Novel isoenzyme of 2-oxoglutarate dehydrogenase is
identified in brain, but not in heart
Victoria Bunik
1,2
, Thilo Kaehne
3
, Dmitry Degtyarev
1
, Tatiana Shcherbakova
2
and Georg Reiser
4
1 Bioengineering and Bioinformatics Department, Lomonosov Moscow State University, Russia
2 Belozersky Institute of Physico-Chemical Biology, Lomonosov Moscow State University, Russia
3 Institute of Experimental Internal Medicine, Otto-von-Guericke University Magdeburg, Germany
4 Institute of Neurobiochemistry, Medical Faculty, Otto-von-Guericke University Magdeburg, Germany
The 2-oxoglutarate dehydrogenase complex (OGDHC)
is a key regulator of a branch point in the tricarboxylic
acid cycle. It belongs to the family of 2-oxo acid dehy-
drogenase complexes which comprise multiple copies
of the three catalytic enzyme components: E1, thia-
mine diphosphate (ThDP)-dependent 2-oxo acid dehy-
drogenase (in OGDHC it is E1o); E2, dihydrolipoyl
acyltransferase with the covalently bound lipoic acid
Keywords
2-oxoglutarate dehydrogenase isoenzyme;
mitochondrial membrane; multienzyme
complex; thiamine; tricarboxylic acid cycle
Correspondence
V. Bunik, Belozersky Institute of Physico-
Chemical Biology, Lomonosov Moscow
State University, Moscow 119992, Russia
Fax: +7 495 939 31 81
Tel: +7 495 939 44 84
E-mail: bunik@belozersky.msu.ru
G. Reiser, Institut fu
¨r Neurobiochemie,
Medizinische Fakulta
¨t, Otto-von-Guericke-
Universita
¨t Magdeburg, Leipziger Straße 44,
39120 Magdeburg, Germany
Fax: +49 391 67 13097
Tel:+49 391 67 13088
E-mail: georg.reiser@med.ovgu.de
(Received 17 April 2008, revised 5 July
2008, accepted 8 August 2008)
doi:10.1111/j.1742-4658.2008.06632.x
2-Oxoglutarate dehydrogenase (OGDH) is the first and rate-limiting com-
ponent of the multienzyme OGDH complex (OGDHC) whose malfunction
is associated with neurodegeneration. The essential role of this complex in
the degradation of glucose and glutamate, which have specific significance
in brain, raises questions about the existence of brain-specific OGDHC iso-
enzyme(s). We purified OGDHC from extracts of brain or heart mitochon-
dria using the same procedure of poly(ethylene glycol) fractionation,
followed by size-exclusion chromatography. Chromatographic behavior
and the insufficiency of mitochondrial disruption to solubilize OGDHC
revealed functionally significant binding of the complex to membrane.
Components of OGDHC from brain and heart were identified using nano-
high performance liquid chromatography electrospray tandem mass spec-
trometry after trypsinolysis of the electrophoretically separated proteins. In
contrast to the heart complex, where only the known OGDH was deter-
mined, the band corresponding to the brain OGDH component was found
to also include the novel 2-oxoglutarate dehydrogenase-like (OGDHL) pro-
tein. The ratio of identified peptides characteristic of OGDH and OGDHL
was preserved during purification and indicated comparable quantities of
the two proteins in brain. Brain OGDHC also differed from the heart com-
plex in the abundance of the components, lower apparent molecular mass
and decreased stability upon size-exclusion chromatography. The func-
tional competence of the novel brain isoenzyme and different regulation of
OGDH and OGDHL by 2-oxoglutarate are inferred from the biphasic
dependence of the overall reaction rate versus 2-oxoglutarate concentra-
tion. OGDHL may thus participate in brain-specific control of 2-oxogluta-
rate distribution between energy production and synthesis of the
neurotransmitter glutamate.
Abbreviations
E1, 2-oxo acid dehydrogenase; E2, dihydrolipoyl acyl transferase; E3, dihydrolipoyl dehydrogenase; nanoLC-MS ⁄MS, nano-high performance
liquid chromatography–electrospray tandem mass spectrometry; OGDH (E1o), 2-oxoglutarate dehydrogenase; OGDHC, 2-oxoglutarate
dehydrogenase complex; OGDHL, 2-oxoglutarate dehydrogenase-like protein; ROS, reactive oxygen species; ThDP, thiamin diphosphate.
4990 FEBS Journal 275 (2008) 4990–5006 ª2008 The Authors Journal compilation ª2008 FEBS
residue (in OGDHC it is E2o); and the terminal com-
ponent E3, FAD-dependent dihydrolipoyl dehydroge-
nase, which is common to all complexes. The
consecutive action of these components within the
multienzyme complex provides for the multistep pro-
cess of oxidative decarboxylation of a 2-oxo acid
(R = -CH
2
-CH
2
-COOH for 2-oxoglutarate; R =
-CH
3
for pyruvate):
According to reaction (1), oxidative decarboxylation
of 2-oxoglutarate produces energy in the form of
NADH and a macroergic acyl thioester bond of succi-
nyl-CoA. Essential for aerobic energy production in all
tissues, the reaction also involves the important
branch-point metabolites 2-oxoglutarate and succinyl-
CoA and may thus be subject to differential regulation
according to the tissue-specific metabolic network. In
particular, succinyl-CoA, which in mammalian mito-
chondria may be used for the substrate-level phosphor-
ylation of GDP or ADP, is preferentially transformed
into ATP in brain [1]. 2-Oxoglutarate is generated both
within the tricarboxylic acid cycle and through gluta-
mate transamination and oxidative deamination. The
ensuing role of OGDHC in the degradation of gluta-
mate, which is neurotoxic in excess, is in accordance
with the known association between reduced OGDHC
activity and neurodegeneration, both age-related [2]
and inborn [3,4]. Furthermore, 2-oxoglutarate takes
part in metabolic signaling [5–10], and therefore its
degradation by OGDHC may affect signal transduc-
tion. Regulated by thioredoxin, OGDHC is at the
intercept of not only energy production and glutamate
turnover, but also mitochondrial production ⁄scaveng-
ing of reactive oxygen species (ROS) [11]. To tune
these pathways to the specific demands of the brain,
the featured integration of OGDHC into the cell-
specific metabolic network is required. This may be
achieved through the expression of isoenzymes, their
structural differences providing for specificity in both
regulation and protein–protein interactions. However,
no tissue-specific isoenzymes of the OGDHC compo-
nents have been isolated to date. Moreover, the insta-
bility of brain OGDHC during purification interferes
with obtaining the brain complex in a homogeneous
state [12]. In addition to general problems known to
arise upon enzyme purification from fat-rich brain tis-
sue, the isolation of functional 2-oxo acid dehydro-
genase multienzyme complexes poses additional
challenges regarding the preservation of non-covalent
protein–protein interactions which determine the native
structure of such megadalton systems. In this study,
we therefore aimed at structural characterization of
brain OGDHC using approaches that do not require
the complex to be purified to homogeneity. In parti-
cular, MS analysis is used to identify the individual
proteins and their relative abundance in complex pro-
tein mixtures [13–15]. Using this technique, we ana-
lyzed a preparation of brain OGDHC which was
purified to an extent that enabled kinetic study of the
complex. As a result, the structure and function of
brain OGDHC were characterized under conditions
that preserved the native state of the complex. Specific
features of brain OGDHC were revealed by compari-
son with OGDHC from heart. We show that, in
contrast to heart, the brain preparation comprises
comparable amounts of both the known 2-oxogluta-
rate dehydrogenase and its novel isoenzyme, a hith-
erto hypothetical 2-oxoglutarate dehydrogenase-like
(OGDHL) protein, with the isoenzyme ratio preserved
during the purification of OGDHC by different proce-
dures. Although the existence of OGDHL has been
inferred from nucleic acid data, with recent structure–
function analysis predicting it to be a novel OGDH
isoenzyme [16], the protein has not been reported in
mammalian mitochondrial proteomes [17–19]. We
show that the presence in brain of the novel isoenzyme
of the first component of OGDHC is accompanied by
a different supramolecular organization and stability
of the complex. Our kinetic study corroborates the cat-
alytic competence of the novel isoenzyme in the overall
OGDHC reaction predicted previously [16], and also
reveals specific regulation of the two isoenzymes by
2-oxoglutarate, which may have implications for brain
glutamate metabolism.
Results
Solubilization and partial purification of OGDHC
from rat brain and heart mitochondria
The 2-oxo acid dehydrogenase complexes are presumed
to be enzymes of the mitochondrial matrix. Accordingly,
given that the mitochondria were disrupted, their purifi-
cation was carried out without detergents [20,21]. Later,
it was found that detergents may improve the solubiliza-
tion of both the pyruvate and 2-oxoglutarate dehydro-
genase complexes from mammalian tissues at different
stages of purification [22–25], although the mecha-
nism(s) of their solubilizing action on the complexes
have not been systematically studied. In order to better
preserve native enzyme regulation and protein–protein
interactions, we attempted to obtain detergent-free
OGDHC from isolated brain mitochondria using soni-
cation only. Solubilization was controlled by following
V. Bunik et al. Novel 2-oxoglutarate dehydrogenase
FEBS Journal 275 (2008) 4990–5006 ª2008 The Authors Journal compilation ª2008 FEBS 4991
the distribution of OGDHC activity between the
supernatant and the detergent extract of the broken
mitochondria pellet. Mitochondrial disruption with the
probe sonicator did not reproducibly solubilize
OGDHC activity. Although disruption was evident
from the appearance in the supernatant of the activity of
the third component of the mitochondrial 2-oxo acid
dehydrogenase complexes, dihydrolipoyl dehydroge-
nase, overall OGDHC activity (Reaction 1) remained in
the broken mitochondria pellet and was solubilized from
the pellet only in the presence of detergent (6%Tri-
ton X-100 or 1% Chaps). By contrast, sonication using
‘Bioruptor’ enabled reproducible solubilization of the
majority (90%) of OGDHC activity from brain
mitochondria without detergents. This preparation is
further referred to as ‘soluble’ OGDHC. A similar
procedure with heart mitochondria left significant
amounts of OGDHC in the pellet. Hence, 1% Chaps
was used to fully solubilize the heart complex from the
pellet. Detergent extraction was also used for brain
OGDHC when its solubilization by sonication was not
efficient or complete. Such preparations are further
called ‘detergent-extracted’ OGDHC. Independent of
the OGDHC extraction details, the majority of the
complex from the two tissues solubilized together with
the integral membrane proteins, such as mitochondrial
ADP ⁄ATP translocase and other transporters (voltage-
dependent anion channel, tricarboxylate, 2-oxoglutarate
and phosphate carriers). These membrane proteins were
identified by nanoLC-MS ⁄MS in bands 8 and 9 of
Fig. 1. Thus, our data on the solubilization of OGDHC
activity and the accompanying proteins indicate that in
mitochondria from both brain and heart OGDHC inter-
acts rather strongly with the membrane fraction.
Unlike the heart complex [23], OGDHC from brain
was much more prone to lose its activity under gel-
filtration conditions which fully resolved it from the
pyruvate dehydrogenase complex. Because of this, rela-
tively rapid gel-filtration on Sephacryl HR300 16 ⁄60 or
Sephacryl S300 12 ⁄30 columns was used to purify the
OGDHC-enriched fraction of the 2-oxo acid dehydro-
genase complexes (molecular mass in the range 10
6
–
10
7
Da) from the proteins of a lower molecular mass
(10
5
–10
6
Da). Although this fraction contained compo-
nents of the pyruvate dehydrogenase complex, as
shown below, the activity peak of the latter complex
was shifted to lower elution volumes compared with
OGDHC. Being rather low even at its peak, the pyru-
vate dehydrogenase reaction rate in the OGDHC-
enriched fraction did not exceed 10% of the rate of
the 2-oxoglutarate dehydrogenase reaction. Impor-
tantly, the elution profile of the common E3 compo-
nent of the two complexes coincided with the elution
of OGDHC, indicating that there was no significant
contribution of the pyruvate dehydrogenase complex-
bound E3 to the E3 content of our OGDHC-enriched
preparation. The latter fraction also lacked the
branched chain 2-oxo acid dehydrogenase complex, as
neither component of the complex was identified by
MS analysis, nor was the activity with 2-oxoisovaleric
acid detected. Components of the glycine cleavage
system, which also includes E3, were not identified in
the OGDHC-enriched fraction. No co-elution of the
glycine cleavage system in the high molecular mass
fraction comprising the pyruvate and 2-oxoglutarate
dehydrogenase complexes was expected, as this com-
plex is much smaller and dissociates easily into its
components [26].
ABC
Fig. 1. Comparison of the SDS electrophoretic patterns of OGDHC preparations from brain and heart mitochondria upon separation on 10%
(A, C) and 7% (B) gels. Molecular mass markers (kDa) are indicated on the right, lane numbers are given in the upper row, protein bands
are numbered on the left. (A) Brain OGDHC solubilized using ‘Bioruptor’ sonication (lane 1); 1% Chaps extract of the pellet from ‘Bioruptor’-
sonicated mitochondria (lane 2); heart OGDHC solubilized by 1% Chaps after ‘Bioruptor’ sonication (lane 3); markers (lane 4). (B) Heart
OGDHC solubilized by 1% Chaps after ‘Bioruptor’ sonication (lane 1); brain OGDHC solubilized by ‘Bioruptor’ sonication (lane 2); markers
(lane 3). (C) Brain OGDHC solubilized by the probe sonicator (lane 1); markers (lane 3).
Novel 2-oxoglutarate dehydrogenase V. Bunik et al.
4992 FEBS Journal 275 (2008) 4990–5006 ª2008 The Authors Journal compilation ª2008 FEBS
A comparison of the SDS electrophoretic patterns of
partially purified heart and brain complexes is shown
in Fig. 1A,B. Varying the concentration of the separat-
ing gel (10% in Fig. 1A and 7% in Fig. 1B) allowed
for a better resolution of some proteins, in particular,
those in band 6. The SDS electrophoretic pattern of
our preparation from rat heart mitochondria (Fig. 1A,
lane 3; Fig. 1B, lane 1) agrees with the known mobility
of the components of bovine heart complexes isolated
from total heart extract [23]. According to the molecu-
lar mass values for the mature proteins, the components
of the 2-oxoglutarate and pyruvate dehydrogenase
complexes were ascribed to the major protein bands of
our preparation as follows: E1o (band 1), E2p (band 2),
E3 (band 3), E2o and the E3-binding component of the
pyruvate dehydrogenase complex (protein X; band 4),
E1pa(band 6), E1pb(band 8). This was confirmed by
nanoLC-MS ⁄MS identification of the components in the
protein bands (Table 1). Our study also showed that
there are two isoenzymes of pyruvate dehydrogenase
kinases in brain (band 6a). Isoenzymes 2 and 3 were
distinguished by three and five specific peptides out of
four and six total peptides identified, respectively
(Table 1).
Interaction of OGDHC with membraneous
proteo-lipid particles and its functional significance
With the sonication parameters fixed, later elution on
size-exclusion chromatography on a Sephacryl HR300
column was observed for OGDHC extracted using
detergent compared with OGDHC solubilized by soni-
cation only. V
e
decreased reproducibly, from 47 to
44 mL for brain OGDHC and from 44 to 42 mL for
heart OGDHC, with standard deviations in V
e
between different chromatographies of a certain prepa-
ration type of < 1 mL. Concomitantly, the shift in V
e
was observed for the high molecular mass opalescent
peak eluted between the column void volume
(V
0
= 38 mL) and the OGDHC activity peak (V
e
between 42 and 47 mL) (Fig. 2A). Elution near the
void volume of the column (Fig. 2A), high opalescence
at a relatively low protein level and the dependence of
V
e
on both the detergent and the sonication mode sug-
gest that this peak comprises membraneous particles.
Membrane vesicles that form spontaneously during
homogenization are known as the microsomal fraction
[27]. A strong dependence of the elution volume of
OGDHC on the elution volume of the opalescent peak
(Fig. 2B, correlation coefficient 1.13) points to OG-
DHC binding to these membrane particles, with their
complex disrupted by the chromatography-accom-
plished trapping of the dissociated intermediates.
Table 2 shows that the better the separation of
OGDHC from microsomes, the more E1o and E3
dissociate from the complex, accompanied by a loss of
total OGDHC activity when subjected to chromatogra-
phy. Increasing dissociation was obvious from the
appearance of the well-defined peak for the compo-
nent activities (DV
e
„0; Table 2), which follows the
Table 1. MS identification of known components of the 2-oxo acid dehydrogenase complexes from brain. Proteins of the bands shown in
Fig. 1 were identified through an NCBI search using MASCOT as described in Experimental procedures. The data for a representative experi-
ment are given. Components of the 2-oxoglutarate dehydrogenase complex were also identified in heart. Unless indicated otherwise,
matches to rat sequences were found. Molecular mass corresponds to the precursor proteins as given in NCBI. NCBI-provided molecular
mass of dihydrolipoyllysine acetyltransferase refers to an incomplete sequence, therefore the true molecular mass from the Expasy data-
base, which corresponds to that in the SDS-electrophoresis (Fig.1), is added (marked by asterisk). NA, not analyzed.
Band
in Fig. 1
Component of the 2-oxo acid dehydrogenase
complexes
NCBI
identifier
Molecular
mass (Da)
Brain Heart
Protein
score
No.
peptides
matched
Protein
score
No.
peptides
matched
1 2-Oxoglutarate dehydrogenase (E1o) 62945278 117 419 1131 28 1647 60
2 Dihydrolipoyllysine acetyltransferase (E2p) 220838 57 645
67 166*
443 16 NA
3 Dihydrolipoyl dehydrogenase (E3) 40786469 54 574 579 12 975 36
4 Dihydrolipoyl succinyl transferase (E2o) 55742725 49 236 400 7 709 28
4 Component X 28201978 mus 54 250 126 2 279 7
6a Pyruvate dehydrogenase kinase, isoenzyme 3 21704122 mus 48 064 196 6 NA
6a Pyruvate dehydrogenase kinase 2 subunit
variant p45
8895958 44 198 151 4 NA
6b Pyruvate dehydrogenase alpha subunit (E1pa) 57657 43 853 716 20 NA
8 Pyruvate dehydrogenase beta subunit (E1pb) 56090293 39 299 519 26 NA
V. Bunik et al. Novel 2-oxoglutarate dehydrogenase
FEBS Journal 275 (2008) 4990–5006 ª2008 The Authors Journal compilation ª2008 FEBS 4993
overall OGDHC activity peak, and an increased ratio
of dissociated to complex-bound activities for E3 and
E1o at the corresponding elution volumes. Importantly,
the chromatography-induced dissociation into compo-
nents and the accompanying loss of total OGDHC
activity were dependent on the separation from micro-
somes rather than on the protein applied (Table 2;
experiment N 1 versus 3). Because of the higher analy-
tical sensitivity of the E3-catalyzed NAD
+
reduction
compared with ferricyanide reduction by E1o, the
E3-catalyzed reaction allowed a better comparison of
the significantly different levels of the component activ-
ities obtained in these experiments. However, a similar
trend was observed for the two components (Table 2),
in good agreement with the known formation of the
E1o–E3 subcomplex upon OGDHC dissociation [28].
Separation from microsomes decreases both the total
and the specific (lmolÆmin
)1
Æmg
)1
of protein) activity
of OGDHC in the peak. Table 2 shows that purifica-
tion of OGDHC by chromatography led to a 30-fold
increase in specific activity with a low degree of
separation from microsomes (experiment 1), but full
separation (experiment 3) resulted in no increase in spe-
cific activity, despite the OGDHC fractions containing
fewer contaminant proteins. Thus, disruption of the
interaction between OGDHC and the microsomal frac-
tion during chromatography destabilizes the complex
structure and function.
At a comparable protein concentration in the
column eluate, the fraction of applied OGDHC activ-
ity found in the eluate differed dramatically for heart
(70%) and brain (10%) complexes. The greater loss of
brain OGDHC activity (90%) compared with that
from heart complex (30%) was not due to a higher
degree of purification, because more proteins co-eluted
with OGDHC from brain. This was evident from the
additional bands on SDS electrophoresis (bands 6a, 7,
8a in Fig. 1) and the greater heterogeneity indicated by
nanoLC-MS ⁄MS analysis of common bands 1, 3, 4, 5.
The tissue specificity of the heterogeneity was mostly
due to synaptosomal proteins in the brain preparation,
Fig. 2. Gel filtration of brain OGDHC on a Sephacryl HR300 16 ⁄60
column. (A) Elution profile, showing attenuance at 280 nm (D
280
)
and the OGDHC activity in arbitrary units (A). (B) Dependence of V
e
of OGDHC on V
e
of membraneous fraction, the line is drawn
according to the equation: y=1.13x)3.05.
Table 2. Dependence of the OGDHC activity yield on the separation of OGDHC from microsomes. Partially purified from ‘Bioruptor’-soni-
cated mitochondria, OGDHC (40–60 mgÆmL
)1
) was applied to the 12 ⁄30 column with Sephacryl S-300. The separation varied due to the dif-
ferences in the sample volume and ⁄or relative content of the microsomes. The interference of the elution volumes of OGDHC and
microsomes, I, was calculated from the elution profiles as the percentage of the microsome-including OGDHC fractions to the total number
of the OGDHC-containing fractions. Separation of E3 or E1o from the complex upon chromatography was characterized by the difference
between the elution volumes, DV
e
, of the peaks of E3 or E1o and OGDHC and the ratio of the component activities at these V
e
(A
non-bound
E3 ⁄E1o
⁄A
bound E3 ⁄E1o
). The OGDHC activity yield is the ratio of the total activity of OGDHC in the eluate to the total activity of the OGDHC
applied to the column. ND, not determined.
No.
Total
protein
applied
(mg)
Separation
of OGDHC
and microsomes,
(100 )I) (%)
Dissociation of E1o
from OGDHC
Dissociation of E3
from OGDHC Total
OGDHC
activity
yield (%)
Specific
OGDHC
activity
increase (%)
DV
e
Anon-bound E1o
Abound E1o DV
e
Anon-bound E3
Abound E3
1 80 25 0 0.3 0 0.8 66 3000
2 40 56 3 0.9 3 1.9 24 300
3 80 100 ND ND 3 2.7 9 100
Novel 2-oxoglutarate dehydrogenase V. Bunik et al.
4994 FEBS Journal 275 (2008) 4990–5006 ª2008 The Authors Journal compilation ª2008 FEBS