Tài liệu Báo cáo Y học: Inactivation of the 2-oxo acid dehydrogenase complexes upon generation of intrinsic radical species pptx

Inactivation of the 2-oxo acid dehydrogenase complexes
upon generation of intrinsic radical species
Victoria I. Bunik
1
and Christian Sievers
2
1
A.N.Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow, Russia;
2
Physiological Chemistry
Institute of Eberhard-Karls-University, Tuebingen, Germany
Self-regulation of the 2-oxo acid dehydrogenase complexes
during catalysis was studied. Radical species as side products
of catalysis were detected by spin trapping, lucigenin fluor-
escence and ferricytochrome c reduction. Studies of the
complexes after converting the bound lipoate or FAD
cofactors to nonfunctional derivatives indicated that radicals
are generated via FAD. In the presence of oxygen, the 2-oxo
acid, CoA-dependent production of the superoxide anion
radical was detected. In the absence of oxygen, a protein-
bound radical concluded to be the thiyl radical of the
complex-bound dihydrolipoate was trapped by a-phenyl-
N-tert-butylnitrone. Another, carbon-centered, radical was
trapped in anaerobic reaction of the complex with 2-oxo-
glutarate and CoA by 5,5¢-dimethyl-1-pyrroline-N-oxide
(DMPO). Generation of radical species was accompanied by
the enzyme inactivation. A superoxide scavenger, superoxide
dismutase, did not protect the enzyme. However, a thiyl
radical scavenger, thioredoxin, prevented the inactivation. It
was concluded that the thiyl radical of the complex-bound
dihydrolipoate induces the inactivation by 1e

–
oxidation of
the 2-oxo acid dehydrogenase catalytic intermediate. A pro-
duct of this oxidation, the DMPO-trapped radical fragment
of the 2-oxo acid substrate, inactivates the first component of
the complex. The inactivation prevents transformation of
the 2-oxo acids in the absence of terminal substrate, NAD
+
.
The self-regulation is modulated by thioredoxin which alle-
viates the adverse effect of the dihydrolipoate intermediate,
thus stimulating production of reactive oxygen species by the
complexes. The data point to a dual pro-oxidant action of
the complex-bound dihydrolipoate, propagated through the
first and third component enzymes and controlled by thio-
redoxin and the (NAD
+
+ NADH) pool.
Keywords: dihydrolipoate; 2-oxo acid dehydrogenase com-
plex; reactive oxygen species; thiyl radical; thioredoxin.
The 2-oxo acid dehydrogenase complexes are key mito-
chondrial enzymes functioning at branch points of
metabolism. They catalyze irreversible oxidation of 2-oxo-
acids (pyruvate, 2-oxoglutarate or branched chain 2-oxo-
acids) yielding CO
2
, acyl-CoAs and NADH via reactions
1–5:
Correspondence to V. Bunik, A.N. Belozersky Institute of Physico-Chemical Biology, Moscow State University, Moscow 119899, Russia.
Tel.: + 7 095 939 14-56, Fax: + 7 095 939 31 81, E-mail:

Abbreviations: E1, 2-oxo acid dehydrogenase; E2, dihydrolipoamide acyltransferase; E3, dihydrolipoamide dehydrogenase; DTPA, diethylene-
triaminepentaacetic acid; DMPO, 5,5¢-dimethyl-1-pyrroline-N-oxide; MNP, 2-methyl-2-nitrosopropane; PBN, a-phenyl-N-tert-butylnitrone;
POBN, a-(4-pyridyl-1-oxide)-N-tert-butylnitrone; ROS, reactive oxygen species; SOD, superoxide dismutase; ThDP, thiamin diphosphate.
(Received 31 May 2002, accepted 23 August 2002)
Eur. J. Biochem. 269, 5004–5015 (2002) Ó FEBS 2002 doi:10.1046/j.1432-1033.2002.03204.x
Depending on the particular 2-oxo acid, R- is CH
3
-(pyru-
vate), HOOC-(CH
2
)
2
- (2-oxoglutarate), CH
3
-CH(CH)
3
-
CH
2
- (2-oxoisovaleriate). Multiple copies of the substrate-
specific 2-oxo acid dehydrogenase (E1), dihydrolipoamide
acyltranferase (E2), and dihydrolipoamide dehydrogenase
(E3) are organized in highly ordered structures [1–4]. E1
catalyzes the rate-limiting step of the whole process [5–7]
and the active sites are coupled through the interacting
network of the lipoyl moieties [3,8]. The two electrons in the
catalytic intermediate of reactions 4 and 5, the 2e
)
reduced
E3, initially were thought to be distributed between flavin

semiquinone and sulfur radical [9,10]. However, at low
temperatures this enzyme form is EPR-silent [11], and an
internal charge transfer complex between a thiolate anion
and oxidized FAD [12,13] is currently accepted as the best
description.
Although the physiological process of the 2-oxo acid
oxidative decarboxylation (Reactions 1–5) occurs through
2e
–
transfers, a number of related reactions involve radical
species as intermediates. The 1e
–
acceptor, ferricyanide, is
able to oxidize catalytic intermediates formed by all the
components of the 2-oxo acid dehydrogenase complexes. In
pyruvate:ferredoxin oxidoreductases [14–16] and in chemi-
cal models [17,18], hydroxyethylthiamine diphosphate (a
product of reaction 1 with pyruvate) and its analogs were
shown to undergo oxidation via a thiazolium radical. Redox
reactions involving radical species were demonstrated for
free and E3-bound FAD, including model reaction with the
lipoic acid radical [19,20] and the reduction of oxygen to
superoxide anion radical [21–23].
Involvement of such processes in catalysis by the 2-oxo
acid dehydrogenase complexes has not been previously
investigated. However, several considerations point to their
regulatory potential. First of all, the high catalytic power
and the significant contribution of E3 to the total flavin
content of mitochondria [24] suggest that the potential input
of complex-bound E3 to the mitochondrial production of

reactive oxygen species (ROS) should not be neglected. ROS
have attracted increasing attention as cellular messengers
involved in differentiation, apoptosis and aging [25–27].
Further, treatment of cells with H
2
O
2
results in selective
oxidation of the 2-oxo acid dehydrogenase complexes [28],
thatcouldalsotakeplacein vivo upon site-specific
production of ROS via the complex-bound FAD. In vitro
and in situ inactivation of the 2-oxo acid dehydrogenase
complexes upon addition of organic hydroperoxides [29–31]
or a superoxide anion radical-generating system [32,33]
correlate with specific targeting and/or impaired function of
the complexes observed in many disorders linked to
mitochondrial and cell damage. These disorders include
poisoning with environmental toxicants [34,35], Alzheimer’s
[36,37] and Parkinson’s [38] diseases, Wernicke–Korsakoff
syndrome [39] and others. However, the mechanisms of
oxidative damage of the complexes and cellular protection
against this damage are not properly understood.
Because the redox state of cellular thiols and disulfides is
an important factor in cellular protection against oxidative
damage, we suggested that the 2-oxo acid dehydrogenase
complexes may be significant not only as catalytic systems,
but also as microcompartments of important biological
thiols, lipoate and CoA. The oligomeric complex cores form
an inner cavity for CoA. Depending on the source and type,
the cores consist of 24 (cube) or 60 (pentagonal dodeca-

hedron) E2 subunits, with each E2 subunit bearing up to
three lipoate residues [2,3]. In vitro, more than a half of the
lipoyl residues of the E2 oligomer may be removed without
significant change in the overall activity [40–42]. If the
residues function only as catalytic intermediates, the reason
for their abundance is not clear. However, in view of the
antioxidant function of lipoate [43–45], its compartmentali-
zation within the complexes may be important for cellular
redox homeostasis through thiol-disulfide interchange.
Indeed, we have identified a flow of redox equivalents
between the complexes and the medium by means of thiol-
disulfide exchange reactions involving the dihydrolipoate
intermediate and cellular thiol-disulfide oxidoreductase,
thioredoxin [46–48]. Further study showed that the first
component of the complexes is inactivated under increased
steady-state concentration of the dihydrolipoate intermedi-
ate [49]. This was surprising, as thiols are usually protective
rather than inactivating. However, the pro-oxidant action
of dihydrolipoate is known [43–45,50,51] and could be
involved in the inactivation. The present study was under-
taken to investigate the relationship between the free radical
chemistry and function of the 2-oxo acid dehydrogenase
complexes. EPR and spin traps were used to study the
complex-catalyzed reactions under a variety of conditions,
including depletion of oxygen, presence of specific radical
scavengers, and selective inactivation of the complex
components. We report that the 2-oxo acid dehydrogenase
complexes produce several radical species of regulatory
significance.
MATERIALS AND METHODS

CoA, diphenyliodonium chloride, succinyl-CoA, b-D-glu-
cose, N-ethyl maleimide, R,S-lipoamide, MNP, DTPA,
glutathione disulfide, copper-zinc superoxide dismutase
(from bovine erythrocytes, 3500 UÆmg
)1
), glucose oxidase
(from Aspergillus niger, 250 UÆmg
)1
), cytochrome c (from
horse heart, type VI) were from Sigma (Deisenhofen,
Germany). 2-Oxoglutarate and pyruvate were from Merck
(Darmstadt, Germany). DMPO, PBN, POBN and lucigenin
were from Molecular Probes (Leiden, the Netherlands).
Catalase (from bovine liver, 260 000 UÆmg
)1
), ThDP,
NAD
+
, and NADH were from Roche Molecular Biochemi-
cals (Mannheim, Germany). Recombinant thioredoxin
from Escherichia coli was from Calbiochem-Novabiochem
GmbH (Bad Soden, Germany). R,S-Dihydrolipoamide was
obtained from R,S-lipoamide by sodium borohydride
reduction as in [47].
Enzyme isolation, assays and modification
2-oxoglutarate and pyruvate dehydrogenase complexes
from pig heart were isolated according to [52] with the
modifications described earlier [46]. The pyruvate dehy-
drogenase complex from E. coli wasisolatedasin[53].
Overall, E1 and E3 enzymatic activities were assayed

spectrophotometrically [49] by NAD
+
reduction, ferricya-
nide reduction and NADH oxidation, respectively. The E2
or E3 components of the 2-oxoglutarate dehydrogenase
complex were selectively modified at room temperature with
freshly prepared solutions of N-ethyl maleimide or diphe-
nyliodonium chloride according to [54] or [47]. N-Ethyl
maleimide (0.3 m
M
, final concentration) was added to the
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5005
mixture of complex (9 mgÆmL
)1
) containing 2-oxoglutarate,
ThDP, MgCl
2
and NAD
+
(1 m
M
each). After 20 min, the
complex activity was not measurable, while activities of the
E1 and E3 components were unchanged. Diphenyliodo-
nium chloride (4 m
M
, final concentration) was added to the
mixture of complex (9 mgÆmL
)1
)withNADH(2m

M
)in
0.1
M
potassium phosphate buffer, pH 7.0. The residual
activity of the E3 component after one hour of modification
was no more than 4% of its initial value. The modified
complexes were separated from the low molecular mass
compounds by desalting on a HiTrap
TM
5 mL column
(Pharmacia, Uppsala, Sweden) in 0.1
M
potassium phos-
phate, pH 7.0.
EPR spectroscopy
Room temperature EPR spectra were recorded in a quartz
flat EPR sample cell at X-Band using a Bruker 300E EPR
spectrometer (Karlsruhe, Germany). The spectrometer was
operated at modulation amplitude 1 mT, modulation
frequency 100 kHz, microwave energy 20 mW. The reac-
tions took place in 0.05
M
potassium phosphate buffer,
pH 7.0. The stock solutions of MNP, DMPO, POBN and
PBN were freshly prepared and kept protected from light.
Controls in each experiment indicated that none of the
components alone produced the EPR signals studied.
Model reactions of thiyl radical trapping PBN took place
in a reaction medium containing excess dithiothreitol and

glutathione in the presence of Ce
4+
, as the latter is known to
oxidize thiol groups through thiyl radicals. FAD radicals
were trapped by PBN in the reaction medium containing
excess FAD in the presence of dithionite. Anaerobic
conditions were created either enzymatically or in a Weidner
glove box (Hardegsen, Germany) connected to the MBraun
GmbH H
2
O/O
2
-Analyzer and Inert gas-System (Garching,
Germany), with the residual O
2
pressure in the glove box
not exceeding 2 p.p.m. To remove oxygen enzymatically,
the reaction mixtures were preincubated for 5 min with the
oxygen-scavenging system including glucose oxidase
(10 UÆmL
)1
), glucose (0.3
M
)andcatalase(26UÆmL
)1
).
Distribution of the EPR signal between the protein and
nonprotein fractions was investigated after precipitation of
2-oxo acid dehydrogenase complex by addition of 0.2 vol.
of 35% polyethylenglycol 6000 and solid ammonium sulfate

to 80% saturation. In anaerobic experiments, manipula-
tions of the samples were performed in the glove box. The
precipitating agents were added to the probe after recording
its EPR spectrum. The precipitated protein was centrifuged
for several minutes in a sealed Eppendorf tube, the pellet
washed with saturated ammonium sulfate and the centri-
fugation repeated. A stable and measurable EPR signal to
compare the reactions in the presence and absence of
O
2
under otherwise equal conditions was achieved by
employing high initial concentration of the 2-oxo acid
dehydrogenase complexes (30–200 mgÆmL
)1
), so that the
concentrations of substrates were comparable to that of the
protein redox centers.
Lucigenin-dependent fluorescence
This was measured with a Berthold LB 9505C Lumino-
meter (Germany). Reaction mixtures of 1 mL contained
2.5 mgÆmL
)1
2-oxoglutarate dehydrogenase complex,
27 l
M
lucigenin, 5 m
M
CoA, 5 m
M
2-oxoglutarate and

0.1 m
M
DTPA. The readings for substrates and for the
enzyme complex with lucigenin without substrates were
used as blanks.
Ferricytochrome
c
reduction
This was monitored at a cytochrome concentration of
16 l
M
by the increase at 550 nm due to formation of
ferrocytochrome c, using a molar extinction coefficient
of 20000
M
)1
Æcm
)1
. The reaction was carried out in 0.1
M
potassium phosphate buffer, pH 7.0, in the presence of
catalase (0.01 mgÆmL
)1
), EDTA (1 m
M
) and 2-oxoglutarate
dehydrogenase complex (1–2 mgÆmL
)1
, specific activity in
the NAD

+
reduction 2–3 lmolÆmin
)1
Æmg
)1
) catalyzing
transformation of either 2-oxoglutarate and CoA (2 m
M
each) or NADH (8 m
M
). The same mixture was used as a
blank except that SOD (0.016 mgÆmL
)1
) was added.
Reactions were started with CoA or NADH. Under these
conditions initial rates of cytochrome c reduction in the
presence of SOD were no more than 20% of those obtained
without SOD.
Enzyme inactivation studies
Time-dependent changes in the activity of the enzyme
complex upon preincubation with its substrates and/or
products was studied during a 20-min preincubation
period in 0.1
M
potassium phosphate buffer, pH 7.0, at
the following final concentrations: enzyme complex,
3mgÆmL
)1
; 2-oxoglutarate, 2 m
M

;CoA,0.05m
M
;succi-
nyl-CoA, 0.3 m
M
; dihydrolipoate, 0.3 m
M
; NADH, 10 m
M
.
Samples were withdrawn at various times during the
preincubation period and assayed for activity. Ferricya-
nide-reductase activity of E1 was measured in all cases
except those involving preincubation with NADH. Because
NADH interferes with the ferricyanide-dependent assay,
overall NAD
+
-reductase activity was measured in the latter
case and 1 m
M
EDTA was added to stabilize the overall
activity upon preincubation. Under the conditions of the
experiment, reversible inhibition of E3 by NADH did not
affect the activity measured. Thiol-containing compounds
did not interfere with the ferricyanide assay due to the
many-fold dilution of the preincubated mixture upon assay.
RESULTS
Trapping of radical species in the course of reactions
catalyzed by the 2-oxo acid dehydrogenase complexes
The spin trapping technique allows one to detect unstable

radical intermediates by converting them to more stable
radicals. Nitrone (PBN, POBN, DMPO) and nitroso
(MNP) spin traps are known to react with short-lived
radicals, resulting in relatively long-lived nitroxide radical
adducts. Together with the characteristic properties of the
adducts formed, differential reactivity of spin traps to
radicals enables selective trapping and identification of the
original radical species.
The spin trap MNP is presumed to efficiently form
adducts with catalytic radical intermediates, as it is small
enough to reach enzyme active sites without major steric
hindrance. Aerobic incubation of MNP with the pyruvate or
5006 V. I. Bunik and C. Sievers (Eur. J. Biochem. 269) Ó FEBS 2002
2-oxoglutarate dehydrogenase complexes and their respect-
ive 2-oxo acid substrates, CoA and NAD
+
, led to formation
of MNP/
•
H, or t-butyl-hydronitroxide. This product of one-
electron reduction [25] was detected from the four line EPR
spectrum with a
N
¼ a
H
¼ 1.44 mT and its characteristic
change observed in 50% D
2
O due to t-butyl-deuteronitrox-
ide (a

N
¼ 1.4 mT, a
D
¼ 0.22 mT). Specific modification of
the complex-bound FAD by diphenyliodonium chloride
prevented the appearance of paramagnetic species. Thus,
the E3-bound FAD catalyzes the 1e
–
oxidation of the
complex-bound dihydrolipoate intermediate by MNP.
MNP/
•
H was also formed in the reaction medium without
NAD
+
. This shows that the reaction may either substitute
or compete with the 2e
–
oxidation by the physiological
substrate, NAD
+
(Reactions 4–5).
Aerobic incubation of the spin trap PBN with the 2-oxo
acid dehydrogenase complexes and their substrates (2-oxo-
acid, CoA and NAD
+
)resultedinanEPRspectrum
characteristic of a PBN adduct with radical species. The
EPR signal (spectrum 1, Fig. 1) was that of a freely rotating
nitroxide, with each line of the nitrogen triplet split into a

doublet due to a hydrogen in the b-position to the nitrogen.
As in the reaction with MNP, the EPR signal persisted
after omitting NAD
+
from the reaction medium (Fig. 1,
spectrum 2). However, omitting either 2-oxo acid or CoA
(i.e. the components leading to the complex-bound dihydro-
lipoate) prevented the adduct formation.
Paramagnetic species were generated in an enzyme-
dependent manner both in the presence of the 2-oxo acid,
CoA (forward reaction) and when the 2-oxo acid dehy-
drogenase complexes catalyzed oxidation of NADH or
external dihydrolipoamide (backward reaction). In any
case, radicals were produced upon reduction of both the
E2-bound dihydrolipoate and E3. The particular contribu-
tion of these components to the production of radicals was
differentiated through their selective inactivation. N-Ethyl
maleimide modification of the lipoate residues of E2 led to
complete loss of the overall activity (Reactions 1–5) due to
E2 inactivation, while the E1 and E3 activities were fully
preserved. With this modified complex, no PBN adduct was
observed in the presence of 2-oxo acid and CoA, but it did
produce PBN adducts upon incubation with the E3
substrates, dihydrolipoate or NADH. Modification of the
tightly bound flavin cofactor of E3 with diphenyliodonium
chloride inactivated E3. This complex gave no PBN adduct
when incubated with either 2-oxo acid and CoA or NADH.
Thus, the E3-bound FAD was responsible for the formation
of radical species at the expense of either complex-bound or
free dihydrolipoate or NADH.

Action of the known radical scavengers and properties of
the adducts obtained were studied to identify original
radical species. No EPR signal was detected in the presence
of both SOD and catalase. SOD alone blocked the
appearance of the paramagnetic species during the first
20–25 min of the reaction, but the EPR signal developed
after the delay. The delayed signal was inhibited by
concomitant addition of SOD and the metal chelator
DTPA. These data show that initial PBN adducts are
dependent on the superoxide anion radical generated in the
system. The delayed paramagnetic species are due to
radicals arising in the presence of adventitious metal ions
from H
2
O
2
, a product of the superoxide dismutation. The
conclusion is supported by the essential role of the
E3-bound FAD in the radical production by the complexes,
as FAD is known to reduce oxygen to superoxide [21–23].
However, comparison of our experimental data to the data
on the previously identified adducts (Table 1) shows that
the stability and spectral characteristics of the PBN adducts
detected in our system (N 1, 2) differ from those with
reactive oxygen species (N 5, 6). The PBN (N 1) and POBN
(N 17) adducts obtained in aerobic reactions with 2-oxo
acids and CoA are similar to those known for thiyl radicals
(N 13–15, 18,19). Precipitation of the protein after the
aerobic reaction did not diminish the EPR signal which
arose from the supernatant, indicating nonprotein PBN

adducts. Probably, the thiyl radical of CoA was trapped
under these conditions. An indirect relationship between the
superoxide and PBN adducts is further supported by the
fact that the hyperfine splitting constants of the aerobic
PBN adducts formed in the forward (Table 1, N 1) and
backward (Table 1, N 2) reactions differed. Hence, secon-
dary reactions with the initially produced superoxide or its
PBN adduct must be invoked to explain formation of the
stable PBN adducts characterized by the EPR spectra
showninFig.1.
Direct detection of the superoxide anion radical
produced by the complexes
Production of superoxide anion radical by the complexes
was also examined by methods other than spin trapping.
Increased luminescence of lucigenin (bis-N-methylacridi-
nium) upon its reaction with superoxide is used for specific
detection of the latter in a number of biological systems [56].
Up to a 10-fold increase in the luminescence occured upon
incubation of the 2-oxoglutarate dehydrogenase complex
with 2-oxoglutarate and CoA in the presence of lucigenin.
However, lucigenin itself may increase formation of super-
oxide in the presence of enzymes that are capable of
catalyzing 1e
–
reduction of lucigenin directly [57], and E3
catalyzes 1e
–
reduction of various compounds [58–60].
Therefore to quantify production of superoxide radical by
Fig. 1. EPR spectra of PBN adducts recorded under equal settings after

15 min incubation of 2-oxoglutarate dehydrogenase complex
(4 mg mL
-1
) with its substrates (2 m
M
each): (1) 2-oxoglutarate, CoA,
NAD
+
(2) 2-oxoglutarate, CoA.
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5007
Table 1. Comparison of the spin trap adducts obtained in the reactions catalyzed by the 2-oxoglutarate or pyruvate dehydrogenase complexes (E) with the known spin trap adducts in water medium. Adducts with no
decay during the time of experiment are referred to as stable.
Spin
trap
Experimental system
involving the 2-oxo acid
dehydrogenase complexes
Radical
group
Previously
identified
radical adducts
a
N
(mT) a
H
(mT)
Stability
of adducts
Reference for

identified radical
adducts
N
PBN E, 2-oxoglutarate
or pyruvate, CoA, O
2
Matched to thiyl radicals 1.56 ± 0.02 0.33 ± 0.01 Stable 1
E, NADH, O
2
Matched to alkyl radicals 1.6 ± 0.004 0.33 ± 0.01 Stable 2
E, 2-oxoglutarate or
pyruvate, CoA, anaerobic
Matched to thiyl radicals 1.57 ± 0.01 0.33 ± 0.01 Transient species,
stable at low substrate
3
E, NADH, anaerobic Matched to thiyl radicals 1.59 ± 0.01 0.33 ± 0.005 Transient species,
stable at low substrate
4
Reactive oxygen species
•
OH
•
OOH
1.53–1.56
1.48
0.27–0.295
0.275–0.295
s to min
Unstable (min)
77–79

78, 79
5
6
Alkoxy (RCO
•
) MethylO
•
MethylO
•
t-ButylO
•
of linoleic acid
1.51
1.51
1.53
1.56
0.35
0.35
0.33
0.35
Stable
Stable
Stable
Stable
80
80
82
82
7
8

9
10
Alkyl (RC
•
) Methyl
•
1.58–1.65 0.36–0.37 Stable 74, 80, 81 11
of linoleic acid 1.62–1.63 0.26–0.3 Stable 82, 83 12
Thiyl (RS
•
) of DTT
of glutathione
of cysteine
of protein Cys in 5
M
urea
1.56
1.56
1.56–1.57
1.6
0.32
0.32
0.34–0.35
0.36
min
min
min under oxidizing
conditions; hours without O
2
Stable

This work
This work
66, 84
67
13
14
15
16
POBN E, pyruvate, CoA, O
2
Matched to thiyl radicals 1.52 0.24 Stable 17
Thiyl (RS
•
) of cysteine
of glutathione
1.50–1.52
1.51
0.26–0.27
0.23
Stable
Stable
84
85
18
19
Alkyl (RC
•
) Methyl
•
1.60 0.28 Stable 86 20

DMPO E, 2-oxoglutarate,
CoA, anaerobic
Matched to carbon-
centered radical
1.58 2.24 Stable 21
Thiyl (RS
•
) of cysteine
of glutathione
of N-acetylcysteine
of CoA
of lipoate
1.52–1.53
1.50–1.54
1.50
1.54
No adduct
1.70–1.73
1.62–1.63
1.68
1.62
min
transition metals with
SH excess catalyze decay
min
min
87
50, 87, 88
87
14

50, 68
22
23
24
25
26
Carbon-centered of ethanol
of formate
1.58
1.59
2.28–2.29
1.93
min
not indicated
14, 89, 90
91
27
28
5008 V. I. Bunik and C. Sievers (Eur. J. Biochem. 269) Ó FEBS 2002
the 2-oxoglutarate dehydrogenase complex, ferricyto-
chrome c reduction [61,62] was used. Both in the forward
(with 2-oxoglutarate plus CoA, 2 m
M
each) and backward
(with NADH, 8 m
M
) partial reactions, the specific activities
of the complex in superoxide production measured as initial
rates of SOD-inhibited ferricytochrome c reduction were
about 1 nmolÆmin

)1
Æmg
)1
. This corresponds to 0.3–0.4% of
the overall NAD
+
-reductase activity of the complex
(reactions 1–5).
Radical species and the catalysis-induced inactivation
of E1
Generation of ROS was documented in the current study
under conditions shown to induce catalysis-dependent
inactivation of the complexes [48,49]. Therefore we exam-
ined the possibility that the inactivation (Fig. 2A) was due
to the superoxide anion radical generated by the system. No
protection from the aerobic inactivation was observed in the
presence of SOD. Besides, the complexes were inactivated
by 2-oxo acid and CoA also under anaerobic conditions
(Fig. 2B). Thus, the enzyme inactivation was not caused by
the ROS produced. On the other hand, radical species were
deteced in the absence of oxygen too (Fig. 3). The spectrum
obtained under anaerobic conditions created with glucose
oxidase, glucose and catalase (Fig. 3, spectrum 2) showed a
weaker doublet at higher field, indicative of adduct decay
during the field sweep. As glucose is a known radical
scavenger [63] and in our experiments it indeed decreased
the signal of the PBN adducts already formed, a more
detailed study was performed under anaerobic conditions
created in a glove box.
Several properties of the anaerobic and aerobic PBN

adducts differed. First, unlike the aerobic spectra, the
anaerobic ones exhibited no significant difference in hyper-
fine splitting constants for the forward (Table 1, N 3) and
backward (Table 1, N 4) reactions. This argued for the same
radical species being trapped, in good agreement with the
limited possibilities of secondary reactions in the absence of
O
2
. Second, in contrast to the aerobic process, formation of
anaerobic adducts was not prevented by SOD and DTPA
(Fig. 3, spectrum 1). Third, the kinetics of the anaerobic and
aerobic reactions were different. Significantly higher inten-
sity of the EPR signal was observed in the absence of O
2
just
after the start of the reaction. However, these species rapidly
disappeared when the substrates were in excess. Under the
same conditions, the aerobic species accumulated with time.
Both the initial accumulation and following disappearance
of the anaerobic species were more pronounced at increased
enzyme concentration. If added substrates were limiting so
that no continuous reduction of the complex redox centers
occurred, the anaerobic PBN adducts were rather stable.
Fourth, the anaerobic reaction led to the PBN adduct being
localized to the protein fraction, whereas the aerobic adduct
under identical conditions was found in the supernatant
(Fig. 4). Different localizations of the EPR signal after the
anaerobic and aerobic reactions with NADH indicated that
a transient protein-bound radical intermediate, not detect-
able in the presence of oxygen, was trapped upon anaerobic

reduction of the complex.
According to the backward catalytic process effected by
NADH (reactions 5 and 4), the protein-bound radical
species (Fig. 4B) could arise from either the E3 redox-active
disulfide and FAD or E2-bound lipoate residues. From
those, only the latter may show no nitroxide immobiliza-
tion, as the lipoyl-lysine side chains are mobile and protrude
from the complex core [1–4]. In particular, their essentially
free rotation was observed upon spin labeling of the lipoyl
Fig. 2. Inactivation of 2-oxoglutarate dehydrogenase complex in the
presence of 2-oxoglutarate and CoA under aerobic (A) and anaerobic (B)
conditions. Concentration of substrates: 0.15 m
M
(1), 1.5 m
M
(2).
Concentration of the complex used (9 mgÆmL
)1
 4.5 l
M
) corres-
ponds to  0.3 m
M
sites for substrate and/or reducing equivalents
(24E1 + 24E2 + 12E3-S
2
+12E3-FAD).
Fig. 3. Spectra of PBN adducts obtained upon anaerobic incubation
of 2-oxoglutarate dehydrogenase complex with 2-oxoglutarate and CoA.
1: Anaerobiosis created in glove box, reaction took place for 18 min in

the presence of 2-oxoglutarate dehydrogenase complex (3 mgÆmL
)1
),
2-oxoglutarate and CoA (2 m
M
each), SOD (0.1 l
M
)andDTPA
(0.1 m
M
); 2: after solutions were preincubated for several minutes with
glucose oxidase, glucose and catalase, the reaction was started by
mixing the 2-oxoglutarate dehydrogenase complex (9 mgÆmL
)1
)with
the substrates (3 m
M
) and PBN and the spectrum was recorded
immediately.
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5009
group with a maleimide carrying a nitroxide label [64,65]. In
contrast, the adducts with the E3 redox centers should show
restricted rotation inherent in the nitroxides being localized
to the protein interior. Indeed, EPR spectra of the PBN
adducts with the protein cysteine residues [66,67] as well as
the spectrum obtained in our model reaction with PBN
trapping intermediates of FAD reduction by dithionite, are
qualitatively different from the spectra of the protein-bound
adduct in Fig. 4B. Thus, catalysis-dependent kinetics of the
anaerobic adduct, its protein localization, spectrum, and

hyperfine splitting constants (Table 1, N 3 and 4) charac-
teristic of the PBN-trapped thiyl radicals (Table 1, N
13,14,15) allow us to conclude that under anaerobic
conditions PBN traps radicals of the complex-bound
dihydrolipoate residues.
Detection of the E2-bound dihydrolipoate thiyl radical
correlated with the inactivation of the complexes by 2-oxo
acid plus CoA (Fig. 2B). In the absence of O
2
, both the
inactivation (Fig. 2B) and the EPR signal stability
decreased with increasing concentration of substrates, i.e.
upon full reduction of the catalytic centers. Dismutation of
the thiyl radicals and reduction of their PBN adducts within
the network of interacting lipoyl moieties of the E2 core
provides a good explanation for these phenomena.
Considering possible mechanisms of the inactivation of
the complex by the thiyl radical of dihydrolipoate, we took
into account that (a) the overall activity (Fig. 2) is decreased
due to the irreversible inactivation of E1 [49], and (b) the
thiyl radical of the lipoyl residue should efficiently interact
with the E1 catalytic intermediate, as the lipoyl-bearing
domain of E2 is designed for this interaction (reaction 2). In
this case, 1e
–
oxidation of the carbanion in the E1 active site
(a product of reaction 1) by highly electrophilic thiyl radical
of the dihydrolipoyl residue of E2 had to be expected. The
reaction was confirmed by anaerobic spin trapping with
DMPO. This spin trap is known to be unreactive towards

the lipoate radical species [50,68]. Addition of DMPO to the
anaerobic reaction mixture containing the 2-oxoglutarate
dehydrogenase complex, 2-oxoglutarate and CoA resulted
in the spectrum (Fig. 5) characteristic of the carbon-
centered DMPO radical, known for the DMPO adducts
with hydroxyalkyl (e.g. ethanol) or formate radicals
(Table 1, N 27 and 28). Because CO
2
and the 1-hydroxy-
3-carboxypropyl moiety bound to ThDP are formed during
the E1-catalyzed decarboxylation of 2-oxoglutarate (reac-
tion 1), the spectrum and hyperfine splitting constants of the
DMPO adduct obtained (Table 1, N 21) are consistent with
the product of 1e
–
oxidation of the 2-oxoglutarate-ThDP-
E1 complex.
Under aerobic conditions, the thiyl radical of dihydro-
lipoate cannot be detected as its PBN adduct due to the
concomitant presence of the superoxide anion radical,
secondary reactions and poor spectral resolution of
different PBN adducts. However, several lines of evidence
point to the presence of the thiyl radical in the aerobic
system. First, the complexes are inactivated by 2-oxo acid
and CoA in the presence of O
2
(Fig. 2A) and the
inactivation is not prevented by SOD. In contrast,
thioredoxin which is a known thiyl radical scavenger [69]
fully protected the enzyme from the inactivation. The

thioredoxin protection was obvious not only when assay-
ing the overall activity (reactions 1–5), but also upon
generation of the paramagnetic species. In the medium
without thioredoxin, the EPR signal intensity reached
saturation after the complex inactivation (10–15 min of
incubation with the substrates, Fig. 2A). When thioredoxin
was added, the initial rate of the EPR signal increase was
the same, but no saturation was obvious during half-an-
hour. Thus, as increased productivity of the complexes in
Fig. 5. DMPO spin trapping of anaerobic reaction medium containing
2-oxoglutarate dehydrogenase complex (9 mgÆmL
-1
), 2-oxoglutarate
and CoA (4 m
M
each).
Fig. 4. Localization of PBN adducts obtained upon incubation of E.coli
pyruvate dehydrogenase complex (27 mgÆmL
-1
)withNADH(0.4m
M
)
under aerobic (A) and anaerobic (B) conditions. 1: Before protein pre-
cipitation. 2: Protein fraction. 3: Non-protein fraction. Fractionation is
described in Materials and methods.
5010 V. I. Bunik and C. Sievers (Eur. J. Biochem. 269) Ó FEBS 2002
generation of radical species was observed in the presence
of thioredoxin.
Another argument for the presence of an oxidizing sulfur-
centered radical species under aerobic conditions is provided

by the accompanying spectral changes of the complex.
Unchanged in the presence of 2-oxoglutarate, the spectrum
exhibited a rapid decrease in absorbance at 450 and 350 nm
after addition of CoA. These changes are known to proceed
upon reduction of E3 with dihydrolipoate. However, a
concomitant increase of comparable magnitude at 400 nm
was also observed. This change does not occur upon
reduction of the isolated E3 or E3 bound to E2 lacking the
lipoyl domain [70]. On the other hand, it involves the
spectral region characteristic of the three-electron bonds
formed with sulfur participation [71]. Similar spectral
change at 400 nm, stable in time and insensitive to oxygen,
was observed upon reaction of E3 with a strongly oxidizing
radical Br
2
• –
(but not with O
2
• –
), which was attributed to
formation of a disulfide radical anion followed by electron
transfer to some other residue [20].
Finally, a series of aerobic inactivation experiments
support the proposed mechanism of the 2-oxo acid plus
CoA-dependent inactivation of E1, involving the complex-
bound thiyl radical. As seen from Table 2, the enzyme
activity was not decreased in the presence of dihydrolipoate,
NADH or succinyl-CoA. This indicates that neither these
products of the overall reaction, nor radical species formed
in the incubation medium with dihydrolipoate or NADH

inactivate E1. However, any combination of the substrates
and/or products providing concomitant presence of the E1
catalytic intermediate and complex-bound dihydrolipoate
(2-oxo acid + CoA; NADH + 2-oxoacid; dihydrolipo-
amide + succinyl-CoA) did lead to inactivation. In partic-
ular, the appearance of the ThDP adduct with the substrate
moiety during the complex-catalyzed succinyl-CoA hydro-
lysis [72] accounts for the inactivation by dihydrolipoamide
in the presence of succinyl-CoA (Table 2).
DISCUSSION
Generation of radical species during catalysis by 2-oxo acid
dehydrogenase multienzyme complexes has been documen-
ted in this work by EPR spectroscopy, ferricytochrome c
reduction and lucigenin fluorescence. The superoxide anion
radical is produced upon the E3-catalyzed 1e
–
oxidation of
the E2-bound dihydrolipoate intermediate. The thiyl radical
of the E2-bound dihydrolipoate formed in this reaction
causes the 1e
–
oxidation of the 2-oxo acid adduct with
ThDP through the site-directed action on E1. This results in
the carbon-centered radical fragment in the E1 active site
and the enzyme inactivation. The inactivation is prevented
by thioredoxin which is a known scavenger of thiyl radicals
[69].
The present work shows that production of radical
species by the 2-oxo acid dehydrogenase complexes under-
lies several phenomena of regulatory significance: (a)

sensitivity of the first component of the complex to the
terminal step of the overall reaction (NADH or superoxide
production), (b) thioredoxin-dependent modulation of this
sensitivity, and (c) the 2-oxoacid, CoA-dependent genera-
tion of a cellular messenger, superoxide anion radical, which
is increased in the presence of thioredoxin. The isolated E3
component was reported to produce superoxide in the
nonphysiological backward reaction of NADH oxidation
[22]. Under physiological conditions, this reaction is unlikely
to contribute to the mitochondrial metabolism significantly:
NADH is a strong inhibitor of E3 and should thus be much
more efficiently oxidized by the enzymes of the respiratory
chain, specifically designed for the NADH oxidation.
However, the current study shows that superoxide is cata-
lytically produced by the complex-bound E3 in the physio-
logical direction of 2-oxo acid oxidation and may take place
concomitantly with the overall reaction (Fig. 1, spectrum 1).
In the presence of 2-oxo acid and CoA the complexes gen-
erate superoxide anion radical at a rate (1 nmolÆmin
)1
Æmg
)1
)
comparable to that of the known superoxide producers:
respiratory chain (0.3–6 nmolÆmin
)1
Æmg
)1
) [73], microsomes
(0.7–4 nmolÆmin

)1
Æmg
)1
), or purified FAD-containing
monooxygenase (3–6 nmolÆmin
)1
Æmg
)1
) [62]. Inability of
the FAD-modified complex to generate paramagnetic
species rules out the direct oxidation of the accumulated
complex-bound dihydrolipoate intermediate as a source of
the superoxide and indicates that the integral complex
structure is required for the radical species production.
The results of the current study and the site-specific
reactivity of O
2
• –
also bear on consideration of the concept
of metabolons, i.e. intracellular compartmentalized func-
tional units. The regulatory potential of superoxide radical
production by the 2-oxoglutarate dehydrogenase complex
should greatly increase in the microenvironment of the citric
acid cycle metabolon. This implies close arrangement of the
2-oxoglutarate dehydrogenase complex and transition
metal-dependent enzymes, such as aconitase and fumarase,
both rapidly reacting with O
2
• –
. Aconitase has been shown

to produce OH
•
radicals upon interaction with the super-
oxide anion radical, which releases its iron(II) [74]. Selective
and simultaneous targeting of aconitase and 2-oxo acid
dehydrogenase complexes under oxidative stress in mito-
chondria [34] favors the interpretation that the former
interacts with the superoxide generated by the latter. Apart
from the induction of the removal of the transition metal
from aconitase via superoxide production, the direct
mobilization of transition metals by dihydrolipoate may
add to its pro-oxidant action in vivo, as the ability of
dihydrolipoate to mobilize ferritin-bound iron, possibly
through a radical species, is known [44,51,75].
Anaerobic reduction of the complexes in the presence of
PBN revealed transient formation of dihydrolipoate thiyl
radicals upon equilibration of the E2 and E3 redox centers.
Because of superoxide production, the aerobic system is too
complicated to allow detection of thiyl radical adduction by
Table 2. Inactivation of 2-oxoglutarate dehydrogenase from pig heart
upon incubation with its substrates and/or products (Reaction conditions
are indicated in Materials and methods).
Substrate(s) and/or product(s) added k
i
(min
)1
)
2-oxoglutarate 0.011 ± 0.005
Dihydrolipoamide 0.009 ± 0.005
CoA No inactivation

Succinyl-CoA No inactivation
NADH No inactivation
2-oxoglutarate, CoA 0.10 ± 0.01
Succinyl-CoA, dihydrolipoamide 0.03 ± 0.01
NADH + 2-oxoglutarate 0.04 ± 0.01
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5011
PBN. However, the 1e
–
oxidation of the reduced complex
by oxygen implicates the redox equilibrium involving the
E3- and E2-bound thiyl radicals and the E3-bound flavin
semiquinone. Under these circumstances, appearance of a
strongly oxidizing thiyl radical of dihydrolipoate (E° of a
number of RS
•
/RS
–
or RS
•
/RSH couples are approximately
+0.75 or +1.33 V, respectively [71]) is supported by the
enzyme spectral changes, the SOD- and O
2
-insensitive
inactivation of the E1 catalytic intermediate (Fig. 2,
Table 2) and the protection by thioredoxin from such an
inactivation. Pro-oxidant action of thiyl radicals is avoided
in the presence of thioredoxin, because the free radical
species of thioredoxin, both disulfide and thiyl, are unreac-
tive [69]. In our system, migration of one electron between

the dihydrolipoate thiyl radical and thioredoxin should
preclude the radical-mediated modification of E1 and
facilitate the dihydrolipoate radical dismutation. A thiore-
doxin mutation which renders the protein sulfur radical
more oxidizing (D30A, numbering of Chlamidomonas
reinhardtii thioredoxin h) [69], leads to a decrease of the
2-oxoglutarate dehydrogenase complex activity, rather than
the protection exhibited by the wild type thioredoxin [48].
Such a modulation of the 1e
–
redox properties of thiore-
doxin by amino-acid substitution may explain the adverse
action of some thioredoxins in the 2-oxo acid dehydroge-
nase system, as well as the specific and highly efficient
protection by mitochondrial thioredoxin [48].
As pointed out by Asmus [71], interaction of oxygen
with thiyl radicals is considerably less efficient than
previously thought. This agrees with our observations that
O
2
does not prevent the E1 damage (Fig. 2) and that the
rates of decay of E1 activity and of the overall reaction are
equal [49]. Irreversible modification of the lipoate residues
by RSOO
•
formation with the deleterious action of the
latter on E1 should have caused a faster inactivation of the
overall reaction compared to the partial E1 decay, because
in this case both E1 and E2 were inactivated. Thus, our
data indicate that (a) oxygen addition to the complex-

bound dihydrolipoate radicals is less efficient than their
interaction with E1, and (b) the dihydrolipoate radical
intermediate survives long enough to be of regulatory
significance. By enabling E1 inactivation in response to the
absence of E3 substrate, the dihydrolipoate radical inter-
mediate transfers information from E3 to E1. As a result,
superoxide production by the complexes is restricted unless
thioredoxin is added. Thioredoxin modulates the self-
regulation of the complexes through abolition of the
deleterious action of the dihydrolipoate thiyl radicals on
E1. This provides an increased performance of the
complexes not only in the overall reaction, but also in
superoxide production.
Thus, the energy-providing oxidative decarboxylation of
2-oxoacids may influence mitochondrial metabolism also by
means of the pro-oxidant action of the dihydrolipoate
intermediate propagated through E3 (superoxide and thiyl
radicals production) and E1 (catalytic intermediate radical
production and inactivation). While formation of the
intrinsic thiyl radical is deleterious for the 2-oxo acid
oxidation, it also is an antioxidant defense mechanism,
preventing the superoxide production by the complexes.
External regulation of these processes by a cellular thiol-
disulfide oxidoreductase, thioredoxin, points to the link
between the complexes and thioredoxin-dependent
pathways in mitochondria. For instance, they may form
an antioxidant defense system, analogous to recently
discovered in mycobacteria where the 2-oxoglutarate dehy-
drogenase complex provides reducing equivalents to the
peroxiredoxin alkyl hydroperoxide reductase through a

thioredoxin-like protein [76]. Multiple levels of regulation
and sensitivity to integral parameters such as substrate
concentrations and NADH/NAD
+
ratio support biological
significance of the complex-catalyzed radical reactions
characterized in the present work.
ACKNOWLEDGEMENTS
This work was partially supported by the Alexander von Humboldt
Foundation. The authors thank Prof. U. Weser and Dr H. Hartman
for their advices concerning the EPR measurements at the beginning of
this work. Critical reading of the manuscript by Prof. J. Mieyal and
Prof. A. J. L. Cooper and the interest of Prof. G. Gibson to this
investigation are greatly acknowledged.
REFERENCES
1. Aevarsson, A., Seger, K., Turley, S., Sokatch, J.R. & Hol, W.G.J.
(1999) Crystal structure of 2-oxoisovalerate dehydrogenase and
the architecture of 2-oxo acid dehydrogenase multienzyme com-
plexes. Nat. Struct. Biol. 6, 785–792.
2. DeKok, A., Hengeveld, A.F., Martin, A. & Westphal, A.H. (1998)
The pyruvate dehydrogenase complex from gram-negative bac-
teria. Biochim. Biophys. Acta 1385, 353–366.
3. Perham, R.N. (1991) Domains, motifs, and linkers in 2-oxo acid
dehydrogenase multienzyme complexes: a paradigm in the design
of a multifunctional protein. Biochemistry 30, 8501–8512.
4. Perham, R.N. (2000) Swinging arms and swinging domains in
multifunctional enzymes: catalytic machines for multistep reac-
tions. Ann. Rev. Biochem. 69, 961–1004.
5. Danson, M.J., Fersht, A.R. & Perham, R.N. (1978) Rapid
intramolecular coupling of active sites in the pyruvate dehy-

drogenase complex of Escherichia coli: mechanism for rate
enhancement in a multimeric structure. Proc.Natl.Acad.Sci.USA
75, 5386–5390.
6. Akiyama, S.K. & Hammes, G.G. (1980) Elementary steps in the
reaction mechanism of the pyruvate dehydrogenase multienzyme
complex from Escherichia coli: kinetics of acetylation and deace-
tylation. Biochemistry 19, 4208–4213.
7. Waskiewicz, D.E. & Hammes, G.G. (1980) Elementary steps in
the reaction mechanism of the a-ketoglutarate dehydrogenase
multienzyme complex from Escherichia coli: kinetics of succiny-
lation and desuccinylation. Biochemistry 23, 3136–3143.
8. Reed, L.J. & Hackert, M.L. (1990) Structure-function relation-
ships in dihydrolipoamide acyltransferase. J. Biol. Chem. 265,
8971–8974.
9. Massey, V., Gibson, Q.H. & Veeger, C. (1960) Intermediates in the
catalytic action of lipoyl dehydrogenase (diaphorase) Biochem. J.
77, 341–351.
10. Massey, V. (1963) Lipoyl dehydrogenase. Enzymes 7, 275–306.
11. Searls, R.L., Peters, J.M. & Sanadi, D.R. (1961) a-Ketoglutaric
dehydrogenase. X. On the mechanism of dihydrolipoyl dehy-
drogenase reaction. J. Biol. Chem. 236, 2317–2322.
12. Matthews, R.G., Ballou, D.P., Thorpe, C. & Williams, C.H. Jr
(1977) Ion pair formation in pig heart lipoamide dehydrogenase.
J. Biol. Chem. 252, 3199–3207.
13. Templeton, D.M., Hollebone, B.R. & Tsai, C.S. (1980) Magnetic
circular dichroism studies on the active-site flavin of lipoamide
dehydrogenase. Biochemistry 19, 3969–3873.
14. Docampo, R., Moreno, S.J. & Mason, R.P. (1987) Free
radical intermediates in the reaction of pyruvate: ferredoxin
5012 V. I. Bunik and C. Sievers (Eur. J. Biochem. 269) Ó FEBS 2002

oxidoreductase in Tritrichomonas foetus hydrogenosomes. J. Biol.
Chem. 262, 12417–12420.
15. Menon,S.&Ragsdale,S.W.(1997)MechanismoftheClostridium
thermoaceticum pyruvate: ferredoxin oxidoreductase. Evidence for
the common catalytic intermediacy of the hydrohyethylthiamine
diphosphate radical. Biochemistry 36, 8484–8494.
16. Chabriere, E., Vernede, X., Guigliarelli, B., Charon, M H.,
Hatchikian, E.C. & Fontecilla-Camps, J.C. (2001) Crystal struc-
ture of the free radical intermediate of pyruvate: ferredoxin oxi-
doreductase. Science 294, 2559–2563.
17. Barletta,G.,Chung,A.C.,Rios,C.B.,Jordan,F.&Schlegel,J.M.
(1990) Electrochemical oxidation of enamines related to the key
intermediate on thiamin diphosphate dependent enzymatic path-
ways: evidence for one-electron oxidation via a thiazolium cation
radical. J. Am. Chem. Soc 112, 8144–8149.
18. Nakanishi, I., Itoh, S., Suenobu, T. & Fukuzumi, S. (1997) Elec-
tron transfer properties of active aldehydes derived from thiamin
coenzyme analogues. Chem. Commun. 1927–1928.
19. Chan, S.W., Chan, P.C. & Bielski, B.H.J. (1974) Studies on the
lipoic acid free radical. Biochim. Biophys. Acta 338, 213–223.
20. Elliot, A.J., Munk, P.L., Stevenson, K.J. & Armstrong, D.A.
(1980) Reactions of oxidising and reducing radical probes with
lipoamide dehydrogenase. Biochemistry 19, 4945–4950.
21. Massey, V., Mueller, F., Feldberg, R., Schuman, M., Sullivan,
P.A., Howell, L.G., Mayhew, S.G., Matthews, R.G. & Foust,
G.P. (1969) The Reactivity of flavoproteins with sulfite. Possible
relevance to the problem of oxygen reactivity. J. Biol. Chem. 244,
3999–4006.
22. Massey, V., Strickland, S., Mayhew, S.G., Howell, L.G., Engel,
P.C., Matthews, R.G., Schuman, M. & Sullivan, P.A. (1969) The

production of superoxide anion radicals in the reaction of reduced
flavins and flavoproteins with molecular oxygen. Biochem.
Biophys. Res. Communs. 36, 891–898.
23. Ballou, D., Palmer, G. & Massey, V. (1969) Direct demonstration
of superoxide anion production during the oxidation of reduced
flavin and of its catalytic decomposition by erythrocuprein.
Biochem. Biophys. Res. Communs. 36, 898–904.
24. Kunz, W.S. & Kunz, W. (1985) Contribution of different enzymes
to flavoprotein fluorescence of isolated rat liver mitochondria.
Biochim. Biophys. Acta 841, 237–246.
25. Khan, A.U. & Wilson, T. (1995) Reactive oxygen species as cel-
lular messengers. Chem. Biol. 2, 437–445.
26. Skulachev, V.P. (2000) Mithochondria in the programmed death
phenomena; a principle of biology. ÔIt is better to die than to be
wrongÕ. JUBMB Life 49, 365–373.
27. Sastre, J., Pallardo, V. & Vina, J. (2000) Mitochondrial oxidative
stress plays a key role in aging and apoptosis. JUBMB Life. 49,
427–435.
28. Cabiscol, E., Piulats, E., Echave, P., Herrero, E. & Ros, J. (2000)
Oxidative stress promotes specific protein damage in Sacharomices
cerevisiae. J. Biol. Chem. 275, 27393–27398.
29. Humphries, K.M. & Szweda, L.I. (1998) Selective inactivation of
a-ketoglutarate dehydrogenase and pyruvate dehydrogenase:
reaction of lipoic acid with 2-hydroxy-2-nonenal. Biochemistry 37,
15835–15841.
30. Rokutan, K., Kawai, K. & Asada, K. (1987) Inactivation of
2-oxoglutarate dehydrogenase in rat liver mitochondrial by its
substrate and T-butyl hydroperoxide. J. Biochem. 101, 415–422.
31. Millar, H.A. & Leaver, C.J. (2000) The cytotoxic lipid peroxida-
tion product, 4-hydroxy-2-nonenal, specifically inhibits dec-

arboxylating dehydrogenases of the matrix of plant mitochondria.
FEBS Lett. 481, 117–121.
32. Tabatabaie, T., Potts, J.D. & Floyd, R.A. (1996) Reactive oxygen
species-mediated inactivation of pyruvate dehydrogenase. Arch.
Biochem. Biophys. 336, 290–296.
33. Andersson, U., Leighton, B., Young, M.E., Blomstrand, E. &
Newsholme, E.A. (1998) Inactivation of aconitase and
2-oxoglutarate dehydrogenase in skeletal muscle in vitro by
superoxide anions and/or nitric oxide. Biochem. Biophys. Res.
Commun. 249, 512–516.
34. Bruschi, S.A., Lindsay, J.G. & Crabb, J.W. (1998) Mitochondrial
stress protein recognition of inactivated dehydrogenases during
mammalian cell death. Proc. Natl Acad. Sci. USA 95, 13413–
13418.
35. Park, L.C.H., Gibson, G.E., Bunik, V. & Cooper, A.J.L. (1999)
Inhibition of select mitochondrial enzymes in PC12 cells exposed
to S-(1,1,2,2-tetrafluoroethyl)-
L
-cysteine. Biochem. Pharmacol. 58,
1557–1565.
36. Gibson, G.E., Sheu, K F.R., Blass, J.P., Baker, A., Carlson,
K.C., Harding, B. & Perrino, P. (1988) Reduced activities of
thiamine-dependent enzymes in brains and peripheral tissues of
patients with Alzheimer’s desease. Arch. Neurol. 45, 836–840.
37. Mastrogiacomo, F., Lindsay, J.G., Bettendorff, L., Rice, J. &
Kish, S.J. (1996) Brain protein and a-ketoglutarate dehydrogenase
complex activity in Alzheimer’s desease. Ann. Neurol. 39, 592–599.
38. Mizuno, Y., Matsuda, S., Yoshino, H., Mori, H., Hattori, N. &
Ikebe, S.I. (1994) An immunohistochemical study on a-keto-
glutarate dehydrogenase complex in Parkinson’s desease. Ann.

Neurol. 35, 204–210.
39. Butterworth, R.F., Kril, J.J. & Harper, C.G. (1993) Thiamine-
dependent enzyme changes in the brains of alcoholics: Relation-
ship to the Wernike–Korsakoff syndrome alcohol. Clin.Exp.Res.
71, 1084–1088.
40. Hackert, M.L., Oliver, R.M. & Reed, L.J. (1983) A computer
model analysis of the active-site coupling mechanism in the pyru-
vate dehydrogenase complex of Escherichia coli. Proc. Natl Acad.
Sci. USA 80, 2907–2911.
41. Collins, J.H. & Reed, L.J. (1977) Acyl Group and electron pair
relay system: a network of interacting lipoyl moieties in the
pyruvate and a-ketoglutarate dehydrogenase complexes from
Escherichia coli. Proc. Natl. Acad. Sci. USA 74, 4223–4227.
42.Guest,J.R.,Ali,S.T.,Artymiuk,P.,Ford,G.C.,Green,J.&
Russel, G.C. (1990) Site-directed mutagenesis of dihy-
drolipoamide acetyltransferase and post-translational modifica-
tion of its lipoyl domains. In Biochemistry and Physiology of
Thiamin Diphosphate Enzymes (Bisswanger, H. & Ullrich, J., eds),
pp. 176–193. Chemie, Weinheim.
43. Scott, B.C., Aruoma, O.I., Evans, P.J., O’Neill, C., Van der Vliet,
A., Cross, C.E., Tritschler, H. & Halliwell, B. (1994) Lipoic and
dihydrolipoic acids as antioxidants. A critical evaluation. Free
Rad. Res. 20, 119–133.
44. Packer, L., Witt, E.H. & Tritschler, H.J. (1995) Alpha-lipoic acid
as a biological antioxidant. Free Rad Biol. Med. 19, 227–250.
45. Biewenga, G.Ph, Haenen, G.R.M.M. & Bast, A. (1997) The
pharmacology of the antioxidant lipoic acid Gen. Pharmacol. 29,
315–331.
46. Bunik, V. & Follmann, F. (1993) Thioredoxin reduction depen-
dent on a-keto acid oxidation by a-keto acid dehydrogenase

complexes. FEBS Lett. 336, 197–200.
47. Bunik, V., Shoubnikova, A., Loeffelhardt, S., Bisswanger, H.,
Borbe, H.O. & Follmann, H. (1995) Using lipoate enantiomers
and thioredoxin to study the mechanism of the 2-oxo acid-
dependent dihydrolipoate production by the 2-oxo acid dehy-
drogenase complexes. FEBS Lett. 371, 167–170.
48. Bunik,V.,Raddatz,G.,Lemaire,S.,Meyer,Y.,Jacquot,J P.&
Bisswanger, H. (1999) Interaction of thioredoxins with target
proteins: Role of particular structural elements and electrostatic
properties of thioredoxins in their interplay with 2-oxo acid
dehydrogenase complexes. Protein Sci. 8, 65–74.
49. Bunik, V. (2000) Increased catalytic performance of the 2-oxo acid
dehydrogenase complexes in the presence of thioredoxin, a thiol-
disulfide oxidoreductase. J. Molec. Catalysis B 8, 165–174.
50. Stoyanovsky, D.A., Goldman, R., Claycamp, H.G. & Kagan,
V.E. (1995) Phenoxyl radical-induced thiol-dependent generation
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5013
of reactive oxygen species: implications for benzene toxicity. Arch.
Biochem. Biophys. 317, 315–323.
51. Anusevicius, Z.J. & Cenas, N.K. (1993) Dihydrolipoate-mediated
redox cycling of quinones. Arch. Biochem. Biophys. 302,420–
424.
52. Stanley, C.J. & Perham, R.N. (1980) Purification of 2-oxo acid
dehydrogenase multienzyme complexes from ox heart by a new
method. Biochem. J. 191, 147–154.
53. Bisswanger, H. (1981) Substrate specificity of the pyruvate dehy-
drogenase complex from Escherichia coli. J. Biol. Chem. 256,815–
822.
54. Ambrose-Griffin, M.C., Danson, M.J., Griffin, W.G., Hale, G. &
Perham, R. (1980) Kinetic analysis of the role of lipoic acid

residues in the pyruvate dehydrogenase multienzyme complex of
Escherichia coli. Biochem. J. 187, 393–401.
55. Kalyanaraman, B. & Mason, R.P. (1979) The reduction of
nitroso-spin traps in chemical and biological systems. A cau-
tionary note. Tetrahedron Lett. 50, 4809–4812.
56. Li, Y., Stansbury, K.H., Zhu, H., & Trush, M.A. (1999)
Biochemical characterization of lucigenin (bis-N-methylacridi-
nium) as a chemiluminescent probe for detecting intramito-
chondrial superoxide anion radical production. Biochem. Biophys.
Res. Commun. 262, 80–87.
57. Skatchkov, M.P., Sperling, D., Hink, U., Aenggard, E. &
Muenzel, T. (1998) Quantificantion of superoxide radical forma-
tion in intact vascular tissue using a Cypridina luciferin analog as
an alternative to lucigenin. Biochem. Biophys. Res. Commun. 248,
383–386.
58. Leskovac, V., Svircevic, J., Trivic, S., Popovic, M. & Radulovic,
M. (1989) Reduction of aryl-nitroso compounds by pyridine and
flavin coenzymes. Int. J.Biochem. 21, 825–834.
59. Vienozinskis, J., Butkus, A., Cenas, N. & Kulys, J. (1990) The
mechanism of the quinone reductase reaction of pig heart lipoa-
mide dehydrogenase. Biochem. J. 269, 101–105.
60. Sreider, C.M., Grinblat, L. & Stoppani, A.O.M. (1992) Reduction
of nitrofuran compounds by heart lipoamide dehydrogenase: role
of flavin and the reactive disulfide groups. Biochem. Internat. 28,
323–334.
61. McCord, J.M. & Fridovich, I. (1968) The reduction of cytochrome
c by milk xantine oxidase. J. Biol. Chem. 243, 5753–5760.
62.Rosen,G.M.,Finkelstein,E.&Rauckman,E.J.(1982)A
method for the detection of superoxide in biological systems. Arch.
Biochem. Biophys. 215, 367–378.

63. Halliwell, B. & Gutteridge, J.M.C. (1986) Oxygen free radicals and
iron in relation to biology and medicine: some problems and
concepts. Arch. Biochem. Biophys. 246, 501–514.
64. Ambrose, M.C. & Perham, R.N. (1976) Spin-label study of the
mobility of enzyme-bound lipoic acid in the pyruvate dehy-
drogenase multienzyme complex of Escherichia coli. Biochem. J.
155, 429–432.
65. Grande, H.J., Van Telgen, H.J. & Veeger, C. (1976) Symmetry
and asymmetry of the pyruvate dehydrogenase complexes from
Azotobacter vinelandii and Escherichia coli as reflected by fluor-
escence and spin-label studies. Eur. J. Biochem. 71, 509–518.
66. Graceffa, P. (1983) Spin labeling of protein sulfhydryl groups by
spin trapping a sulfur radical: application to bovine serum albu-
min and myosin. Arch. Biochem. Biophys. 225, 802–808.
67. Gatti, R.M., Radi, R. & Augusto, O. (1994) Peroxynitrite-medi-
ated oxidation of albumin to the protein-thiyl free radical. FEBS
Lett. 348, 287–290.
68. Romero, F.J., Ordonez, I., Arduini, A. & Cadenas, E. (1992) The
reactivity of thiols and disulfides with different redox states of
myoglobin. J. Biol. Chem. 267, 1680–1688.
69. Hanine, L.C., El, Conte, D., Jacquot, J P. & Houee-Levin, C.
(2000) Redox properties of protein disulfide bond in oxidized
thioredoxin and lysozyme: a pulse radiolysis study. Biochemistry
39, 9295–9301.
70. Westphal, A.H., Fabisz-Kijowska, A., Kester, H., Obels, P.P. &
DeKok, A. (1995) The interaction between lipoamide dehy-
drogenase and the peripheral-component-binding domain from
Azotobacter vineladii pyruvate dehydrogenase complex. Eur. J.
Biochem. 234, 861–870.
71. Asmus, K D. (1990) Sulfur-centered free radicals. Methods Enz-

ymol. 186, 168–180.
72. Frey, P.A., Flournoy, D.S., Gruys, K. & Yang, Y S. (1989)
Intermediates in reductive transacetylation catalyzed by pyruvate
dehydrogenase complex. Ann. NY Acad. Sci. 573, 21–35.
73. Konstantinov, A.A., Peskin, A.V., Popova, E.Yu, Khomutov,
G.B. & Ruuge, E.K. (1987) Superoxide generation by the
respiratory chain of tumor mitochondria. Biochim. Biophys. Acta
894, 1–10.
74. Vasquez-Vivar, J., Kalyanaraman, B. & Kennedy, M.C. (2000)
Mitochindrial aconitase is a source of hydroxyl radical. An elec-
tron spin resonance investigation. J. Biol. Chem. 275, 14064–
14069.
75. Bonomi, F., Cerioli, A. & Pagani, S. (1989) Molecular aspects of
the removal of ferritin-bound iron by
DL
-dihydrolipoate. Biochim.
Biophys. Acta V. 994, 180–186.
76.Bryk,R.,Lima,C.D.,Erdjument-Bromage,H.,Tempst,P.&
Nathan, C. (2002) Metabolic enzymes of mycobacteria linked to
antioxidant defense by a thioredoxn-like protein. Science 295,
1073–1077.
77. Kotake, Y. & Janzen, E.G. (1991) Decay and fate of the hydroxyl
radical adduct of a-phenyl-N-tert-butylnitrone in aqueous media.
J. Am. Chem. Soc. 113, 9503–9506.
78. Harbour, J.R., Chow, V. & Bolton, J.R. (1974) An electron spin
resonance study of the spin adducts of OH and HO
2
radicals with
nitrones in the ultraviolet photolysis of aqueous hydrogen per-
oxide solutions. Can. J. Chem. 52, 3549–3553.

79. Janzen, E.G., Nutter, D.E. Jr, Davis, E.R., Blackburn, B.J., Poyer,
J.L., & McCay, P.B. (1978) On spin trapping hydroxyl and
hydroperoxyl radicals. Can. J.Chem. 56, 2237–2242.
80. Britigan, B.E., Coffman, T.J. & Buettner, G.R. (1990) Spin trap-
ping evidence for the lack of significant hydroxyl radical produc-
tion during the respiration burst of human phagocytes using a spin
adduct resistant to superoxide-mediated destruction. J. Biol.
Chem. 265, 2650–2656.
81. Britigan, B.E., Pou, S., Rosen, G.M., Lilleg, D.M. &
Buettner, G.R. (1990) Hydroxyl radical is not a product of the
reaction of xanthine oxidase and xanthine. J. Biol. Chem. 265,
17533–17538.
82. Osipov, A.N., Savov, V.M., Yachyaev, A.V., Zubarev, V.E.,
Azizova, O.A., Kagan, V.E. & Vladimirov, Y.A. (1984) Spin
trapping study of radicals generated in the reaction of organic
peroxides with ferrous ion. Biophysics (Russ.). 29, 533–536.
83. Azizova, O.A., Osipov, A.N., Savov, V.M., Zubarev, V.E.,
Kagan, V.E. & Vladimirov, YuA. (1984) Spin trapping study of
linoleic acid radicals during the initiation of lipid peroxidation in
Fenton’s reagent. Biophysics (Russ.) 29, 766–769.
84. Graceffa, P. (1988) Spin trapping the cysteine thiyl radical
with phenyl-N-T-butylnitrone. Biochim. Biophys. Acta 954,227–
230.
85. Shi, X. & Dala, N.S. (1988) On the mechanism of the chromate
reduction by glutathione. ESR evidence for the glutathionyl
radical and an isolable Cr(V) intermediate. Biochem. Biophys. Res.
Commun. 156, 137–142.
86. Stoyanovsky, D.A., Melnikov, Z. & Cederbaum, A.I. (1999) ESR
and HPLC-EC analysis of the interaction of hydroxyl radical with
DMSO. Rapid reduction and quantification of POBN and PBN

nitroxides. Anal. Chem. 71, 715–721.
87. Ross, D., Albano, E., Nilsson, U. & Moldeus, P. (1984) Thiyl
radicals formation during peroxidase-catalyzed metabolism of
acetaminophen in the presence of thiols. Biochem. Biophys. Res.
Commun. 125, 109–115.
5014 V. I. Bunik and C. Sievers (Eur. J. Biochem. 269) Ó FEBS 2002
88. Mason, R.P. & Ramakrishna Rao, D.N. (1990) Thyil free radical
metabolites of thiol drugs, glutathione, and proteins. Methods
Enzymol. 186, 318–329.
89. Buettner, G.R. (1984) Thiyl free radical production with hem-
atoporphyrin derivative, cysteine and light: a spin-trapping study.
FEBS Lett. 177, 295–299.
90. Saez,G.,Thornalley,P.J.,Hill,H.A.O.,Hems,R.&Bannister,
J.V. (1982) The production of free radicals during the autoxidation
of cysteine and their effect on isolated rat hepatocytes. Biochim.
Biophys. Acta 719, 24–31.
91. Sankarapandi, S. & Zweier, J.L. (1999) Evidence against the
generation of free hydroxyl radicals from the interaction of cop-
per, zinc-superoxide dismutase and hydrogen peroxide. J. Biol.
Chem., 274, 34576–34583.
Ó FEBS 2002 Radicals upon oxidation of 2-oxo acids (Eur. J. Biochem. 269) 5015