Understanding and Increasing Alcohol Tolerance in Wine Yeasts

The goal of our project is to measure the cell membrane composition of a wide range of wine yeasts and correlate this composition with strain characteristics such as growth and ethanol tolerance, thus developing a rational approach to strain modification. In this way, new strains can be constructed that are highly ethanol tolerant or ethanol tolerance can be introduced into existing strains with favorable flavor characteristics. To achieve these goals, we first identified a group of yeast strains representing a wide range of ethanol tolerances. We then did an in-depth examination of the growth characteristics and ethanol tolerance of these strains, and confirmed that we do, in fact, have a wide range of characteristics. We have completed development of a new HPLC-mass spectrometer (LC-MS) assay for determining cell membrane lipid composition. The main classes of lipids in yeast membranes are well separated with this method, and we are able to quantify over 80 species of fatty acids making this significantly higher resolution than the traditional methods for lipid analysis that have been used previously in studies on Saccharomyces. We are nearly finished with the associated assay for sterols. We used this assay, in its current state, to examine the membrane composition of six of the strains with fairly different ethanol tolerances from a point during small scale fermentations. These data were used, along with multivariate statistical approaches (e.g. Principal Component Analysis), to begin to examine how membrane composition varies with fermentation properties and ethanol tolerance levels. Studies to examine yield of cells from nitrogen were also initiated.

Adaptive Evolution of Commercial Wine Strains for Reduced Ethanol

The goal of this research project was to explore the use of the impact of chronic exposure to a moderately inhibitory concentration of furfural under fermentative conditions on cellular metabolism. Furfural inhibits yeast metabolism and therefore growth by competing with acetaldehyde for the reduced cofactor, NADH, generated during sugar catabolism. The reaction of acetaldehyde with NADH produces ethanol. Similarly the reduction of furfural produces a much less toxic alcohol. The yeast cell is therefore faced with a challenging metabolic problem: in order to detoxify furfural in the environment the cells must make less acetaldehyde and therefore less ethanol. Adaptive evolution is a process that exposes yeast to chronic inhibitory concentrations for multiple generations thereby imposing selective pressure to mutate to become more resistant to the inhibitory conditions. The first aim of this project was to screen a collection of commercial wine strains for tolerance to furfural and 5-hydroxymethylfurfural. We expected higher tolerance given that furfural is found during barrel fermentation and aging and resistance to this compound would be expected across wine strains. All 27 of the wine strains tested displayed some level of resistance to furfural in contrast to the 3 laboratory strains that displayed sensitivity. Upon more detailed analysis of sensitivity, the wine strains were grouped into six clusters depending upon their patterns of resistance. Ethanol yields were evaluated for strains in each of the clusters and one cluster appeared to contain strains that were low ethanol yielders, validating this approach as a mechanism to generate strains with reduced ethanol yields. Adaptive evolution experiments are underway to hopefully lead to even further reductions in ethanol production.

Understanding and Increasing Alcohol Tolerance in Wine Yeasts

The goal of our project is to measure the cell membrane composition of a wide range of wine yeasts and correlate this composition with strain characteristics such as growth and ethanol tolerance, thus developing a rational approach to strain modification. In this way, new strains can be constructed that are highly ethanol tolerant or ethanol tolerance can be introduced into existing strains with favorable flavor characteristics. To achieve these goals, we first identified a group of yeast strains representing a wide range of ethanol tolerances. We then took a 12-member subset of this group and did a preliminary examination of the growth characteristics and ethanol tolerance of these strains, confirming that we do, in fact, have a wide range of characteristics. We have nearly completed development of a new HPLC/mass spectrometer (LC/MS) assay for determining cell membrane composition, and are in the process of moving this assay over to a more sensitive instrument for quantification. The main classes of lipids in yeast membranes are well separated with this method. We used this assay, in its current state, to examine the membrane composition of two strains with fairly different ethanol tolerances at several points throughout small scale fermentations. We demonstrated that membrane composition does, in fact, change over the course of fermentation, and is also quite different for the different strains. In addition, our initial experiments have shown that ethanol tolerance may be a dual function of both membrane composition and nitrogen utilization efficiency.

Impact of malolactic fermentation on red wine color

The color of a red wine is an important sensory attribute that originates primarily from anthocyanins. However, development of stable red wine color is impacted by compounds such as p-coumaric acid, caffeic acid, catechin, and quercetin that are involved in copigmentation reactions as well as acetaldehyde and pyruvic acid. While it is known that yeast can alter the concentrations of some of these compounds, little is known regarding the impact malolactic bacteria may have on red wine color development. This project is investigating the effect of the malolactic fermentation (MLF) on red wine color and the ability of malolactic bacteria to degrade compounds important to the development of stable red wine color.

Pinot noir and Merlot wines were produced using grapes from the Oregon State University vineyard and were fermented with Saccharomyces cerevisiae VQ15. Concurrently, a third of Pinot noir musts were inoculated with Oenococcus oeni. At dryness, wines were pressed and filtered (0.45 µm nominal) with a second third of the wines being inoculated with O. oeni VFO (remaining third was was not inoculated). Some of the wine that had not undergone MLF was pH adjusted to the same final pH of wines that had completed MLF. Samples were taken before and after MLF for analysis and wines were sterile filtered, bottled, and stored at 55° F.

Prior to MLF, all Pinot noir wines had very similar concentrations of acetaldehyde and pyruvic acid. This was also the case with the Merlot wines. However, all wines that had undergone MLF, including the simultaneously fermented Pinot noir wine, had lower acetaldehyde and pyruvic acid concentrations. Pinot noir wines that had undergone MLF also had lower wine color and polymeric pigment values compared to wines that had not gone through MLF. For the Merlot wines, wines that had undergone MLF also had lower wine color, copigmentation, anthocyanins, and polymeric pigment than wines that had not undergone MLF with the differences being more pronounced then what was observed for the Pinot noir wines. The phenolic composition of wines that underwent MLF was different from wines that had not. Both Pinot noir and Merlot wines that had undergone MLF had lower levels of caftaric acid and higher levels of caffeic acid than wines that had not undergone MLF. Wines that did not undergo MLF also had lower malvidin glucoside and monomeric anthocyanin concentrations than wines that had undergone MLF. Finally, the concentration of tannin in Pinot noir wines that had undergone MLF was lower than in wines that had not. These results demonstrate that MLF as well as time of bacterial inoculation can effect the concentration of phenolic and non-phenolic compounds involved in red wine color development.

Interactions Between Nitrogen and Vitamins on Fermentation Rate and H2S Production by Saccharomyces

Sluggish fermentation and hydrogen sulfide production are currently serious problems facing the wine industry. Besides nitrogen deficiency, a lack of certain vitamins such as thiamine and pyridoxine, can also impact H2S formation, coenzymes involved in yeast metabolism, play an important role in sulfur production. A comprehensive and systematic research approach was conducted to determine how nitrogen and vitamins (thiamine and pyridoxine) influences yeast growth rate, fermentation rate, and hydrogen sulfide production. Synthetic grape juice based on the amino acid composition of Cabernet Sauvignon grape must was used for fermentation was inoculated with Saccharomyces cerevisiae strain Montrachet (UCD 522). Fermentation rate (decrease in soluble solids), yeast viability, and H2S production were all affected by the availability of nitrogen and these vitamins.

Kinetics of Flavor Formation During Grape Juice Fermentations

Using solid phase microextraction coupled with gas chromatotgraphic analysis, we are able to “continuously” monitor ester production throughout grape juice fermentations. In previous studies we used this technique to monitor differences in production of acetate and fatty acid ethyl esters that could be related to the progression of the fermentation. In addition a multi-peak pattern of ester production was observed which had not previously been reported. During the past year (2001-2002) our studies showed that:

  • Ethanol concentrations did not have a significant effect on measured concentrations of most esters studied using the SPME technique. However, at ethanol levels greater than 5{aed9a53339cdfc54d53cc0c4af03c96668ab007d9c364a7466e3349a91bf0a23}, measured concentrations of ethyl decanoate were significantly decreased. This may indicate that SPME analysis underestimates concentrations of this ester as fermentations proceed and ethanol concentrations increase.
  • Carbon dioxide flow rates at levels approximating those occurring at the height of fermentation had only a minimal effect on measured ester concentrations. These results suggest that SPME sampling provides an accurate picture of total ester concentrations throughout fermentation, even when volatilization rates are expected to be high (i.e,. Logarithmic yeast growth).
  • Yeast inoculation level did not significantly impact the concentrations or production profiles of the ethyl esters and acetate esters studied, except for ethyl acetate. The reason why ethyl acetate production responds differentially to yeast inoculum levels is unknown.

PDF: Kinetics of Flavor Formation During Grape Juice Fermentations

Analysis of Sacchoromyces During Normal and Problem Fermentations

The aims of the first three years of this proposal were to acquire, develop and optimize technologies for the analysis of problem fermentations. The goal of this work is to develop better fermentation management strategies to reduce and hopefully eliminate the incidence of slow and incomplete fermentations. In this first phase of the research we have successfully adapted functional genomic analysis to Saccharomyces grown under enological conditions. We have identified several key differences in the physiology of yeast grown under nutrient sufficient versus nitrogen-limited conditions. We have begun identifying molecular markers associated with healthy or robust fermentations and those associated with nutritional or environmental stress. The project is well poised to complete this analysis in the next year and to identify key yeast strain and physiological input factors needed for full optimization of the predictive potential of neural networks. We have developed bacterial-specific primers for direct analysis of bacterial strains in wine. In addition, we have developed and tested several yeast specific primers and employed them on samples obtained from commercial wine fermentations. This approach has resulted in direct identification of viable but non-culturable yeast populations, a potential factor in stuck fermentations. The project is well poised to complete this analysis in the next year and to identify key yeast strain and physiological input factors needed for prediction of fermentation kinetics. In addition to the molecular and physiological work, we are currently completing detailed analysis of samples from over 200 commercial Chardonnay fermentations from the 2001 harvest. Analysis of the juice and wine from these fermentations, which ranged from normal to sluggish and stuck, will allow us to identify juice characteristics and processing choices that are critical in determining fermentation kinetics. We have also developed bacterial-specific primers for direct analysis of bacterial strains in wine. In addition, we have developed and tested several yeast specific primers and employed them on samples obtained from commercial wine fermentations. This approach has resulted in direct identification of viable but non-culturable yeast populations, a potential factor in stuck fermentations. With all yeast physiological and microbial ecology factors, juice characteristics, and processing parameters identified that are critical in determining wine fermentation kinetics, we will be able to predict problem fermentations and their resolution early in the fermentation process.

PDF: Analysis of Sacchoromyces During Normal and Problem Fermentations

Analysis of Sacchoromyces During Normal and Problem Fermentations

The aims of the first three years of this proposal were to acquire, develop and optimize technologies for the analysis of problem fermentations. The goal of this work is to develop better fermentation management strategies to reduce and hopefully eliminate the incidence of slow and incomplete fermentations. In this first phase of the research we have successfully adapted functional genomic analysis to Saccharomyces grown under enological conditions. We have identified several key differences in the physiology of yeast grown under nutrient sufficient versus nitrogen-limited conditions. We have begun identifying molecular markers associated with healthy or robust fermentations and those associated with nutritional or environmental stress. The project is well poised to complete this analysis in the next two years and to identify key yeast strain and physiological input factors needed for full optimization of the predictive potential of neural networks. In addition, we have demonstrated that artificial neural networks can be used to predict wine fermentation kinetics when all critical juice characteristics and processing are known. A means of using simple optical density measurements one to two days into a fermentation in order to predict problems has been identified. We have also adapted TGGE and DGGE technologies for the analysis of the microbial complement of wine samples. This capability now allows us to detect the presence of all common wine microbes in a juice, must or wine sample without the need for cultivation. This permits a more statistically robust sampling of a fermentation and will provide data of sufficient quality to be useful in the development of neural networks for the prediction of fermentation behavior.

PDF: Analysis of Sacchoromyces During Normal and Problem Fermentations

Identification of Yeast Strain Genetic Factors in the Formation of Volatile Sulfur Compounds

In this current grant year, the analysis of the impact of over-expression of two genes involved in consumption of reduced sulfur, CYS4 and MET17, on H2S formation in commercial and natural wine strains of Saccharomyces was completed. Interestingly, increasing the level of expression of the CYS4 gene completely eliminated hydrogen sulfide production in four strains, had no effect in others, and in a few resulted in an increase in H2S. Similar results were obtained for MET17. So far, strains that showed reduced volatile sulfur formation with CYS4 did not show any effect with MET17 and those showing an effect with MET17 showed no or increased H2S formation with over-expression of CYS4. Strains that were high produces of H2S tended to decrease sulfide release when CYS4 was present, while the moderate producers showed a stronger response with MET17. Thus, there are multiple underlying genetic causes for the production of hydrogen sulfide. This analysis does indicate that once the cause of H2S release is known for a given strain, it can be corrected genetically. It will also be possible to screen for strains naturally possessing alleles leading to reduced sulfide production to be used in conventional breeding programs.

This research has further clarified the basis for the two phases of hydrogen sulfide release observed during fermentation. The early phase of hydrogen sulfide production occurs shortly after maximal cell biomass is attained, within the first few days of active fermentation, and is related to the relative activities of the enzymes generating and consuming reduced sulfur. The later stage, which occurs at the end of fermentation, is related to the nitrogen recycling behavior of the culture. Genomic data indicates that at this point in time numerous pathways have been induced that shunt nitrogen between amino acid components. When this occurs, there is a net shift of nitrogen from the sulfur containing amino acids to the non-sulfur containing amino acids. If nitrogen levels are in ample supply, this is prevented from occurring. Interestingly, analysis of the pattern of production of hydrogen sulfide of the 12 strains used in this study revealed that many of the strains produce hydrogen sulfide continuously during fermentation. Over-expression of MET17 and CYS4 has the highest impact on the continual producers versus the transient producers.

PDF: Identification of Yeast Strain Genetic Factors in the Formation of Volatile Sulfur Compounds

Analysis of Sacchoromyces During Normal and Problem Fermentations

The aims of the first three years of this proposal were to acquire, develop and optimize technologies for the analysis of problem fermentations. The goal of this work is to develop better fermentation management strategies to reduce and hopefully eliminate the incidence of slow and incomplete fermentations. In this first phase of the research we have successfully adapted functional genomic analysis to Sacchoromyces grown under enological conditions. We have identified several key differences in the physiology of yeast grown under nutrient sufficient versus nitrogen-limited conditions. We have begun identifying molecular markers associated with healthy or robust fermentations and those associated with nutritional or environmental stress. The project is well poised to complete this analysis in the next two years and to identify key yeast strain and physiological input factors needed for full optimization of the predictive potential of neural networks. In addition, we have demonstrated that artificial neural networks can be used to predict wine fermentation kinetics when all critical juice characteristics and processing are known. A means of using simple optical density measurements one to two days into a fermentation in order to predict problems has been identified. We have also adapted TGGE and DGGE technologies for the analysis of the microbial complement of wine samples. This capability now allows us to detect the presence of all common wine microbes in a juice, must or wine sample without the need for cultivation. This permits a more statistically robust sampling of a fermentation and will provide data of sufficient quality to be useful in the development of neural networks for the prediction of fermentation behavior.