Abstract
The application of high hydrostatic pressure (HHP) to inactivate microorganisms in a grape must (crushed grapes), coupled with autoclavable microscale fermentors, was investigated as a potential tool for studying the impact of yeast on red wine aroma and flavor. A Pinot noir grape must was inoculated with Saccharomyces cerevisiae, Brettanomyces bruxellensis, Kluveromyces thermotolerans, Lactobacillus hilgardii, Oenococcus oeni, and Acetobacter aceti at ~1 × 105 cfu/mL and subjected to HHP treatment for 10 min at 551 MPa (5510 bar). After HHP treatment no viable cells were detected in the grape must. Autoclavable microscale fermentors were then used to conduct replicate fermentations of HHP-treated and untreated Pinot noir grapes producing sufficient wine for chemical and sensory analysis. No differences were observed in fermentation rate between treatments, and variability between replicates was very low. No significant differences in color or hue were observed, but wine produced from HHP-treated grapes contained higher total phenolics. Sensory analysis of the wines by a trained panel revealed that other than a slight increase in overall fruit aroma there were no significant differences between any aroma and flavor descriptors. Results suggest that HHP processing of grapes, in conjunction with autoclavable microscale fermentors, could be used to conduct experimental red wine fermentations without the influence of native yeast and bacteria present on grapes and fermentation equipment.
Studies investigating the effect of specific yeast species or strains on wine aroma and flavor have generally focused on white wines such as Riesling (Egli et al. 1998, Henick-Kling et al. 1998), Chardonnay (Egli et al. 1998), and Sauvignon blanc (King et al. 2008, Swiegers et al. 2009), with few performed on red wine. This emphasis on white wines is partially due to the importance of aroma to white wine quality, but aroma can also be an important quality parameter for red wines such as Pinot noir (Fang and Qian 2006). However, one of the challenges when investigating the effect of yeast on red wine aroma is the presence of naturally occurring yeast and bacteria on the grapes prior to fermentation. Previous studies have attempted to minimize the impact of these background microorganisms through various means including high inoculation rates of commercial yeast strains combined with thermal processing (Clemente-Jimenez et al. 2004), rehydrated juice (Pickering et al. 2008), and large amounts of SO2 (Reynolds et al. 2007), and other chemicals such as dimethyl dicarbonate (Mateos et al. 2006, Gonzalez et al. 2007). Unfortunately, each method has drawbacks in terms of wine aroma and flavor research. For example, rehydrated juice does not accurately mimic fermentation of fresh grapes and thermal processing may result in the modification or destruction of volatile compounds in wine (Girard et al. 1997). While SO2 will reduce populations of naturally occurring microorganisms, it will not completely inactivate them and high residual SO2 may present problems during alcoholic fermentation. Dimethyl dicarbonate has been shown to be effective against yeast, but it is less effective against bacteria (Delfini et al. 2002, Costa et al. 2008) and is hazardous to work with. Finally, inoculation with a commercial strain of Saccharomyces cerevisiae does not guarantee a monoculture of the inoculated yeast strain (Howell et al. 2004).
High hydrostatic pressure (HHP) processing is an alternative method for inactivating microorganisms present in a grape must while minimizing sensory changes and has been investigated in food and beverage systems (Cheftel 1995, Balasubramaniam and Farkas 2008, Lamela and Torres 2008). HHP inactivates microorganisms through disruption of higher structures of cellular proteins and enzymes and by causing damage to cell wall structure, resulting in cell leakage (Smelt 1998). In one of the few studies on HHP treatment of grapes, HHP microbiologically stabilized a Barbera grape juice without impacting sensory properties (Delfini and Conterno 1995). However, wine was not produced from this grape juice and so no conclusions could be drawn about the impact on the resulting wine.
Small-scale fermentation is often used when studying wine yeast as it allows multiple strains to be evaluated in a replicated system. However, for red winemaking, microscale fermentations can be difficult to conduct due to the need to maintain skin/juice contact while fermenting. In addition, if microbiological studies are performed, then fermentations should be conducted under sterile conditions with aseptic sampling to avoid contamination. One study used a relatively simple microscale vessel that allowed fermentation of 3.5 kg fruit and a submerged cap method to maintain skin/juice contact (Sampaio et al. 2007). Although this fermentor allowed investigation of phenolic extraction during fermentation, it was not sterile and could not be sampled aseptically, making it less useful for microbiological studies. On a smaller scale, one study investigated deficit irrigation effects on Shiraz using 800 g fermentations (Chalmers et al. 2010), while another study reported using 450 mL red grape must in sterile 500 mL flasks, but no further description of the fermentation vessels, fermentation management, or sampling protocol was given (Gonzalez et al. 2007).
The purpose of this study was to investigate the effectiveness of HHP to inactivate microorganisms present in a grape must and to assess whether HHP treatment of grapes impacted wine sensory properties. Autoclavable microscale fermentors that allowed aseptic sampling were developed and used to produce the wine. If effective, HHP-treated grapes could be used to study the effects of yeast species and strains on red wine aroma and flavor without interference from yeast and bacteria naturally present on grapes and fermentation equipment.
Materials and Methods
Microorganisms.
An active-dry form of Saccharomyces cerevisiae MERIT.ferm (Chr. Hansen, Hørsholm, Denmark) was obtained. Kluyveromyces thermotolerans was obtained from the Symphony yeast starter culture (Chr. Hansen; a blend of S. cerevisiae MERIT.ferm and K. thermotolerans) by streaking for single colonies on WL agar (Difco, Franklin Lakes, NJ) and distinguishing from S. cerevisiae based on colony morphology and color (Pallmann et al. 2001). Acetobacter aceti, Lactobacillus hilgardii, and Brettanomyces bruxellensis were obtained from C.G. Edwards (Washington State University, Pullman, WA). Oenococcus oeni VFO (Chr. Hansen) was obtained in freeze-dried form. All yeast were streaked for single colonies and then maintained on potato dextrose agar (Difco) slants while bacteria were maintained in de Man, Rogosa, and Sharpe (MRS) stabs or slants (pH 4.5) and stored at 4°C. When needed, yeast were transferred from slants to YPD broth (Difco) and grown aerobically at 30°C for 48 hr. Bacterial cultures were transferred from stabs to MRS broth (pH 4.5) and grown aerobically at 30°C for 5 days. Cells were harvested by centrifugation (4,000 × g for 20 min) and resuspended in 0.2 M phosphate buffer (pH 7.0) before inoculation.
Grapes.
Grapes were Vitis vinifera L. cv. Pinot noir harvested at the Oregon State University (OSU) Woodhall Vineyard (Alpine, OR) on 15 Oct 2008. After harvest, grapes were stored at 4°C overnight before being hand-sorted and destemmed. Grapes were pooled and divided into 3 kg aliquots. To assess the ability of HHP to inactivate wine microorganisms from a grape must, four aliquots were inoculated with S. cerevisiae, B. bruxellensis, K. thermotolerans, Lb. hilgardii, O. oeni, and A. aceti at ~1 × 105 cfu/mL. After mixing, samples were taken and used to assess initial yeast and bacterial populations by plating on YPD for total yeast counts, lysine (Difco) for non-Saccharomyces counts, and MRSC (MRS + 100 mg/L cycloheximide or actidion) for bacterial counts. To the remaining 3 kg aliquots, 30 mg/L SO2 was added. All aliquots were then placed in Food Saver bags (Jarden Corp., Boca Raton, FL) and vacuum-sealed. Three bags were put aside and stored at 4°C (non-HHP treatment) and all remaining bags were treated by HHP.
High hydrostatic pressure processing.
Aliquots of destemmed grapes were processed 10 min at 551 MPa (5510 bar). The HHP unit was custom made (National Forge Company, Irvine, PA) with a 22 L maximum capacity and a 689 MPa (6890 bar) maximum pressure. The high-pressure intensifier pump had a maximum capacity of 620 MPa (6200 bar) (model 7XS-6000; Flow International Corporation, Kent, WA).
Enumeration.
After HHP treatment, samples were taken from the four bags which had been inoculated with wine microorganisms and viable microbial populations were determined by plating on YPD, lysine, and MRSC agar after appropriate dilutions (0.1% peptone). Plates were incubated aerobically at 25°C for 48 hr (yeast) or 7 days (bacteria) prior to counting.
Microscale fermentors.
Microscale fermentors were based on Sampaio et al. (2007) (Figure 1). Fermentation vessels (Q Glass, Towaco, NJ) were fabricated from autoclave safe glass with a 4 L capacity and a flanged top with an o-ring channel. The headpiece was a custom cut polycarbonate disk (Ridout Plastics, San Diego, CA), and o-ring gaskets were high-temperature, food-grade silicone (Applied Industrial Technologies, Cleveland, OH). The headpiece was secured to the flanged top of the fermentation vessel with spring clamps (Irwin Tools, Wilmington, OH). All bungs and airlocks used were high-temperature, food-grade silicone (Vin Table, Ambler, PA), and sample tubes were constructed from food-grade stainless-steel tubing. Fermentors were equipped with a perforated polycarbonate punch-down disc (Ridout Plastics) attached by a food-grade stainless-steel punch-down handle. Aseptic sampling during the fermentation was accomplished using a nitrogen siphon system. Microscale fermentors were autoclaved at 121°C for 20 min and cooled before use.
Fermentations.
Three kg aliquots of destemmed HHP-processed or unprocessed grapes were transferred to autoclaved microscale fermentors in a laminar flow hood and inoculated with S. cerevisiae MERIT.ferm at ~1 × 106 cfu/mL. An uninoculated control was also prepared. Fermentations were conducted in triplicate in a temperature-controlled room at 27°C. During fermentation, the punch-down disc was pushed halfway down so that the grapes were fully submerged. Samples were taken daily after mixing, which was performed by moving the plunger up and down five times at a set distance. 70% ethanol was sprayed on the exposed plunger handle before and after mixing to prevent contamination. Brix was monitored using an Anton-Paar DMA 35N Density Meter (Graz, Austria) and viable cell populations were monitored by plating on YPD, MRSC, and lysine agar after appropriate dilutions. After all fermentations were complete (<0.5 g/L reducing sugar as measured by CliniTest), wines were pressed using a small, modified basket press that applied a constant pressure of 1 bar for 5 min, allowing consistent pressing. No malolactic fermentation was carried out. Samples were taken from each replicate and frozen at −20°C until required for analysis. Preliminary sensory evaluations revealed no apparent sensory differences between replicates and these were combined for sensory analysis. Wines had 30 mg/L SO2 added before being settled at 4°C for 5 days, racked, sterile-filtrated (0.45 μm cartridge filter; Pall Corp., East Hills, NY), bottled (375 mL bottles with crown-cap closure), and stored at 13°C.
Color and total phenolics.
Color was determined by spectrophotometric analysis (Genesys 10 UV Spectrophotometer; Thermo Scientific, Madison, WI) at 420 and 520 nm in 1 mm path-length cuvettes after pH adjustment to 3.60. Total phenolics were determined using the Folin–Ciocalteu assay (Boulton et al. 1996).
Wine sensory analysis.
Sensory panelists and training.
Twelve panelists were recruited from the OSU Food Science and Technology Department and from the community (Corvallis, OR). Panelists included five men and seven women and all were experienced with wine sensory evaluation or trained sensory analysis. Six training sessions were conducted. Panelists tasted several Pinot noir wines that were of similar age as the experimental wines as well as the actual experimental wines and created a list of descriptors for wine aroma and flavor. During training the list was amended until all panelists agreed on a final list of descriptors and their definitions. References were created based on previous publications (Noble et al. 1987, Guinard and Cliff 1987) to help panelists understand and define descriptors. Nine aroma descriptors (overall intensity, overall floral, overall fruity, citrus, red fruit, dark fruit, jammy/cooked, spicy, leather) and five retronasal aroma/tactile (flavor) descriptors (overall intensity, fruit, spicy, sour, astringency) were chosen. A 16-point intensity scale was used and intensity standards were created to aid panelists in rating intensities. The standards used and their agreed upon intensities were fresh safflower oil (3 = slight), fresh orange juice (9 = moderate to large), and cinnamon gum (15 = extreme). At the beginning of each training and tasting session, panelists were encouraged to refamiliarize themselves with reference and aroma intensity standards. At the conclusion of training, two panelists were removed from the panel as examination of the data showed them to be inconsistent with the rest of the panel in the use of some descriptor terms.
Sensory evaluation.
Wines were assessed by a sensory panel 6 months postbottling and kept at 13°C until the day before testing. All wines were allowed to equilibrate to room temperature and poured 30 min before evaluation. 25 mL samples were served in 240 mL INOVA tulip glasses (St. George Crystal, Jeannette, PA) and covered with plastic lids. Glasses were coded and evaluated in a completely randomized order with panelists tasting each wine treatment twice. Panelists rated the samples based on the list of descriptors and rated each descriptor using a 16-point intensity scale. Because of the potential astringency of the samples, panelists were instructed to rinse with a 0.1 g/L pectin solution between samples. Panelists were allowed 3 min to evaluate each wine and were given a mandatory 1 min break between samples.
Statistical analyses.
Statistical analysis of the wine chemical data was performed using SAS version 9.1 (Cary, NC) with Tukey’s HSD test for mean separation. The coefficient of variation (CV) was calculated as the relative percentage of the standard deviation (Excel 2003, Microsoft, Redmond, WA). Sensory data was analyzed using univariate analysis of variance (ANOVA) performed by general linear model (GLM) in SAS ver. 9.1. The ANOVA model comprised two main effects (panelist (PAN) and wine (WINE)), a nested effect (replication, which was nested in PAN or REP(PAN)), and a two-way interaction effect between PAN and WINE (WINE*PAN). The PAN, REP(PAN), and (WINE*PAN) were treated as random effects and WINE was treated as a fixed effect. Significant differences detected by ANOVA were subjected to post-hoc Tukey HSD multiple comparison to test least squares means of WINE (means) at the p < 0.05 significance level.
Results and Discussion
Before HHP treatment, high populations of microorganisms were present in the grape must (Table 1), with counts on YPD, lysine, and MRSC agar greater than 1 × 105 cfu/mL. Mold growth was also observed on YPD plates (data not shown). However, after HHP treatment no growth was recorded on any of the plates (Table 1), demonstrating the ability of HHP to inactivate high concentrations of microorganisms present in a grape must. These results are in agreement with studies of HHP in different food systems (Oey et al. 2008) and in grape juice (Daoudi et al. 2002) and wine (Delfini and Conterno 1995, Mok et al. 2006). Furthermore, although HHP treatment was sufficient to inactivate microorganisms in the grape must, it did not cause any significant change in the basic juice parameters Brix, pH, and titratable acidity (Table 2).
Pinot noir grapes that had or had not been treated by HHP were fermented in autoclavable microscale fermentors to determine the sensory properties of the resulting wines. Saccharomyces cerevisiae MERIT.ferm grew well in both HHP- and non-HHP-treated musts, reaching maximum populations of 1.0 × 109 cfu/mL (Figure 2). All fermentations were completed after 8 days. In fermentations of HHP-treated grapes, no non-Saccharomyces yeast or bacteria were detected. In contrast, in the non-HHP-treated fermentations, non-Saccharomyces populations were initially present at ~4 × 105 cfu/mL and rose close to 1 × 108 cfu/mL after 48 hr. At this point populations rapidly decreased below 1 × 102 cfu/mL. At no time during the course of the fermentation were any viable microorganisms detected in the uninoculated control (data not shown), demonstrating the effectiveness of the HHP treatment. Furthermore, Brix did not decrease in the uninoculated control during the course of the fermentation (Figure 2).
No significant differences for color or color hue were observed between wines produced from grapes treated or not treated with HHP (Table 3). However, wine produced from HHP-treated grapes contained significantly higher total phenolics than wine made from non-HHP-treated grapes. This increase in total phenolic content may be due to cell wall breakdown caused by the HHP treatment, leading to greater extraction of phenolic compounds during fermentation. Cell wall breakdown due to HHP has previously been observed in spinach and cauliflower (Prestamo and Arroyo 1998). In an effort to reduce the excessive extraction of phenolics during fermentation, wines were made in a subsequent year (2009) from Pinot noir grapes treated by HHP for 5 min rather than 10 min. In addition, wines were produced where the punch-down device in the fermentation vessel was fully raised to allow a cap of grape skins to form, and the cap was mixed into the fermenting must twice daily (simulating a punch-down). Wine was also made with a submerged cap as previously described (Table 4). Although still higher, total phenolics in wines made with HHP-treated grapes (5 min) were more similar with wine made from non-HHP-treated grapes than wine made with grapes treated by HHP for 10 min. The difference in total phenolics between wines made with a submerged cap versus a punch down was not statistically significant. These results indicate that the length of time grapes are processed by HHP influences the extraction of phenolic compounds during fermentation and should be taken into account if HHP-treated grapes are used to study wine phenolic compounds.
Wines produced from HHP-treated or non-HHP-treated grapes were analyzed by the trained sensory panel to determine what, if any, sensory differences existed between the wines. Confirming other report on HHP in food and beverage systems (Cheftel 1995, Delfini and Conterno 1995, Balasubramaniam and Farkus 2008), there were few sensory differences between the wines. The only significant difference was a slight increase in overall fruity aroma for wine made from HHP-treated grapes (Table 5). However, none of the secondary fruit aroma descriptors (citrus, red fruit, dark fruit, jammy/cooked fruit) were significantly different. Surprisingly, no significant difference was found between the astringency of the two wines despite the higher total phenolic content of wine made with HHP-treated grapes, possibly because all wines evaluated were relatively young and were all rated highly for astringency. In addition, the presence of significant populations of non-Saccharomyces yeast early during the fermentation of non-HHP-treated grapes also did not result in sensory differences between the wines. Any changes in sensory characteristics due to non-Saccharomyces yeast growth may have been overpowered by the inoculated Saccharomyces yeast or may have been too subtle for the sensory panel to discern. Although non-Saccharomyces yeast have been reported to impact the overall sensory characteristics of a wine (Lema et al. 1996, Henick-Kling et al. 1998, Soden et al. 2000), the majorirty of these studies have been performed in white wines where aroma differences are likely to be more noticable.
Fermentations using the autoclavable microscale fermentors allowed aseptic sampling and produced wine in sufficient quantity for chemical and sensory analysis. Variability across microscale fermentor replicates was very low. The coefficient of variation (CV) was less than 0.5% for Brix during fermentation other than one sampling point (non-HHP-treated grapes at 96 hr). In addition, color and total phenolic analysis of the wines resulted in a CV of less than 0.5%, suggesting that variability introduced due to mixing during fermentation and pressing was minimal. These results further support other findings regarding the potential usefulness of microscale fermentors in wine research (Sampaio et al. 2007). However, because the microscale fermentors in the present study can be autoclaved and aseptically sampled, they could be used for microbiological studies during red winemaking without the risk of microbial contamination.
Conclusions
The presence of background microorganisms on grapes and fermentors can make it difficult to determine the effects of specific yeast strains or species on red wine parameters. In this study, HHP treatment was effective in inactivating microorganisms present in a Pinot noir grape must, allowing fermentation to be conducted by a specific inoculated yeast strain. Sufficient wine for sensory and chemical analysis was produced using autoclavable microscale fermentors with low variation between replicates. Wine made from HHP-treated grapes contained higher total phenolics, suggesting that a more detailed understanding of how HHP treatment impacts phenolic wine composition would be required before using HHP for wine phenolic research. Minimal sensory differences were noted between the wines, demonstrating that HHP-processed grapes could be used to produce representative red wines for sensory evaluation. In conjunction with autoclavable microscale fermentors, HHP processing could be used to allow experimental red wine fermentations without the influence of naturally present yeast and bacteria.
Acknowledgments
Acknowledgments: The Oregon Wine Board provided funding for this project.
The authors thank Cindy Lederer for her assistance with sensory and statistical analysis.
- Received March 1, 2011.
- Revision received June 1, 2011.
- Accepted July 1, 2011.
- Published online December 1969
- © 2011 by the American Society for Enology and Viticulture