A potential mechanism of energy-metabolism oscillation in an aerobic chemostat culture of the yeast Saccharomyces cerevisiae


K. Tsurugi, Department of Biochemistry 2, University of Yamanashi, Faculty of Medicine, 1110 Shimokato, Tamaho, Yamanashi, 409–3898 Japan
Tel/Fax: +81 55 2736784
E-mail: ktsurugi@yamanashi.ac.jp


The energy-metabolism oscillation in aerobic chemostat cultures of yeast is a periodic change of the respiro-fermentative and respiratory phase. In the respiro-fermentative phase, the NADH level was kept high and respiration was suppressed, and glucose was anabolized into trehalose and glycogen at a rate comparable to that of catabolism. On the transition to the respiratory phase, cAMP levels increased triggering the breakdown of storage carbohydrates and the increased influx of glucose into the glycolytic pathway activated production of glycerol and ethanol consuming NADH. The resulting increase in the NAD+/NADH ratio stimulated respiration in combination with a decrease in the level of ATP, which was consumed mainly in the formation of biomass accompanying budding, and the accumulated ethanol and glycerol were gradually degraded by respiration via NAD+-dependent oxidation to acetate and the respiratory phase ceased after the recovery of NADH and ATP levels. However, the mRNA levels of both synthetic and degradative enzymes of storage carbohydrates were increased around the early respiro-fermentative phase, when storage carbohydrates are being synthesized, suggesting that the synthetic enzymes were expressed directly as active forms while the degradative enzymes were activated late by cAMP. In summary, the energy-metabolism oscillation is basically regulated by a feedback loop of oxido-reductive reactions of energy metabolism mediated by metabolites like NADH and ATP, and is modulated by metabolism of storage carbohydrates in combination of post-translational and transcriptional regulation of the related enzymes. A potential mechanism of energy-metabolism oscillation is proposed.


alcohol dehydrogenase


aldehyde dehydrogenase


dissolved oxygen


energy-metabolism oscillation






glyceraldehyde-3-phosphate dehydrogenase


oxygen uptake rate


pyruvate dehydrogenase complex


protein kinase A

Biological rhythms are considered to be ubiquitous in eukaryotic and prokaryotic organisms and are synchronized so the organism can adapt to environmental changes. Circadian rhythm is the most common of the biological rhythms and the molecular clock mechanism has been studied extensively. However, no circadian rhythm has been discovered in yeast and instead two kinds of ultradian rhythms of energy metabolism have been reported. One is a KCN-induced oscillation of the glycolytic pathway and the other is an energy-metabolism oscillation (EMO) found in aerobic chemostat cultures. The KCN-induced oscillations were evoked after addition of glucose by inhibiting mitochondrial respiration with cyanide [1] and show a periodicity of 1–2 min as monitored by measuring the level of NAD(P)H. The glycolytic pathway has been physically proven to be a sustained oscillator by Prigogine and co-workers [2,3]. Theoretically, the glycolytic pathway oscillates under the primary control of phosphofructokinase 1 (Pfk1p) which is activated auto-catalytically by its own product ADP leading to a nonlinear accumulation of NADH in combination with glyceraldehyde-3-phosphate dehydrogenase (GAPDH). NADH acts as the feed-forward activator of a glycolytic pathway facilitating the fermentation and ATP produced by fermentation acts as a feedback inhibitor by inhibiting the kinase reaction of the enzymes hexokinase and Pfk1p. Experimentally, the ATP level oscillated with an inversed phase to that of the NADH level, supporting the theory [4]. Although the NADH oscillation ceases within 30 min in batch cultures, it lasts as long as several days in chemostat cultures [5] proving that the glycolytic pathway is a sustained oscillator comprising a dissipative structure.

The other ultradian oscillation called EMO is usually monitored by measuring the dissolved oxygen (DO) level and shows a periodicity of approximately 4 h [6–9], while a similar EMO with a short-period (about 40 min) has been studied [10,11]. EMO arises spontaneously under glucose- and nitrogen-limited conditions dependent on a high cell-density (∼ 5 × 108 cells·mL−1) [12] and is sustained synchronization of the cell division cycle and a periodic change in the factors involved in energy metabolism such as CO2 production, O2 uptake, glucose and ethanol concentrations, and amounts of storage carbohydrates [6–9]. EMO is considered to be a periodic change between respiratory and respiro-fermentative phases in which oxygen demands are relatively high and low, respectively. We previously reported that the NADH level oscillated in the same phase as the DO level [12] being high in the respiro-fermentative phase and low in the respiratory phase. However, we found later that the ATP level, which is supposed to act as a feedback inhibitor, was low in the respiratory phase [13], suggesting that EMO oscillates with a distinct, if similar, mechanism from the KCN-induced oscillation of the glycolytic pathway. On the other hand, in the short-period EMO, Lloyd et al. [14] found that the NADH level oscillated in the same phase as the DO level and Klevecz et al. [15] reported that the mRNA levels of most genes show peaks during the two distinct time windows within the reductive phase, and only 10% of genes are transcribed in the oxidative phase suggesting that gene expression is influenced by redox states of cells. They also reported that the oscillation become synchronized by the rhythmic secretion of metabolites like acetaldehyde and hydrogen sulfide [16]. Therefore, it seems likely that a periodic change of intracellular redox potentials and/or levels of some metabolites underlie the control of EMO. However, as even the central pathway of the energy metabolism contains many oxidoreductase reactions and many intermediates, the mechanism by which EMO oscillates remains largely unknown.

We isolated the gene GTS1 as a candidate rhythm-related gene from a yeast cDNA library using oligonucleotides that encoded an Ala-Gln repeat in Gts1p [17,18] and, during the course of studying the role of GTS1, we found that EMO is coupled with a fluctuation in the trehalose level and that the trehalose level began to increase in the early respiro-fermentative phase and decrease in the late respiratory phase after the elevation of the cAMP level [13,19]. Furthermore, we found that the transcription of TPS1 encoding trehalose-6-phosphate synthase 1 (Tps1p), a regulatory enzyme in trehalose synthesis, was periodically regulated in EMO peaking in the late respiratory phase [13,20] and that EMO was destabilized in a transformant in which the synthesis of trehalose was inhibited. So, considering that Tps1p has been reported to regulate glucose influx for glycolysis [21,22], we suggested that the metabolism of storage carbohydrates like trehalose is also involved in the stabilization of EMO [13]. Recently, Jules et al. [23] reported that yeast cells in a batch culture on trehalose generate transient short-period oscillations of respiration and cell division and that intracellular degradation of glycogen and trehalose plays an important role for the generation of EMO. However, the mechanism by which the regulation of trehalose and glycogen metabolism stabilize EMO remains to be elucidated.

In this report, to investigate the mechanism of EMO, we determined profiles of the levels of metabolic variables and found that EMO is the periodic change between reductive and oxidative states with respect to the NAD+/NADH ratio, thereby forming a feedback loop of dehydrogenase reactions. Furthermore, we found that the synthesis and breakdown of storage carbohydrates play an important role in regulation of EMO.


Energy metabolism in the respiro-fermentative phase of EMO

In our chemostat culture, the DO oscillation shows a periodicity of about 4.5 h and the oxygen uptake rate (OUR) in the respiratory phase was estimated to be twice that in the respiro-fermentative phase, showing fairly rapid transitions between the two phases (Fig. 1A). The glucose consumption rates in respiration per cell (fmol glucose·min−1·cell−1) were estimated to be 0.03–0.04 in the respiro-fermentative phase and 0.07–0.08 in the respiratory phase. On the other hand, the NAD(P)H level of culture samples, which was measured fluorimetrically, showed a sharp decrease in the early respiratory phase, (Fig. 1A) suggesting a change in the redox state in the cells. As NADH is claimed to be a feed-forward activator of the glycolytic pathway [1–3], we first determined the change of the intracellular level of NADH by the enzyme cycling method. The NADH level increased in the early respiro-fermentative phase and decreased slowly until the beginning of the respiratory phase when it sharply decreased causing an increase in the NAD+/NADH ratio, as the levels of total NAD cofactor were almost constant during EMO (Fig. 1B). The result suggested that NADH was rapidly consumed on the activation of some dehydrogenases to reduce substrates in the early respiratory phase. In agreement, the level of ethanol, the main reductive product of glycolysis including the ethanol production in yeast, increased rapidly with the decrease of the NADH level peaking in the early respiratory phase while the ethanol concentration remained low (1 mm or 0.05 fmol·cell−1) during the respiro-fermentative phase (Fig. 2A). The level of glycerol, another main product of glycolysis, also increased together with ethanol, although it started to increase earlier than that of ethanol in the late respiro-fermentative phase (compare the time points indicated by arrows in Fig. 2A,B). However, the maximal amount of glycerol in the culture was estimated to be 1/20th that of ethanol, as the extracellular concentration of glycerol was minimal (∼ 1/40), different from ethanol, to which the cell membrane is highly permeable. Therefore, NADH is mostly used to produce ethanol, supporting the notion that NADH acts as a feed-forward activator of glycolysis. It should be added that the acetaldehyde level was low and almost constant during EMO (Fig. 2A), suggesting that acetaldehyde does not play an important role in the regulation of EMO different from the cases of the short-period EMO [16] and KCN-induced glycolytic oscillation [24].

Figure 1.

Time courses of the change in oxygen uptake rate (OUR) and NADH levels during energy-metabolism oscillation (EMO). S. cerevisiae S288C was continuously cultured at a dilution rate of 0.1 h−1 in a synthetic medium added with 1% glucose and 0.1% yeast extract at 30 °C, pH 5.0. Energy metabolism oscillation was monitored by measuring dissolved oxygen (DO) using an oxygen electrode. (A) The NAD(P)H level (O) and OUR (•) estimated from the DO level (halftone line). The NAD(P)H level of culture samples was determined by measuring fluorescence by a spectro-fluorimeter (excitation 316, emission 400 nm) after diluting culture samples to 107 cells·mL−1 with distilled water. OUR was calculated on KLa, the volumetric oxygen transfer coefficient and the oxygen concentration in medium. R-F and R indicate approximate periods corresponding to the respiro-fermentative and respiratory phases, respectively. (B) The intracellular levels of NADH (○) and the sum of NAD+ and NADH (•), and the NAD+/NADH ratio (bsl00084). The NADH level was determined by the enzymatic cycling method. A solution of 0.05 m NaOH/1 mm EDTA (500 µL) was added to 20 µL of culture samples and the extract was divided into two 200-µL aliquots. One of the aliquots was neutralized by adding 100 µL of a solution containing 100 mm Tris/HCl, pH 8.1, and 0.05 m HCl and used to determine the sum of NAD+ and NADH levels. The remaining alkali extract was incubated at 60 °C for 30 min to destroy NAD+ and neutralized with the same solution to determine the NADH level. Neutralized cell extract (5 µL) was used for the assay of the enzymatic cycling method. The halftone line indicates DO.

Figure 2.

Time courses of the change in concentrations of some metabolic intermediates during EMO. Sample of cell lysate was deproteinized with perchloric acid and supernatant was neutralized with KOH before assay. The concentrations of ethanol, acetaldehyde, glycerol and acetate were determined by measuring the fluorescence of NAD(P)H produced or consumed enzymatically after incubation of culture samples with substrate-specific dehydrogenases using kits. (A) The extracellular ethanol (○) and acetaldehyde levels (•). An arrow indicates the approximate time point when the ethanol level increased to an intermediate value between the minimal and the maximal level during EMO. (B) The intracellular (○) and extracellular glycerol levels (•). An arrow indicates an approximate time point when the intracellular glycerol level increased to its intermediate value. (C) The intracellular (○) and extracellular acetate levels (•). An arrow indicates an approximate time point when the intracellular acetate level increased to its intermediate value. Halftone lines in (A) to (C) indicate DO. R-F and R indicate approximate periods corresponding to the respiro-fermentative and respiratory phases, respectively.

Role of storage carbohydrates in EMO

At the transition to the respiratory phase of EMO, the glucose influx into glycolysis has to be rapidly facilitated to increase respiration twice (Fig. 1A). To address this issue, we determined the glucose flux rate from storage carbohydrates (Fig. 3), as we knew that the glucose level in the medium remained too low to be detected (< 10 µm) throughout EMO [13]. Thus, it cannot be supplied from the medium. The time course of the change in storage carbohydrate levels showed that glycogen and trehalose persistently accumulated in the respiro-fermentative phase and three-times more glycogen than trehalose accumulated by the end of the phase (Fig. 3A). The synthesis of storage carbohydrates started early in the respiro-fermentative phase, reached a plateau after 1 h of the start and accelerated again 1 h before the beginning of breakdown in the late respiro-fermentative phase, as reflected in the biphasic influx rate (Fig. 3B). The flux rate of glucose into storage carbohydrates (0.04 fmol glucose·min−1·cell−1) is almost the same as the consumption rate of glucose by respiration (0.03–0.04) in the early respiro-fermentative phase. On the other hand, breakdown started in the late respiro-fermentative phase when the cAMP level increased, and continued throughout the respiratory phase (Fig. 3A). The rate at which glucose was released from storage carbohydrates was estimated to be about 0.1 fmol·min−1·cell−1 (Fig. 3B), which is high enough to make the intracellular glucose level 10 times that observed (1.55 mm or 0.08 fmol·cell−1 on average) if the released glucose remained intact for 1 min in the cytoplasm. The fact that the intracellular glucose level did not change significantly, despite the breakdown of storage carbohydrates (Fig. 3B), indicated that most of the glucose fluxed into the glycolytic pathway.

Figure 3.

Time courses of the change in concentrations of intracellular glucose, storage-carbohydrates and biomass during EMO. (A) The intracellular concentrations of trehalose (○), glycogen (•) and cAMP (bsl00084). Trehalose and glycogen levels were determined according to a method described elsewhere [55]. The cAMP level was determined using the cAMP enzyme-immunoassay system (Amersham Pharmacia Biotech) [17]. The concentrations of trehalose and glycogen were expressed in glucose equivalents per cell. (B) The glucose-flux rate related to storage carbohydrates estimated from the changes in trehalose and glycogen levels (○) in (A) and the intracellular glucose level (•). Sample of cell lysate were deproteinized with perchloric acid and supernatant was neutralized with potassium hydroxide (KOH) before assay. The concentration of intracellular glucose was determined by measuring the fluorescence of NADPH produced enzymatically after incubation of samples with substrate-specific dehydrogenases using a kit. Glucose-flux rates associated with storage carbohydrates in a cell (fmol glucose·min−1·cell−1) were calculated as the difference of the sum of trehalose and glycogen (fmol glucose·cell−1) between two samples divided by the sampling interval (30 min). Positive values for the glucose flux rate indicate the flux of glucose into the energy-metabolism pathway and negative values indicate that into storage carbohydrates. Numbers indicate the averages of the glucose flux rates in the early respiro-fermentative and respiratory phases. (C) The total biomass of the culture (○) and that with storage-carbohydrate mass subtracted (•), and budding rate (bsl00084). Biomass was determined using nitrocellulose filters as described previously [56]. A solid line indicates the total mass of storage carbohydrates and a halftone line indicates DO.

It should be pointed out that the increase in the flux of glucose also causes the increase in the NADH level via the GAPDH reaction in the glycolytic pathway. However, the NADH level was even decreased in the respiratory phase (Fig. 1), suggesting that NADH was rapidly reoxidized during the production of glycerol and ethanol (Fig. 2A,B), resulting in the increase of the NAD+/NADH ratio (Fig. 1).

Energy metabolism in the respiratory phase of EMO

In the respiratory phase, the accumulated ethanol and glycerol decreased with a rapid increase in the level of acetate which decreased by the end of the respiratory phase (Fig. 2C). The result suggested that ethanol and glycerol were oxidized to acetate as NAD+, an oxidative cofactor of the related dehydrogenases, was increased (Fig. 1) and that acetate was oxidized through the citric acid cycle in mitochondria after conversion to acetyl-CoA [25], as respiration was activated in this phase.

Under aerobic conditions, the oxidative decarboxylation of pyruvate catalyzed by the pyruvate dehydrogenase complex (PDC) is the main pathway connecting glycolysis with the citric acid cycle and PDC is known to be regulated positively by AMP and negatively by ATP and NADH. We previously reported that either the ADP/ATP or AMP/ATP ratio increased in the respiratory phase (Table 1) and therein found that the NAD+/NADH ratio increased (Fig. 1), suggesting that the metabolites coordinately activated PDC to facilitate respiration in combination with the accelerated glucose flux from the storage carbohydrates which persisted during the full respiratory phase (Fig. 3B). Furthermore, the biomass measurement showed that the biomass, except for storage carbohydrates, was mostly formed during the respiratory phase in parallel with the increase in the budding rate (Fig. 3C). Thus, these results suggested that the glucose released from the storage carbohydrates was used to activate catabolism through fermentation and respiration to support the biomass formation in the respiratory phase.

Table 1.  Changes of metabolic variables during EMO between the respiro-fermentative (R-F) and respiratory phases (R) in the wild type. The levels were determined at 30-min intervals over three waves of EMO and those belonging to the respiro-fermentative and respiratory phases were used for calculation. The intracellular levels of ATP, ADP and AMP were determined by high performance liquid chromatography as described previously [13], and OUR and the NADH level were as described in Fig. 1.
 R-F phaseR phaseR/R-F
  1. a The units for the NADH level are µmol ·µg−1 protein. b The units for OUR are µmol·min−1·L−1 of culture medium.

ADP/ATP ratio0.534 ± 0.062 (n = 18)0.659 ± 0.052 (n = 12)1.25
AMP/ATP ratio0.195 ± 0.035 (n = 18)0.289 ± 0.060 (n = 12)1.48
NADH level a0.297 ± 0.028 (n = 7)0.168 ± 0.061 (n = 6)0.57
OUR b120 ± 6.1 (n = 10)213 ± 90 (n = 6)1.78

To confirm that the increase in the supply of glucose triggered the transition to the respiratory phase, we examined whether the phase transition can be triggered prematurely by adding glucose to the medium before the increase in the breakdown of storage carbohydrates. Glucose was injected into the medium at a final concentration of 1 mm which corresponds to the amount supplied by the storage carbohydrates for 20 min in the respiratory phase (Fig. 4). The result showed that, when glucose was injected 30 min prior to the transition, a complete profile of the respiratory phase was obtained, while this was not the case when it was injected 60 and 90 min before the transition (Fig. 4A). On the other hand, ethanol added at 2 mm did not trigger the respiratory phase when injected at 30 min before the transition (Fig. 4B). Thus, the increased influx of glucose plays an important role in the transition in a limited phase of EMO, that is, the late respiro-fermentative phase in which the level of cAMP is increasing.

Figure 4.

The induction of the respiratory phase by exogenous glucose and ethanol. (A) Glucose was added to the medium of the continuous culture of S288C at a final concentration of 1 mm at 30 (○), 60 (•) and 90 min (bsl00084) prior to the transition to the respiratory phase. (B) Ethanol was added to the medium at 2 mm 30 min before the transition (○). Halftone lines in (A) and (B) indicate DO in the culture without addition of glucose or ethanol.

Expression of genes related to the synthesis and breakdown of storage carbohydrates

We previously reported that the synthesis of trehalose was promoted by expressing TPS1 at an appropriate time to stabilize EMO, but whether or not the trehalase genes NTH1 and NTH2[26] are expressed in the late respiro-fermentative phase when they are required to function is unknown. The northern blot analysis showed that the mRNA levels of NTH1 and NTH2 increased about three- and sixfold, respectively, peaking at the late respiratory phase (Fig. 5), similar to the TPS1 mRNA [13], and suggesting that NTH1 and NTH2 were expressed out of the phase in which neutral trehalases are required to function. It should be pointed out that NTH2, which was little expressed in the exponentially growing phase [26], was expressed more actively than NTH1 and that the NTH2 RNA level remained high until the late respiro-fermentative phase (Fig. 5A), suggesting that NTH2 is the major neutral trehalase in cells showing EMO. Then, we determined the mRNA levels of GSY2 which encodes one of the two glycogen synthases mainly expressed in the stationary phase [27], and GPH1 which encodes the only enzyme involved in the breakdown of glycogen [28] (Fig. 5B). The result showed that GPH1 and GSY2 were periodically expressed in a similar phase of EMO between the respiratory and early respiro-fermentative phases. Thus, the genes involved in both the synthesis and breakdown of the storage carbohydrates were expressed in a similar phase of EMO.

Figure 5.

The mRNA levels of genes involved in the synthesis and degradation of the storage carbohydrates. Total RNA (20 µg) was electrophoresed on 1% agarose formaldehyde gel and hybridized with PCR-amplified and DIG-labeled cDNA probes. Northern blots were visualized with a lumino-image analyzer. (A) The mRNA levels of NTH1 (○) and NTH2 (•) encoding neutral trehalases during EMO. (B) The mRNA levels of GSY2 (○) and GPH1 (•) encoding glycogen synthase 2 and glycogen phosphorylase 1, respectively, during EMO. Northern blot analysis was performed in duplicate using samples from two independent cultures and averages of the levels relative to the ACT1 mRNA level were determined. The relative levels for each mRNA were also normalized so that the lowest levels were given a value of 1.0.

Continuous culture of gsy2Δ

In cells showing EMO, glycogen is a main storage carbohydrate and thought to play an important role in supplying glucose at the transition to the respiratory phase for stabilization of EMO. To confirm this presumption, we examined whether the GSY2-deleted transformant gsy2Δ shows a stable EMO in continuous culture. The growth of gsy2Δ was delayed in the batch culture reaching about half the critical density (∼ 5 × 108 cells·mL−1) by the start of the continuous culture and could not generate any oscillation (Fig. 6). However, as the culture continued, the cell density increased gradually and reached the critical density after about 18 h. Thus, the culture was restarted after stopping the supply of the fresh medium for 2 h as this procedure usually evokes EMO again in the culture of the wild-type and gts1Δ cells. However, the gsy2Δ culture showed only a few small waves (Fig. 6), supporting the notion that glycogen plays an important role in the stabilization of EMO.

Figure 6.

The DO pattern of gsy2Δ in the continuous culture. GSY2 encoding glycogen synthase 2 was deleted in S288C using a deletion cassette loxP-kanMX-loxP and the resulting transformant gsy2Δ was applied to a chemostat culture. A batch culture of gsy2Δ started at the time point 1 and ceased at 2 when all carbon sources were exhausted. After about 2 h, a continuous culture started at time point 3, continued until the cell density reached a critical level (∼ 5 × 108 cells·mL−1) at time point 4, and re-started at time point 5. Arrowheads indicate the time points when cell densities were measured and numbers indicate cell density in cells·mL−1.


In this communication, we studied on the mechanism of EMO in yeast and showed that the oscillation of the energy metabolism is basically regulated by the periodic change between reductive and oxidative states of NADH, thereby forming a feedback loop of dehydrogenase reactions in the energy metabolism. NADH produced from the upper part of the glycolytic pathway (from glucose phosphorylation to pyruvate production) activates ethanol production in the lower part (from pyruvate decarboxylation to ethanol production) and then the resulting increase of NAD+ activates oxidation of ethanol and mitochondrial respiration. We further showed that metabolism of trehalose and glycogen plays an important role in the regulation of EMO. The synthesis of them attenuates the glucose flux for glycolysis in the early respiro-fermentative phase and the breakdown induces the facilitation of respiration on the shift to the respiratory phase.

The potential mechanism of EMO is suggested as follows (Fig. 7). First, in the respiro-fermentative phase shown in Figure 7A, glucose is catabolized mainly by respiration and anabolized into storage carbohydrates at similar flux rates when calculated from OUR (Fig. 1A) and the storage carbohydrate-related glucose flux rate (Fig. 3B). Although this phase has been called ‘respiro-fermentative’, ethanol production is suppressed (Fig. 2A) so that reoxidation of NADH decreases. It has been proposed that a metabolic function of trehalose synthesis is to restrict the flux of glucose for glycolysis by inhibiting the early steps of the glycolytic pathway. The glycolytic pathway is designed based on the ‘turbo’ principle, as mutants having a defect in TPS1 cannot grow on glucose-accumulating glucose-phosphates until depletion of ATP, its own product, occurs [29]. Thus, it is likely that the attenuation of the early steps of glycolysis is caused, at least in part, by the flux of glucose into storage carbohydrates suppressing the overflow of pyruvate to the ethanol production which may lead to premature shift to the respiratory phase. In addition, the synthesis of trehalose is involved in the recovery of inorganic phosphate, which is required for the GAPDH reaction to produce NADH [21], via trehalose-6-phosphate phosphatase. The synthesis of glycogen and trehalose is facilitated by expression of synthetic genes encoding regulatory enzymes like Tps1p [13] and Gsy2p in this phase (Fig. 5B). Our result showed that the degradative enzymes of storage carbohydrates were also expressed in this phase, but were thought to remain inactive until the transition to the respiratory phase when they are activated by phosphorylation [20]. Although the production of trehalose increased earlier than that of glycogen in the early respiro-fermentative phase, glycogen was three times more abundant than trehalose, suggesting that glycogen is a main carbohydrate reservoir in cells during EMO (Fig. 3).

Figure 7.

Summary of energy metabolism during EMO. Schematic presentation of the metabolism of glucose in the respiro-fermentative (A), respiratory (C) and transitional phases (B,D). Numbers indicate the approximate flux rate of glucose expressed in fmol·min−1·cell−1. Enzymes involved are indicated by the genes encoding them: TDHs, GAPDHs; GDP1/2, Glycerol-3-phosphate dehydrogenases (DH); PYK1, pyruvate kinase; ADHs, alcohol DHs; GCY1, glycerol DH; ALDs, aldehyde DHs; PFK2, phosphofructokinase 2; TPS1, trehalose-6-phosphate synthase 1; GSY2, glycogen synthase 2; NTH1/2, neutral trehalases and GPH1, glycogen phosphorylase 1. Pi indicates inorganic phosphate. T-shaped arrows indicate inhibitory action.

Secondly, during the transition to the respiratory phase shown in Figure 7B, ethanol as well as glycerol were rapidly accumulated consuming NADH as a reductive force for related dehydrogenases; ethanol was produced through the reduction of acetaldehyde by NADH-dependent alcohol dehydrogenases (encoded by the ADH gene family). The total amount of glycerol was much less than that of ethanol (< 1/20) although the intracellular concentration of glycerol was higher than that of ethanol at the peaks. The result suggested that the role of glycerol as a carbohydrate reservoir is considered to be small but rather glycerol, known as a major osmolyte in cells, may increase the cellular ‘turgor pressure’ which is proposed to facilitate bud formation [30]. Alternatively, as either hyper- or hypo-osmotic pressure induces calcium pulse responses in cells [31,32], glycerol production may be involved in the signal-transfer pathway considering that calcium is known to act in coupling with the Ras-cAMP pathway [33]. The fact that glycerol production started earlier than that of ethanol in the late respiro-fermentative phase (Fig. 2), when the bud formation and cAMP production started (Fig. 3), supported these speculations. To start the acceleration of ethanol and glycerol production, the glucose influx into the glycolytic pathway has to be facilitated. We suggested that glucose was supplied from the storage carbohydrates mediated by cAMP which increased prior to this phase (Fig. 3A). The cAMP-dependent PKA was activated leading to the degradation of trehalose and glycogen by activating neutral trehalases (encoded by NTH1/2) and glycogen phosphorylase (GPH1). In contrast, the Ras/PKA signaling pathway has been reported to inactivate glycogen synthase [34] and the mutation of RAS2 in which the cAMP level decreased induced hyperaccumulation of glycogen [35]. Furthermore, the activity of purified glycogen synthase was inhibited after phosphorylation by PKA in vitro[36]. Thus, it is probable that the degradative and synthetic enzymes for glycogen were inversely regulated by PKA in this phase. Furthermore, a few lines of evidence have been reported which support the acceleration of the flux rate of the glycolytic pathway by cAMP in this phase. (1) In addition to the conventional proposal that Pfk1p is activated by its own product ADP in an autocatalytic fashion [2,3], phosphofructokinase 2 (Pfk2p encoded by PFK2) which produces fructose-2,6-bisphosphate (F-2,6-BP), a potent activator of Pfk1p, is exclusively dependent on PKA [37,38]. (2) The pyruvate kinase is reportedly activated by PKA in the presence of the activator fructose-1,6-bisphosphate (FBP) [39], increasing production of pyruvate. In this respect, Wittmann et al. [40] reported recently that FBP oscillated peaking at the early respiratory phase out of phase of 2- and 3-phosphoglycerate, and phosphoenolpyruvate, suggesting that FBP activates pyruvate kinase leading to progressively increasing glycolytic flux. These results suggested that the released glucose flowed into the glycolytic pathway leading to ethanol production which was facilitated by PKA, in combination with NADH which is not only an feed-forward activator for the ethanol production in glycolysis, but also an inhibitor of PDC suppressing the flux of puruvate into the citric acid cycle.

Thirdly, in the early respiratory phase (Fig. 7C), NADH was mostly oxidized to NAD+ as a result of the ethanol and glycerol production and the resulting increase in the NAD+/NADH ratio activated mitochodrial respiration in combination with the increased flux of glucose from storage carbohydrates for glycolysis. Ethanol and glycerol produced in the early respiratory phase are oxidized to acetate via acetaldehyde using NAD+ as an oxidative cofactor of glycerol-3-phosphate dehydrogenase (GPD1/2), alcohol dehydrogenases (ADHs) and aldehyde dehydrogenases (ALDs), and acetate is further oxidized in mitochondria via acetyl-CoA [25]. Alternatively, glycerol may be oxidized by mitochondria using FAD not to influence the NAD+/NADH ratio in the cytoplasm via the glycerol-3-phosphate shuttle or external NADH dehydrogenase on mitochondria [41]. As a result, OUR increased more than twofold that in the respiro-fermentative phase, suggesting that the internal environment of cells became oxidative. There have been reports that GAPDH (encoded by the TDH gene family), mildly oxidized by either H2O2 or nitric oxide, exhibits acyl phosphatase activity with which GAPDH directly produces 3-phosphoglycerate without ATP production, and thus facilitates glycolysis [42–44]. Therefore, it is possible that glycolysis in the respiratory phase is facilitated by this modification due to reactive oxygen species in addition to the cAMP action described above. The doubling of OUR in this phase means that ATP production was also doubled. However, the fact that the ATP level decreased while the AMP and ADP levels increased in the respiratory phase [13] suggested that ATP was consumed at a higher rate than it was produced. The decrease in the ATP level is required together with the increase in the NAD+ level to stimulate the activation of mitochondrial respiration via PDC. We found that ATP is mainly used for the biomass formation for budding (Fig. 3). The predominant formation of biomass in the respiratory phase of EMO is in good agreement with a previous result [9].

Finally (Fig. 7D), by the end of the respiratory phase, both ATP [14] and NADH levels (Fig. 1B) recover because the biomass formation and acetate oxidation were completed, and act as feedback inhibitors to suppress respiration in mitochondria. In addition, the trehalose and glycogen synthesis started attenuating the glucose influx into glycolysis, which promotes the shift to the respiro-fermentative phase.

It should be noted that EMO is a dissipative structure that spontaneously operates in chemostat cultures, obeying the second law of thermodynamics, which states that spontaneously occurring reactions increase entropy (energy specifying the amount of randomness or disorder) in the universe. Dissipative structures are defined as sustained oscillators operating far from equilibrium of energy according to the second law of thermodynamics [2,3]. In other words, they are oscillators that spontaneously operate by dissipating energy so that they self-organize vivid structures (or systems). The dissipation of energy means a series of reactions consisting of uptake of high-ordered (energy-rich) materials from the environment and subsequent decomposition of these materials into low-ordered materials releasing free energy and entropy, which are used for work and are excreted into the environment, respectively. Living organisms are complex examples of dissipative structures. EMO in yeast is a sustained oscillator that operates spontaneously along with the decomposition of glucose into carbon dioxide and water, synchronizing various metabolic pathways in an oscillatory fashion, including the cell-replication cycle, metabolism of storage carbohydrates and glycerol, gene expression, and so on, all of which require ATP, a product of EMO.

We have previously reported that in gts1Δ, in which expression of TPS1 is attenuated, EMO is destabilized and disappears [13], and it is indicated herein that GSY2-deleted transformants are unable to generate EMO (Fig. 6), suggesting that the normal synthesis of storage carbohydrates is required for stabilization of EMO. So far, it seems likely that trehalose synthesis functions as a primary source of inorganic phosphate, while glycogen is a primary source of glucose upon the transfer to the respiratory phase. However, whether the roles of glycogen and trehalose in the stabilization of EMO are distinct from one another remains to be clarified. In addition, although we suggested that cAMP/PKA is deeply involved in the control of EMO by regulating glycogen synthesis and various enzyme activities, it is likely that many other kinases are also involved. For example, PAS and Pho85p kinases inhibit glycogen synthase via phosphorylation and promote mRNA translation or the cell cycle [45]. Furthermore, it was recently reported that the TOR (target of rapamycin) and cAMP/PKA signaling pathways coordinately stimulate the transcription of ribosomal proteins and rRNA [46] which accounts for over 60% of all transcription. Studies on the involvement of these kinases in EMO should be conducted in the future.

In this communication, we for the first time presented a potential mechanism of EMO by systemically determining profiles of the levels of metabolic variables. It should be mentioned, however, that there have been several studies on the metabolism of some intermediates in EMO [8,40,47]. There are no significant differences between previous results and ours: for example, ethanol and acetate were produced and degraded in a sequential manner in the respiratory phase [8,47]; the cAMP level increased at the shift to the respiratory phase activating trehalases [47], and trehalose and glycogen were metabolized similarly peaking in the late respiro-fermentative phase [8]. On the other hand, there are some significant differences between the metabolism of EMO and the short-period EMO [10,11]. For example, the ethanol level did not oscillate significantly while the acetaldehyde level oscillated with a high amplitude and the trehalose level did not significantly oscillate [11]. So, possibly, there are some differences between the mechanisms of short- and long-period EMO.

Very recently, Tu et al. [48] reported the results of microarray analysis of gene expression in a long-period EMO using a diploid yeast strain CEN.PK. They showed that over half of genes were expressed periodically during EMO and that genes encoding proteins having a common function exhibit similar temporal expression patterns. Roughly, the genes involved in the protein synthesis and cell division are expressed in the respiratory (named the OX phase in their report) and early respiro-fermentative phases (R/B phase), respectively, and those involved in nonrespiratory modes of metabolism and protein degradation are in the late respiro-fermentative phase (R/C phase). Thus, it is likely that about half of genes were expressed in the phase when their encoded proteins are required to function. Furthermore, their supplemental data indicated that the genes encoding the synthetic and degradative enzymes of storage carbohydrates are all expressed in the late respiro-fermentative. Although the phase of expression of the genes seems earlier than that shown in this report, their result also indicated that post-transcriptional regulation is required for the enzymes to function properly. The problem whether and how the gene expression and post-transcriptional regulation are influenced by the redox change of EMO remains to be elucidated.

Experimental procedures

Yeast and chemostat culture

Saccharomyces cerevisiae S288C (MATαmal gal2 SUC2) was used. Chemostat (continuous) culture was performed using a bench-top fermenter type MDL-6C (Marubishi Bioengineering, Tokyo, Japan) with a constant volume of 500 mL. Cells were first batch-cultured in a synthetic medium as defined elsewhere [8] added with 1% glucose and 0.1% yeast extract (w/v) at 30 °C, pH 5.0. Then, continuous cultures were started 2 h after the consumption of the carbon source from the medium and continued at a dilution rate of 0.1 inflow (ml·h−1)/culture volume (ml) with the same medium [6,49]. The culture was aerated with a flow of 1 L·min−1 and stirred at an agitation speed of 420 r.p.m. Energy metabolism oscillation was monitored by measuring DO using an oxygen electrode.

Production of GSY2-deleted mutant

GSY2 encoding glycogen synthase 2 was deleted in S288C using a deletion cassette loxP-kanMX-loxP which was amplified by polymerase-chain reaction (PCR) directed on the template plasmid pUG6 (a gift from J. H. Hegemann, Washington University, St. Louis, MO, USA) [50] using synthetic oligonucleotides 5′-CGTGACCTACAAAACCATTTGTTATTCGAGACTGCGACTGAGGCAGCTGAAGCTTCGTACGC-3′ named Gsy2-Fknock 7 and 5′-CTGTCATCAGCATATGGGCCATCGTCGTCATCGTCAGCTGCAGGGCATAGGCCACTAGTGGATCTG-3′ named Gsy2-Rknock 2114 as the forward and reverse primers, respectively. Bold letters in the primers indicate the sequence homologous to pUG6 and standard letters in Gsy2-Fknock7 and Gsy2-Rknock2114 indicate the sequences homologous to the region from seven nucleotides downstream of the start codon and the upstream region of the stop codon of GSY2, respectively. S288C was transformed with the PCR product and GSY2-deleted transformants (gsy2Δ) were selected by culturing cells on a YPAD plate (1% yeast extract, 2% polypeptone, 40 µg·mL−1 adenine sulfate and 2% glucose, w/v) containing the kanamycin derivative G-418 at 500 mg·L−1[50]. Deletion of GSY2 was verified by PCR using Gsy2-Fknock7 and the kanR-specific primer Kan-R378 (5′-CAGGAACACTGCCAGCGCATC-3′).

Determination of the quantities of various metabolites

The NADH level was determined by the enzymatic cycling method as described previously [51–53]. The cycling steps were performed using malic and alcohol dehydrogenases and the cycled product, malate, was converted to NADH by adding an indicator reagent containing malic dehydrogenase and glutamate oxaloacetate transaminase [52]. The NADH generated from malate was measured fluorimetrically with a spectro-fluorimeter (Hitachi F-4500, Tokyo, Japan) (excitation 325 nm, emission 463 nm) and the NADH level in the cell samples were estimated based on a calibration curve obtained by adding known amounts of NADH (0, 5 and 10 µmol) to the mixture of the cycling reaction in the place of cell samples. The NAD+ level was calculated by subtracting the NADH level from the sum of the NAD+ and NADH levels, as the NAD+ level determined by the acid extraction (in 0.05 m HCl, at 60 °C for 30 min) [53] was fairly low suggesting some destruction of NAD+ during the treatment.

Ethanol, acetaldehyde, glycerol, acetate and glucose concentrations were determined enzymatically using kits (F-kit, numbers 176290, 668613, 148270, 148261 and 716251; JK International, Tokyo, Japan) according to the manufacturer's instructions. To determine the intracellular concentrations, a known number of cells (∼8 × 108 cells) were harvested, washed with ice-cold distilled water and added with 200 µL of 0.4 m perchloric acid containing silica beads. Cells were lysed by vortexing and supernatant was obtained after a centrifugation in the cold room. The supernatant was neutralized by adding 1/10 volume of 4.2 m KOH and the resulting supernatant was stored at −80 °C until use. To determine the extracellular concentrations, a known volume of medium was used.


The oxygen uptake rate (OUR) of cultures (RX in µmol·min−1·L−1) was first determined by turning off the air according to the following equation [54]:


where C is the concentration of dissolved oxygen (µmol· L−1); X, the concentration of cells (cells·L−1) and R, the oxygen uptake rate of a cell. However, as the DO level oscillates, OUR could not be measured when the DO level decreased. So, OUR at a defined time Tn (min) during the DO oscillation was tentatively calculated according to the following equation:


where KLa is the volumetric oxygen transfer coefficient (1.33 min−1) determined as described previously [54] and C*, the oxygen concentration at saturation (0.228 µmol·L−1). OUR calculated with the equation was similar to values measured by turning off the gas at the time points with high DO levels.

The glucose consumption rate through respiration (fmol glucose·min−1·cell−1) was calculated from OUR with the following reaction formula:


To estimate the intra- and extracellular concentrations of metabolites, necessary parameters were determined: average cell volume, 50 µm3; wet weight, 45 pg·cell−1 and protein content, 4.5 pg·cell−1. The cell volume was determined as described previously [17] and protein amount was measured with a protein assay kit (Bio-Rad, Hercules, CA, USA).

Northern blot analysis

Total RNA was isolated with Isogen (Nippon Gene, Tokyo, Japan) and 20 µg of RNA was electrophoresed on 1% agarose formaldehyde gel. Probes for hybridization to mRNAs of NTH1/2 encoding neutral trehalase 1 and 2, GPH1 encoding glycogen phosphorylase 1, GSY2 encoding glycogen synthase 2 and ACT1 encoding actin were PCR-amplified directed on genomic DNA using forward (F) and reverse primers (R) (Table 2). Probe labeling and hybridization were performed with the PCR DIG-labeling mix and DIG system according to the manufacturer's instructions (Roche, Mannheim, Germany). The mRNA levels were normalized to the level of ACT1 mRNA as a control.

Table 2.  Synthetic oligonucleotides used for primers of PCR. F and R in the primer names indicate forward and reverse primers, respectively, and numbers indicate the first and last nucleotide numbers of each gene in the F and R primers, respectively.
Primer nameNucleotide sequence (5'- to 3')


We thank S. Akiyama, Center for Life Science Research, University of Yamanashi, for his valuable advice in the physical aspect of this work.