Identification of the Essential Cysteine Residue in the Active Site of Bovine Pyruvate Dehydrogenase*

Pyruvate dehydrogenase (El), the first catalytic com- ponent of the bovine pyruvate dehydrogenase complex, is composed of two nonidentical subunits in a te- trameric aZfiz form. The sulfhydryl-specific reagent N-ethylmaleimide (NEM) was used to identify the reactivi- ties and function of cysteinyl residues and subsequent identification of these residues in the active site of bo- vine E l . Treatment of E l with 0.2 n m NEM resulted in loss (90%) of enzymatic activity; the inactivation fol- lowed bimolecular reaction kinetics. The inactivation was almost entirely prevented by thiamin pyrophos- phate (TPP) and pyruvate; protection is probably due to formation of the hydroxyethylidene-TPP intermediate. To identify the reactive cysteinyl residues in the active site region, the nonessential SH groups in E l were first modified with NEM in the presence of TPP and pyruvate. After quenching with dithiothreitol and removal of the substrate and cofactor by dialysis, the modified E l was treated with [l4C1NEM to label the exposed cys- teinyl residue(s) in or near the active site region. The data indicate that NEM reacted in the active site region of the E l component with a stoichiometry of 2 mol of [14C]NEM bound per mol of E l

Pyruvate dehydrogenase ( E l ) , the first catalytic component of the bovine pyruvate dehydrogenase complex, is composed of two nonidentical subunits in a tetrameric aZfiz form. The sulfhydryl-specific reagent Nethylmaleimide (NEM) was used to identify the reactivities and function of cysteinyl residues and subsequent identification of these residues in the active site of bovine E l . Treatment of E l with 0.2 n m NEM resulted in loss (90%) of enzymatic activity; the inactivation followed bimolecular reaction kinetics. The inactivation was almost entirely prevented by thiamin pyrophosphate (TPP) and pyruvate; protection is probably due to formation of the hydroxyethylidene-TPP intermediate. To identify the reactive cysteinyl residues in the active site region, the nonessential SH groups in E l were first modified with NEM in the presence of TPP and pyruvate. After quenching with dithiothreitol and removal of the substrate and cofactor by dialysis, the modified E l was treated with [l4C1NEM to label the exposed cysteinyl residue(s) in or near the active site region. The data indicate that NEM reacted in the active site region of the E l component with a stoichiometry of 2 mol of [14C]NEM bound per mol of E l tetramer. The initial rapid labeling of E l with [l4C1NEM established that incorportaion was predominantly into the Q subunit. A single radiolabeled peptide was isolated following V8 protease digestion of radiolabeled E l by [14C]NEM. Sequence analysis of the labeled peptide derived from bovine E l demonstrated that the labeled cysteinyl residue was equivalent to Cys-62 in the Q subunit (mature form) of human E l .
The pyruvate dehydrogenase (El)' component of the mammalian pyruvate dehydrogenase complex (PDC) catalyzes the oxidative decarboxylation of pyruvate in a two-step process. In the first step, it forms COz and hydroxyethylidene-thiamin pyrophosphate (HE=TPP) intermediate. step appears to be the rate-limiting step in the overall PDC reaction (Cate et al., 1980). The acetyl group and reducing equivalents formed are transferred sequentially from the dihydrolipoamide acetyltransferase to CoA and to the dihydrolipoamide dehydrogenase component, respectively (Reed, 1974). Mammalian El is composed of two nonidentical subunits a and p with a tetrameric structure (az&) and is regulated by reversible interconversion between a n active nonphosphorylated form a n d a n inactive phosphorylated form by specific phospho-Ela phosphatase and Ela kinase, respectively. The available evidence suggests that Ela plays a crucial role in catalysis (Roche and Reed, 1972;Stepp and Reed, 19851, and E1P is involved in the binding of El to the structural core of dihydrolipoamide acetyltransferase (Rahmatullah et al., 1989) and may participate in the second partial reaction of PDC (Roche and Reed, 1972). From chemical modification studies, the amino acid residues which play important role in catalysis have been characterized in pigeon breast muscle El (Khailova et al., 1989) and Escherichia coli E l (Schwartz and Reed, 1970). A more detailed investigation of the pigeon enzyme revealed the presence of a highly reactive cysteine residue, located within or near the active site of El contributing directly to the transformation of reactants to products (Khailova et al., 1985). Flournoy and Frey (1989) suggested a nucleophilic role for cysteine residues in E.
coli El. El requires reducing agents such as dithiothreitol or P-mercaptoethanol for maximal activity (Khailova et al., 1983) and is shown to be sensitive to thiol reagents (Schwartz and Reed, 1970). This evokes the question as to which cysteine residue(s1 is essential for the catalytic activity and what are the optimal conditions for making this identification. Knowledge of the primary structure creates the potential for precise placement of critical cysteine residues.
The primary structures of both a and P subunits of human El were deduced from their cDNA sequences (Ho et al., 1989;Ho and Patel, 1990). There are 13 and 6 cysteine residues in the primary sequences of a and /3 subunits of h u m a n E l , respectively. We undertook a chemical modification study to identify the cysteine residue(s1 that is likely to be involved in the catalysis of bovine El using TPP and pyruvate as protecting agents at the catalytic site. It was possible to specifically label a cysteine residue by radioactive NEM which caused inactivation o f E l . T h e critical cysteinyl residue was identified and was found to be equivalent to Cys-62 of human Eia. This report provides the first localization of cysteine residue at or near the active site of El component of the PDC.

EXPERIMENTAL PROCEDURES
Materials-NEM was purchased from Sigma. Acetonitrile was from Fisher. Tnfluoroacetic acid and V8 protease were from Pierce Chemical Co. [1-l4C1NEM (38.7 mCi/mmol) from Du Pont-New England Nuclear was diluted with unlabeled NEM and stored at -20 "C.
Assay o f E l Activity-Bovine kidney PDC was purified as described previously (Roche and Cate, 1977). The PDC components were resolved 22353 Active Site of Pyruvate Dehydrogenase according to established procedures (Linn et al., 1972;Roche and Cate, 1977). Reconstitution assays were conducted with an excess of E2 and E 3 added together with El and the overall PDC activity was measured (Roche and Reed, 1972). Protein concentrations were determined spectrophotometrically using the Bio-Rad protein assay reagent with bovine serum albumin as a standard. Reaction of El with NEM-The reaction mixture contained 20 m~ potassium phosphate buffer (pH 7.5), 10 nmol NEM and 0.6 nmol E l in a final volume of 0.2 ml. At specified intervals, 5-pl aliquots of the reaction mixture were removed and assayed for activity as described. To investigate the protective effect of compounds against NEM-induced inactivation, El was preincubated with 50 PM TPP, 100 PM pyruvate, or with the combination at 30 "C for 10 min prior to incubation with NEM. Specific Labeling of the Cysteine Residues in El-In a typical experiment, E l (3 nmol) was preincubated with 50 l.m TPP, 100 J~M pyruvate, and 5 mM MgCI, in 0.25 ml of 20 mM phosphate buffer (pH 7.5) for 10 min a t 30 "C followed by addition of NEM (50 nmol) to alkylate the cysteine residues not essential for activity. After 30 min, dithiothreitol (2 mM) was added to stop the alkylation reaction as well as to remove bound reactants (probably HE=TPP) (Khailova et al., 1989). The reaction mixture was then dialyzed against 20 m~ phosphate buffer (pH 7.5) (500 ml x 4) for 6 h to remove dithiothreitol, TPP, and reaction products. The residual specific activity of El was determined and was found to be approximately 80% of the original value. E l (2 nmol) was then incubated with 30 nmol of [I4CINEM (specific radioactivity, 2.5 x lo4 c p d nmol) in a final volume of 0.25 ml for 12 min a t 30 "C. The extent of incorporation of [l4C1NEM through alkylation o f E l was assessed using trichloroacetic acid precipitation procedure at various times of reaction. Four p1 of the reaction mixture was spotted onto filter paper discs (Whatman No. 3") and these were washed three times with 10% (w/v) cold trichloroacetic acid (30 m i d , twice with 95% ethanol, and once with diethyl ether. To determine the protein bound radioactivity, the filter papers were air-dried and counted in 5 ml of scintillation fluid in a Beckman LS-8100 liquid scintillation counter. Aliquots were also withdrawn to measure the E l activity and to identify the labeled polypeptide by SDS-polyacrylamide gel electrophoresis as described below. In studies in which the labeled enzyme was subjected to protease digestion, the alkylation reaction was quenched by addition of 10-fold excess cysteine, followed by extensive dialysis against 50 mM phosphate buffer (pH 7.5).
SDS-Polyacrylamide Gel Electrophoresis-Polyacrylamide gel electrophoresis in the presence of 0.1% SDS with 12% acrylamide was performed as described by Laemmli (1973). Gels were stained with Coomassie Blue R-250, destained, and dried onto Whatman No. 3MM paper with Bio-Rad gel dryer, and the autoradiograph was developed with Kodak XAR-5 x-ray film.
Protease Digestion and Peptide M~pping-['~CINEM-labeled El (1.1 nmol) was digested with V8 protease overnight a t room temperature at a ratio of 10:1 ( E l : V8 protease) in presence of 2 M guanidinium chloride. Approximately 3 pl of trifluoroacetic acid was added to stop the reaction. The separation of peptides generated by VS protease was performed on a Shimadzu high performance liquid chromatography (HPLC) system (model LC-600) with a Synchropack C-4 reverse phase column (25 x 0.46 cm). The column was equilibrated with 0.1% trifluoroacetic acid in triply distilled water, and peptides were eluted with a 0-60% acetonitrile gradient containing 0.1% trifluoroacetic acid developed over a period of 120 min at a flow rate of 0.9 mumin. Fractions were collected using a Shimadzu fraction collector (model SF-2120) and analyzed for radioactive peptidds).
Amino Acid Sequence Analysis-Automated sequence analysis of the 14C-labeled peptide was performed using a gas phase sequencer (Applied Biosystems model 477A). The phenylthiohydantoin-derivative obtained at each cycle of Edman degradation was identified automatically with HPLC (Applied Biosystems model 120 phenylthiohydantoin analyzer) which was attached to the sequencer. The amino acid residue containing the I 4 C was identified by measuring the radioactivity of the cleaved residue after each cycle of degradation.
Circular Dichroism Spectral Analysis-The circular dichroism spectra in the far ultraviolet region (190-260 nm) of both native and modified E l were measured in a 5-600 automatic spectropolarimeter from JASCO (Tokyo). The instrument was calibrated with d-10-camphorsulfonic acid. Spectra were obtained with solutions containing 0.2 mg/ml E l in 20 mM phosphate buffer (pH 7.5) in a cell with an optical path of 1 mm. The ellipticity values were corrected for solvent and background noise using an automated computer program.

RESULTS
Reaction of El with NEM-E1 was inactivated in a timedependent manner when incubated with excess NEM. The loss in activity was pseudo first order at a given NEM level (Fig. 1).

The linear relationship between kapp min-' and [NEM] shows a
simple bimolecular relationship for the reaction between enzyme and reagent (Fig. 1, inset). The slope of the curve yields a second order rate constant of 1.2 mM-l min-l.
Protection of El against Znactivation by NEM-Substrate and cofactor protection against NEM inactivation of El was evaluated and the results are shown in Fig. 2. Neither TPP nor pyruvate alone afforded protection against inactivation of El by NEM; however, TPP (50 p~) plus pyruvate (100 p~) allowed 70430% of El activity to be retained even after a 20-min treatment with 0.2 mM NEM. It is likely that catalytic turnover led to the active site being occupied by the HE=TPP intermediate under these conditions. Circular Dichroism-Circular dichroism spectra (190-260 nm) were collected for both the native and the modified El and were used to calculate the secondary structure of El (results not shown). The data suggested that the secondary structure of E l was not significantly perturbed by NEM modification. Specific Radiolabeling of Essential Cysteine Residues-We have taken advantage of a two-step modification procedure ( T u and Weiner, 1988) to preferentially label a cysteine residue in or near the catalytic site. In the first step, an essential cysteine residue(s) in El was treated with TPP plus pyruvate, prior to reaction of excess NEM with the nonessential cysteine residues followed by quenching with dithiothreitol. The reactants were then removed by dialysis, and the residual catalytic activity was determined to be approximately 80% of the original activity. This modified enzyme was then incubated with [l4C1NEM to specifically label any protected cysteine residue(s). The stoichiometry of incorporation of NEM and residual activities were determined as a function of time (Fig. 3). There was a direct correlation between the fractional loss of El activity and the NEM incorporation into E l . Ninety percent inactivation of El was achieved with the introduction of two mol of [l4C1NEM per mol of E l tetramer, suggesting that either the a or p of E l subunit contains an essential cysteine residue. This was further confirmed by SDS-gel electrophoresis and autoradiography of the NEM-labeled E l (Fig. 3, inset). During a 4-min  NEM. E l was first incubated with unlabeled NEM in the presence of TPP (50 p~) plus pyruvate (100 PM) in order to allow the modification of the nonessential cysteine residues for 30 min. After dialysis, the modified enzyme was incubated with [I4C1NEM in the absence of TPP and pyruvate as described under "Experimental Procedures." Aliquots were taken at indicated time intervals to assay residual activity (0) and radioactivity incorporated in E l (0). Inset, IT-labeled E l was subjected to SDSpolyacrylamide gel electrophoresis followed by autoradiography. E l was labeled with ['TINEM for 4 min (lane I ) and 12 min (lane 2) as described above. The migration of a and p subunits ofEl in the gel are identified.
incubation the radioactivity appeared only in the a subunit, but upon allowing the alkylation reaction to proceed for 12 min, a small amount of radioactivity also appeared in the p subunit. However, the small level of incorporation into the p subunit did not constitute alkylation of even one site per tetramer, and thus it cannot explain the observed loss of activity.
Zdentification of Cysteine Residues-To determine the location of 14C-labeled NEM within the protein, radiolabeled E l was digested with V8 protease. Since it was noted earlier that the p subunit of El was resistant to trypsin digestion (Barrera et al., 1972), we have chosen V8 protease to digest El because of its high specificity for cleavage of peptide bonds on carboxyl side of glutamic and aspartic acid residues. The resulting peptide mixture was subjected to analysis by reverse phase HPLC. Fig. 4 shows the separation profile for the peptides and the amount of radioactivity associated with each fraction. Radioactivity identified in A (indicated by plus sign) corresponded to fraction 62 in Fig. 4B. No radioactivity was detected in the fractions beyond fraction 80 in Fig. 4B (results not shown).
The NEM-labeled peptide was further analyzed by Edman degradation sequencing procedures along with measuring the radioactivity released following each cycle of degradation (Fig.  5). Radioactivity appeared only in the 16th cycle which corresponded to cysteine. The specificity of this modification is evident because the second cysteinyl residue (corresponding to cycle 19 in Fig. 5) was not protected by TPP plus pyruvate and hence did not contain radioactivity when El was deprotected and labeled with [14C]NEM. Comparison of the observed amino acid sequence (LKADQLYKQKIIRGF*CHLCD) of bovine Ela peptide with the deduced amino acid sequence of human Ela established that the cysteine equivalent to the cysteine a t position 62 of human Ela reacted with NEM following deprotection of the active site. The sequence obtained also shows that there is complete identity between bovine and human proteins in this region. The sequence obtained spans amino acids 47-66 of human Ela (Ho et al., 1989).

DISCUSSION
In the present work, the involvement of cysteine residues in the catalytic function of bovine El was indicated by the inactivation of the enzyme by the thiol-specific reagent NEM. Kinetic analysis of the reaction of NEM supports the conclusion that modification of only 1 cysteine/single subunit resulted in loss of activity. Protection studies with TPP plus pyruvate indicated the involvement of a unique thiol in maintaining the El activity. Since TPP plus pyruvate protects the enzyme from NEM inactivation, it is suggested that the essential cysteine is protected by the HE=TPP intermediate. Following deprotection of the active site, 2 cysteine residues/tetramer of El are found to be selectively alkylated by [14CJNEM. This result further supports the presence of two active centers per E l tetramer and that each of these centers has equivalent catalytic efficiency (Khailova and Korochkina, 1982).
The sequence of the peptide generated from bovine El containing the labeled cysteine residue is LKADQLYKQKIIRGF*-CHLCD. The complete identity between human and bovine Ela sequences establishes that the modified cysteine residue corresponds to position 62 in the human Ela sequence. The fact that only one cysteine was labeled with [I4CJNEM demonstrates a high degree of specificity was achieved. The relatively high rate of enzyme inactivation indicates a high reactivity of Cys-62, suggesting a unique role of Cys-62 in the reaction catalyzed by El. This cysteinyl residue is conserved in the aligned amino acid sequences of Ela from the PDC of rat (Matuda et al., 1991), pig (Urata et al., 19911, human (Ho et al., 1989), and yeast (Behal et al., 1989). Among El that are a& tetramers, the only known exception to this pattern comes from Ela of Bacillus stearothermophilus (Hawkins et al., 1990) which contains tyrosine at the aligned position. We suggest that this reactive cysteine has an essential role in catalysis. The possibility of a cysteine at the active site was first indicated by spectroscopic data for an acyl intermediate formed by pigeon breast El (Khailova et al., 1985). According to the proposed mechanism, the interaction of holo-El with pyruvate leads to an intramolecular oxidative transfer of the substrate bound to the protein moiety with the formation of the acetyl-thio enzyme. Frey et al. (1989)   In summary, we have identified the essential cysteinyl residue in or near the active site of bovine El, The identification of the essential cysteine residue represents the first of several other amino acid residues implicated in El catalysis. The support for the role of this cysteinyl residue in El catalysis will come from future studies involving site-directed mutagenesis and overexpression of mutant mammalian El.