Monitoring Stress-Related Genes during the Process of Biomass Propagation of Saccharomyces cerevisiae Strains Used for Wine Making

Abstract

Physiological capabilities and fermentation performance of Saccharomyces cerevisiae strains to be employed during industrial wine fermentations are critical for the quality of the final product. During the process of biomass propagation, yeast cells are dynamically exposed to a mixed and interrelated group of known stresses such as osmotic, oxidative, thermic, and/or starvation. These stressing conditions can dramatically affect the parameters of the fermentation process and the technological abilities of the yeast, e.g., the biomass yield and its fermentative capacity. Although a good knowledge exists of the behavior of S. cerevisiae under laboratory conditions, insufficient knowledge is available about yeast stress responses under the specific media and growth conditions during industrial processes. We performed growth experiments using bench-top fermentors and employed a molecular marker approach (changes in expression levels of five stress-related genes) to investigate how the cells respond to environmental changes during the process of yeast biomass production. The data show that in addition to the general stress response pathway, using the HSP12 gene as a marker, other specific stress response pathways were induced, as indicated by the changes detected in the mRNA levels of two stress-related genes, GPD1 and TRX2. These results suggest that the cells were affected by osmotic and oxidative stresses, demonstrating that these are the major causes of the stress response throughout the process of wine yeast biomass production.

Production of Saccharomyces cerevisiae biomass is an economically important process. Industrial strains of S. cerevisiae are used by many food companies as starters for fermentative processes in the making of bread, wine, beer, and other alcoholic beverages. The technological characteristics of commercially produced yeast determine, to a high extent, the quality and fermentation performance of those processes. Several studies have evaluated the energetic, kinetic, and yield parameters of the yeast biomass production process (6, 8, 28, 29). However, the evaluation of the yeast transient response to various environmental challenges which occurs during stressful conditions in the propagation process and how, ultimately, the cells alleviate or remedy them remains to be addressed.

In the wine and baker's yeast production processes, aerated molasses supplemented with nutrients is inoculated with selected strains. Then cells grow through various transient stages during the batch and fed-batch phases of the process. In a sequence of consecutive fermentations, yeast biomass grown in small fermentors is used to inoculate larger tanks (6, 8, 28, 29). In the initial batch phase, cells are exposed to increased osmotic pressure due to the high concentration of sugars present in the molasses. Also, high aeration and aerobic metabolism can lead to oxidative stress. During the fed-batch phase, the feed rate is set to limit the sugar concentration to maintain respiratory metabolism and increase the biomass yield. In the case of wine yeast, the biomass is concentrated and dehydrated at the end of the process to obtain active dry wine yeasts that can be stored for long periods of time (8, 28, 29). During this process, wine yeast cells endure various stressful situations that induce multiple intracellular changes (1, 24). Subsequently, in a period of several hours during maturation and final processing, yeast cells suffer nutrient limitation and a complex mix of different stresses (thermic and osmotic, etc.) caused by the drying process. As a result, these dynamic environmental injuries affect the biomass yield, the fermentative capacity, the vitality, and the viability of the cells (1, 24).

Eukaryotic cells have developed molecular mechanisms to sense stressful situations, transfer information to the nucleus, and adapt to the new conditions (9, 14, 15). Protective molecules are rapidly synthesized and transcriptional factors are activated, thus changing the transcriptional profile of the cells. Many stress response genes are induced by several adverse conditions through the sequence element STRE (general stress responsive element), which targets the transcriptional factors Msn2p and Msn4p (17, 20). This pathway, also known as the “general stress response pathway,” increases the expression of many different genes, including the well-studied HSP12 and GSY2 genes which are involved in protein folding and glycogen metabolism, respectively (2, 9). Furthermore, yeasts cells have been shown to respond specifically to certain stresses. During thermic stress, the transcriptional factor Hsf1p activates transcription of genes, such as STI1, which code for proteins that counteract protein denaturation and aggregation (18, 30). Aerobic growth and prooxidants generate reactive oxygen species (ROS), leading to oxidative damage on the cells. To neutralize the harmful effects of oxidative stress, proteins are generated that participate in two major functions: antioxidants (such as GSH1, TRX2, CUP1, and CTT1) to reduce proteins and eliminate ROS damage and metabolic enzymes (such as PGM1 and TDH2) that redirect metabolic fluxes to synthesize NADPH, slowing down catabolic pathways like glycolysis (12). Another well-known specific stress response is the high-osmolarity glycerol response pathway (3), which induces genes involved in glycerol synthesis (GPD1, GPP2) and methylglyoxal detoxification (GLO1). Intracellular accumulation of glycerol counteracts hyperosmotic pressure to avoid water loss (14, 15). There also are some stress response pathways still poorly understood, such as those involved in the adaptation to nutrient limitation. Large groups of glucose-repressed genes, nitrogen-regulated genes, and others of unknown function, such as YGP1, are induced after exposing cells to nutrient limitation (14). During the study of the stress response systems, it has been demonstrated that exposure to one kind of stress can activate protective mechanisms against other different stresses, a phenomenon known as cross-protection (7,23, 32, 33).

Previous studies have demonstrated the suitability of marker gene expression as a tool for the study of yeast stress responses in industrial processes (13, 16, 21). In this work, we performed stress experiments and bench-top growth trials of wine yeast biomass production to demonstrate that the induction of specific stress-related genes may enable us to determine the environmental disturbances to which the cells are dynamically exposed. The data indicate that osmotic and oxidative stresses are two of the main adverse conditions that Saccharomyces cerevisiae senses during this process.

MATERIALS AND METHODS

Organisms and cultivation media.

The strains used in this study were S.cerevisiae industrial strain T73 (CECT1894), a natural diploid strain isolated from Alicante (Spain) musts (26), which has been commercialized by Lallemand, Inc. (Montreal, Canada). This strain has been previously used in several studies and has proven to be a good wine yeast model (11, 21, 22, 25). The S. cerevisiae strain IFI87 is a natural wine strain isolated from Montilla (Spain) (supplied by the Instituto de Fermentaciones Industriales). Cultures and precultures were prepared in YPD medium (1% yeast extract, 2% bacteriological peptone, 2% glucose) or SD medium (0.17% yeast nitrogen base, 0.5% ammonium sulfate, 2% glucose) and incubated at 30°C with shaking (250 rpm). Culture growth was monitored by measuring the optical density at 600 nm (OD₆₀₀). YPD precultures were used to inoculate (OD₆₀₀ = 0.1) industrial media. Molasses medium (diluted to 60 g of sucrose liter⁻¹ for batch phase or 100 g of sucrose liter⁻¹ for fed-batch phase) was supplemented with 7.5 ml of (NH₄)₂SO₂ liter⁻¹, 3.5 g of KH₂PO₄ liter⁻¹, 0.75 g of MgSO₄ · 7H₂O liter⁻¹, 10 ml of vitamin solution liter⁻¹, and 1 ml of antifoam 204 (Sigma, St. Louis, Mo.) liter⁻¹. Molasses and mineral solutions were autoclaved separately. The vitamin solution containing 50mg of d-biotin liter⁻¹, 1 g of calcium pantothenate liter⁻¹, and 1 g of thiamine hydrochloride liter⁻¹ was filter sterilized (0.2-μm pore size) prior to use in the molasses medium.

Stress time course experiments.

Cultures were inoculated to an initial optical density of 0.1 (OD₆₀₀ = 0.1) from overnight YPD precultures, harvested after they reached the exponential phase of growth (OD₆₀₀ = 0.5 to 0.7), washed with cold distilled water, and transferred to 30°C prewarmed YPD medium containing 0.5 M KCl or 1 M sorbitol (osmotic stress) or 39°C prewarmed YPD medium (thermic stress). In glucose deprivation experiments, cells were transferred from SD medium (0.17% yeast nitrogen base, 0.5% ammonium sulfate, 2% glucose) to S medium (0.17% yeast nitrogen base, 0.5% ammonium sulfate). In oxidative stress experiments, 2 mM hydrogen peroxide (H₂O₂) or 0.5 mM menadione (final concentrations) were added to the YPD medium after the cells reached the exponential phase of growth. For nutrient limitation stress experiments, SD medium or SDn medium (0.17% yeast nitrogen base, 0.05% ammonium sulfate, 10% glucose) for glucose or nitrogen limitation stresses, respectively, were inoculated (OD = 0.1) from SD medium overnight precultures. Aliquots were taken at several time points for RNA analysis. Experiments were carried out at 39°C for thermic stress and 30°C for other stress experiments.

Apparatus and operation.

Propagation growth experiments were conducted in a 5-liter reactor model BIOFLO III equipped with proportional, integral, and derivative (PID) control units for pH, temperature, oxygen, and agitation speed (NBS, New Jersey). Experiments started with a working volume of 2,000 ml at 30°C. The initial pH was 4.5, and it was allowed to freely vary between 4 and 5 during the batch phase. During the fed-batch phase, the reactor pH was maintained at 4.5 by the automatic addition of 42.5% H₃PO₃ or 1 M NaOH. The reactor was continuously fed with medium by a type 501 peristaltic pump (Watson-Marlow, Falmouth, United Kingdom) at the desired flow rate. Dissolved oxygen, measured with an electrode (Mettler-Toledo), was maintained above 20% by a PID control system that allowed the automatic modification of the agitation speed between the range limits of 300 to 500 rpm. Cell growth was monitored by measuring the OD₆₀₀. Cell dry weight determination was carried out by cell centrifugation at 5,000 × g for 10 min, washing the cells with distilled water, and drying the cells at 80°C until they reached a constant weight.

Determination of glucose, ammonium, and sucrose contents.

Residual glucose was determined using a glucose oxidase-peroxidase assay (Boehringer Mannheim GmbH). Extracellular ammonium was determined with a commercial kit (Boehringer Mannheim GmbH). Samples were centrifuged for 1 min at 13,000 × g, and the glucose and ammonium concentrations were determined in the supernatants. For sucrose determination, diluted samples were incubated for 10min at 30°C in 200 μl of 50 mM sodium acetate (pH 5.0) and 2.5 U of invertase (Sigma). The reactions were stopped by adding 100 μl of 0.4 M K₂HPO₄ and boiling the reaction mixtures for 3 min. Samples were centrifuged for 1 min at 13,000 × g, and the glucose concentration was determined in supernatants with a glucose oxidase-peroxidase assay. In all of the determinations, the final concentrations were obtained by interpolation of the sample results using the slope of the fitting linear regression from the respective standards.

Analysis and quantification of mRNA.

Total RNA from standardized yeast cell pellets (10 mg) obtained during either stress or fed-batch time course experiments was extracted with an automated device for multisample processing (Fast-Prep; Savant), separated by electrophoresis in formaldehyde-containing agarose gels, and analyzed by Northern blotting. Restriction fragments from the STI1 gene (a 1.5-kb KpnI-SacI fragment), from the HSP12 gene (a 0.6-kb EcoRI fragment), from the CTT1 gene (1.4-kb and 1.1-kb EcoRI fragments), and from the GSY2 gene (a 1.7-kb BanII fragment) were used as probes. Specific primers used in PCRs to synthesize all the other probes are shown in Table 1. Probes obtained by restriction digestion or PCR were labeled by random priming using [α-³²P]dCTP. Finally, the quantification of each specific mRNA was carried out by direct measurement of radioactivity on the filters with an Instant Imager FLA-3000 (FujiFilm). The results of gene transcription were normalized to rRNA levels, thus allowing the comparison of the different mRNA levels between successive hybridizations. Normalization was also performed versus probe-specific radioactivity and internal hybridization controls. Results were expressed and presented as the relative mRNA gene levels. All experiments were carried out at least twice, and the results differed by less than 20% on the median coefficient of variation.

TABLE 1.

Genes and primers used for the amplification of DNA probes

Gene	Primer	Sequence (5′-3′)
TRX2	TRX2-1	AAATCCGCTTCTGAATAC
	TRX2-2	CTATACGTTGGAAGCAATAG
GSH1	GSH1-1	CCCGATGAAGTCATTAACA
	GSH1-2	GGAAAAGGTCAAAATGCT
CUP1	CUP1-1	TGTTCAGCGAATTAATTAAC
	CUP1-2	CATTTCCCAGAGCAGCA
GLO1	GLO1-1	GAACACTTCGGTATGAAG
	GLO1-2	GGTATTCCTGACCCTCTC
YGP1	YGP1-1	ACTTTGCCGGCATGGGATG
	YGP1-2	GTGAAGACACCGGAGTAC
GPD1	GPD1-1	TTGAATGCTGGTAGAAAG
	GPD1-2	TGACCGAATCTGATGATC

RESULTS

Specificity study of stress responsive genes.

To identify genes transcribed differentially after cell exposure to stress and validate them for their use as signal molecules along bench-top growth trials, we cultivated S. cerevisiae T73 and performed time course experiments to evaluate gene induction in response to different stress conditions (heat shock, oxidative, osmotic, and nutrient limitation). Subsequently, the transient responses evaluated included those of the following genes: TRX2, STI1, GPD1, CUP1, GLO1, CTT1, GSH1, YGP1, and GSY2. Figure 1 shows the averaged data of normalized mRNA induction levels for STI1 as an example of the results. The normalization of the mRNA induction data allowed us to compare the results of gene transcription between independent experiments. Data normalizations increase the background noise; therefore, to avoid interpretation artifacts, we have considered that inductions below fourfold were not significant.

FIG. 1.

Induction of one selected gene marker, STI1, after individual stress treatments. mRNA levels were determined under the following stress conditions: glucose deprivation in S medium, 1 M sorbitol, 0.5 M KCl, 39°C, 0.5 mM menadione, and 2 mM H₂O₂. Results were normalized with mRNA and with a common gene probe for different hybridizations of the same filter.

Our results indicate that differential transcription of STI1 (Fig. 1) occurred during heat shock stress but not in response to oxidative, osmotic, or starvation stresses. These results are in agreement with previous findings (16). As can be seen in the results summarized in Table 2, TRX2 showed induction in yeast grown on YPD medium containing 2 mM H₂O₂. However, there was no response when cells were challenged with 0.5 mM menadione in YPD, during heat shock at 39°C, under osmotic stress (0.5 mM KCl or 1 M sorbitol), or during glucose depletion (transferring cells from SD minimal medium to S medium). Glucose (SD medium) or nitrogen (SDn medium) exhaustion did not increase TRX2 expression levels. CUP1 appears to be specifically induced in response to 0.5 mM menadione (Table 2). As opposed to TRX2, CUP1 was not significantly induced by the presence of 2 mM H₂O₂. Other stress conditions did not affect CUP1 gene expression in wine yeasts. We did not see induction throughout glucose or nitrogen exhaustion experiments. Data on GPD1 expression in the presence of the same stressors are shown in Table 2. During osmotic stress, GPD1 revealed the strongest response of all genes analyzed in the presence of any stress condition. GPD1 mRNA levels increased 16-fold in response to the presence of 0.5 mM KCl, while an 8-fold induction was seen in response to 1 M sorbitol. A smaller response (∼4-fold) was also observed in response to 2 mM H₂O₂. Other stress conditions did not induce GPD1 expression.

TABLE 2.

Summary of gene expression under laboratory conditions

Stress challenge	Result for genea:
	CUP1	GSH1	CTT1	TRX2	GLO1	GPD1	STI1	YGP1	GSY2
0.5 mM menadione	+	−	+	−	−	−	−	−	−
2.0 mM H2O2	−	−	++	+	−	+	−	−	−
1 M sorbitol	−	−	+	−	++	+	−	−	−
0.5 M KCl	−	−	+++	−	+	+++	−	+	−
39°C	−	−	−	−	+	−	+	+	+
Glucose depletion	−	−	+++	−	+	−	−	+	+
Glucose exhaustion	−	−	++	−	−	−	−	+++	+++
Nitrogen exhaustion	−	−	−	−	−	−	−	+	−

^a−, no induction; +, 4 to 8 times induction; ++, 8 to 12 times induction; +++, ≥12 times induction.

Other genes analyzed in this study were less specifically induced (Table 2). GSY2 expression increased in response to heat shock and glucose exhaustion. YGP1 was induced by glucose depletion and by 0.5 mM KCl osmotic stress. GLO1 increases its mRNA levels in response to osmotic stress, heat shock, and glucose depletion. Finally, gene expression analysis showed that CTT1 is the most pleiotropic stress response gene, responding to every condition studied, except nitrogen exhaustion and high temperatures.

Bench-top trials of yeast biomass production.

The yeast proliferation method used during industrial processes is a combination of two growth configurations: a batch stage followed by a fed-batch stage. Here, we performed bench-top trials of yeast biomass production by scaling down the process and reproducing the environmental conditions (temperature, pH, aeration, growth medium) usually employed during the industrial processes. Under these conditions, two independent experiments were performed for each of the two chosen wine yeast strains (T73 and IFI87). Figure 2 shows the evolution of critical variables in one of the experiments cultivating S. cerevisiae T73. The turbidity, cell dry weight, and parameters that define strain growth were determined during the experiments and are summarized in Table 3.

FIG. 2.

Critical parameters during a typical bench-top scale experiment of Saccharomyces cerevisiae biomass production. (A) Feed rate (—) and pH evolution (▪); (B) percentage of O₂ saturation (▪) and the coupled agitation (□).

TABLE 3.

Kinetic and yield parameters during bench-top trials of S. cerevisiae biomass production

Strain	Result for stagea:
	Batch		Fed batch
	μ	Y	μ	Y
T73	0.37 ± 0.02	0.076 ± 0.016	0.025 ± 4.1 × 10−4	0.18 ± 0.017
IFI87	0.42 ± 5.0 × 10−3	0.093 ± 4.1 × 10−4	0.048 ± 5.0 × 10−3	0.15 ± 7.6 × 10−3

^aμ, maximum specific growth rate in h⁻¹; Y, biomass yield in g (dry weight) of cells g of sucrose⁻¹. Averages ± standard deviations of the results from at least two experiments are shown.

During the batch stage, S. cerevisiae carried out an alcoholic fermentation when grown in an excess carbon source. Immediately afterwards and in the presence of oxygen, cells produced carbon dioxide and water by aerobic respiration from the generated ethanol. These differential stages could be monitored by monitoring the pH, dissolved oxygen consumption, and required agitation speed during the process (Fig. 2). Growth kinetics (not shown) and ethanol accumulation (Fig. 3C and 4C) were used to double-check the achievement of this condition. Overall biomass yield and kinetic parameters of cellgrowth obtained in these bench-top scale experiments (Table 3) were close to the expected theoretical values and those values obtained in the industrial production of wine yeasts (8, 28, 29).

FIG. 3.

Pattern of gene expression during a typical bench-top scale experiment of growth for Saccharomyces cerevisiae T73 industrial strain. The expression of the gene markers HSP12, GPD1, STI1, GSY2, and TRX2 is shown during the batch (A) and fed-batch (B) stages of growth. Sucrose consumption (▪) and ethanol concentration (○) profiles are shown in panel C.

FIG. 4.

Pattern of gene expression during a typical bench-top scale experiment of growth for Saccharomyces cerevisiae IFI87 industrial strain. The expression of the gene markers HSP12, GPD1, STI1, GSY2, and TRX2 is shown during the batch (A) and fed-batch (B) stages of growth. Sucrose consumption (▪) and ethanol concentration (○) profiles are shown in panel C.

During the fed-batch stage, the medium feed rate is a critical parameter. Each biomass factory develops its own feed rate pattern depending on the yeast strain and other factors to achieve the desired biomass yield. Based on data from various sources, we set a feed rate pattern for the wine yeast strains T73 and IFI87 tested in this research. The feed rate pattern used in our experiments (Fig. 2) enabled sucrose aerobic respiration to yield carbon dioxide and water. Thus, limiting the feed of sucrose to the fermentor allowed us to maintain the value of dissolved oxygen above 20% and the agitation at a maximum of 500 rpm. Therefore, we were able to achieve this critical value, avoiding cell shear due to agitation, oxygen limitation, and the fermentative metabolism of sucrose. The data (Table 3) do not show significant variations in the biomass yield and in the kinetic parameters of cell growth. The reproducibility of our growth experiment results allowed us to analyze the specific cellular stress responses in the obtained biomass by applying molecular tools as described below.

Differential transcriptional response of stress-related genes of wine yeast during biomass proliferation.

The transcriptional time course response of five stress-related genes during yeast propagation was evaluated using bench-top fermentation experiments as described above. The TRX2, STI1, and GPD1 genes were used to analyze the yeast stress-specific responses to oxidative, thermic, and osmotic stresses, respectively. Expression of the Msn2p/Msn4p-regulated genes HSP12 and GSY2 were also studied to distinguish between stress-specific changes in gene expression and the induction of the general stress response. The averages of the data from the proliferation experiments are shown in Fig. 3 and 4. These figures display the gene induction profiles during the batch stage (Fig. 3A and 4A) and from the start of the fed-batch stage (Fig. 3B and 4B) for S. cerevisiae T73 (Fig. 3) and S. cerevisiae IFI87 (Fig. 4). Panels C in both figures show sucrose consumption and ethanol concentration profiles for the corresponding strains.

As seen in Fig. 3A, there is a clear response to osmotic stress in the first hour of fermentation in strain T73. The high sugar content in molasses (60 g of sucrose liter⁻¹) provokes an increase in mRNA levels for the osmotic stress marker GPD1. The same results were also observed in strain IFI87 at these early times (data not shown). In Fig. 4A, the first time point in the experiment was taken after 2.5 h, so only the end of this induction peak was observed. After the lag phase, sugar consumption reaches its highest rate (Fig. 3C and 4C). Subsequently, a second stress response was triggered, and yeasts started to express the oxidative stress marker TRX2. Both strains show this oxidative response but at different times. While IFI87 responds during the first part of sucrose consumption (Fig. 4A and C), T73 reacts close to the point of sucrose exhaustion (Fig. 3A and C). Interestingly, as expected from the stress experiments, the osmotic stress marker GPD1, weakly expressed during oxidative stress conditions (Table 2), was also transcribed in parallel to TRX2.

Coinciding with ethanol exhaustion, HSP12 transcription reaches its maximum for both wine strains (at 24 h for IFI87 and 25 h for T73) 4 h before the end of the batch stage (Fig. 3A and 4A). At this point, a new oxidative stress response is observed as well. After sucrose exhaustion, a strain-specific response of HSP12 is observed for strain T73 compared to strain IFI87.

During the fed-batch process, just after sucrose was supplied to the medium, a clear osmotic stress response was observed in both strains, while only strain T73 showed a link to the oxidative stress response (Fig. 3B and 4B). GPD1 and TRX2 were significantly transcribed in strain T73; however, only GPD1 was induced in strain IFI87. HSP12 expression increases continuously in strain IFI87 until the end of the entire process and coincides with GPD1 expression. TRX2 shares this pattern, but it only reaches significant levels near the end of the fed-batch stage. In strain T73, the same genes are induced but with different patterns, e.g., there is an initial induction of GPD1, HSP12, and TRX2 and a slow decrease in the mRNA levels during the fed-batch stage. A significant induction in the general stress response gene GSY2 and the specific heat shock marker STI1 was not observed under our experimental conditions.

DISCUSSION

Living cells respond to environmental challenges with global changes in gene expression. Extensive data on stress-associated genes in Saccharomyces cerevisiae are available in the literature, and a few studies have analyzed and compared the expression of each one of these genes in response to exposure to specific kinds of stresses. A remarkable study by Gasch et al. (10) provides important information concerning the transitory responses of laboratory Saccharomyces strains developed during the very early stages of the growth cycle. However, limited information is available about cell responses to stress during the propagation process of polyploid wine-making strains. Moreover, similar studies have shown important differences in the patterns of gene expression between industrial and laboratory Saccharomyces strains (4, 5, 25, 27).

Our results indicate that four of the nine analyzed genes can be used as specific stress response markers (Table 2). STI1 was selected as a heat shock gene marker because the mRNA levels increased after heat shock, but a significant induction was not found in response to osmotic, oxidative, or nutrient limitation stresses. Ivorra et al. (16) reported increased expression of STI1 near the end of microvinification experiments using S.cerevisiae T73 when a mix of nutrient limitation and ethanol stress occurred during the late stationary phase of growth. An equivalent situation was not present under the conditions evaluated in our research; therefore, STI1 was selected as a heat shock marker. TRX2 is another good gene marker in wine yeasts. Expression of this gene increases only after exposure to oxidative stress in 2 mM H₂O₂. Similarly, CUP1 is significantly induced only in the presence of the superoxide radical generator menadione (0.5 mM) and could also be used as an oxidative marker. However, low expression levels were obtained during our studies, while previous results of other researchers have indicated stress cross-induction (19, 31). Therefore, TRX2 was selected as the best oxidative stress marker. Our results agree with published data regarding differential expression of oxidative response genes depending on the oxidative agent (19). Another useful gene marker is GPD1, which has been shown to be induced in response to osmotic stress in wine yeast (21). In our study, GPD1 mRNA reaches high levels in tested osmotic stress conditions and also responds to 2 mM H₂O₂. However, its expression under oxidative conditions is lower than in the osmotic stress conditions shown in Table 2. Thus, GPD1 was selected as a suitable marker for osmotic stress using an oxidative marker in parallel. Other analyzed genes do not exhibit specific induction and respond to more than two of the stress tested situations. Consequently, they were not considered as adequate markers for monitoring processes where mixed stresses could occur.

Batch and fed-batch processes have been traditionally evaluated from the point of view of energy-yielding metabolism. Here, we have studied the cell response by analyzing the changes in expression levels of stress-related genes during a scaled-down version of the industrial yeast production process. Although other studies have been performed to identify molecular markers in industrial yeast strains (13, 16, 21), this is the first report of application to the analysis of the stress response in bench-top trials of yeast biomass production. The technological relevance of these results greatly depends on the similarity of the bench scale experiments to the real industrial process. Therefore, we have developed a two-stage experiment that fits the industrial parameters of media composition, feed rate, aeration, and pH (6, 8, 28). Growth and biomass yield were close to those theoretically expected during batch industrial propagation processes (8). Interestingly, during the mainly respiratory fed-batch phase, the results show a lower yield than that described for baker's yeast (6, 8, 28). This may indicate a strong adaptation of wine yeast strains for wine making, where they perform mainly a fermentative metabolic role.

The expression of stress gene markers was analyzed during the batch and fed-batch stages of the production experiments with two yeast wine strains. Besides the specific gene markers selected in this work, two general stress gene markers (HSP12 and GSY2) were also included. The study of these additional genes allowed us to follow the induction of the general stress response pathway and helped to identify stress-specific responses. Clear differences between strains can be seen in the level and in the timing of gene induction. However, several conclusions can be taken from these studies. The induction of HSP12 shows that yeast sense stress during different stages of yeast production, even under the controlled environmental proliferation conditions. Induction is stronger for HSP12 than for other stress genes in wine yeasts, as has been described previously (16, 21). As expected, STI1 expression analysis shows that heat shock stress does not occur during this well-controlled fermentation process. The inoculation of yeasts into the molasses medium generates an osmotic shock, as is shown by changes in the GPD1 expression. The osmotic stress response of GPD1 to the high sugar content in the medium has been previously described for laboratory strains (14, 15) as well as for wine yeast strains during wine fermentations (21). More interesting is the yeast response to oxidative stress during the initial 20 h of the batch stage. It is well known that respiratory metabolism of sugars produces the generation of ROS. Therefore, during yeast growth and in the presence of a high aeration, respiratory metabolism is facilitated and may increase intracellular oxidation of proteins and lipids. This subsequently leads to cellular damage and may halt the cellular cycle. As a result, an oxidative stress response is required, as was observed by the increased expression of the TRX2 molecular marker in both strains, especially when sucrose consumption is reaching its maximal rate for strain IFI87 and at sucrose exhaustion for strain T73. This situation may also be present during the ethanol consumption phase and during most of the respiratory fed-batch stage. This indicates the relevance of oxidative stress resistance in wine yeast for its endurance along the industrial process of biomass production. The oxidative stress response is more intense for strain T73 than for IFI87. This difference might be related to the adaptation of the commercial strain T73 to industrial growth.

A better understanding of the adaptation to oxidative stress may allow biomass producers to improve cell survival, vitality, and performance by controlling oxidative conditions and also by using selected strains with increased resistance to oxidative stress.