Báo cáo khoa học: Novel isoenzyme of 2-oxoglutarate dehydrogenase is identified in brain, but not in heart potx

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 dehydro

identified in brain, but not in heart

Victoria Bunik1,2, Thilo Kaehne3, Dmitry Degtyarev1, Tatiana Shcherbakova2and Georg Reiser4

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.

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 = -CH2-CH2-COOH for 2-oxoglutarate; R =

-CH3for 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

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 106–

107Da) from the proteins of a lower molecular mass (105–106Da) 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]

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).

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 Ve decreased reproducibly, from 47 to

44 mL for brain OGDHC and from 44 to 42 mL for heart OGDHC, with standard deviations in Ve between different chromatographies of a certain prepa-ration type of < 1 mL Concomitantly, the shift in Ve was observed for the high molecular mass opalescent peak eluted between the column void volume (V0= 38 mL) and the OGDHC activity peak (Ve 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

Veon 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 (DVe „ 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)

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*

3 Dihydrolipoyl dehydrogenase (E3) 40786469 54 574 579 12 975 36

4 Dihydrolipoyl succinyl transferase (E2o) 55742725 49 236 400 7 709 28

6a Pyruvate dehydrogenase kinase, isoenzyme 3 21704122 mus 48 064 196 6 NA

6a Pyruvate dehydrogenase kinase 2 subunit

variant p45

6b Pyruvate dehydrogenase alpha subunit (E1pa) 57657 43 853 716 20 NA

8 Pyruvate dehydrogenase beta subunit (E1pb) 56090293 39 299 519 26 NA

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)1Æmg)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 (D280)

and the OGDHC activity in arbitrary units (A) (B) Dependence of Ve

of OGDHC on Ve 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, DVe, of the peaks of E3 or E1o and OGDHC and the ratio of the component activities at these Ve(Anon-bound E3 ⁄ E1o ⁄ Abound 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

OGDHC activity yield (%)

Specific OGDHC activity increase (%)

DVe

Anon-bound E1o Abound E1o DVe

Anon-bound E3 Abound E3

pointing to the presence of synaptosome-derived

microsomes in the membraneous fraction

accompany-ing brain OGDHC

Structural differences between OGDHC from

brain and heart

An essential difference between brain and heart

OGDHC was revealed by nanoLC-MS⁄ MS analysis of

band 1 In the brain preparation, this band contained

both OGDH and OGDHL, a hypothetical isoenzyme

of OGDH predicted from the nucleic acid data [16]

Our analysis of 10 band 1 samples from 9 different

brain preparations identified the structures of 10–17

peptides which were specific for OGDH and 5–10

pep-tides specific for OGDHL (Table 3, Fig 3) Although

direct quantification of proteins from the

nanoLC-MS⁄ MS peak intensities is difficult, there is a general

correlation between the number of protein peptides

identified and the amount of protein present in the

mixture, if protein size is normalized [13–15] For

OGDH and OGDHL, which have similar molecular

masses, the ratio of identified peptides may be taken as

an estimate of the relative abundance of the

isoen-zymes in the analyzed sample We calculated this ratio

for OGDH and OGDHL, using either the number of

all peptides identified or only those specific for the sequences and that were non-redundant (when peptides with the same primary sequence were counted as one) The latter excludes a possible bias due to common peptides, and is thus a better measure of the specific sequence coverage However, with the high sequence coverage for each of the isoenzymes (Fig 3), both cal-culations give a similar ratio The ratio points to a comparable amount of the two isoenzymes in the brain preparation ( 60% OGDH and 40% OGDHL; Table 3) No reproducible enrichment of OGDHC with one of the isoenzymes could be detected in the different OGDHC preparations, for example, isolated with or without detergents, before or after gel-filtra-tion, precipitated by either poly(ethylene glycol) or

pH, and collected from different pools of column elu-ate, which may vary in the OGDHC saturation by peripheral components E1o and E3 (Table 3) It is worth noting that the same isoenzyme ratio was observed in both the crude poly(ethylene glycol) frac-tion of the mitochondrial extract and the chromatogra-phy-purified OGDHC (Table 3) Co-purification of the novel isoenzyme with the high molecular mass OGDHC fraction points to OGDHL being the com-plex component, in good agreement with predictions based on the structural analysis [16]

Table 3 Ratio of the peptides characteristic of OGDH and OGDHL isoenzymes in different preparations of brain OGDHC Samples isolated under the indicated conditions (details in Experimental procedures) were subjected to SDS electrophoresis, and the OGDH ⁄ OGDHL band of

110 kDa was analyzed using nanoLC-MS ⁄ MS The indicated number of specific peptides refers to the non-redundant peptides only, i.e the same peptide modified or of a reduced length was not counted The total number of peptides found by MASCOT search, as described in Experimental procedures is given in parentheses.

Isolation conditions

OGDH-specific (total) peptides

OGDHL-specific (total) peptides

Specific (total) peptide ratio (% OGDH : OGDHL)

‘Bioruptor’ + PEG before chromatography 10 (17) 6 (11) 60 : 40 (60 : 40)

‘Bioruptor’ + PEG

‘Bioruptor’ + pH

Ve = 44–48 mL

‘Bandelin’ + PEG

Ve= 43–46 mL

OGDHC solubilized by sonication only 8–17 (17–28)

average 13 (22)

5–10 average 7 (13)

Average 65 : 35 (60 : 40) (50 : 50 to 80 : 20)

‘Bandelin’, 6% Triton X-100 extract + pH

Ve = 42–48 mL

‘Bioruptor’, 1% Chaps extract + pH

Ve= 45–48 mL

‘Bioruptor’, 1% Chaps extract + PEG

Ve= 45–48 mL

Detergent-solubilized OGDHC 11–13 (17–23)

average 12 (20)

6–9 (12–19) average 7 (15)

Average 60 : 40 (60 : 40) (60 : 40 to 70 : 30)

OGDHL_rat_h MSQLRLLLFRLGP QARKLLATRDIAAFG GRRRSSGPPTTIPRSRGGVSPSYVEEMYFAWLENPQSVHKSWDNFF 74

OGDHL_rat_b MSQLRLLLFRLGP QARKLLATRDIAAFG GRRR SSGPPTTIPR SRGGVSPSYVEEMYFAWLENPQSVHK SWDNFF 74

OGDHL_rat_h QRATKEASVGPAQPQPP -AVIQESRASVSSCTKTSKLVEDHLAVQSLIRAYQIRGHHVAQLDPLGILDADLDSF 147

OGDH_rat_b RNTNAGAPPGTAYQSPLSLSR SSLATMAHAQSLVEAQPNVDK LVEDHLAVQSLIR AYQIRGHHVAQLDPLGILDADLDSS 160 OGDHL_rat_b QR ATK EASVGPAQPQPP - AVIQESR ASVSSCTKTSK LVEDHLAVQSLIR AYQIRGHHVAQLDPLGILDADLDSF 147

OGDHL_rat_h VPSDLITTIDKLAFYDLQEADLDKEFRLPTTTFIGGSENTLSLREIIRRLESTYCQHIGLEFMFINDVEQCQWIRQKFET 227 OGDH_rat_h VPADIISSTDK LGFYGLHESDLDK VFHLPTTTFIGGQEPALPLREIIRRLEMAYCQHIGVEFMFINDLEQCQWIR QKFET 240 OGDH_rat_b VPADIISSTDKLGFYGLHESDLDKVFHLPTTTFIGGQEPALPLREIIRRLEMAYCQHIGVEFMFINDLEQCQWIRQKFET 240

OGDHL_rat_h PGVMKFSIEEKR TLLAR LVRSMRFEDFLARKWSSEKRFGLEGCEVMIPALKTIIDKSSEMGVENVILGMPHRGR LNVLAN 307 OGDH_rat_h PGIMQFTNEEK R TLLAR LVRSTR FEEFLQR KWSSEKR FGLEGCEVLIPALK TIIDMSSANGVDYVIMGMPHRGR LNVLAN 320 OGDH_rat_b PGIMQFTNEEKRTLLARLVRSTR FEEFLQR KWSSEKR FGLEGCEVLIPALK TIIDMSSANGVDYVIMGMPHRGR LNVLAN 320 OGDHL_rat_b PGVMKFSIEEKRTLLARLVRSMRFEDFLARKWSSEKR FGLEGCEVMIPALK TIIDK SSEMGVENVILGMPHR GR LNVLAN 307

OGDH_rat_h VIR KELEQIFCQFDSKLEAADEGSGDMKYHLGMYHR RINRVTDRNITLSLVANPSHLEAADPVVMGKTK AEQFYCGDTEG 400 OGDH_rat_b VIR K ELEQIFCQFDSKLEAADEGSGDMK YHLGMYHRRINRVTDRNITLSLVANPSHLEAADPVVMGKTK AEQFYCGDTEG 400 OGDHL_rat_b VIR KDLEQIFCQFDPK LEAADEGSGDVK YHLGMYHERINRVTNRNITLSLVANPSHLEAVDPVVQGKTKAEQFYRGDAQG 387

OGDH_rat_h K KVMSILLHGDAAFAGQGIVYETFHLSDLPSYTTHGTVHVVVNNQIGFTTDPRMAR SSPYPTDVAR VVNAPIFHVNSDDP 480 OGDH_rat_b K KVMSILLHGDAAFAGQGIVYETFHLSDLPSYTTHGTVHVVVNNQIGFTTDPRMAR SSPYPTDVAR VVNAPIFHVNSDDP 480 OGDHL_rat_b RKVMSILVHGDAAFAGQGVVYETFHLSDLPSYTTNGTVHVVVNNQIGFTTDPRMAR SSPYPTDVAR VVNAPIFHVNADDP 467

OGDH_rat_h EAVMYVCK VAAEWRNTFHK DVVVDLVCYR R NGHNEMDEPMFTQPLMYK QIRKQKPVLQKYAELLVSQGVVNQPEYEEEIS 560

OGDH_rat_h KYDK ICEEAFTR SKDEK ILHIK HWLDSPWPGFFTLDGQPRSMTCPSTGLEEDILTHIGNVASSVPVENFTIHGGLSRILK 640

OGDH_rat_h TRRELVTNRTVDWALAEYMAFGSLLKEGIHVR LSGQDVER GTFSHR HHVLHDQNVDKR TCIPMNHLWPNQAPYTVCNSSL 720 OGDH_rat_b TRRELVTNRTVDWALAEYMAFGSLLKEGIHVR LSGQDVER GTFSHR HHVLHDQNVDKR TCIPMNHLWPNQAPYTVCNSSL 720 OGDHL_rat_b GRADMTKKRTVDWALAEYMAFGSLLKEGIHVR LSGQDVER GTFSHR HHVLHDQDVDRR TCVPMNHLWPDQAPYTVCNSSL 707

OGDHL_rat_h SEYGVLGFELGYAMASPNALVLWEAQFGDFHNTAQCIIDQFISTGQAKWVRHNGIVLLLPHGMEGMGPEHSSARPERFLQ 787 OGDH_rat_h SEYGVLGFELGFAMASPNALVLWEAQFGDFNNMAQCIIDQFICPGQAKWVRQNGIVLLLPHGMEGMGPEHSSARPERFLQ 800 OGDH_rat_b SEYGVLGFELGFAMASPNALVLWEAQFGDFNNMAQCIIDQFICPGQAKWVRQNGIVLLLPHGMEGMGPEHSSARPERFLQ 800 OGDHL_rat_b SEYGVLGFELGYAMASPNALVLWEAQFGDFHNTAQCIIDQFISTGQAKWVRHNGIVLLLPHGMEGMGPEHSSARPERFLQ 787 OGDHL_rat_h MSNDDSDAYP-VFTEDFEVSQLYDCNWIVVNCSTPASYFHVLRRQVLLPFR KPLIVFTPK SLLRHPDAKSSFDQMVSGTS 866 OGDH_rat_h MCNDDPDVLPNLQEENFDISQLYDCNWIVVNCSTPGNFFHVLR R QILLPFR KPLIVFTPK SLLRHPEAR TSFDEMLPGTH 880 OGDH_rat_b MCNDDPDVLPNLQEENFDISQLYDCNWIVVNCSTPGNFFHVLRRQILLPFR KPLIVFTPK SLLRHPEAR TSFDEMLPGTH 880 OGDHL_rat_b MSNDDSDAYP-VFTEDFEVSQLYDCNWIVVNCSTPASYFHVLRRQVLLPFR KPLIVFTPK SLLRHPDAK SSFDQMVSGTS 866

OGDHL_rat_h FQRMIPEDGPAAQSPERVERLIFCTGKVYYDLVKERSSQGLEKQVAITRLEQISPFPFDLIMREAEKYSGAELVWCQEEH 946 OGDH_rat_h FQRVIPEDGPAAQNPDK VKR LLFCTGKVYYDLTR ERKAR DMAEEVAITR IEQLSPFPFDLLLK EAQKYPNAELAWCQEEH 960 OGDH_rat_b FQRVIPEDGPAAQNPDKVK RLLFCTGK VYYDLTR ERKAR DMAEEVAITR IEQLSPFPFDLLLK EAQKYPNAELAWCQEEH 960

OGDHL_rat_h KNMGYYDYISPRFMTLLGHSRPIWYVGREPAAAPATGNKNTHLVSLRKFLDTAFNLKAFEGKTF 1010

OGDH_rat_h K NQGYYDYVKPRLRTTIDRAKPVWYAGRDPAAAPATGNKK THLTELQRFLDTAFDLDAFKK FS- 1023

OGDH_rat_b K NQGYYDYVKPRLRTTIDRAKPVWYAGRDPAAAPATGNK KTHLTELQR FLDTAFDLDAFKKFS- 1023

OGDHL_rat_b K NMGYYDYISPR FMTLLGHSRPIWYVGREPAAAPATGNKNTHLVSLRK FLDTAFNLK AFEGKTF 1010

*

Fig 3 Sequence alignment of rat OGDH and OGDHL showing (in color) the peptides identified by nanoLC-MS ⁄ MS in the OGDHC prepara-tion from heart (two upper sequences marked by ‘h’) and brain (two lower sequences marked by ‘b’) Common peptides for the two sequences are shown in red The sequence-specific peptides are in bold: pink for the OGDH and blue for the OGDHL The N-terminal cleav-age site, as determined by the sequencing of the truncated bovine E1o [35], is marked by an asterisk above the alignment.

In contrast to brain OGDHC, no peptide specific

for OGDHL was identified in the heart complex,

despite the higher protein load and the purity of the

E1o band (band 1, lane 3 versus lane 1; Fig 1A),

which resulted in an increase in the sequence coverage

(26–36 non-redundant or 45–60 total peptides in

inde-pendent determinations) As shown in Fig 3, heart

preparation exhibits either OGDH-specific peptides

(pink) or peptides common to the two proteins (red),

but OGDHL-specific peptides (blue) were found in the

brain preparation only Thus, whereas only the known

OGDH component coded by chromosome 7 in

humans [29,30] was identified by nanoLC-MS⁄ MS in

OGDHC from heart, brain complex, purified using

the same procedure, contained comparable amounts

of both OGDH (chromosome 7) and OGDHL

(human chromosome 10) [31–33] proteins (Table 3,

Fig 3), which were identified even at lower

purifica-tion yields

Another structural feature of brain 2-oxo acid

dehy-drogenase complexes is seen from SDS electrophoresis

The E3 component (band 3), the majority of which is

associated with OGDHC as shown above, is hardly

visible in the brain preparation (Fig 1A, lanes 1–2)

compared with the heart preparation (Fig 1A, lane 3)

Despite the low E3 level, under standard assay

condi-tions we did not observe any activation of brain

OGDHC in the presence of or following preincubation

with at least a 10-fold protein excess of E3

(commer-cial bovine enzyme) Thus, even the low levels of E3

seen in the brain preparation were able to support

maximal OGDHC reaction rates This is in accordance

with published data on the rate-limiting role of the

E1o component in Reaction (1) catalyzed by the

com-plex [34] It is known that the binding of E3 to

OG-DHC is mediated by E1o, with the proteolytic removal

of a small N-terminal fragment of E1o impairing

bind-ing [28,35] However, the lower E3 level in brain

OGDHC was not due to E1o proteolysis, because

several peptides preceding the cleavage site (marked by

asterisk in Fig 3) were identified in both isoenzymes

by MS analysis This was in good agreement with the

mobility of the E1o band on SDS electrophoresis

(Fig 1), which corresponded to the molecular mass of

non-proteolysed E1o (110 kDa), being higher than that

of truncated E1o with an apparent molecular mass of

94 kDa [28,35] Because full extraction of the OGDHC

activity from heart mitochondria required 1% Chaps,

we checked whether the E3 deficiency of brain

OGDHC could be due to the membrane binding of its

E3 Figure 1A shows that 1% Chaps extract of the

pellet fraction (lane 2) obtained after removal of

E3-deficient OGDHC (lane 1) did not contain E3 By

contrast, when the activity of E3 and OGDHC was followed in parallel upon sonication, a significant portion of the E3 activity solubilized before the overall activity of OGDHC Taken together, these findings indicate that the E3 deficiency of brain 2-oxo acid dehydrogenase complexes (Fig 1) is not due to mem-brane binding of the E3 component Compared with heart complexes, easier dissociation of this component appears to occur upon sonication of brain mitochon-dria Indeed, E3 was better presented in complexes that were detergent-extracted after a less efficient soni-cation (Fig 1C) Sonisoni-cation by ‘Bioruptor’ (Fig 1A,B) was nevertheless preferred for the isolation, because it gave reproducible results and did not lead to the high molecular mass aggregates (150–300 kDa) observed in the SDS electrophoresis of OGDHC solubilized with the probe sonicator (Fig 1C)

A different supramolecular organization for OG-DHC from brain and heart was further supported by size-exclusion chromatography, in which proteins of a higher molecular mass are eluted more rapidly, i.e at

a lower elution volume Ve As mentioned above, under the same sonication conditions the activity peak of OGDHC from brain eluted later than that of OGDHC from heart: 44 versus 42 mL for soluble OGDHC and

47 versus 44 mL for Chaps-extracted OGDHC The later elution corresponds to a lower molecular mass for the purified brain complex, which agrees with its lower saturation with peripheral E3 component, as dis-cussed above As inferred from both SDS electropho-resis (Fig 1A,B) and size-exclusion chromatography, the different supramolecular organization of heart and brain OGDHC was further supported by the MS-based estimate of the relative abundance of the complex components in the preparation (Table 4) Abundance coefficients were calculated as described in Experimental procedures according to the previously developed approach of comparative proteomics [13– 15] As indicated by the standard deviation values for the preparations from one tissue, these ratios showed good agreement in different experiments However, the values were clearly different for OGDHC from heart and brain Table 4 shows that in OGDHC from brain the E2o⁄ E1o and E3 ⁄ E1o ratios (40 and 70%, respec-tively) were no more than half those in the heart com-plex (120 and 140%, respectively) Because the heart preparation did not possess OGDHL, we also com-pared the abundance coefficients for brain OGDHC when based on the OGDH content only The brain ratios remained lower than those of heart (Table 4) Thus, compared with the heart complex, OGDHC isolated from brain showed an excess of the first component over the second and third The decrease in

the MS-based abundance of E2o and E3 components

in brain OGDHC correlated with the low intensity of

the E3 band in SDS electrophoresis and a lower

molecular mass of OGDHC from brain versus heart

upon size-exclusion chromatography Thus, the data

obtained using the three independent approaches

sug-gest differences in the supramolecular organization of

OGDHC isolated from heart and brain

Saturation of brain OGDHC with 2-oxoglutarate

Kinetic analysis of the dependence of the overall

activ-ity of brain OGDHC on the saturation with

2-oxoglut-arate agrees with the presence in the preparation of

two isoenzymes of 2-oxoglutarate dehydrogenase

which are functionally competent in Reaction (1)

Sol-ubilized with or without detergents, brain OGDHC

did not exhibit standard Michaelis–Menten kinetics

(Fig 4) That is, simulations using the parameters

yielded by the double reciprocal linearization of the

experimental data showed a systematic shift in the

the-oretical curves to lower rates at high 2-oxoglutarate

saturation (Fig 4A) In view of the identification of

the second isoenzyme of OGDH by nanoLC-MS⁄ MS,

we introduced a second saturation function into the

equation As shown in Fig 4B, this abolished the

inconsistencies between the experiment and the

simula-tion, resulting in a satisfactory description of the

sys-tem behavior at both low and high substrate

saturation The better correspondence between the

experimental data and the two-saturation model is

obvious not only from visual inspection of the

coinci-dence between the experimental points and theoretical

curves in Fig 4B compared with Fig 4A, but also

from an increase in the correlation coefficients (from

0.804 and 0.918 in Fig 4A to 0.986 and 0.997 in

Fig 4B) The biphasic saturation parameters provided

in the legend to Fig 4 show that Km,1and Km,2values,

as well as the contributions of V1 and V2 to the

maximal reaction rate (V = V1+ V2), were similar

for soluble and Chaps-extracted OGDHC Based

on three independent experiments, the following

parameters were obtained: Km,1= 0.07 ± 0.02 mm;

Km,2=0.40 ± 0.07 mm; V1⁄ (V1+ V2) = 45 ± 4%;

V2⁄ (V1+ V2) = 55 ± 4% It is worth noting that the simulation-derived partial contributions of V1 and V2

to the overall V value are close to the relative abun-dance of the isoenzymes as determined by

nanoLC-MS⁄ MS (35–40% of OGDHL and 60–65% of OGDH; Table 3) Moreover, detergents are known to desensi-tize cooperative and allosteric enzymes to effectors, but they do not significantly change the kinetic param-eters of brain OGDHC (Fig 4), in accordance with the lack of change in the isoenzyme ratio caused by detergents (Table 3) Thus, the parameters obtained by simulation of the v(S) dependence according to the model suggested by the MS identification of the two isoenzymes are reproducible and in a good agreement with the MS-based abundance of the isoenzymes in the OGDHC preparation Taken together, the kinetic and

MS data support functional competence of the novel isoenzyme in the overall OGDHC reaction and differ-ent saturation of the two isoenzymes of OGDH with 2-oxoglutarate

Discussion

Identification of novel OGDH isoform and its implication in brain metabolism

Distinguishing proteins with highly similar primary structures, such as the products of alternative splic-ing or of different genes (isoforms or isoenzymes), represents one of the challenges in characterizing the cellular proteome [15] The modern development of

MS analysis provides strong advantages over immu-nological approaches to address this challenge, because determination of isoform-specific peptides distinguishes unambiguously between isoforms which may show cross-reactivity to antibodies [36] In this study, we successfully applied nanoLC-MS⁄ MS to identify both the known OGDH and the hypotheti-cal OGDHL in OGDHC partially purified from brain mitochondria At the same time, only the

Table 4 Relative abundance of the OGDHC components in the preparations Abundance index, A, corresponds to the number of peptides detected by nanoLC-MS ⁄ MS, normalized to the molecular mass of the OGDHC component (see Experimental procedures) The E2o and E3 abundance indexes were related to that of either E1o (the sum of OGDH + OGDHL) or OGDH taken as 100% The data are presented as the average values ± SD.

Tissue

Brain 0.34 ± 0.03 100 0.22 ± 0.04 100 0.14 ± 0.05 40 ⁄ 60 0.23 ± 0.01 70 ⁄ 100

known OGDH was determined in a similar

prepara-tion of the complex from heart Expression of the novel

OGDHL component of OGDHC in brain is in accord

with the isolation of OGDHL cDNA from brain tissue

[31–33], whereas earlier cloning of the OGDH gene

used a fetal liver cDNA library [29] Thus, apart from

housekeeping OGDH, OGDHL is synthesized in

brain Integration of the OGDHL isoenzyme into the

complex, which was predicted by our structure– function analysis [16], is evident from the constant ratio of OGDH and OGDHL during the purification

of brain OGDHC (Table 1), elution of OGDHL in the high molecular mass fraction corresponding to OGDHC, and biphasic saturation with 2-oxoglutarate (Fig 4), indicative of a functional competence of the two isoenzymes in Reaction (1) catalyzed by the complex

Identification of the two isoenzymes of OGDHC by

MS was taken into account in the kinetic modeling of the dependence of the overall reaction rate of brain OGDHC on the 2-oxoglutarate concentration (Fig 4) Indeed, the dependence can be better described by the sum of two saturation processes than by standard Michaelis–Menten kinetics (Fig 4), which is in accord with the contribution to the overall reaction rate of the two isoenzymes having different affinities to 2-oxo-glutarate Moreover, simulation of this model revealed that partial contributions of each of the isoenzymes,

V1 and V2 into the overall reaction rate V are in a good agreement with the MS-based relative abundance

of the isoenzymes (Table 3), suggesting that OGDH and OGDHL have similar catalytic rates The compat-ibility of parameters derived from kinetic modeling and MS analysis strongly supports the plausibility of the model assuming two isoenzymes for interpretation

of the kinetic data High correlation between the simu-lated dependence and experimental data within this model (Fig 4B) did not justify further refinement of the model Thus, kinetic analysis of OGDHC from brain provides experimental evidence in support of an earlier prediction from genome data that OGDHL is a functionally active isoenzyme of OGDH [16] Further-more, the kinetics is indicative of an approximately sixfold difference between Km,1and Km,2characterizing saturation of the two isoenzymes with 2-oxoglutarate Compared with OGDHC from heart and adrenal glands, which are half-saturated with 2-oxoglutarate at 0.2 mm [37–39], OGDHC from brain requires higher concentrations for full saturation (Km,2= 0.40

± 0.06 mm), being sensitive to lower concentrations of 2-oxoglutarate (Km,1= 0.07 ± 0.02 mm) Possessing the two isoenzymes which provide the different Km values, brain OGDHC may thus respond to an expanded interval in the 2-oxoglutarate levels The differential regulation of brain OGDH isoenzymes by the substrate may also address the physiological needs

of brain tissue to establish different steady-state concentrations of 2-oxoglutarate, depending on cellular conditions, compartment or type Compared with other tissues, physiological concentrations of glutamate

in brain differ not only between regions and cell types,

Fig 4 Kinetic analysis of brain OGDHC saturation with

2-oxogluta-rate Hollow circles, soluble OGDHC; filled circles,

detergent-extracted OGDHC Dependence of the reaction rate v (arbitrary

units of the fluorescence change dFÆmin)1Æmg)1of protein) on the

2-oxoglutarate concentration ([S]) was approximated by a single

Michaelis–Menten curve v = 170*[S] ⁄ (0.07 + [S]), r 2 = 0.804 for

soluble OGDHC and v = 480*[S] ⁄ (0.09 + [S]) for

detergent-extracted OGDHC, r 2 = 0.918 (A) or the sum of the two Michaelis–

Menten curves v = 110*[S] ⁄ (0.07 + [S]) + 170*[S] ⁄ (0.47 + [S]),

r2= 0.986 for the soluble OGDHC and v = 350*[S] ⁄ (0.09

+ [S]) + 320*[S] ⁄ (0.42 + [S]), r 2 = 0.997 for detergent-extracted

OGDHC (B) Details of the simulation procedure are given in

Experi-mental procedures.