Gaseous environments modify physiology in the brewing yeast Saccharomyces cerevisiae during batch alcoholic fermentation


Gilles Feron, UMR1129 FLAVIC, ENESAD, INRA, Université de Bourgogne, 17 rue Sully BP 86510, F-21065 Dijon, France.


Aims:  To investigate the impact of different gaseous atmospheres on different physiological parameters in the brewing yeast Saccharomyces cerevisiae BRAS291 during batch fermentation.

Methods and Results:  Yeasts were cultivated on a defined medium with a continuous sparging of hydrogen, helium and oxygen or without gas, permitting to obtain three values of external redox. High differences were observed concerning viable cell number, size and metabolites produced during the cultures. The ethanol yields were diminished whereas glycerol, succinate, acetoin, acetate and acetaldehyde yields were enhanced significantly. Moreover, we observed major changes in the intracellular NADH/NAD+ and GSH/GSSG ratio.

Conclusions:  The use of gas led to drastic changes in the cell size, primary energy metabolism and internal redox balance and Eh. These changes were different depending on the gas applied throughout the culture.

Significance and Impact of the Study:  For the first time, our study describes the influence of various gases on the physiology of the brewing yeast S. cerevisiae. These influences concern mainly yeast growth, cell structure, carbon and redox metabolisms. This work may have important implications in alcohol-related industries, where different strategies are currently developed to control better the production of metabolites with a particular attention to glycerol and ethanol.


Changes in the parameters during alcoholic fermentation process can lead to major modifications concerning the quality of the final product. Researchers as well as industries have unceasingly searched to improve the quality of production by modifying cell metabolism. Studies have focused mainly on the following targets: extension of substrate range, improvement of productivity and yield, elimination of undesirable by-products, improvement of process performance, improvement of cellular properties and enhancement of characteristics of interest. For this purpose, two principal approaches have been usually undertaken consisting of molecular engineering and modification of environmental parameters.

Molecular engineering of Saccharomyces cerevisiae has developed widely. The different works with different aims were reviewed by Ostergaard et al. (2000). Metabolic engineering of yeast focused on particular metabolism such as malolactic fermentation (Volschenk et al. 2001; Husnika et al. 2006) and xylose fermentation (Kuyper et al. 2004; Jeffries 2006; Chu and Lee 2007). A large number of works were conducted on alcoholic fermentation. They consisted of modifying carbon flux which impacted mainly on glycerol, acetate or ethanol productions (Nissen et al. 2000; B.V. Alfenore et al. 2002;Nevoigt et al. 2002; Taherzadeh et al. 2002; Geertman et al. 2006; Heux et al. 2006a,b; Wang and Hatzimanikatis 2006; Cordier et al. 2007).

However, genetic manipulations have not been often promoted in the industry because of the GMO (genetically modified organism)-phobia that occurs in the consumer attitude. In place, metabolic modifications could be performed by applying changes to the physicochemical parameters of the environment where the yeast is cultivated. These solutions appear more appropriate to the consumer demand for safe product. Modifications of growth and metabolism were observed by changing pH and citric acid concentrations (Nielsen and Arneborg 2007) or assimilable nitrogen (Wang et al. 2003; Bohlscheid et al. 2007) in the medium. Similarly, a dynamic model of alcoholic fermentation in wine-making conditions with relation to temperature and nitrogen additions (Malherbe et al. 2004) was proposed to have a better control of the fermentation process. In this model, change in temperature and/or assimilable nitrogen concentration allowed to enhance the quality of the final product. Addition of electron acceptors influenced the by-products of alcoholic fermentation (Roustan and Sablayrolles 2003). Modification in oxygen addition during an oenological fermentation has impact on sterol contents of S. cerevisiae as well as their reactivity towards oxygen (Fornairon-Bonnefond et al. 2003).

Compared with other different environmental parameters such as pH, temperature, water activity, etc., the oxidoreduction potential, Eh, was studied more recently. Effects of redox potential on microbial physiology have been demonstrated early in aerobic as well as anaerobic micro-organisms (Wimpenny and Necklen 1971; Andreeva and Rabotnova 1978). Recently, many works concentrated on the influence of redox environment on Escherichia coli growth and metabolism (Silva and Wong 1995; George and Peck 1998; Gill et al. 1998; Bagramyan et al. 2000; Riondet et al. 2000). In Lactococcus lactis, recent works showed an effect of the redox environment on the synthesis of flavour during growth and fermentation (Kieronczyk et al. 2006).

In yeast, few studies have described the influence of redox environment on physiology of S. cerevisiae (Cachon et al. 2002), Yarrowia lipolytica (Husson et al. 2006) and more recently Sporidiobolus ruinenii (Feron et al. 2007). Husson et al. (2006) showed that reducing conditions were favourable to hydroperoxide lyase (HPL) synthesis as well as its activity but it was unfavourable to yeast growth. In Spo. ruinenii, lowering the Eh7 (Eh at pH 7) to neutral conditions favoured the production of γ-decalactone, an aroma compound, but this production was very low under reducing conditions (Feron et al. 2007). These studies suggest that external Eh is able to modify yeast physiology and thus modify its metabolism. With regard to the importance of S. cerevisiae in different food processes (wine making, brewing), it appears essential to estimate the impact of Eh modifications on its physiological behaviour and then on the overall quality of the fermented products.

Previous study reported different means to modify redox environment. It concerned the use of chemical compounds such as dithiothreitol (DTT), potassium ferricyanide [FeK(CN)6] or 2,6-dichloroindophenol (DCPIP) (George and Peck 1998; Roustan and Sablayrolles 2003; Husson et al. 2006). However, it appears difficult to apply this technique to industrial production. Other studies described the use of different gases (nitrogen, oxygen, hydrogen) to modify the Eh environment (George et al. 1998; Riondet et al. 1999, 2000; Laurinavichene et al. 2001; Ouvry et al. 2002; Alwazeer et al. 2003; Bourel et al. 2003; Husson et al. 2006; Feron et al. 2007). This technique is more flexible and easier to be applied in the industry. Moreover beyond the sole modulation of Eh, the use of gas during alcoholic fermentation could be considered as a valuable technology.

The purpose of this study is to evaluate the effect of different gases on the physiology in the yeast S. cerevisiae. We will examine the consequences of different gaseous conditions on the growth, cell structure and carbon metabolism by following the formation of major fermentative metabolites and the internal redox state of the yeast. Our findings will help to understand the effect of gas on S. cerevisiae redox and fermentative metabolism and to propose a new strategy for the development of processes modifying some of the metabolite yields in alcoholic fermentation.

Materials and methods


Brewing yeast strain S. cerevisiae BRAS291 (bottom fermentation) was kindly provided by the BRAS Collection of the Unité de Brasserie et des Industries Alimentaires, Université Catholique de Louvain (Louvain-la-Neuve, Belgium). The yeast strain was stored at –80°C in a 10% v/v glycerol solution.


Compressed gases (oxygen, hydrogen and helium) were obtained from Air Liquide (France). Gas purities were about 99·99%. Hydrogen and oxygen gases have been selected as reducing and oxidizing agents, respectively. Helium has been selected because of its neutral property and its similarity to hydrogen in term of size of the molecule and diffusivity (Air Liquide 2002).

Culture conditions

Subcultures of S. cerevisiae BRAS291 were grown in YPGM medium (1% w/v yeast extract, 0·5% w/v peptone, 5% w/v glucose and 5% w/v maltose) at 28°C on a rotary shaker (120 rev min−1). The inoculation at 1 × 106 cells ml−1 in a fully defined medium was performed by using yeast suspension collected from a 24-h subculture.

The fully defined medium constituted by Kennedy et al. (1997) and Taidi et al. (2003) was used as previously described (Pham et al. 2008). The cultures were carried out at 28°C during 13 days with permanent stirring (250 rev min−1) in a Biostat Q batch fermentation system.

Three gaseous environments were obtained by sparging the bioreactor with hydrogen, helium and oxygen with a gas flow at 0·03 (v/v/min). The off gas passed through a cooled condenser (4°C) to avoid evaporation and stripping of volatile compounds (Heux et al. 2006a,b; Berovic and Herga 2007), except ethanol. The control condition was the gas-free culture. The pH value of the culture conducted under oxygen was regulated close to that of the other conditions to ensure a normal yeast growth. This regulation was started from the second day of culture by automatic addition of NaOH (10 mol l−1) to obtain a target pH value of 4·0 during the second day, and 3·8 until the end of culture.

To measure the dissolved oxygen concentration, the equipments were calibrated with air. The gas-free culture started at 100% of dissolved oxygen; the oxygen concentration decreased to 0% after 4 h of fermentation. The hydrogen and helium cultures started at 0% and the oxygen culture at 400% of dissolved oxygen using pure oxygen. These concentrations were constant throughout the cultures.

Data acquisition

To test sensitivity towards the gaseous environment, a Biostat Q multiple fermentor unit (B. Braun Biotech International, Melsungen, Germany) was used to follow simultaneously the pH, the oxidoreduction potential measured (Em in mV) and the dissolved oxygen content of the medium during the course of fermentation. Additionally, pH electrodes (405-DPAS-SC K8S/200; Mettler Toledo SARL, Paris, France), redox electrodes (Pt 4805-DPAS-SC K8S/200; Mettler Toledo SARL) and oxygen sensors (InPro 6100/1200/T/N; Mettler-Toledo) were connected to the Biostat Q unit.

The pH electrodes were calibrated using two buffers: pH 4·0 and 7·0. The redox electrodes were polished with alumina powder (aluminium oxide; Prolabo, Lyon, France) to restore the platinum surface (Jacob 1970). A software system (MFCS win 2·0; B. Braun Biotech International) run by a PC recorded automatically pH, Em and dissolved oxygen values during fermentation.

Based on the standard electrode value (Eref) at the temperature of fermentation (Eref = 205 mV), the measured values of the redox potential (Em, expressed in comparison with reference electrode Ag/AgCl) were converted into Eh values (redox potential, expressed in comparison with hydrogen electrode, Eh = Em + Eref). The decrease of medium pH (such as done by yeast in alcoholic fermentation) indirectly modifies Eh (Nernstian response of the redox electrode); then Eh was expressed as Eh at a given value of pH, pH 7 (Eh7) according to the following equation: Eh7 = Eh – α × (7 – pHx) where α is the Eh–pH correlation factor determined experimentally as 41, 52, 59, 42 for the control, hydrogen, helium and oxygen cultures respectively, and pHx is the pH of the medium.

Analytical methods

Viable cells number

Viable yeast cell concentration was monitored using a methylene blue staining method as described previously (Pham et al. 2008). Methylene blue solution consisted of: 0·025% methylene blue (Sigma-Aldrich, USA), 0·9% NaCl, 0·042% KCl, 0·032% CaCl2, 0·02% NaHCO3 and 1% glucose. Yeast slurries (1 ml) were diluted in sterile physiological water (0·9% NaCl). Equal volumes of methylene blue solution and diluted yeast cells were wet mounted on a Malassez count cell (Poly Labo, Paul Block & Cie, France). The analysis was performed by light microscopy, using a microscope Leitz Laborlux 513558 (×40) (Ernst Leitz Wetzlar gmbh, Leitz Portugal). The number of counted yeast cells was 300 at a minimum. Cells which are viable were not blue.

Growth curves during the first 24 h of culture were fitted by using program developed by SigmaPlot (version 2000; spss Inc., USA) (Feron et al. 1996) (function inline image where X is the number of cells, a and d are the asymptotic minimum and maximum values, b is the slope parameter, c is the t value at the asymptotic point and e is a symmetry parameter. After curve fitting, specific growth rate (μ) was calculated as described by Jason (1983) and is given byinline image where ΔX and inline image are respectively the difference and the mean between the two values selected, and Δt is the interval time.

Dry biomass determination

Culture samples were vacuum filtered over Durapore membrane filters (pore size 0·45 μm; Millipore Corporation, Ireland). After removal of the medium the filters were washed with permuted water and dried in a dry-heater at 63°C for 48 h and then weighed. The cell dry weight was determined by the difference between the weights of the membrane filter before and after drying.

Cell size measurement

Cell measurement by direct visualization under the microscope was described by Poirier et al. (1998). The volume of yeast cells was measured using a light microscope Leitz Laborlux 513558 (×40) (Ernst Leitz Wetzlar GmbH), a digital camera Canon Power Short S40 (Canon inc., Japan) and an image analysis system (series 151; Imaging Technology, Woburn, MA, USA). The cells were analysed individually by determination of their area (S). Assuming cells of S. cerevisiae BRAS291 to be spherical, the volume (V) was then calculated from the measured area. Measurement was performed on 100 cells at a minimum (Pham et al. 2008).

Substrates and metabolites analysis

Culture samples were centrifuged at 4°C and 5000 g for 10 min and the supernatants were used for determination of sugars and metabolites.

High-performance liquid chromatography (HPLC) analysis. Fermentable sugars (maltose, glucose, fructose, maltotriose) and metabolite products (ethanol, glycerol, succinate, lactate, acetate) were analysed on a Merck HPLC apparatus (Merck, France) containing an Aminex HPX 87H column (Bio-Rad, France). The column was eluted at 65°C with 0·5 mmol l−1 of sulfuric acid at a flow rate of 0·6 ml min−1. Detection was by means of a Bischoff IR 8110 refractive detector.

Enzymatic analysis. Pyruvate was determined by the method described by Martin (1998). The assay for pyruvate was carried out directly in a 1-cm (light path) cuvette containing 2·15 ml of pH 7 phosphate buffer (0·3 mmol l−1), 250 μl of 0·002 mol l−1 NADH solution and 100 μl of lactate dehydrogenase solution (0·4 units in 10 ml of pH 7 phosphate buffer). The reaction is started by adding up to 500 μl of test sample, quickly mixing and the reaction progress was monitored by measuring the decrease of absorbance at 340 nm against a control using a Perkin-Elmer Lambda 15 UV/VIS spectrophotometer (Perkin-Elmer, USA). Malate was measured by using an enzymatic kit no. 11 215 558 035 (Boehringer Mannheim/R-Biopharm, Roche, France).

Measurement of ethanol stripping

Ethanol accumulated in high concentrations during fermentation and the application of gas continuously throughout the culture led to its evaporation (Duboc and Stockar 1998). The effect of gases on ethanol stripping was measured experimentally by sparging at 0·03 vvm the different gases through 800 ml of culture medium supplemented with 40 g of ethanol and by measuring the changes in ethanol concentration in the fermentor during 13 days. On these conditions, the maximum rate of ethanol stripped was estimated as 0·014, 0·016 and 0·025 g of ethanol per day and per gram of ethanol in the medium for hydrogen, helium and oxygen conditions, respectively. The stripping effect was taken into account for calculating all the ethanol concentrations presented in this paper.

Determination of intracellular cofactors

Oxidized and reduced cofactors were extracted using the method described by Mailinger et al. (1998) with little modifications. Cell suspensions (3·5 ml) were injected immediately into either 1·0 ml of 35% v/v perchloric acid (–20°C) to extract oxidized cofactors or a 1 mol l−1 of alcoholic KOH (–20°C, 50% v/v in ethanol) to extract reduced cofactors. For optimal extraction, three freeze-thaw cycles (between –80°C and 0°C) were performed. Maximal recovery of reduced cofactors was achieved when alkaline extracts were further incubated for 7 min at 70°C. After cooling to 0°C and incubating for 10 min, cofactor extracts were neutralized to pH 7 by carefully adding and thoroughly vortexing with either 4 mol l−1 of KOH or 1 mol l−1 of HCl (final volume of 5 ml). The KClO4 precipitate was removed by filtration. Because of the interference of reducing components in the alkaline extracts, reduced cofactors were converted to the oxidized forms with the aid of 5 μl of glutamate dehydrogenase (2 mg ml−1) in the presence of 50 μl of 2-oxoglutarate (0·1 mol l−1 in 0·2 mol l−1 NH4Cl). The reaction proceeded for 20 min at 25°C and was terminated with 100 μl of HCl (5 mol l−1) at 50°C (5 min). The extracts were neutralized to pH 7·0 by adding 5 mol l−1 (qsp 2·5 ml) of NaOH. The extracts were conserved at –80°C before analysis.

The analytic system was composed of a chromatographic system (Waters, USA) containing a Kromasil C18 5U 100 A column (AIT, France), a WatersTM 600 controller and a Waters 717plus autosampler. The column was eluted at room temperature with tetrabutyl ammonium hydroxide (TBAH) solvents (solvent A: 10 mmol l−1 TBAH in 0·125% methanol and 10 mmol l−1 KH2PO4; solvent B: 2·8 mmol l−1 TBAH in 30% methanol and 100 mmol l−1 KH2PO4) at a flow rate of 0·9 ml min−1. The following gradient was used: 100% A: 0% B from 0 to 12 min (0·9 ml min−1), then 80% A: 20% B from 12 to 14 min (0·9 ml min−1) and 80% A: 20% B from 14 to 14·01 min (1·1 ml min−1); 80% A: 20% B to 70% A: 30% B from 14·01 to 23 min (1·2 ml min−1); 70% A: 30% B to 55% A: 45% B from 23 to 35 min (1·2 ml min−1); 55% A: 45% B to 40% A: 60% B from 35 to 46 min (1·2 ml min−1); 40% A: 60% B to 35% A: 65% B from 46 to 55 min (1·2 ml min−1); 35% A: 65% B to 25% A: 75% B from 55 to 65 min (0·9 ml min−1); 25% A: 75% B to 0% A: 100% B from 65 to 90 min (0·9 ml min−1); 0% A: 100% B to 100% A; 0% B from 90 to 130 min (0·9 ml min−1). Detection at 267 nm was by means of a WatersTM 996 Photodiode Array Detector. Data acquisition was performed by Millenium32TM software (Waters). A standard curve generated with the known amount of cofactors was used to determine specimen concentrations from pick surfaces.

Intracellular GSH/GSSH determination

Intracellular glutathione (GSH, GSSG) was extracted using the method described by Perrone et al. (2005) with little modifications. Yeast cells were harvested by centrifugation at 5000 g at 4°C for 10 min and washed twice with distilled water. Cells were then suspended again in ice-cold 1·3% (w/v) 5-sulfosalicylic acid and broken by ultrasonic sound (Vibracell 72412; Biolock Scientific, France) for three 30-s periods each at 1-min intervals, followed by incubation on ice (15 min) to precipitate the proteins. Cell debris and proteins were eliminated by centrifugation (15 min, 13 000 g at 4°C) and the glutathione levels were determined in the resulting supernatant by enzymatic method.

Glutathione concentration was measured using GSH Assay Kit WPI No. 062404 (World Precision Instruments), by measuring the absorbance of reaction solutions at 412 nm using a spectrophotometer (Perkin-Elmer Lambda 15 UV/VIS; Perkin-Elmer). Total glutathione refers to total concentration of free glutathione (does not include protein-bound glutathione), i.e. [GSH] + 2 × [GSSG].

GSH reacted with 5,5′-dithiobis(nitrobenzoic acid) or DTNB to produce a chromophore TNB with maximal absorbance at 412 nm and oxidized glutathione GSSG which was reduced to GSH by enzyme glutathione reductase (GR) in the presence of NADPH. Exactly 100 μl of the sample was added to cuvette containing 30 μl of supplied glutathione reductase solution and 30 μl of supplied DTNB solution. The reaction was started by adding of 100 μl of supplied NADPH solution and the absorbance was recorded for 3 min every 15 s. A standard curve generated with the known amount of glutathione was used to determine specimen concentrations. For quantification of oxidized glutathione, GSSG, samples were pretreated with 5% (v/v) 2-vinylpyridine for 1 h at room temperature before analysis.

Statistical analysis

The maximal specific growth rate (μmax) and the time of μmax data were analysed by a one-way analysis of variance (anova), along with the Turkey–Fisher Honestly Difference (HSD). The metabolite yields data were analysed by one-way anova, along with the Multiple Comparisons with the Best (MCB; Hsu 1996). Analyses based on three replicates were carried out using the Minitab software package (version 13; Minitab Inc., PA, USA). The measure of variation was a standard deviation. Differences were considered significant when P ≤ 0·05.


Effect of gases on Eh7 during cultures of Saccharomyces cerevisiae BRAS291

Eh (expressed as Eh7) parameter was followed during 13 days of culture performed under different gaseous conditions (hydrogen, helium and oxygen), namely the ‘gas-in’ conditions, and without gas, namely the ‘gas-free’ condition.

The introduction of gases in the culture medium changed the environmental Eh7 in different ways. The use of hydrogen and oxygen helped in maintaining Eh7 throughout the culture at about +515 mV and about –385 mV, respectively (Fig. 1). In the gas-free culture, an important decrease, from around +500 mV to around –175 mV, was observed during the first 24 h. Then Eh7 was stable for 3 days but increased afterwards to reach c.–128 mV at the end of the cultures. A similar profile was observed on helium culture: a drop from an initial Eh7 value of +350 mV to a value of –195 mV after 2 days, and then a slight increase to reach +40 mV at the end of the culture. Consequently, the use of gases allowed creating three levels of redox environment: (i) high oxidizing condition under oxygen; (ii) high reducing condition under hydrogen; and (iii) low reducing condition under helium.

Figure 1.

 Changes in Eh7 during the cultures of Saccharomyces cerevisiae BRAS291 under different gaseous conditions: hydrogen (□); helium (◊); oxygen (bsl00084); gas-free (bsl00001). Data points represent the mean from three independent experiments. Error bars omitted for reasons of clarity. Maximum observed relative standard deviation of Eh7 was 29 mV.

Growth and cell size of Saccharomyces cerevisiae BRAS291 under different gaseous environments

Viable yeast cells number was measured during cultures performed with the three gaseous environments and without gas (Fig. 2a). In the gaseous environments, the number of cells was smaller than that in the gas-free condition. It was between 26 × 10cells ml−1 after 1 day of culture and 20 × 106 cells ml−1 at the end of the culture in gas-in conditions, against 40 × 106 and 32 × 106 cells ml−1, respectively, in gas-free culture. These data showed that gaseous environments affected unfavourably viable yeast cell numbers.

Figure 2.

 Viable cell number (a) and cell size changes (b) of Saccharomyces cerevisiae BRAS291 during cultures under different gaseous conditions: hydrogen (□); helium (◊); oxygen (bsl00084); gas-free (bsl00001). Data points represent the mean and standard deviation from three independent experiments.

Relationship between yeast growth and cell size has been studied. Figure 2b presents changes in cell size during cultures performed with various gaseous environments. Three different cell size profiles were observed during the cultures. In the gas-free culture, cell volume remained nearly stable (500 μm3) until the sixth day. Afterwards, it increased greatly to double at the eighth day and remained nearly stable over the time. Under hydrogen and helium conditions, we did not observe a significant change in cell volume. It remained almost constant in the entire culture in comparison with the initial cell volume (500 μm3). On the contrary, under the oxygen condition, yeast cell volume changed significantly with time. During the first two days of culture, it was almost stable. Then, after 2 days, cell volume decreased strongly to reach approximately a value of 220–250 μm3. On the sixth day, this volume remained stable until the end of the culture. Compared with yeast cell sizes at the end of the gas-free culture, the cell volumes under hydrogen and helium conditions were two times smaller and those under oxygen were four times smaller. These results showed that gaseous environments were unfavourable to the development of cell size. Moreover, observation of yeast cells under transmission electron microscope at the beginning of fermentation showed that gaseous environments led to drastic changes in cellular ultrastructure (data not shown). This could be linked to the changes in cell size described before.

Additionally, we focused on the effect of gaseous environments during the first 24 h of culture. The results showed that maximal specific growth rates (μmax) and times of μmax (Tμmax) were not significantly different between applied conditions (P > 0·05) (Table 1). However, there is an obvious tendency for μmax to occur at different times, the earliest of which happened in gas-free culture (4·69 h) and the latest in the hydrogen culture (6·28 h). Apparently, reducing and oxidizing conditions seemed to delay the start of yeast growth. This delay corresponds to almost one generation of cells. It seems a likely link to the diauxic shift from glucose to maltose which was affected probably by the Eh7 environment as shown next. After the diauxic shift and when glucose was exhausted, this modification on yeast growth was preserved all over the culture. This reinforces the hypothesis that hydrogen and oxygen may affect growth when cells were cultivated on polysaccharide.

Table 1.   Maximum specific growth rates and times to reach maximum specific growth rate observed in culture of Saccharomyces cerevisiae BRAS 291 cultivated under different gaseous environments
 Culturesμmax (h−1)Tμmax (h)
  1. Data represent the mean and standard deviation from three separate experiments. Mean values within the same columns followed by the same letters are not statistically different (P > 0·05).

Gas-free0·48 ± 0·02a4·69 ± 0·70b
Hydrogen0·54 ± 0·10a6·28 ± 0·89b
Helium0·52 ± 0·14a5·16 ± 0·30b
Oxygen0·47 ± 0·04a5·72 ± 0·35b

Fermentation kinetics under different gaseous environments

An influence of different gaseous environments was observed before on the yeast growth and cell size. Such modifications may have a high impact on carbon metabolism because the exchanges between internal cellular space and external environment can be modified with the change in cell size. This phenomenon is explored in this part of the study by following sugar consumption in S. cerevisiae BRAS291 cultivated under different gaseous conditions (Fig. 3).

Figure 3.

 Consumption of maltose (a) and total sugar (b) during the first day of cultures of Saccharomyces cerevisiae BRAS291 under different gaseous environments: hydrogen (□); helium (◊); oxygen (bsl00084); gas-free (bsl00001). Hundred per cent corresponds to initial concentration of sugars: glucose (10·4 g l−1), fructose (4·6 g l−1), sucrose (3·5 g l−1), maltose (115·5 g l−1); total sugar (133·5 g l−1). Data points represent the mean from three independent experiments. Error bars omitted for reasons of clarity. Maximum observed relative standard deviation was 6·6%.

During the first 24 h, no differences were observed concerning the consumption of simple sugar, i.e. fructose and glucose (data not shown). However, a different phenomenon occurred when yeast shifted from simple sugars to polysaccharides. The consumptions of polysaccharide by yeast were delayed in hydrogen and oxygen cultures (Fig. 3a): 11% and 8% of maltose were consumed in hydrogen and oxygen cultures, respectively whereas this amount reached about 16% in helium and gas-free cultures. After 24 h, 26% of the total sugars were consumed during helium and gas-free cultures compared with 20% and 14% during hydrogen and oxygen cultures, respectively (Fig. 3b).

The influence of gas on sugar consumption was also followed during 13 days (Fig. 4). Simple sugars were consumed in totality after 3 or 4 days with a little delay observed for oxygen culture (data not shown). However, beyond 24 h, the profile of maltose consumption changed. In oxygen condition, maltose concentration decreased slowly during 5 days and then stabilized until the end of the fermentation (Fig. 4a). At this time, only 17% of maltose was consumed. Contrary to the observations made during the first 24 h of culture, the consumption of maltose was higher in the cultures performed with hydrogen and helium than that observed in culture performed without gas. After 13 days of culture, the amounts of maltose consumed were about 86% for the hydrogen and helium conditions whereas it was 72% for the gas-free condition. A similar profile was observed for the consumption of maltotriose (data not shown). In totality, yeast cultivated under hydrogen and helium conditions consumed more sugars compared with yeast cells cultivated without gas (89% contrary to 76%). In oxygen condition, only 29% of the total sugars were consumed (Fig. 4b).

Figure 4.

 Consumption of maltose (a) and total sugar (b) during the whole cultures of Saccharomyces cerevisiae BRAS291 under different gaseous environments: hydrogen (□); helium (◊); oxygen (bsl00084); gas-free (bsl00001). Data points represent the mean and standard deviation from three independent experiments.

Effect of gaseous environments on metabolite profiles

Fermentation by-products (ethanol, glycerol, acetate, succinate, malate, lactate, acetaldehyde, pyruvate, acetoin, 2,3-butanediol and 2-oxoglutarate) were followed during batch fermentations performed on the different conditions described before. Results are presented in Table 2 and they correspond to the data obtained after 13 days of fermentation. At this time of culture, the fermentation progress was about 0·78 for the gas-free condition whereas it was 0·36 under oxygen. In contrast, under helium and hydrogen the fermentation progress was higher with values of about 0·9 for both the gases. In oxygen culture, the biomass yield was significantly higher than that in the others cultures (Table 2).

Table 2.   Product yields, biomass and carbon balance on the thirteenth day cultures of Saccharomyces cerevisiae BRAS291 in different gaseous environments
Cultures EhNo gas (–175 ÷ –125 mV)Hydrogen (–385 mV)Helium (–195 ÷ +40 mV)Oxygen (+515 mV)
  1. Data represent the mean and values in brackets represent standard deviation from three independent experiments.

  2. *Significantly different at < 0·05 from the gas-free condition.

  3. †Carbon dioxide was calculated from ethanol concentration determined.

  4. ‡Significantly different at P < 0·05 from oxygen condition.

  5. §Significantly different at P < 0·05 from helium condition.

  6. ** ND, not determined.

  7. ††Significantly different at P < 0·05 from hydrogen condition.

1 –S/So 0·78 0·90 0·88 0·36
Product yield (% carbon product formed/sugar consumed)
Ethanol57·423 (3·811)52·434 (4·396)49·210* (3·594)40·630* (2·039)
Carbon dioxide†28·712 (1·905)26·217 (2·198)24·610 (1·797)20·315 (1·019)
Biomass2·287‡ (0·526)1·539‡ (0·129)1·821‡ (0·088)5·767 (0·105)
Glycerol3·191§ (0·093)4·034 (0·165)4·209 (0·361)2·848§ (0·286)
Acetate0·288‡ (0·087)0·221‡ (0·013)0·230‡ (0·054)2·509 (0·807)
Succinate0·332‡ (0·029)0·453‡ (0·034)0·467‡ (0·063)0·745 (0·106)
Malate0·133 (0·023)0·077 (0·010)0·079 (0·010)ND**
Acetaldehyde0·0039†† (0·0018)0·0060 (0·0014)0·0036†† (0·0004)0·0035†† (0·000)
Pyruvate0·085‡ (0·014)0·049‡ (0·006)0·059‡ (0·009)0·288 (0·117)
Acetoin 0‡ 0‡ 0‡1·784 (0·075)
2,3-butanediol0·009‡ (0·008)0·012‡ (0·002)0·024‡ (0·003)0·657 (0·010)
2-oxoglutarate0·002 (0·000)0·040 (0·010)0·043 (0·009)ND**
Carbon balance92·865 (5·325)85·283 (6·666)80·977* (5·353)75·546* (3·908)

The yields of ethanol were significantly lower in helium and oxygen cultures – more oxidizing conditions than that in the gas-free and hydrogen cultures – more reducing conditions (Table 2). The ethanol yield in the hydrogen culture decreased by 5% compared with the gas-free culture and this decrease was 8% in and 16% in helium and oxygen cultures, respectively.

The yield of glycerol was not significantly different between the oxygen and gas-free cultures (2·8% against 3·2%; Table 2). However, these yields increased significantly when yeast strains are cultivated with hydrogen or helium, with values of 4% and 4·2%, respectively.

The yields of other by-products were much lower than the yields of ethanol, glycerol and biomass. Acetate, succinate, pyruvate, acetoin and 2,3-butanediol levels were significantly higher in the oxygen condition compared with the others. It was increased 10-fold for acetate, 5-fold for pyruvate and 2-fold for succinate. The yield of 2,3-butanediol was 0·66% in the oxygen culture and was negligible in the others. Acetoin was reduced totally into 2,3-butanediol in gas-free, hydrogen and helium cultures and partly in the oxygen culture. Malate and 2-oxoglutarate were not determined in the oxygen culture. Moreover, among the three other conditions of culture, acetate, pyruvate and malate yields were slightly higher in the gas-free culture than in the others. Contrasted results were observed with succinate, 2,3-butanediol and 2-oxoglutarate for which the yields were slightly higher in hydrogen and helium cultures. The yield of acetaldehyde was significantly different in hydrogen culture – high reducing condition and the others – from low reducing to high oxidizing conditions.

As observed for yields of ethanol, a global influence of gases was shown on the carbon balance. It was significantly different between the groups of helium and oxygen cultures – more oxidizing conditions and the group of gas-free and H2 cultures – more reducing conditions, as well as observations on ethanol yield. A loss of carbon was observed through the carbon balance estimated at 85·3%, 81% and 75·5% under hydrogen, helium and oxygen conditions respectively, against 92·9% in the gas-free culture. The effects of gaseous environments on the production of some principal metabolites were studied deeply by measuring ethanol and glycerol throughout the cultures (Fig. 4).

Under hydrogen environment, ethanol production increased continuously to attain 60 g l−1 at the end of the culture (Fig. 5a). In gas-free and helium cultures, ethanol concentrations were similar until mid-culture (fermentation progress of 0·5). Then, in the gas-free culture, ethanol production pursued to attain a level similar to that observed in the hydrogen culture at the same fermentation progress (0·78). Unlike hydrogen culture, the production of ethanol in helium culture increased progressively until 0·75 of fermentation progressed and then, was stationary at 48 g l−1 until the end of the culture (fermentation progress of 0·88). Ethanol production on oxygen culture increased slowly during the first five days of culture and remained constant at 11·5 g l−1 until the thirteenth day. When glycolysis was not supplied with glucose, ethanol formation was not completed. This fact explains the lower ethanol concentration found in the oxygen culture. During fermentative metabolism, in addition to the production of ethanol and carbon dioxide, the yeasts produced glycerol that is considered as the third important metabolite. We observed an influence of gases on the concentration of glycerol in the medium. In hydrogen and helium cultures, the concentrations of glycerol evolved differently (Fig. 5b). They reached 4·5–5 g l−1 whereas it was only 3 g l−1 in the gas-free culture. On the contrary, the concentration of glycerol in the oxygen culture was lower, stopping at c. 1 g l−1. This observation can be linked to a decrease of cellular metabolism caused by the lowering of sugar consumption. Moreover, this result agrees with previous works done in continuous cultures of S. cerevisiae reporting decreases in ethanol and glycerol production when oxygen concentration increased (Kuriyama and Kobayashi 1993; Franzén 2003; Frick and Wittmann 2005).

Figure 5.

 Changes in ethanol (a) and glycerol (b) production during the cultures of Saccharomyces cerevisiae BRAS291 under different gaseous environments: hydrogen (□); helium (◊); oxygen (bsl00084); gas-free (bsl00001). Data points represent the mean from three independent experiments. Error bars omitted for reasons of clarity. Maximum observed relative standard deviation was 8%.

Variation of redox balance under different gaseous environments

Intracellular changes in redox balance can influence strongly the metabolism of micro-organisms (van Dijken and Scheffers 1986). In the literature, the redox state of cells was determined by measuring the amounts and ratio of the interconvertible, reduced and oxidized forms of various redox couples, such as GSH/GSSG and NADH/ NAD+ (van Dijken and Scheffers 1986; Penninckx 2000, 2002; Schafer and Buettner 2001). Therefore, to find an explanation for the differences observed on by-products, the intracellular redox balance and Eh were estimated by measuring the intracellular levels of cofactors (NADH/ NAD+) and glutathione (GSH/GSSG) at different times of the culture. The results are presented in Table 3.

Table 3.   Effect of gaseous environments on intracellular redox balance in Saccharomyces cerevisiae BRAS291 yeast
Conditions of culture/ culture durationIntracellular level
NADH (mmol l−1)NAD+ (mmol l−1)Eh by NADH/ NAD+ (mV)GSH (mmol l−1)GSSG (mmol l−1)Eh by 2GSH/ GSSG (mV)
  1. Data represent the mean and values in brackets represent standard deviations from three independent experiments.

No gas
 1 day1·45 (1·26)0·77 (0·53)–353 (1·7)2·44 (1·11)0·005 (0·000)–231 (12·3)
 7 days0·68 (0·57)0·86 (0·23)–314 (9·1)1·24 (0·35)0·003 (0·000)–225 (5·2)
 13 days0·68 (0·07)0·85 (0·14)–317 (1·2)1·04 (0·33)0·005 (0·001)–211 (12·5)
 1 day1·16 (0·06)0·41 (0·02)–333 (1·1)2·72 (0·99)0·004 (0·000)–238 (9·7)
 7 days3·62 (1·59)1·73 (0·16)–326 (0·3)0·66 (0·35)0·001 (0·000)–217 (10·9)
 13 days6·71 (0·12)2·76 (0·60)–330 (0·2)0·30 (0·09)0·008 (0·001)–174 (3·4)
 1 day1·13 (0·17)0·31 (0·05)–337 (1·6)3·62 (1·18)0·009 (0·000)–234 (8·6)
 7 days2·10 (0·31)1·51 (0·52)–326 (4·0)0·60 (0·23)0·003 (0·000)–210 (7·5)
 13 days1·54 (0·23)1·83 (0·13)–318 (1·3)0·57 (0·15)0·013 (0·002)–185 (4·7)
 1 day6·63 (0·55)0·55 (0·16)–356 (7·4)1·80 (0·78)0·002 (0·000)–233 (11·6)
 7 days7·42 (0·84)0·69 (0·09)–350 (0·1)0·57 (0·07)0·023 (0·016)–203 (9·0)
 13 days8·36 (2·07)1·25 (0·21)–347 (2·5)0·18 (0·14)0·022 (0·014)–145 (11·8)

Cultivation of yeast under a gaseous environment caused a substantial modification of the intracellular level of NAD(H), H+. On gas, the level of NAD+ increased continuously from about 0·4 to 1·2–2·8 mmol l−1, depending on the gas used. For gas-free culture, this level was stable at around 0·8 mmol l−1. Major differences were observed on intracellular concentrations of NADH, especially from cells cultivated with hydrogen and oxygen. For the last one, they reached a similar level around 7–8 mmol l−1. It is about 5–10-fold more than the concentrations observed in the gas-free and helium cultures. The difference between oxygen and hydrogen conditions concerned mainly changes in the nucleotide concentration. Over time, it increased dramatically under hydrogen condition while it remained constant at a high level under oxygen condition. Such effect on cofactor concentrations led to differences in values of Eh when it was estimated from NAD(H), H+ ratio. Under gas-free and helium conditions, a significant increase of the intracellular Eh was observed (35 and 20 mV, respectively) while it stabilized throughout the culture under hydrogen and oxygen conditions (about –330 and –350 mV, respectively). However, measurement of Eh from NAD(H), H+ ratio might be subject to controversy.

The glutathione is considered as a better indicator of intracellular redox level. It is the most important and abundant redox buffer present in the cells (Perrone et al. 2005) and makes the greatest contribution to the redox environment of the cell (Schafer and Buettner 2001). For this reason we measured intracellular concentration of GSH/GSSG during the culture. On one hand, in all cases, the intracellular level of GSH decreased with the age of the culture and this tendency was observed particularly when cells were cultivated with gas. After 13 days of culture, it decreased to about one-sixth to one-ninth of the previous level in cells cultivated under gaseous environments, but the decrease was only one half for the gas-free culture. Globally, these concentrations of GSH were kept in the range of values reported in the literature, i.e. 1–10 mmol l−1 (Powis et al. 1995). On the other hand, the amount of GSSG increased throughout the culture (particularly under oxygen conditions, in which the increase was about 10 times, compared with about two times under hydrogen and helium). These changes certainly led to modifications of intracellular Eh as observed in other studies (Schafer and Buettner 2001). In all conditions, Eh increased with the age of the culture indicating an oxidation of the intracellular redox environment. Eh increased from about –230 to –210 mV in the gas-free culture and from about –235 to –185 mV in the helium culture. In the hydrogen condition, it was an increase of about 65 mV and in the oxygen environment about it was about 75 mV. These values are different from those calculated on the basis of NAD(H), H+ ratio.


Applying gases in our work allowed creating different levels of redox environment compared with the gas-free culture. As predicable, a constant value of redox during the entire culture was observed under hydrogen and oxygen and a strong decrease in redox value at the beginning of the culture, followed by stabilization until the end of culture was observed under helium and in the gas-free condition.

Our Eh7 values created by hydrogen and oxygen gases were similar to those described in the literature, i.e. between –300 and –400 mV under hydrogen atmosphere (George et al. 1998; Riondet et al. 1999, 2000; Laurinavichene et al. 2001; Ouvry et al. 2002; Bourel et al. 2003) and between +300 and +400 mV under oxygen atmosphere depending on the oxygen concentration (George and Peck 1998; George et al. 1998). The decrease of redox potential during microbial growth was principally described in aerobic, facultative anaerobic and anaerobic bacteria (Jacob 1970). On yeast, there were only a few observations on redox potential changes in anaerobic and aerobic cultures (Hewitt 1936). However, our results are similar to the observations made by Roustan and Sablayrolles (2003) during the fermentation of S. cerevisiae K1 ICV-INRA in oenological condition in gas-free and ferricyanide-added cultures. In both cases, the redox potential fell continuously during the growth phase and a plateau (between –220 and –250 mV in gas-free culture and between –100 and –150 mV in ferricyanide-added culture) was then reached which was stable until the end of cultures. Recently, Husson et al. (2006) observed also a decrease of redox potential only during the first 12 h of the gas-free and ferricyanide-added cultures in the yeast Y. lipolytica. The mechanism of the change of redox potential during yeast growth was not explained. However, the decrease of redox potential might have resulted from the consumption of dissolved oxygen by yeast and also the synthesis and release in the medium of reducing compounds as sulfites (Hansen and Kielland-Brandt 1996a,b; Park and Bakalinsky 2000). Redox stabilization in the control and helium cultures could result from the fact that oxidized and reduced forms are present in the cultures in about equal amounts. The increase in redox potential observed at the end of these cultures can be because of the lowering of metabolism resulting from the beginning of cell lyses (Jacob 1970).

Influence of gaseous environment and Eh on yeast growth was reported previously in the literature. However, Thom and Marquis (1984) showed that hydrogen and helium gases used for compression of culture were not growth inhibitory for S. cerevisiae. Influence of gas and/or Eh on yeast growth could be dependent on yeast strain and/or medium culture. Recently, in oenological medium, the reducing condition obtained by using hydrogen gas was shown to be favourable to the growth of S. cerevisiae (Cachon et al. 2002), increasing the biomass as well as the specific growth rate.

With regard to yeast cell size, results found in the gas-free condition were close to those described in other studies on the cultivation of brewing yeast under classical condition of fermentation, i.e. between 100 and 900 μm3 (Murray and Cahill 2000; Barker and Smart 2005). The size of microbial cells has been known to vary with cultural conditions. Effect of the condition of aeration on cell morphology of Candida utilis during cultivation has been described previously (Straskrabova et al. 1980). Previous studies showed the dependency of S. cerevisiae cell size on growth rates. In chemostat culture, the size at initiation of budding was proportional to the growth rate for values comprised between 0·33 and 0·23 h−1. However, at growth rates lower than 0·23 h−1, cells displayed a minimum cell size at bud initiation independent of growth rate. In batch condition, reduced growth rates led also to a decrease in cell size (Johnston et al. 1979).

In response to osmotic shifts, cell volume variations of S. cerevisiae have been shown to depend on cell physiological state (de Maranon et al. 1996). Variations in the volume of microbial cells depended also on agitation rates of chemostat cultures. Mean cell volume of S. cerevisiae was found to be directly proportional to the agitation rate in the fermentor (Wase and Patel 1985). With the same agitation rate applied to the fermentor, introduction of gases in the medium led to an increase in the agitation level and thus influenced cell size variation. Influence of physical stresses (thermal stress, osmotic stress) on morphological characteristics of the yeasts S. cerevisiae and Saccharomyces pombe was investigated (Adya et al. 2006). Their mean cell volume and their viability decreased in conjunction with an increase of thermal or osmotic shocks. Interestingly, influence of hyperbaric stress generated by gas on morphology of the yeast S. cerevisiae was also studied. The effect of pressure on cell activity depends strongly on the nature of gas used for pressurization. While nitrogen and air, used at a maximum of 0·6 MPa of pressure, were innocuous to yeast, oxygen and carbon dioxide pressure caused cell inactivation, which was confirmed by the reduction of bud cell size within time. Similarly, the reduction in the cell size was explained by an inhibition of growth in combination with damage on cellular membrane (Coelho et al. 2004). The modifications on cell size could be linked to changes in the ultrastructure of cells as observed by transmission electron microscope (data not shown). Although these observations have been only made at an early stage of the culture, they strongly showed the effect of gas on yeast morphology and it seems likely that these effects will be preserved over the entire culture.

This study showed that modifications on cell growth and size were accompanied by changes in sugar assimilation, carbon and redox metabolisms. Previous works on sugar consumption in yeast focused almost on glucose metabolism and maltose transport mechanism (van Dijken et al. 1993; Lagunas 1993; Weusthuis et al. 1994a; Ostergaard et al. 2000; Brondijk et al. 2001). In yeast, monosaccharides like glucose and fructose are transported inside the cell by facilitated diffusion mechanism driven by a sugar concentration gradient. In contrast, the fermentation of maltose by S. cerevisiae requires first the action of a maltose permease to transport maltose inside the cell and second a maltase to hydrolyse maltose to glucose which is fermentable by the yeast. Moreover, maltose transport system is a proton symport system which requires metabolic energy (Lagunas 1993). In our work, compared with the gas-free culture, hydrogen and helium conditions were favourable to maltotriose and maltose consumption whereas the consumption of these polysaccharides by the cells diminished significantly under oxygen. This observation could be explained by an inhibitory effect of oxygen on maltose permease and/or maltase. The effect of external oxidative state on maltose transport and consumption is poorly documented. It was principally studied through the influence of oxygen limitations (Weusthuis et al. 1994b). The authors showed that the different oxygen rates (0–100 ml min−1) did not influence maltose metabolism in S. cerevisiae CBS8066. However, in the same study, at all oxygen rates tested in C. utilis CBS 621 cultures, alcoholic fermentation did not occur and the amount of maltose metabolized was dependent on the oxygen supply.

Moreover, it is well known that Eh can affect cell energetic system (ΔpH, proton motive force) (Riondet et al. 1999; Bagramyan et al. 2000). We can suppose here that a change in external redox state was not favourable to energetic of the maltose transport system, especially under oxygen. In addition, we have observed that this phenomenon was reversible. When the flux of oxygen was stopped after 2 days of culture, sugar consumption started again and the fermentation pursued. When oxygen was added again after 5 days of culture, the fermentation stopped again (data not shown).

The increase in sugar consumption as well as the fermentation progress for cells cultivated on helium or hydrogen is more difficult to explain. This phenomenon can be linked to the cell size which was 50% of that observed for cell cultivated in gas-free conditions. As a considerable increase of exchange surface between the internal and external environments of the cell might result from these changes in cell size, the cell metabolism would be enhanced as well.

In parallel with the modification of sugar consumption, we observed a change in the intracellular carbon metabolism. Globally, it was shown a modification of carbon balance recovery on gas. This observation indicates a metabolic shift in carbon metabolism. Even the differences observed were not highly significantly different; modifications were due principally to a decrease of ethanol yield which led also to a decrease of carbon dioxide yield calculated from ethanol production. Moreover, the tendency of a modification of carbon flux inside the cells was confirmed by other by-products analysis which showed significant differences between the gas-free condition on the one hand and helium and hydrogen conditions on the other hand. In particular, the glycerol yields were higher for the gas. In yeast, glycerol is the second major by-product of glucose-fermenting cells of S. cerevisiae and the cytosolic redox balance is restored mainly by its formation (Rigoulet et al. 2004). In the current study, glycerol yield increased in spite of a decrease in the biomass yield. Similar observations were made on S. cerevisiae cultivated at different aeration conditions and the author attributed the phenomenon to an increase of cellular content of RNA and proteins. In hydrogen and helium cultures, succinate, malate and 2-oxoglutarate were also increased significantly suggesting that flux through the oxidative branch of tri-carboxylic acid had been modified.

Under oxygen, the low formation of ethanol was due mainly to a metabolic shift towards the formation of oxidized products, i.e. acetate, succinate, pyruvate, acetoin and 2,3-butanediol. A similar phenomenon was observed on fermentative and respiro-fermentative growth during chemostat cultures of S. cerevisiae ATCC 32167 (Frick and Wittmann 2005). In particular, the authors showed that acetate concentrations during respiro-fermentative growth were higher than those measured during fermentative growth. The higher yields of acetate and 2,3-butanediol that we observed in oxygen culture were also similar to the observations made by Franzén (2003). The author showed that, in aerobic condition, the yields of acetate and 2,3-butanediol increased at most six and two times, respectively compared with the yields measured in anaerobic condition. In our study, this tendency was more pronounced.

These modifications in the different by-product yields led to a decrease in carbon recovery for all the cultures performed with gas. Similarly, Cachon et al. (2002) showed also a loss in carbon balance during fermentation of S. cerevisiae under gaseous environments and in oenological conditions. The carbon balance recoveries were significantly lower in cultures conducted under hydrogen or a mixture of hydrogen and nitrogen than those measured in the gas-free condition. However, in these experiments, the flux of gas added to the culture was high (0·3 vvm). Then, an alcohol-related stripping phenomenon cannot be ignored and could explain simply the obtained results. Recently, during similar studies on bacteria, Zigha et al. (2006) suggested the possibility of a gas-related metabolic shift towards metabolites not yet determined, i.e. extracellular proteins. Similarly, this led to a loss of carbon as observed before. In parallel, Heux et al. (2006a) observed a shift of carbon metabolism in genetic-engineered S. cerevisiae strains overexpressing an NADH oxidase. The authors suggested that the balance of intracellular cofactors such as NADH/NAD+ could be involved in this phenomenon.

The importance of redox balance in the metabolism of sugars by yeast has been discussed in relation to energy metabolism and production of metabolites (van Dijken and Scheffers 1986). In Candida versatilis (halophila), the redox state involves cofactor balances (NADH/NAD+, NADPH/NADP+) and this redox state affected, in turn, the rate of fermentation and the activity of the mannitol metabolic pathway (Silva-Graça et al. 2003). Moreover, tolerance of S. cerevisiae against thermal shock and oxidative stress was shown to be regulated by intracellular redox level estimated via the intracellular glutathione ratio (GSH/GSSG) (Espindola et al. 2003; Ueom et al. 2003). Recently, Feron et al. (2007) showed that β-oxidation of the yeast Spo. ruinenii was influenced unfavourably by a drop in the NAD+/NADH observed in low redox environments.

In S. cerevisiae, the ADH reaction to produce ethanol from acetaldehyde plays a major role in reoxidizing the NADH formed by glycolysis. Globally, alcoholic formation of glucose is itself a redox-neutral process and ethanol formation cannot account for the reoxidation of assimilatory NADH. Saccharomyces cerevisiae solves this redox problem by glycerol production, used as a NADH redox sink by reoxidizing the excess of NADH generated during the formation of biomass and other oxidized compounds (Oura 1977; van Dijken and Scheffers 1986; Ansell et al. 1997; Bakker et al. 2001).

In oxygen culture the higher levels of acetaldehyde, pyruvate and especially acetate found (Table 2) could explain the higher level of NADH observed. In fact, in glucose metabolism, another NADH-producing pathway is the production of acetate. It may occur through the action of cytosolic, NAD+-dependent aldehyde dehydrogenases, which reduced NAD+ to NADH from acetaldehyde to acetic acid. The expression of these enzymes depend on a variety of stress, including osmotic shock, heat shock, glucose exhaustion, oxidative stress and drugs (Navarro-Aviño et al. 1999). In our case, oxygen fermentation progress stopped at 0·36 because the substrates were not consumed. As a consequence, metabolic pathways in general and reactions to reoxidize NADH in particular could not function. This could explain the slight change in NADH and NAD+ level after the seventh day of culture observed before and also the excess level of intracellular NADH under oxygen culture. In addition, a high NADH/NAD+ ratio has been shown to inhibit the activity of pyruvate dehydrogenase (Nelson and Cox 2000; Zhu et al. 2006). In our context, this result could explain the high level of pyruvate and thus the small level of ethanol observed under oxygen condition.

More surprising was the effect of hydrogen on intracellular redox balance. An effect of low oxidoreduction potential, fixed by hydrogen gas, was shown on NADH-producing flux in Bacillus cereus F4430/73 cultivated under anaerobic condition (Zigha et al. 2006). NADH-producing flux through pyruvate dehydrogenase complex increased in a low redox condition, i.e.–148 ÷ –222 mV obtained with hydrogen, compared with a high redox condition, +45 ÷–42 mV obtained with nitrogen gas. Studies on other anaerobic and facultative anaerobic bacteria showed also that hydrogen generated low extracellular Eh and greatly increased intracellular NADH content (Lovitt et al. 1988; Sridhar and Eiteman 2001). As a consequence, the change in intracellular cofactor led to the modification of the intracellular redox state in E. coli (de Graef et al. 1999), lactic bacteria (Garrigues et al. 1997), Clostridium cellulolyticum (Guedon et al. 1999) and Sporidiobolus ruinenii (Feron et al. 2007).

The estimation of intracellular Eh through NAD+/ NADH, H+ or GSSG/GSH ratio demonstrated clearly the better validity of the second method. In fact, NAD(H), H+ concentrations in the cells were changing extremely rapidly because of the involvement of the nucleotides in numerous biochemical reactions. Consequently, the ratio is more an indication of the equilibrium inside a metabolic pathway than an indication of the internal redox state (Bakker et al. 2001). For this reason, we estimated the intracellular redox state by measuring GSH/GSSG concentrations. Globally, the different values calculated here are close to those estimated previously (Hwang et al. 1995). In yeast, 90% of the glutathione is normally present in the reduced form. The oxidized form is formed upon oxidation of GSH and can in turn be reduced to GSH by GR and the presence of NADPH (Carmel-Harel and Storz 2000). Intracellular GSSG found in our yeast was between 0·001 and 0·023 mmol l−1, i.e. <10% of the total glutathione. The different values of intracellular Eh calculated here via GSH/GSSG concentrations are close to estimations previously done (Hwang et al. 1995; Schafer and Buettner 2001; Drakulic et al. 2005). The increase observed in internal Eh during cultures of S. cerevisiae BRAS291 appeared to depend on the biological state of the cells. It became more oxidized throughout the cultures, especially under gaseous atmospheres. This observation corresponds to a decrease in the intracellular level of reduced glutathione. Previous studies showed that changes in the Eh of the GSSG/2GSH couple appeared to be correlated with the biological status of the cell: proliferation Eh ≈ −240 mV; differentiation Eh ≈ −200 mV; or apoptosis Eh ≈ −170 mV (Schafer and Buettner 2001). More recently, Drakulic et al. (2005) showed that intracellular Eh of different strains of S. cerevisiae decreased during culture at a maximum of about 40%, from the exponential phase to the stationary phase. These changes in the redox state of the cells, involving the change in GSH/GSSG concentrations, can be in response to differentiation and enzyme inducers (Kirlin et al. 1999). Recent works have also shown that glutathione was involved in the response of the yeast to nutritional, environmental and oxidative stress (Carmel-Harel and Storz 2000; Penninckx 2000, 2002). That could explain the more pronounced oxidation of redox state of the cells cultivated under gas in comparison with that of the gas-free culture.

In conclusion, our results showed clearly that gas induced drastic changes in the primary energy metabolism of S. cerevisiae but the effects were different considering the gas used. Under oxygen, the metabolism was principally oriented to the production of compounds related to oxidative reactions. This phenomenon was linked to a modification of intracellular redox balance with an increase of the intracellular Eh estimated. One surprising result concerned the inhibitory effect of oxygen on maltose utilization. Up to our knowledge, this last effect was not described in the past. We suggested here that it could be linked to a modification of the energetic of maltose transport.

Hydrogen and helium lowered ethanol yield and increased glycerol yield. However, they increased the fermentation progress and thus the overall production of ethanol and glycerol. They lowered the amount of biomass and thus they might lower the ATP yield. These metabolic changes were not Eh-dependant. They seemed more linked to the use of gas, may be via a stripping of some volatile compounds (e.g. carbon dioxide or sulfite, considered frequently as important metabolites in the regulation of yeast metabolism). The redox balance and internal Eh were modified differently by the gas. Under helium, they evolved similarly to what was observed during gas-free fermentation whereas a reducing environment seemed to lower intracellular Eh.

The physiological mechanisms regulating these metabolic changes need further elucidation. In particular, attention must be focused on the maltose transport system that seems to be one of the main targets of a modification of external Eh. This is currently under investigation in our laboratory. The changes in cell size in relation to ultrastructural modifications that seems to be affected by gas must be also studied.

More practically, our study shows the possibility to modulate ethanol and glycerol production by the use of gas. This is of interest for the alcohol beverage industries where different strategies are currently developed to lower the level of alcohol. Another promising issue could be found in technologies on ethanol production where one of the major bottlenecks that limit industrial fermentation processes is the industrial performances in terms of ethanol production and concentration. The use of gas, particularly helium, can constitute a simple, original and flexible means to reach these objectives.


The authors are thankful to Prof. Sonia Collin, Unité de Brasserie et des Industries Alimentaires, Université Catholique de Louvain, Louvain-la-Neuve, Belgium, for providing the yeast train and useful discussions on obtained results.