EUKARYOTIC CELL, Feb. 2003, p. 170–180 Vol. 2, No.

1
1535-9778/03/$08.00⫹0 DOI: 10.1128/EC.2.1.170–180.2003
Copyright © 2003, American Society for Microbiology. All Rights Reserved.

Metabolic-Flux Profiling of the Yeasts Saccharomyces cerevisiae

and Pichia stipitis
Jocelyne Fiaux,1 Z. Petek Çakar,2† Marco Sonderegger,2 Kurt Wüthrich,1
Thomas Szyperski,3* and Uwe Sauer2
Institute of Molecular Biology and Biophysics1 and Institute of Biotechnology,2 ETH Zürich, CH-8093 Zürich,
Switzerland, and Department of Chemistry, State University of New York,
Buffalo, New York 142603
Received 13 August 2002/Accepted 28 October 2002

The so far largely uncharacterized central carbon metabolism of the yeast Pichia stipitis was explored in
batch and glucose-limited chemostat cultures using metabolic-flux ratio analysis by nuclear magnetic reso-
nance. The concomitantly characterized network of active metabolic pathways was compared to those identified
in Saccharomyces cerevisiae, which led to the following conclusions. (i) There is a remarkably low use of the
non-oxidative pentose phosphate (PP) pathway for glucose catabolism in S. cerevisiae when compared to P.
stipitis batch cultures. (ii) Metabolism of P. stipitis batch cultures is fully respirative, which contrasts with the
predominantly respiro-fermentative metabolic state of S. cerevisiae. (iii) Glucose catabolism in chemostat
cultures of both yeasts is primarily oxidative. (iv) In both yeasts there is significant in vivo malic enzyme
activity during growth on glucose. (v) The amino acid biosynthesis pathways are identical in both yeasts. The
present investigation thus demonstrates the power of metabolic-flux ratio analysis for comparative profiling of
central carbon metabolism in lower eukaryotes. Although not used for glucose catabolism in batch culture, we
demonstrate that the PP pathway in S. cerevisiae has a generally high catabolic capacity by overexpressing the
Escherichia coli transhydrogenase UdhA in phosphoglucose isomerase-deficient S. cerevisiae.

The two yeasts Saccharomyces cerevisiae and Pichia stipitis
exhibit fundamentally different modes of metabolic regulation
in glucose-containing media. At high extracellular concentra-
tions of glucose, one observes simultaneous fermentation and
respiration (respiro-fermentative metabolism) in S. cerevisiae
at high growth rates even under fully aerobic conditions (11,
34, 53). In Crabtree-positive yeasts, such as S. cerevisiae, ele-
vated glucose concentrations induce the carbon catabolite re-
pression response (16), resulting in low levels of transcription
of genes involved in respiration and in the tricarboxylic acid
(TCA) cycle. In contrast, the Crabtree-negative P. stipitis ex-
hibits predominantly respirative metabolism even at high glu-
cose concentrations (33). These obvious differences in meta-
bolic regulation between the two yeasts provide motivation for
investigation of potential differences in the central carbon
pathways. Stable isotope labeling experiments appear to be a
promising approach, since it reveals in vivo activity of pathways
and reactions (10, 43). Biosynthetically directed fractional
(BDF) 13C labeling of amino acids can provide comprehensive
insight into central carbon metabolism, yielding a network of
active pathways and quantification of intracellular flux ratios
(36, 37, 44, 46); its use in the present study promises to expand
on the results of earlier labeling studies in yeast that focused
on intermediates or products of individual pathways or reac-
tions (14, 17, 20, 22, 41, 57).
BDF 13C labeling is achieved by growing cells on mixtures of
* Corresponding author. Mailing address: Department of Chemis-
try, State University of New York, Buffalo, NY 14260. Phone: (716)
645-6800, ext. 2245. Fax: (716) 645-7338. E-mail: szyperski@chem
.buffalo.edu.
† Present address: Department of Molecular Biology and Genetics,
Istanbul Technical University, TR-80626 Maslak, Istanbul, Turkey.
unlabeled and uniformly 13C-labeled [U-13C6]glucose (40, 44).
Using two-dimensional (2D) nuclear magnetic resonance
(NMR) spectroscopy, the multiplet fine structures due to 13C-
13
C scalar coupling between adjacent carbons in proteinogenic
amino acids are analyzed in hydrolysates from such fractionally
labeled cells. Decomposition of the fine structures into their
multiplet components allows calculation of the relative abun-
dance of contiguous carbon fragments originating from a single
source molecule of glucose (44). Such intact carbon fragments
are then balanced within a metabolic network to identify the
network of active biosynthetic pathways and to determine
metabolic-flux ratios (44). Although it was initially conceived
for studies of bacterial metabolism, metabolic-flux ratio
(METAFoR) analysis has recently been applied also for the
analysis of eukaryotic metabolism (27), where the spatial sep-
aration of metabolic subnetworks in cytosol and mitochondria
as well as the intercompartmental transport fluxes of pyruvate
(PYR), acetyl coenzyme A (ACoA), and oxaloacetate (OAA)
were taken into consideration.
Here we use METAFoR analysis by NMR for metabolic
profiling of S. cerevisiae (27) and the much less well character-
ized xylose-fermenting yeast P. stipitis in both glucose-limited
chemostat and batch culture. The present investigation focuses
on the metabolic impact of glucose repression (the Crabtree
effect) and pentose phosphate (PP) pathway operation. To
gain further insight into the PP pathway, we further investi-
gated phosphoglucose isomerase (Pgi)-deficient S. cerevisiae,
which catabolizes glucose exclusively via the PP pathway (3).
MATERIALS AND METHODS
Strains and growth conditions. The laboratory yeast strains S. cerevisiae
CEN.PK 113.7D (MATaMAL2-8c SUC2) and P. stipitis CBS6054 were used for

170
VOL. 2, 2003 METABOLIC-FLUX PROFILING OF YEASTS 171

all labeling experiments. The former was obtained from P. Kötter (Institute of
Microbiology, Johann Wolfgang Goethe-University, Frankfurt, Germany) and
the latter was from B. Hahn-Hägerdal (Department of Applied Microbiology,
Lund University, Lund, Sweden). The strains ENY.WA-1A (MAT␣ ura3-52
leu2-3 trp1-289 his3-⌬1 MAL2-8c MAL3 SUC3) and EBY44 (ENY.WA-1A
pgi1-1⌬::URA3) (kindly provided by E. Boles, Institut für Mikrobiologie, Hein-
rich Heine Universität, Düsseldorf, Germany) were used for UdhA overexpres-
sion experiments. The soluble Escherichia coli transhydrogenase UdhA (4) was
cloned as a BamHI-HindIII fragment from an E. coli overexpression vector (7)
under the control of the constitutive, truncated HXT7 promoter of p425HXT7
(LEU2) (21) (provided by E. Boles). The resulting construct, p425-udhA, was
transformed in S. cerevisiae ENY.WA-1A and in its Pgi-deficient derivative.
Batch cultures of 100 ml were grown in 1-liter baffled shake flasks on a rotary
shaker at 30°C and 300 rpm in yeast minimal medium, which contained 0.5%
(wt/vol) glucose and 6.7 g of yeast nitrogen base without amino acids (Difco) per
liter. Uracil (50 mg/liter), tryptophan (50 mg/liter), histidine (50 mg/liter), and
leucine (250 mg/liter; to avoid hidden Leu limitations [6]) were supplemented
where necessary.
Glucose-limited chemostat cultures were grown in a 1.5-liter bioreactor (Bio-
engineering, Wald, Switzerland) at 30°C with a working volume of 1 liter. The
filter-sterilized chemostat medium contained the following components (per liter
of distilled water): glucose, 3.6 g; (NH4)2SO4, 5 g; MgSO4 䡠 7H2O, 0.5 g;
KH2PO4, 3 g; 1 ml of vitamin solution (biotin, 0.05 mg/liter; calcium pantothe-
nate, 1 mg/liter; nicotinic acid, 1 mg/liter; inositol, 25 mg/liter; thiamine-HCl, 1
mg/liter; pyridoxine-HCl, 1 mg/liter; para-aminobenzoic acid, 0.2 mg/liter); and 1
ml of trace element solution (EDTA, 15 mg/liter; ZnSO4 䡠 7H2O, 4.5 mg/liter;
CoCl2 䡠 6H2O, 0.3 mg/liter; MnCl2 䡠 4H2O, 1 mg/liter; CuSO4 䡠 5H2O, 0.3 mg/
liter; CaCl2 䡠 2H2O, 4.5 mg/liter; FeSO4 䡠 7H2O, 3 mg/liter; NaMoO4 䡠 2H2O; 0.4
mg/liter; H3BO3, 1 mg/liter; KI, 0.1 mg/liter) (51). To prevent foaming, 2 ml of
polypropylene glycol 2000 was added (1:10 diluted in H2O) per liter of medium.
Anaerobic cultures were supplemented with filter-sterilized Tween 80-ergosterol
solution (1.25 ml/liter) that contained 8 g of ergosterol and 336 g of Tween 80 per
liter of ethanol. The pH was maintained at 5.5 by the addition of 3 M KOH, and
the fermentation volume was kept constant by using a weight-controlled pump. FIG. 1. Biochemical reaction network for yeast central carbon me-
Aerobic conditions were achieved with a constant airflow of 0.5 liters/min and an tabolism. The arrows indicate reaction directionality. Letters in bold-
agitation speed of 1,200 rpm. To establish anaerobic conditions, air was substi- face type indicate metabolites for which the 13C-labeling pattern can be
tuted with nitrogen at the same flow rate and the agitation speed was reduced to accessed through METAFoR analysis. Abbreviations: G6P, glucose-6-
600 rpm. phosphate; F6P, fructose-6-phosphate; P5P, pentose-5-phosphates;
Analytical procedures. Cell growth was monitored spectrophotometrically by E4P, erythrose-4-phosphate; S7P, seduheptulose-7-phosphate; G3P,
determining the optical density of the cultures at 600 nm. Cellular dry weight glyceraldehyde-3-phosphate; PGA, 3-phosphoglycerate; ICT, isocitrate;
(cdw) was determined from six parallel 10-ml culture aliquots that were centri- OGA, oxoglutarate; SUC, succinate; MAL, malate; GOX, glyoxylate.
fuged for 20 min in preweighed glass tubes at 3,000 ⫻ g, washed once with water,
and dried at 90°C for 24 h to a constant weight. Concentrations of glucose and
fermentation products in the culture broth were determined with commercial of 13C labeling was in close agreement with the value calculated from the
enzymatic kits (Beckman) or by high-performance liquid chromatography. Phys- composition of the minimal medium, and was further confirmed by analysis of
iological parameters were calculated as described previously (37). the 13C scalar coupling fine structure of Leu-␤ (44).
13
C labeling experiments. In batch culture, BDF 13C-labeling of cellular amino R. Glaser’s program FCAL (version 2.3.0) (46) was used for integration of
acids was achieved by growth on a 5-g/liter glucose mixture consisting of 90% 13
C-13C scalar coupling fine structures observed for the resonances of 48 carbon
(wt/wt) unlabeled and 10% (wt/wt) uniformly labeled [U-13C6]glucose (degree of positions in the amino acids. The relative abundances, f, of intact carbon frag-
13
C labeling ⬎ 99% [wt/wt]; Martek Biosciences Corp., Columbia, Md.). These ments originating from a single source molecule of glucose were calculated from
cultures were inoculated with less than 1% (vol/vol) of a mid-exponential-phase the relative intensities, I, of the multiplets in the 13C-13C scalar coupling fine (27,
culture in minimal medium so that the presence of unlabeled biomass could be 43, 44). These f values (see the appendix for supplementary material) provide
neglected. Labeled biomass aliquots were taken from cultures in the mid-expo- information on the metabolic origin of the amino acid precursor molecules in the
nential phase at an optical density at 600 nm of 1 (37). central metabolism. Tracing intact labeled carbon fragments in these metabolites
Chemostat cultures in physiological steady state were BDF 13C-labeled by leads to identification of the active metabolic pathways and to quantification of
substituting unlabeled glucose in the feed medium with the above-mentioned the ratios of fluxes converging to one particular substrate (45). In cases where
mixture containing 10% (wt/wt) [U-13C6]glucose. Labeled biomass aliquots were only two reactions contribute to one metabolite pool and where one of the
withdrawn after about one culture volume change so that about 60% of the contributions is assessed from the NMR data, the remaining fraction of the total
biomass was fractionally 13C-labeled. pool can be attributed to the competing reaction(s).
Biomass aliquots of about 100 mg (cdw) were harvested from all experiments Biochemical reaction network for S. cerevisiae and P. stipitis. The biochemical
by centrifugation of the culture broth at 3,000 ⫻ g and 4°C for 10 min, washed reaction network considered for S. cerevisiae growing on glucose in batch culture
once with 20 mM Tris-HCl (pH 7.6), centrifuged again at the previous settings, was recently described (27). Eukaryotic compartmentation into mitochondrial
and resuspended in 3 ml of 20 mM Tris-HCl (pH 7.6). After addition of 6 ml of and cytosolic subsystems was included in the model by considering distinct pools
6 M HCl, hydrolysis was performed in sealed glass tubes at 110°C for 24 h, and of PYR, OAA, and ACoA in both compartments (Fig. 1). Firstly, the transport
the solutions were filtered through 0.2-␮m-pore-size filters (Millex-GP; Milli- of PYR into the mitochondria is driven by the proton motive force and was thus
pore), lyophilized, and dissolved in 700 ␮l of 0.1 M 2HCl in 2H2O for the NMR considered to be unidirectional (15). Secondly, although likewise proton driven,
measurements. the transport of OAA into the mitochondria is reversible (27, 32), and intercom-
NMR spectroscopy and data analysis. 2D proton-detected heteronuclear sin- partmental exchange between the two OAA pools may additionally occur via
gle-quantum 13C-1H correlation NMR spectroscopy (COSY) was employed to shuttle transport mechanisms of TCA cycle intermediates (1, 26, 31). Hence,
detect aliphatic and aromatic resonances of amino acids at a 1H resonance OAA transport was a priori considered to be bi-directional. Thirdly, mitochon-
frequency of 500 MHz on a Bruker DRX500 spectrometer as described previ- drial transport of ACoA was considered reversible because it is mediated by
ously (37, 44). The overall degree of 13C labeling was determined by integrating facilitated diffusion via the so-called carnitine shuttle (48). Although much less is
resolved 13C satellites in 1D 1H NMR spectra (27, 44). The thus-obtained degree known about central carbon metabolism of P. stipitis compared to that of baker’s
172 FIAUX ET AL. EUKARYOT. CELL

TABLE 1. Origins of metabolic intermediates during aerobic

exponential growth of P. stipitis and S. cerevisiae batch culturesf
% Fraction of total pool
Metabolite (mean ⫾ 2.5 SD)

P. stipitis S. cerevisiaea

Cytosolic
PEP derived through at least one TKb 57 ⫾ 9 0–4
(ub)
P5P from glucose (lb) 43 ⫾ 2 32 ⫾ 2
P5P from G3P and S7P (TK reaction) 57 ⫾ 2 68 ⫾ 2
P5P from E4P (TK and TAc reactions) 35 ⫾ 2 10 ⫾ 2
E4P from F6P (lb) 9⫾6 15 ⫾ 4
PEP from OAAcyt (PEP carboxykinase ⬍5 NDd
reaction)
OAAcyt from PYRcyt 41 ⫾ 3e 88–100e
OAAcyt reversibly converted to FUM at 24 ⫾ 10 0–4
least once (cytosolic or
intercompartmental exchange)

Mitochondrial
PYRmit from MAL ⬍6 25–30
OAAmit from PEP (anaplerosis) 36 ⫾ 2 76 ⫾ 4
OAAmit reversibly converted to FUM 58 ⫾ 12 35 ⫾ 4
at least once (in the TCA cycle)
a
Data taken from Maaheimo et al. (27).
b
TK, transketolase.
c
TA, transaldolase.
d
ND, not detectable because the fragment needed for tracing this activity is
absent.
e
Values calculated assuming absence of cytosolic OAA-to-FUM conversion.
f
The experimental error was estimated from the analysis of redundant 13C-13C
scalar coupling fine structures and the signal-to-noise ratio of the [13C,1H]-
COSY spectra employing the Gaussian law of error propagation. In certain
cases, the NMR data permit only the determination of upper (ub) or lower (lb)
bounds on the origin of metabolites. Abbreviations are explained in the legend
to Fig. 1.
yeast, we found no evidence in the literature that the metabolic network of S.
cerevisiae needs to be modified for P. stipitis. In fact, the extensive body of
13
C-labeling data obtained for P. stipitis in the present study was in agreement
with the network proposed for S. cerevisiae.
RESULTS
Central metabolism during growth in batch culture.
METAFoR analysis provides direct information on the biosyn-
thetic pathways for amino acids (23), which are required to
infer the 13C-labeling patterns of metabolic intermediates from
those detected in the amino acids. For P. stipitis we obtain the
result that amino acids are synthesized along the same path-
ways as in S. cerevisiae. As a continuation of our previous
investigation of S. cerevisiae (27), we first compared the results
of this earlier study with the flux ratios in central carbon me-
tabolism of exponentially growing P. stipitis in aerobic batch
culture. In these batch cultures with excess glucose concentra-
tions, the maximum specific growth rates were 0.4 and 0.3 h⫺1
for S. cerevisiae and P. stipitis, respectively. Using BDF 13C-
labeling, [13C,1H]-COSY, and METAFoR analysis, we quan-
tified for several intracellular metabolites the fraction of the
total pool that was derived from the specified substrates (Table
1). In turn, this quantifies the ratio of the fluxes that converge
to a particular metabolite (Fig. 2). In cases where only two
reactions contribute to one metabolite pool and where one of
contributions is assessed from the NMR data, the remaining
fraction of the total pool can be attributed to the competing
FIG. 2. Metabolic-flux ratios in S. cerevisiae (top values in the in-
sets) and P. stipitis (bottom values) during batch growth with glucose as
the sole carbon source. Connected ratios of fluxes that converge to a
particular metabolite are presented in identical geometrical shapes.
For abbreviations see the legend to Fig. 1.
reaction. Five key differences between the two yeast species in
batch culture could be identified.
First, the upper bound on the fraction of phosphoenol-
pyruvate (PEP) molecules derived through at least one trans-
ketolase reaction was widely different (Table 1). This value
provides information about the relative contributions of the
PP pathway and the glucose catabolism via glycolysis to the
generation of trioses. Only an upper bound for the relative
catabolic activity of the PP pathway can be obtained because
glycolytically derived G3P may be exchanged with interme-
diates of the nonoxidative PP pathway via the transketolase
reaction. This exchange reaction generates the same intact
fragments in PEP as glucose catabolism through the PP
pathway. In P. stipitis, the upper bound on the relative PP
pathway flux is less than 57% (Table 1), which is well in the
range of values typically seen in prokaryotic cultures (15, 36,
37). In S. cerevisiae, in contrast, PEP synthesis from pentoses
was not observed, which demonstrates that glucose catabo-
lism proceeds almost exclusively through glycolysis (Fig. 2)
(27). This key difference in glucose catabolism between the
two yeast species becomes readily apparent by visual inspec-
tion of the 13C-13C scalar coupling fine structures of Phe C␣,
which reflects the labeling pattern of PEP (Fig. 3A). The
doublet component with the smaller coupling constant (Fig.
3A) corresponds to PEP molecules in which the C1 carbon
and the C2-C3 fragment originate from different glucose
source molecules, and its presence thus provides evidence
for the contribution of transketolase to glucose catabolism.
Despite the absence of glucose catabolism through the PP
pathway in the batch culture of S. cerevisiae (27), evidence
VOL. 2, 2003
FIG. 3. 13C-13C scalar coupling fine structures along ␻1 (13C) of a
2D [13C,1H]-COSY spectrum at the resonance frequency of Phe C␣
(A) and His C␤ (B). The spectra were recorded with amino acids
obtained from exponentially growing aerobic S. cerevisiae (left panels)
and P. stipitis (right panels) batch cultures. The corresponding carbon
atom in the precursor metabolites is given in parentheses. Carbon
fragments that can be inferred primarily from a given multiplet com-
ponents are depicted below the left cross-section. The arrow (A) in-
dicates the multiplet component used to trace the generation of PEP
from the PP pathway (see text). Preserved bonds from the source
molecule are shown in bold. The relative abundances of the various
carbon fragments calculated from integration of the fine structures (44,
46) are indicated below each spectrum. Abbreviations: s, singlet; d and d*,
doublets with different coupling constants; dd, doublet of doublets.
was obtained for significant anabolic flux via the oxidative
PP pathway to the formation of the biomass precursor P5P:
more than 32% of the P5P molecules originate directly from
glucose (Table 1) (27). This anabolic activity of the PP
pathway is directly apparent in the cross-sections of His C␤,
where an additional doublet splitting of the doublet of dou-
blets (Fig. 3B) documents the presence of intact C5-frag-
ments from glucose in the pool of pentoses (43, 44). Con-
sistent with the putative increase of PP activity in P. stipitis,
the minimal fraction of P5P derived from glucose is 43%
(Table 1).
Second, the fraction of OAAmit (where the superscript
indicates mitochondrial localization of the metabolite pool)
that is derived from PEP via the anaplerotic PYR carboxy-
lase reaction differs by a factor of 2 in the two yeasts:
anaplerosis contributes 70% in S. cerevisiae but only 36% in
P. stipitis to OAAmit synthesis (Table 1). This value specifies
METABOLIC-FLUX PROFILING OF YEASTS 173
the relative flux to OAAmit required to satisfy the biosyn-
thetic demands of the TCA cycle. The competing flux from
ACoAmit to OAAmit is used for complete oxidation of car-
bon substrates to CO2 within the TCA cycle. In yeast, PYR
carboxylase is localized exclusively in the cytosol (49). Fol-
lowing (27), anaplerosis is calculated from the fraction of
OAAmit derived from PEP (Fig. 1), since the PYRcyt (where
the superscript indicates cytosolic localization of the metab-
olite pool) pool is not directly assessable by BDF labeling
because PYR-derived amino acids are synthesized from the
PYRmit pool. We conclude that in P. stipitis the TCA cycle
operates predominantly for respiration (Fig. 2). The 13C-13C
scalar coupling fine structures of Glu C␣ provide direct
evidence for this salient difference (Fig. 4A, left panel). The
higher contribution of the anaplerotic reaction to glutamate
synthesis in S. cerevisiae is reflected by the higher intensity
of the doublet with the smaller coupling constant and the
doublet of doublets arising predominantly from carboxyla-
tion of triply 13C-labeled PYR (44).
Third, while synthesis of OAAcyt, which is required for cy-
tosolic biosynthesis of the Asp amino acid family and anaple-
rosis of the TCA cycle, originates nearly exclusively from PEP
in S. cerevisiae (27), less than half of OAAcyt originates from
PEP in P. stipitis (Table 1; Fig. 2). This becomes readily ap-
parent when comparing the 13C-13C scalar coupling fine struc-
tures of Asp C␣ (Fig. 4B, left panel). Specifically, the large
doublet of doublets in the S. cerevisiae cross section demon-
strates that a high proportion of intact C3-fragments arise from
carboxylation of PYR. On the other hand, the doublet d in Fig.
4B documents that most of the C2-C3 bonds in OAAcyt of P.
stipitis were newly formed. This is characteristic for OAAcyt
synthesized either from oxoglutarate in the TCA cycle or
through the glyoxylate pathway.
Fourth, the fraction of cytosolic or mitochondrial OAA
that was once reversibly converted to fumarate (FUM) is
much higher for P. stipitis than for S. cerevisiae (Table 1).
For OAAmit, this conversion occurs via malate dehydroge-
nase and fumarase in the TCA cycle. The increased ex-
change flux might relate to a predominantly respirative me-
tabolism in P. stipitis. The reversible conversion of OAAcyt,
however, may occur either in the cytosol via the glyoxylate
pathway enzyme MAL dehydrogenase in combination with
the cytosolic isoform of fumarase (28) or by intercompart-
mental exchange of OAA molecules (Fig. 1). The small
extent of reversible conversion of OAAcyt to FUM in S.
cerevisiae thus demonstrates the absence of glyoxylate shunt
activity and indicates that the flux of OAAcyt to OAAmit is
unidirectional (27). In P. stipitis, however, about 20% of the
OAAcyt molecules are converted to FUM at least once, and
thus one or both of the two above-mentioned flux scenarios
must be invoked.
Fifth, although malic enzyme contributes significantly to mi-
tochondrial PYR biosynthesis in batch cultures of S. cerevisiae
(27), malic enzyme activity is low in P. stipitis (Table 1). The
manifestation of this reaction in the labeling pattern of PYRmit
would be consistent with mitochondrial localization of the
malic enzyme not only in S. cerevisiae (2, 27) but in both yeasts.
174 FIAUX ET AL. EUKARYOT. CELL

FIG. 4. 13C-13C scalar coupling fine structures along ␻1 (13C) of a 2D [13C,1H]-COSY spectrum at the resonance frequency of Glu C␣ (A) and
Asp C␣ (B). The spectra were recorded with amino acids obtained from batch and chemostat cultures of S. cerevisiae and P. stipitis. The
corresponding carbon atom in the precursor metabolites is given in parentheses. As indicated in the left-most cross-sections, all these multiplets
consist of a singlet (s) representing the 12C⬘–13C␣–12C␤ isotopomer, a doublet (d) originating from 12C⬘–13C␣–13C␤, a second doublet (d*) for
13
C⬘–13C␣–12C␤, and a doublet of doublets (dd) arising from 13C⬘–13C␣–13C␤.

TABLE 2. Origins of metabolic intermediates during growth of

Central metabolism during growth in glucose-limited che- P. stipitis and S. cerevisiae in glucose-limited chemostat
cultures at a dilution rate of 0.1 h⫺1c
mostat cultures. While catabolic fluxes are maximal during
exponential growth in batch cultures, growth and metabolic % Fraction of total pool
(mean ⫾ SD)
fluxes are reduced in chemostat cultures by the limited avail-
Metabolite
ability of glucose. Flux patterns obtained from these differ- P. stipitis S. cerevisiae
ent types of growth conditions have been shown to differ (aerobic) Aerobic Anaerobic
significantly in most microbes (15, 18, 37). To investigate
flux responses to glucose limitation under conditions when Cytosolic
PEP derived through at least 61 ⫾ 11 40 ⫾ 8 8⫾5
no Crabtree effect is active, we cultivated both yeasts in one TK (ub)
aerobic chemostat cultures at a dilution rate of 0.1 h⫺1. P. P5P from glucose (lb) 28 ⫾ 2 41 ⫾ 2 11 ⫾ 2
stipitis grew somewhat more efficiently, as evidenced by bio- P5P from G3P and S7P (TK 72 ⫾ 2 59 ⫾ 2 89 ⫾ 2
mass yields per g of glucose consumed (means ⫾ standard reaction)
P5P from E4P (TK and TA 43 ⫾ 2 33 ⫾ 2 15 ⫾ 2
deviations) of 0.45 ⫾ 0.03 g and 0.38 ⫾ 0.02 g (cdw) for P. reactions)
stipitis and S. cerevisiae, respectively. For both yeasts the E4P from F6P (lb) 27 ⫾ 5 6⫾6 44 ⫾ 4
growth was predominantly respirative, with little ethanol PEP from OAAcyt (PEP 0–3 2⫾8 NDa
formation (data not shown). When the S. cerevisiae culture carboxykinase reaction)
was switched to anaerobic conditions (note that P. stipitis OAAcyt from PYRcyt 24 ⫾ 3b 62 ⫾ 4b 40 ⫾ 10b
OAAcyt reversibly converted to 47 ⫾ 16 0–8 37 ⫾ 7
does not grow under such conditions), metabolism became FUM at least once (cytosolic
entirely fermentative, with a strongly reduced biomass yield or intercompartmental
(mean ⫾ standard deviation) of 0.09 ⫾ 0.00 g (cdw) per g of exchange)
glucose and ethanol as the primary metabolic product. Glyc-
Mitochondrial
erol, acetate, succinate, and PYR were also found in appre- PYRmit from MAL ⬍7 ⬍13 ⬎10
ciable amounts in the anaerobic culture (data not shown). OAAmit from PEP 32 ⫾ 2 31 ⫾ 2 98 ⫾ 2
METAFoR analyses were performed with hydrolyzed bio- (anaplerosis)
mass samples that were harvested from these chemostat OAAmit reversibly converted 58 ⫾ 14 56 ⫾ 14 43 ⫾ 4
cultures in physiological steady-state. to FUM at least once (in
the TCA cycle)
In aerobic chemostats, the upper bound of PEP molecules
a
that were derived through at least one transketolase reaction Not detectable because of degeneracy of the f values.
b
Values calculated assuming absence of cytosolic OAA-to-FUM conversion
was higher in P. stipitis than in S. cerevisiae (Table 2), but for S. (27).
cerevisiae in both aerobic and anaerobic growth it was much c
See Table 1 footnotes and legend to Fig. 1 for abbreviations.
VOL. 2, 2003
TABLE 3. Comparison of the flux ratios in S. cerevisiae batch and
chemostat cultures determined by metabolic flux balancing (18)
and by METAFoR analysis (Tables 1 and 2)
% Fraction of total pool (mean ⫾ SD)
Pathway and Batch culture Chemostat culture
metabolite
Flux METAFoR Flux METAFoR
balancing analysis balancing analysis
Glucose catabolism 14a 0–4 52a ⬍40 ⫾ 8
through the PP pathway
(PEP derived through at
least one TK reaction)
Anaplerosis
OAAmit from PEP 100 76 ⫾ 4 42 31 ⫾ 2
OAAcyt from PYRcyt 82 88–100 60 62 ⫾ 4
Malic enzyme (PYRmit 2 28–42 5 2–13
from MAL)
a
Note that Gombert et al. estimated the split ratio of glycolysis to PP pathway.
To compare this value to the value of PEP via TK, one needs to subtract the
biosynthetic withdrawal of PP pathway intermediates.
higher than previously described for batch cultures of (27)
(compare Tables 1 and 2). Thus, the difference in catabolic PP
pathway usage is much less pronounced under glucose-limited
growth conditions than in batch cultures because the chemo-
stat cultures grew at only about one third of the maximum
specific growth rate. Under anaerobic conditions, the fraction
of PEP derived through at least one transketolase reaction is
significantly reduced (Table 2), which indicates reduced cata-
bolic use of the PP pathway, as was shown previously also for
anaerobic batch cultures (27).
During aerobic chemostat growth, both yeasts exhibit a com-
parable contribution of about 30% from the anaplerotic reac-
tion to the synthesis of OAAmit (OAAmit from PEP in Table
2), while the remaining OAAmit originates from the TCA cycle.
This flux partitioning is very similar to the value obtained for
the P. stipitis batch culture, but it is in contrast to the values
obtained for the S. cerevisiae batch culture (Table 1). Thus, P.
stipitis appears to exhibit predominantly respirative metabo-
lism in both glucose-limited chemostat and batch culture,
whereas S. cerevisiae metabolism is respirative in the chemostat
but respiro-fermentative in the batch culture, which is consis-
tent with the physiological data. In anaerobic chemostat cul-
ture, the TCA cycle operates through two branches, which both
serve exclusively anabolic functions (Table 2). As described above
for anaplerosis, these flux responses are readily apparent by in-
spection of the Glu C␣ cross-sections (Fig. 4A, right panel).
In contrast to the S. cerevisiae batch culture (Table 1),
OAAcyt is not derived exclusively from PYRcyt in the che-
mostat cultures (Table 2). As discussed above for the batch
cultures, this observation could be explained either by sig-
nificant exchange between the OAAcyt and OAAmit pools,
by high activity of the glyoxylate pathway, or by a combina-
tion of these two factors. Consistently, a significant fraction
of OAAcyt was reversibly converted at least once to FUM in
aerobic P. stipitis and anaerobic S. cerevisiae cultures (Table
2). Furthermore, the data indicate that a small fraction of
PYRmit originates from MAL (Tables 1 and 2), thus dem-
onstrating activity of the malic enzyme in S. cerevisiae and P.
stipitis chemostat cultures. The relative contribution of
METABOLIC-FLUX PROFILING OF YEASTS 175
malic enzyme to the synthesis of PYRmit in S. cerevisiae is,
however, significantly reduced in the chemostat culture
when compared with the batch culture (27) (Tables 1 and 2).
The malic enzyme activity in aerobic and anaerobic S. cer-
evisiae cultures is comparable (Table 2).
Transhydrogenase expression compensates for Pgi knock-
out in S. cerevisiae. The flux ratios in batch culture might
suggest that S. cerevisiae has relatively low catabolic capacity
of the PP pathway, which would be consistent with the
apparent difference in pentose utilization by the two yeasts
considered here (20, 25, 39, 52). To address this hypothesis,
we used a Pgi-deficient S. cerevisiae strain which catabolizes
glucose exclusively via the PP pathway. Since S. cerevisiae
does not contain a transhydrogenase that could transfer
electrons from NADPH to NAD⫹ (47), this mutant cannot
reoxidize PP pathway-produced NADPH and, consequently,
it cannot grow on glucose as the sole carbon source (3). To
enable recycling of NADPH, we overexpressed the soluble
transhydrogenase UdhA from E. coli (4, 7) in this Pgi mu-
tant. As expected, the resulting Pgi⫺ UdhA⫹ strain grows
aerobically on glucose with a specific growth rate of 0.15
h⫺1, which is half the rate observed for the control strain
with a functional Pgi. Hence, S. cerevisiae is fully capable of
exclusive glucose catabolism via the PP pathway at a rela-
tively high metabolic rate, provided that concomitantly pro-
duced NADPH is recycled.
DISCUSSION
Here, we report a metabolic network analysis of P. stipitis
central carbon metabolism, which shows that S. cerevisiae
network topology (27) allows consistent data interpretation
also for P. stipitis. The following key factors were identified
for the flux distributions: (i) There is significant in vivo
activity of the nonoxidative PP pathway. (ii) There is evi-
dence for PYR carboxylase activity in the anaplerotic reac-
tion. (iii) There must be intercompartmental exchange of
OAA if one assumes that the glyoxylate pathway is inactive.
(iv) There is evidence for malic enzyme activity in the mi-
tochondria, as was also reported for S. cerevisiae (2, 27).
Furthermore, there was no evidence in the aerobic regime
for activity of the gluconeogenic PEP carboxykinase in ei-
ther of the two yeasts, which is consistent with the reported
glucose repression of the PEP carboxykinase-encoding gene
in S. cerevisiae (13, 35, 42). The present data thus suggest
that a similar regulation operates in P. stipitis, and that the
pathways for amino acid biosynthesis in P. stipitis are iden-
tical to those used in S. cerevisiae.
Overall, comparative flux profiling by METAFoR analysis
revealed distinctly different metabolic regimes for cultures
growing either in batch or in glucose-limited chemostat
mode. Metabolic differences between the two yeast species
were minor in glucose-limited chemostats, where cultures
exhibited a predominantly respirative mode of metabolism,
but were significant in batch culture. First, in batch culture,
S. cerevisiae exhibited only 24% cyclic (respiratory) TCA
cycle flux (from oxoglutarate) compared to 76% anaplerotic flux
(from OAAcyt) into the OAAmit pool of the TCA cycle, but the
inverse ratio for these two fluxes was found for P. stipitis (Fig.
2), as was also reported for other yeasts (29). This flux ratio is
176 FIAUX ET AL.
consistent with the generally held view that sugar catabolism in
aerobic batch cultures is respiro-fermentative in S. cerevisiae
(53), and predominantly respirative in Crabtree-negative
yeasts such as P. stipitis (33). Second, during respiro-fermen-
tative metabolism of S. cerevisiae in batch culture, cytosolic
OAA originates exclusively from PYR and is transported uni-
directionally into the mitochondria (27). The predominantly
respirative metabolism in all other aerobic cultures, however,
generates the OAAcyt pool containing significant fractions of
molecules that do not originate from PYRcyt but come either
from intercompartmental exchange of TCA cycle intermedi-
ates or from the cytosolic glyoxylate pathway. Third, although
some in vivo activity of the mitochondrial malic enzyme is
detected in all cultures, the highest relative contribution to
PYRmit synthesis was observed for respiro-fermentative S. cer-
evisiae in the batch culture. Fourth, our results reveal a major
difference in the contribution of the PP pathway to glucose
catabolism in batch cultures of P. stipitis and S. cerevisiae. The
fraction of PEP molecules derived through at least one trans-
ketolase provides the sum of the net flux through the PP
pathway and the reversible exchange of trioses with interme-
diates of the nonoxidative PP pathway. This value is close to
zero in the S. cerevisiae batch culture (27), indicating absence
of significant in vivo activity of the nonoxidative PP pathway
branch in S. cerevisiae batch culture, which is thus different
from P. stipitis batch culture (Table 1; Fig. 2).
From this exceptionally low catabolic PP pathway flux in S.
cerevisiae one might conclude that the in vivo activity of this
pathway limits catabolism of pentoses in recombinant S. cer-
evisiae (20, 25, 52), but this appears to be incompatible with the
comparatively rapid growth of Pgi-deficient, transhydrogenase-
overexpressing S. cerevisiae. Since it depends on exclusive glu-
cose catabolism via the PP pathway, this strain grows at a
specific rate that is half of that observed with the control. As a
reference, a corresponding E. coli strain grew at about one
third of the wild type rate (7). Thus, sufficient PP pathway
capacity is potentially available in S. cerevisiae, although it
might not be fully activated during pentose catabolism. The
conclusion that the nonoxidative pathway does not principally
limited pentose catabolism in recombinant S. cerevisiae agrees
well with the results of a recent study to simultaneous overex-
press all four pathway enzymes (24). Our results, however, do
not exclude the possibility that the rate of pentose catabolism
is affected by the activity of one or more PP pathway enzymes.
Like in E. coli, our results reveal that UdhA catalyzes the
electron transfer from NADPH to NAD⫹, as was observed
previously during overexpression of another soluble transhy-
drogenase in S. cerevisiae (30).
Using a different methodology, Gombert et al. recently
reported on network identification and flux quantification in
the same S. cerevisiae strain that was used here (18), leading
to a mutual identity of the metabolic network models in both
cases. To facilitate direct comparison between the results of
both studies, we calculated several metabolic-flux ratios
from the net flux data of Gombert et al. (18) (Table 3). The
flux ratios obtained with the two different approaches are in
good agreement; i.e., both reveal increased glucose catabo-
lism via the PP pathway and reduced anaplerosis in glucose-
limited chemostat culture when compared to batch culture.
Both cultures show some malic enzyme activity and very
EUKARYOT. CELL
little if any glycine cleavage to CO2 (data not shown). In
particular, the consistent flux distributions obtained for glu-
cose-limited S. cerevisiae by these two and other methods
support that both techniques provide reliable results (8, 9, 18).
Three apparent differences between the results obtained
with the method used by Gombert et al. and with our ap-
proach concern primarily the batch growth of S. cerevisiae.
First, Gombert et al. report exclusive operation of the TCA
cycle as a two-branched pathway in batch culture that does
not contribute to glucose catabolism, while we found previ-
ously that cyclic TCA cycle fluxes contribute about 24% of
the total flux to OAAmit (27). Since the absolute flux to
OAAmit is relatively low in batch culture, this minor net
TCA cycle flux may have been missed by the flux estimation
procedure due to its small contribution to the global target
function of the isotopomer balancing. Second, our results
suggest a significant contribution of malic enzyme to PYRmit
synthesis in batch culture (27), where this activity is not
reported by Gombert et al. (18). Similar to the above case,
absolute fluxes to PYRmit are low under these respiro-fer-
mentative conditions, so that the higher relative contribu-
tion of malic enzyme to PYRmit synthesis not necessarily
reflect a higher absolute flux. Third, our data show that the
catabolic flux through the PP pathway in batch cultures of S.
cerevisiae contributes less than 4% to PEP biosynthesis (27),
compared to an estimate of 14% calculated from the data of
Gombert et al. (Table 3). Both estimates are consistent with
the calculated minimal flux into the PP pathway of 2% of the
total glucose consumption for NADPH formation with am-
monia as the sole nitrogen source (5, 47).
Global interpretation of 13C-labeling and physiological
data on the basis of isotopomer balancing (12, 38, 50, 56) as
used by Gombert et al. extracts the maximum information
from labeling experiments, but local features of the flux
distribution may be inaccurately reflected as a result of the
global error minimization function used. METAFoR analy-
sis differs from this approach by providing a strictly local
analysis of 13C-labeling data, which is independent of abso-
lute fluxes and physiological data (15). Additionally, varia-
tions in the growth conditions used (i.e., 0.5 units pH dif-
ference, different composition of the minimal medium, and
bioreactor versus shake flask cultivation) may have led to
slightly different flux distributions. A possible dependence
of catabolic PP pathway fluxes on environmental conditions
may also explain, at least in part, previously reported dif-
ferences in PP pathway flux estimates for S. cerevisiae on the
basis of isotope labeling studies (17, 18, 27, 54; P. M. Bru-
inenberg, G. W. Waslander, J. P. van Dijken, and W. A.
Scheffers, unpublished data [presented at the Physiological
and genetic modulation of product formation, Como,
1986]). Overall, the combined use of global isotopomer bal-
ancing and local METAFoR analysis promises to further enhance
the reliability of future flux determinations (10, 15, 55).
ACKNOWLEDGMENTS
We thank Jay Bailey for his continuous support and Eckhard Boles
for sharing the Pgi mutant and the HXT promoter with us.
Financial support for this work was partly obtained from the Swiss
Priority Program in Biotechnology.
 13
TABLE A1. Relative abundances of intact C2 and C3 fragments in the biosynthesis of fractionally C-labeled amino acids for chemostat
cultivations of S. cerevisiae and P. stipitisa
Relative abundance
Carbon
Chemostat S. cerevisiae P. stipitis
position
(1) (2) (2ⴱ) (3) (1) (2)
f f f f f f f(2ⴱ) f(3)

Ala C␣ Aerobic 0.04 0.11 0.04 0.81 0.01 0.15 0.03 0.81
Anaerobic 0.00 0.07 0.00 0.94

Ala C␤ Aerobic 0.08 0.92 0.06 0.94

Anaerobic 0.01 0.99

Arg C␤ Aerobic 0.66 0.32 0.00 0.02 0.63 0.00 0.33 0.04
Anaerobic 0.07 0.93 0.00 0.00

Arg C␦ Aerobic 0.14 0.86 0.09 0.91

Anaerobic 0.26 0.74

Asp C␣ Aerobic 0.14 0.10 0.19 0.57 0.24 0.14 0.42 0.20
Anaerobic 0.01 0.22 0.01 0.76

Asp C␤ Aerobic 0.17 0.65 0.16 0.02 0.26 0.25 0.41 0.08
Anaerobic 0.01 0.79 0.00 0.20

Glu C␣ Aerobic 0.28 0.22 0.42 0.08 0.24 0.22 0.45 0.09
Anaerobic 0.02 0.75 0.03 0.20

Glu C␤ Aerobic 0.69 0.31 0.00 0.68 0.32 0.00

Anaerobic 0.00 1.00 0.00

Glu C␥ Aerobic 0.08 0.00 0.92 0.00 0.02 0.00 0.98 0.00
Anaerobic 0.14 0.00 0.86 0.00

Gly C␣ Aerobic 0.19 0.81 0.18 0.82

Anaerobic 0.10 0.90

His C␣ Aerobic 0.04 0.02 0.04 0.90 0.03 0.00 0.02 0.95
Anaerobic 0.02 0.01 0.00 0.97

His C␤ Aerobic 0.10 0.49 0.00 0.41 0.12 0.60 0.00 0.28
Anaerobic 0.07 0.82 0.00 0.11

His C␦2 Aerobic 0.32 0.68 0.43 0.57

Anaerobic 0.15 0.85

Ile C␣ Aerobic 0.25 0.00 0.75 0.00 0.37 0.01 0.62 0.00
Anaerobic 0.23 0.00 0.77 0.00

Ile C␥2 Aerobic 0.12 0.88 0.06 0.94

Anaerobic 0.03 0.97

Ile C␥1 Aerobic 0.82 0.18 0.00 0.52 0.48 0.00

Anaerobic 0.82 0.18 0.00

Ile C␦1 Aerobic 0.84 0.16 0.53 0.47

Anaerobic 0.91 0.09

Leu C␣ Aerobic 0.10 0.00 0.90 0.00 0.05 0.00 0.95 0.00
Anaerobic 0.11 0.00 0.89 0.00

Leu C␤ Aerobic 0.95 0.01 0.04 1.00 0.00 0.00

Anaerobic 1.00 0.00 0.00

Leu C␦1 Aerobic 0.24 0.76 0.17 0.83

Anaerobic 0.13 0.87

Leu C␦2 Aerobic 1.00 0.00 1.00 0.00

Anaerobic 1.00 0.00

Lys C␣ Aerobic 0.07 0.05 0.85 0.03 0.06 0.03 0.91 0.00
Anaerobic 0.18 0.00 0.82 0.00

Lys C␤ Aerobic 0.70 0.30 0.00 0.70 0.29 0.01

Anaerobic 0.03 0.97 0.00
Continued on following page

177
 TABLE A1—Continued
Relative abundance
Carbon
Chemostat S. cerevisiae P. stipitis
position
f(1) f(2) f(2ⴱ) f(3) f(1) f(2) f(2ⴱ) f(3)

Lys C␥ Aerobic 0.66 0.34 0.00 0.68 0.32 0.00

Anaerobic 0.04 0.96 0.00

Lys C␦ Aerobic 0.11 0.79 0.10 0.05 0.95 0.00

Anaerobic 0.22 0.77 0.01

Lys Cε Aerobic 0.14 0.86 0.07 0.93

Anaerobic 0.24 0.76

Met C␣ Aerobic 0.14 0.12 0.19 0.55 0.20 0.22 0.32 0.26
Anaerobic 0.03 0.20 0.02 0.75

Phe C␣ Aerobic 0.04 0.09 0.01 0.86 0.04 0.10 0.00 0.86
Anaerobic 0.03 0.02 0.00 0.95

Phe C␤ Aerobic 0.03 0.97 0.00 0.00 0.02 0.98 0.00 0.00
Anaerobic 0.02 0.98 0.00 0.00

Pro C␦ Aerobic 0.12 0.88 0.09 0.91

Anaerobic 0.22 0.78

Pro C␣ Aerobic 0.30 0.22 0.43 0.05 0.27 0.28 0.33 0.12
Anaerobic 0.07 0.75 0.00 0.18

Pro C␤ Aerobic 0.66 0.34 0.00 0.67 0.33 0.00

Anaerobic 0.02 0.18 0.80

Pro C␥ Aerobic 0.09 0.88 0.03 0.06 0.91 0.03

Anaerobic 0.20 0.76 0.04

Ser C␣ Aerobic 0.06 0.07 0.30 0.57 0.05 0.08 0.23 0.64
Anaerobic 0.02 0.02 0.34 0.62

Ser C␤ Aerobic 0.39 0.61 0.30 0.70

Anaerobic 0.39 0.61

Thr C␣ Aerobic 0.15 0.11 0.18 0.56 0.25 0.14 0.41 0.20
Anaerobic 0.02 0.23 0.02 0.73

Thr C␤ Aerobic 0.17 0.41 0.42 0.27 0.68 0.05

Anaerobic 0.04 0.82 0.14

Thr C␥2 Aerobic 0.83 0.17 0.50 0.50

Anaerobic 0.83 0.17

Tyr C␣ Aerobic 0.05 0.10 0.00 0.85 0.04 0.12 0.00 0.84
Anaerobic 0.03 0.03 0.00 0.94

Tyr C␤ Aerobic 0.034 0.96 0.00 0.00 0.05 0.95 0.00 0.00
Anaerobic 0.03 0.97 0.00 0.00

Tyr C␦1 Aerobic 0.05 0.95 0.00 0.04 0.96 0.00

Anaerobic 0.04 0.96 0.00

Tyr Cε1 Aerobic 0.31 0.08 0.18 0.43 0.24 0.05 0.16 0.55
Anaerobic 0.29 0.00 0.27 0.44

Val C␣ Aerobic 0.14 0.00 0.83 0.03 0.12 0.01 0.83 0.04
Anaerobic 0.10 0.00 0.90 0.00

Val C␥1 Aerobic 0.11 0.89 0.06 0.94

Anaerobic 0.02 0.98

Val C␥2 Aerobic 0.97 0.03 0.96 0.04

Anaerobic 0.96 0.04
a
The fractions f(i), named according to reference 44 and corresponding to the relative amount of Ci intact fragments, were calculated from the relative intensities
of the 13C multiplet components using the probabilistic equations described by Szyperski (44). Note that in reference 44 f(2a) and f(2b) correspond to f(2) and f(2ⴱ),
respectively.

178
VOL. 2, 2003
APPENDIX
The supplementary material mentioned in Materials and Methods is
presented in Tables A1 and A2.
REFERENCES
1. Bakker, B. M., K. M. Overkamp, A. J. A. van Maris, P. Kötter, M. A. H.
Luttik, J. P. van Dijken, and J. T. Pronk. 2001. Stoichiometry and compart-
mentation of NADH metabolism in Saccharomyces cerevisiae. FEMS Micro-
biol. Rev. 25:15–37.
2. Boles, E., P. de Jong-Gubbels, and J. T. Pronk. 1998. Identification and
characterization of MAE1, the Saccharomyces cerevisiae structural gene en-
coding mitochondrial malic enzyme. J. Bacteriol. 180:2875–2882.
3. Boles, E., W. Lehnert, and F. K. Zimmermann. 1993. The role of the
NAD-dependent glutamate dehydrogenase in restoring growth on glucose of
a Saccharomyces cerevisiae phosphoglucose isomerase mutant. Eur. J. Bio-
chem. 217:469–477.
4. Boonstra, B., C. E. French, I. Wainwright, and N. C. Bruce. 1999. The udhA
gene of Escherichia coli encodes a soluble pyridine nucleotide transhydro-
genase. J. Bacteriol. 181:1030–1034.
5. Bruinenberg, P. M., J. P. van Dijken, and W. A. Scheffers. 1983. A theoret-
ical analysis of NADPH production and consumption in yeasts. J. Gen.
Microbiol. 129:953–964.
6. Çakar, Z. P., U. Sauer, and J. E. Bailey. 1999. Metabolic engineering of
yeast: the perils of auxotrophic hosts. Biotechnol. Lett. 21:611–616.
7. Canonaco, F., T. A. Hess, S. Heri, T. Wang, T. Szyperski, and U. Sauer. 2001.
Metabolic flux response to phosphoglucose isomerase knock-out in Esche-
richia coli and impact of overexpression of the soluble transhydrogenase
UdhA. FEMS Microbiol. Lett. 204:247–252.
8. Christensen, B., T. Christiansen, A. K. Gombert, and J. Nielsen. 2001.
Simple and robust method for estimation of the split ratio between the
oxidative pentose phosphate pathways and the Embden-Meyerhof-Arnas
pathway in microorganisms. Biotechnol. Bioeng. 74:517–523.
9. Christensen, B., A. K. Gombert, and J. Nielsen. 2002. Analysis of flux
estimates based on 13C-labeling experiments. Eur. J. Biochem. 269:2795–2800.
10. Christensen, B., and J. Nielsen. 1999. Metabolic network analysis. Adv.
Biochem. Eng. Biotechnol. 66:209–231.
11. Cortassa, S., and M. A. Aon. 1998. The onset of fermentative metabolism in
continuous culture depends on the catabolite repression properties of Sac-
charomyces cerevisiae. Enz. Microb. Technol. 22:705–712.
12. Dauner, M., J. E. Bailey, and U. Sauer. 2001. Metabolic flux analysis with a
comprehensive isotopomer model in Bacillus subtilis. Biotechnol. Bioeng.
76:144–156.
13. de Jong-Gubbels, P., P. A. Vanrolleghem, J. J. Heijnen, J. P. van Dijken, and
J. T. Pronk. 1995. Regulation of carbon metabolism in chemostat cultures of
Saccharomyces cerevisiae grown on mixtures of glucose and ethanol. Yeast
11:407–418.
14. Dickinson, J. R., S. J. Harrison, and M. J. Hewlins. 1998. An investigation
of the metabolism of valine to isobutyl alcohol in Saccharomyces cerevisiae.
J. Biol. Chem. 273:25751–25756.
15. Emmerling, M., M. Dauner, A. Ponti, J. Fiaux, M. Hochuli, T. Szyperski, K.
Wüthrich, J. E. Bailey, and U. Sauer. 2002. Metabolic flux responses to
pyruvate kinase knockout in Escherichia coli. J. Bacteriol. 184:152–164.
16. Gancedo, J. M. 1998. Yeast carbon catabolite repression. Microbiol. Mol.
Biol. Rev. 62:334–361.
17. Gancedo, J. M., and R. Lagunas. 1973. Contribution of the pentose-phos-
phate pathway to glucose metabolism in Saccharomyces cerevisiae: a critical
analysis on the use of labelled glucose. Plant Sci. Lett. 1:193–200.
18. Gombert, A. K., M. M. dos Santos, B. Christensen, and J. Nielsen. 2001.
Network identification and flux quantification in the central metabolism of
Saccharomyces cerevisiae under different conditions of glucose repression. J.
Bacteriol. 183:1441–1451.
19. Gupta, R. N., T. Hemscheidt, B. G. Sayer, and I. D. Spenser. 2001. Biosyn-
thesis of vitamin B6 in yeast: incorporation pattern of glucose. J. Am. Chem.
Soc. 123:11353–11359.
20. Hahn-Hägerdahl, B., C. F. Wahlbom, M. Gardonyi, W. H. van Zyl, R. R.
Cordero Otero, and L. J. Jönsson. 2001. Metabolic engineering of Saccha-
romyces cerevisiae for xylose utilization. Adv. Biochem. Eng. Biotechnol.
73:53–84.
21. Hauf, J., F. K. Zimmermann, and S. Müller. 2000. Simultaneous genomic
overexpression of seven glycolytic enzymes in the yeast Saccharomyces cer-
evisiae. Enz. Microb. Technol. 26:688–698.
22. Herve, M., B. Buffin-Meyer, F. Bouet, and S. Tran-Dinh. 2000. Detection of
modifications in the glucose metabolism induced by genetic mutations in
Saccharomyces cerevisiae by 13C- and 1H-NMR spectroscopy. Eur. J. Bio-
chem. 267:3337–3344.
23. Hochuli, M., H. Patzelt, D. Oesterhelt, K. Wüthrich, and T. Szyperski. 1999.
Amino acid biosynthesis in the halophilic archaeon Haloarcula hispanica. J.
Bacteriol. 181:3226–3237.
24. Johansson, B., and B. Hahn-Hägerdahl. 2002. The non-oxidative pentose
METABOLIC-FLUX PROFILING OF YEASTS 179
phosphate pathway controls the fermentation rate of xylulose but not of
xylose in Saccharomyces cerevisiae TMB3001. FEMS Yeast Res. 2:277–282.
25. Kötter, P., and M. Ciriacy. 1993. Xylose fermentation by Saccharomyces
cerevisiae. Appl. Microbiol. Biotechnol. 38:776–783.
26. Lancar-Benda, J., B. Foucher, and M. Saint-Macary. 1996. Characterization,
purification and properties of the yeast mitochondrial dicarboxylate carrier.
Biochimie 78:195–200.
27. Maaheimo, H., J. Fiaux, Z. P. Çakar, J. E. Bailey, U. Sauer, and T. Szyper-
ski. 2001. Central carbon metabolism of Saccharomyces cerevisiae explored
by biosynthetic fractional 13C labeling of common amino acids. Eur. J. Bio-
chem. 268:2464–2479.
28. McAlister-Henn, L., and W. C. Small. 1997. Molecular genetics of yeast TCA
cycle isoenzmyes. Prog. Nucleic Acid Res. 57:317–339.
29. Møller, K., B. Christensen, J. Förster, J. Piskur, J. Nielsen, and L. Olsson.
2002. Aerobic glucose metabolism of Saccharomyces kluyveri: growth, me-
tabolite production, and quantification of metabolic fluxes. Biotechnol. Bio-
eng. 77:186–193.
30. Nissen, T. L., M. Anderlund, J. Nielsen, J. Villadsen, and M. C. Kielland-
Brandt. 2001. Expression of a cytoplasmic transhydrogenase in Saccharomy-
ces cerevisiae results in formation of 2-oxoglutarate due to depletion of the
NADPH pool. Yeast 18:19–32.
31. Palmeri, L., F. M. Lasorsa, A. De Palma, F. Palmeri, M. J. Runswick, and
J. E. Walker. 1997. Identification of the yeast ACR1 gene product as a
succinate-fumarate transporter essential for growth on ethanol or acetate.
FEBS Lett. 417:114–118.
32. Palmeri, L., A. Vozza, G. Agrimi, V. De Marco, M. J. Runswick, F. Palmieri,
and J. E. Walkers. 1999. Identification of the yeast mitochondrial transporter
for oxaloacetate and sulfate. J. Biol. Chem. 274:22184–22190.
33. Passoth, V., M. Zimmermann, and U. Klinner. 1996. Peculiarities of the
regulation of fermentation and respiration in the crabtree-negative xylose-
fermenting yeast Pichia stipitis. Appl. Biochem. Biotechnol. 57–58:201–212.
34. Postma, E., C. Verduyn, W. A. Scheffers, and J. P. van Dijken. 1989. Enzymic
analysis of the crabtree effect in glucose-limited chemostat cultures of Sac-
charomyces cerevisiae. Appl. Environ. Microbiol. 55:468–477.
35. Proft, M., D. Grzesitza, and K.-D. Entian. 1995. Identification and charac-
terization of regulatory elements in the phosphoenolpyruvate carboxykinase
gene PCK1 of Saccharomyces cerevisiae. Mol. Gen. Genet. 246:367–373.
36. Sauer, U., V. Hatzimanikatis, J. E. Bailey, M. Hochuli, T. Szyperski, and K.
Wüthrich. 1997. Metabolic fluxes in riboflavin-producing Bacillus subtilis.
Nat. Biotechnol. 15:448–452.
37. Sauer, U., D. R. Lasko, J. Fiaux, H. M., R. Glaser, T. Szyperski, K. Wüthrich,
and J. E. Bailey. 1999. Metabolic flux ratio analysis of genetic and environ-
mental modulations of Escherichia coli central carbon metabolism. J. Bac-
teriol. 181:6679–6688.
38. Schmidt, K., M. Carlsen, J. Nielsen, and J. Villadsen. 1997. Modeling iso-
topomer distributions in biochemical networks using isotopomer mapping
matrices. Biotechnol. Bioeng. 55:831–840.
39. Senac, T., and B. Hahn-Hägerdal. 1990. Intermediary metabolite concentra-
tions in xylulose- and glucose-fermenting Saccharomyces cerevisiae cells.
Appl. Environ. Microbiol. 56:120–126.
40. Senn, H., B. Werner, B. A. Messerle, C. Weber, R. Traber, and K. Wüthrich.
1989. Stereospecific assignment of the methyl 1H-NMR lines of valine and
isoleucine in polypeptides by nonrandom 13C labeling. FEBS Lett. 249:113–
118.
41. Skoog, K., B. Hahn-Hägerdal, H. Degn, J. P. Jacobsen, and H. S. Jacobsen.
1992. Ethanol reassimilation and ethanol tolerance in Pichia stipitis CBS6054
as studied by 13C nuclear magnetic resonance spectroscopy. Appl. Environ.
Microbiol. 58:2552–2558.
42. Stückrath, I., H. C. Lange, P. Kötter, W. M. van Gulik, K.-D. Entian, and
J. J. Heijnen. 2002. Characterization of null mutants of the glyoxylate cycle
and gluconeogenic enzymes in S. cerevisiae through metabolic network mod-
eling verified by chemostat cultivation. Biotechnol. Bioeng. 77:61–72.
43. Szyperski, T. 1998. 13C-NMR, MS and metabolic flux balancing in biotech-
nological research. Q. Rev. Biophys. 31:41–106.
44. Szyperski, T. 1995. Biosynthetically directed fractional 13C-labeling of pro-
teinogenic amino acids: an efficient analytical tool to investigate intermedi-
ary metabolism. Eur. J. Biochem. 232:433–448.
45. Szyperski, T., J. E. Bailey, and K. Wüthrich. 1996. Detecting and dissecting
metabolic fluxes using biosynthetic fractional 13C labeling and two-dimen-
sional NMR spectroscopy. Trends Biotechnol. 14:453–459.
46. Szyperski, T., R. W. Glaser, M. Hochuli, J. Fiaux, U. Sauer, J. E. Bailey, and
K. Wüthrich. 1999. Bioreaction network topology and metabolic flux ratio
analysis by biosynthetic fractional 13C-labeling and two-dimensional NMR
spectroscopy. Metab. Eng. 1:189–197.
47. van Dijken, J. P., and W. A. Scheffers. 1986. Redox balances in the metab-
olism of sugars by yeasts. FEMS Microbiol. Rev. 32:199–224.
48. van Roemund, C. W. T., E. H. Hettema, M. van der Berg, H. F. Tabak, and
R. J. A. Wanders. 1999. Molecular characterization of carnithine-dependent
transport of acetyl-CoA from peroxisomes to mitochondria in Saccharomyces
cerevisiae and identification of a plasma membrane carnithine transporter,
Agp2p. EMBO J. 18:5843–5852.
49. van Urk, H., D. Schipper, G. J. Breedveld, P. R. Mak, W. A. Scheffers, and
180 FIAUX ET AL. EUKARYOT. CELL

13
TABLE A2. Relative abundances of intact C2 and C3 fragments in the biosynthesis of fractionally C-labeled amino acids for batch
cultivations of S. cerevisiae and P. stipitisa
Relative abundance
Carbon
S. cerevisiae P. stipitis
positionb
(1) (2) (2ⴱ) (3) (1) (2)
f f f f f f f(2ⴱ) f(3)

Ala C␣ 0.03 0.05 0.05 0.87 0.05 0.10 0.02 0.83

Ala C␤ 0.09 0.91 0.08 0.92
Arg C␤ 0.26 0.00 0.70 0.04 0.59 0.00 0.36 0.05
Arg C␦ 0.05 0.95 0.08 0.92
Asp C␣ 0.02 0.02 0.02 0.94 0.19 0.15 0.27 0.39
Asp C␤ 0.03 0.96 0.00 0.01 0.20 0.46 0.27 0.07
Glu C␣ 0.05 0.61 0.21 0.13 0.27 0.24 0.39 0.10
Glu C␤ 0.25 0.75 0.00 0.66 0.34 0.00
Glu C␥ 0.03 0.00 0.97 0.00 0.07 0.00 0.93 0.00
Gly C␣ 0.07 0.93 0.23 0.77
His C␣ 0.02 0.00 0.00 0.98 0.07 0.00 0.04 0.89
His C␤ 0.06 0.62 0.00 0.32 0.13 0.44 0.00 0.43
His C␦2 0.11 0.89 0.35 0.65
Ile C␣ 0.05 0.00 0.94 0.01 0.32 0.00 0.68 0.00
Ile C␥2 0.12 0.88 0.08 0.92
Ile C␥1 0.98 0.00 0.02 0.00 0.69 0.00 0.31 0.00
Ile C␦1 1.00 0.00 0.68 0.32
Leu C␣ 0.03 0.00 0.96 0.01 0.07 0.00 0.93 0.00
Leu C␤ 0.99 0.00 0.01 1.00 0.00 0.00
Leu C␦1 0.18 0.82 0.18 0.82
Leu C␦2 1.00 0.00 1.00 0.00
Lys C␣ 0.04 0.00 0.96 0.00 0.05 0.02 0.92 0.01
Lys C␤ 0.26 0.74 0.00 0.67 0.33 0.00
Lys C␥ 0.18 0.78 0.04 0.60 0.38 0.02
Lys C␦ 0.04 0.95 0.01 0.06 0.90 0.04
Lys Cε 0.05 0.95 0.10 0.90
Met C␣ 0.03 0.05 0.05 0.87 0.20 0.20 0.19 0.41
Phe C␣ 0.02 0.01 0.00 0.97 0.04 0.13 0.00 0.83
Phe C␤ 0.02 0.98 0.00 0.00 0.04 0.96 0.00 0.00
Pro C␦ 0.04 0.96 0.07 0.93
Pro C␣ 0.07 0.63 0.22 0.08 0.25 0.24 0.41 0.10
Pro C␤ 0.26 0.74 0.00 0.67 0.33 0.00
Pro C␥ 0.02 0.93 0.05 0.04 0.96 0.00
Ser C␣ 0.01 0.01 0.19 0.79 0.05 0.11 0.12 0.72
Ser C␤ 0.21 0.79 0.21 0.79
Thr C␣ 0.02 0.02 0.03 0.93 0.21 0.13 0.28 0.38
Thr C␤ 0.01 0.99 0.00 0.22 0.74 0.04
Thr C␥2 1.00 0.00 0.65 0.35
Tyr C␣ 0.03 0.01 0.00 0.96 0.04 0.12 0.00 0.84
Tyr C␤ 0.03 0.97 0.00 0.00 0.04 0.96 0.00 0.00
Tyr C␦1 0.04 0.96 0.00 0.04 0.96 0.00
Tyr Cε1 0.31 0.02 0.25 0.42 0.33 0.06 0.23 0.38
Val C␣ 0.09 0.01 0.89 0.01 0.17 0.01 0.81 0.01
Val C␥1 0.10 0.90 0.08 0.92 0.01
Val C␥2 0.93 0.07 0.98 0.02 0.07
a
The data for S. cerevisiae are those of reference 27 for comparison. The fractions f(i), named according to reference 46 and corresponding to the relative amount
of Ci intact fragments, were calculated from the relative intensities of the 13C multiplet components using the probabilistic equations described by Szyperski (44). Note
that in reference 46 f(2a) and f(2b) correspond to f(2) and f(2ⴱ), respectively.
b
All cultures aerobic.

J. P. van Dijken. 1989. Localization and kinetics of pyruvate-metabolizing
enzymes in relation to aerobic alcoholic fermentation in Saccharomyces
cerevisiae. Biochim. Biophys. Acta 992:78–86.
50. van Winden, W., P. Verheijen, and J. J. Heijnen. 2001. Possible pitfalls of flux
calculations based on 13C-labeling. Metab. Eng. 3:151–162.
51. Verduyn, C., E. Postma, W. A. Scheffers, and J. P. van Dijken. 1992. Effect
of benzoic acid on metabolic fluxes in yeasts: a continuous-culture study on
the regulation of respiration and alcoholic fermentation. Yeast 8:501–517.
52. Wahlbom, C. F., A. Eliasson, and B. Hahn-Hägerdal. 2001. Intracellular
fluxes in a recombinant xylose-utilizing Saccharomyces cerevisiae cultivated
anaerobically at different dilution rates and feed concentrations. Biotechnol.
Bioeng. 72:289–296.
53. Walker, G. M. 1998. Yeast physiology and biotechnology. John Wiley &
Sons, Ltd., Chichester, United Kingdom.
54. Wang, C. H., I. Stern, C. M. Gilmour, S. Klungsoyr, D. J. Reed, J. J. Bialy,
B. E. Christensen, and V. H. Cheldelin. 1958. Comparative study of glucose
catabolism by the radiorespirometric method. J. Bacteriol. 76:207–216.
55. Wiechert, W. 2001. 13C metabolic flux analysis. Metab. Eng. 3:195–206.
56. Wiechert, W., C. Siefke, A. A. de Graaf, and A. Marx. 1997. Bidirectional
reaction steps in metabolic networks: II. Flux estimation and statistical anal-
ysis. Biotechnol. Bioeng. 55:118–135.
57. Woldman, Y., and D. R. Appling. 2002. A general method for determining
the contribution of split pathways in metabolite production in the yeast
Saccharomyces cerevisiae. Metab. Eng. 4:170–181.