In the present work comparative 13C metabolic flux analysis of S. cerevisiae at different defined physiological states revealed that the transition from oxidative to fermentative utilization of glucose is reflected by substantial flux changes throughout the whole central carbon metabolism. The different physiological conditions, i. e. purely oxidative, mixed respiro-fermentative and mainly fermentative growth were established by variation of the dilution rate in chemostat culture. Hereby, the extent of the fermentation could be precisely adjusted and run into metabolic and isotopic steady-state.
We observed the previously described transition from respiration to fermentation as indicated by the redirection of flux from the TCA cycle towards the fermentation pathways [8
]. It becomes, however, clear from the present study, that the regulatory changes linked to the metabolic transition were not restricted to these reactions, but affected a number of additional pathways.
(i) The metabolic shift had a strong effect on the flux partitioning between glycolysis and PPP. Whereas, at purely oxidative metabolism, the major carbon flux was channeled into the PPP, the opposite was observed for fermentative metabolism. The data of the present work together with previous flux studies of S. cerevisiae
under various conditions clearly show that the flux partitioning between glycolysis and PPP is growth rate dependent (Figure ). A similar relation is found for the flux partitioning and the biomass yield. The reduced biomass yield at higher growth rates results in a reduced demand for erythrose 4-phosphate, ribose 5-phosphate and NADPH, respectively, which are all supplied by the PPP [45
]. In the present work a significant amount of carbon was recycled from the PPP into glycolysis via fructose 6-phosphate and glyceraldehyde 3-phosphate by the reversible reactions of the PPP. This indicates that the demand for NADPH, but not for carbon precursors, determined the flux towards the PPP. A detailed discussion of the NADPH metabolism is given below. Previous enzymatic studies of S. cerevisiae
in chemostat culture revealed that the in vitro
the capacity, of glucose 6-phosphate dehydrogenase slightly increased with the dilution rate between 0.1 h-1
and 0.4 h-1
]. The reduced PPP flux is therefore not due to a lack of capacity of glucose 6-phosphate dehydrogenase. In comparison to other studies the relative PPP flux observed under purely oxidative growth was slightly higher [27
]. Due to the fact that this flux parameter could be determined with relatively high precision (Table ) it seems likely that differences in cultivation conditions or the used strains are responsible for this.
Figure 4 Correlation of specific growth rate and relative carbon fluxes in the central metabolism of S. cerevisiae: glucose 6-phosphate dehydrogenase (A), citrate synthase (B), and malic enzyme (C). Data shown comprise fluxes of S. cerevisiae ATCC 31267 determined (more ...)
(ii) Pyruvate decarboxylase, the key enzyme of fermentative metabolism, was found to already operate at a high activity under conditions in which alcoholic fermentation was absent (Figure ). The substantial amount of acetaldehyde formed under these conditions was completely converted into cytosolic acetyl CoA which is either used for anabolism or transported into the mitochondrion. These findings disprove the common perception that only a small fraction of pyruvate is channeled through the fermentative by-pass at low glycolytic flux [30
], but rather support previous speculations on fluxes around the pyruvate node linked to the Crabtree effect in S. cerevisiae
]. Based on enzyme profiles, substrate affinities and intracellular pyruvate level studied in chemostat cultivation of S. cerevisiae
at varied dilution rate it was previously supposed that pyruvate decarboxylase might be active even during purely oxidative metabolism leading to the formation of cytosolic acetyl CoA and its transport into the mitochondrion. The increased pyruvate level observed during fermentative growth [7
] could also be an explanation for the observed flux redirection from mitochondrial to cytosolic alanine biosynthesis.
(iii) We found evidence for substantial in vivo
activity of malic enzyme, as was previously reported for S. cerevisiae
]. Additionally we could show a strong flux increase for malic enzyme, when the cells switched from pure oxidation to fermentative metabolism. The flux through malic enzyme increased almost linearly with increasing growth rate (Figure ). Although the present work and previous studies differ, concerning e.g.
the investigated strain, the pH value or the nutrient status, the flux through malic enzyme is generally increased at high growth rate. As exception only a minor flux of 7 % was observed for malic enzyme in a batch culture of S. cerevisiae
grown on glucose [27
]. This might arise from the above mentioned differences in the experimental setup or the generally high uncertainty for the quantification of this flux parameter (Table ). The exact physiological role of malic enzyme in S. cerevisiae
so far remained quite intriguing [27
]. Also this study does not really provide a satisfying insight into the function of malic enzyme. What at least can be stated is that the flux through malic enzyme seems to be related to the flux through the cytosolic pyruvate carboxylase and the oxaloacetate transporter. One might conclude that the increased activity of malic enzyme results indirectly from the flux redirection at the cytosolic pyruvate node.
(iv) The obtained fluxes allow a detailed inspection of the NADPH metabolism during the metabolic shift. Hereby, it should be noted that the performed flux analysis did not include a balance for NADPH. The fluxes reactions through NADPH supplying reactions were determined only from the fit of the labeling data and the metabolite balances given below. Generally S. cerevisiae
can recruit different enzymes for the formation of NADPH involving the glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase in the PPP [31
], NADPH specific acetaldehyde dehydrogenase [32
], and NADPH specific isocitrate dehydrogenase [33
]. Also malic enzyme is potentially capable to supply NADPH [34
]. Clear differences resulted for the different physiological conditions (Figure ). Linked to the strong decrease of the biomass yield, the anabolic NADPH demand decreased with increasing growth rate. To some extent this is attenuated by the higher protein content of S. cerevisiae
at high growth rate and the correspondingly increased stoichiometric NADPH demand (Tables , ). We additionally found that sources of NADPH varied with the physiological state.
Figure 5 NADPH metabolism of S. cerevisiae cultivated in chemostat under aerobic glucose-limited conditions at different growth rates. The fluxes correspond to purely oxidative (μ = 0.15 h-1), respiro-fermentative (μ = 0.30 h-1), and mainly fermentative (more ...)
Macromolecular composition of Saccharomyces cerevisiae in glucose-limited chemostat culture at different dilution rates. The data given in g (g cell dry mass)-1 are calculated from literature data [27, 47-49].
During purely oxidative growth the PPP alone can account for the total anabolic NADPH demand. Under these conditions an additional NADPH supply by other reactions seems not necessary. Because no intercompartmental transport system for NADPH has been described for S. cerevisiae
, it is likely that also mitochondrial enzymes are involved in the supply of NADPH. A potential candidate is mitochondrial NADPH specific isocitrate dehydrogenase [33
]. This enzyme can, however, account only for a certain fraction of the total flux through the TCA cycle. An exclusive activity of NADPH specific isocitrate dehydrogenase can be excluded. Assuming this the total amount of NADPH supplied would be almost twice as high as the anabolic demand. Thus we conclude that to large extent mitochondrial NADH specific isoenzymes of isocitrate dehydrogenase contribute to the TCA cycle flux during oxidative growth of S. cerevisiae
Under predominantly fermentative utilization of glucose only about 60 % of the totally required NADPH was provided by the PPP. This is an indication for an additional cytosolic NADPH source required during fermentative metabolism in S. cerevisiae
. S. cerevisiae
possesses several isoenzymes for acetaldehyde dehydrogenase with different specificity for NADH and NADPH [35
]. Possibly the relative contribution of the different isoenzymes changes with the on-set of fermentation, so that instead of mainly NADH under oxidative growth, also NADPH is formed. Recently it was found that especially the NADPH dependent acetaldehyde dehydrogenase isoforms are involved in anaerobic fermentation of S. cerevisiae
]. The increased NADPH formation by acetaldehyde dehydrogenase would probably lead to an increased level of intracellular NADPH, known to inhibit the PPP enzymes glucose 6-phosphate dehydrogenase and 6-phosphogluconate dehydrogenase [31
]. The reduction of the PPP flux might therefore not only be caused by the reduced anabolic demand, but also by the activation of alternative cytosolic pathways for NADPH supply.
An insight into the limiting reactions, involved in the complex flux redirection during the metabolic shift, can be derived from absolute carbon fluxes. As can be calculated from the relative fluxes and the specific glucose uptake rate (Figure ) the transition from oxidative to fermentative metabolism in S. cerevisiae
was accompanied by a more than 5-fold increase of the specific glucose uptake flux. It is interesting to see, how the increased influx of carbon is distributed throughout the metabolic network. The absolute carbon flux of enzymes involved in fermentation and the transport of cytosolic acetyl CoA into the mitochondrion are shown in Figure . Pyruvate decarboxylase exhibits a tremendous flux increase with increasing growth rate. A completely different picture results for the enzymes catalyzing the subsequent metabolic steps, i.e.
acetaldehyde dehydrogenase, acetyl CoA synthase and the transporter. The absolute carbon flux through acetyl CoA synthase remained constant for all physiological states investigated. Obviously this enzyme was already working at maximum capacity during purely oxidative growth. The saturation of acetyl CoA synthase obviously initiated the secretion of acetate into the medium with increasing growth rate. Additionally a strong decrease of the intracellular acetyl CoA level appears very likely. The insufficient activity of this enzyme had also consequences for the intercompartmental acetyl CoA transport. As shown, the absolute carbon flux through the transporter decreased to practically zero (Figure ). The cytosolic acetyl CoA was in fact completely channeled into anabolic reactions. Probably the reactions withdrawing cytosolic acetyl CoA into anabolism are favored by a higher affinity of the catalyzing enzymes. The insufficient capacity of acetaldehyde dehydrogenase and acetyl CoA synthase could even limit the supply of cytosolic acetyl CoA for anabolism, i.e.
the formation of lysine and lipids. In this regard a significant decrease of the in vitro
activity, i. e. the capacity, of acetaldehyde dehydrogenase and acetyl CoA synthase was previously observed during the metabolic shift from oxidative to fermentative metabolism [9
]. It was concluded that probably these enzymes display the limiting steps that cause the secretion of ethanol and acetate at high growth rate. The present study provides further evidence that indeed these reactions are key points with respect to increased production of ethanol and acetate. The same principle can be applied to study the reactions downstream of the pyruvate pool (Figure ). The absolute carbon flux of pyruvate dehydrogenase was low under purely oxidative growth, but showed no further increase towards higher growth rate, despite the carbon flux entering the cytosolic pyruvate pool, and thus potentially available, was much higher. This indicates that pyruvate dehydrogenase already operates at maximum capacity during respiro-fermentative growth at μ = 0.3 h-1
. The excess carbon entering the cytosolic pyruvate pool is obviously directed towards pyruvate carboxylase and fermentation. Probably also the activation of cytosolic alanine synthesis is related to this phenomenon. Pyruvate dehydrogenase thus displays an additional bottleneck. The insufficient capacity of pyruvate dehydrogenase seems to have an influence on the transport of pyruvate into the mitochondrion. Due to the fact that, during fermentative growth, a substantial fraction of mitochondrial pyruvate is already supplied via malic enzyme, the absolute carbon flux of the pyruvate transport is lower (Figure ). The absolute carbon flux through glucose 6-phosphate dehydrogenase increased about 1.5-fold when the dilution rate was increased from 0.15 h-1
to 0.4 h-1
. In a previous study it was shown that, linked to the same shift in dilution rate, the intracellular level of this enzyme was increased about 1.6-fold [9
]. Obviously, both glucose 6-phosphate dehydrogenase level and NADPH level are involved in the regulation of carbon flux into the PPP. In case of acetaldehyde dehydrogenase and acetyl CoA synthase the limited carbon flux at high dilution rate seems to be related to a decrease in enzyme level [9
]. Additionally inhibitory effects, e.g.
of acetaldehyde on acetaldehyde dehydrogenase [9
], might be involved. The present study suggests that the flux redirection, caused by the regulatory network linked to the Crabtree effect, leads to a saturation of the capacity, i.e.
the absolute flux, of acetaldehyde dehydrogenase, acetyl CoA synthase, and pyruvate dehydrogenase. It is the limitation of these enzymes that is then mainly responsible for the onset of the production of fermentative by-products during growth of S. cerevisiae
at high growth rate.
Figure 6 Absolute carbon fluxes of S. cerevisiae cultivated in chemostat under aerobic glucose-limited conditions at different growth rates: Fluxes of pyruvate decarboxylase, acetaldehyde dehydrogenase and acetyl CoA synthase (A), and of intercompartmental pyruvate (more ...)