Abstract
Saccharomyces cerevisiae can make use of a nonproteogenic amino acid, γ-aminobutyric acid (GABA), one of the various nitrogen sources present in grape juice. The effect of this amino acid on yeast fermentation kinetics and by-product formation in winemaking was investigated. The GABA content of musts ranged from 2 to 580 mg/L as a function of variety, year, and geographic origin of the grapes. The γ-amino acid content may account for up to 20% of the assimilable nitrogen in grape juice. The commercial wine yeast efficiently metabolized exogenous GABA during wine fermentation. The assimilation of this γ-amino acid increased yeast growth, fermentation rate, and glycerol production, but only when nitrogen was limiting. Results demonstrate that GABA can act as a source of succinate in wine, regardless of the total nitrogen content of the must. During fermentation, succinate is primarily produced from sugars via the reductive and oxidative branches of the tricarboxylic acid cycle. The yield of succinate from GABA ranged from 0.75 to 1 mol succinate/mol GABA as a function of yeast genetic background. Up to 50% of succinate in wine may be derived from GABA, depending on the initial concentration in grape juice. These results provide new insight into GABA metabolism and offer an opportunity to improve the control of wine acidity.
A wide variety of nitrogen sources are present in grape juice, including ammonium ions and free amino acids. Their concentration depends on grape variety, time of harvest, climate, and cultivation method (Jiranek et al. 1995b). The total nitrogen content of grape juice ranges from 60 to 2400 mg/L (Ough and Amerine 1988, Henschke and Jiranek 1993), but only certain nitrogen sources can be metabolized by Saccharomyces cerevisiae. The concentration of assimilable nitrogen affects the efficiency of wine fermentation. The duration of fermentation is related to the availability of assimilable nitrogen (Henschke and Jiranek 1993). Insufficient levels of nitrogen in the must limit both biomass yield and sugar catabolism kinetics and may cause sluggish fermentation (Blateyron and Sablayrolles 2001).
Ammonium ions and primary amino acids constitute the major source of assimilable nitrogen in the must (Ough and Amerine 1988, Bisson 1991). These compounds vary in their ability to support growth because their metabolism produces more or less readily ammonium ions, glutamate, and glutamine, all of which are key components of yeast nitrogen metabolism (Cooper 1982, Jiranek et al. 1995a, 1995b). Nitrogen sources favoring high growth rates are preferentially assimilated and are referred to as good or preferred nitrogen sources. Assimilable nitrogen compounds have been classified as a function of their rate and extent of removal from the medium (Jiranek et al. 1995b). However, these parameters may depend on yeast strain and the relative abundance of the different nitrogen sources in the grape juice (Monteiro and Bisson 1991a, Manginot et al. 1998). Ammonium ions, glutamine, and asparagine are rapidly assimilated before other sources such as arginine, alanine, glutamate, and aspartate. The poorer nitrogen sources (glycine, tryptophan, and tyrosine) are assimilated after depletion of the preferred nitrogen sources and assimilation is rarely complete. Finally, proline is only assimilated to a limited extent because the restricted supply of oxygen limits the activity of proline oxidase, which is involved in proline assimilation (Wang and Brandriss 1987).
The sequential utilization of nitrogen sources by S. cerevisiae is mediated through a regulatory mechanism known as nitrogen catabolite repression (NCR). The presence of preferred nitrogen sources in the medium induces the expression of genes encoding permeases involved in the uptake and utilization of these preferred sources but decreases the level of enzymes required for the transport and metabolism of poor nitrogen sources (ter Schure et al. 1995, Magasanik and Kaiser 2002). Saccharomyces cerevisiae is also able to use nonprotein amino acids, such as ornithine and γ-aminobutyric acid (GABA), as nitrogen sources (Dubois et al. 1978, Ramos et al. 1985). Insignificant amounts of ornithine have been found in grape juice, whereas GABA has been detected at concentrations ranging from 50 to 250 mg/L (Monteiro and Bisson 1991b, Guitart et al. 1997, Torrea-Goni and Ancin-Azpilicueta 2002, Soufleros et al. 2003, Hernandez-Orte et al. 2006, Bisson 1991) and may thus account for a significant proportion of the assimilable nitrogen in grape juice. It has been reported that, during wine fermentation, the rate of GABA assimilation is lower than that of preferred nitrogen sources (Monteiro and Bisson 1991a, Torrea-Goni and Ancin-Azpilicueta 2002), consistent with the NCR control exerted on the genes involved in GABA transport (Daugherty et al. 1993, Talibi et al. 1995). It has also been shown that GABA is not completely removed from the medium during the winemaking process (Monteiro and Bisson 1991a, Torrea-Goni and Ancin-Azpilicueta 2002). However, the contribution of GABA to wine fermentation remains unclear.
The GABA transported into yeast cells is assimilated through the GABA pathway (Figure 1⇓), which is also known as the UGA regulon (Ramos et al. 1985, Andre and Jauniaux 1990, Coleman et al. 2001). In the first step of this pathway, the GABA transaminase Uga1p catalyzes the transfer of the amine group from GABA to α-ketoglutarate (α-KG), producing glutamate. This reaction generates succinate semi-aldehyde (SSA), which is further oxidized into succinate by the SSA dehydrogenase Uga2p. We have recently reported that S. cerevisiae cells convert up to 4 g/L GABA into succinate, with a molar yield of 0.7 mol/mol, during fermentation on synthetic medium (Bach et al. 2009), suggesting that GABA may represent a significant source of succinate in wine. During wine fermentation, this organic acid is produced mostly by the reductive branch of the tricarboxylic acid (TCA) cycle and, to a lesser extent, by oxidative decarboxylation of α-KG (Camarasa et al. 2003). The possibility of converting GABA into succinate through the GABA pathway is of great interest in enology because succinate, which provides a combination of saltiness, bitterness, and acidity (Peynaud 1984), contributes to the sensory characteristics and microbiological stability of wines.
Saccharomyces cerevisiae can also synthesize endogenous GABA from glutamate through the action of glutamate decarboxylase (GAD), encoded by GAD1 (Ramos et al. 1985, Coleman et al. 2001). We recently reported that the endogenous production of GABA through Gad1p is strongly limited during fermentation (Bach et al. 2009). However, as all genes of the GAD/GABA pathway in commercial wine yeasts are induced by high sugar stress (Erasmus et al. 2003), the contribution of this route should be investigated in wine fermentation, one of the main characteristics of which is high sugar content.
Our goal was to better understand the effect of GABA on yeast fermentation and wine quality by examining the variability of GABA concentration in musts, as a function of grape origin, and how the assimilation of exogenous GABA, in concentrations similar to those found in grape juice, affect fermentation rate, growth, and metabolite production of wine yeast during the winemaking process. We also evaluated the ability of wine yeast strains to synthesize endogenous GABA.
Materials and Methods
Yeast strains.
Three commercial wine yeast strains were used: EC1118 (Lalvin, Champagne, France), D254 (ICV-Lalvin, Languedoc, France), and K1M (ICV-INRA, Languedoc, France). The model wine yeast V5 (MATa, ura3), a haploid derivative of commercial yeast, was used as a reference. The deletion mutant gad1 was generated in S. cerevisiae strain V5 (Bach et al. 2009). Saccharomyces cerevisiae strains were grown on rich YPD medium (1% Bacto yeast extract, 2% bactopeptone, 2% glucose).
Grape juice preparation.
Ripe grapes (Vitis vinifera) of different origins (summarized in Table 1⇓), were randomly harvested from 10 vines in the same domain, on the same date. The grapes (200 g) were washed, dried, and crushed. The must was then centrifuged (8,000 g, 20 min, 4°C) and used for amino acid determinations.
Fermentation conditions.
Initial cultures of yeast in YPD medium were grown in 50-mL flasks at 28°C, with shaking (150 rpm) for 12 hr. These cultures were used to inoculate subsequent cultures at a density of 1 x 106 cells/mL. Fermentation was carried out in synthetic MS medium, containing 200 g/L glucose, 6 g/L malic acid, 6 g/L citric acid, at pH 3.5 (Bely et al. 1990). Nitrogen was provided as a mixture of amino acids (75%) and NH4Cl (25%), at concentrations of 460 mg N/L for nitrogen-rich (MSH) and 105 mg N/L for nitrogen-limited (MSL) media. Ergosterol (7.5 mg/L), oleic acid (2.5 mg/L), and Tween 80 (0.21 g/L) were provided as anaerobic growth factors. When appropriate, GABA was added to the culture medium. Fermentation was carried out in 1.1-L fermentors equipped with fermentation locks, at 28°C, with continuous magnetic stirring (500 rpm). The CO2 released was monitored by automatic measurements of the weight lost by the fermentor every 20 min, and the rate of CO2 production was calculated by polynomial smoothing of the curve for CO2 release. Each fermentation was carried out in triplicate.
Analytical methods.
An electronic particle counter (ZM, Coultronics, Margency, France) was used to count the yeast cells. Glucose and fermentation products (acetate, succinate, hydroxyglutarate, α-KG, glycerol, and ethanol) were analyzed in culture supernatants by high-pressure liquid chromatography (HPLC 1100, Agilent Technologies, Santa Clara, CA) on an HPX-87H Aminex column (Bio-Rad, Hercules, CA). Dual detection was performed with a refractometer and a UV detector (Hewlett Packard, Palo Alto, CA).
GABA and other amino acids were quantified with a specific amino acid analyzer (Biochrom 20, Pharmacia, Peapack, NJ). Amino acids were separated by liquid chromatography on an ion exchange column (Ultrapac-8 Lithium form, Amersham Pharmacia Biotech, Peapack, NJ), with subsequent detection by the ninhydrin reaction followed by absorbance measurement at 570 nm, except for proline, which was detected by measuring absorbance at 440 nm. Norleucine (0.5 mM) was added to the samples and used as internal standard.
Liquid chromatography-mass spectroscopy (LC-MS) was used to assay γ-hydroxybutyric (GHB) acid in a Waters 2690 system (Waters, Milford, MA) equipped with a reversed-phase LiChrospher 100-RP18 column (250 mm x 2 mm i.d., 5 μm, Merck, Darmstadt, Germany) (Allan et al. 2003). After 1 minute of elution with 10 mM ammonium formate and 1% methanol in distilled water, separation was achieved with a linear gradient from 0 to 100% methanol in 4 min. Elution with 100% methanol was continued for 8 min. The flow rate was 0.25 mL/min. GHB was identified and quantified by mass spectrometry (negative ion mode, 50–300 arbitrary mass units) using a Thermo Finnigan LCQ Advantage MS detector (Thermo Scientific, Waltham, MA) equipped with an ESI ion source and an ion mass trap analyzer.
Results
GABA content of grape juice.
The composition and concentration of nitrogen sources in grape musts is variable, and no systematic study of the GABA content of musts has been conducted. A few studies have reported GABA concentrations, ranging from 50 to 250 mg/L (Bisson 1991, Soufleros et al. 2003, Hernandez-Orte et al. 2006). The effects of grape variety, geographic origin, and vintage on the presence of GABA and its role as a nitrogen source in musts were determined by analyzing (1) musts obtained from 21 grape varieties (same year, same domain), (2) two independent samples of Chardonnay musts from seven regions, harvested the same year, and (3) juice from Chardonnay, Alicante, Sauvignon blanc, and Syrah grapes harvested from the same vineyard for three consecutive years. There was considerable variation in GABA concentration as a function of must origin (Figure 2⇓). The GABA content of most grape juices was between 30 and 80 mg/L. However, concentrations <5 mg/L were found in some musts, whereas GABA concentrations of 300 to 500 mg/L were found in juice from Sauvignon blanc, Chardonnay, Carignan, and Cardinal varieties. In addition to grape variety, geographic origin and vintage had a large impact on GABA concentration.
GABA, at concentrations above 30 mg/L, was one of the most abundant amino acids in the grape juice, together with arginine, alanine, and serine (data not shown). The nitrogen from GABA accounted for ~10% of the assimilable nitrogen in most of the musts analyzed and for up to 25% in some grape juices (Figure 2⇑). The presence of GABA may, therefore, have an effect on yeast growth and fermentation rate.
GABA as a nitrogen source during fermentation.
The fermentation rate profile in winemaking depends on the genetic background of the yeast strain and the composition of the grape must. Typically, the fermentation rate increases rapidly to a maximum during yeast growth, and then progressively decreases during the stationary phase. Yeast growth depends mostly on the availability of nutrients, including nitrogen in particular. A minimum of 120–140 mg N/L is reportedly required for satisfactory fermentation (Agenbach 1977, Bely et al. 1990, Alexandre and Charpentier 1998). Nitrogen compounds differ in their ability to sustain growth and are sequentially removed from the medium (Monteiro and Bisson 1991a, Jiranek et al. 1995b).
The efficiency of GABA as a nitrogen source during wine fermentation under nitrogen-limited and nitrogen-rich conditions was investigated using two synthetic musts: MSL and MSH (Bely et al. 1990), containing 107 and 460 mg total nitrogen/L nitrogen (as ammonium [25%] and amino acids [75%]), respectively. The media were both supplemented or not supplemented with 500 mg/L GABA, which provided 68 mg/L of extra nitrogen. Wine yeast strains have an unusual genetic background and their properties are different than nonindustrial strains. Furthermore, wine yeast strains differ in their nitrogen requirements (Manginot et al. 1998). We used the haploid wine yeast model V5, which has previously been used to investigate GABA metabolism (Bach et al. 2009), and the three commercial wine yeast strains K1M, D254, and EC1118, which have high, medium, and low nitrogen requirements, respectively (Manginot et al. 1998, Julien et al. 2000; www.lallemandwine.us/). Fermentation by V5 in MSL supplemented with 260 mg/L NH4Cl was used as a control. Final yeast population and fermentation kinetics were monitored for three independent series of fermentations.
GABA was entirely consumed when added to nitrogen-limited medium at the beginning of fermentation (Figure 3A⇓). Under these conditions, the final population was much greater (15% for EC1118 and D254 and up to 42% for K1M) than for fermentations without GABA. The efficiency of GABA utilization depended on genetic background. Previous studies have reported differences between strains in the patterns of utilization of and yields from nitrogen compounds (Cooper 1982, Boulton et al. 1996). The increase in the final population of V5 resulting from the addition of GABA was smaller than that generated by the addition of ammonium salts, possibly due to the lower nitrogen availability resulting from the incomplete conversion of GABA into succinate (molar yield: 0.87 mol/mol; Table 2⇓).
When grown in MSL medium (Figure 4⇓), the three commercial yeast strains had similar maximal fermentation rates (expressed as rate of CO2 release: 1.1 to 1.25 g CO2/L/hr), fermentation profiles, and durations (230 hr). The fermentation profile of the haploid strain V5 was different and the fermentation duration was shorter than for the other strains (130 hr). The addition of 500 mg/L GABA increased the maximum fermentation rate by 44, 59, 66, and 68% for V5, K1M, EC1118, and D254, respectively. All the strains had a higher CO2 production rate throughout the stationary phase in media containing GABA, and the fermentation duration was shortened by ~80 hr. The addition of GABA during V5 fermentation was as efficient as the addition of NH4Cl.
GABA addition to nitrogen-rich medium (MSH) had no observed effect on yeast population (Figure 3B⇑) or the rate of CO2 production for the four strains studied (data not shown).
Impact of GABA assimilation on by-product formation.
The effect of GABA on yeast metabolism and the production of succinate was investigated by measuring the residual GABA content and the concentration in culture supernatants of the main fermentation products (ethanol, acetate, pyruvate, glycerol, succinate) after complete sugar exhaustion in the fermentation experiments described above. GABA was entirely consumed during fermentation on MSL. In contrast, only 77 to 88% of GABA, depending on yeast strain (Figure 3B⇑), was consumed in nitrogen-rich conditions (MSH). Only the production of succinate and glycerol was affected by the presence of GABA in MSL and MSH media. There was no effect on the formation of other metabolites (acetate, pyruvate, ethanol; data not shown). The addition of GABA at the beginning of fermentation resulted in a significant increase (by a factor of about two) in the production of succinate by the four strains, in both MSL and MSH media (p values between 0.001 and 1.10−5) (Figure 3⇑). The supplementation of MSL medium with 260 mg/L NH4Cl, however, had no effect on succinate formation, as shown for strain V5. The increase in succinate content was directly related to the conversion of GABA into succinate catalyzed by Uga1p and Uga2p, which has been identified as the major route of GABA catabolism in S. cerevisiae (Ramos et al. 1985, Coleman et al. 2001). The molar yields for conversion of GABA into succinate by the four studied strains during fermentations with MSL medium were up to 15% higher than those obtained during cultures in MSH (Table 2⇑). This result and the incomplete consumption of GABA during fermentation in nitrogen-rich conditions (Figure 3B⇑) indicate that GABA is used more efficiently in conditions of nitrogen limitation.
Glycerol formation, which serves as a redox valve for disposing of the excess reducing power generated during yeast growth (Van Dijken and Scheffers 1986), is another consequence of GABA assimilation. Glycerol was produced at higher concentrations in nitrogen-limited MSL medium supplemented with GABA or NH4Cl (p values between 0.02 and 0.05) (Figure 3A⇑). Higher levels of biomass production were also observed under these conditions. However, glycerol production and population growth were unaffected by the addition of GABA to MSH medium. The higher levels of glycerol production in MSL supplemented with GABA may be an indirect consequence of the assimilation of this γ-amino acid, resulting from the need to reoxidize the surplus NADH generated by the extra biomass formed.
We previously demonstrated that a metabolic alternative to the oxidation of GABA into succinate was its reduction into γ-hydroxybutyric acid (GHB), which can be further catabolized into polyhydroxybutyrate (PHB) (Bach et al. 2009). As GHB is toxic (Sporer et al. 2003), the possibility that wine yeast strains may excrete this metabolite during fermentation was examined. However, GHB was not detected at the end of fermentation (detection threshold 1 mg/L).
Formation of succinate from endogenous GABA.
Based on microarray analysis of the response to high sugar stress, researchers have suggested that the GAD/GABA pathway may contribute to the formation of succinate from glutamate during winemaking under sugar stress conditions (Erasmus et al. 2003). Indeed, GAD1 (which encodes the glutamate decarboxylase catalyzing the conversion of glutamate into GABA), UGA1, and UGA2 (responsible for GABA catabolism; Figure 1⇑) are more strongly induced when yeasts are grown in grape juice containing 40% (w/v) sugar than in fermentations on 22% (w/v) sugar (Erasmus et al. 2003). However, these authors did not investigate the consequences of high sugar content in the must on the excretion of succinate.
We tested this hypothesis by deleting GAD1 (encoding the glutamate dehydrogenase) from strain V5 and assessing the effect of this deletion on the production of succinate during fermentation in MSH medium containing 20%, 30%, or 40% (w/v) glucose (Figure 5⇓). The succinate yield (mg/g glucose) did not increase in the presence of high sugar concentrations and no difference in the yield of succinate was found between strains V5 and V5 gad1. This indicated that the glutamate decarboxylase and, consequently, the GAD/GABA pathway (Figure 1⇑) are not involved in the production of succinate by the strain V5 during wine fermentation, even in the presence of high sugar concentrations.
Discussion
In this study, three commercial wine yeasts with different genetic backgrounds efficiently metabolized exogenous GABA. This finding is consistent with previous observations showing that the profile of consumption of this γ-amino acid by the V5 strain is similar to that for total nitrogen (Bach et al. 2009). In addition, the presence of GABA in grape juice with low nitrogen content had a positive effect on yeast performance, considerably improving the fermentation rate and final population density and reducing fermentation time. These results underscore that GABA, which may account for 20% of the assimilable nitrogen in grape juice, should be considered an efficient nitrogen source in winemaking. However, this compound is only partly removed during fermentation. Under nitrogen-rich conditions, the presence of GABA affects neither the growth nor the fermentation performance of wine yeasts.
Our findings highlight the importance of GABA catabolism as a source of succinate during winemaking. Succinate contributes to the sensory characteristics by increasing the vinous, sapid, salty, and sometimes bitter taste of wine (Peynaud 1984). Moreover, grape juices from warm climate regions or prepared from fully ripened grapes often contain very high sugar concentrations and insufficient amounts of malic and tartaric acids, resulting in a lack of acidity in wine. High levels of succinate may counter this imbalance. Succinate is the main carboxylic acid produced by yeast during wine fermentation and is usually found in concentrations from 0.5 (white wines) to 1.2 g/L (red wines) and occasionally up to 2 g/L (Coulter et al. 2004). During wine fermentation, succinate is produced from sugars through the reductive branch of the TCA cycle and, to a lesser extent, by α-ketoglutarate decarboxylation (Albers et al. 1996, Camarasa et al. 2003) (Figure 1⇑). The amount of succinate derived from the TCA cycle during wine fermentation in MS medium was between 300 and 700 mg/L, depending on yeast strain (Remize et al. 2000). We found that the yield for the conversion of GABA into succinate varied from 0.75 to 1.0 mol/mol, depending on the genetic background of the yeast and on initial nitrogen content of the medium. As reported here and in other studies (Bisson 1991), GABA concentration of grape juice may reach 580 mg/L. Therefore, up to 665 mg/L succinate may be produced from exogenous GABA in addition to the usual formation of succinate from sugars. A comparison of the concentrations of succinate generated by these two metabolic routes showed that GABA could serve as a major source of succinate in wine, depending on its concentration in grape juice. A considerable variation in the succinate content of wines has been reported and has been attributed to several factors, including yeast strain, fermentation temperature, oxygenation, must clarity, and composition (Coulter et al. 2004). Results here suggest that GABA should be considered an important factor in the variation of succinate concentration and that high concentrations of GABA in grape juice could account for the high levels of succinate found in some wines. However, the factors favoring the accumulation of GABA in grapes have not been clearly identified. Carbonic maceration produces grape juice with higher GABA levels because anaerobic stress results in GABA storage in the grapes (Tesniere et al. 1994). GABA has also been reported to accumulate in cells when the plants are grown under anoxic or water stress conditions (Bouche and Fromm 2004).
According to our previous study, the small proportion of GABA that was not converted into succinate could either be stored intracellularly or reduced into γ-hydroxybutyrate (GHB), which could then be further polymerized into polyhydroxybutyrate (PHB) (Bach et al. 2009). GHB has been detected in wines at concentrations ranging from 4 to 20 mg/L (Elliott and Burgess 2005). We therefore investigated the possibility that this compound might be excreted into the medium as a by-product of GABA catabolism. However, we failed to detect GHB after fermentation by commercial wine yeast strains in the presence of 500 mg/L GABA, which is consistent with previous data obtained with extremely high GABA concentrations (Bach et al. 2009) and suggests that commercial wine yeasts are not responsible for the presence of GHB in wines. A bacterial provenance of GHB cannot be excluded, as higher levels of GHB were found in red wines subject to malolactic fermentation than in other wines.
Conclusion
Results show that the presence of GABA in grape must has a positive effect on fermentation kinetics and metabolite production, particularly for succinate. Consequently, our results highlight the importance of taking into account the GABA content of musts for the optimal control of fermentation and wine acidity. The range of GABA concentrations in grape juice is broad, reaching 580 mg/L. GABA concentration varies with grape variety, year, and geographic origin. However, the physiological mechanisms responsible for GABA storage in grapes should be investigated. The identification of methods of cultivation and/or environmental factors favoring the production of grape juice with high levels of GABA might make it possible to develop wines with higher succinate levels.
Footnotes
↵2 (present address) Inter-Rhône F-84100 Orange, France.
- Received March 2009.
- Revision received May 2009.
- Accepted June 2009.
- Published online December 2009
- Copyright © 2009 by the American Society for Enology and Viticulture