Abstract
This study investigated the effect of select non-Saccharomyces yeast strains on Hanseniaspora uvarum growth and acetic acid and ethyl acetate production during prefermentation cold soak. We tested commercially available non-Saccharomyces yeasts for their ability to reduce H. uvarum growth and acetic acid production during a simulated cold soak in a grape juice-based medium. All tested non-Saccharomyces yeast reduced H. uvarum growth and acetic acid production, with some yeast having a greater impact than others. Following the screening of non-Saccharomyces yeast, we tested 14 different H. uvarum isolates against a selected non-Saccharomyces yeast, Metschnikowia fructicola, and found that all H. uvarum isolates had reduced growth and acetic acid production when grown in co-culture with M. fructicola, with variation between isolates noted. Finally, we evaluated the effect of M. fructicola on H. uvarum during prefermentation cold soak of Pinot noir grapes. Pinot noir grapes were inoculated with a combination of H. uvarum and M. fructicola and cold soaked for six days at 8°C. At the end of cold soaking, treatments inoculated with M. fructicola contained lower populations of H. uvarum and significantly lower acetic acid and ethyl acetate concentrations compared with treatments not inoculated with M. fructicola. After the completion of alcoholic fermentation, wines where M. fructicola was added contained significantly lower ethyl acetate but no differences in acetic acid concentration. These results suggest that adding select non-Saccharomyces yeast may be another method to reduce the risk of spoilage by H. uvarum during prefermentation cold soaking.
Prefermentation cold soaking, commonly known simply as cold soaking, is a winemaking method where red grapes are harvested and held at cold temperatures (5 to 10°C) for several days prior to the start of alcoholic fermentation (Sacchi et al. 2005). This process may improve phenolic extraction and color stability in red wine (Casassa and Sari 2015, González-Neves et al. 2015) and can alter the concentrations of wine aroma compounds such as acetate esters, ethyl esters, and higher alcohols (Moreno-Pérez et al. 2013, Hall et al. 2017). Cold soaking may positively contribute to wine quality, but it can also increase the risk of microbial spoilage because the conditions of cold soak can allow the growth of spoilage yeast such as Hanseniaspora uvarum (anamorph Kloeckera apiculata) (Mendoza et al. 2009). H. uvarum is often the most abundant yeast on grapes at the time of harvest and can produce large quantities of ethyl acetate and acetic acid (Domizio et al. 2011). Ethyl acetate is a common wine ester that lends pleasant fruity aromas at concentrations below 50 mg/L (Zoecklein et al. 1995). Above this concentration, ethyl acetate can mask other aromas and is considered a sensory defect at concentrations >150 mg/L (Jackson 2014). High concentrations of acetic acid can also cause sensory defects in wine, lending harsh acidic flavors and vinegar-like aromas above 0.7 g/L (Jackson 2014).
Though H. uvarum is commonly present in fresh grape must and juice, growth of this yeast is typically suppressed by adding sulfur dioxide (SO2) during grape processing, with higher doses recommended if grapes are damaged (Divol et al. 2012). Aside from SO2, other chemical preservatives such as chitosan and dimethyl dicarbonate may offer some protection from spoilage microorganisms such as H. uvarum (Delfini et al. 2002, Elmaci et al. 2015). While these preservation methods can be effective at decreasing H. uvarum growth, their use can also result in losing beneficial non-Saccharomyces yeast, which are increasingly associated with improved wine quality and complexity (Jolly et al. 2014, Liu et al. 2017). For example, Henick-Kling et al. (1998) observed that wines made with 50 mg/L SO2 had noticeably less pronounced fruity aromas than wines made without SO2 addition. These changes were attributed to decreases in non-Saccharomyces populations upon SO2 addition. Furthermore, there is a trend toward reducing the addition of SO2 during winemaking due to potential health concerns (Vally et al. 2009). For these reasons, among others, there has been recent interest in using protective microbial cultures as an alternative means of preventing prefermentation wine spoilage (Ciani et al. 2016). This concept has been termed “biocontrol” or “bioprotection” and, although relatively new to the wine industry, is a common practice in other food production industries (Gaggia et al. 2011, Spadaro and Droby 2016).
To date, studies investigating antagonistic non-Saccharomyces yeast have primarily focused on isolating and characterizing novel species (Ciani et al. 2016). Limited work has been done to determine the effectiveness of antagonistic yeast to reduce microbial spoilage during wine production. Studies focused on spoilage prevention have investigated non-Saccharomyces yeast as a means of limiting acetic acid accumulation during sweet wine production (Bely et al. 2008, Rantsiou et al. 2012) or limiting volatile phenol production by Brettanomyces bruxellensis (Mehlomakulu et al. 2015). Simonin et al. (2018) reported on the biocontrol activity of Torulaspora delbrueckii during red wine production but only monitored differences in microbial populations and not the production of spoilage compounds. Others have looked at remediating spoilage by refermenting high acetic acid wines with fresh grape must (Vilela-Moura et al. 2008).
Several commercially available non-Saccharomyces cultures are suitable for use as biocontrol cultures during prefermentation cold soaking. However, very few reports in scientific literature report the effectiveness of these cultures during cold soaking. Therefore, the objective of this study was to investigate the ability of commercially available non-Saccharomyces wine yeasts to suppress H. uvarum growth and acetic acid and ethyl acetate production during prefermentation cold soaking.
Materials and Methods
Microorganisms
H. uvarum strains OSU2a and OSU2b were sourced from Oregon State University and initially isolated from Pinot noir grapes undergoing cold soak (Hall et al. 2017). H. uvarum strains UCD997, UCD998, UCD1000, UCD2100, UCD2207, UCD2117, UCD2119, UCD2832, UCD2853, UCD2968, and UCD3876 were obtained from the Viticulture and Enology Culture Collection housed in the University of California, Davis, Department of Viticulture and Enology. H. uvarum strain MUCL31704 was obtained from BCCM/MUCL Agro-food & Environmental Fungal Collection (Louvain-la-Neuve, Belgium). Zymaflore Egide (mixture of Metschnikowia pulcherrima and T. delbrueckii) and T. delbrueckii Zymaflore Alpha were sourced from Laffort; Lachancea thermotolerans Concerto and T. delbrueckii Prelude were sourced from Chr. Hansen; T. delbrueckii Primaflora VB Bio, M. pulcherrima Primaflora VR Bio, and L. thermotolerans Levulia Alcomeno were sourced from AEB-USA; and Metschnikowia fructicola Gaïa and T. delbrueckii Biodiva were sourced from Lallemand.
Yeast were grown in yeast peptone dextrose broth (10 g/L yeast extract, 20 g/L peptone, 20 g/L D-glucose, pH 6.5) and streaked for single colonies on Wallerstein Laboratory (WL) differential agar (Difco). Single colonies were inoculated in yeast peptone dextrose broth and incubated for five days at 25°C prior to storage at -80°C in 25% glycerol. Commercial yeast strains were isolated from commercially available lyophilized cultures. Other yeast cultures were isolated from agar slants. Yeast cultures used for the screening assays were grown in De Man, Rogosa, and Sharpe broth (20 g/L tryptone, 5 g/L peptone, 5 g/L yeast extract, 5 g/L D-glucose, 1.0 mL/L 5% [w/w] Tween 80, 200 mL/L apple juice, pH 4.50) for five days at 25°C. Yeast cultures used for Pinot noir winemaking experiments were grown in acidic grape juice broth (0.0025 g/L MnSO4, 0.125 g/L MgSO4, 5 g/L yeast extract, 1.0 mL/L 5% Tween 80, 250 mL/L white grape juice [Santa Cruz Organic], pH 3.50) for five days at 25°C. Cultures were concentrated by centrifugation (4200 rpm for 5 min), followed by resuspension in a pH 7.0, 0.05 M sodium phosphate buffer. Cultures were then immediately used for inoculation.
Screening assays
All screening assays were performed in a grape juice medium (MGJ), which was prepared by adding 44.5 g/L D-glucose, 44.5 g/L D-fructose, and 2.7 g/L yeast extract to white grape juice (Santa Cruz Organic). The resulting MGJ had a sugar content of ~23 Brix, 150 mg/L yeast assimilable nitrogen (YAN), and a pH of 3.40. Sugar content was assessed by an Anton-Paar DMA 35N. YAN was calculated by assessing the primary-amino nitrogen content using the NOPA assay (Dukes and Butzke 1998) and by the enzymatic analysis of ammonia (Boehringer). pH was adjusted using 25% (v/v) phosphoric acid and measured by an ion selective electrode (Thermo Scientific). MGJ was sterile-filtered through a 0.45 μm polyethersulfone membrane (Nalgene) into sterile media bottles and then aliquoted (20 mL each) into sterile culture-tubes (PYREX No. 9826, Corning). Fresh MGJ was prepared for each assay and stored overnight at 8°C prior to use.
For the assessment of yeast-yeast interactions, H. uvarum strains and commercial non-Saccharomyces yeast cultures were prepared as previously detailed. H. uvarum was inoculated either alone or co-inoculated with single commercial non-Saccharomyces yeast culture in MGJ. All yeast were inoculated at ~106 cfu/mL and incubated for six days at 8°C. Culturable yeast cell populations were determined on Days 0 and 6 by plating on WL agar with or without 10 mg/L cycloheximide (Sigma), after suitable dilution. At the completion of the experiment (Day 6), samples were taken and frozen at -20°C until needed for acetic acid analysis. Acetic acid was measured enzymatically (Vintessential).
Grape processing
Pinot noir wine was produced at the Oregon State University Research Winery from Vitis vinifera L. cv. Pinot noir grapes harvested at Woodhall Vineyard (Alpine, OR, USA) on 21 Sept 2017. Grapes were kept overnight in refrigerated storage (8°C) prior to being sorted, destemmed, pooled, and aliquoted (2.5 kg each) into autoclaved microfermenters, as described by Takush and Osborne (2012). Samples were taken, and basic grape must analysis was performed. pH was measured by ion-selective electrode (Thermo Scientific). Titratable acidity (TA) was determined by titration with 0.1 N NaOH. Brix was measured by an Anton-Paar DMA 35N. YAN was measured by assessing the primary-amino nitrogen content using the NOPA assay (Dukes and Butzke 1998) and by the enzymatic analysis of ammonia (Boehringer). The basic parameters of the Pinot noir grapes were 22.1 Brix, pH 3.51, TA 4.35 g/L, and YAN 179.6 mg/L.
Winemaking
Prior to cold soaking, the grape must was inoculated with ~103 or 106 cfu/mL H. uvarum OSU2a either alone or co-inoculated with M. fructicola Gaïa at ~106 cfu/mL. H. uvarum OSU2a and M. fructicola Gaïa were prepared from frozen stock cultures, as previously described. A treatment was also prepared with only M. fructicola Gaïa added to the grape must. All treatments were performed in triplicate. Microfermenters were held at 8°C for six days in a walk-in cooler, and daily samples were taken aseptically and either analyzed immediately or stored at -80°C until required for further analysis. H. uvarum populations were monitored daily throughout cold soak and for the first four days of alcoholic fermentation by plating on WL agar supplemented with 150 mg/L biphenyl (Sigma) and 10 mg/L cycloheximide (Sigma) or on WL agar with biphenyl alone. H. uvarum were identified based on growth and appearance on WL media (Pallmann et al. 2001). Brix was monitored daily using an Anton-Paar DMA 35N.
At the end of cold soak, fermenters were warmed to 27°C in a temperature-controlled room. A “no cold soak” control treatment was also performed in triplicate, starting on the first day of grape processing. All treatments were inoculated with S. cerevisiae RC212 at a rate of ~106 cfu/mL, prepared as previously described. Fermaid K (0.25 g/L; Lallemand) was also added at this point. Brix was monitored daily using an Anton-Paar DMA 35N; after the completion of alcoholic fermentation (<0.5 g/L reducing sugar as measured by Clinitest [Bayer]), grapes were pressed using a modified basket press with an applied constant pressure of 0.1 MPa for 5 min. Pressed wines were transferred to 2-L glass bottles with airlocks, and samples were taken immediately and stored at -80°C until required for analysis.
Yeast community ITS meta-barcoding analysis
Yeast community compositions were determined at the beginning and end of cold soak by high-throughput sequencing of pooled internal transcribed spacer (ITS) region polymerase chain reaction (PCR) amplicons, as described by Comeau et al. (2017) with modifications. Juice samples used for DNA extraction were taken prior to inoculation with the non-Saccharomyces yeast strains (Day 0) and again on the last day of cold soak (Day 6) prior to inoculation with S. cerevisiae RC212. DNA extractions were performed on replicate samples using the DNeasy PowerFood kit (Qiagen) and an Omni Bead Ruptor 24 (Omni International, Inc.), with 10 cycles of a 15 sec pulse at 8 m/sec and a 55 sec rest. Barcoded amplicon libraries were generated in duplicate by PCR in 96-well plates using pairwise combinations of forward (BITS) and reverse (B583) ITS primers (Bokulich and Mills 2013) (Supplemental Table 1) and the Platinum Hot Start PCR Master Mix (Thermo Fisher). DNA extractions and PCR reactions were validated using the ZymoBIOMICS Microbial Community Standard and Microbial Community DNA Standard (Zymo Research Corp.). Duplicate libraries were pooled, cleaned, and normalized using a SequalPrep 96-well plate (Applied Biosystems). Amplicon sequencing was carried out by the Oregon State University Center for Genome Research and Biocomputing on the Illumina MiSeq, with 2 × 300 bp v3 chemistry.
Computational analysis of the paired amplicon sequence data was carried out using QIIME1 and python scripts by following the Microbiome Helper virtual box v1 SOP workflow (Comeau et al. 2017), with minor adaptation for ITS analysis using BITS and B58S3 primers. Briefly, paired amplicon sequences were stitched, combined, and then filtered by quality and length (quality score 30, 100 bp cut-off, 90% threshold). Chimera identification was performed using the uchime ITS reference dataset (Nilsson et al. 2015). Microbiome Helper QIIME1 Python scripts were used to create a table of operational taxonomic units (OTUs) from nonchimeric sequences. Taxonomic identities were assigned using the UNITE open source database with a cut-off of 0.1% low confidence OTUs. The resulting OTU reads were then rarified at 2000 counts per sample, and the OTU table was exported for further analysis.
Ethyl acetate analysis
Ethyl acetate was quantified using a headspace (HS) gas chromatography (GC) with flame ionization detection (FID) method because of its high concentrations in wine. One mL of juice or wine sample was pipetted into a 20 mL autosampler vial, and 20 μL of the internal standard (2.5 mg/mL of methyl propionate) was added. The vial was tightly capped with a Teflon-faced silicone septa. Sample mixture was then incubated at 50°C for 15 min with 250 rpm agitation to achieve the equilibrium between the sample mixture and HS. A Varian CP-3800 GC equipped with a flame ionization detector (Varian, Inc.) was used for sample analysis. The GC system was combined with a CombiPAL autosampler (CTC Analytics) coupled with a 1 mL syringe (Hamilton Company) for sample injection. The syringe was also kept at 50°C, the same as the incubation temperature. After equilibration, 0.5 mL of sample HS was taken by syringe and directly injected into the GC system with 1:10 split ratio. Separation of analytes was achieved by a DB-WAX capillary column (30 m × 0.25 mm × 0.5 μm; Agilent Technologies., Inc). A constant nitrogen flow of 2 mL/min was applied. The initial oven temperature was 35°C, held for 4 min, then ramped up to 150°C at a rate of 10°C/min and held for 5 min. The inlet temperature was 200°C and the FID temperature was 250°C.
Peak identification of the volatile components was conducted by comparing retention time with authentic pure standards. Standard curves for grape juice and wine analysis were built with either devolatilized juice or synthetic wine (3.5 g/L tartaric acid, 12% v/v ethanol, pH 3.5) containing known amounts of the analytes. For juice devolatilization, 2 g of Li-Chrolut EN (Sigma) was stirred with 200 mL of grape juice collected at the time of processing for 17 hr and then removed by suction dilution. Standard calibration curve was obtained through GC software and used to calculate the concentrations of ethyl acetate in the samples.
Data analysis
Statistical analysis was conducted using either SPSS version 20 (SPSS, Inc.) or JMP Pro version 13.0 (SAS Institute, Inc.) and a significance value (α) of 0.05. For the screening experiments, statistical differences among treatments were determined by analysis of variance, followed by Tukey’s honest significant difference if appropriate. A two-sample student’s t-test was used to determine the influence of M. fructicola addition on production of acetic acid and ethyl acetate during Pinot noir winemaking.
Results
Commercially available non-Saccharomyces yeast cultures were tested for their effect on H. uvarum growth and acetic acid production under cold-soak conditions (8°C for six days) in an MGJ. Co-inoculation of H. uvarum OSU2a with any of the tested non-Saccharomyces yeast resulted in significantly lower acetic acid concentrations at the end of cold soak compared with the H. uvarum monoculture control (Table 1). An average reduction of 83 mg/L acetic acid during cold soak was measured, though notable differences in the extent of acetic acid reduction were observed between yeast strains. For example, a 52% reduction in acetic acid occurred when H. uvarum OSU2a was co-inoculated with L. thermotolerans Concerto, whereas only a 29% reduction occurred when H. uvarum OSU2a was co-inoculated with T. delbrueckii Biodiva (Table 1). H. uvarum OSU2a growth was also lower in the co-inoculated treatments, except for M. fructicola Gaïa (Table 1). Overall, co-inoculation with a non-Saccharomyces yeast resulted in an average 39% reduction in H. uvarum OSU2a growth, with the greatest impact from T. delbrueckii Alpha TD and L. thermotolerans Concerto (Table 1).
Changes in Brix, yeast assimilable nitrogen (YAN), acetic acid, and growth of Hanseniaspora uvarum OSU2a in monoculture or co-culture with various commercial non-Saccharomyces cultures after a six-day cold soak in grape juice medium.
Inoculation of H. uvarum OSU2a into the MGJ resulted in a decrease in Brix and YAN by the end of the six-day simu lated cold soak. On average, YAN concentrations decreased from ~150 mg/L in the juice prior to cold soak to 103 ± 26 mg/L at the end of cold soak (Table 1). At the end of cold soak, no significant differences in YAN concentrations were observed between the H. uvarum monoculture and co-culture treatments. In contrast, significant differences in sugar utilization were observed between the co-culture and monoculture treatments (Table 1). Except for Zymaflore Égide, co-inoculated treatments had significantly lower concentrations of sugar at the end of cold soak compared with the H. uvarum monoculture control. This finding was particularly true in the treatment containing L. thermotolerans Concerto, which experienced a reduction from ~23 to 17.4 Brix at the end of cold soak (Table 1).
M. fructicola Gaïa was used to test if similar reductions in acetic acid and H. uvarum growth would be observed across several different H. uvarum strains. To do so, 14 strains of H. uvarum were grown with or without M. fructicola Gaïa under the same model conditions described previously. Up to a two-fold difference in acetic acid production was observed between H. uvarum strains when grown alone (Table 2). Co-inoculation with M. fructicola Gaïa resulted in a significant decrease in acetic acid concentrations in 10 of the 14 H. uvarum co-cultures compared with their monoculture counterpart (Table 2). The extent of acetic acid reduction by M. fructicola correlated (r = 0.949, p < 0.0001) with the amount of acetic acid produced in the H. uvarum monoculture treatments, meaning H. uvarum strains that produced the highest concentrations of acetic acid in monoculture experienced the greatest reduction in acetic acid production when grown in co-culture. For example, we observed a 41% reduction in acetic acid production for the highest acetic acid-producing H. uvarum strain (UCD997), while the lowest acetic acid-producing strain (UCD2110) experienced an insignificant 4% reduction (Table 2).
Change in acetic acid concentration and growth of Hanseniaspora uvarum in monoculture or co-culture with Metschnikowia fructicola Gaïa after a six-day cold soak in grape juice medium.
Similar to acetic acid production, the effect of M. fructicola Gaïa on H. uvarum growth varied between H. uvarum isolates (Table 2). Compared to growth in monoculture, growth of H. uvarum isolates was significantly reduced when grown in co-culture with M. fructicola Gaïa. For example, when isolate H. uvarum OSU2a was grown in monoculture, there was a 1.1 ± 0.1 log increase in cfu/mL during the six-day cold soak. However, when grown in co-culture with M. fructicola Gaïa, only a 0.33 ± 0.1 log increase in cfu/mL occurred (Table 2). All H. uvarum isolates, except for MUCL31704, had reduced growth when co-inoculated with M. fructicola Gaïa (Table 2), with no increase in H. uvarum populations observed in the OSU2b and UCD2853 co-culture treatments (Table 2).
M. fructicola Gaïa had a similar impact on H. uvarum growth and acetic acid production during cold soaking of Pinot noir grapes. Background H. uvarum present on the grapes resulted in H. uvarum populations of ~105 cfu/mL at the start of cold soak, even when no H. uvarum OSU2a was added (Figure 1A). Adding M. fructicola Gaïa to the Pinot noir cold-soak treatments resulted in an ~1 log reduction in H. uvarum growth in the treatment where no addition (Figure 1A) or a low addition (~103 cfu/mL; Figure 1B) of H. uvarum OSU2a had been made. In both treatments, total H. uvarum populations rose from ~105 cfu/mL to 1.5 × 107 cfu/mL if M. fructicola Gaïa was not added, but they reached only ~1.5 × 106 cfu/mL if M. fructicola Gaïa was added (Figure 1A, 1B). The impact of M. fructicola on H. uvarum growth was less pronounced if the initial population of H. uvarum was high (106 cfu/mL), though differences in populations were observed (Figure 1). In this case, H. uvarum populations at the end of cold soak were 4.3 × 106 cfu/mL without M. fructicola Gaïa and 2.0 × 106 cfu/mL with M. fructicola Gaïa. H. uvarum populations increased in all treatments when the temperature of the fermentations was raised (8°C to 27°C) at the end of the six-day cold soak (Figure 1). The trend seen at the end of cold soak carried over into alcoholic fermentation, where treatments without M. fructicola Gaïa reached higher H. uvarum populations than those treatments where M. fructicola Gaïa was added (Figure 1).
Hanseniaspora uvarum culturable cells during cold soak and alcoholic fermentation of Pinot noir grapes with or without the addition of Metschnikowia fructicola Gaïa and inoculated with (A) 0, (B) 103, or (C) 106 cfu/mL of H. uvarum OSU2a. The transition between cold soak and alcoholic fermentation is denoted by a vertical line. Error bars represent standard deviation from the mean.
ITS metabarcoding analysis showed that the fungal community present in the initial grape must prior to cold soaking and yeast inoculations comprised ~67% H. uvarum and 7% Metschnikowia species, with the other 27% corresponding mainly to filamentous fungi of the genera Cladosporium (Figure 2). In agreement with the plating data, adding M. fructicola Gaïa reduced the proportion of H. uvarum populations in the high and low H. uvarum treatments present at the end of the cold-soak period compared with the respective controls (Figure 2). At the end of the six-day cold soak, H. uvarum made up 90% of the fungal community present in the treatments where M. fructicola Gaïa was not added. In contrast, H. uvarum made up only 7.5% and 16% of the fungal community in treatments where M. fructicola Gaïa was added at the beginning of cold soak, while Metschnikowia species represented 76% and 70%, respectively (Figure 2). Predominant fungal species in the “other” category at the end of cold soak were filamentous fungi belonging to the genera Cladosporium and the yeast-like fungus, Aureobasidium pullulans.
Relative abundance of fungal operational taxonomic units present at the beginning of a prefermentation cold soak of Pinot noir grapes (initial grape must) and after six days following inoculation with Metschnikowia fructicola Gaïa and/or Hanseniaspora uvarum OSU2a. H. uvarum OSU2a was inoculated at either a low (103 cfu/mL) or high (106 cfu/mL) population.
Adding M. fructicola Gaïa also affected acetic acid concentration in the Pinot noir must at the end of cold soak. In treatments inoculated with either a low or high population of H. uvarum OSU2a, a significantly lower concentration of acetic acid was present at the end of cold soak if M. fructicola Gaïa was also added (Figure 3A). For example, 90 mg/L acetic acid was present in the treatment with a low addition of H. uvarum OSU2a, whereas only 55 mg/L was present if M. fructicola Gaïa was added at the beginning of cold soak (Figure 3A), a 60% reduction. In treatments where a high inoculation of H. uvarum OSU2a was added at the beginning of cold soak, 198 mg/L acetic acid was present, but if M. fructicola Gaïa was added, only 159 mg/L was present (Figure 3A).
Concentration of acetic acid (A) and ethyl acetate (B) at the end of a six-day cold soak (CS) of Pinot noir grapes inoculated with a low (103 cfu/mL) or high (106 cfu/mL) population of Hanseniaspora uvarum OSU2a with and without Metschnikowia fructicola Gaïa. Different letters within a treatment indicate significant differences at p ≤ 0.05, n = 3.
Ethyl acetate concentrations in all treatments were also significantly lower at the end of cold soaking if M. fructicola Gaïa had been added (Figure 3B). For example, in the treatment with a low addition of H. uvarum, 33 mg/L of ethyl acetate was present at the end of cold soak. When M. fructicola was added to this treatment, only 6 mg/L of ethyl acetate was present, an 80% reduction. When a high population of H. uvarum OSU2a was added, ethyl acetate production was reduced from 33 mg/L to 19 mg/L, a 43% decrease if M. fructicola was added at the start of cold soak (Figure 3B).
Differences in acetic acid at the end of cold soak were no longer present at the end of alcoholic fermentation (Figure 4A), with all wines containing greater than 200 mg/L acetic acid, including wine produced without cold soaking. However, significant differences in ethyl acetate concentrations were still present in wines where M. fructicola Gaïa had been added (Figure 4B). Wines that were made with or without the addition of M. fructicola Gaïa and low populations of the inoculated H. uvarum OSU2a contained 70 mg/L and 49 mg/L ethyl acetate, respectively. In contrast, the addition of M. fructicola Gaïa did not reduce ethyl acetate concentration if a high population of H. uvarum OSU2a had been added prior to cold soak (Figure 4B).
Concentration of acetic acid (A) and ethyl acetate (B) at the end of alcoholic fermentation of Pinot noir grapes that did not undergo a cold soak or grapes that underwent a six-day cold soak (CS) after inoculation with a low (103 cfu/mL) or high (106 cfu/mL) population of Hanseniaspora uvarum OSU2a, with and without Metschnikowia fructicola Gaïa. Different letters within a treatment indicate significant differences at p ≤ 0.05, n = 3.
Discussion
A focus on reducing the use of SO2 during winemaking has led to exploring alternative methods to control spoilage microorganisms. One such method is inoculating a high population of a microorganism that acts as a protective culture by repressing the growth of other naturally present microorganisms. This biocontrol method has been successfully used in the food industry (Gaggia et al. 2011, Spadaro and Droby 2016) but has only recently been applied in the wine industry, with few peer-reviewed studies demonstrating the effectiveness of biocontrol during winemaking. The present study presents evidence that current commercially available non-Saccharomyces yeast cultures can be utilized to reduce H. uvarum growth and the production of spoilage compounds during prefermentation cold soaking. This biocontrol activity was demonstrated both in an MGJ and during prefermentation of Pinot noir grapes.
In an MGJ, we noted that all tested non-Saccharomyces cultures reduced H. uvarum acetic acid production with differences between cultures. For example, co-inoculation of H. uvarum with either L. thermotolerans culture resulted in sig nificantly lower acetic acid production under cold-soak conditions compared with other non-Saccharomyces cultures used. Reduced acetic acid production has been observed in co-culture treatments of S. cerevisiae and non-Saccharomyces yeast (Bely et al. 2008, Rantsiou et al. 2012) but has not been shown to occur during co-culture of two non-Saccharomyces yeast. These previous studies investigated co-inoculation of non-Saccharomyces yeast and S. cerevisiae during fermentation of high-sugar grape musts and attributed the lower concentration of acetic acid to reduced osmotic stress during S. cerevisiae growth due to sugar consumption by the non-Saccharomyces yeast strain (Rantsiou et al. 2012). Under high-osmotic stress, S. cerevisiae upregulates glycerol production and, concurrently, acetic acid production as a means of maintaining cellular redox potential (Yang et al. 2017). Compared to studies by Bely et al. (2008), who used a 36 Brix must, and Rantsiou et al. (2012), who used a 40 Brix must, a high Brix must was not used in this study (22 to 23 Brix). While this sugar concentration may still place S. cerevisiae under some osmotic stress, it is unlikely to result in the same stress as the conditions used by Bely et al. (2008) and Rantsiou et al. (2012) that resulted in reduced acetic acid production. Furthermore, non-Saccharomyces yeast are reported to not respond to high-osmotic stress in the same manner as S. cerevisiae (Bely et al. 2008, Rantsiou et al. 2012).
The reduction of acetic acid in co-cultures of H. uvarum OSU2a and commercial non-Saccharomyces yeast cultures was more likely due to the reduced growth of H. uvarum OSU2a during the six-day simulated cold soak. All commercial cultures, aside from M. fructicola Gaïa, significantly reduced H. uvarum growth following the six-day cold soak. Reduced growth rate and cellular biomass is a reported consequence of co-inoculations of S. cerevisiae and non-Saccharomyces yeasts and is normally accompanied by an increased persistence of the non-Saccharomyces strain (Mendoza et al. 2007, Domizio et al. 2011). A similar effect on growth occurred in the present study when two different non-Saccharomyces yeast were co-inoculated. The inhibition of non-Saccharomyces yeast by other non-Saccharomyces yeast has been previously reported by Simonin et al. (2018), who noted that a high population of T. delbrueckii added at the beginning of alcoholic fermentation of a white wine resulted in a decrease in the overall ratio of other non-Saccharomyces species such as Hanseniaspora, Metschnikowia, and Candida. The findings from the present study support the findings of Simonin et al. (2018) and provide information on the effect of individual non-Saccharomyces yeast species on H. uvarum. In addition, results from the screening experiment demonstrate that growth inhibition of H. uvarum was not specific to one yeast species because H. uvarum OSU2a growth was inhibited by Metschnikowia, Torulaspora, and Lachancea yeast cultures. Future studies should seek to confirm these findings during wine production because, aside from M. fructicola Gaïa, non-Saccharomyces yeast were evaluated only for their effect on H. uvarum in an MGJ.
To determine if different strains of H. uvarum would react in a similar manner to H. uvarum OSU2a, many different H. uvarum strains were inoculated into MGJ with or without M. fructicola Gaïa. When grown in monoculture, the H. uvarum strains produced various amounts of acetic acid, ranging from a high of 255.8 mg/L to a low of 133.7 mg/L. Why these differences in acetic acid production occurred is unclear. Differences do not appear related to the growth of the individual H. uvarum strains as there was no clear trend between the relative increase in cfu/mL and acetic acid production. Unfortunately, information is limited regarding variation in acetic acid production by H. uvarum strains, with Romano et al. (1992) providing one of the few reports on variable acetic acid production between K. apiculata strains. However, regardless of whether the strains produced higher or lower amounts of acetic acid when grown alone, co-inoculation with M. fructicola Gaïa caused decreased production of acetic acid for the majority of the H. uvarum strains tested. In addition, co-inoculation with M. fructicola Gaïa reduced the growth of all H. uvarum strains tested.
Results from the screening experiments suggest that high inoculation of a commercial non-Saccharomyces yeast can reduce H. uvarum growth and acetic acid production under cold-soak conditions. This finding was confirmed during a six-day cold soaking of Pinot noir grapes. Here, adding a high population of M. fructicola Gaïa at the beginning of cold soaking resulted in reduced growth of H. uvarum relative to the control. At the end of cold soaking, Metschnikowia was the dominate yeast species in treatments where M. fructicola Gaïa had been added, while H. uvarum dominated when no addition was made. Interestingly, if a high population of H. uvarum OSU2a was also added to the grape must, then the impact of M. fructicola Gaïa on H. uvarum growth was reduced. This finding suggests that the initial population of H. uvarum may influence the effectiveness of adding a non-Saccharomyces yeast as a biocontrol agent. Domizio et al. (2011) reported similar behavior in co-fermentations of S. cerevisiae and a single non-Saccharomyces yeast strain, where differences in the inoculation ratio of the two different yeasts affected the subsequent growth of each yeast.
The influence of the initial H. uvarum population was also apparent when considering the production of acetic acid. While a 40% reduction in acetic acid occurred in co-inoculated cold soaks if H. uvarum initial populations were ~105 cfu/mL, only a 20% reduction occurred if the initial H. uvarum population was above 106 cfu/mL. Adding M. fructicola Gaïa at the start of cold soak reduced the amount of ethyl acetate produced in a similar manner seen with acetic acid, but the effect was more pronounced. At the end of cold soaking treatments where M. fructicola Gaïa had been added, there was an 80% reduction in ethyl acetate if the initial population of H. uvarum was low and a 40% reduction if the initial H. uvarum populations were high. In the case of the high H. uvarum treatment, the reduction in ethyl acetate occurred even though M. fructicola Gaïa had little effect on H. uvarum growth during cold soaking. Benucci et al. (2018) also reported that cold soaking Sangiovese grapes with M. fructicola reduced ethyl acetate in the finished wine, but they did not report the differences at the end of cold soaking. Benucci et al. (2018) noted that wines made from grapes cold-soaked at 5°C for 24 or 72 hr had 24 and 57% less ethyl acetate, respectively, than the cold-soaked control wines. Similarly, Benito et al. (2015) reported that sequential fermentation of Riesling with M. pulcherrima Flavia and S. cerevisiae resulted in a 24% reduction in ethyl acetate compared with the S. cerevisiae control.
After alcoholic fermentation, significant differences in ethyl acetate concentrations remained between wines that had been cold soaked with or without M. fructicola Gaïa, except where a high inoculation of H. uvarum had been made at the start of cold soaking. In all treatments, ethyl acetate increased during alcoholic fermentation, which could have been from continued ethyl acetate production by H. uvarum during alcoholic fermentation but was also likely due to production by S. cerevisiae, as this yeast can produce ethyl acetate during alcoholic fermentation (Hall et al. 2017). While differences in ethyl acetate remained after alcoholic fermentation, acetic acid concentrations were no longer significantly different between treatments with or without M. fructicola Gaïa, which is in contrast to several studies that reported increased acetic acid concentrations due to cold soaking (Gordillo et al. 2010, Maturano et al. 2015). The lack of difference in acetic acid concentrations might be explained by differences in the yeast strains and cold-soak temperatures used in these studies. Acetic acid production can vary considerably between yeast strains, including strains of H. uvarum and S. cerevisiae (Romano et al. 1992). Cold-soak temperature will also affect H. uvarum growth and potentially the production of ethyl acetate and acetic acid. For example, Maturano et al. (2015) reported that volatile acidity significantly increased when cold soaking at 14°C but not at 8°C or 2.5°C. Temperature differences may also account for the relatively low concentrations of ethyl acetate and acetic acid measured in the present study where none of the wines had ethyl acetate or acetic acid concentrations above their reported spoilage sensory threshold of 150 mg/L and 0.7 g/L, respectively (Jackson 2014). Additional experiments at higher cold-soak temperatures should be conducted to determine if the trends observed during an 8°C cold soak also occur at higher temperatures where greater growth of H. uvarum would be expected.
In addition to temperature, cold-soak conditions such as SO2 and initial microbial populations will likely also influence the effectiveness of non-Saccharomyces yeast as biocontrol agents. In the present study, no SO2 additions were made, but non-Saccharomyces yeast varied in their sensitivity to this antimicrobial compound (Constantí et al. 1998, Henick-Kling et al. 1998), which is commonly added during a cold soak. It is unknown whether using SO2 in combination with a non-Saccharomyces culture will enhance or reduce the effectiveness of the non-Saccharomyces culture to reduce H. uvarum growth. The initial microbial population present at the beginning of the cold soak may also influence how effective the non-Saccharomyces yeast will be at reducing H. uvarum growth; in the present study, a higher initial starting population of H. uvarum reduced the effectiveness of M. fructicola Gaïa. We also do not know what impact high populations of yeast species other than H. uvarum would have on the ability of an added non-Saccharomyces yeast to act as a biocontrol because in the present study, the dominant species in the initial grape must was H. uvarum. Additional studies investigating a combination of factors such as temperature, SO2, and microbial load are needed to determine the effectiveness of non-Saccharomyces yeast as biocontrol agents over a range of cold-soak conditions typical of commercial winemaking.
While this study did not specifically seek to determine how non-Saccharomyces yeast repress H. uvarum growth, the results do provide some insight into a possible mechanism. Biocontrol of a target microorganism by a culture of another microorganism is typically due to either competition for resources, such as nutrients or physical space, or production of inhibitory substances. For example, some non-Saccharomyces yeast are reported to produce killer toxins that can inhibit other non-Saccharomyces yeast species, such as B. bruxellensis (Comitini and Ciani 2011, Villalba et al. 2016). In addition, some strains of M. fructicola can produce the iron-sequestering compound pulcherrimic acid, which inhibits the growth of certain yeasts and molds (Csutak et al. 2013, Oro et al. 2014, Spadaro and Droby 2016). Given that the repression of H. uvarum growth was not specific to a particular non-Saccharomyces species suggests that inhibition was not due to an inhibitory substance, as the production of such substances is typically species-specific. However, this mechanism cannot be ruled out and could be a contributing factor to the repression of H. uvarum growth.
Competition for nutrients such as nitrogen or carbon is also unlikely to be the mechanism by which non-Saccharomyces cultures repressed H. uvarum growth because sugar concentrations remained abundant at the end of cold soaking and no significant differences in YAN were observed between treatments with or without an addition of a non-Saccharomyces culture. However, given that individual nutrients were not assessed in this study (only ammonia nitrogen and free amine nitrogen), we cannot completely rule out the possibility that a non-Saccharomyces culture depleted of a key nutrient required by H. uvarum resulted in reduced growth. However, based on the results of this study, the most likely explanation for the reduction of H. uvarum growth is competition for physical space. Physical entrapment of H. uvarum by high populations of S. cerevisiae has been reported to lead to lower growth rates of H. uvarum (Arneborg et al. 2005). This is also a proposed mechanism for the early death of certain non-Saccharomyces yeast species observed in mixed fermentations with high populations of S. cerevisiae (Nissen and Arneborg 2003, Nissen et al. 2003, Renault et al. 2013, Rossouw et al. 2018). Adding a high population of a non-Saccharomyces yeast at the start of cold soaking may have repressed H. uvarum in a similar manner. This mechanism also explains why similar growth and acetic acid reduction was observed in all yeast species and strains tested in this study. Further studies are needed to elucidate how non-Saccharomyces yeast repress H. uvarum so that the use of these cultures as biocontrol agents can be optimized.
Conclusions
Because of the interest in utilizing non-Saccharomyces cultures for biocontrol during winemaking, this study compared the biocontrol capabilities of several commercially available non-Saccharomyces species under model cold-soak conditions and during the production of Pinot noir wine. All non-Saccharomyces cultures tested reduced H. uvarum growth and acetic acid production in an MGJ, demonstrating their potential use for reducing spoilage during cold soaking. This potential was confirmed during cold soaking of Pinot noir grapes where the addition of M. fructicola Gaïa repressed H. uvarum growth and reduced production of acetic acid and ethyl acetate. After alcoholic fermentation, differences in ethyl acetate remained but there were no longer differences in acetic acid concentrations. Additional studies will be required to determine the mechanism by which the non-Saccharomyces yeast used in this study repressed H. uvarum growth, but results from the present study suggest it is likely due to competition for physical space rather than competition for nutrients or production of inhibitory compounds. Future studies should also focus on the influence of cold-soak conditions such as temperature, SO2 concentration, and initial microbial populations on the effectiveness of non-Saccharomyces yeast as biocontrol agents, as this will help optimize their use during cold soaking.
Acknowledgments
The authors gratefully acknowledge the Oregon Wine Board and the Oregon Wine Research Institute for providing financial support for this study.
Footnotes
Supplemental data is freely available with the online version of this article at www.ajevonline.org.
By downloading and/or receiving this article, you agree to the Disclaimer of Warranties and Liability. The full statement of the Disclaimers is available at http://www.ajevonline.org/content/proprietary-rights-notice-ajev-online. If you do not agree to the Disclaimers, do not download and/or accept this article.
- Received January 2020.
- Revision received April 2020.
- Accepted April 2020.
- Published online October 2020
- © 2020 by the American Society for Enology and Viticulture
Literature Cited
Vol 71 Issue 4
Jump to section
Related Articles
Cited By...
More from this TOC section
Similar Articles
