OUP user menu

Metabolome dynamic responses of Saccharomyces cerevisiae to simultaneous rapid perturbations in external electron acceptor and electron donor

Mlawule R. Mashego, Walter M. van Gulik, Joseph J. Heijnen
DOI: http://dx.doi.org/10.1111/j.1567-1364.2006.00144.x 48-66 First published online: 1 January 2007

Abstract

Rapid perturbation experiments are highly relevant to elaborate the in vivo kinetics for mathematical models of metabolism, which are needed for selecting gene targets for metabolic engineering. Perturbations were applied to chemostat-cultivated biomass (D=0.05 h−1, aerobic glucose/ethanol-limited) using the BioScope of Saccharomyces cerevisiae CEN. PK 113-7D over time span of 90 and 180 s. The availability of the external electron acceptor oxygen was decreased from fully aerobic to anaerobic conditions. It was observed that the changes in metabolome response under these conditions were limited to the pyruvate node. Acetaldehyde supply was used as an extra external electron acceptor during glucose perturbation under fully aerobic conditions. This had a strong effect on the metabolome dynamics and resulted in a significantly higher initial glycolytic flux. Dynamic response of the adenine nucleotides indicated that their behavior is not dictated by the glycolytic flux but is much more coupled to the cytosolic NADH/NAD+ ratio through the equilibrium pool of fructose 1,6-bisphosphate and 2/3-phosphoglycerate. Also, the electron donor availability (glucose) was decreased. This did not result in significant changes in the concentrations of the glycolytic and tricarboxylic acid cycle metabolites, whereas the adenine nucleotides, especially ADP and AMP, showed the opposite response to that observed in a glucose pulse experiment. Surprisingly, trehalose was not mobilized in the time frame of 180 s.

Keywords
  • metabolomics
  • Saccharomyces cerevisiae
  • glucose perturbation
  • acetaldehyde
  • O2
  • BioScope

Introduction

A rational approach towards the identification of gene targets for the metabolic engineering of (micro)organisms requires quantitative information on the in vivo kinetics of metabolism (Bailey, 1991; Stephanopoulos, 1998; Nielsen, 2001).

Stimulus response experiments pioneered by Theobald (1993) performed on Saccharomyces cerevisiae grown under carbon-limited chemostat conditions have provided a platform for studying the in vivo kinetics and regulation of primary central carbon metabolism. Such stimulus–response experiments (Oldiges & Takors, 2005) traditionally involve an instantaneous increase in residual glucose concentration followed by rapid sampling; instant quenching; intracellular metabolite extraction and measurements of the intracellular metabolite concentrations over a short time period. The short time scale is important for the assumption of constant enzyme levels, because these experiments are designed to study the in vivo kinetics of key regulatory enzymes, whose response to the increased metabolite levels must be dissected from the increase or decrease in enzyme levels. The increase or decrease in enzyme levels as a regulatory mechanism in cells is believed to occur at a longer time frame of several minutes (Stephanopoulos, 1998). These techniques are important for success in quantitative analysis and modeling of the in vivo kinetics of metabolism (Harrison & Maitra, 1969; de Koning & van Dam, 1992; Theobald, 1993, 1997; Gonzalez, 1997; Schaefer, 1999; Buchholz, 2002; Buziol, 2002; Schmitz, 2002; Visser, 2002; Mashego, 2004).

However, it should be recognized that although glucose perturbation experiments have paved the way and provided valuable information for understanding the kinetics of primary carbon metabolism (especially glycolysis), different stimulus–response experiments using alternative perturbing agents or a combination thereof are of major interest for increasing our current understanding of the in vivo kinetics of primary cellular metabolism, as has recently been demonstrated (Buchholz, 2002; Visser, 2004). Standard glucose perturbations (Theobald, 1993, 1997; Visser, 2002; Mashego, 2004) perturb primary metabolism from the top, i.e. at glucose entrance. This traditional approach leads to a quite strong kinetic response of glycolysis, whereas the tricarboxylic acid (TCA) cycle response is only moderate. Furthermore, the important anaplerotic reactions are addressed less in a glucose perturbation experiment. Hence, other points of perturbation are of general interest, as has been most recently attempted in the perturbation of the aromatic amino acid pathway in Escherichia coli (Oldiges, 2004). For example, perturbation of O2 uptake can be expected to directly interfere with redox metabolism and the TCA cycle. Another possibility for perturbing redox metabolism is to offer an additional external electron acceptor. In yeast, acetaldehyde can act as an additional electron acceptor, because it can be reduced to ethanol, thus draining the NADH produced from glycolysis.

Acetaldehyde and glucose perturbation of oscillating yeast culture under fermentation conditions has been performed before (Richter, 1975). It was found that the NADH level dropped rapidly, and ATP accumulated, leading to reduced activity of phosphofructokinase. In turn, glucose 6-phosphate and fructose 6-phosphate accumulated, whereas fructose 1,6-bisphosphate (F-1,6-BP) Glyceraldehyde-3-phosphate (GAP) and Dihydroxyacetonephosphate (DAP) concentrations decreased and remained low (Richter, 1975). Furthermore, the anaplerotic route can be expected to be sensitive to CO2 levels; therefore, the response to different CO2 levels is of interest too. In the standard glucose perturbation, the glucose concentration is significantly increased, leading to a strong increase in the glycolytic flux. From a kinetic point of view, a lower glycolytic flux is also of major interest. This can be achieved by sudden termination of the glucose feed to cells grown under glucose-limited conditions.

The recently developed modified BioScope (a mini-plug flow reactor) allows the performance of different perturbations on steady-state biomass obtained from the same chemostat culture.

In the present work, perturbations of both the electron acceptor and the electron donor were carried out in the modified BioScope, with the aim of elucidating the kinetics of central carbon metabolism in S. cerevisiae. First, perturbation experiments were performed whereby a pulse of the electron donor (glucose) was combined with reduced (O2-limited and anaerobic) as well as with increased (through addition of acetaldehyde) electron acceptor availability.

Second, a perturbation experiment was performed with reduced electron donor availability (through termination of the glucose feed).

Materials and methods

Yeast strain and maintenance

The haploid, prototrophic S. cerevisiae strain CEN.PK113-7D was kindly provided by Dr P. Kötter (Frankfurt, Germany). Precultures were grown to stationary phase in shake-flasks on mineral medium (Verduyn, 1992), adjusted to pH 6.0 and containing 2% (w/v) glucose. After addition of glycerol (30% v/v), 2-mL aliquots were stored at −80°C in sterile vials. These frozen stocks were used to inoculate precultures for chemostat cultivation.

Preculture conditions

Precultures were grown in 100 mL of mineral medium (Verduyn, 1992), contained in 500-mL Erlenmeyer flasks at 30°C overnight with shaking at 220 r.p.m. in an incubator shaker (New Brunswick, NJ).

High-density and low-density chemostat cultivation

Aerobic glucose/ethanol-limited chemostat cultures of S. cerevisiae were grown at a dilution rate (D) of 0.05 h−1 in 7-L laboratory fermenters (Applikon, Schiedam, The Netherlands) with a working volume of 4 L at 30°C and pH 5.0, as reported previously (Lange, 2001; Mashego, 2005). The molar composition of the feed medium used has been described previously (Lange, 2001). This medium contains a mixed carbon source of 150 mM glucose and 31 mM ethanol. The addition of ethanol successfully prevented the cultures from oscillations. The biomass concentration supported by this medium is c. 14.5 g L−1±2%. Cultures grown on this medium are henceforth described as high-cell-density cultures. The second medium used had the same composition as described above, except that the concentrations of all medium components were halved, including the mixed carbon source; for example, 75 mM glucose and 15.5 mM ethanol were used instead. The biomass concentration supported by this medium composition is c. 7.22 g L−1±2%. Cultures grown on this medium are henceforth described as low-cell-density cultures. For the low-cell-density cultures, the aeration rate was also halved compared to the high-density protocol (Lange, 2001), in order to minimize differences in the dissolved-O2 and -CO2 concentrations.

Experimental design

BioScope glucose perturbations with reduced external electron acceptor (O2) availability were performed using chemostat-cultured biomass at high cell density (14.5 g L−1). In these experiments, it was necessary to establish a fast transition from fully aerobic conditions in the chemostat to conditions of reduced O2 availability or fully anaerobic conditions in the BioScope. To achieve this, the steady-state dissolved-O2 tension (DOT) in the chemostat was reduced from c. 80% to 27%. This was accomplished using a 1 : 1 mixture of air and N2 gas for aeration of the chemostat, thereby maintaining the same gassing rate of 3.33 L min−1. Because the gas flow was not changed, the dissolved-CO2 concentration in the fermenter was expected to remain the same. The O2 concentration of this gas mixture was confirmed by passing it through a Rosemount NGA 2000 off-gas analyser (Fisher-Rosemount, Germany), and it did indeed contain 10.47% O2. Second, it was necessary to verify whether the physiological state of the culture was not changed when the DOT was lowered to 27%. To this end, intracellular metabolite levels of the steady-state culture growing at a reduced DOT of 27% were compared with metabolite levels of the same steady-state culture growing at 80% DOT.

Glucose perturbations with reduced electron acceptor (O2) availability

The first set of three experiments (Table 1) was performed using the high-cell-density chemostat.

View this table:
Table 1

Comparison of specific uptake rates of glucose (qglucose) and specific oxygen uptake rate (qO2), with a glucose perturbation (5.8 mM) under fully aerobic, O2-limited and anaerobic conditions in the BioScope fed with broth from the high-cell-density chemostat; values in parentheses indicate the 95% confidence intervals

BioScope perturbation experimentsq glucose [μmol (g DW s)–1]qO2μmol (g DW s)–1
Phase I (0–22 s)Phase II (22–90 s)
Fully aerobic2.73 (2.34–3.12)1.00 (0.85–1.16)0.57
O2-limited (fixed qO2)2.37 (1.88–2.86)1.47 (1.20–1.74)0.19
Anaerobic1.67 (0.99–2.34)1.28 (1.20–1.37)0.00
Steady state0.140.140.38

Experiment IA

In this experiment, a glucose perturbation was carried out in the BioScope under fully aerobic conditions. A continuous flow of culture broth from the chemostat was introduced into the BioScope and mixed at a ratio of 1 : 9 with a continuous flow of 56 mM glucose solution maintained at 30°C to obtain an initial glucose concentration at the entrance of the BioScope of 5.6 mM. The flow rate of the glucose solution was 0.2 mL min−1 and that of the culture broth was 1.8 mL min−1, giving a total flow rate of 2.0 mL min−1. With this flow rate, the total residence time in the BioScope was 90 s. The broth and the glucose solution were pumped using two independent high-precision peristaltic pumps (Ismatec, ISM 930, ISMATEC SA, Labortechnik, CH-8152, Glattenbrugg-Zürich, Switzerland). During the experiment, the gas channel of the BioScope was continuously flushed with air (20.95% O2) at a constant flow rate of 130 mL min−1. At this flow rate, the O2 concentration was close to 21% and the CO2 concentration was close to 0%, allowing maximum O2 transfer and CO2 removal. The DOT in the fermenter during steady state was kept constant at 80% air saturation by continuous gassing with air at a constant flow rate of 3.33 L min−1 and a reactor overpressure of 0.3 bar. During the glucose pulse experiment, the DOT in the BioScope was monitored at various sample ports using a dissolved-O2 probe mounted onto a flow cell with a volume of 200 μL. This flow cell was also used for measurements of dissolved CO2, except that a dissolved-CO2 probe was used (Mettler Toledo CO2 measuring system, In Pro 5100e, Mettler-Toledo GmbH, Switzerland).

Experiment IB

In this experiment, a glucose perturbation was carried out under O2-limited conditions. The experimental conditions were as described above, except that the gas composition in the gas channel of the BioScope was a mixture of 25% air and 75% N2. In this way, the O2 concentration was lowered to 5.25%. The gas flow rate was maintained at the same value of 130 mL min−1. The solubility of O2 for this gas composition and under the experimental conditions applied was C*OL=0.061 mmol L−1. For these experiments, the DOT of the chemostat was reduced from 80% to 27% as described above, to minimize O2 carryover from the chemostat broth to the BioScope. The dissolved-O2 tension at the last sample port of the BioScope was measured with the dissolved-O2 probe as described above, and found to be zero. Using the O2 transfer characteristics of the BioScope, the maximum O2 transfer rate was 9.25 mmol L−1 h−1 (volumetric mass transfer rate [KLa]=148 h−1, C*OL=0.061 mmol L−1) or qO2=0.69 mmol [g dry weight (DW) h]−1 [or 0.192 μmol (g DW s)−1; Table 1]. This is about 45% of the qO2 in the steady-state chemostat.

Experiment IC

In this experiment, a glucose perturbation was carried out under fully anaerobic conditions. The experimental conditions were the same as described above for the O2-limited (fixed qO2) glucose perturbation experiment, except that the gas channel of the BioScope was continuously flushed with N2 at a constant gas flow rate of 130 mL min−1. During the glucose pulse, the DOT at the sample ports of the BioScope was equal to zero, thereby ensuring that anaerobic conditions were satisfied.

Glucose perturbations with increased electron acceptor (acetaldehyde) availability

In the previous experiments, reduced availability of the electron acceptor (O2) during glucose perturbation was performed. However, increased electron acceptor availability during a glucose perturbation is also of interest, because of the known redox stress originating from high flux of glyceraldehyde-3-phosphate dehydrogenase. Therefore, a second set of two experiments was performed in the BioScope, using biomass from the low-cell-density (7.22 g L−1) chemostat.

Experiment IIA

In this experiment, a glucose perturbation was carried out with a 50% lower initial glucose concentration of 2.8 mM under fully aerobic conditions. The time frame of the experiments was extended, which was achieved by setting the feed rate of the glucose solution (28 mM) at 0.1 mL min−1, whereas that of the culture broth was set at 0.9 mL min−1, leading to a final total flow rate of 1 mL min−1. In this way, the residence time in the BioScope was increased from 90 to 180 s. The gas channel part of the BioScope was continuously flushed with air (20.95% O2) at 130 mL min−1. The DOT in the fermenter during steady state was maintained constant above 80% saturation (0.196 mM) by continuous aeration at a flow rate of 1.67 L min−1 with a reactor overpressure of 0.3 bar. The dissolved-CO2 concentration in the chemostat (0.52 mM) was identical to that of the high-cell-density culture because the airflow was halved.

Experiment IIB

In this experiment, the availability of the external electron acceptor was increased by supplying acetaldehyde during the glucose pulse experiment. The reduction of one acetaldehyde molecule to one ethanol molecule leads to the consumption of one NADH molecule. Glucose (56 mM) and acetaldehyde (33 mM) solutions were premixed at a 1 : 1 ratio with a high-precision multichannel peristaltic pump (Ismatec, ISM 930, ISMATEC SA, Labortechnik, CH-8152, Glattenbrugg-Zürich, Switzerland), using two independent pump tubings (0.64-mm internal diameter) connected via a Y-piece connector prior to the BioScope entry port. The flow rates of the two solutions were independently found to be both equal to 0.05 mL min−1, thus giving a total measured and flow rate of 0.1 mL min−1 for the glucose/acetaldehyde mixture. The total flow rate of 0.1 mL min−1 was verified prior to the commencement of the experiment. In this way, the initial glucose concentration at the entrance of the BioScope was the same as in glucose pulse experiment IIA, i.e. 2.8 mM. The initial acetaldehyde concentration at the entrance of the BioScope was calculated to be 1.5 mM. The experiment was carried out under fully aerobic conditions by flushing air (130 mL min−1) through the gas channel of the BioScope.

Experiment IIIC

In this experiment, the availability of the electron donor was reduced under fully aerobic conditions, using biomass from the low-cell-density chemostat. This was achieved by feeding the BioScope with steady-state broth from the chemostat at a flow rate of 0.9 mL min−1 without supplying any glucose. Because only a small amount of glucose is available, i.e. the steady-state residual concentration of 0.11 mM as the broth leaves the chemostat, the dissolved-CO2 concentration in the BioScope would be expected to drop, due to the strongly decreased CO2 production combined with CO2 mass transfer to the gas phase via the silicone membrane. In order to maintain a constant CO2 concentration in the BioScope, dissolved CO2 was added to the broth taken from the low-cell-density chemostat such that the dissolved-CO2 concentration was tripled instantaneously in the BioScope, i.e. instant increase from 0.52 to 1.7 mM. This was achieved by continuously mixing water saturated with 40% CO2 at 0.1 mL min−1 liquid flow rate with 0.9 mL min−1 broth in the BioScope. The gas channel part of the BioScope was continuously flushed with a CO2-enriched gas mixture containing 5% CO2 and 20.95% O2, the rest being N2. The 5% CO2 in the gas phase corresponds to a solubility of 1.45 mM CO2 in the liquid phase. Therefore, flushing the gas channel of the BioScope with an enriched (5% CO2) gas mixture ensured near-zero CO2 transfer or limited CO2 escape from the liquid phase to the gas channel of the BioScope. The dissolved-CO2 in the liquid stream was measured at all sampling ports of the BioScope using a flow-through cell in which the dissolved-CO2 probe (In Pro 5100e; Mettler-Toledo GmbH; Switzerland) was mounted. The perturbation experiments using either the high-cell-density or low-cell-density chemostat were performed c. 24 h after each other over a total period of 3 days in the sequence in which they are presented above.

Culture dry weight determinations

The biomass dry weight concentration was determined as described previously (Mashego, 2005).

Sampling procedure and analysis for extracellular metabolites (glucose, acetaldehyde, ethanol and acetate)

The applied procedures for rapid sampling, quenching and analysis of extracellular metabolites (glucose, ethanol and acetate) from the BioScope samples were as described previously (Mashego, 2003). For validation of the procedure for the determination of the residual acetaldehyde in the glucose + acetaldehyde perturbation, an acetaldehyde stock solution (33 mM) kept at 30°C in the temperature-controlled BioScope cabinet was fed through the BioScope, and samples were collected with simultaneous quenching as described previously for extracellular metabolites (Mashego, 2003). The acetaldehyde concentration in these samples was found to be 32 mM, indicating that no significant loss of acetaldehyde during sampling/quenching occurred. At the end of the experiment (c. 6 h), a sample from the acetaldehyde stock solution bottle was also collected in order to check the stability and the concentration of the acetaldehyde at the end of the experimental period. This was necessary because acetaldehyde is known to be very unstable at room temperature and tends to trimerize/evaporate. The concentration of the acetaldehyde stock solution was found to be 25 mM, representing a loss of about 20% in 6 h at 30°C.

Sampling procedure for intracellular metabolites

The BioScope system was flushed with a mixture of broth and the perturbing agent for at least five residence times (7.5–15 min) before sampling was started through computer-aided automatic activation of the sequentially positioned, two-way pinch valves. The sampling and quenching procedure for determination of intracellular metabolites was the same as described previously (Visser, 2002).

Trehalose extraction, enzymatic hydrolysis and analysis

Samples for intracellular trehalose determination were extracted using the ethanol boiling method (Mashego, 2005). To a 100-μL sample from the trehalose extraction procedure, 50 μL of trehalase (Sigma-Aldrich, Steinheim, Germany) 0.1 U mL−1 and 850 μL of 80 mM sodium acetate buffer (pH 5.2) were added. The mixture was incubated at 37°C for 6.5 h with shaking at 220 r.p.m. in 1.5-mL Eppendorf tubes. The glucose concentration in the trehalase-hydrolyzed samples was measured enzymatically as described previously (Mashego, 2005).

Intracellular metabolite analysis procedure

Intracellular metabolite concentrations were measured in the cell extracts as described previously using isotope dilution mass spectrometry (van Dam, 2002; Mashego, 2004; Wu, 2005).

Results and discussion

Comparison of the physiological steady state of aerobic glucose/ethanol-limited chemostat culture of S. cerevisiae at 27% and 80% DOT

It is necessary to decrease the dissolved-O2 concentration in the BioScope to very low levels as soon as possible for the perturbation experiments involving different O2 supplies. To achieve this, the chemostat (14.5 g L−1 biomass) dissolved O2 was reduced from 0.196 mol m−3, corresponding to 80% DOT, to 0.066 mol m−3, corresponding to 27% DOT. From the off-gas analysis in the chemostat, it follows that this does not cause any change in qCO2 and qO2. Specific fluxes: carbon dioxide evolution rate (qCO2), oxygen uptake (qO2), qglucose and qethanol were 0.039±2% and 0.042±2% mol (C-mol biomass h)−1, 0.013 mol (C-mol biomass h)−1 and 0.0028 mol (C-mol biomass h)−1, respectively. The carbon and redox balance closed at 100% and 101%, respectively.

Furthermore, intracellular metabolite concentration levels representing the physiological steady state of the culture were also compared under both 27% and 80% DOT conditions. The measured intracellular metabolite concentrations for both conditions are shown in Fig. 1. All values are averages of four steady-state samples analyzed in duplicate. It can be seen from Fig. 1 that lowering of the DOT from 80% to 27% air saturation has no significant effect on the intracellular metabolite concentrations. Apparently, a decrease of the dissolved-O2 level from 0.196 to 0.066 mol m−3 has no effect on the physiological state of S. cerevisiae. Using the mass transfer characteristics of the BioScope, it was calculated that the dissolved-O2 concentration in the BioScope drops to near zero at 0.4-m length of the BioScope channel, corresponding to 6 s (see Appendix). This is true provided that the broth from the chemostat contains 0.066 mol m−3 (27% DOT). For a DOT of 80%, it takes about three times longer to achieve a DOT of zero in the BioScope (see Appendix). O2 limitation in the BioScope can therefore be assumed after a few seconds.

Figure 1

Comparison of the concentrations of intracellular glycolytic and tricarboxylic acid (TCA) cycle metabolites, 6-phosphogluconate, mannose 6-phosphate and trehalose 6-phosphate, in aerobic glucose/ethanol-limited steady-state chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D grown at a dilution rate of 0.05 h−1 at 27% and 80% dissolved-oxygen tension (DOT). Error bars represent standard deviations between analyses of four samples.

Perturbation of external electron acceptor during a glucose pulse

The first set of experiments concerned the lower availability of the electron acceptor O2. A glucose perturbation was applied under fully aerobic, O2-limited and anaerobic conditions (Table 1). The response will be discussed with respect to the effect on uptake and secretion fluxes. Under standard glucose perturbation conditions in the BioScope, there is sufficient dissolved oxygen (>0.05 mmol L−1), as shown in Fig. 2a. During the perturbation experiment, the CO2 concentration increased from 0.50 to 1 mM.

Figure 2

Dissolved carbon dioxide (DCO2), dissolved-oxygen concentration, residual glucose, excreted ethanol and acetate concentration profiles before and during a glucose perturbation of high-cell-density fully aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D under aerobic, O2-limited and anaerobic conditions in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

Using the measured dissolved O2 concentration of 0.05 mmol L−1 and the O2 transfer characteristics of the BioScope (KLa=148 h−1, C*OL=0.245 mmol L−1 and biomass concentration [Cx] =13.05 g L−1) allows to calculate that the average qO2 during the perturbation experiment equals 0.57 μmol (g DW s)−1 (Table 1). This is about 50% higher than the steady-state chemostat value of 0.38 μmol (g DW s)−1 (Table 1). A similar increase in qO2 has been found before in a glucose perturbation in the fermenter (Bloemen, 2003).

Figure 2b shows the time patterns of the glucose concentration during glucose perturbation experiments carried out in the BioScope under fully aerobic, O2-limited and anaerobic conditions. Upon visual observation of the measured patterns of the glucose concentration (Fig. 2b), an initial phase of rapid glucose consumption can be distinguished during the first 22 s after the perturbation, followed by a second phase of less rapid glucose uptake. The glucose uptake rates during both phases were calculated for the three perturbation experiments by means of weighted linear regression. The results are shown in Table 1. It can be seen from Table 1 that the glucose uptake rate during phase I is significantly higher than in phase II, except for the anaerobic perturbation experiment. Furthermore, the glucose uptake rate in phase I of the anaerobic perturbation is lower than under aerobic conditions. However, it appears that the effect of reduced O2 availability on the glucose uptake rate, and hence the rate of glycolysis, is limited. With respect to secreted ethanol and acetate, Fig. 2c shows that the results of the O2 perturbation experiments (O2-limited and anaerobic) are quite similar, and when compared to the standard fully aerobic perturbation, the ethanol production is much higher and acetate production is much lower.

The effect of the O2 supply perturbation on intracellular metabolites is shown in Fig. 3, which represents the concentration profiles of the well-known intracellular glycolytic intermediates 6-phosphogluconate (6-PG), mannose 6-phosphate (M6P) and trehalose 6-phosphate (T6P). There was no significant difference for most metabolite concentration profiles under the different specific O2 uptake rates imposed. Only pyruvate seems to be higher for the O2-limited condition, which suggests less pyruvate flux into the TCA cycle. The above-mentioned higher ethanol and lower acetate levels are in accord with this higher pyruvate level and the flux being redirected from respiration (TCA cycle) to fermentation. Figure 3 shows that the dynamic metabolite concentration responses resemble the very well-known dynamic patterns of glucose perturbations (Theobald, 1997; Buziol, 2002; Visser, 2002; Mashego, 2004). Because the glycolytic flux is hardly different, no difference is to be expected in glycolytic intermediates.

Figure 3

Concentration profiles of intracellular glycolytic metabolites, 6-phosphogluconate and trehalose 6-phosphate, before and during a glucose perturbation of high-cell-density fully aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D under aerobic, O2-limited and anaerobic conditions in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

The dynamic responses of the TCA cycle metabolites citrate, fumarate and malate are shown in Fig. 4. The fumarate level seems to be higher in the anaerobic glucose perturbation, which suggests a more reduced state of the cytosol. This further supports the hypothesis of a slightly higher cytosolic NADH/NAD+ ratio under O2-limited perturbation conditions. This higher redox level with the higher pyruvate level can further explain the observed higher ethanol and lower acetate secretion. In conclusion, O2 perturbation mainly seems to affect the pyruvate node fluxes via pyruvate level and the redox level, and affects the glycolytic metabolites and fluxes much less.

Figure 4

Concentration profiles of intracellular tricarboxylic acid (TCA) cycle metabolites (citrate, fumarate and malate) before and during a glucose perturbation of high-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D under fully aerobic, O2-limited and anaerobic conditions in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

In the second set of glucose perturbation experiments, the availability of the external electron acceptor was increased by supplying acetaldehyde under fully aerobic conditions. Acetaldehyde readily permeates the cell membrane and is easily reduced to ethanol, consuming 1 mol of NADH per mol of acetaldehyde reduced.

In the glucose perturbation experiment, the extracellular glucose concentration was increased instantaneously from the steady-state value of 0.11–2.8 mM (Fig. 5a). Also, in this case two phases can be distinguished, an initial phase of rapid glucose uptake (0–30 s), followed by a phase of less rapid uptake (30–180 s). The glucose uptake rates during both phases were calculated for both conditions by means of weighted linear regression. The results are shown in Table 2. The observation of rapid initial and subsequent reduced glucose uptake fits very well with the apparent inhibition of hexokinase by an increase of T6P (Fig. 6), thus controlling the glycolytic flux (Blazquez, 1993; Hohmann, 1996; François & Parrou, 2001; Gancedo & Flores, 2004; Guillou, 2004). The specific glucose uptake rates in phase I for glucose only and glucose + acetaldehyde perturbation experiments were 1.66 and 2.80 μmol (g DW s)−1, respectively, leading to 12-fold and 20-fold increases for glucose only and glucose + acetaldehyde, respectively, relative to the steady-state qglucose flux (Table 2). The higher glycolytic flux in phase I under glucose + acetaldehyde perturbation conditions is apparently due to the electron removal by acetaldehyde, which freely diffused from the medium into the cell as an external electron acceptor. Figure 5b shows that the acetaldehyde level in the BioScope increased instantaneously from 0 to 1 mM, 5 s after the perturbation, and sharply decreased linearly in the first 100 s. There is a rapid consumption of acetaldehyde. Figure 5a shows that the presence of acetaldehyde also increases the glucose uptake rate, leading to much higher ethanol levels (Fig. 5c) and also slightly higher acetate levels (Fig. 5d).

Figure 5

Concentration profiles of residual glucose, acetaldehyde, ethanol and acetate before and during fully aerobic glucose-only and glucose + acetaldehyde perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

View this table:
Table 2

Comparison of specific uptake rates of glucose (qglucose) and acetaldehyde (qacetaldehyde) under steady-state culture conditions as well as during fully aerobic glucose-only and glucose + acetaldehyde perturbations with a glucose perturbation (2.8 mM) under fully aerobic conditions and feeding with broth from the low-cell-density chemostat; values in parentheses are standard deviations

BioScope perturbation experimentsGlucose qglucoseμmol (g DW s)−1q acetaldehyde [μmol (g DW s)−1]
Phase I (0–30 s)Phase II (30–180 s)
Glucose1.66 (1.02–2.30)0.95 (0.88–1.02)0.00
Glucose + acetaldehyde2.80 (2.51–3.09)0.94 (0.86–1.02)1.38
Steady state0.140.140.00
Figure 6

Comparison of concentration profiles of intracellular glycolytic metabolites, mannose 6-phosphate, 6-phosphogluconate, trehalose 6-phosphate and trehalose, before and during fully aerobic glucose-only and glucose + acetaldehyde perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

The specific acetaldehyde uptake rate was calculated to be 1.38 μmol (g DW s)−1, which is equivalent to 0.69 μmol O2 (g DW s)−1. As the qO2 under standard fully aerobic glucose perturbation conditions is 0.57 μmol O2 (g DW s)−1 (Table 1), it can be calculated that the electron sink is increased from 0.57 to 1.26 μmol O2 equivalent (g DW s)−1, due to acetaldehyde consumption. Both glucose and acetaldehyde were taken up simultaneously by the cells, leading to almost twice as much ethanol and acetate accumulation as was found in glucose-only perturbation (Fig. 5c and d).

Ethanol production started in both cases 5 s after addition of the glucose pulse. Figure 5c shows that the final ethanol concentration in the case of the glucose + acetaldehyde perturbation is 2.5 mM, which is 1.3 mM more than observed for the glucose-only perturbation (1.2 mM). The consumption of the 1 mM acetaldehyde explains the increase of ethanol production by maximally 1 mM. The remaining increase is due to the increased fermentation of glucose. Figure 5d shows that acetate production is increased, leading to an increase in concentration from about 0.1 to 0.25 mM in the presence of acetaldehyde, which indicates that a small part of the added acetaldehyde is oxidized to acetate.

This increase is probably not due to the increased acetaldehyde concentration, because the Km of acetaldehyde dehydrogenase is known to be very low (Wang, 1998). A likely explanation is that the cytosolic NADH/NAD+ ratio is lowered due to the acetaldehyde–alcohol conversion, as explained below. In summary, it appears that the presence of an extra cytosolic electron drain strongly (50%) stimulates the initial glycolytic flux (first 30 s) and leads to increased ethanol excretion rates.

The effects of the presence of the additional external electron acceptor acetaldehyde on the intracellular glycolytic and TCA cycle metabolites are shown in Figs 6 and 7, respectively. Adenine nucleotides are shown in Fig. 8. First, it was observed that the mass action ratios of phosphoglucose isomerase (PGI) and phosphoglucomutase (PGM) were perturbed but quickly returned to their steady-state values, whereas those of enolase and fumarase were also perturbed but remained at lower levels (data not shown). Comparing the effect of the presence of the additional external electron acceptor acetaldehyde, it follows that the concentrations of the hexose phosphate pools are lower, and that of F-1,6-BP is much lower, while the C3 pool (2/3-phosphoglycerate [2/3PG] and phosphoenolpyruvate [PEP]) decreased less quickly and pyruvate increased less quickly. These differences are remarkable in view of the increased glycolytic flux, which normally results in higher levels of the hexose phosphates and F-1,6-BP and a faster drop in the C3 pool. The behavior of the adenine nucleotides (Fig. 8) is even more remarkable. A glucose perturbation normally leads to a strong decrease in the ATP level and increases in ADP and AMP, as shown in Fig. 8. This is usually explained by the increased glycolytic flux, which consumes large amounts of ATP to achieve the hexose phosphorylation.

Figure 7

Comparison of concentration profiles of intracellular tricarboxylic acid (TCA) cycle metabolites before and during fully aerobic glucose-only and glucose + acetaldehyde perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

Figure 8

Intracellular adenine nucleotide (ATP, ADP and AMP) concentration profiles, sum of nucleotides, cellular energy charge and adenylate kinase mass action ratio during fully aerobic glucose-only and glucose + acetaldehyde perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represents standard deviations between duplicate analyses of the same sample.

In the presence of the extra external electron acceptor acetaldehyde, it can be seen that despite the strongly increased glycolytic flux, the nucleotides behave completely differently. Initially, ATP increases, whereas ADP and AMP decrease. This is most surprising as ATP production is supposed to be less, because glycolytically produced NADH is diverted through acetaldehyde reduction to produce ethanol, and therefore does not enter the electron transport phosphorylation pathway. This ATP increase is in agreement with the observations by Richter (1975). On the basis of these observations, it seems unlikely that the usually observed ATP decrease is caused by the high glycolytic flux. An alternative explanation is that the behavior of ATP and ADP is controlled by the set of equilibrium reactions between F-1,6-BP and 2/3PG (see Eqn [1]). Embedded Image 1

This set of reactions is known to operate close to equilibrium (Visser, 2004), even during a perturbation with increased fluxes. It would be expected that in the presence of acetaldehyde, the cytosolic NADH/NAD+ ratio would not increase very much. Indeed, the formation of glycerol 3-phosphate, which usually strongly responds to cytosolic redox stress, was observed to increase much more slowly in the presence of acetaldehyde (Fig. 6). Moreover, despite the higher glycolytic flux, the level of F-1,6-BP appeared to be much lower (Fig. 6) in the presence of acetaldehyde. This is also in agreement with the above equilibrium and the expected lower cytosolic NADH/NAD+ ratio. The fold change in the NADH/NAD+ ratio compared to the steady state can be calculated from the fold change of the intracellular metabolites. The intracellular phosphate concentration and the pH were not measured. However, in previous glucose pulse experiments (Wu, 2006), only a slight decrease of the intracellular phosphate concentration of c. 20% has been observed. With respect to the pH, Ramos (1989) observed a decrease from 6.8 to 6.4 as a response to a much higher glucose pulse of 20 mM. For our calculation of the fold change of the NADH/NAD+ ratio in the glucose pulse experiments with and without acetaldehyde, the pH and phosphate concentration were assumed to remain constant. As a result, the absolute values of the calculated NADH/NAD+ ratios are probably somewhat overestimated. The result is shown in Fig. 9. It can be seen from Fig. 9 that the presence of acetaldehyde leads to a much lower calculated fold change of the NADH/NAD+ ratio during the glucose pulse. This shows that acetaldehyde effectively balances the rapid production of glycolytic NADH. It is concluded that because of the equilibrium pool of F-1,6-BP and 2/3PG, the fold change in redox (NADH/NAD+) is coupled to the fold change in ATP/ADP ratio. A similar conclusion was drawn by Richard (1996) from measurements of glycolytic intermediates, adenine nucleotides and phosphate in oscillating yeast cultures grown in chemostats. Redox stress (high fold increase of NADH/NAD+) therefore leads to high fold decrease in the ATP/ADP ratio.

Figure 9

Comparison of NADH/NAD+ ratio before and during fully aerobic glucose-only and glucose + acetaldehyde perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope.

It is remarkable that the mass action ratio of adenylate kinase does not remain constant but decreases to the same extent in both experiments. A (pseudo) steady state is achieved 80 s after the perturbation at a value almost half of the value before the perturbation. This is an important observation, as this reaction is usually modeled as an equilibrium during dynamic conditions.

The presence of acetaldehyde during the glucose pulse also appears to have a strong effect on 6-PG, leading to a significant decrease in the concentration (Fig. 6) compared to the pulse without acetaldehyde. This might indicate that less NADPH is required from the pentose phosphate pathway, due to the oxidation of acetaldehyde to acetate as a source of NADPH, thus partly replacing the pentose phosphate pathway in the glucose + acetaldehyde pulse.

From the concentration profiles of the TCA cycle intermediates shown in Fig. 7, it can be seen that fumarate and malate increase much less in the presence of acetaldehyde. Their increase in a standard glucose perturbation is considered to be due to the increase of the NADH/NAD+ ratio, leading to a more reduced C4 pool (aspartate, oxaloacetate, malate and fumarate). This is in agreement with the much lower increase in the acetaldehyde experiment, where the change in NADH/NAD+ is much less. In contrast to this, the increase of the succinate concentration is much more pronounced in the presence of acetaldehyde. This could be due to an increased TCA cycle flux, which might be less inhibited due to the lower redox ratio (NADH/NAD+). In conclusion, the use of acetaldehyde seems to be an excellent method of creating highly informative perturbations in redox, energy and carbon of all major primary pathways.

Perturbation by lower electron donor supply (no glucose addition)

Figure 10 compares the fully aerobic standard glucose perturbation (2.8 mM) with a similar experiment without glucose addition. The glucose and CO2 concentration in the BioScope are shown in Fig. 10a. It can be seen that the glucose concentration slowly decreases from 0.11 to 0.03 mM. The dissolved-CO2 concentration is between 1.5 and 2 mM, which is comparable to the CO2 concentration that occurs in a fully aerobic glucose perturbation (1–1.5 mM). Figure 10b shows the calculated glucose uptake rate, obtained from differentiating a polynomial function that was fitted to the measured glucose concentration profile. It can be seen from Fig. 10 that the glucose consumption rate decreases sharply, reaching a value of zero after 60 s.

Figure 10

Residual glucose concentration, dissolved CO2 and specific glucose uptake rate, before and during fully aerobic, low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope with no-glucose pertubation. Error bars represent standard deviations between duplicate analyses of the same sample.

The concentration patterns of the glycolytic and TCA cycle intermediates and adenine nucleotides measured in the experiment without glucose addition are compared with the results from the standard glucose perturbation in Figs 1113. It was calculated that all known mass action ratios (PGI, PGM, phosphomannose isomerase [PMI], enolase, and fumarase) were maintained close to equilibrium (data not shown).

Figure 11

Comparison of concentration profiles of intracellular glycolytic metabolites, mannose 6-phosphate, 6-phosphogluconate, trehalose 6-phosphate and trehalose, before and during fully aerobic glucose-only and no-glucose perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

Figure 12

Intracellular tricarboxylic acid (TCA) cycle metabolites concentration profiles before and during fully aerobic glucose-only and no-glucose perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

Figure 13

Intracellular adenine nucleotide (ATP, ADP, and AMP) concentration profiles, sum of nucleotides, cellular energy charge and adenylate mass action ratio during fully aerobic glucose-only and no-glucose perturbation of low-cell-density aerobic glucose/ethanol-limited chemostat culture of Saccharomyces cerevisiae CEN.PK 113-7D in the BioScope. Error bars represent standard deviations between duplicate analyses of the same sample.

It can be seen from Figs 11 and 12 that the concentrations of most of the metabolites slowly but constantly decrease in the absence of glucose supply. Exceptions are 2/3PG and PEP, for which the concentrations remain more or less constant during the first 60 s. This could be caused by the steady decrease of the ATP level in the equilibrium between F-1,6-BP and 2/3PG. Also, the lower F-1,6-BP levels lead to less stimulation of pyruvate kinase.

Another exception is glycerol 3-phosphate, which hardly changes, showing the absence of redox stress. With respect to the TCA cycle intermediates, slightly increased initial levels of the C4 pool metabolites, (malate, fumarate and succinate) were observed, which might be related to the higher CO2 level in the experiment and the near-equilibrium reaction of pyruvate carboxylase, which converts pyruvate, CO2 and ATP to the C4 pool via oxaloacetate.

In contrast to this, the adenine nucleotides (Fig. 13) show different patterns, with a slow decrease in ATP and increases in ADP and AMP. Furthermore, a slow decrease of the cellular energy charge is observed. Most notably, there is hardly any change in the sum of the adenine nucleotides, and the adenylate kinase equilibrium is maintained. Intracellular trehalose was also quantified during both perturbations (Fig. 11). It was found that there was no significant difference in steady state during a perturbation in both cases. It is remarkable that as well as the absence of glucose addition, no decrease in the trehalose concentration was observed. This is surprising, as S. cerevisiae would be expected to mobilize storage carbohydrates (trehalose and glycogen) under starvation conditions (François & Parrou, 2001; Gancedo & Flores, 2004; Guillou, 2004). It is concluded that the glucose starvation period of 180 s is not long enough to trigger mobilization of storage compounds, e.g. trehalose.

Conclusions

Glucose perturbation experiments carried out at different levels of O2 availability (fully aerobic to anaerobic) did not result in large differences in increase in the glycolytic flux or in the concentration profiles of the measured primary metabolites. The use of acetaldehyde as an additional external electron acceptor under fully aerobic conditions and glucose perturbation shows a highly significant impact on fluxes and the metabolite levels of glycolysis and the TCA cycle, and especially the adenine nucleotides. The results strongly indicate that the dynamic response of the adenine nucleotides is determined by the equilibrium coupling between the NADH/NAD+ and the ATP/ADP ratios via the equilibrium pool of F-1, 6-BP and 2/3PG. In the absence of glucose supply, most glycolytic and TCA cycle metabolite levels show a steady decrease. With respect to the adenine nucleotides, the overall decrease of the ATP level is similar in the presence and absence of a glucose pulse within the time frame of observation of 180 s. However, the levels of AMP and ADP, which decrease after a short initial increase in response to a glucose pulse, were observed to increase in the absence of glucose supply. As a result, the sum of the adenine nucleotides was found to decrease in response to a glucose pulse but remained constant in the absence of a pulse. It is remarkable that there is no indication of mobilization of trehalose under these conditions.

Acknowledgements

We would like to acknowledge Rob Kerste and Dirk Geerts for technical assistance, and Johan Knoll, Cor Ras, Jan van Dam and Max Zomerdijk for analytic work. Applikon BV, Schiedam, The Netherlands, DSM Anti-Infectives, Delft, The Netherlands and The Netherlands Organization for Scientific Research (NWO) are acknowledged for financing the project.

Appendix

Appendix

In the case of a constant O2 consumption rate, rO2 (mol m−3 s−1) in the BioScope, the O2 balance can be written as follows: Embedded Image (A1) Integration of Eqn (A1), using x=0, COL=COL (0), leads to the expression Embedded Image (A2) Using COL (0)=0.066 mol m−3 (DOT=27%), C*OL=0 (N2 gas in the gas channel of the BioScope), rO2=0.007 mol m−3 s−1, Di=1.2 × 10−3 m, Koverall=1.9 × 10−5 m s−1 and φv=2 mL min−1=3.33 × 10−8 m3 s−1 leads to Embedded Image Here, x is the BioScope channel length (m), and COL is mol O2 m−3.

This equation shows that COL=0 at x=0.45 m, corresponding to 6 s.

Footnotes

  • Editor: Barbara M. Bakker

References

View Abstract