Peculiar H⁺ Homeostasis of Saccharomyces cerevisiae during the Late Stages of Wine Fermentation

Abstract

Intracellular pH (pH_in) is a tightly regulated physiological parameter, which controls cell performance in all living systems. The purpose of this work was to evaluate if and how H⁺ homeostasis is accomplished by an industrial wine strain of Saccharomyces cerevisiae while fermenting real must under the harsh winery conditions prevalent in the late stages of the fermentation process, in particular low pH and high ethanol concentrations and temperature. Cells grown at 15, 25, and 30°C were harvested in exponential and early and late stationary phases. Intracellular pH remained in the range of 6.0 to 6.4, decreasing significantly only by the end of glucose fermentation, in particular at lower temperatures (pH_in 5.2 at 15°C), although the cells remained viable and metabolically active. The cell capability of extruding H⁺ via H⁺-ATPase and of keeping H⁺ out by means of an impermeable membrane were evaluated as potential mechanisms of H⁺ homeostasis. At 30°C, H⁺ efflux was higher in all stages. The most striking observation was that cells in late stationary phase became almost impermeable to H⁺. Even when these cells were challenged with high ethanol concentrations (up to 20%) added in the assay, their permeability to H⁺ remained very low, being almost undetectable at 15°C. Comparatively, ethanol significantly increased the H⁺ permeability of cells in exponential phase. Understanding the molecular and physiological events underlying yeast H⁺ homeostasis at late stages of fermentations may contribute to the development of more robust strains suitable to efficiently produce a high-quality wine.

INTRODUCTION

Winemaking is a human activity dating back to prehistoric times that leads to the conversion of grape juice (must) into wine as the result of a complex network of interactions established along the winemaking process (29). Saccharomyces cerevisiae is the main yeast used in wine fermentation. This makes S. cerevisiae wine strains an attractive target for the improvement of wine technology and quality.

During wine fermentation, yeasts are exposed to an extremely wide gradient of stress conditions. After a short stage of active increase in the population, under osmotic stress, growth slows down, the cells enter stationary phase, and fermentation is carried out by resting cells. As fermentation proceeds, the sugar concentration decreases and the ethanol concentration increases, the pH drops, and temperature rises, enhancing the deleterious effect of the other forms of stress. While transforming must into wine, the physiological state of the yeast suffers important changes; resting cells are physiologically very different from the cells at the first stages of the fermentation process (31). Several authors reported that ethanol affects key features of yeast physiology, impairing the major membrane transport systems assayed under lab conditions (14, 20), reducing metabolic activity (16, 21), decreasing cell growth (4), and inducing apoptotic responses and cell death (12). Although ethanol toxicity in yeasts is a complex mechanism, the main target of such stress seems to be cell membranes (1, 2). Ethanol causes an increase in membrane fluidity and a loss in its structural integrity (7), with an accompanying increase of leakiness (13, 20) and changes in lipid composition (10), generating deleterious nonmediated diffusion processes (30, 33). Some of the above biochemical changes induced in yeast by ethanol stress are similar to those caused by thermal stress (27), and a synergistic effect between ethanol and temperature is well documented (17).

In recent years, attempts have been made to identify specific genes responsible for tolerance to ethanol and to other stress factors commonly found during alcoholic fermentation (19, 37). Numerous genes have been reported to be up- or downregulated, revealing the possible implication of several pathways, such as the regulatory HOG and TOR pathways (8, 31), transcription machinery and RNA processing (28), protein trafficking and vacuolar function and transport (5, 37), peroxisome import protein machinery (34), and lipid membrane and wall biosynthesis (15). However, most of these studies were centered on active growing cells, with less attention being paid to late stages of fermentation.

In all forms of life, cells require a suitable physiological pH so that metabolic reactions can proceed. In yeast, under acidic pH, proton homeostasis is essentially maintained by a balance between the flux of H⁺ pumped out by the PMA1 ATPase and the influx of H⁺ by passive diffusion through the lipid bilayer and by secondary transporters (20). This balance is also supported by the cell buffering capacity (22) and vacuolar transport (34). Ethanol interferes with this delicate equilibrium by inhibiting mediated transport and enhancing passive diffusion (20, 35), disturbing both intracellular pH (pH_in) and proton motive force across the plasma membrane (18). In the past, the ability of yeasts to acidify the environment has been suggested to be an indicator of good metabolic performance and ethanol tolerance (11, 23).

The aim of this work was to analyze parameters involved in H⁺ homeostasis in S. cerevisiae during the process of grape must fermentation, in particular pH_in and H⁺ fluxes across the plasma membrane. We focused on the physiological characterization of an enological yeast strain, comparing exponential and early- and late-stationary-phase cells, with special emphasis on H⁺ homeostasis as a requirement for survival capacity and “health” while fermenting white grape must. The most relevant result was that late-stationary-phase cells became almost impermeable to H⁺ and were able to maintain their metabolic activity and survival capacity, even when they were challenged with high extracellular ethanol concentrations.

MATERIALS AND METHODS

Yeast strain.

An enological Saccharomyces cerevisiae strain (ISA1000), isolated from a commercial active dry yeast starter (FERMIVIN), from the Instituto Superior de Agronomia (ISA) yeast culture collection, was used. Stock cultures were maintained at 4°C on YPD agar (1% yeast extract, 2% peptone, 2% glucose, and 2% agar).

Must and fermentation conditions.

Unless otherwise stated, the yeast strain was grown in white grape must from the ISA vineyards, supplemented with sulfur dioxide (50 ppm), which corresponds to standard enological treatments. Must was centrifuged (Eppendorf 5810R centrifuge) at 4°C, 5 min, 18,000 × g, and sterilized by filtration (0.22-μm Millipore filter, type GSWP). Cells were precultured in the same must at 28°C in orbital shaken flasks for 24 h. Inoculation was standardized at 1 × 10⁶ cells ml⁻¹ in 800 ml of white grape must contained in 1-liter Erlenmeyer flasks with a cotton cap. Fermentations were performed at 15, 25, and 30°C with low magnetic stirring (150 rpm). For comparison with standard lab conditions, cells were also grown up to an optical density at 640 nm (OD₆₄₀) of ≈1 (mid-exponential phase) in mineral medium (20) with 2% glucose at 25°C with low magnetic stirring (150 rpm).

Growth monitoring and sampling.

Growth was monitored by measuring optical densities at 640 nm in an Ultrospec 2100 pro (Amersham Biosciences) spectrophotometer. Viability was determined by counting CFU on YPD medium after 2 to 3 days at 28°C. The total cell number was assessed by counting in a hemocytometer. Samples for further assays were collected at the mid-exponential phase and at early and late stationary phases.

Analytical techniques.

In order to monitor the evolution of the fermentation process, the glucose concentration was roughly estimated with a rapid detection assay using a commercially available dipstick (Diabur-Test 5000; Boehringer, Mannheim, Germany) commonly used for measuring glucose in urine (range, 1 to 50 g liter⁻¹). After removing the cells by centrifugation (12,000 × g, 3 min, 4°C), glucose and fructose, as well as ethanol, concentrations in supernatants were measured by enzymatic assays (10.139.106.035 and 0.176.290, respectively; Boehringer Mannheim/R-Biopharm, Roche), to establish accurate fermentation profiles.

Intracellular pH.

pH_in was estimated according to the method of Rottenberg (32), as described in reference 25, based on the relative distribution of radiolabeled benzoic acid between the cell interior and the extracellular medium, using the classical equation:

$$[H^{+}]in=\frac{1}{\frac{A^{T}in}{A^{T}out}\left(\frac{1}{[H^{+}]out}+\frac{1}{K_d}\right)-\frac{1}{K_d}}$$

where A^T_in and A^T_out are the intracellular and extracellular total concentrations of the acid, respectively, and K_d is the dissociation constant of benzoic acid (6.40 × 10⁻⁵ mol liter⁻¹). The method is based on the assumption that the plasma membrane is impermeable to benzoate but freely permeable to the undissociated form of the acid, which, in the equilibrium, reaches identical concentrations on both sides of the membrane. A solution of [7-¹⁴C]benzoic acid (18.5 MBq, 588.3 MBq mmol⁻¹; NEN) was prepared by adding 1.5 ml of 5 μM NaOH to the commercial vial. From cultures in white grape must (as previously described), samples were collected in mid-exponential and early and late stationary phases at the three temperatures tested. Two identical 50-ml samples of each suspension were transferred to 100 ml Erlenmeyer flasks. [7-¹⁴C]benzoic acid solution was added to one of the flasks (approximately 10 μl, to keep the ratio between the amount of radioactive acid and amount of cells in the sample constant). The concentration of benzoic acid was below 8 μM in all cases, and it was assumed not to disturb the cells. Both flasks were immediately incubated, with magnetic stirring, at 15°C for 150 min or at 25 and 30°C for 90 min. The incubation time was established in preliminary experiments as the time necessary to reach a steady radioactivity counting. Aliquots containing approximately 5 mg dry biomass were filtered through glass fiber filters (Whatman GF/C, 25-mm diameter) and immediately washed with cold water (5 ml); the filter was then immersed in 5 ml of Optiphase Hisafe 2 scintillation cocktail (Perkin Elmer), and the radioactivity was counted in a Beckman LS 6000LL scintillation counter. To calculate the total concentration of benzoic acid inside the cells, a value of internal cell volume of 1.9 μl (mg dry biomass)⁻¹ (36) was used. In order to measure the extracellular concentration of benzoic acid, 500 μl of the cell suspension was centrifuged for 5 min at 16,000 × g, and the radioactivity in 100 μl of the supernatant was also counted in the same system. The flask prepared in parallel without radioactivity was used to measure external pH in a PHM220 standard pH meter (Radiometer Copenhagen) and for determination of dry biomass. Values for pH_in are the means of at least two determinations using cells from at least two independent growth experiments. When indicated, pH_in determinations were performed with cell suspensions prepared in appropriate buffers, following the same methodology.

Dry biomass was determined by filtering the same volume of suspension as used to count intracellular radioactivity through membrane filters (pore size, 0.22 μm); filters were rapidly rinsed with 10 ml of distilled water and desiccated at 80°C for 24 h.

Proton fluxes.

H⁺ movements were evaluated in cells collected throughout the fermentation process. Cells were harvested (12,000 × g, 3 min, 4°C) and washed twice with ice-cold water. Suspensions were prepared in distilled water at 60 to 70 mg of dry biomass ml⁻¹ and kept on ice for at least 1 h. To determine dry biomass, 0.1 ml of cell suspension was desiccated as above in small pieces of preweighed aluminum foil. H⁺ movements were measured by recording the pH of unbuffered cell suspensions in a water-jacketed cell of 2 ml with magnetic stirring, with a standard pH meter (PHM62; Radiometer Copenhagen) connected to a potentiometer recorder (BBC-Goerz Metrawatt SE460) (20). For H⁺ efflux measurements, 0.1 ml of the cell suspension was mixed with 0.8-ml ethanol solutions in water (final concentrations, 0 to 20% [vol/vol]). pH was adjusted to 5. The addition of 20% glucose (0.1 ml) initiated the H⁺ extrusion, observed as the acidification of the unbuffered environment. The maximum rate of increase in the extracellular H⁺ concentration, calibrated with 10 mM NaOH, was taken as a measure of H⁺ extrusion activity (20). For passive H⁺ influx, 0.2 ml of the cell suspension was mixed with 0.8-ml ethanol solutions in water. H⁺ net influx corresponds to the overall movement of protons in both directions: H⁺ entering the cells by passive diffusion and H⁺ being pumped out by the ATPase. To minimize the activity of the ATPase, H⁺ influx assays were performed in the presence of 2-deoxy-d-glucose (1 mM) (to decrease the level of ATP) and antimycin (2 μg ml⁻¹) (to inhibit the respiratory chain) (20). The rate of H⁺ influx was calculated as the steady rate of decrease in the concentration of extracellular H⁺, recorded immediately after adjusting pH to 4 (13) for at least 10 min. Calibrations were performed with 10 mM HCl. The assays were always performed at the same temperature as the cells were grown.

Reproducibility of results.

All experiments were performed at least three times. Mean values or results of typical experiments are presented, as appropriate.

RESULTS

Effect of temperature on fermentation profiles.

Yeast fermentation profiles were established at 15, 25, and 30°C, up to glucose exhaustion. An isolate from a wine commercial starter was used in white grape must to simulate real winemaking conditions as closely as possible. As expected, the duration of fermentation decreased with temperature (Fig. 1 and Table 1; data for 25°C are not shown in the figure).

Fig 1.

Fermentation kinetics of S. cerevisiae ISA1000 in white grape must at 15°C (A) or 30°C (B). Dashed lines (X, Y, and Z) indicate sample points for pH_in determination and H⁺ movement assays, corresponding to exponential and early and late stationary phases, respectively. Abs, absorbance.

Table 1.

Macrokinetic characteristics for fermentations performed at different temperatures

Macrokinetic parameter	Value at temp (°C)a
	15	25	30
Duration of glucose fermentation (h)	300	142	72
Lag phase (h)	15	6	<2
Doubling time (h)	6.5 ± 0.2	2.4 ± 0.1	1.3 ± 0.1
Final dry biomass (g liter−1)	5.8 ± 0.3	6.0 ± 0.6	5.6 ± 0.1

^aDoubling time and biomass are expressed as means ± SD.

At all temperatures tested, after a few doublings, growth slowed down, cultures entered stationary phase, and most of the glucose and fructose fermentation was carried out by resting cells (at least 70% of the overall fermentation time).

At lower temperatures, the lag phase and the doubling time were longer, resulting in a delay in reaching the maximum population (10⁸ CFU ml⁻¹). Once it was reached, it remained high during the whole process. The maximum biomass value was quite stable with temperature, with only a slight reduction observed at 30°C.

The relative amount of viable cells in the total population remained quite stable during must fermentation for all temperatures tested. Approximately 50% of the total population was able to form colonies when plated under nonstress conditions (Fig. 1). Specific glucose and fructose consumption rates increased with temperature, as expected. Values of 0.24, 0.25, and 0.77 g glucose (g dry biomass)⁻¹ h⁻¹ and 0.13, 0.15, and 0.29 g fructose (g dry biomass)⁻¹ h⁻¹ were estimated at the beginning of stationary phase at 15, 25, and 30°C, respectively. As expected, fructose was fermented at a lower rate, and when glucose was exhausted, the cells were still fermenting fructose. For both sugars, there was a steep increase in specific consumption rates when the temperature rose from 25 to 30°C, which is noteworthy. The initial pH of the must was 3.0, and a drop of 0.2 U was observed at the beginning of stationary phase for all temperatures tested. The fermentation profiles were established as a preliminary stage for H⁺ homeostasis studies. In Fig. 1, the dashed lines indicate sample points for subsequent experiments, representative of exponential phase (X) and early (Y) and late (Z) stationary phases.

Intracellular pH during fermentation.

Intracellular pH of cells collected in exponential and early and late stationary phases was estimated by the relative distribution of radiolabeled benzoic acid. As Fig. 2 illustrates, pH_in was not affected by temperature (the value remained between 6.0 and 6.4) up to the beginning of stationary phase. When cells were grown under standard lab conditions (mineral medium, 25°C, exponential phase), a similar value (pH_in 6.4) was obtained. For all the temperatures tested, pH_in fell as the fermentation proceeded further. Nevertheless, this drop was more pronounced for stationary-phase cells at lower temperatures, in particular at 15°C, reaching a value of pH_in 5.2 at the end of glucose fermentation. At 30°C only a minor drop occurred (pH_in 5.8).

Fig 2.

Intracellular pH of S. cerevisiae ISA1000 during white grape must fermentation at 15, 25, and 30°C. Cells were collected at representative sample points for exponential and early and late stationary phases.

Proton fluxes during fermentation.

Proton efflux rates can be correlated with the activity of plasma membrane ATPase (23). In the present study, these rates were estimated in cells collected at exponential and early and late stationary phases throughout fermentations carried out at 15, 25, and 30°C. The maximum plasma membrane H⁺ efflux rate was reached for cells cultivated at 30°C, at early stationary phase (2.41 mmol [g dry biomass]⁻¹ h⁻¹), and decreased at late stationary phase (Fig. 3). At 15 and 25°C, a progressive drop in H⁺ efflux occurred as long as the fermentation proceeded (Fig. 3). Clearly, late-stationary-phase cells exhibited the lowest rate of H⁺ efflux.

Fig 3.

Rates of net H⁺ efflux in S. cerevisiae ISA1000. Cells were collected at exponential and early and late stationary phases during white grape must fermentation at 15, 25, and 30°C. The assays were performed at the same temperatures at which the cells were grown.

Proton inward movements were registered after the application of an HCl pulse to the cell suspension, which promoted a sudden decrease in the external pH (from 5 or 6 to 4). The decrease in external pH was followed by an immediate steady increase due to H⁺ influx into the cells (13, 20), which was observed for a few minutes. Figure 4 illustrates that the H⁺ influx rate at 30°C was higher than that at 15°C for all fermentation stages. The H⁺ influx rate was highest in exponential-phase cells at 25°C and was particularly low throughout fermentations performed at 15°C (maximum value, 0.10 mmol [g dry biomass]⁻¹ h⁻¹ in cells harvested at exponential phase). In all cases, it decreased as the fermentation proceeded. The surprising result is that H⁺ influx was almost insignificant in late-stationary-phase cells, suggesting that cells were almost impermeable to H⁺ at this stage at the three temperatures tested. We still considered that another hypothesis to explain this observation could be very high leakiness of the cells. If cells at late stages of fermentation were very leaky, this would cause the absence of membrane ΔpH, implying no driving force for H⁺ influx. To test this hypothesis, early- and late-stationary-phase cells were preincubated in weak buffer solutions at pH 5.0 and pH 6.3, and the pH was very rapidly adjusted to 4.0 in the reactor. If cells were leaky, their pH_in would enter equilibrium with extracellular pH (pH_out) during the incubation time in buffer. Therefore, the expected ΔpH during the assay would be higher in cells preincubated at pH 6.3, and a significantly higher H⁺ influx would be observed in this case. The results did not support this hypothesis. Permeability was low even under these conditions (data not shown). Accordingly, we concluded that plasma membranes from late-stationary-phase cells were impermeable to H⁺.

Fig 4.

Rates of passive H⁺ influx in S. cerevisiae ISA1000. Cells were collected at exponential and early and late stationary phases during white grape must fermentation at 15, 25, and 30°C. Each assay was performed at the same temperature at which the cells were grown.

Effect of ethanol on proton influx.

Ethanol has been described as affecting plasma membrane permeability, resulting in leaky membranes and increasing the passive fluxes through the bilayer (13, 20). The effect of ethanol on H⁺ influx was evaluated in this study as an estimation of plasma membrane integrity.

Figure 5 illustrates the H⁺ influx rates in the presence of increasing concentrations of ethanol, at 15°C (Fig. 5A) and 30°C (Fig. 5B) for cells collected at exponential and late stationary phases.

Fig 5.

Effect of ethanol on the rates of passive H⁺ influx in S. cerevisiae ISA1000 cells collected at exponential and late stationary phases during white grape must fermentation performed a 15°C (A) or 30°C (B). Each assay was performed at the same temperature at which the cells were grown.

For both temperatures, influx rates were extremely low for stationary-phase cells compared to those for exponential cells (Fig. 5A and B) although always higher at 30 than at 15°C. As expected, ethanol in the assay increased the rate of passive H⁺ influx. Amazingly, this was not the case for stationary-phase cells at 15°C, since even with 20% (vol/vol) ethanol, H⁺ influx was almost not detected.

DISCUSSION

In previous decades, many studies of the physiology of S. cerevisiae growing in nonnatural environments have been performed, and there is currently a good understanding of the processes that occur inside an Erlenmeyer flask when cells are grown in synthetic medium. However, our comprehension of what happens under “real-life” conditions, like those during grape must fermentation, is still poor, and it is often difficult to extrapolate useful interpretations from lab studies to explain yeast performance during winemaking. This is particularly true for cells in late stages of the fermentation process, since most of the data collected under lab conditions result from assays performed with cells harvested in the exponential phase. There have been, so far, very limited data collected for the “less-conventional” stationary cells, whose performance is certainly critical for the fine tailoring of wine quality. Nevertheless, it is well known that the entry into a “resting state,” typically accompanied by a dramatic decrease in the overall growth rate, is characterized by an increased resistance to a variety of stresses (9). An important point is that yeast cultures are dynamic systems, dealing with variable problems and conditions in different stages of wine fermentation.

The aim of this work was to study yeast proton homeostasis during wine fermentation, keeping in mind that metabolism and the regulation of pH_in are interdependent processes. Cells require a suitable pH_in both in the context of general cellular functioning and as a powerful signaling tool (24). Additionally, metabolic activity (ATP production) is required so that the ATPase can accomplish its function of pumping H⁺ out, thereby regulating pH_in. This is particularly relevant in harsh environments of low external pH and high ethanol concentrations, which presumably could significantly increase passive H⁺ influx. In the present study, white grape must was fermented by an enological S. cerevisiae strain at temperatures used in wineries: 15°C (commonly used for white wine), 25°C (standard lab conditions), and 30°C (commonly used for red wine). We anticipated that temperature would also affect H+ homeostasis, acting synergistically with ethanol, as described in classical works by van Uden and coworkers (17, 35).

Yeast populations that are fermenting must spend most of their lifetime in stationary phase. In our work, this was most evident at 15°C (Fig. 1). Although resting cells in late stages of fermentation face harsh, deleterious conditions, we confirmed that cells maintained the ability to form colonies (approximately 50%) throughout the whole fermentation period and kept their capacity to ferment glucose up to exhaustion. These results indicate that cells seem to adapt fairly well to stress conditions occurring during late stages of the wine fermentation process.

Remarkably, the pH of the grape must used in this study (pH 3.0) did not vary significantly during fermentation, which reflects its high buffer capacity. Also, in exponential-phase and early-stationary-phase cells, pH_in was quite stable and did not vary with the temperature of fermentation. The values measured for pH_in were in the range usually found in the literature for cytosolic pH of S. cerevisiae (6, 25), although there is still the possibility that the data reflect a nonhomogeneous population with a mix of metabolically active and inactive cells with different pH_in values. At the end of glucose fermentation, in particular at 15 and 25°C (Fig. 2), the situation was different. In this case, pH_in was close to the values previously reported as the minimum at which fermentation would proceed (pH 5) (25), creating challenging conditions for metabolic activity (although cells were still viable). As expected, fermentation was slower at lower temperatures, not only because most enzymes are less active at lower temperatures but most probably also because low pH_in affects their activity. It was previously shown that this is true in particular for hexokinase, enolase, and glucose-6-Pi isomerase (26).

Our first hypothesis to explain the lower pH_in at the end of fermentation was that the cells would become leaky to H⁺ in this situation, since ethanol disturbs the integrity of the lipid bilayer (3, 7), thereby enhancing passive diffusion (13, 20). In addition, the H⁺-ATPase could have been impaired by ethanol (20). In this stage, the H⁺ pump plays an important role in maintaining pH_in values, as confirmed by the highest pH_in value obtained at 30°C, when the pumping activity was not so severely affected. The good yeast performance at this temperature (30°C) should be emphasized: the fermentation proceeded much faster (the lag phase was not detected, the duration of fermentation was shorter, and accordingly the specific sugar consumption rate was much higher than at 25°C), and probably as a consequence, the H⁺ efflux was more intense, resulting in higher pH_in (more evident in late stationary phase) than at other temperatures.

To our surprise, cells in late stationary phase showed a very low H⁺ influx compared with cells harvested in early stages (Fig. 4). This result might be explained either by the lower proton motive force across the plasma membrane in these cells or by a lower permeability coefficient. We confirmed that the permeability was significantly low even when a higher ΔpH was artificially imposed by preincubating the cells at higher external pH. When we evaluated the effect of ethanol present in the assay of H⁺ influx, rates were extremely low for stationary-phase cells compared with those for exponential-phase cells for all ethanol concentrations tested. It was demonstrated that in stationary-phase cells, even 20% (vol/vol) ethanol had no effect on H⁺ influx rates at 15°C, and at 30°C, only the highest ethanol concentration used (20% [vol/vol]) was effective in slowing down H⁺ influx. Our results suggest that during the long fermentation process, the cells have gradually changed to become less permeable to H⁺, presumably due to changes in membrane composition.

Considering the current trends toward the production of high-quality wines with distinctive and very characteristic properties and taking into account the economic requirements for efficient fermentation processes, the demand of wine makers for “special yeasts for special traits” able to perform an economically viable fermentation process still remains to be satisfied. An in-depth understanding of the molecular and physiological events underlying the response and adaptation of yeast cells to the various stress conditions occurring at late stages of fermentation is of the utmost importance and will contribute to the development of more robust strains suitable to perform a more efficient wine fermentation.