Yeast stress and death caused by the synergistic effect of ethanol and SO2 during the second fermentation of sparkling wines

Problems can often arise at the beginning of the second fermentation (prise de mousse) of sparkling wines, such as no start, a long lag period or slow fermentation. These problems are generally associated with yeast stress when inoculated in a base wine with high ethanol content and low pH. However, few studies focus on sulphites, which are often added to base wines to prevent malolactic fermentation, microbiological instability, and wine oxidation. This study aimed to evaluate the joint effect of ethanol and sulfur dioxide on yeasts during the second fermentation. For this purpose, yeasts (Saccharomyces cerevisiae EC1118) were subjected to ethanol, sulfur dioxide and ethanol/sulfur dioxide at the beginning of fermentation, and their vitality and viability, as well as the accumulation of intracellular reactive oxygen species and intracellular pH, were evaluated by flow cytometry. Furthermore, the expression of genes involved in sulfur transport and metabolism was determined. The results showed high mortality, ROS accumulation and intracellular pH reduction in fermentations with both ethanol and sulfur dioxide. The negative effect of ethanol, sulfur dioxide and ethanol/sulfur dioxide on yeasts was found to be dose-dependent and high in those commonly found in some base wines. Cells treated with ethanol/sulfur dioxide showed over-expression of genes involved in sulphite transport (SUL1 and SUL2), efflux pump (SSU1 and FZF1) and metabolism of sulfur amino acids (MET14). Altogether, our data indicate that ethanol and sulfur dioxide have a synergistic effect on yeasts, which may be the root cause of the problems encountered at the beginning of the second fermentation of sparkling wines, and should thus be seriously taken into consideration by winemakers.


Introduction
Sparkling wines develop over two consecutive fermentations: the first one converts grape must into base wine, and the second favours CO 2 incorporation (Di Gianvito et al., 2019). The base wines are characterised by an alcohol content of 9.5 % -11.5 % (v/v) and relatively higher acidity (compared to still wines), with a pH index of 2.8 -3.3 (Togores, 2018). The second alcoholic fermentation or "prise de mousse" (literally "foam creation"), is carried out in closed vessels (hermetic tanks or bottles) causing the incorporation of carbon dioxide into the liquid (Carrascosa et al., 2011). This second fermentation can be performed using either the traditional method (fermentation in the bottle) or the Charmat method (fermentation in a tank). For this purpose, together with the base wine, a mix (liqueur de tirage) is added with a specific amount of sugars per litre of wine (usually sucrose cane or beet sugar at 20 -25 g/L) and yeast starter culture for the fermentation (Di Gianvito et al., 2019).
The chemical composition of the base wine and the conditions of the second fermentation usually stressing factors for the yeast inoculum: for example, high ethanol concentration, presence of glycerol and low pH (Borrull et al., 2015), presence of sulfur dioxide (SO 2 ) (Sudraud et al., 1985), presence of acetic acid (Giannattasio et al., 2013), relatively low fermentation temperatures and concentration of nutrients (Kemp et al., 2020;Martí-Raga et al., 2015), and presence of endogenous CO 2 (Porras-Agüera et al., 2020). Therefore, the adaptation of the yeasts to the base wine is important for the success of the second fermentation (Benucci et al., 2016;Borrull et al., 2016;Martí-Raga et al., 2015). This adaptation procedure is known as pied-de-cuve (in the French language) and consists of two important phases: adaptation to the alcoholic medium and active growing phase (Benucci et al., 2016). This process can last for several hours or days, and usually with a gradually increasing concentration of ethanol and sugar (Benucci et al., 2016;Borrull et al., 2016;Martí-Raga et al., 2016). Traditional yeast adaptation and inoculation protocols for the second fermentation recommend a final inoculum of approximately 1.5 x 10 6 cells/mL of base wine (Ribéreau-Gayon et al., 2006). According to these authors, it is possible for levels well below this to cause sluggish fermentations and leave unfermented sugars. Conversely, higher levels (above 2 OENO One -Template for authors Latest update: April 2021, 28 th 2 x 10 6 cells/mL) speed up fermentations; however, depending on the strain, they can result in excessive yeast taste in the final product due to yeast autolysis.
Wine is a hydrolytic solution in which ethanol is the second most abundant compound (Waterhouse et al., 2016) and an inhibitor for several microorganisms. In the conditions of the second fermentation of sparkling wines, ethanol is described as the main environmental factor to have an influence on yeast transcriptional responses (Penacho et al., 2012). Among several factors tested in a study by, ethanol in the base wine was considered to be the main stress factor for yeasts in the second fermentation of sparkling wines. Tolerance to ethanol varies widely among yeast species , and within the Saccharomyces cerevisiae species, tolerance may vary depending on the strain (Borrull et al., 2015). The presence of ethanol can cause structural changes in yeasts cells, which can impact the fluidity of the plasmatic membrane (Navarro-Tapia et al., 2018), and consequently cell morphology (Dinh et al., 2008).
Another important compound pointed out as being a stress factor in base wines is SO 2 , which is used as a preservative in these wines until the second alcoholic fermentation. Sulfur dioxide is added to grape musts and wines to reduce the medium (antioxidant activity) and to inhibit undesirable microorganisms (antimicrobial action) (Blouin and Peynaud, 2006;Gould and Russell, 2003). Although the use of SO 2 for the conservation of wines is an old practice (Gould and Russell, 2003), ingestion of sulphites through the consumption of food and drink can cause some related adverse clinical effects (Vally et al., 2009) ; for this reason, there is a worldwide movement towards decreasing the concentration of sulphites in wines. Although several studies have shown alternatives for SO 2 and sought to reduce its use (Capece et al., 2020;Christofi et al., 2021;Marchante et al., 2019;Shih et al., 2020;Simonin et al., 2020;Zara and Nardi, 2021), to date, no other physical technique or chemical additive can provide the efficacy and broad spectrum of action of this compound (Lisanti et al., 2019).
The microbial inhibition by SO 2 in yeasts has been attributed to several cellular changes, such as the modification of membrane transport activity by binding to membrane proteins (Divol et al., 2012); the inhibition of glyceraldehyde-3-phosphate dehydrogenase (GAPDH) -a critical enzyme in the glycolysis pathway (Hinze and Holzer, 1986) -and other enzymes like ATPase, alcohol dehydrogenase and NAD-glutamate dehydrogenase (Maier et al., 1986), causing a decrease of the ATP content in cells (Hinze and Holzer, 1986;Maier et al., 1986;Schimz and Holzer, 1979); the modification of the expression of many genes correlated with cell metabolism (Park and Hwang, 2008); the degradation of available thiamine (Labuschagne and Divol, 2021); the binding of 14 metabolites (acetaldehyde, pyruvate, glucose, dihydroxyacetone-phosphate, oxaloacetic acid, and αketoglutaric acid), thereby preventing their further use as substrates for metabolic pathways (Rankine and Pocock, 1969). However, wine yeasts, particularly S. cerevisiae, have a certain tolerance to SO 2 (García-Ríos and Guillamón, 2019) and this tolerance varies between strains (Nadai et al., 2016). SO 2 tolerance is a desired trait in wine yeasts which has been unconsciously selected in wine-making practices over time (Zimmer et al., 2014), and can be considered an evolutionary advantage .
To reduce the toxicity of sulfur dioxide, yeasts use several mechanisms, such as increasing the production of acetaldehyde to bond with SO 2 and thus reducing the free fraction (Cheraiti et al., 2010;Park and Hwang, 2008); activating the Ssu1p sulphite pump encoded with the SSU1 gene (Marullo et al., 2020;Zara and Nardi, 2021); activating sulfur amino acid biosynthesis (Divol et al., 2012) ; and/or modifying the overall metabolic and cell cycle that lead to a "viable but nonculturable" cell behaviour ( Divol and Lonvaud-Funel, 2005;Salma et al., 2013).
The effect of ethanol and sulfur dioxide on the individual antimicrobial action and resistance to stress of wine yeast of the Saccharomyces cerevisiae species has been studied in the past decades; however, relatively few studies discuss the joint effect of these compounds on yeast cells (Chandra et al., 2015). Long lag phases and lazy fermentations during the second fermentation of sparkling wines are often reported in the industry. The results of a study on the preparation of starter cultures (pied de cuve) for sparkling wine production indicate that differences in fermentative kinetics may be caused by differences in the content of ethanol and SO 2 in base wines (Benucci et al., 2016). As reported in spoiling wine yeasts (Chandra et al., 2014;Edwards and Oswald, 2018), the presence of ethanol and SO 2 can modify yeast behaviour and viability.
The second fermentation of sparkling wines is a delicate step in the process, which can impact the final product. Predicting the progress of this step and any problems that may occur can be crucial for obtaining a quality sparkling wine. Many empirical reports express the concern of winemakers regarding this initial stage of the process, as well as their lack of full and sound understanding of what causes viability reduction and the increase in lag phase. For these reasons, we decided to explore the topic more deeply by determining whether the presence of both ethanol and SO 2 molecules causes synergism to the extent of modifying homeostasis and physiological stress responses in yeasts during inoculation and at the beginning of the second alcoholic fermentation of sparkling wines.

Materials and methods
The Saccharomyces cerevisiae yeast strain Lalvin EC-1118® (Lallemand, Canada) was used in all assays. This yeast strain is recommended and conventionally used in the second fermentation of sparkling wines. Moreover, its genome is already sequenced (Novo et al., 2009), a factor that facilitated molecular analyses.

Synthetic wine and inoculum preparation
All experiments evaluating the synergism between SO 2 and ethanol were conducted in a synthetic wine (Martí-Raga et al., 2016) with 4 g/L tartaric acid, 0.5 g/L citric acid, 0.5 g/L malic acid, 0.134 g/L sodium acetate, 1.7 g/L YNB (with ammonium sulfate and without amino acids) and 4.0 g/L glycerol. The final pH of the solution was adjusted to 3.1 using sodium hydroxide. Different ethanol and sulfur dioxide concentrations were added to this medium following the experimental design.
For the adaptation of yeast (pied de cuve) a modified protocol was followed (Benucci et al., 2016).
Briefly, a yeast colony (Lalvin EC-1118) was added to the YPD broth (2 % yeast extract, 2 % glucose, 1 % peptone, pH 6.5) and grown and shaken (150 rpm) at 28 °C for 20 hr. Next, for the adaptation of the inoculum, 10 % (v/v) ethanol, 50 g/L of sucrose and 3.5 g/L of dibasic ammonium phosphate were added to the synthetic wine. An equal volume of the base wine was added to the initial culture (1:1 v/v) and kept static at 20 °C for 24 h. This culture was then diluted with the synthetic wine to a ratio of 1:3 and maintained at a temperature of 20 °C for another 24 h. After this, in all treatments, 24 g/L of sucrose (for fermentation) was added to the synthetic wine and inoculated with 1.4 x 10 6 cells/mL of adapted yeasts.

The effect of ethanol and SO 2 during fermentation
The treatments were divided into four groups in triplicate: i) no EtOH or SO 2 (the control), ii) with 20 mg/L SO 2 (from potassium metabisulphite solution), iii) with 10 % v/v EtOH, and iv) with 20 mg/L SO 2 + 10 % (v/v) EtOH. The fermentations were monitored at a controlled temperature of 15 °C for 28 days (Figure 1). In this experiment, the yeast growth and viability were microscopically monitored as described in Section 2.1.

Evaluation of stress markers
The treatments were applied as previously described. In this experiment, the fermentations were also carried out under a controlled temperature of 15 °C for 6 days ( Figure 2). Yeast growth and viability were determined by microscopic methods and colony-forming units (CFU) evaluated on YPD agar (methods described in Section 2.1.). Cell membrane integrity, intracellular ROS and intracellular pH were analysed by flow cytometry as described in Section 2.2.
All the treatments were inoculated with 1.4 x 10 6 cells/mL of adapted yeasts. The fermentation temperature was held at 20 °C for 96 h (Figures 3, 4, 5 and 6; Table 2). The cell membrane integrity/cellular vitality, intracellular ROS, intracellular pH, concentrations of reducing sugars, free SO 2 and acetaldehyde were evaluated as described in Sections 2.2. and 2.3. Gene expression was also quantified (see Section 2.4. for the method) and related to the presence of sulfur dioxide in the fermentation environment at the sample collection points closest to the time of inoculation (after 12 h and 24 h).

Second fermentation on an industrial scale
In the second industrial scale fermentation (traditional method), the base wine comprised a blend (assemblage) of wines from the Chardonnay grape varieties (36 %), Riesling Italic (30 %), and Pinot Noir (34 %) (white vinified). The wine alcohol concentration was 11.2 % v/v, and it had a pH of 3.27, 80 mg/L of total SO 2 and 16.5 mg/L of free SO 2 . Approximately 22 g/L of sucrose (liqueur de tirage) and an adapted inoculum were gradually added to the base wine following a specific company protocol. The percentage of ethanol in the inoculum at the time of inoculation was 13.3 % (v/v) and the total yeast population was 3.5 x 10 7 cells/mL with a viability of approximately 70 %.
The base wine was inoculated with 1.4 x 10 6 cells/mL; that is 4 % (v/v) of the final fermentation 6 26 27 28 155 156 volume. The bottles were kept at 12 °C for 21 days and three bottles were analysed weekly. The cell membrane integrity/cellular vitality and free SO 2 were analysed (methods described in Sections 2.2. and 2.3.).

Yeast growth and viability assays
Growth and viability of yeast cells (exclusion tests) were performed using a light microscope (Olympus IX71) with a x 400 magnification using a Neubauer chamber, and viability was determined by staining with a 0.1 % Trypan Blue solution (Thermo Fisher Scientific, MA, EUA) (McGahon et al., 1995). Moreover, viable and culturable cells were determined by serial dilution, plating and colony counting on YPD agar (1 % yeast extract, 2 % peptone, 2 % glucose and 2 % agar). The plates were incubated at 28 °C for 24 h, the colony units were counted and data expressed as colony-forming units (CFU/mL) (Fugelsang and Edwards, 2007).

Flow cytometer analyses
To carry out the analyses with the flow cytometer, samples were centrifuged to separate the cells (4629 x g for 5 min). Once separated, the yeasts were washed in phosphate buffered saline with a pH of 7.2 (PBS) and stained with specific fluorescent dyes.
Flow cytometry analyses were performed in a FACSCalibur flow cytometer (Becton-Dickinson, CA, USA) equipped with an argon-ion laser emitting at 488 nm. The flow cytometer data of 20,000 cells were acquired using CellQuest Pro software (BD Bioscience) and data analysis was carried out using FlowJo v.10 software (TreeStar, Inc). All samples were incubated for 30 min in the dark before analysis.
The cellular vitality and cell membrane integrity were determined using the LIVE/DEAD™ formation of a fluorescent product, dichlorofluorescein (DCF) (Kalyanaraman et al., 2012) ; the fluorescence intensity of DCF is related to the amount of ROS within the cells. Stock solutions were prepared by dissolving DCFH at 5 mg/ml in DMSO (dimethylsulfoxide). Staining was performed in 500 µl of sample and 5 µg/ml of dye solution.
The intracellular pH was determined using the fluorescent probe pHrodo Green AM Intracellular pH Indicator (Thermo Fisher Scientific). The pHrodo Green is slightly fluorescent at neutral pH and fluorescent in acid conditions. The sample preparation procedure was performed according to the manufacturer's recommendations. The results were expressed in relative fluorescence, because, due to the low pH of the wines, part of the treatments remained outside the ideal range of pH variation detectable by pHrodo Green (pH 9-4).
The sizes of the cells were compared using forward scatter measurement (FSC); the intensity of the FSC is proportional to the diameter of the cell and is mainly due to the diffraction of light around the cell (Leif, 1986).

Physicochemical analyses
These analyses were performed together from samples of the supernatant (samples centrifuged 4629 x g for 5 min) which had been frozen at the time point of each collection. Total reducing sugars (g/L) were evaluated via the hydrolysis of sucrose in an acid medium and colorimetric method using 3,5-dinitrosalicylic acid (DNS) and microplate reader (absorbance 595 nm) (Dos Santos et al., 2017). The acetaldehyde (ethanal) concentration (mg/L) was quantified using the colorimetric method (acetaldehyde reacts with sodium nitroferricyanide and piperidine solution) using a spectrophotometer (570 nm absorbance), and the calibration curve was obtained directly using acetaldehyde in different concentrations, as adapted from the OIV method (OIV, 2009). Free SO 2 (mg/L) was estimated according to the Ripper titrimetric method using iodate (Zoecklein et al., 1999).

Gene expression
RNA extraction was performed according to a protocol specific for Saccharomyces cerevisiae (Shedlovskiy et al., 2017). The extracted RNA was treated with DNase I (Thermo Fisher Scientific) following the manufacturer's protocol. The absence of contaminant genomic DNA in the RNA was checked before cDNA synthesis using RNA as a template for a PCR assay. The RNA was reversed- following the manufacturer's protocol. The primers for qRT-PCR are shown in Table 1. Their sequences were obtained from published studies and by using the Primer designing tool system on the NCBI website (Primer-BLAST). Primers were purchased from Thermo Fisher Scientific. With the values obtained from the expression of each gene in the treatments, ΔCт was calculated by subtracting C T (threshold cycle) of the reference gene (ACT1) from C T of the target gene.
Afterwards, ΔΔCт was calculated [∆C T (target sample) -∆C T (sample reference)], subtracting the ΔCт of the genes from the treatment samples (20 mg/L SO 2 without EtOH, -10 % EtOH without SO 2 , 10 % EtOH + 20 mg/L SO 2 ) by the control sample (without EtOH and without SO 2 ). Then the formula for the comparative method was applied (2 -ΔΔCт ) (Livak and Schmittgen, 2001), giving a value for the relative expression of the genes or RQ.

Statistical analysis
Statistical analyses were performed by two-way ANOVA and Tukey's multiple comparisons test, using a level of significance of 95 %. Graphs and statistical analyses were performed using the Graphpad Prism® software (GraphPad, CA, USA).

The effect of ethanol and SO 2 during fermentation and evaluation of stress markers
In the first experiment, it was possible to verify the impact caused by SO 2 and ethanol on yeast viability and cell growth. As can be observed in Figure 1, yeast cells grown on the control medium showed high viability throughout the experiment and exhibited a typical growth behaviour, attaining the stationary phase with approximately 5 x 10 7 cells/mL after 8 days. On the medium supplemented with SO 2 (20 mg/L), the yeast population underwent a small (approximately 20 %) reduction in viability in the first few days ( Figure 1A), which was maintained until the end of fermentation. This initial reduction in yeast viability caused a delay in population growth, but it reached almost the same final cell density as the control ( Figure 1B). However, in the medium with EtOH (10 %), cell viability remained high throughout the fermentation, but yeast growth was drastically reduced ( Figure 1A and 1B).
Conversely, in the medium containing both EtOH and SO 2 (10 % EtOH + 20 mg/L SO 2 ), a drastic reduction in yeast viability during the first few days was observed, reaching 2 % on the sixth day  Figure 1A). After this point, the cell population started to grow slowly ( Figure 1B), remaining relatively low compared to the control. At the end of the experiment (28 days) the yeast population in EtOH + SO 2 medium was just 1/6 of the control and SO 2 treatments, and 1/2 of the EtOH treatment ( Figure 1B). These data indicate that EtOH and SO 2 have a synergistic effect on yeast viability and growth. The cell viability results obtained using trypan blue ( Figure 2C) and PI ( Figure 2D) showed high viability (> 95 %) in the control-and ethanol-containing medium. Yeasts cultivated in the medium supplemented with SO 2 exhibited an initial reduction in viability (after 24 and 48 h), followed by an increase in viability associated with population growth. Conversely, yeasts cultivated in EtOH + SO 2 medium showed a drastic and rapid reduction in viability.
The cytometric analysis of intracellular ROS using the DCFH-DA dye showed a basal low ROS concentration for yeast cells grown in the control medium and in the media that contained just ethanol or SO 2 . However, a peak in intracellular ROS occurred on the first day of the treatment supplemented with both ethanol and SO 2 ( Figure 2E), suggesting that together these compounds cause greater oxidative stress and may have also influenced the drastic reduction in cell viability.
The analysis of intracellular pH and treatments with SO 2 and with EtOH + SO 2 revealed considerable changes in the fluorescence of the dye (the higher the fluorescence, the lower the pH).
However, the EtOH + SO 2 treatment exhibited the highest fluorescence (

The interaction of different concentrations of ethanol and sulfur dioxide concentrations in yeasts
The control, 10 mg/L SO 2 , 10 % EtOH and 10 % EtOH +10 mg/L SO 2 treatments showed almost the same behaviour for all variables, except for sugar consumption ( Figure 3B Interestingly, the yeasts in the more severe treatment (12 % EtOH + 20 mg/L SO 2 ) did not accumulate intracellular ROS -a fact that may be related to the high mortality (> 95 %) observed in this treatment -indicating that ROS accumulation depends on metabolic activity.
Regarding the intracellular pH, data in Figure 3D show that yeast cells cultivated in the highest concentrations of EtOH and SO 2 (10 % EtOH + 20 mg/L and 12 % EtOH + 20 mg/L SO 2 showed greater fluorescence (lower pH intracellular) than the control and the other treatments. This decrease seems to be correlated with the percentage of cells with a damaged membrane, which may The comparison of cell size using direct dispersion measurement (FSC) by flow cytometry analysis ( Figure 5) showed that yeast cells in the control and the 10 and 20 mg/L SO 2 treatments exhibited a "normal" distribution, while yeast cells subjected to 10 % EtOH showed a small percentage of larger cells ( Figure 5A). The treatments that cause a more marked reduction in yeast viability (10 % EtOH + 20 mg/L SO 2 , 12 % EtOH + 20 mg/L SO 2 ), showed a high percentage of cells with reduced size ( Figure 5D), thus indicating modifications having been made to cell permeability and water content or a possible modification to the metabolism of the membrane lipids (phospholipids). One of the most important compounds involved in sulfur dioxide neutralisation is acetaldehyde, which is produced during alcoholic fermentation and combines with sulfur dioxide, thus reducing the amount of free SO 2 . The results of the analyses of free SO 2 and acetaldehyde (Table 2) show that only the treatments with considerable fermentative activity ( Figure 3B) showed a large decrease in free SO 2 ; meanwhile, the treatment with 12 % EtOH + 20 mg/L SO 2 , which exhibited a dramatic inactivation of yeast, showed the lowest acetaldehyde production and maintained high levels of free SO 2 throughout the fermentation process. The treatments with SO 2 showed an increase in acetaldehyde concentration (mainly within 12 h) in comparison to the control, as did those with ethanol alone. The treatments with both molecules (10 % EtOH + 10 mg/L SO 2 and 5-10 % EtOH + 20 mg/L SO 2 ) showed higher concentrations of acetaldehyde than the other treatments. Yeast responds to a stress factor (in this case SO 2 and ethanol) with a differential expression of genes that contribute to cell viability and homeostasis. As shown in previous figures, in mid-or non-lethal concentrations of ethanol and sulfur dioxide, a yeast population can adapt and grow, albeit slowly. This "adaptation" involves the differential expression of genes, the products of which contribute to yeast cell homeostasis, metabolism and eventual division. In this context, we evaluated the expression of several genes associated with sulfur transport and metabolism. As can be observed in Figure 6A, the SSU1, MET4, and FZF1 genes were hyper-expressed in yeasts grown in 10 % EtOH + 20 mg/L SO 2 compared with the treatments that contained just ethanol or SO 2 . However, the SUL1 gene showed higher expression in both 10 % EtOH and 10 % EtOH + 20 mg/L SO 2 treatments, while SUL2 exhibited higher expression in the treatments that contained SO 2 . The COM2 gene expression did not vary among treatments. After 24 h ( Figure 6B), the yeast cells subjected to 10 % EtOH and 10 % EtOH + 20 mg/L SO 2 showed higher expression of SSU1 and, in particular, MET14 genes than the control and the SO 2 treatments. However, the high expression of SUL1 and SUL2 observed in the first few hours was not evident after cell adaptation to ethanol and sulfur dioxide stress.

Discussion
Empirical observations in wineries report frequent problems at the start of and during the second fermentation of sparkling wines in both traditional and Charmat processes. In general, the practical solutions adopted in these cases are pre-adaptation and reinoculation with yeasts, which is costly and laborious, especially for traditional sparkling wines. Furthermore, industrial data show that these problems are associated with a considerable reduction in yeast viability. In an attempt to explain the occurrence of these problems, we carried out a series of experiments in synthetic wine to determine the effects of high concentrations of ethanol and the presence of free sulfur dioxide, and the synergism between the two, on the vitality, viability and other parameters of yeasts.
The present study shows that the synergism between ethanol and SO 2 can be considered to be the main stress factor for yeasts at the beginning of the second fermentation ( The data obtained for the synthetic wine that were supplemented with ethanol, sulfur dioxide and ethanol/sulfur dioxide (Figures 1, 2, 3 and 4), and real base wines (Figure 7) showed the synergistic effect of ethanol and sulfur dioxide on the vitality and viability of yeasts. Despite the yeasts having been adapted, the results show that whether inoculated in an environment with an ethanol concentration close to that of base wines or in an industrial base wine, yeast grows slowly but maintains high viability. This behaviour may be due to fluidisation (Huffer et al., 2011;Jones and Greenfield, 1987), the depolarisation of membrane potential by an increased passive proton flux and inhibition of nutrient uptake (Casey and Ingledew, 1986) and a delay in the cell cycle (Kubota et al., 2004). Moreover, cells grown in synthetic wine without ethanol and supplemented with 20 mg/L of SO 2 maintained a normal growth rate and high viability, as has been observed in grape must fermentations (Ferreira et al., 2017). Conversely, yeast inoculated in wine containing both ethanol and sulfur dioxide exhibited a rapid and strong decrease in cell viability, indicating their synergistic effect on the yeast. The low number of cells that remained viable after the initial shock slowly resumed growth, resulting in a long lag phase and slow fermentation.
The results of the experiment on synthetic musts supplemented with different concentrations of sulfur dioxide show that there was a dose-dependent longer lag phase; the yeasts started to multiply when free SO 2 concentration decayed to approximately 5 mg/L on combination with yeasts and fermentation-derived products, such as acetaldehyde (Ochando et al., 2020). in the intracellular concentration of ROS ( Figures 2E and 3C). Experimental data ( Figure 3C) indicates that ROS accumulation in ethanol/sulfur dioxide interaction depends on cell metabolism.
The accumulation of ROS is one of the main determinants of apoptotic cell death (Farrugia and Balzan, 2012). Apoptosis can function as a defence and preservation mechanism of cell populations in the face of stress (viral pathogens, homeostasis change, nutrient insufficiency and other adverse conditions), ensuring that part of the cells survive to propagate their genome (Fröhlich and Madeo, 2000). When comparing both results obtained regarding ROS accumulation ( Figures 2E and 3C It is known that the intracellular pH of yeasts decreases in the presence of sulfur dioxide (Pilkington and Rose, 1988). In an aqueous solution, SO 2 (acid oxide-sulfur dioxide) in its molecular form (SO 2 ) enters cells more easily, because it has no charge, and the molecule rapidly dissociates to form bisulphite (HSO 3 -) and sulphite (SO 3 2-) anions (Divol et al., 2012). SO 2 behaves like a weak acid in aqueous environments (Waterhouse et al., 2016). Moreover, studies of changes to cell membrane caused by ethanol have shown that it increases passive water transport (Madeira et al., 2010), modifies cell membrane fluidity and decreases H + -ATPase activity, which is responsible for maintaining intracellular pH (Aguilera et al., 2006). This modification to membrane structure may facilitate the diffusion of molecular SO 2 into the cytoplasm, which has no charge.  (Divol et al., 2012;Gould and Russell, 2003;Ribéreau-Gayon et al., 2006;Waterhouse et al., 2016). The reduction of intracellular pH (loss of internal buffering capacity) may be the result of a "snowball effect", because the lower the cell's intracellular pH, the higher the percentage of the molecular form of SO 2 that would theoretically remain active within the cell, leading to the disruption of cell homeostasis and thus cell death. This hypothesis is reinforced by ethanol having a dose-dependent effect when in association with sulfur dioxide (Figure 3).
Another interesting fact is the size of yeast cells depends on their environment. In the presence of ethanol, there is a tendency for cell structure to change and for part of the cell population to become larger (it becomes swollen) ( Figure 5); such an effect on the diameter of Saccharomyces cerevisiae cells has been described in other studies (Dinh et al., 2008;Kubota et al., 2004). However, death caused by the presence of SO 2 and ethanol decreases cell size ( Figure 5). Our results indicate that there may be a relationship between cell size and loss of plasma membrane integrity (PI + cells); the percentage of inactivated cells is almost the same as that of cells that become smaller in treatments with 10 % EtOH + SO 2 20 mg/L and 12 % EtOH + SO 2 20 mg/L.
In alcoholic fermentation, the production of acetaldehyde by yeasts plays an important role when in combination with free SO 2 (Rankine and Pocock, 1969), because it decreases the antimicrobial action of SO 2 (Liu and Pilone, 2000). The excess of acetaldehyde produced by the decarboxylation of pyruvate during fermentation is secreted, instead of being used in the production of ethanol or acetic acid (Liu and Pilone, 2000). Increased acetaldehyde has been reported in the second fermentation of sparkling wines (Pozo-Bayón et al., 2003), and more acetaldehyde has been found to form in some yeasts (including the EC1118 strain) in the presence of exogenous SO 2 in must fermentations (Li and Mira de Orduña Heidinger, 2020;Li and Mira de Orduña, 2017). However, in the literature, no correlation between the increase in extracellular acetaldehyde and greater stress during the second fermentation of wines in the presence of EtOH and EtOH + SO 2 has been made, in contrast to the present study (Table 2). Our results point to a greater formation of acetaldehyde when ethanol and SO 2 are present together, which corroborates with the results that show an increase in stress when both molecules are present. Another important factor is the relationship between the formation of acetaldehyde by yeasts and their overcoming stress and consequently leaving the lag phase. This was clearly seen in the treatment with 10 % EtOH + 20 mg/L of SO 2 in which there was a delay in acetaldehyde production, probably due to reduced viability ( Figures 3A   and 4). In the treatment with 12 % EtOH + 20 mg/L of SO 2 practically no acetaldehyde was formed to combine SO 2 and overcome stress, and almost the entire yeast population was inactivated (Figures 3 and 4).
The regulation of gene expression plays an important role in an organism's development and its response to physiological and environmental changes; it can be responsible for the survival of a group of cells and its growth after a period of latency. The results obtained in terms of relative gene expression showed that, compared to the other treatments, there was an increase in the expression of several of the genes involved in sulfur sensing and metabolism in the 10 % EtOH + 20 mg/L SO 2 treatment after 12 and 24 hours ( Figure 6). This agrees with the results of the other analyses that show a synergistic increase in stress in this treatment.
The SUL1  . While the SUL2 was the most expressed in treatments with SO 2 only and SUL1 the most expressed in ethanol only, both these genes were hyper expressed in the EtOH + SO 2 treatment ( Figure 6). Moreover, in the EtOH + SO 2 treatment, FZF1 and SSU1 are more expressed than in the other treatments. The FZF1 gene encodes a plasma membrane protein (Fzf1p) involved in the expression of the SSU1 sulphite efflux pump (Avram et al., 1999;Park and Bakalinsky, 2000). In turn, the SSU1 gene, a marker of adaptive evolutionary advantage that is found in oenological yeast as a result of the use of sulphites in winemaking Zimmer et al., 2014), is positively correlated with the tolerance and detoxification mechanism of sulphite (Marullo et al., 2020;Zara and Nardi, 2021).
The MET14 gene, which encodes the APS kinase enzyme involved in the sulfur assimilation pathway that reduces sulphate to sulphide (Donalies and Stahl, 2002;Noble et al., 2015), was hyper-expressed in the first 12 h of the EtOH + SO 2 treatment. However, after 24 h, gene expression seems to have been more influenced by the presence of ethanol only. As it is a metabolic route gene linked to the biosynthesis of sulfur amino acids, other factors -in addition to the presence of exogenous SO 2 -may influence its greater expression, such as the formation of sulphites by the yeast (Donalies and Stahl, 2002). Surprisingly, no change in expression was found for COM2, a gene whose product controls, directly or indirectly, the expression of more than 80 % of the genes activated by SO 2 (Lage et al., 2019).
In general, the increase in the expression of these genes may indicate a response to a greater amount of intracellular SO 2 caused by the synergistic effect of ethanol/sulfur dioxide, as, theoretically, more proteins related to this stress would need to be synthesised to create an efficient sulphite efflux in order to survive.
In practical terms, the data obtained in this study show that winemakers should be particularly careful when adding sulphites to base wines for the production of sparkling wines; depending on their concentrations, free sulfur dioxide in the presence of ethanol can lead to the death of a large part of yeast population, and consequently to either the non-development of the second fermentation or a slow fermentation with a long latency phase. It is worth remembering that, when added to base wines to avoid microbial instability and oxidation, sulfur dioxide will remain as free SO 2 , since most of the compounds which commonly bind to sulfur dioxide are removed from the lees of the first fermentation and during the stabilisation and filtering processes. Moreover, to attain 11 to 12 % ethanol in sparkling wines, the ethanol concentration of base wines must contain more than 10 % problems in the second fermentation. Therefore, sulphites may literally be responsible for headaches in consumers (Silva et al., 2019), but their high concentrations in the production of sparkling wines can also be the cause of "headaches" for winemakers.

Conclusions
Depending on their concentrations, the presence of both ethanol and SO 2 in the fermentation environment causes synergism and increases stress on yeasts by modifying intracellular homeostasis, deregulating intracellular pH and increasing oxidative stress, thus leading to cell death. Moreover, ethanol/sulfur dioxide causes yeasts to hyper-express genes related to sulphite tolerance in an attempt to overcome their negative effects. A reduction in the vitality and viability of a yeast population can either prevent the second fermentation of sparkling wines from occurring at all or increase its lag phase, thus causing serious technological problems that are difficult to overcome.