Partial Removal of Ethanol during Fermentation to Obtain Reduced-Alcohol Wines

Transcription

1 Partial Removal of Ethanol during Fermentation to Obtain Reduced-Alcohol Wines Evelyne Aguera, 1 Magali Bes, 1 Aline Roy, 1 Carole Camarasa, 2 * and Jean-Marie Sablayrolles 2 Abstract: The demand for wines with lower alcohol content has increased over recent years. There are various techniques for lowering alcohol content, including ethanol extraction from the wine and reduction of sugar content of the musts. The feasibility and potential importance of ethanol extraction techniques during wine fermentation were determined. In 100 L pilot-scale fermentations, 2% ethanol was removed halfway through fermentation, either by distillation under vacuum or by stripping with CO 2. Despite being stressful for the yeasts, none of these treatments had a negative effect on fermentation kinetics. On the contrary, the fermentation rate increased, probably due to the extraction of inhibitory compounds and, consequently, the fermentation was easier to complete. Removing 2% ethanol during the course of fermentation slightly affected the final concentration of compounds derived from yeast redox metabolism (glycerol, acetate) and aromatic molecules. Both treatments lowered the volatile compounds content (by 25% and 45% for fusel alcohol and esters, respectively), but these losses were partly compensated for by synthesis in the second part of fermentation. The final concentrations were thus close to or even higher for glycerol and isobutanol (up to 19% and 32% depending on the treatment, respectively) than those of the control fermentation. In addition, sensory analysis detected no significant difference between wines produced with or without dealcoholization treatment, except for a slight negative effect with stripping. Removing 2% ethanol by stripping or by distillation during the fermentation are promising alternatives for reducing the alcohol content of wine without altering the sensory quality of the product. Key words: wine fermentation, alcohol extraction, ethanol content of wines, stripping, distillation under vacuum, sensory quality There is growing demand worldwide for wines containing lower concentrations of alcohol. However, current winemaking practices favor the production of wines with strong flavors, prepared from fully matured grapes. The juice obtained from such grapes may contain very high sugar concentrations, resulting in wines with excessive alcohol concentration. Many winemakers are seeking methods to slightly decrease the alcohol content of their wine, often by 1 or 2%, without lowering the concentration of other compounds involved with wine quality, especially aromatic compounds. One microbiological strategy to reduce alcohol is to use yeast strains that give low ethanol yields. Although 1 UE0999 Pech Rouge, INRA, F Gruissan, France; 2 UMR1083 Sciences pour l Oenologie, INRA-SupAgro, 2 place Viala, F Montpellier, France. *Corresponding author ( camarasa@supagro.inra.fr; tel: (33) ; fax: (33) ) Acknowledgments: This work was supported by the Agence Nationale de la Recherche through the quality wines reduced alcohol content program (no. ANR-05-PNRA-011). The authors thank M. Mikolajczak for help in carrying out the experiments, S. Caille and A. Samson for organizing the sensory analysis and the treatment of sensory data, and M. Angénieux for the analysis of the volatile compounds in all samples. Manuscript submitted May 2009, revised Sept 2009, accepted Nov Publication costs of this article defrayed in part by page fees. Copyright 2010 by the American Society for Enology and Viticulture. All rights reserved. Saccharomyces cerevisiae has a wide range of phenotypic diversity, the difference in percentage ethanol production between wild strains does not exceed 0.5% (Dequin 2007), so selection is not a feasible option. An alternative would be to use genetic engineering to reduce the ethanol yield of Saccharomyces cerevisiae by diverting sugar metabolism into by-products other than ethanol. For example, yeast strains producing more glycerol and less ethanol have been obtained by overexpressing GPD1, a gene that codes for glycerol 3-phosphate dehydrogenase (GPDH) (Michnick et al. 1997, Nevoigt and Stahl 1996, Nevoigt et al. 2002, Remize et al. 1999). Other strategies include using lactate dehydrogenase in yeast, resulting in the simultaneous conversion of pyruvate into ethanol and lactate (Dequin and Barre 1994, Dequin et al. 1999); using a yeast strain producing glucose oxidase to catalyze the oxidation of glucose to form gluconolactone in the presence of molecular oxygen (Malherbe et al. 2003); and expressing an H 2 O-forming NADH oxidase, leading to a marked decrease in the intracellular NADH pool and consequently to a decrease in ethanol production (Heux et al. 2006). Although promising, none of these strains has been marketed because consumers are hesitant in accepting genetically modified organisms and significant reduction of ethanol yield in yeast strains results in the accumulation of undesirable by-products, which affect the sensory properties of wine. Physical processes are already available for producing wines containing less alcohol. Several strategies have been studied, all based on volatility or diffusion (Moutounet et 53

2 54 Aguera et al. al. 2007). A review article summarized the various techniques based on distillation or evaporation principles (Pickering 2000). The spinning cone column (SCC) is the main evaporation process, used industrially in several countries. It is a gas-liquid contact device consisting of a vertical countercurrent flow system that contains a succession of alternate rotating and stationary metal cones with upper surfaces coated with a thin film of liquid (Makarytchev et al. 2004). Membrane processes are also available, and reverse osmosis is probably the most widely used technique. The wine is filtered under pressure through a nonporous membrane that is permeable to alcohol and water, but not to most of the dissolved components (Bui et al. 1986, Villettaz 1987, Chinaud et al. 1991). Osmotic evaporation by membrane contactor can be used on wine, coupled with reverse osmosis or nanofiltration (Hogan et al. 1998, Diban et al. 2008). Removal of fermentable sugar, using a multistep membrane process, is an alternative method (Rektor et al. 2006). Despite their efficiency, these procedures are not selective enough and can also affect flavor balance through the loss of aromatic compounds. The possibility and advantages of removing ethanol during the fermentation process to allow de novo synthesis of aromatic compounds were analyzed in this study. Few studies have been done in this area. Two different techniques were compared: (1) distillation under vacuum, a thermalbased method, and (2) gas stripping, a potentially powerful but simple technology (Scott and Huxtable 1995) not yet tested in winemaking. The second method is based on the transfer of volatile compounds from the liquid to the gas, driven by the gradient of concentrations between the two phases. The effect of diluting the initial must with dealcoholized wine was also tested. The dilution procedure, despite important technical advantages, had been considered as reference conditions rather than a possible technological solution to decrease ethanol content in wine, since diluting grape juice is prohibited in most countries. The effects of these techniques on fermentation kinetics, yeast metabolism, and wine sensory properties were investigated. Materials and Methods Culture conditions. Yeast strain and musts. The enological yeast strain K1 ICV-INRA (Lallemand, Montreal, Canada), commercially available worldwide, was used in this study. Fermenters were inoculated with 200 mg/l active dry yeast. Most fermentations were run using a natural Chardonnay must from southern France, containing 212 g/l total reducing sugars, 231 mg/l assimilable nitrogen, and a total acidity of 3.35 g/l (H 2 SO 4 equivalent). An additional must for white wine (another Chardonnay) and two musts for red wine (a Syrah and a mixture of Syrah and Carignan) were also tested. The musts were flash-pasteurized and stored under sterile conditions (Aguera and Sablayrolles 2005). Fermentation conditions. Fermentations were run in 100 L stainless-steel tanks (Aguera et al. 2005). The temperature was controlled at 20 C. Oxygen addition. Dissolved oxygen was added during alcoholic fermentation to limit the risk of stuck fermentation (Blateyron et al. 1998). Fermenting must was pumped through a circulation loop at a 2000 L/hr flow rate. Pure oxygen was injected into this loop at a 5 L/hr flow rate through a stainless-steel diffuser (N10A; Air Liquide, Toulouse, France). A transfer of 7 mg/l oxygen was obtained after 6.5 min of pumping and oxygen injection, based on overall volumetric oxygen transfer calculations. The must was oxygenated when the ethanol concentration reached 6% (at the beginning of the stationary phase). Partly dealcoholized wines were aerated just after the removal of 2% ethanol. Monitoring of fermentation kinetics. Fermentation kinetics were monitored online by mass flow meters (Brooks Instruments, Hatfield, PA). There is a direct correlation between CO 2 production, sugar degradation, and alcohol production (El Haloui et al. 1988). Postfermentation treatments. After fermentation, 50 mg/l SO 2 was added. Following tartaric stabilization, the wine was filtered and bottled. Lowering ethanol concentration. Extraction of ethanol during fermentation. The fermenting must was partially dealcoholized by removing 2% ethanol when the concentration reached 6% (beginning of stationary phase). The fermenting must was transferred from the fermentation tank to the dealcoholization device under inert gas to avoid oxygenation. After ethanol removal it was returned to the fermentation tank under the same conditions. Fermentation and ethanol extraction were run in triplicate. Distillation under vacuum. Total distillation of one-third of the fermenting must volume, carried out when the alcohol level was ~6%. The total duration of the treatment was approximately 2 hours (45 min heating). The distillation was at a temperature of 38 to 40 C and an absolute pressure of 60 mbars in a 35-L working volume boiler. The boiler was heated by three 2 kw microwave generators (frequency 2450 Mhz). The 40 theoretical plates distillation column was fitted with a two-stage condenser at temperatures of 4 C and -10 C. Stripping. A total of 2 ± 0.2% ethanol was stripped by CO 2. The gas was pumped off and injected into the must (total volume) using a sparger at a flow rate of 50 m 3 /hr and then passed through a condenser. The solution recovered in the condenser was a mixture of ethanol (~23% v/v) and water. The stripping lasted approximately 6 hours. A volume of water equal to the volume lost during stripping was added to the fermenting must. Addition of dealcoholized wine. A wine obtained from the must to be tested was distilled under vacuum to remove all the ethanol. The test must was diluted with this dealcoholized wine before fermentation to lower the initial concentration of sugar by 34 g/l, therefore reducing the final ethanol concentration in the wine by 2%. Analytical methods. Cells were counted with an electronic Coulter Counter (model Z2, Beckman-Coulter, Margency, France) fitted with a 100 µm aperture probe. A

3 Removing Ethanol during Wine Fermentation 55 histogram of the cell size distribution was also obtained. Cell viability was measured using methylene blue. The concentration of residual reducing sugars after alcoholic fermentation was measured by reacting the reducing sugars with copper in alkaline solution and then titrating the excess copper with iodine (Schoorl 1929). Glycerol, acetate, and succinate concentrations were determined by HPLC (HPLC 1100; Agilent Technologies, Santa Clara, CA) on a HPX-87H Aminex column (Bio- Rad, Hercules, CA). Long-chain alcohols and esters in the headspace were assayed by gas chromatography, using a E DB-wax 30 m/0.53 mm/1 µm column (Agilent Technologies). Ethanol concentration in the samples was standardized at 14%. After vapor stripping, the ethanol concentration was determined by density, the volatile acidity by the bromophenol blue method, SO 2 concentration by iodometry, and total acidity by titrimetry. Sensory analysis. A panel of 17 to 22 experts analyzed the sensory properties of wines. Samples (40 ml) were served in dark wineglasses for triangular tests to avoid any visual bias of the participants evaluations and in clear glasses for descriptive analysis. Triangular tests were used for testing the triplicates run for each experimental procedure. A linear scale was used for descriptive analysis. A common vocabulary was generated during the first session. The panel was then trained for two sessions and finally the wines were scored. Fizz software (Biosystèmes, Couternon, France) was used for balanced and randomized testing and data collection. Results Effects on fermentation kinetics. The effect of the removal of 2% ethanol, either by distillation under vacuum, stripping, or initial dilution of must with dealcoholized wine, on the fermentation kinetics of the wine yeast K1 growing in Chardonnay must was investigated. For the four experimental procedures (including the control), oxygen was added when the ethanol concentration reached 6%. Three separate series of fermentations were run. The kinetic profiles of the three fermentations with ethanol removed by stripping were close (Figure 1). The triplicates for the other conditions were also close and the variability of maximum fermentation rate and fermentation duration was consistently less than 5% for experiments carried out under similar conditions. For ease in reading graphs, only one representative curve for each type of procedure was plotted (Figure 2). Comparison of the fermentation kinetics for the two techniques used to remove the 2% ethanol indicated that fermentation by the wine yeast was not inhibited by the removal of ethanol. On the contrary, stripping, and to a lesser extent distillation under vacuum, increased the rate of CO 2 production in the stationary phase, resulting in a substantial decrease in fermentation time of 90 hr and 60 hr, respectively. Effects on cell population and activity. The addition of oxygen to control fermentations at the beginning of the stationary phase had no effect on the yeast population but led to a substantial increase in the rate of CO 2 production (Figure 2). By contrast, combining this oxygenation with the removal of 2% ethanol, either by distillation under vacuum or by stripping, increased yeast population by ~10% (Figure 3). Cell mortality after stripping was close to 20%, similar to that in the control fermentation. Conversely, the distillation procedure was detrimental to yeast viability because of increased mortality in the boiler (up to 50%). However, only one-third of the total volume was distilled under vacuum, so the overall mortality was ~30%. Consistent with these observations, the cell size distribution of yeasts was changed by the removal of 2% ethanol using both methods, while oxygenation alone did not affect this repartition (Figure 4). Figure 1 Change in the rate of CO 2 production during fermentation with ethanol removal by stripping. The commercial wine yeast strain K1 was grown in triplicate on Chardonnay must. When 6% ethanol concentration was reached, 2% ethanol was removed by stripping and oxygen (7 mg/l) was added to the medium. Figure 2 Effect of ethanol removal by stripping or distillation on the change in fermentation rate. Rate of CO 2 production by the K1 strain on Chardonnay must was measured during fermentation. Ethanol concentration was lowered by stripping, distillation under vacuum, dilution of the must with dealcoholized wine, or no dilution (control). For the four experimental procedures, 7 mg/l oxygen was added when 6% ethanol was reached.

4 56 Aguera et al. The greatest proportion of small size cells, mainly observed after distillation, can be explained by the increase in yeast population and by the decrease in cell viability, as dead cells are smaller. The specific CO 2 production rate, which is proportional to the average fermentation activity of viable cells, has been calculated to analyze the effect of the different treatments on fermentation rate. The addition of oxygen increased the specific fermentation rate of the control by 16 ± 1%. The effect of oxygenation on yeast activity was greater when combined with distillation or stripping procedures; the specific CO 2 production rates increased by 29 ± 8% and 28 ± 4%, respectively. Because the rate of fermentation by viable yeast was increased to the same extent by distillation and stripping, the difference in fermentation rate between the two treatments (Figure 2) was probably due to a difference in the number of viable cells after dealcoholization. Furthermore, the change in specific CO 2 production rate was plotted versus the ethanol concentration of the fermenting must, after alcohol removal and oxygenation (Figure 5). The specific rate of CO 2 production by cells from fermentations with stripping or distillation was markedly higher than that of yeasts from control or diluted fermentations, for the same concentration of ethanol. This finding clearly indicates that the activation of the cells after the dealcoholization treatment was not solely due to the decrease in ethanol concentration. Testing a different must under white winemaking conditions and two musts under red winemaking conditions confirmed the positive effect of removing 2% ethanol by stripping on fermentation kinetics. In all cases, the specific CO 2 production rates were increased by ~40%, while oxygenation alone had a much lower effect (Table 1). Effects on by-product synthesis. The effect of the two dealcoholization techniques on yeast metabolism were investigated by comparing, in the experiments described above, the concentrations of six fermentation by-products measured before and after dealcoholization and at the end of fermentation (Table 2). These compounds were selected because they are either marker molecules of redox me- Figure 3 Effect of dealcoholization treatments (stripping, distillation, or initial dilution of the must) on the growth of K1 strain during fermentation on Chardonnay must. Oxygenation (7 mg/l) and ethanol removal by distillation or stripping were performed when 6% ethanol concentration was reached. Yeast population was measured before and 24 hr after oxygen addition. All values are means ± standard deviation from three independent experiments. Figure 4 Effect on cell size distribution of oxygenation with or without 2% ethanol removal during fermentation. Cell size distribution profile of the K1 strain was determined before and 24 hr after the addition of 7 mg/l oxygen to fermentations performed under four conditions: removal of 2% ethanol by stripping (A), distillation (B), initial dilution of the must with dealcoholized wine (C), and control (D).

5 Removing Ethanol during Wine Fermentation 57 tabolism (glycerol and acetate) or aromatic compounds, which may affect the sensory properties of wine (fusel alcohols [isobutanol and isoamyl alcohol], acetate esters [isoamyl acetate], and ethyl esters [ethyl hexanoate]). Most of these compounds were produced before treatment, in the first half of fermentation, except acetate, where 50% of total production was in the stationary phase. During dealcoholization, a slight increase in the glycerol production was observed, probably because of an increase in biomass (Figure 3). Both stripping and distillation treatments, however, lowered the concentration of volatile compounds such as fusel alcohols and esters by 25 and 45%, respectively. For distillation treatments, measurements were made after reassembling the fermenting must, suggesting the boiler had a more drastic effect. Unexpectedly, stripping did not affect the concentration of isobutanol, but had a marked effect on acetate, which was totally removed by the treatment. After distillation or stripping, the rate of resynthesis in the second part of fermentation was much greater, although the final concentrations were generally lower than those in control conditions. Isobutanol was an exception, with a much higher final concentration after stripping. Effects on wine quality. The consequences of stripping or distillation during wine fermentation on the sensory properties of wines were assessed. Triangular tests were Figure 5 Specific fermentation activity of strain K1 after oxygenation with or without ethanol removal during fermentation. The rate of CO 2 production was plotted against the total amount of released CO 2, which is proportional to the production of ethanol. Fermentations by K1 strain on Chardonnay must were performed under four conditions: control, initial dilution of the must with dealcoholized wine, stripping, and distillation. Table 1 Ratio between the specific rates of CO 2 production determined before and after oxygenation with or without stripping. Must Fermentation conditions Oxygenation Oxygenation with stripping Syrah+Carignan 1.15 ± ± 0.08 Syrah 0.99 ± ± 0.08 Chardonnay 1.17 ± ± 0.05 used to check the reproducibility between replicates and to discriminate between wines produced by fermentation with oxygenation with or without dilution, stripping, or distillation treatment. No significant difference was found between the wines, except between the wine obtained after stripping and its control (Table 3). Despite these small differences, 12 wines obtained using the four different treatments performed in triplicate were analyzed through descriptive analysis. The panel generated 16 different descriptors, the most discriminatory being color intensity, alcohol content, and bitterness and amylic, fruity, floral, empyreumatic, and animal aromas, as shown on the two main axes of principal component analysis (Figure 6). Most of the triplicates were close to each other, Table 2 Effect on fermentation products of oxygenation and 2% alcohol removal during fermentation by stripping, distillation or initial dilution of the must. All values are means ± standard deviation of three independent experiments. Fermentation time Compound a / Fermentation condition b 90 hr 117 hr End Ratio c Glycerol (g/l) Control 4.8 ± ± Distillation 4.7 ± ± ± Stripping 4.8 ± ± ± Dilution 5.3 ± ± Acetate (mg/l) Control 82 ± ± 10 - Distillation 58 ± ± ± Stripping 81 ± ± 21 - Dilution 59 ± ± 2 - Isobutanol (mg/l) Control 21.7 ± ± Distillation 21. ± ± ± Stripping 21.4 ± ± ± Dilution 20.3 ± ± Isoamyl alcohol (mg/l) Control 149 ± ± 4 34 Distillation 137 ± ± ± Stripping 141 ± ± ± Dilution 174 ± ± Isoamyl acetate (mg/l) Control 6.8 ± ± Distillation 6.4 ± ± ± Stripping 6.6 ± ± ± Dilution 7.3 ± ± Ethyl hexanoate (mg/l) Control 0.65 ± ± Distillation 0.69 ± ± ± Stripping 0.65 ± ± ± Dilution 0.69 ± ± a Concentrations of compounds were determined before (90 hr) and 24 hr after treatment (117 hr) and at the end of fermentation (wine). b Oxygenation with or without dealcoholization treatment performed at 6% ethanol concentration during K1 fermentation on Chardonnay must. c The ratio between concentrations at the end of fermentation and after the treatment (oxygenation and dealcoholization) represents the change in concentrations in the second part of fermentation.

6 58 Aguera et al. confirming the reproducibility of the experiments. Axis 1 was positively correlated to fruity and floral descriptors and negatively correlated to animal and empyreumatic descriptors. Despite some heterogeneity in the results, most of the conditions could be separated and samples associated with stripping were on the left side of axis 1, that is, with negative descriptors. Discussion Feasibility of partial removal of ethanol during fermentation. Distillation under vacuum and stripping by CO 2 were used to remove 2% ethanol halfway through fermentation at a pilot scale (100 L). These physical treatments, which involve heating or use of high gas flows, could increase mortality in yeast population and stress the yeast. Thus, a negative effect on fermentation rate of these procedures, even when combined with the addition of 7 mg/l oxygen, might be expected. However, fermentation rates increased after alcohol removal and the fermentation durations after distillation and stripping treatments were, therefore, similar to and shorter than those of diluted must, respectively. The removal of 2% ethanol during fermentation appears to efficiently reduce fermentation time by up to 90 hours and limit the risk of stuck fermentation, representing a technological advantage. Impact on yeast activity and fermentation kinetics. The increase in fermentation rate during the stationary phase caused by removal of 2% ethanol involved both a 10% increase in yeast population and an increase in specific yeast fermentation activity. Adding ethanol reportedly causes a cell-cycle delay associated with a transient dispersion of F-actin cytoskeleton, resulting in an increase in cell size (Kubota et al. 2004). Since ethanol concentration decreased by 2% after the stripping and distillation treatments, the changes in the number of cells and in the cell size distribution pattern observed under these conditions are consistent with an enhanced ability of yeast to undergo new cell cycles. Progressive addition of ethanol during fermentation showed that the rate of sugar assimilation was exclusively controlled by the ethanol concentration in the must (Ansanay-Galeote et al. 2001). This finding agrees with the detrimental effect of alcohol on fermentation activity, as previously described (Casey et al. 1984). Conversely, the increase in rate of fermentation observed after alcohol removal procedures cannot be explained by the lowering of ethanol concentration alone: for the same alcohol concentration, the specific CO 2 production rates measured after distillation or stripping were higher than those observed in the control fermentation. During wine fermentation, yeast metabolism produces a wide variety of compounds, including medium-chain fatty acids. These compounds, present at low concentrations (from 0.5 to 10 mg/l), are very important to wine quality. However, it has been reported that these medium-chain fatty acids inhibit the growth and fermentation activity of S. cerevisiae (Lafon-Lafourcade et al. 1984), since they alter the spatial organization of the plasma membrane (Sá-Correia 1989, Viegas et al. 1989). Thus, the increase in fermentation rate is almost certainly due to the extraction of these volatile inhibitory compounds from the fermenting must during the ethanol removal procedures. Indeed, yeast activity is not affected by distillation treatment and is higher, for the same ethanol concentration, in a dealcoholized must than in must produced under control Table 3 Sensory analysis of the wines from the four experimental conditions by triangular tests of difference. Test Type 1 risk (α) Type II risk (β) Control/dilution Control/distillation Control/stripping < a a Dilution/distillation Dilution/stripping Distillation/stripping a Values in bold represent tests showing a significant difference. Figure 6 Descriptive analysis of wines obtained by fermentation with oxygenation with or without stripping or distillation during fermentation; or initial dilution of the must. Descriptors (A) and wines (B) are represented on the two main axes of a principal component analysis.

7 Removing Ethanol during Wine Fermentation 59 conditions (Barbirato, Camarasa, Sablayrolles, unpublished data, 2006). Impact on metabolism and wine quality. The techniques used for removing 2% ethanol are not specific and also lowered the concentration of other volatile compounds, such as fusel alcohols and esters, by 25 to 45%. These compounds are important contributors to the sensory properties of wine and their loss could be detrimental to final wine quality. However, most of these losses were partly compensated for by a de novo synthesis after dealcoholization treatments. In particular, the synthesis of higher alcohols in the second part of fermentation was more than twice that of the control. Ethyl hexanoate, however, was not resynthesized after stripping, likely because of the elimination of its volatile precursor, hexanoic acid. In the distillation procedure, only one-third of the volume was distilled and residual hexanoic acid in the final mixture allowed the production of ethyl hexanoate in the second part of fermentation. Acetate was partially or completely removed by distillation or stripping, respectively. However, final concentrations of acetate in fermentations with 2% ethanol extraction were only 15% lower than those of the control, because large quantities of acetate are synthesized in the stationary phase. Acetate is indeed synthesized throughout the fermentation process (Saint-Prix et al. 2004, Michnick 1997) for acetyl- CoA and NADPH requirements for yeast growth and maintenance (Flikweert et al. 1996, van den Berg et al. 1995). Osmotic pressure is one factor that controls production of acetate in winemaking (Pigeau and Inglis 2005, Caridi et al. 1999). Indeed, using diluted must, which reduces osmotic stress, acetate production was 40% less than that of the control. During fermentation, glycerol synthesis is mainly produced during the growth phase and serves as a redox valve removing excess reducing power generated in biomass formation (van Dijken and Scheffers 1986). Glycerol concentration was not lowered by distillation or stripping because of its very low volatility. The production of glycerol, even limited, in the second part of fermentation was higher after dealcoholization, an indirect consequence of the additional growth that occurred. Partial dealcoholization has only a minor effect on final concentrations of the main fermentation by-products, including aromatic compounds, so great differences in the sensory characteristics of wines obtained using these treatments were not expected. Only one pair of wines (control and produced using stripping) was considered significantly different by our panel: descriptive analysis was consistent with the partial correlation between stripping and negative descriptors. This was confirmed by tests with another must (data not shown). Distillation treatment, however, had no detrimental effect on wine quality and could provide a solution to limiting the alcohol content of the wines and facilitating the fermentation of musts with very high sugar content. Unfortunately, specialized equipment is required which is not usually available in wineries and, like any treatment during fermentation, timing has to be precise, which represents a constraint. Conclusion The extraction of approximately 2% ethanol during fermentation represents a feasible new method of lowering alcohol concentration in wine. Treatments based on stripping and distillation had little effect on the wine characteristics because by-products of sensory importance were subsequently resynthesized. An unexpected advantage of these procedures, on a pilot scale, was an increase in fermentation rate and hence a decrease in fermentation time. However, routine application of even these simple techniques can be difficult in the wine industry because they require specialized equipment and must be performed at a specific time in fermentation. Furthermore, in the case of stripping, additional treatment is required to recover the stripped water. Yeast responses to the environmental stresses caused by the use of distillation and stripping during fermentation to remove ethanol were also highlighted. An increase in the sugar consumption rate and modifications in yeast metabolism, including resynthesis of aromatic compounds, resulted from the removal of 2% ethanol by these treatments. However, a global postgenomic approach would be necessary to understand fully the complex molecular mechanisms involved in this response. Literature Cited Aguera, E., and J.M. Sablayrolles Pilot scale vinifications (100 L). III Controlled fermentations. Wine Internet J. 6 ( com/default.asp?scheda=2784). Aguera, E., C. Picou, M. Perez, and J.M. Sablayrolles Pilot scale vinifications (100 L). I The controlled fermentation facilities at the INRA in Pech Rouge. Wine Internet J. 6 ( Default.asp?scheda=2584). Ansanay-Galeote, V., B. Blondin, S. Dequin, and J.M. Sablayrolles Stress effect of ethanol on fermentation kinetics by stationary-phase cells of Saccharomyces cerevisiae. Biotechnol. Lett. 23: Blateyron, L., E. Aguera, C. Dubois, C. Gerland, and J.M. Sablayrolles Control of oxygen additions during alcoholic fermentations. Vitic. Enol. Sci. 53:13l-135. Bui, K., R. Dick, G. Moulin, and P. Galzy A reverse osmosis for the production of low ethanol content wine. Am. J. Enol. Vitic. 37: Caridi, A., P. Crucitti, and D. Ramondino Winemaking of musts at high osmotic strength by thermotolerant yeasts. Biotechnol. Lett. 21: Casey, G., C. Magnus, and W.M. Ingledew High-gravity brewing: Effects of nutrition on yeast composition, fermentative ability, and alcohol production. Appl. Environ. Microbiol. 48: Chinaud, N., P. Broussous, and G. Ferrari Application de l osmose inverse à la désalcoolisation des vins. Int. J. Sci. Vig. Vin 25: Dequin, S Managing excess alcohol in wine: A new challenge for wine yeast. In Proceedings of Les XIX es Entretiens Scientifiques Lallemand. Global Warming: New Oenological Challenges, pp Lallemand, Blagnac, France. Dequin, S., and P. Barre Mixed lactic acid-alcoholic fermentation by Saccharomyces cerevisiae expressing the Lactobacillus casei L (+)-LDH. Nat. Biotechnol. 12:

8 60 Aguera et al. Dequin, S., E. Baptista, and P. Barre Acidification of grape musts by Saccharomyces cerevisiae wine yeast strains genetically engineered to produce lactic acid. Am. J. Enol. Vitic. 50: Diban, N., V. Athès, M. Bes, and I. Souchon Ethanol and aroma compounds transfer study for partial dealcoholization of wine using membrane contactor. J. Memb. Sci. 311: El Haloui, N., D. Picque, and G. Corrieu Alcoholic fermentation in winemaking: On-line measurement of density and carbon dioxide evolution. J. Food Eng. 8: Flikweert, M.T., L. van Der Zanden, W.M. Janssen, H.Y. Steensma, J.P. van Dijken, and J.T. Pronk Pyruvate decarboxylase: An indispensable enzyme for growth of Saccharomyces cerevisiae on glucose. Yeast 12: Heux, S., R. Cachon, and S. Dequin Cofactor engineering in Saccharomyces cerevisiae: Expression of a H 2 O-forming NADH oxidase and impact on redox metabolism. Metab. Eng. 8: Hogan, P.A., R.P. Canning, P.A. Peterson, R.A. Johnson, and A.S. Michaels A new option: Osmotic distillation. Chem. Eng. Prog. 94: Kubota, S., I. Takeo, K. Kume, M. Kanai, A. Shitamukai, M. Mizunuma, T. Miyakawa, H. Shimoi, H. Iefuji, and D. Hirata Effect of ethanol on cell growth of budding yeast: Genes that are important for cell growth in the presence of ethanol. Biosci. Biotechnol. Biochem. 68: Lafon-Lafourcade, S., C. Geneix, and, P. Ribereau-Gayon Inhibition of alcoholic fermentation of grape musts by fatty acids produced by yeasts and their elimination by yeast ghosts. Appl. Environ. Microb. 47: Makarytchev, S.V., T.A. Languish, and D.F. Fletcher Mass transfer analysis of spinning cone columns using CFD. Chem. Eng. Res. Des. 82: Malherbe, D.F., M. du Toit, R. Cordero-Otero, P. van Rensburg, and I.S. Pretorius Expression of the Aspergillus niger glucose oxidase gene in Saccharomyces cerevisiae and its potential applications in wine production. Appl. Microbiol. Biotechnol. 61: Michnick, S., J.L. Roustan, F. Remize, P. Barre, and S. Dequin Modulation of glycerol and ethanol yields during alcoholic fermentation in Saccharomyces cerevisiae strains overexpressed or disrupted for GPD1 encoding glycerol 3-phosphate dehydrogenase. Yeast 13: Moutounet, M., M. Bes, and J.L. Escudier Las tecnologías de elaboración de vinos con bajo nivel de etanol. ACE Rev. Enol. 24:5-10. Nevoigt, E., and U. Stahl Reduced pyruvate decarboxylase and increased glycerol-3-phosphate dehydrogenase [NAD + ] levels enhance glycerol production in Saccharomyces cerevisiae. Yeast 12: Nevoigt, E., R. Pilger, E. Mast-Gerlach, U. Schmidt, S. Freihammer, M. Eschenbrenner, L. Garbe, and U. Stahl Genetic engineering of brewing yeast to reduce the content of ethanol in beer. FEMS Yeast Res. 2: Pickering, G.H Low- and reduced-alcohol wine: A review. J. Wine Res. 11: Pigeau, G.M., and D.L. Inglis Upregulation of ALD3 and GPD1 in Saccharomyces cerevisiae during icewine fermentation. J. Appl. Microbiol. 99: Rektor, A, G. Vatai, and E. Bekassy-Molnar Multi-step membrane processes for the concentration of grape juice. Desalination 191: Remize, F., J.L. Roustan, J.M. Sablayrolles, P. Barre, and S. Dequin Glycerol overproduction by engineered Saccharomyces cerevisiae wine strains leads to substantial changes in by-product formation and to a stimulation of fermentation rate in stationary phase. Appl. Environ. Microbiol. 65: Sá-Correia, I., S.P. Salgueiro, C.A. Viegas, and J.M. Novais Leakage induced by ethanol, octanoic and decanoic acids in Saccharomyces cerevisiae. Yeast 5 S1989A: Saint-Prix, F., L. Bönquist, and S. Dequin Functional analysis of the ALD gene family of Saccharomyces cerevisiae during anaerobic growth on glucose: The NADP + -dependent Ald6p and Ald5p isoforms play a major role in acetate formation. Microbiology 50: Schoorl, N Suiker titratic. Chem. Weekbl. 26:13. Scott, J.A., and S.H. Huxtable Removal of alcohol from beverages. J. Appl. Bacteriol. 79: van den Berg, M.A., and H.Y. Steensma ACS2, a Saccharomyces cerevisiae gene encoding acetyl-coenzyme A synthetase, essential for growth on glucose. Eur. J. Biochem. 231: van Dijken, J.P., and W.A. Scheffers Redox balances in the metabolism of sugars by yeasts. FEMS Microbiol. Rev. 32: Viegas, C.A., M. Fernanda Rosa, I. Sá-Correia, and J. Novais Inhibition of yeast growth by octanoic and decanoic acids produced during ethanolic fermentation. Appl. Environ. Microbiol. 55: Villettaz, J.C A new method for the production of low alcohol wines and better balanced wines. In Proceedings of Sixth Australian Wine Industry Technical Conference. T. Lee (ed.), pp Australian Industrial Publishers, Adelaide.