Time-dependent regulation analysis dissects shifts
between metabolic and gene-expression regulation
during nitrogen starvation in baker’s yeast
Karen van Eunen
1
, Jildau Bouwman
1,
*, Alexander Lindenbergh
1
, Hans V. Westerhoff
1,2
and Barbara M. Bakker
1,3
1 Department of Molecular Cell Physiology, Vrije Universiteit Amsterdam, The Netherlands
2 Manchester Centre for Integrative Systems Biology, University of Manchester, UK
3 Department of Pediatrics, University of Groningen, The Netherlands
Introduction
Living organisms have the option to regulate their
molecular activities by altering expression of the cor-
responding genes. For example, in the yeast Saccharo-
myces cerevisiae changes in glycolytic flux have
frequently been found to be accompanied by changes
in enzyme capacities [1–3] or amounts [4]. However, a
change in flux through a certain enzyme can also be
regulated through the interaction of that enzyme with
altering concentrations of its substrate(s), product(s)
and ⁄ or modifier(s) (metabolic properties). To quantify
the extent to which the change in flux through an
individual enzyme is regulated by a change in enzyme
Keywords
fermentative capacity; glycolysis; regulation

metabolic regulation is also known to act fast. To analyse the interplay
between autophagy and metabolism, we examined the first 4 h of nitrogen
starvation in detail using time-dependent regulation analysis. Some
enzymes were initially regulated more by a breakdown of enzyme capacity
and only later through metabolic regulation. However, other enzymes were
regulated metabolically in the first hours and then shifted towards regula-
tion via enzyme capacity. We conclude that even initial regulation is subtle
and governed by different molecular levels.
Abbreviations
ADH, alcohol dehydrogenase; ALD, aldolase; GAPDH, glyceraldehyde 3-phosphate dehydrogenase; GPM, phosphoglycerate mutase; HXK,
hexokinase; PDC, pyruvate decarboxylase; PFK, phosphofructokinase; PGI, phosphoglucose isomerase; PGK, 3-phosphoglycerate kinase;
PYK, pyruvate kinase.
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5521
capacity ( V
max
) and by changes in the interactions of
the enzyme with the rest of metabolism, regulation
analysis was developed [5–7].
To date, regulation analysis has been applied to
compare two steady states. Previous studies have
revealed a diversity of regulation which remained visi-
ble after the cells ultimately adjusted their enzyme
capacities to the new steady state [5,8,9]. In order to
obtain insight into adaptation strategies of organisms,
it would be more informative to follow the patterns of
regulation during the transition from one steady state
to another. To this end, time-dependent regulation
analysis has been developed [10].
Regulation analysis has the rate through an enzyme
(v) vary proportionally to a function f that depends on

h
quantifies the relative contri-
bution of changes in enzyme capacity (V
max
) to the
regulation of the flux through the enzyme of interest.
The hierarchical regulation coefficient is associated
with changes in the entire gene expression cascade all
the way from transcription to protein synthesis, stabil-
ity and modification [8,9], hence the name ‘hierar-
chical’. The relative contribution of changes in the
interaction of the enzyme with the rest of metabolism
is reflected in the metabolic regulation coefficient q
m
.
Together the two regulation coefficients should
describe regulation completely, i.e. add up to 1.
Experimentally, the hierarchical regulation coeffi-
cient is the one that is more readily determined,
because it requires only measurements of the V
max
of
the enzyme and the flux through it, under two condi-
tions, according to:
q
h
¼
D log V
max
D log J

ðt
0
Þ
log vðtÞÀlog vðt
0
Þ
ð5Þ
We denote the in vivo rate through the enzyme with
v rather than J because we are now considering
transient rather than steady states.
In this study, we applied time-dependent regulation
analysis to the case of the nitrogen starvation of yeast
cells. A brief period of nitrogen starvation is applied at
the end of the production process of industrial baker’s
yeast (S. cerevisiae) in order to increase its carbohy-
drate content, which in turn increases the storage sta-
bility of the yeast [11,12]. This period of nitrogen
starvation leads to partial loss of the fermentative
capacity, which is defined as the specific rate of carbon
dioxide and ethanol production immediately upon
introduction of the yeast into an anaerobic, glucose
excess environment (i.e. the dough). The production of
carbon dioxide plays a major role in leavening of the
dough and gives bread its open structure. It is believed
that the loss in fermentative capacity is mainly caused
by the degradation of proteins. Unspecific bulk degra-
dation of cytosolic proteins and small organelles via
autophagy is enhanced [13,14] within 30 min of nitro-
gen starvation and protein half-lives of < 1 h are mea-
sured [15,16]. If autophagy is the primary cause of the

we hypothesize that the initial regulation will be purely
hierarchical. Such ‘multisite regulation’ [19] would lead
to initial metabolite homeostasis and a lack of meta-
bolic regulation. Alternatively, metabolic regulation
may be involved from the beginning, which will
become visible as a mixed regulation or even a com-
plete metabolic regulation in the early time points. To
our knowledge, this is the first experimental study ever
in which regulation is studied in this way with quanti-
tative time resolution.
Results
Growth and perturbation condition
S. cerevisiae strain CEN.PK113-7D was grown in aero-
bic glucose-limited chemostat cultures at a dilution
rate of 0.35 h
)1
. Under these conditions, a respiro-
fermentative metabolism was observed (Table 1), in
agreement with literature data [20]. To induce nitrogen
starvation, cells were transferred from steady-state
chemostat cultures to a batch culture in medium lack-
ing nitrogen but with excess glucose. The addition of
glucose served to prevent additional starvation for the
carbon source. To discriminate between the effects
caused by nitrogen starvation and by the shift from
glucose limitation to glucose excess, control experi-
ments were performed in which cells were shifted to
glucose excess, but in the continued presence of nitro-
gen. Samples were taken from steady-state cultures
and at 0, 1, 2, 3, 4 and 24 h after the start of the per-

the different time points. The cells were washed and
transferred to an anaerobic vessel containing fresh and
complete (with 38 mm ammonium sulfate) defined min-
eral medium [21] with an excess amount of glucose
(56 mm). This condition mimics the situation of
baker’s yeast in dough [2]. Apart from the ethanol
flux, the fluxes of glucose, glycerol, acetate, succinate,
Table 1. Physiological parameters of the aerobic glucose-limited chemostat cultures from which cells were taken to be subjected to nitro-
gen starvation and glucose excess conditions or glucose and nitrogen excess conditions. Dilution (growth) rate was set to 0.35 h
)1
. Errors
represent SEM of seven independent chemostat cultures.
Yield
glu,X
(gÆg
)1
) q
O
2
a
q
CO
2
b
RQ
c
q
glucose
a
q

produced, within the bounds of experimental error
(Table S1). In the experiment in which cells were
shifted to glucose excess in the presence of nitrogen,
the carbon balance matched only in the 0-h sample. In
the other samples the assessed carbon production rates
were 17–21% lower than the carbon consumption rates
(Table S2). The assumption that the difference is in
the glycogen flux is not realistic in this case, because
glycogen is usually consumed rather than produced
during glucose excess conditions. The most likely
explanation is that the missing carbon ends up in bio-
mass and biomass-related CO
2
. Note that CO
2
was not
measured in the fermentative-capacity assay and the
reported CO
2
flux is calculated based on the catabolic
fluxes. We recalculated the fluxes through the enzymes
by assuming that the gap in the carbon balance was
caused by a flux from pyruvate to biomass. Although
this had an effect on the absolute fluxes, it had little
impact on the regulation analysis reported below.
However, if the gap was caused by drainage at other
points in glycolysis and if the relative flux through
such a branch differed between time points, this may
somewhat affect the reported regulation coefficients in
the control experiment.

0
t
1
t
2
t
3
t
4
t
24
Nitrogen starvation Glucose )0.40 ± 0.02 )0.40 ± 0.02 )0.37 ± 0.00 )0.33 ± 0.02 )0.31 ± 0.02 )0.17 ± 0.02
Ethanol 0.66 ± 0.04 0.60 ± 0.02 0.56 ± 0.01 0.54 ± 0.01 0.56 ± 0.03 0.53 ± 0.01
Glycerol 0.08 ± 0.00 0.08 ± 0.00 0.09 ± 0.00 0.08 ± 0.00 0.08 ± 0.00 0.05 ± 0.01
Trehalose 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 )0.01 ± 0.00 )0.01 ± 0.00 )0.04 ± 0.01
Glucose excess Glucose )0.37 ± 0.03 )0.52 ± 0.01 )0.56 ± 0.02 )0.57 ± 0.02 )0.60 ± 0.06 n.d.
Ethanol 0.62 ± 0.04 0.71 ± 0.02 0.77 ± 0.04 0.84 ± 0.04 0.89 ± 0.10 n.d.
Glycerol 0.08 ± 0.00 0.09 ± 0.00 0.09 ± 0.00 0.09 ± 0.01 0.09 ± 0.01 n.d.
Trehalose 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 0.00 ± 0.00 n.d.
Experimental time-dependent regulation analysis K. van Eunen et al.
5524 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS
corresponds to the fermentative capacity. Upon the
shift from glucose limited to glucose excess conditions
(in the presence of nitrogen) the fermentative capacity
increased by 40%. When the same shift was accompa-
nied by the shift to nitrogen starvation a 20% decrease
in fermentative capacity was observed. This suggests
that the decrease in fermentative capacity is an effect
of the nitrogen starvation itself, but was counteracted
by the shift from glucose-limited to glucose excess

(ALD), PGK, GPM, PYK and pyruvate decarboxylase
(PDC) were upregulated. The capacity of alcohol dehy-
drogenase (ADH) was downregulated and the capaci-
ties of PGI, phosphofructokinase (PFK) and GAPDH
remained constant. PGI was only downregulated at
Fig. 2. Fluxes through the glycolytic and fermentative pathways under anaerobic glucose excess conditions in cells that had undergone the
shift to nitrogen starvation and glucose excess or to glucose excess conditions in the presence of nitrogen. Cells were transferred to the off-
line assay system at various time points during nitrogen-starvation and glucose-excess (closed circles) or during glucose- and nitrogen-excess
conditions (open circles). In this simplified scheme of the glycolytic and fermentative pathways, enzymes with the same flux are depicted in
the same box. Measured fluxes are depicted in bold. Branching metabolites connect the boxes. Fluxes were calculated based on the stoichi-
ometry of the glycolytic and fermentative pathways (described under Experimental procedures). In the graphs, the fluxes through the glyco-
lytic and fermentative pathways are plotted as a function of time. Fluxes are depicted in percentage with respect to the flux at t
0
. The error
bars represent the SEM of three independent experiments carried out on cells from different chemostat cultures (two for t
24
in the nitrogen-
starvation experiment).
K. van Eunen et al. Experimental time-dependent regulation analysis
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5525
4 h. The trend, that more enzymes were upregulated
than downregulated, parallels the observed upregula-
tion of the fluxes under this condition.
Time-dependent regulation analysis during the
first 4 h
If the initial regulation during nitrogen starvation was
dominated by unspecific bulk degradation of cytosolic
proteins and small organelles, all hierarchical regula-
tion coefficients should be equal to 1 initially. Accord-
ing to the summation theorem (Eqn 4) all metabolic

close to 1 and a hierarchical regulation coefficient (q
h
)
close to 0 were found in cells adjusting to nitrogen
starvation, as well as in cells accommodating excess
glucose. The changes in fluxes through these enzymes
ABC
EFGH
D
IJ
Fig. 3. The V
max
values of the glycolytic and fermentative enzymes expressed as percentages with respect to their values at t
0
, during shift
to nitrogen-starvation and glucose-excess (closed circles) or to glucose-excess conditions in the presence of nitrogen (open circles). Error
bars represent the SEM of three (two for t
24
in nitrogen-starvation experiment) independent experiments carried out on cells from different
chemostat cultures. Absolute values are reported in Tables S3 and S4.
Experimental time-dependent regulation analysis K. van Eunen et al.
5526 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS
were regulated purely by interactions with their sub-
strate(s), product(s) or other metabolites and not by
changes of V
max
. GAPDH was regulated metabolically
in both perturbations, PGI only upon nitrogen starva-
tion and PFK only after the shift to glucose-excess
conditions in the presence of nitrogen.

h
from 0 to $ 1, but
also any other time profile in which q
h
increased. This
means that, as time progressed, changes in V
max
became more important at the cost of metabolic regu-
lation. The enzymes PFK, GPM and ADH belonged
to this category when the cells were starved of nitro-
gen. PGK was regulated in this way in the cells shifted
to glucose excess in the presence of nitrogen. HXK,
ALD, PYK and PDC showed increasing hierarchical
regulation upon both perturbations. However, upon
the shift from limiting to excess glucose with excess
nitrogen throughout, all these enzymes showed
decreased hierarchical regulation after 3 or 4 h.
ABCD
EFGH
IJ
Fig. 4. Hierarchical regulation coefficients quantifying the regulation upon shift to nitrogen-starvation and glucose-excess conditions. Regula-
tion coefficients were calculated according to the integrative time-dependent regulation analysis (see Introduction). The error bars represent
SEM of three independent experiments carried out on cells from four different chemostat cultures. The dashed lines indicate a q
h
of 1.0 and
the dotted lines indicate a q
h
of 0.
K. van Eunen et al. Experimental time-dependent regulation analysis
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5527

mutase; HXK, hexokinase; PDC, pyruvate decarboxylase; PFK, phosphofructokinase; PGI, phosphoglucose isomerase; PGK, 3-phosphoglycer-
ate kinase; PYK, pyruvate kinase.
Category of regulation HXK PGI PFK ALD GAPDH PGK GPM PYK PDC ADH
Purely metabolic • o • o
Purely hierarchical • o
Antagonistic directed by metabolism o
Towards hierarchical regulation • o ••oo••o • o •
Towards metabolic regulation o
A
E
I
B
F
J
C
G
D
H
Fig. 5. Hierarchical regulation coefficients quantifying the regulation upon the transition to glucose-excess conditions in the presence of
nitrogen. Regulation coefficients were calculated according to the integrative time-dependent regulation analysis (see Introduction). The error
bars represent SEM of three independent experiments carried out on ditto-different chemostat cultures. The dashed lines indicate a q
h
of
1.0 and the dotted lines indicate a q
h
of 0.
Experimental time-dependent regulation analysis K. van Eunen et al.
5528 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS
precedes metabolic regulation or vice versa (Fig. 4).
Apparently, both mechanisms contribute from the begin-

under the two growth conditions. Apparently, the regu-
lation of the flux through these enzymes upon the intro-
duction of nitrogen starvation is sensitive to the growth
conditions prior to nitrogen starvation.
Transcript levels
The diversity in the time profiles of the V
max
values
suggested that, apart from unspecific bulk degradation
of proteins, other more specific regulation mechanisms
of protein regulation were involved in the response to
nitrogen starvation. To investigate the extent to which
such regulation took place at the mRNA level, we
measured the transcript levels of nearly all glycolytic
and fermentative genes using qPCR (Fig. 6). First, the
V
max
levels of PGI and GAPDH remained constant.
We wondered whether (possible) degradation of these
proteins would be compensated for by increased syn-
thesis driven by increased transcription, but we found
no increase in the mRNA levels of these enzymes.
Figure 6A shows that the transcript level of PGI1 did
not change significantly. The transcript levels of the
TDH genes, which code for GAPDH, were changed
significantly (Student’s t-test, P < 0.05). TDH1 was
increased, and TDH2 and TDH3 were both decreased
(Fig. 6B). However, because TDH3 was the most
abundant of the three, the total transcript level of the
TDH genes was decreased. Second, trends observed in

and cells that
started off as growing exponentially in a batch culture [9]. The
errors represent, SEM of two independent experiments carried out
on different chemostat cultures (this study) and SEM of four inde-
pendent experiments carried out on different batch cultures. ADH,
alcohol dehydrogenase; ALD, aldolase; GAPDH, glyceraldehyde
3-phosphate dehydrogenase; GPM, phosphoglycerate mutase;
HXK, hexokinase; PDC, pyruvate decarboxylase; PFK, phosphofruc-
tokinase; PGI, phosphoglucose isomerase; PGK, 3-phosphoglycer-
ate kinase; PYK, pyruvate kinase.
Enzyme
Respiro-fermentative
growing cells (this
study)
Exponential growing
cells Rossell et al. [9]
q
h
SEM q
m
q
h
SEM q
m
HXK 0.8 0.1 0.2 1.0 0.2 0.0
PGI 3.5 0.9 )2.5 0.8 0.3 0.2
PFK 1.9 0.2 )0.9 0.4 0.2 0.6
ALD 2.0 0.3 )1.0 1.1 0.5 )0.1
GAPDH 0.0 0.2 1.0 0.7 0.5 0.3
PGK 0.1 0.0 0.9 0.0 0.2 1.0

(inverse) correlation between the degree of downregu-
lation under nitrogen starvation and the degree of
upregulation upon glucose excess.
We compared the measured flux and V
max
data to
earlier reports. Both fermentative capacity and enzyme
capacities measured at time point 0 h (nonstarved
yeast cells) were highly comparable to the data
obtained by Van Hoek et al. for yeast grown under
identical conditions [20]. In addition, we calculated
whether the measured V
max
values can support the
fluxes measured under both perturbations. This is true
for all enzymes, with the exception of PFK in the
nitrogen-excess experiment. The fact that PFK has
quite a few allosteric regulators, i.e. ATP, citrate,
fructose-2,6-bisphosphate, etc., might complicate mea-
suring the actual V
max
. However, fructose-2,6-bisphos-
phate is no longer commercially available, which limits
the possibilities for rapid further measurements. Alto-
gether our results were similar to literature data and
make sense to the yeast cell physiology.
Because both metabolic and hierarchical regulation
played a role in the adaptation of the yeast cell to
nitrogen starvation, we discuss the mechanisms acting
at each level. The hierarchical regulation can be

).
Experimental time-dependent regulation analysis K. van Eunen et al.
5530 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS
of the V
max
values of PGK and PYK after 24 h could
not be predicted from changes in their mRNA concen-
trations. Similarly, the constant V
max
values of PGI
and GAPDH could not be simply predicted from the
corresponding mRNA levels. We did observe an
increase in the TDH1 mRNA level, encoding one of
the GAPDH isoenzymes, in line with earlier reports of
induction of this transcript in heat-shocked cells and
under glucose starvation conditions [23]. However, in
our experiments, the decreased expression of TDH3,
the most abundant of the three TDH transcripts, prob-
ably caused an overall decrease in TDH mRNA levels.
The fact that hardly any correlation was observed
between the transcript levels and the enzyme capacities
is consistent with earlier observations [8,24,25].
The lack of correlation between the transcript levels
and enzyme capacities, suggests that the regulation of
the V
max
values may be at the post-transcriptional
level, i.e. protein synthesis, degradation and modifica-
tion. It has been shown that during nitrogen starvation
the rate of protein synthesis is limited by the size of

If the different time profiles of the V
max
values are
caused by differences in the degradation rates of the
corresponding proteins, there are two main scenarios.
Either, some proteins are hidden from the protein-
degradation machinery or this machinery recognizes
the different proteins and distinguishes between them.
There is evidence for both mechanisms. First, GAP-
DH, one of the enzymes with a stable capacity during
the first hour of nitrogen starvation, can be incorpo-
rated into the cell wall under stress conditions such as
starvation and ⁄ or a temperature upshift. This incorpo-
ration of GAPDH into the cell wall in response to
stress does not require de novo protein synthesis [33],
indicating that this mechanism could work under
nitrogen starvation and shield GAPDH effectively
from unspecific breakdown of cytosolic proteins by
autophagy. In our study, we did not distinguish
between different subcellular localizations of the glyco-
lytic proteins, but it should be noted that such relocal-
ization may also preclude participation of the enzyme
in glycolysis and may therefore provide an additional
layer of regulation.
Second, specificity of protein degradation is also a
plausible mechanism to explain our data. In general,
not all proteins are degraded to the same extent and at
the same rate. The autophagy route to degradation,
which is often considered to be unspecific, has been
reported to exhibit some specificity. For example, one

which part of the V
max
regulation is caused by post-
K. van Eunen et al. Experimental time-dependent regulation analysis
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5531
translational modifications of proteins, such as phos-
phorylation. This extension of the method [8] requires
precise measurements of protein concentrations.
Despite improvements in quantitative proteomics, the
accuracy is probably insufficient to dissect the initial
kinetics in this study. The flux reduction is only 20%
in the first 4 h and for example, to quantify a partial
regulation by posttranslational modification at 20%
accurately, would require 4% accuracy in protein
measurements.
To quantitatively explain the metabolic regulation
observed for several enzymes would require not only
metabolite concentrations, but also quantitative assess-
ment of their impact on the enzyme rates. Two studies
showed decreased levels of fructose-1,6-bisphosphate
during the first hour of nitrogen starvation [17,18].
Because fructose-1,6-bisphosphate is known as an allo-
steric activator of PFK and PYK [46,47], this is in line
with the decreased fermentative capacity found under
nitrogen starvation. Another allosteric regulator cit-
rate, which is an inhibitor of PFK [48], is increased
when yeast cells are starved of nitrogen [18]. However,
this is a small sample of all metabolites that might pos-
sibly affect the rates of glycolytic enzymes and the cor-
relation is only qualitative. We are currently working

respiro-fermentative conditions at a dilution rate of
0.35 h
)1
. The stirring speed was 800 rpm. The pH was kept
at 5.0 ± 0.1 by an ADI 1010 controller, via automatic
addition of aliquots of 2 m KOH. The fermentor was aer-
ated by flushing with air at a flow rate of 30 LÆh
-1
. Chemo-
stat cultures were assumed to be in a steady state when,
after at least five volume changes, the culture dry weight,
the specific carbon dioxide production rate and the oxygen
consumption rate had changed by < 2% after at least one
volume change. The number of generations after the start
of the chemostat cultivation was kept < 20, because it is
known that changes in the cell population occur during
prolonged chemostat cultivation [50,51]: the perturbation
was performed after 18–19 generations.
Perturbation conditions
For the nitrogen-starvation experiments, the same defined
mineral medium was used as for the chemostat culture,
except that ammonium sulfate was lacking and glucose was
in excess (195 mm). Also in the case of the nitrogen-excess
conditions, the same defined mineral medium was used but
now in the presence of ammonium sulfate (38 mm) and
with 195 mm glucose. Yeast cells were harvested from the
steady-state chemostat as described above, washed with
equal volumes of ice-cold (4 °C) nitrogen starvation or
nitrogen excess medium, and resuspended in the corre-
sponding medium to a volume equal to that harvested from

1 m NaOH, incubated at 100 °C for 10 min and subse-
quently cooled on ice. The protein concentration was deter-
mined according to the Lowry method with BSA
(2 mgÆmL
)1
stock solution, Pierce, Thermo Fisher Scienti-
fic, Rockford, IL, USA) in 1 m NaOH as standard (final
concentration of BSA stock solution in 1 m NaOH was
1.8 mgÆmL
)1
).
Fermentative capacity and steady-state fluxes
The fermentative capacity is measured as the rate of etha-
nol production in an off-line assay in which cells are trans-
ferred to a complete growth medium under anaerobic
conditions at excess of glucose. Culture samples were taken
and cells were washed and taken up in defined mineral
medium [21] lacking glucose. In previous studies, we did
not observe significant alterations in enzyme activities dur-
ing the washing of cells and transfer to the new medium.
The fermentative capacity and the steady-state fluxes were
measured under anaerobic conditions with an excess of glu-
cose (56 mm, added at time 0) for 30 min in a 6% wet
weight cell suspension at 30 °C. The set-up used for the
determination of fermentative capacity was as described in
Van Hoek et al. [2], with the modification that the head-
space was flushed with water-saturated N
2
(0.6 LÆh
)1

the consumed carbon did not completely match the pro-
duced carbon, the difference was in the glycogen flux,
which we did not measure for reasons of limited accuracy.
In Fig. 2, enzymes with same flux are boxed together. The
flux through HXK is equal to the glucose flux. Fluxes
through PGI, PFK and ALD were calculated by dividing
the sum of the glycerol and ethanol fluxes by 2. The fluxes
through the enzymes from GAPDH to ADH were taken to
be equal to the measured ethanol flux. As the fluxes were
determined under anaerobic conditions, there was no flux
into the citric acid cycle and respiration. Other fluxes,
which may have contributed (acetate, pyruvate and
biomass) were negligible (see Results).
Enzyme capacity measurements
To prepare the cell-free extracts, samples were harvested,
washed twice with 10 mm potassium phosphate buffer (pH
7.5) containing 2 mm Na
2
H
2
-EDTA, concentrated 10-fold
and stored at )20 °C. Samples were thawed, washed and
resuspended in an equal volume of 100 mm potassium
phosphate buffer (pH 7.5) containing 2 mm MgCl
2
and
1mm dithiothreitol. Cell-free extracts were prepared by
using the FastPrep
Ò
method with acid-washed glass beads

(t)] were
calculated according to Eqn (5) (see Introduction). Time
point t
0
is defined as the time at which the perturbation
was started after washing the cells. In total four experi-
ments, in which the cells were shifted to nitrogen starvation
with excess of glucose, were carried out starting from inde-
pendent chemostat cultures and the cultures were moni-
tored during the first 4 h of starvation. V
max
values were
determined in three of the nitrogen-starvation experiments
K. van Eunen et al. Experimental time-dependent regulation analysis
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5533
and for three parallel experiments, the steady-state fluxes
were estimated. Averages and SD were calculated separately
for the numerator and the denominator of Eqn (5). Based
on the SD of the numerator and the denominator the SEM
of q
h
was computed, assuming statistical independence of
the two. The time-dependent metabolic regulation coeffi-
cients [ q
m
(t)] were calculated according to the summation
law (Eqn 4). The same procedure was followed for time
point 24 h of the nitrogen starvation, based on two data-
sets, and for the nitrogen-excess conditions, based on three
datasets.

faster than, for example, differences in protein levels. Disso-
ciation curves (dissociation curves 1.0 f. software, PE
Applied Biosystems) of PCR products were run to verify
that only the correct product was amplified.
DDCt ¼ ÀððCt
X;t
À Ct
PDI1;t
ÞÀðCt
X;ss
À Ct
PDI1;ss
ÞÞ ð6Þ
½mRNA
X

t
½mRNA
X

ss
¼ 2
DDCt
ð7Þ
Acknowledgements
This project was supported financially by the IOP
Genomics program of Senter Novem. The work of
BM Bakker and HV Westerhoff is further supported
by STW, NGI-Kluyver Centre, NWO-SysMO, BBSRC
(including SysMO), EPSRC, AstraZeneca, and EU

283, 23524–23532.
6 Rossell S, van der Weijden CC, Kruckeberg AL,
Bakker BM & Westerhoff HV (2005) Hierarchical and
metabolic regulation of glucose influx in starved
Saccharomyces cerevisiae. FEMS Yeast Res 5, 611–619.
7 ter Kuile BH & Westerhoff HV (2001) Transcriptome
meets metabolome: hierarchical and metabolic regula-
tion of the glycolytic pathway. FEBS Lett 500, 169–
171.
8 Daran-Lapujade P, Rossell S, van Gulik WM, Luttik
MA, de Groot MJ, Slijper M, Heck AJ, Daran JM, de
Winde JH, Westerhoff HV et al. (2007) The fluxes
through glycolytic enzymes in Saccharomyces cerevisiae
are predominantly regulated at posttranscriptional
levels. Proc Natl Acad Sci USA 104, 15753–15758.
9 Rossell S, van der Weijden CC, Lindenbergh A, van
Tuijl A, Francke C, Bakker BM & Westerhoff HV
(2006) Unraveling the complexity of flux regulation: a
new method demonstrated for nutrient starvation in
Saccharomyces cerevisiae. Proc Natl Acad Sci USA 103,
2166–2171.
Experimental time-dependent regulation analysis K. van Eunen et al.
5534 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS
10 Bruggeman FJ, de Haan J, Hardin H, Bouwman J,
Rossell S, van Eunen K, Bakker BM & Westerhoff HV
(2006) Time-dependent hierarchical regulation analysis:
deciphering cellular adaptation. Syst Biol (Stevenage)
153, 318–322.
11 Caron C (1995) Commercial production of baker’s yeast
and wine yeast. In Enzymes, Biomass, Food and Feed

19 Fell DA & Thomas S (1995) Physiological control of
metabolic flux: the requirement for multisite modula-
tion. Biochem J 311, 35–39.
20 Van Hoek P, Van Dijken JP & Pronk JT (2000) Regu-
lation of fermentative capacity and levels of glycolytic
enzymes in chemostat cultures of Saccharomyces cerevi-
siae. Enzyme Microb Technol 26, 724–736.
21 Verduyn C, Postma E, Scheffers WA & Van Dijken JP
(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.
22 van den Brink J, Canelas AB, van Gulik WM, Pronk
JT, Heijnen JJ, de Winde JH & Daran-Lapujade P
(2008) Dynamics of glycolytic regulation during adapta-
tion of Saccharomyces cerevisiae to fermentative metab-
olism. Appl Environ Microbiol 74, 5710–5723.
23 Boucherie H, Bataille N, Fitch IT, Perrot M & Tuite
MF (1995) Differential synthesis of glyceraldehyde-3-
phosphate dehydrogenase polypeptides in stressed yeast
cells. FEMS Microbiol Lett 125, 127–133.
24 Greenbaum D, Colangelo C, Williams K & Gerstein M
(2003) Comparing protein abundance and mRNA
expression levels on a genomic scale. Genome Biol 4,
117.
25 Griffin TJ, Gygi SP, Ideker T, Rist B, Eng J, Hood L
& Aebersold R (2002) Complementary profiling of gene
expression at the transcriptome and proteome levels in
Saccharomyces cerevisiae. Mol Cell Proteomics 1, 323–
333.

phosphate dehydrogenase protein in yeast. FEMS Yeast
Res 4, 297–303.
34 Onodera J & Ohsumi Y (2004) Ald6p is a preferred
target for autophagy in yeast, Saccharomyces cerevisiae.
J Biol Chem 279, 16071–16076.
35 Bernales S, McDonald KL & Walter P (2006) Auto-
phagy counterbalances endoplasmic reticulum expan-
sion during the unfolded protein response. PLoS Biol 4,
e423.
36 Sakai Y, Oku M, van der Klei IJ & Kiel JA (2006)
Pexophagy: autophagic degradation of peroxisomes.
Biochim Biophys Acta 1763, 1767–1775.
K. van Eunen et al. Experimental time-dependent regulation analysis
FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS 5535
37 Tal R, Winter G, Ecker N, Klionsky DJ & Abeliovich
H (2007) Aup1p, a yeast mitochondrial protein phos-
phatase homolog, is required for efficient stationary
phase mitophagy and cell survival. J Biol Chem 282,
5617–5624.
38 Kissova I, Salin B, Schaeffer J, Bhatia S, Manon S
& Camougrand N (2007) Selective and non-selective
autophagic degradation of mitochondria in yeast. Auto-
phagy 3, 329–336.
39 Wang CW, Kim J, Huang WP, Abeliovich H, Stromh-
aug PE, Dunn WA Jr & Klionsky DJ (2001) Apg2 is a
novel protein required for the cytoplasm to vacuole
targeting, autophagy, and pexophagy pathways. J Biol
Chem 276, 30442–30451.
40 Klionsky DJ (1998) Nonclassical protein sorting to the
yeast vacuole. J Biol Chem 273, 10807–10810.

colysis be understood in terms of in vitro kinetics of the
constituent enzymes? Testing biochemistry Eur J
Biochem 267, 5313–5329.
50 Jansen ML, Diderich JA, Mashego M, Hassane A, de
Winde JH, Daran-Lapujade P & Pronk JT (2005) Pro-
longed selection in aerobic, glucose-limited chemostat
cultures of Saccharomyces cerevisiae causes a partial
loss of glycolytic capacity. Microbiology 151, 1657–
1669.
51 Mashego MR, Jansen ML, Vinke JL, van Gulik WM &
Heijnen JJ (2005) Changes in the metabolome of
Saccharomyces cerevisiae associated with evolution in
aerobic glucose-limited chemostats. FEMS Yeast Res 5,
419–430.
52 Postma E, Kuiper A, Tomasouw WF, Scheffers WA &
van Dijken JP (1989) Competition for glucose between
the yeasts Saccharomyces cerevisiae and
Candida utilis.
Appl Environ Microbiol 55, 3214–3220.
53 Schmitt ME, Brown TA & Trumpower BL (1990) A
rapid and simple method for preparation of RNA from
Saccharomyces cerevisiae. Nucleic Acids Res 18, 3091–
3092.
54 Spijker S, Houtzager SW, De Gunst MC, De Boer WP,
Schoffelmeer AN & Smit AB (2004) Morphine exposure
and abstinence define specific stages of gene expression
in the rat nucleus accumbens. FASEB J 18, 848–850.
Supporting information
The following supplementary material is available:
Table S1. C-flux in mmolÆCÆmin

this journal provides supporting information supplied
by the authors. Such materials are peer-reviewed and
may be re-organized for online delivery, but are not
copy-edited or typeset. Technical support issues arising
from supporting information (other than missing files)
should be addressed to the authors.
Experimental time-dependent regulation analysis K. van Eunen et al.
5536 FEBS Journal 276 (2009) 5521–5536 ª 2009 The Authors Journal compilation ª 2009 FEBS