with the treatment inoculated with M. pulcherrima. The impact of non-Saccharomyces yeast on ester production has been reported in previous studies (Lema et al. 1996, Viana et al. 2008), but many of these studies focused on ester production during alcoholic fermentation of white wines. To our knowledge, our study is one of only a few that have reported the impact of non-Saccharomyces yeasts on ester content in Pinot noir. In addition, the present study reports the impact of yeast growth during CS on ester production and is unique in this respect.
Aromatic alcohols were generally higher in the wines that did not undergo a CS or that underwent CS without any microorganisms. Higher alcohols are synthesized by yeasts and are derived from branch-chained amino acids (Swiegers et al. 2005), with their production being impacted by both the amino acid content of the grape must and the yeast strain used (Hernández-Orte et al. 2006). The elevated concentrations of higher alcohols may be due to the presence of lower total yeast population, and therefore to a greater pool of available amino acids. Treatments with increased higher alcohol content received only an inoculation of S. cerevisiae RC212, whereas the other treatments also had been inoculated with one or more yeast species, along with S. cerevisiae RC212. Furthermore, non-Saccharomyces yeasts typically produce lower amounts of higher alcohols than does S. cerevisiae (Jolly et al. 2006) although this may vary between species. This observation is consistent with our findings with the CS treatments. Treatments inoculated with the non-Saccharomyces isolates (M. pulcherrima, H. uvarum, and L. thermotolerans) had lower concentrations of higher alcohols than the treatments inoculated with the S. cerevisiae isolates. This difference included 2-phenyl ethanol, a compound with a characteristic rosy aroma that is an important aroma attributor in Pinot noir (Fang and Qian 2005)
In addition to generating aroma-active compounds during fermentation, some yeasts also possess β-glucosidase activity that may result in the release of sugar-bound, grape-derived aroma compounds (Charoenchai et al. 1997, McMahon et al. 1999, Swangkeaw et al. 2011). The yeast isolates used in this study displayed varying levels of β-glucosidase activity in a broth-based system under CS conditions. When these yeasts were used in the production of Pinot noir, the concentrations of some grape-derived aroma compounds were impacted by the yeast isolate used, but others were not. For example, the concentrations of the C13-norisoprenoids β-ionone and β-damascenone were very similar among all the treatments, but the concentrations of terpene alcohols differed. Although the concentrations of these compounds were within the range of what has been reported for Pinot noir (Fang and Qian 2006), their concentrations were not impacted by yeasts that had previously demonstrated high β-glucosidase activity. In addition, the concentrations of β-ionone and β-damascenone were not impacted by the CS itself, whether yeasts were present or not. Furthermore, two additional volatile aroma compounds that are often glycosylated in grapes, benzyl alcohol and 2-phenyl ethanol, were present at significantly lower amounts in many of the wines made with a CS with or without yeasts present. These findings suggest that performing a CS would not increase the concentrations of these glycosylated compounds in Pinot noir.
In contrast to the findings for the C13-norisoprenoids and benzenoids, differences in the concentrations of terpene alcohols among the treatments were observed. Linalool, α-terpineol, nerol, and geraniol were present in the highest concentrations in wines where a prefermentation cold maceration was conducted without any microorganisms. The concentrations measured were similar to those previously reported in Pinot noir (Fang and Qian 2006). However, the concentration of β-citronellol was as much as twice as high in the treatments inoculated with the β-glucosidase–producing S. cerevisiae isolate 1, S. cerevisiae isolate 2, and a combination of all the yeasts. To our knowledge, this is the first report of yeasts increasing the free terpene alcohol content of Pinot noir. It should be noted, however, that the concentrations of β-citronellol observed, although high for Pinot noir, were still below the published sensory threshold in wine (Guth 1997).
Conclusion
In summary, performing a CS of Pinot noir grapes resulted in significant differences in wine volatile aromas compared with wines made without CS. The volatile aromas were impacted by the CS process itself as well as by the presence and growth of various yeast species. Aside from differences in esters and higher alcohols, the concentrations of certain grape-derived compounds were also impacted. Significant changes in the concentration of the terpene β-citronellol occurred when yeast with demonstrated β-glucosidase activity were present during the CS. Future work should involve sensory evaluation of the wines produced with different yeast isolates present during CS as well as of wines made from fermentations of grape varieties containing higher concentrations of terpene alcohols, such as Riesling.
Acknowledgments
This work was made possible in part or whole by a grant from the Oregon Wine Board. The authors would also like to acknowledge Rich DeScenzo at ETS Laboratories for conducting the VNTR analysis of yeast isolates.
- Received April 2016.
- Revision received September 2016.
- Accepted September 2016.
- Published online December 1969
- ©2017 by the American Society for Enology and Viticulture