On the mechanism of action of the antifungal agent propionate
Propionyl-CoA inhibits glucose metabolism in
Aspergillus nidulans
Matthias Brock
1
and Wolfgang Buckel
2
1
Laboratorium fu
¨
r Mikrobiologie, Universita
¨
t Hannover;
2
Laboratorium fu
¨
r Mikrobiologie, Fachbereich Biologie,
Philipps-Universita
¨
t Marburg, Germany
Propionate is used to protect b read and animal feed from
moulds. The mode of action of this short-chain fatty acid
was s tudied using Aspergillus nidulans as a model organism.
The filamentous fungus is able to grow slowly on propio-
nate, which is oxidized to acetyl-CoA via propionyl-CoA,
methylcitrate and pyruvate. Propionate inhibits growth of
A. nidulan s on glucose but not on acetate; the l atter was
shown to inhibit propionate oxidation. When grown on
glucose a methylcitrate synthase deletion mutant is much
more sensitive towards the presence of propionate in the
medium as compared to the wild-type and accumulates

methylcitrate [2]. Isomerization of this tricarboxylic acid,
most likely via cis-2-methylaconitate [6], yields (2R,3S)-2-
methylisocitrate, w hich is cleaved t o succinate and pyruvate
[3]. Studies with
13
C-labelled propionate indicated that in
Escherichia coli the 2-oxo acid is further oxidized to
acetyl-CoA, which is either funnelled i nto the citrate cycle
or used for biosyntheses [7].
A clue to the mechan ism of propionate toxicity was
the construction of an A. nidulans methylcitrate synthase
deletion strain (DmcsA), which was unable to grow on
propionate as sole carbon and energy source. Unex-
pectedly, growth of DmcsA on glucose was more
inhibited by propionate than that of a wild-type strain
[2]. This result indicated that (2S,3S)-methylcitrate or
(2R,3S)-2-methylisocitrate are unlikely to b e responsible
for t his inhibitory effect. At h igh l evels of propionyl-
CoA yeast citrate synthase catalyses the slow formation
of three of the four stereoisomers of methylcitrate [8],
but their concentrations (< 10 l
M
) are very low and it
remains controversial whether they may be able to act
as significant inhibitors. Therefore, whether the finding
that methyl citrate might be the c ausative agent o f
propionate toxicity in Salmonella enterica [9] is also true
for eukaryotic cells, is questionable. Nevertheless, the
identification of these isomers b y GLC/MS i s used for
diagnosis of disorders in human propionate metabolism

propionyl-CoA in rat liver hepatocytes led to a decrease in
the activity of pyruvate dehydrogenase [13].
In this investigation we e xamined c arbon balances under
different growth c onditions. We found that growth of
A. nidulan s on glucose + propionate, especially of the
DmcsA strain, led to the excretion of pyruvate a nd to high
intracellular c oncentrations of propionyl-CoA, which inhi-
bited pyruvate d ehydrogenase, succinyl-CoA s ynthetase
(GDP forming) and ATP-citrate lyase. We conclude that
these observations can explain the toxicity of propionate
towards cells growing on glucose as sole carbon and energy
source. Furthermore, we were able to show a correlation
between inhibition of polyketide formation and intracellular
propionyl-CoA content.
Experimental procedures
Materials
Chemicals were from Sigma-Aldrich. Enzymes u sed for
determination of acetate, gluco se and pyruvate were fr om
Roche. Columns a nd chromatographic m edia were, if not
otherwise indicated, from Amersham Pharmacia Biotech.
A. nidulans
strains, growth conditions and carbon
balances
The A. nidulans strainsusedinthisstudyarelistedin
Table 1 . Supplem ented minimal and complete media
were prepared as described previously [14]. For the deter-
mination of specific enzyme activities on different carbon
sources, growth t imes were strain and m edium specific.
Approximately 10
8

enzymatically. On 100 m
M
acetate and 100 m
M
acet-
ate + 100 m
M
propionate all strains, with the exception
of strain SMB/acuA, were grown for 36 and 41 h,
respectively. To determine enzyme activities during growth
on 100 m
M
propionate, we added 10 m
M
glucose to the
medium to support initial growth. After total c onsumption
of glucose cells were grown further for at least 12 h.
Therefore, the w ild-type strain was grown f or 42 h, whereas
the methylcitrate synthase deletion strain and the facB
multi-copy strain were incubated for 94 h. Strain SMB/
acuA was always grown in t he presence of glucose, because
the strain did not grow on acetate and growth on
acetate/propionate was very poor. Therefore, we used
the following composition of media and g rowth t imes:
10 m
M
glucose + 100 m
M
acetate harvest after 27 h ;
10 m

NaOH followed by
sterile water to avoid the t ransfer o f NaOH t o t he growth
medium. The CO
2
produced was trapped in a fourth wash
bottle containing 400 mL 0.2
M
Ba(OH)
2
. The insoluble
BaCO
3
that formed was dried at 60 °C for 20 h a nd weighed.
Residual glucose and acetate contents in the g rowth medium
were determined by enzymatic methods (see below). The
mycelium was pressed to remove any liquid, frozen with
liquid n itrogen, lyophilized, w eighed, and ground to a fine
powder. The CHN content o f the mycelium was determined
by elemental analysis (Zentrale Routineanalytik, Philipps-
Universita
¨
t Marburg, Lahnberge, Germany). Results from
Table 1. A. nidulans strains used in this s tudy. Strain RYQ11 was used throughout all e xperiments. Strain SDmcsA1 was used in a previous work
was t aken as a control t o confirm the re sults of s pore co lour formation, enzyme activities and carbon consumption.
Strain Genotype Source
SMB/acuA facA303, yA2; veA1 [2]
MH2671 pabaA1; prn-309, cnxJ1 [46]
Fab4-J3 MH2671 cotransformed with pFAB4 and pAN222 (approx. 4–8 copies facB) [46]
A637 yA2, pabaA1, pdhA1 FGSC, Kansas City, KS, USA
A634 yA2, pabaA1; pdhB4 FGSC, Kansas City, KS, USA

M
K
2
CO
3
. A fter incubation on ice for 15 min m ost of t he
perchloric acid was precipitated as insoluble KClO
4
.The
solution w as c entrifuge d at 120 00 0 g for 25 min and
the supernatant was collected. For concentration and
partial purification of the CoA-thioesters, the supernatant
was applied on a C18-cartridge ( Chromafix C18 ec
Ò
,
510 m g; Macherey-Nagel, Du
¨
ren, Germany), previously
rinsed with methanol and washed w ith 0.1% trifluoroacetic
acid. The supernatant was slowly applied to the column and
washed with 10 mL 0.1% trifluoroacetic acid. Elution was
carried out with 1.5 mL 50% acetonitrile/0.1% trifluoro-
acetic acid and samples were collected in 2-mL micro
centrifuge c ups. T he acetonitrile was evaporated in a Speed
Vac Concentrator (Bachofer GmbH, Reutlingen, Germany)
without heating and the residual volume o f 200–500 lLwas
measured with an accuracy of ± 2 lL using a micropipette.
An aliquot of the samples was used for the enzymatic
determination of acetyl-CoA and propionyl-CoA concen-
trations.

Amicon chamber o ve r a PM 30 membrane (Millipore,
Eschborn, Germany). Purity was sufficient for inhibition
studies. Succinyl-CoA synthetase was partially purified as
described above, except that buffer A did not contain
dithiothreitol. No f urther column purification was necessary
for the described activity measurements.
Enzymatic determination of glucose, acetate and
pyruvate in the growth medium
Glucose concentrations were determined by the combined
action of glucose o xidase (GOD, from A. niger), peroxidase
(POD, from horseradish) and 2,2¢-azinobis(3-ethylbenzo-6-
thiazolinesulfonic acid). The test was a modification of a
described procedure [ 15]. The composition of the test
reagent was: 130 m
M
sodium phosphate, pH 7.0; 400 U
POD (2 mg; 200 UÆmg
)1
), 80 0 U GOD (4 m g; 200 U Æmg
)1
)
and 2 5 mg 2,2¢-azinobis(3-ethylbenzo-6-thiazolinesulfonic
acid), final volume 50 mL. Each assay, which contained
900 lL rea gent and 100 lL sample, was incubated for
15 min at 3 7 °C and measured at 436 nm in a spectropho-
tometer. The a ssay was linear in a range of 0–30 l
M
glucose.
A standard was run for every freshly prepared reagent.
Pyruvate concentrations were determined by the use of

phosphate, pH 7.0; 10
L
-malate, 0.2 CoASH, 2 NAD
+
,
2ATP, 4MgCl
2
, 0.5 dithiothreitol, 0.5 U MDH, 0 .5 U
citrate synthase, 0 .1 U ACS and 50–100 lL diluted
medium. All components were added with t he exception
of MDH a nd citrate synthase and the r esulting absorbance
at 340 nm was measured (A
1
). MDH was added and the
absorbance after r eaching t he equilibrium was taken as A
2
.
Citrate synthase was added and the reaction w as monitored
until no further c hange in absorbance w as visible (A
3
).
Concentrations were calculated by the formula below
[e, absorbance (extinction) coefficient; d, length of light
path of the cuvette], which con siders the decrease of the
concentration of oxaloacetate in equilibrium with
L
-malate
during t he formation o f NADH (the concentrations of
malate and NAD
+

ML
-malate, 2 m
M
NAD, 0.5 U MDH, 0.5 U citrate
synthase, 0.5 U methylcitrate synthase and 50–100 lL
sample. The concentration of acetyl-CoA was determined
first by the u se of citrate synthase. The reaction w as
followed at 340 nm until no further change in absorbance
was detected. Methylcitrate synthase was added and the
second change in absorbance was monitored.
The second method was based on the formation of a
nitrothiophenolate (2-mercapto-5-nitrobenzoate dianion)
during the reaction of 5,5 ¢-dithiobis-(2-nitrobenzoate)
(DTNB) with CoAS H, which was released during the
condensation of oxaloacetate with acetyl-CoA or propio-
nyl-CoA. The assay contained, in a final volume of 1 mL,
50 m
M
Tris/HCl, pH 8.0; 1 m
M
oxaloacetate, 1 m
M
DTNB,
0.5 U citrate synthase, 0.5 U methylcitrate synthase and
20–100 lL sample. Change in absorbance was m onitored at
412 nm; e ¼ 14.2 m
M
)1
Æcm
)1

M
MgCl
2
,5m
M
potassium phosphate, enzyme sample and
water to a final volume of 1 mL. One unit of e nzyme a ctivity
was defined as th e amount of enzyme producing 1 lmol
CoASHÆmin
)1
under the assay conditions.
Isocitrate lyase. The assay [ 22] contained (m
M
) 50 potas-
sium phosphate, pH 7.0; 1 threo-isocitrate, 10 phenylhydr-
azine HCl, 2 dithiothreitol, 2 MgCl
2
, 10–100 lLenzyme
sample and water to a final v olume of 1 mL. T he formation
of glyoxylate phenylhydrazone was followed at 324 n m;
e ¼ 16.8 m
M
)1
Æcm
)1
. O ne unit of enzyme a ctivity w as
defined as the amount of enzyme producing 1 lmol
glyoxylate phenylhydrazoneÆmin
)1
under the assay condi-

producing 1 lmol CoASHÆmin
)1
under the assay condi-
tions.
Pyruvate dehydrogenase. Pyruvate dehydrogenase (PDH)
activity was measured according t o a pro cedure described
previously [23] with some modifications. The assay con-
tained (in m
M
), in a final volume of 1 mL, 50 T ris/HCl,
pH 8.0; 2 pyruvate, 0.8 thiamine pyrophosphate, 2.5 cys-
teine/HCl, 2 NAD, 2 MgCl
2
, cell-free extract and water t o a
final volume of 990 lL. The reaction was started by the
addition of 0.02–0.17 m
M
CoASH and reduction of NAD
+
to NADH was followed at 340 nm. One unit of enzyme
activity was defined as the amount of enzyme producing
1 lmol NADHÆmin
)1
under t he assay c onditions. The
activity of 2-oxoglutarate dehydrogenase was determined by
the analogous procedure, in which pyruvate was replaced by
2-oxoglutarate [24].
Acetyl-CoA synthetase. Acetyl-CoA synthetase activity
was determined in a coupled assay by the use of MDH
and citrate synthase. In this method the acetyl-CoA

accumulation of NADH [25]. Lineweaver–Burk diagrams
were obtained b y use of the worksheet of the program
EXEL
98 (Microsoft Inc.).
Propionyl-CoA synthetase. Propion yl-CoA synthetase
activity was determined by the same method as described
for the determination of a cetyl-CoA synthetase activity,
except s odium acetate was replaced by sodium propionate
and c itrate synthase by methylcitrate s ynthase ( 0.8 U) f rom
A. nidulan s [2].
CoA-Transferase. CoA-Transferase activity was deter-
mined by u sing succinyl-CoA or propionyl-CoA as the
CoA-donor and acetate or propionate as the acceptor.
3230 M. Brock and W. Buckel (Eur. J. Biochem. 271) Ó FEBS 2004
When acetate was the acceptor the assay was monito re d b y
the use of citrate s ynthase, which released CoASH upon the
condensation of n ewly generated acetyl-CoA with oxalo-
acetate as described for t he determination of citrate synthase
activity. When p ropionate w as used as the acceptor, purified
methylcitrate synthase was used to measure the CoASH
release upon the condensation of p ropionyl-CoA w ith
oxaloacetate as described for t he determination of methyl-
citrate synthase activity. A t ypical ass ay contained ( in m
M
),
in a final volume of 1 mL, 50 M ops, pH 7.5; 0.4 CoA-
donor (succinyl-CoA or propionyl-CoA, respectively), 2 U
citrate synthase or 0.8 U m ethylcitrate synthase, r espect-
ively, 1 oxaloacetate and 10 C oA-acceptor (acetate or
propionate, respectively) and cell-free extract.

KCl and transferred to fresh glucose minimal medium
containing cycloheximide (200 lgÆmL
)1
), which inhibits
eucaryotic protein biosynthesis. The c ultures w ere in cubated
for further 9 h at 37 °C and 240 r.p.m . The mycelium was
dried and the biomass was compared to that of control
samples. Glucose concentrations before and after the
incubation with cycloheximide were measured as described
above.
Results
Carbon balances on different growth media
Initial experiments s howed that growth on glucose + prop-
ionate resulted in significant excr etions of pyruvate i nto t he
medium (Table 2). I n o rder to exclude substantial excretions
of other c arbon compounds, w e measured t he total c arbon
balances of wild-type and methylcitrate synthase deletion
strain (DmcsA). Therefore, the consumption of substrates,
formation of CO
2
, as well a s excretion of pyruvate and t he
final pH were determined in media in which c ells had been
grown on different carbon sources. The measured carbon
balances add up to almost 100% (T able 3) indicating that
there was no substantial excretion of compounds other than
CO
2
and pyruvate or a significant consumption of prop-
ionate. The increase in the final pH (Table 2 ) correlated
with the consumption of t he carboxylates, by w hich protons

of mycelium)
Glucose consumption
(mmol/g)
Acetate consumption
(mmol/g)
Pyruvate excretion
(mmol/g)
Growth time
(h)
Wild-type * Glucose (6.6) 10.6 – 0.140 20
DmcsA * Glucose (6.7) 9.9 – 0.054 20
Wild-type Glucose/acetate (7.3) 10.0 1.0 0.070 22
DmcsA Glucose/acetate (7.9) 8.0 9.0 0.087 22
Wild-type * Glucose/propionate (6.8) 14.6 – 1.27 44
DmcsA * Glucose/propionate (6.3) 16.2 – 2.21 72
Wild-type Glucose/propionate/acetate (7.5) 7.0 24 0.37 30
DmcsA Glucose/propionate/acetate (7.4) 11.0 19 0.50 30
Wild-type Acetate/propionate (8.0) – 49 0.144 47
DmcsA Acetate/propionate (8.5) – 62 0.035 47
Wild-type Acetate (8.2) – 54 < 0.01 40
DmcsA Acetate (8.2) – 55 < 0.01 40
Ó FEBS 2004 Propionyl-CoA inhibits glucose metabolism (Eur. J. Biochem. 271) 3231
Addition of acetate to a medium contai ning glucose did
not change the growth rate significantly, but the lack of
methylcitrate synthase i n the mutant strain induced acetate
consumption (Table 2). This observation is similar t o strain
Fab4-J3, which carries multiple copies of the transcriptional
activator FacB of the acetate utilization genes. FacB is
induced by a cetate and acet ylcarnitine [27]. G rowth experi-
ments w ith s train Fab4-J3 revealed that in the p resence of

2
(Table 3).
Probably the increase in CO
2
production caused by
propionate (doubled with the w ild-type and tripled with
the mutant) was due to energy production required for
maintenance ( see b elow) d uring the extended g rowth times.
Upon addition of acetate to the media containing glucose
and propionate, the growth rate of both strains increased
and the effect of propionate became less apparent. Finally,
in media containing acetate and propionate but no glucose,
there was only a small delay ( 30%) in growth of the mutant
as compared to the wild-type [2]. The higher acetate
consumption of the mutant strain was probably due to
higher maintenance requirement (see below) or to the action
of a CoA-transferase, which is induced by propionate and
seems to transfer the CoA-moiety from succinyl-CoA
preferentially to acetate (see below and Table 5).
The observed excretion of p yruvate prompted u s to c heck
strains, in each of which another of the three genes encoding
pyruvate dehydrogenase [29] was mutated (A637, pdhA1-
mutant ¼ lipoate acetyltransferase; A634, pdhB4 ¼ b-sub-
unit of p yruvate decarboxylase; A627, pdhC1 ¼ a-subunit
of pyruvate decarboxylase). All three strains were unable to
grow on glucose or propionate, but grew well on acetate.
Growth of strain A627 on 50 m
M
acetate yielded 239 mg
dried m ycelium a fter 23 h (59 mmol acetateÆg myc elium

biosynthesis. The mycelium of pregrown cultures was
washed and tran sferred to f resh medium containing
cycloheximide (200 lgÆmL
)1
), which was sufficient to
prevent biomass formation. Cultures were incubated for
8 h and dry mass as well as glucose consumption was
determined. I n this e xperiment significant g lucose con-
sumption was observed ( 8.75 ± 0.1 mmolÆh
)1
Ægdried
cells
)1
). We conclude that indeed the prolonged growth
time of both the wild-type and DmcsA strains on glucose/
propionate medium led to the increased consumption of
glucose as determined.
Intracellular acetyl-CoA and propionyl-CoA contents
To investigate w hether propionyl-CoA a ccumulates in t he
methylcitrate s ynthase deletion strain during growth on
Table 3. Carbon balances of wild-type and DmcsA strain. Balan ces are calculated f or 1 g of dried m yce lium. The concentrations of the sub strates are
indicated in Table 2 (marked by asterisks). The wil d-type strain was SMI45 and DmcsA strains were RYQ11 and SDmcsA1.
Strain/C-source
Glucose
consumed
(mmol C)
Pyruvate
(mmol C)
CO
2

Acs A100 153 ± 5 205 ± 10 124 ± 2 17 ± 1
Acs G10/P100 133 ± 4 128 ± 10 150 ± 10 22 ± 2
Acs A100/P100 135 ± 10 289 ± 15 167 ± 10 18 ± 2
Pcs G50 10 ± 1 9 ± 1 8 ± 1 1 ± 0.5
Pcs G50/A100 16 ± 2 50 ± 5 13 ± 2 2.3 ± 0.2
Pcs G50/P100 10 ± 1 21 ± 1 10 ± 1 6 ± 0.4
Pcs G50/A100/P100 26 ± 2 38 ± 4 13 ± 1 3.5 ± 0.5
Pcs A100 58 ± 1 67 ± 3 42 ± 2 29 ± 2
Pcs G10/P100 77 ± 2 63 ± 3 76 ± 6 31 ± 1
Pcs A100/P100 59 ± 1 90 ± 2 74 ± 1 30 ± 1
Icl G50 0.2 ± 0.1 0.1 ± 0.1 0.2 ± 0.1 0.6 ± 0.2
Icl G50/A100 23 ± 1 108 ± 4 14 ± 2 7 ± 1
Icl G50/P100 35 ± 2 62 ± 9 41 ± 1 24 ± 2
Icl G50/A100/P100 85 ± 3 170 ± 5 34 ± 3 26 ± 1
Icl A100 86 ± 5 225 ± 5 63 ± 3 71 ± 4
Icl G10/P100 130 ± 5 107 ± 7 294 ± 10 67 ± 5
Icl A100/P100 161 ± 1 287 ± 15 180 ± 10 82 ± 5
Micl G50 7 ± 1 6 ± 2 6 ± 2 4 ± 1
Micl G50/A100 10 ± 1 12 ± 1 9 ± 2 11 ± 1
Micl G50/P100 30 ± 2 31 ± 2 62 ± 4 44 ± 1
Micl G50/A100/P100 26 ± 1 27 ± 1 29 ± 2 33 ± 1
Micl A100 26 ± 1 20 ± 1 29 ± 2 24 ± 1
Micl G10/P100 74 ± 5 28 ± 2 132 ± 1 64 ± 1
Micl A100/P100 35 ± 1 36 ± 2 46 ± 1 63 ± 3
McsA G50 1 ± 0 2 ± 1 0 1 ± 0
McsA G50/A100 5 ± 1 14 ± 2 0 7 ± 1
McsA G50/P100 55 ± 2 52 ± 2 0 57 ± 4
McsA G50/A100/P100 37 ± 1 38 ± 1 0 41 ± 1
McsA A100 38 ± 2 20 ± 1 0 42 ± 3
McsA G10/P100 147 ± 6 72 ± 3 0 153 ± 5

The suitability of this method was checked by mixing
16.5 nmol acetyl-CoA and 1 6.1 nmol of propionyl-CoA
and performing the identical procedure as for the partial
purification of the acyl-CoA e ster from lyophilized
mycelium, including addition of perchloric acid , neutral-
ization, centrifugation, C
18
-cartridge and c oncentration.
The recovery was 15.1 nmol (91.5%) acetyl-CoA and
14.4 nmol (89.5%) propionyl-CoA w hich showed that the
method gave reliable results. Therefore we can conclude
that the r atio between a cetyl-CoA and p ropionyl-CoA
remained constant during the procedure and the total
yield was about 90% assuming that all cells were opened
by the procedure described above.
After 2 0 h of growth on glucose as the sole carbon
source, neither the wild-type nor the methylcitrate
synthase deletion strain showed significant accumulation
of propionyl-CoA (Fig. 1). Addition of propionate to the
glucose medium led to an increase of the propionyl-CoA
level in the wild-type strain. The methylcitrate synthase
deletion strain showed an up to tenfold higher accumu-
lation of propionyl-CoA under these conditions, as the
thioester cannot be oxidized further. Addition of acetate
to the glucose/propionate medium reduced the propionyl-
CoA level of the cells, whereas an increase was observed
again a fter growth on acetate + propionate without
glucose. Despite this high level of propionyl-CoA, which
was most probably due to an unspecific action of acetyl-
CoA synthetase (described below), only a slight growth

used this latter value for the calculation of the internal
propionyl-CoA c oncentration o f t he methylcitrate synthase
mutant and the wild-type after growth on 5 0 m
M
glu-
cose + 100 m
M
propionate. Thus the DmcsA strain accu-
mulated 0.21 m
M
propionyl-CoA, whereas i n t he wild-type
strain only 0.03 m
M
propionyl-CoA could be found.
Nevertheless, concentrations given here are just a simple
mathematical calculation. Due to t he very high concentra-
tion of macromolecules within the cell, accompanied by
high viscosity, local concentrations may differ from that
shown here. In addition, propionyl-CoA is supposed to be
generated i n t he cytoplasm. For transport to t he mitochon-
dria a conversion into a carnitine-ester and a back-
conversion to the CoA-ester inside the mitochondria has
to be involved, which is most likely performed by cytoplas-
mic and mitocho ndrial ac yl-carnitine transferases (AcuJ [33]
and FacC [27]). The transporter involved in t hat process is
most likely AcuH [34]. Mutants of the corresponding genes
were unable to grow on propionate as sole carbon and
energy source (data not shown). This transport m echanism
Fig. 1. Intracellular contents o f acetyl-CoA
and pro pionyl-CoA from A. nidulans wild-type

chondria and cytoplasm. Since the equilibrium constant
between these two propionate esters is close to 1.0, we
assume for our calculations that the concentration of
propionyl-CoA is similar in all compartments.
Formation of acetyl-CoA and propionyl-CoA
For the determination of the substrate specificity of acetyl-
CoA synthetase and a putative propionyl-CoA s ynthetase
we u sed the a cetate-grown strain Fab4-J3 and g lucose/
propionate grown S MB/acuA cells (10 m
M
glucose/100 m
M
propionate; 29 h ). The high expression of the acetate
utilization g enes in th e Fab4-J3 strain seemed to be suitable
to measure mainly the acetate and propionate activating
activity of acetyl-CoA synthetase. I n comparison SMB/
acuA carries a defective acetyl-Co A synthetase gene, which
means that the activating activity must derive from alter-
native acyl-CoA synthetases, most likely a propionyl-CoA
synthetase.
The kinetic constants were determined with an extract
from acetate grow n Fab4-J3 cells with acetate as substrate:
V
max
¼ 20 5 mUÆmg
)1
protein and K
m
¼ 44 l
M

ison, an extract from propionate grown SMB/acuA cells
gave following values with acetate as substrate: V
max
¼
22 mUÆmg protein
)1
and K
m
¼ 880 l
M
(V
max
/K
m
¼ 25
UÆg
)1
Æm
M
)1
) and with propionate as substrate: V
max
¼
31 mU mg protein
)1
and K
m
¼ 90 l
M
(V

To determine the extent of acetate activation in compar-
ison to propionate activation in the p resence o f both
substrates we used the wild-type strain A26 grown o n a
medium containing 100 m
M
acetate + 100 m
M
propionate
(Table 6). The cell-free extract was used to determine the
inhibition of acetyl-CoA synthetase activity by propionate.
The acetyl-CoA formed was measured in a coupled assay
with citrate synthase, which displays no significant activity
with propionyl-CoA. Therefore we exclusively monitored
the activity for activation of acetate. In the presence of
0.5 m
M
acetate and 10 m
M
propionate (ratio 1 : 20) we
observed still 50% acetyl-CoA syntheta se activ ity. There-
fore we conclude that in a wild-type b ackground the
activation of acetate is much favoured over the activation of
propionate or, vice versa, acetate inhibits the formation of
propionyl-CoA. This observation readily explains the
decreased propionyl-CoA levels found in cells grown on
glucose/acetate/propionate as compared to glucose/pro-
pionate.
Inhibition of CoASH-dependent enzymes of glucose
metabolism
The high levels of propionyl-CoA in the mutant strain

for C oASH (7.2 l
M
)increasedinthe
presence of 0.1 m
M
propionyl-CoA 3 .6-fold (25 l
M
),
whereas V
max
was r educed only b y 3 0%, which demonstra-
ted a mainly competitive inhibition with an apparent K
i
of
% 50 l
M
. Addition of high concentrations of propionate
Table 6. Acetyl-CoA synthetase activity from wild-type strain A26
grown on 100 m
M
acetate + 100 m
M
propionate. 100% acetyl-CoA
synthetase activity refers to (135 ± 10) mUÆmg protein
)1
.
Substrates
Activity
(%)
Acetate

CoA synthetase more precisely, we partially purified both
enzymes by c hromatography over a Q-Sepharose column.
Inhibition of ATP citrate lyase by a cetyl-CoA, propionyl-
CoA and butyryl-CoA was measured by addition of
different concentrations of single acyl-CoA to the in vitro
assay i n t he presence of 0.34 m
M
CoASH. Activity without
addition of acyl-CoA (10 mUÆmL
)1
)wassetto100%
(Fig. 2 A). Propionyl-CoA showed the strongest inhibitory
effect, followed by acetyl-CoA and butyryl-CoA.
Succinyl-CoA synthetase. Succinyl-CoA s ynthetase
(10 mUÆmL
)1
) was assayed with succinyl-CoA, inorganic
phosphate and GDP by trapping the liberated CoASH w ith
5,5¢-dithiobis-2-nitrobenzoate (Fig. 2 B). At concentrations
of 0.4 m
M
acetyl-CoA or 0.4 m
M
propionyl-CoA the
succinyl-CoA synthetase was inhibited by 70%. A combi-
nation of 0.2 m
M
acetyl-CoA and 0.4 m
M
propionyl-CoA,

succinyl-CoA + propionate and propionyl-CoA + acetate
as substrates (Table 4). In both strains highest CoA-
transferase activity was determined by use of succinyl-
CoA as the CoA-donor and a cetate as the acceptor,
followed by the transfer from succinyl-CoA to propionate
(% 35% of the former activity) and the transfer from
propionyl-CoA to acetate (% 11%). The enzyme was
most active in strains grown in the presence of propionate
and always higher in the DmcsA strain as compared to
the wild-type. These CoA-transferase levels resemble the
expression pattern o f th e gene encoding 2-methylisocitrate
lyase, a specific enzyme of the m ethylcitric a cid cycle
(compare Tab le 4 to Table 3). Ther efore, we conclude
that an efficient transfer of the CoA-moiety f rom
succinyl-CoA to acetate in the presence of both acetate
and propionate is possible. In addition this might explain
the low accumulation of propionyl-CoA during growth
on glucose/acetate/propionate medium especially of the
Dmc sA strain, which is consistent with the higher growth
rate and the elevated acetate consumption of both strains
(Table 1). In the absence o f acetate (glucose/propionate
medium) t he CoA-moiety, however, can only be trans-
Fig. 2. Inhibition of ATP c itrate lyase (A) and
succinyl-CoA synthe tase (B) f rom A. nidulans
by differ ent CoA-thioesters. Both enzymes
were partially purified by chroma tography
over Q- Sepharose. Ac tivity without ad dition
of CoA-thioesters ( % 10 m UÆmL
)1
)wassetas

processes such as fatty acid and nucleotide s ynthesis. If only
NADPH is required, glucose can be completely oxidized via
this pa thwa y to CO
2
without ATP formation. It was
demonstrated that glucose-6-phosphate dehydrogenase, the
first enzyme of this pathway, is essential for the viability of
fungal cells, most likely due to its important biosynthetic
role [37,38]. As shown in Table 7, A. nidulans contains
relatively high amounts of glucose-6-phosphate dehydro-
genase and gluconate-6-phosphate dehydrogenase, which
weremeasuredtogetherinthesameassay.Thedataindicate
that the p resence of propionate in the medium reduces the
activity by % 50% in the w ild-type as well a s in the DmcsA
strain. T herefore it appears unlikely that a n enhanced
oxidation of glucose via the pentose phosphate cycle is
responsible for the observed uncoupling of glucose oxida-
tion and g rowth inhibition caused in the presence of
propionate.
Correlation of spore colour formation to propionyl-CoA
levels and enzymatic activities
The spore c olour of conidia from A. nidulans derives f rom
the polyketide naphtopyrone [39]. We have assumed a
negative effect of propionyl-CoA on spore colour formation
in an earlier study, w ithout the knowledge about the
accumulation of propionyl-CoA [2]. Recently, by screen ing
for A. nidulans mut ants with a defect in the synthesis of the
polyketide sterigmatocystin (ST) a methylcitrate s ynthase
deletion strain was i dentified. Further analysis of this
mutant showed that it was not only disturbed in ST

activity. The fa cB multi-co py strain shows elevated propio-
nyl-CoA synthetase a ctivity w ithout increasing methyl-
citrate synthase activity and th erefore reacts more sensitive ly
than the wild-type. However, it is noteworthy that strain
Fab4-J3 in c omparison to the wild-type s hows similar
activities of Acs, Pcs and Icl on propionate med ium (G10/
P100 of Table 3) but reduced levels of propionate specific
enzyme activities such as methylcitrate synthase and meth-
ylisocitrate lyase ( a canditate gene is AN8755.2 from the
conceptual translation of the A. nid ulans genome, which
shows 46% ide ntity to the m e thylisocitrate lyase f rom
Sc. cerevisiae ). This implies that the activating effect on
glyoxylate cycle enzymes mediated by propionate is FacB
independent and furthermore, higher basal levels of FacB
seem to ha ve a n egative effect on methylcitrate cycle
enzymes.
The inability of the methylcitrate synthase mutant to
remove propionyl-CoA via the m ethylcitrate pathway l eads
to loss of spore colour formation even at low propionate
concentrations. A s s hown i n lines VII a nd IX of Fig. 3, the
addition of acetate to glucose/propionate medium releases
suppression of spore c olour formation especially in the
methylcitrate synthase mutant and the w ild-type. The facB
multi-copy strain Fab4-J3, however, i s inhibited e ven more.
This strain shows strongly increased acetyl-CoA and
Table 7. Determination of the o xidative steps o f the pentose phosphate
pathway. Wild-type and DmcsAweregrownondifferentcarbon
sources an d t he combined activity o f glucose-6-phosphate dehyd ro-
genase and g luco nate-6-pho sphate dehydrogenase was determined.
One u nit (U) is defined as the reduction of 1 lmol of NADP

100 m
M
acetate and 100 m
M
propionate, the levels rose to
20 nmol acetyl-CoA and 66 nmol propionyl-CoA Ægdry
weight
)1
(ratio 1 : 3.3). Furthermore lines VII and VIII
show that this strain behaved very similarly on media
without glucose, which i s in agreement w ith the observation
that the strain hardly uses glucose and acetate in parallel (see
section entitled Carbon balances on different growth
media). Nevertheless, the spore colour of strain Fab4-J3 in
lanes IX and X is hard t o visualize, because the number o f
spores at these g rowth conditions is greatly reduced. This is
also true for the DmcsA-strain on G50/P100, which i mplies
that at high propionyl-CoA concentrations not only spore
colour formation but also conidiation is affected.
Utilization of acetate by the acetyl-CoA synthetase
mutant is s trictly dependent on the activity of t he
predicted propionyl-CoA synthetase. Strain SMB/acuA
shows better growth on media containing only 10 m
M
propionate and 50 m
M
acetate instead of equimolar
Fig. 3. Spore colour formation of different A. nidulans str ain s. Growth conditio ns a re g iven on the right (G, g lu cose; P , p ropiona te, A , a cetate; e .g.
G50/P10 ¼ the medium contained 50 m
M

Æm
M
)1
). From these re-
sults we conclude that the acetyl-CoA/propionyl-CoA
ratio and also the ability to activate propionate to
propionyl-CoA has to be well balanced with the methyl-
citrate synthase activity for successful spore colour
formation and growth.
The data in Table 3 further imply that propionyl-CoA
might be a direct inducer of methylcitrate cycle genes. On
media G50/P100 and G10/P100, which lead to a strong
accumulation of propionyl-CoA in the DmcsA-strain, the
activity of methylisocitrate lyase is twice as high than that of
the wild-type. The addition of acetate to these media not
only lowered the propionyl-CoA l evel, but also that of
methylisocitrate lyase activity. Therefore, a putative tran-
scriptional a ctivator of the methylcitrate cycle genes see ms
to be activated by propionyl-CoA (or propionyl-carnitine)
rather than by methylcitrate as suggested for the procaryotic
regulator of the propionate utilization genes from
S. typhimurium [41].
Discussion
Growth of A. nidulans on glucose medium is inhibited by
propionate in a concentration-dependent manner. In a
strain carrying a defective methylcitrate synthase g ene, this
effect is even much more pronounced. When a cetate was the
main carbon source, addition of propionate had no growth
inhibitory effect on the w ild-type and little effect on the
methylcitrate synthase deletion strain. One might assume

have to focus on the ac tivity p attern of this enzyme on
propionate containing media. We cannot evaluate the d irect
effect of a p artial inhibition of ATP citrate lyase b y
propionyl-CoA on the metabolism, because this enzyme is
not involved in glucose degradation. It is also not clear
whether inhibition of ATP citrate lyase indirectly diminishes
polyketide synthesis or whether a direct interaction of
propionyl-CoA with polyketide s ynthetase i s responsible for
this effect.
Succinyl-CoA synthetase is directly involved in the
degradation of g lucose, acetate and p ropionate via the
Krebs c ycle. Therefore an inhibition of this enzyme would
block the oxidation of all three substrates, which was not
observed with a cetate. An elegant way to bypass the
inhibition of this synthetase is the transfer of the CoA-
moiety from succ inyl-CoA to either acetate or p ropionate.
We were able to show the existence of such a CoA-
transferase, which indeed seems to be i nduced by propionate
but prefers acetate to propionate as CoA acceptor (Table 4).
Hence, the CoA-transferase explains the higher growth rate,
which was always observed when acetate was added to a
medium containing propionate. In the absence of acetate,
however, t he transferase enhances t he formation of p ropio-
nyl-CoA, which traps the system into a loop.
A very i mportant inhibition is attributed to the p yruvate
dehydrogenase c omplex. The low K
i
of % 50 l
M
propionyl-

(dehydration, lethargy, nausea and vomiting as well as a
risk for neurologic sequelae) might be caused not only by
metabolites derived from propionyl-CoA a s are propionate,
Ó FEBS 2004 Propionyl-CoA inhibits glucose metabolism (Eur. J. Biochem. 271) 3239
b-hydroxypropionate, b-hydroxybutyrate, methylmalonyl-
CoA and methylcitrate, but also directly by propionyl-CoA
inhibiting pyruvate dehydrogenase as described in this
study.
Besides the impairment caused by propionyl-CoA we
cannot exclude a depletion of free CoASH, which would
also lead to a strong disturbance o f the metabolism a nd a
reduction of pyruvate oxidation. However, the fact t hat the
DmcsA strain also accumulates significant amounts of
propionyl-CoA on acetate/propionate medium without
showing a significant reduction in biomass formation
compared to acetate as sole carbon source [2] seems to
exclude this effect.
In order to get further i nsights into the mechanism of
growth inhibition mediated by propionate, future work will
focus o n the phenotypic characterization of other mutants
carrying defe ctive g enes of the methylcitrate cycle. Analysis
of the fatty acid composition f rom the DmcsA strain grown
on different c arbon sources m ight also give an insight into
substrate specificity of acetyl-CoA carboxylase and fatty
acid synthases, depending on the existence of branched and
odd chain fatty acids. Furthermore, we are trying to identify
and purify the transcriptional activator of the propionate
utilization genes and analyse its DNA recognition sequence.
Knowledge of t his sequence w ill facilitate the screening of
other promoters for putative regulation by propionate,

citrate synthase from Aspergillus nidulans: implications for p ro-
pionate as an antifungal age nt. Mol. Microbiol. 35, 961–973.
3. Brock, M., Darley, D., Textor, S. & B uckel, W. (2001) 2-Methy-
lisocitrate lyases fro m th e b acterium Escherichia coli and the
filamentous fu ngus Aspe rgillus nidulans: characterization and
comparison of both enzymes. Eu r. J. Biochem. 268, 3577–3586.
4. Armitt, S., McCullough, W. & Roberts, C.F. (1976) Analysis of
acetate non-utilizing (acu)mutantsinAspergillus nidulans. J. Gen.
Microbiol. 92, 263–282.
5. Sandeman, R.A. & Hynes, M.J. (1989) Isolation of the facA
(acetyl-coenzyme A synthetase) and acuE (ma late synthase) genes
of Aspergillus n idulans. Mol. Gen. Genet. 218, 8 7–92.
6. Brock, M., Maerk er, C ., Sc hutz, A ., Vo
¨
lker, U. & Buckel, W.
(2002) Oxidation o f p ropionate to pyruvate in Escherichia coli.
Involvement of m ethylcitrate dehydratase and aconitase. Eur. J.
Biochem. 26 9, 6184–6194.
7. Textor, S., Wendi sch, V.F., De Graaf, A.A., Mu
¨
ller, U., Linder,
M.I., Linder, D. & Buckel, W. (1997) Propionate oxidation in
Escherichia coli: evidence for operation of a methylcitrate cycle in
bacteria. Arch. M icrobiol. 168, 428 –436.
8. vanRooyen,J.P.,Mienie,L.J.,Erasmus,E.,DeWet,W.J.,Ket-
ting, D., Duran, M. & Wadman, S.K. (1994) Identification of the
stereoisomeric configurations of methylcitric acid produced by
si-citrate synthase and methylcitrate synthase using capillary gas
chromatography-mass spectrometry. J. Inherit. Metab. Dis. 17,
738–747.

of beta-NADH a nd beta-NADPH. Clin. C hem. 22, 1 51–160.
17. Buckel, W . & Eggerer, H. (1965) Zur optischen Bestimmung von
Citrat-Synthase und Acetyl-Coenzym A. Biochem. Zeitschr. 343,
29–43.
18. Buckel, W., Ziege rt, K. & E ggerer, H . (1973) Acetyl-CoA-
dependent cleavage of citrate on i nactivated citrate l yase. Eur. J.
Biochem. 37 , 295–304.
19. Riddles, P.W., Blakeley, R.L. & Zerner, B. (1979) Ellman’s rea-
gent: 5 ,5¢-dithiobis(2-nitrobenzoic acid) – a re-examination. Anal.
Biochem. 94 , 75–81.
20. Takeda, Y ., Suzuki , F. & Inoue, H. (1969) A TP: citrate lyase. In
Methods in Enzymology (Lowenstein, J.M., ed.), pp. 153–163.
Academic Press I nc, London.
21. Srere, P.A. (1966) Citrate-condensing enzyme-oxaloacetate binary
complex. studies o n i ts physical and chemical properties. J. Biol.
Chem. 241, 2157–2160.
22. McFadden, B.A. (19 69) Isocitrate lyase. In Methods in Enzymol-
ogy (Lowenstein, J.M., ed.), pp. 163–170. Academic Press Inc,
London.
3240 M. Brock and W. Buckel (Eur. J. Biochem. 271) Ó FEBS 2004
23. Millar, A.H., Leaver, C.J. & Hill, S.A. (1999) Charac terization of
the d ihydrolipoamide acetyltransferase of the mitochondrial
pyruvate dehydrogenase complex from potato a nd comparisons
with sim ilar enzymes in diverse plant species. Eur. J. Biochem. 264,
973–981.
24. Reed, L.J. & Mukherjee, B.B. (1969) Alpha-ketoglutarate dehy-
drogenase c omplex from Escherichia coli.InMethods E nzymol.
(Lowenstein, J.M., e d.), pp. 55–61. A cademic P ress In c., Londo n.
25. Buckel, W. & Eggerer, H. (1969) Intramolecular nucleophilic
catalysis on t he hydro lysis of citryl-CoA. Hoppe S eylers Z. Phy-

tional analysis of mutations i n the human carnitin e/acylcarnitine
translocase in Aspergillus nidulans. Fungal Genet. Biol. 39,211–
220.
35. van den Berg, M .A., de Jong-Gubbels, P., Kortland, C.J., van
Dijken, J.P., Pronk, J.T. & Steensma, H.Y. (1996) The two ace tyl-
coenzyme A s yn thetases of Saccharomyces cerevisiae differ with
respect to kinetic properties and transcriptional regulation. J. Biol.
Chem. 271 , 28953–28959.
36. Horswill, A.R. & Escalante-Semerena, J.C. (1999) The prp Egene
of Salmonella typhimurium LT2 encodes propionyl-CoA s ynthe-
tase. Microbiology 145, 1381–1388.
37. van den Broek, P., Goosen, T., Wennekes, B . & van den Broek, H.
(1995) Isolation and characterization of the glu cose-6-ph osphate
dehydrogenase encoding g ene ( gsdA) f rom A spergillus niger. Mol.
Gen. Genet. 247, 229–239.
38. Loiudice, F.H., Silva, D.P., Zanchin, N.I., Oliveira, C.C. & P es-
soa, A. Jr (2001) Overexpression of glucose-6-phosphate dehy-
drogenase in genetically modified Saccharomyces c erevisiae. Appl.
Biochem. Biote chnol. 91–93, 161–169.
39. Watanabe, A ., Fujii, I., Sankawa, U., Mayorga, M.E., Timber-
lake, W.E. & Ebizuka, Y. (1999) Re-identification of Aspergillus
nidulans wA-gene for a polyketid e synthase of naphtho pyrone.
Tetrahedron Le tt. 40, 9 1–94.
40. Zhang, Y.Q. & K eller, N.P. (2004) Blockage of methylcitrate cycle
inhibits p olyketide production i n Aspergillus nidulans. Mol. Mi-
crobiol. 52, 541–550.
41. Tsang, A.W., Horswill, A.R. & Escalante-Semerena, J.C. (1998)
Studies of r egulation of expression of the propionate (prpBCDE)
operon provide insights into h ow Salmonella typhimurium LT2
integrates its 1,2-propanediol and propionate catab olic pathways.