Purification and Characterization of 2-Methyl-branched Chain Acyl Coenzyme A Dehydrogenase, an Enzyme Involved in the Isoleucine and Valine Metabolism, from Rat Liver Mitochondria*

2-Methyl-branched chain acyl-CoA dehydrogenase was purified to homogeneity from rat liver mitochon- dria. The native molecular weight of the enzyme was estimated to be 170,000 by gel filtration. On sodium dodecyl sulfate-polyacrylamide gel electrophoresis both with and without 2-mercaptoethanol, the enzyme showed a single protein band with M, = 41,500, sug-gesting that this enzyme is composed of four subunits of equal size. Its isoelectric point was 5.50 2 0.2, and A?&, nm was 12.5. This enzyme contained protein-bound FAD. The purified enzyme dehydrogenated S-2-methyl- butyryl-CoA and isobutyryl-CoA with equal activity. The activities with each of these compounds were co- purified throughout the entire purification procedure. This enzyme also dehydrogenated R-2-methylbutyryl- CoA, but the specific activity was considerably lower (22%) than that for the S-enantiomer. The enzyme did not dehydrogenate other acyl-CoAs, including isoval-eryl-CoA, propionyl-CoA, butyryl-CoA, octanoyl-CoA, and palmitoyl-CoA, at any significant rate. Apparent K,,, and V,,, values for S-2-methylbutyryl-CoA were 20 p~ and 2.2 pmol min” mg”, respectively, while those for isobutyryl-CoA were 89 p~ and 2.0 pmol min” mg” using as an out using the purified 2-methyl-branched chain acyl-CoA dehydrogenase preparation as the antigen and the four individual antibodies to other acyl-CoA dehydrogenases mentioned above. of

the /+oxidation cycle. These are butyryl-CoA-(EC 1.3.99.2), general acyl-CoA-(EC 1.3.99. 3), and long chain acyl-CoA dehydrogenases; they are most active with butyryl-CoA, octanoyl-CoA, and palmitoyl-CoA, respectively, as substrate. These enzymes are localized in the mitochondria of various tissues in mammals. However, the enzymes which dehydrogenate branched chain acyl-CoAs such as isovaleryl-CoA, 2methylbutyryl-CoA, and isobutyryl-CoA have not been extensively studied until recently. These branched chain acyl-CoAs are produced as intermediates in the metabolism of the branched chain amino acids, leucine, isoleucine, and valine, respectively. Previously, butyryl-CoA dehydrogenase (short chain acyl-CoA dehydrogenase) had been thought to catalyze the dehydrogenation of all short branched chain acyl-CoAs (1,2). However, several years ago, our biochemical observations on patients with isovaleric acidemia, an inborn error of leucine metabolism, suggested the existence of a dehydrogenase which is specific for isovaleryl-CoA (3)(4)(5). Using rat liver mitochondria as an enzyme source, we subsequently demonstrated that a specific isovaleryl-CoA dehydrogenase indeed exists and that it is distinct from butyryl-CoA dehydrogenase (6)(7)(8). Furthermore, we have recently reported the purification and characterization of isovaleryl-CoA dehydrogenase (8,9). This enzyme is biochemically and immunologically distinct from butyryl-CoA dehydrogenase (9).
In the course of these studies, we demonstrated for the first time that 2-methylbutyryl-CoA and isobutyryl-CoA were not dehydrogenated by either isovaleryl-CoA dehydrogenase or butyryl-CoA dehydrogenase. Instead, these two 2-methyl-substituted acyl-CoAs were dehydrogenated by an enzyme which was distinct from the other four acyl-CoA dehydrogenases (8). We have previously reported partial purification of this enzyme from rat liver mitochondria by a sequence of DEAE-Sephadex and hydroxyapatite column chromatographies and isoelectric focusing, and designated it 2-methyl-branched chain acyl-CoA dehydrogenase (8). In the present paper: we report the purification to homogeneity of 2-methyl-branched chain acyl-CoA dehydrogenase from rat liver mitochondria. We also describe here the molecular characteristics, kinetic parameters, requirement for ETF,' susceptibility to various types of inhibitors, and immunochemical properties of this enzyme.

Methods
Synthesis of Coenzyme A Thioesters of Sand R-2-Methylbutyric Acids, and Isoualeric Acid-S-and R-2-methylbutyric acids were prepared from L-isoleucine and L-allo-isoleucine, respectively, according to the method described previously (8). The purities of both carboxylic acids were 99% as determined by gas chromatographic analysis and mass spectrometry.
Coenzyme A thioesters of these carboxylic acids were synthesized by the mixed anhydride synthesis (10). These acyl-CoAs were purified by paper chromatography to remove unreacted coenzyme A and other reagents, using ethanol, 0.1 M potassium acetate, pH 4.5 (l:l), as a developing solvent. Coenzyme A thioester of isovaleric acid was also synthesized by the same method because the commercial isovaleryl-CoA from P-L Biochemicals contained 15% 2-methylbutyryl-CoA as determined by gas chromatography and mass spectrometry of the acyl group.
Assay of Acyl-CoA Dehydrogenases-Assays for the various acyl-CoA dehydrogenase activities were performed spectrophotometrically using PMS and DCIP as intermediate and terminal electron acceptors, respectively, and an appropriate acyl-CoA as substrate according to the method described previously (8,9). The incubation medium was composed of 0.1 M potassium phosphate buffer (pH 8.0), either 1.5 mM' or 3 mM3 PMS, 0.048 mM DCIP, 0.1 mM FAD, and 0.1 mM acyl-CoA unless otherwise mentioned. The final volume was 1 rnl.
The enzyme reaction was carried out a t 32 "C and the reaction was started with the addition of substrate. Bleaching of DCIP was followed at 600 nm for at least 2 min using a Beckman model 3600 double-beam spectrophotometer. Enzyme activity was cxpressed as micromoles or nanomoles of DCIP reduced/ml of enzyme solution/ min. The extinction coefficient of DCIP (21 mM" cm") at 600 nm was used as the basis for computation of the amount of DCIP reduced.
The suitability of E T F for supporting the purified 2-methylbranched chain acyl-CoA dehydrogenase activity was tested using the E T F preparation purified from rat liver mitochondria as described previously (8, 9). An appropriate amount of E T F (100-200 pg) was added to the above assay mixture, replacing PMS. The enzyme activity was again assayed for monitoring DCIP bleaching.
Preparation of Rat Liuer Mitochondria and the First Four Steps of Purification-In a typical experiment, 100 adult male Charles River CD rats, weighing 200 to 280 g, were killed by decapitation and the liver mitochondria were isolated by the method of de Duve et al. (11). The first four steps of the purification are the same as those utilized to isolate crude 8-methyl-branched chain acyl-CoA dehydrogenase as described previously (8). The mitochondria (52.0 g of protein) were sonicated and centrifuged a t 105,000 x g for 60 min (step 1). The supernatant was fractionated by ammonium sulfate (40-80%) precipitition (step 2).
The areciaitate was redissolved in 10 mM KPO, buffer. DH 8.0, containing 0:5 mM EDTA and dialyzed against the same buffer. The dialyzed solution (14.4 g of' protein) was applied to four DEAE-Sephadex A-50 columns (4.6 X 20 cm) equilibrated with 10 mM KPO, buffer, pH 8.0, 0.5 mM EDTA and the adsorbed fraction on each column was eluted with a linear NaCl gradient (0-0.6 M ) in 2 liters of the same buffer (step 3). The fractionation on this column was essentially identical with that described in our previous paper (8). Fractions from the column chromatography were separated into two which dehydrogenated isobutyryl-CoA, S-2-methylbutyryl-CoA, n-major preparations. Preparation B (tubes 62-85) contained activities butyryl-CoA, n-octanoyl-CoA, and palmitoyl-CoA.
Preparation B (1500 mg of protein) was applied to two hydroxyapatite columns (4.6 X 18 cm) equilibrated with 10 mM KPO4 (pH 7.5). The adsorbed fraction on each column was eluted with a linear gradient of 2 liters of KPO, buffer, p H 7.5 (0.01-0.33 M) (step 4). This column pattern was also similar to that described in our previous paper (8). Fractions from hydroxyapatite column chromatography ' 1.6 mM PMS was used for assay of 2-methyl-branched chain acyl-CoA-, isovaleryl-CoA-, and butyryl-CoA dehydrogenase activities (8).
were pooled into four major preparations (preparations D, E, F, and G). Preparation E (tubes 63-90) contained most of the S-2-methylbutyryl-CoA-and isobutyryl-CoA-dehydrogenating activities, and it was used for further purification.
Matrex Gel Blue A Chromatography (Step 5)"Preparation E (108 mg of protein in 25 ml) was applied to a Matrex Gel Blue A column (1.5 X 8 cm) equilibrated with 10 mM KPO, buffer, pH 8.0, containing 10% glycerol and 0.5 mM EDTA. The column was washed with the same buffer until the absorbance at 280 nm returned to the base-line, and then was washed with the same buffer containing 0.35 M NaC1. The adsorbed proteins were eluted with a linear gradient of 140 ml of the buffer containing 0.35 M NaCl and the same buffer containing 0.4 M NaCl and 7 mM FAD. Elution was done a t a flow rate of 0.42 ml/min.
Agarose-Hexane-CoA Chromatography (Step 6)"The sample solution (5 mg of protein in 7 ml) from step 5 was applied to an agarosehexane-CoA (type I) column (1 X 7 cm) equilibrated with 10 mM K P 0 4 buffer (pH 8.0) containing 10% glycerol and 0.5 mM EDTA. The column was washed with the same buffer until the absorbance a t 280 nm returned to the base-line. The adsorbed proteins were eluted with an 80-ml linear gradient from 0-0.5 M NaCl in equilibration buffer (pH 8.0). Elution was done a t a flow rate of 0.42 ml/min.
Purification of Electron Transfer Flauoprotein-ETF was purified to homogeneity from rat liver mitochondria according to the method described in our previous papers (8,9). The final preparation gave two protein hands on SDS-PAGE in both the absence and presence of 2-mercaptoethanol. These subunit M, = 30,000 and 35,500, in close agreement with those reported by Furuta et af (12). The ratios of absorbance at the maxima, 270:375:435:460 nm, were 6.9:0.04:1.0:0.8, as described previously (9).
Protein Determinations-Protein concentrations were determined by the method of Lowry et al. (13) unless otherwise indicated. Because the fractions from the Matrex Gel Blue A column contained a large amount of FAD, protein concentrations were determined by the method of Bradford (14). Determination of the protein concentrations of the pure enzyme preparation was done by the microbiuret method according to Itzhaki and Gill (15). Bovine serum albumin was used as standard in all of the assay methods.
Identification of Reaction Product-The reaction conditions for this purpose were similar to those utilized in the dye reduction assay except for the following modifications: the reaction mixture contained 100 mM phosphate buffer (pH 8.0), 3 mM PMS, 0.1 mM FAD, and 0.4 mM isobutyryl-CoA or 0.2 mM Sor R-2-methylbutyryl-CoA. No DCIP was added. The total volume was 0.5 ml. The mixture was incubated a t 37 "C for 2, 5, 10, and 20 min. After termination of the reaction by the addition of 0.05 ml of 3 M perchloric acid, the reaction product was hydrolyzed and steam-distilled according to the method previously described (9). The evaporated residues of the alkalinized distillate were redissolved in 50 p1 of 10% aqueous formic acid. One pl of the solution was injected into a Hewlett-Packard 5840A gas chromatograph equipped with flame ionization detectors and a 18850A terminal/data system. A coiled glass column (2 mm x 1.8 m) packed with SP-1200 (Supelco: Bellafonte, PA) was utilized for analysis. The temperature of the column oven was 110 "C, and nitrogen gas was used as a carrier gas. The recovery of acyl groups throughout these procedures was approximately 75%.
For mass spectral identification, the evaporated residues of the alkalinized distillate were dissolved in 1 ml of H20, acidified with 6 N HCl, and extracted four times with 1 ml of diethyl ether. The ether extracts were combined, dried over anhydrous MgS04, carefully concentrated to 0.5 ml on ice under a gentle nitrogen stream, and methylated with gaseous diazomethane according to the method previously described (16). A Finnigan 4510 automated gas chromatography/mass spectrometer/computer was used for analysis with electron impact ionization mode. A 10% OV-17 column (1.8 m X 2 mm) was used as the inset gas chromatographic column. The initial oven temperature was 40 "C, and it was raised at a rate of 6 "C/min. The ionizing voltage was 70 eV.
Electrophoretic Procedures-PAGE without SDS was performed in a 5.0% gel using Tris-glycine buffer (pH 8.9) according to the method  Summary of purification steps for 2-methyl-branched chain acyl-CoA dehydrogenase from rat liver mitochondria Rat liver mitochondria (52 g of protein) were fractionated to purify the enzyme. Both isobutyryl-CoA (iC4CoA)and S-2-methylbutyryl-CoA (S-2-MeC4CoA)-dehydrogenating activities were determined by the dye reduction assay. The activity of each preparation was assayed after it was concentrated and dialyzed.
A c t i v i t i e s could not be determined due t o i n t e r f e r e n c e of t h e dye-reduction assay by non-specific r e d u c t a n t s .
of Davis (17). SDS-PAGE was carried out in 7.5% gels according to the method of Weber and Osborn (18). Gels were stained with 0.25% Coomassie brilliant blue and were destained in 7.5% acetic acid and 5% methanol solution. The proteins used as calibration standards Amino Acid Analysis-The amino acid composition of the purified enzyme preparation was determined after HC1 hydrolysis by the method of Stein and Moore (19) using a Beckman 121M amino acid analyzer. Total half-cystine content was determined as cysteic acid after performic acid oxidation according to the method of Hirs (20). Tryptophan content was determined after hydrolysis with 3 N mercaptoethanesulfonic acid according to the method of Penke et al. (21).
Immunoreactions-Antibody raised against the purified isovaleryl-CoA dehydrogenase (9) and those raised against short chain acyl-CoA-, medium chain acyl-CoA-(22), and long chain acyl-CoA dehydrogenases4 purified from rat liver mitochondria were used in immunoreactivity experiments. From the immune titration curve for purified short chain acyl-CoA-, medium chain acyl-CoA-, or long chain acyl-CoA dehydrogenase, the corresponding antibody (100 pl) precipitated 18 p g of the pure short chain acyl-CoA-, 14 pg of the pure medium chain acyl-CoA-, or 7 pg of the pure long chain acyl-CoA dehydrogenases under the conditions described in our previous paper (9). Immunotitrations and Ouchterlony double diffusion experiments were carried out using the purified 2-methyl-branched chain acyl-CoA dehydrogenase preparation as the antigen and the four individual antibodies to other acyl-CoA dehydrogenases mentioned above.

RESULTS
Purification of 2-Methyl-branched Chain Acyl-CoA Dehydrogenase-The entire purification procedure is summarized in Table I. Technical details are described under "Experimental Procedures." The first four steps of the purification are the same as previously described (8). Rat liver mitochondria were solubilized by sonication (step 1) and the supernatant Details of purification of short chain acyl-CoA-, medium chain acyl-CoA-, and long chain acyl-CoA dehydrogenases will be published elsewhere.
was fractionated by the sequence of ammonium sulfate precipitation (40-80%) (step 2), DEAE-Sephadex A-50 (step 3), and hydroxyapatite (step 4) chromatography. Fractionation patterns at steps 3 and 4 are shown in our previous publication (8). Preparation E, which was obtained from hydroxyapatite chromatography, contained other acyl-CoA dehydrogenase activities in addition to those dehydrogenating S-2-methylbutyryl-CoA and isobutyryl-CoA. In particular, butyryl-CoA dehydrogenase activity was very high while isovaleryl-CoA dehydrogenase activity was undetectable.
Relative specific activities using S-2-methylbutyryl-CoA, isobutyryl-CoA, isovaleryl-CoA, n-butyryl-CoA, n-octanoyl, and palmitoyl-CoA as substrates in preparation E were 1.0, 0.95, 0, 9.4, 0.68, and 1.5, respectively. The octanoyl-CoA-dehydrogenating activity in preparation E was due to the co-existing long chain acyl-CoA dehydrogenase; medium chain acyl-CoA dehydrogenase was not present in this preparation. In order to separate 2methyl-branched chain acyl-CoA dehydrogenase from other acyl-CoA dehydrogenases, preparation E (108 mg of protein in 25 ml) was applied to a Matrex Gel Blue A column (step 5). When elution was done with a linear FAD gradient (0-7 mM), S-2-methylbutyryl-CoA-and isobutyryl-CoA-dehydrogenating activities were co-eluted as a sharp single peak at 1-2 mM FAD, while only a small amount of n-butyryl-CoAdehydrogenating activity was eluted in this region; no significant activities for n-octanoyl-CoA and palmitoyl-CoA were detectable. Most of the butyryl-CoA-and long chain acyl-CoA dehydrogenases were eluted as very broad peaks a t FAD concentrations higher than 3 mM. When the column was further eluted with 10 mM KPO, buffer (pH 8.0) containing 0.8 M NaCl and 3 mM FAD, the butyryl-CoA-and long chain acyl-CoA dehydrogenases which still remained in the column were eluted as a sharp peak (Fig. 1). Fractions 18 to 35, containing both S-2-methylbutyryl-CoA-and isobutyryl-CoA-dehydrogenating activities, were pooled together. Relative specific activities with S-2-methylbutyryl-CoA, isobutyryl-CoA, n-butyryl-CoA, n-octanoyl-CoA, and palmitoyl- CoA in the pooled fraction were 1.0, 0.92, 0.23, 0, and 0, respectively. After concentration, the sample preparation (5 mg of protein in 7 ml) was applied to an agarose-hexane-CoA column (step 6).
When the agarose-hexane-CoA column was eluted with a linear NaCl gradient (0-0.4 M), both S-2-methylbutyryl-CoAand isobutyryl-CoA-dehydrogenating activities were again coeluted as a single peak in fractions 25 to 45; these activities were well separated from butyryl-CoA dehydrogenase (Fig. 2). After concentration, the sample solution (0.8 mg of protein in 1.5 ml) from step 6 was applied to a Bio-Gel A-0.5m column (step 7). As shown in Fig. 3, both S-2-methylbutyryl-CoAand isobutyryl-CoA-dehydrogenating activities were still coeluted as a single peak in fractions 50 to 70. These fractions were combined and then concentrated; this sample represents the final preparation. The specific activities of the final preparation for S-2-methylbutyryl-CoA and isobutyryl-CoA were 2.4 and 2.3 Fmol min" mg", respectively. The overall yield of the enzyme was 2.5%. An identical value was obtained when the recoveries of either of the activities were utilized for this computation. The purified enzyme was stable for at least 30 days when stored in 50% glycerol at -20 "C.
Purity, Molecular Weight, and Subunit Structure-The purity of the final 2-methyl-branched chain acyl-CoA dehydrogenase preparation was determined by PAGE with and without SDS. When boiled, the purified enzyme gave a single protein band with M , = 41,500 in SDS-PAGE in both the absence and presence of 2-mercaptoethanol (Fig. 4, B and C).
When the sample was subjected to SDS-PAGE without boiling, the enzyme gave an apparent single protein band with M, = 85,000, both in the presence and absence of 2-mercaptoethanol, indicating a dimeric form of the protein (Fig. 4A).
In PAGE without SDS, the purified enzyme also gave a single protein band in 5.0% gel ( R p value of 0.55) (Fig. 4E). The native molecular weight of the enzyme was estimated to be 170,000 by gel filtration on Bio-Gel A-0.5m chromatography (Fig. 3 ) . These data indicated that the enzyme is composed of four subunits of identical size. Isoelectric Point-The isoelectric point of the purified enzyme was 5.5 ? 0.2 as determined by sucrose discontinuous isoelectric focusing (LKB) using a 1:4 mixture of pH 3.5-10 and 4-6 ampholytes. The pH of each fraction was determined using a pH meter at 0 "C. This PI value was almost identical with our previous results using a crude 2-methyl-branched chain acyl-CoA dehydrogenase preparation from the hydroxyapatite chromatography step (8). Chromatofocusing (Pharmacia) was also carried out using PBE 94 and Polybuffer 74.
This enzyme was eluted from the chromatofocusing at pH 5.1 t 0.2.
Amino Acid Composition-The amino acid composition is shown in Table 11. The number of cysteine residues was estimated to be 5/subunit. The subunit molecular weight of the enzyme was calculated to be 42,400 from the amino acid composition, in close agreement with the value (41,500) determined by SDS-PAGE. The specific volume of the enzyme was 0.72 as computed from the amino acid composition.
Absorption and Fluorescence Spectra, and Prosthetic Group-The visible and ultraviolet absorption spectrum of the purified enzyme is shown in Fig. 5. The major absorption maxima were found a t 275, 340, and 435 nm. The ratios of absorbance at 275, 340, and 435 nm were 10.3:1.3:1.0. The fluorescence emission spectrum of the purified enzyme excited at 450 nm showed a peak at 520 nm as in the case of authentic FAD; its intensity was 28% of the equivalent amount of authentic FAD, indicating quenching due to FAD-protein interaction. The excitation spectrum of the enzyme as moni-   Table 111. The enzyme exhibited its highest activity with either S-2-methylbutyryl-CoA or isobutyryl-CoA as a substrate. When R-2-methylbutyryl-CoA was used as a substrate, its activity was 22% of that observed with S-2-methylbutyryl-CoA. In contrast, the activity was extremely low or not detectable when the following compounds were used as a substrate: isovaleryl-CoA, propionyl-CoA, n-butyryl-CoA, nvaleryl-CoA, n-hexanoyl-CoA, n-octanoyl-CoA, palmitoyl-CoA, glutaryl-CoA, and sarcosine. The apparent V,,,,, and K,,, values for isobutyryl-CoA were 2.0 pmol min" mg" and 89 PM, respectively, and the apparent V,,, and K , values for S-2-methylbutyryl-CoA were 2.2 pmol min" mg" and 20 p~, respectively (Tables 111 and VI).
The reaction products were identified by gas chromatographic analysis as shown in Fig. 6. Under the conditions were not separated. The reaction product of the purified enzyme with isobutyryl-CoA was identified as methacrylyl-CoA by detection of its hydrolysis product, methacrylic acid (Fig. 60). The reaction product of the same enzyme with S-2methylbutyryl-CoA was identified as tiglyl-CoA (Fig. 6b). When R-2-methylbutyryl-CoA was used as a substrate, a compound which had a considerably shorter retention time (7.6 min) than that of tiglic acid (9.8 min) was detected (Fig.  6c). After conversion to a methyl ester, this compound was identified as ethylacrylic acid using mass spectroscopy by the identity of its mass spectrum to that of t,he authentic standard (23).
The reaction rates as assessed by measuring the amount of product in the first 5 min using gas chromatography were 12.  min/0.5 ml of reaction mixture for both isobutyryl-CoA and S-2-methylbutyryl-CoA) obtained by the dye reduction assay using the same amount of the enzyme preparation. These results verify the validity of the dye reduction assay for acyl-CoA dehydrogenase.
The time course of the reaction was studied by gas chromatographic analysis of the products to estimate an apparent K , using isobutyryl-CoA and S-2-methylbutyryl-CoA. The reaction mixture was incubated a t 37 "C for 2, 5, 10, and 20 min. The reaction products from isobutyryl-CoA and S-2methylbutyryl-CoA were produced linearly with time for at least 5 min, but the reaction rate diminished after this point. The decrease of isobutyryl-CoA and the increase of methacrylyl-CoA both plateaued, revealing an equilibrium after 20 min. A similar time course was observed with S-2-methylbutyryl-CoA as a substrate. The apparent K , was determined as the ratio of the product concentration to the substrate concentration at the equilibrium. The apparent K , for these two substrates differed greatly: K , for isobutyryl-CoA was 1.0 while that for S-2-methylbutyryl-CoA was 4.0.
The inhibitory effects of various acyl-CoAs on 2-methylbranched chain acyl-CoA dehydrogenase activity were investigated by using isobutyryl-CoA or S-2-methylbutyryl-CoA as substrates. The results are summarized in Table IV. No difference between the two substrates was observed. The most notable finding was that tiglyl-CoA strongly inhibited both the isobutyryl-CoA-and S-2-methylbutyryl-CoA-dehydro-

TABLE Ill
Substrate specificity of the purified 2-methyl-branched chain acyl-CoA dehydrogenaye The dehydrogenating activity was determined by the dye reduction assay in the presence of 100 p M FAD. The substrate concentration was 100 p~ except for isobutyryl-CoA which was used a t 200 pM.
ETF as Electron Acceptor-The purified E T F preparation equal specific activity with S-2-methylbutyryl-CoA. In these experiments, we could not determine an apparent K, for ETF, because saturation was not observed with the amounts of E T F employed.
Effects of Various Inhibitors-The effects of various inhibitors on both S-2-methylbutyryl-CoA-and isobutyryl-CoAdehydrogenating activities of the purified enzyme are shown in Table V. Essentially identical inhibitory effects on these two activities were observed using various inhibitors. These dehydrogenating activities were severely inhibited by sulfhy-

TABLE V The effects of various compounds on 2-methyl-branched chain acyl-
CoA dehydrogenase activity The purified enzyme (10 pg of protein) was preincubated with each compound at the concentration indicated for 5 min at 32 "C. The dehydrogenating activity was determined by the dye reduction assay in the presence of 100 p~ FAD. 2-meC4CoA, 2-methylbutyryl-CoA iC4CoA, isobutyryl-CoA.
Residual activity with Preincubated with inhibited the enzyme activity by 50%, but iodoacetamide (2 mM) did not significantly inhibit the activity. The enzyme activity was completely inhibited by heavy metals such as Hg'+, Cu'+, and Ag' (0.1 mM each) which are known to affect thiol groups in proteins. Ca", Zn2+, Pb", and Fe3+ did not inhibit this enzyme activity at all.
Immunological Properties-In both immunoprecipitation and Ouchterlony double diffusion experiments, the purified 2-methyl-branched chain acyl-CoA dehydrogenase (10 yg of protein; specific activity = 2.2 pmol min" mg" for S-2methylbutyryl-CoA) was reacted with individual antiserum raised against isovaleryl-CoA-, short chain acyl-CoA-, medium chain acyl-CoA-, or long chain acyl-CoA dehydrogenase, respectively. The 2-methyl-branched chain acyl-CoA dehydrogenase did not exhibit any cross-reaction with these four antibodies in Ouchterlony double diffusion experiments; its enzyme activity was not precipitated by the four antibodies with either isobutyryl-CoA or S-2-methylbutyryl-CoA as a substrate (data not shown). These results indicate that 2methyl-branched chain acyl-CoA dehydrogenase is immunologically distinct from the four other acyl-CoA dehydrogenases and that the final preparation was not contaminated by the other enzymes.

DISCUSSION
In the present study, we purified 2-methyl-branched chain acyl-CoA dehydrogenase from rat liver mitochondria to homogeneity in seven steps including affinity chromatographies with Matrex Gel Blue A and agarose-hexane-CoA, which were used at the fifth and sixth steps, respectively. The activity to dehydrogenate S-2-methylbutyryl-CoA and that to dehydrogenate isobutyryl-CoA were co-purified throughout the entire seven steps of purification (Table I). The specific activity of the final preparation was enriched 90-fold over that of the preparation obtained after the DEAE-Sephadex step. The activities in the crude preparations such as mitochondrial sonic supernatant and (NH4),S04 precipitates could not be accurately measured due to interference by nonspecific reductants. The tritium release assay which is free of such interference was not available for these activities. In our previous study (9), the specific activity of isovaleryl-CoA dehydrogen-ase preparation after the DEAE-Sephadex stage was enriched approximately 20 times over that of the mitochondrial sonic supernatants as measured by the tritium release assay, and a similar degree of purification can be expected for 2-methylbranched chain acyl-CoA dehydrogenase at these steps. Thus, the final preparation of 2-methyl-branched chain acyl-CoA dehydrogenase is probably enriched nearly 1800-fold over that in the mitochondrial sonic supernatant.
The fact that isobutyryl-CoA-and S-2-methylbutyryl-CoAdehydrogenating activities co-purified throughout all steps of purification suggests that a single enzyme catalyzes the dehydrogenation of both isobutyryl-CoA and S-2-methylbutyryl-CoA. We have also shown in this report that both isobutyryl-CoA-and S-2-methylbutyryl-CoA-dehydrogenating activities of this enzyme were competitively inhibited by tiglyl-CoA, the product from S-2-methylbutyryl-CoA (Tables  IV and VI The purified enzyme exhibited a high substrate specificity (Table 111). It dehydrogenated isobutyryl-CoA and S-2-methylbutyryl-CoA with high specific activities. The rates for these two substrates were approximately equal. This enzyme also dehydrogenated the R-enantiomer of 2-methylbutyryl-CoA, but the rate of the reaction with this substrate was only 22% of that with the S-enantiomer. The reaction products from isobutyryl-CoA and S-2-and R-2-methylbutyryl-CoA by this enzyme were identified as methacryl-CoA, tiglyl-CoA, and ethylacrylyl-CoA, respectively, by the detection of their hydrolysis products (Fig. 6). In contrast, this enzyme did not dehydrogenate any other straight chain acyl-CoAs, or a branched one, isovaleryl-CoA, at any significant reaction rate. This substrate specificity is very narrowly limited to those substrates with a methyl substitution at the a-carbon. Among the substrates with an a-methyl substitution, S-2-methylbutyryl-CoA and isobutyryl-CoA were dehydrogenated with high efficiencies while R-2-methylbutyryl-CoA was dehydrogenated at a considerably slower rate. These results on the substrate specificity and the identification of the products suggest that the reaction of this enzyme proceeds by elimination of one hydrogen each, respectively, from the a-methine group and the /%methylene (methyl) group taking the a position as illustrated in Fig. 7. Whether the substitution on the a position is a methyl or an ethyl does not significantly affect the rate of reaction. In contrast, when the substitution on the c position is an ethyl, the rate of reaction was significantly slower than that when it was a methyl. This suggests that the size of the substitution directed to the c position, although it does not participate in the dehydrogenase reaction, is important in defining the fitness of the substrate to the conformation of this enzyme at the active site.
The product inhibition by tiglyl-CoA was also specific. 2-Methyl-branched chain acyl-CoA dehydrogenase activity was inhibited by neither 3-methylcrotonyl-CoA nor isovaleryl-CoA. However, it was moderately inhibited by n-butyryl-CoA, n-valeryl-CoA, or crotonyl-CoA. This suggests that the enzyme can bind n-butyryl-CoA, valeryl-CoA, or crotonyl-CoA as substrate analogs, although the enzyme does not dehydrogenate them at a significant rate (Table 111).
The substrate specificity of 2-methyl-branched chain acyl-CoA dehydrogenase and the inhibition -of this enzyme by tiglyl-CoA are of particular interest in view of the regulation of the branched chain amino acid metabolism. The three -" Antibody to IVD no-cross reaction positive reaction no-cross reaction Antibody to SC-AD no-cross reaction no-cross reaction positive reaction a Described in detail elsewhere ( 9 ) .
These values were determined spectrophotometrically on the purified enzyme preparations. FAD might have been partially lost in the purification procedures. The FAD content per subunit of both enzymes in the native form is estimated to be 1 mol per subunit because activities of the final preparations were enhaned 1.5-2.5 times by the addition of exogenous FAD.
Emission spectra were monitored with exitation of 450nm, and excitation spectra were taken with emission at 530nm.
e The activity of isovaleryl-CoA dehydrogenase was not inhibited by tiglyl-CoA.
The activity of short chain acyl-CoA dehydrogenase was not inhibited by tiglyl-CoA. branched chain amino acids, leucine, isoleucine, and valine, are first transaminated to the corresponding 2-oxo acids. These three 2-oxo analogs are then oxidatively decarboxylated to isovaleryl-CoA, S-2-methylbutyryl-CoA, and isobutyryl-CoA, respectively, by a single common enzyme, branched chain 2-oxo acid dehydrogenase. This enzyme is subject to inhibition by any of the three branched chain acyl-CoAs (24). Thus, the branched chain 2-oxo acid dehydrogenase step has been considered to be the site for metabolic regulation which is common for the three branched chain amino acids. We have shown in the previous report that isovaleryl-CoA is specifically dehydrogenated by isovaleryl-CoA dehydrogenase and this reaction is specifically inhibited by 3-methylcrotonyl-CoA (9). Isobutyryl-CoA and S-2-methylbutyryl-CoA were dehydrogenated commonly by 2-methyl-branched chain acyl-CoA dehydrogenase, which is subject to the inhibition by tiglyl-CoA as shown in the present paper. Thus, after the 2oxo acid decarboxylation, the metabolism of isoleucine and valine may be commonly regulated while the leucine metabolism is independently controlled.
The ability of this enzyme to dehydrogenate R-2-methylbutyryl-CoA may be of more than a theoretical interest. It has previously been shown that when experimental animals were given RS-2-methylbutyric acid labeled with stable isotopes, they excreted labeled 2-ethyl-3-hydroxyproprionic acid (2-ethylhydracrylic acid) into urine (25-26). The results from detailed mass spectroscopic analyses of the urinary metabolites indicated that ethylhydracrylic acid was further oxidized by a pathway (R-pathway) which is analogous to the valine pathway (25). It was hypothesized that R-2-methylbutyryl-CoA was dehydrogenated on the shorter acyl chain producing 2-ethylacrylyl-CoA which was then hydrated to 2-ethylhydracrylyl-CoA while S-2-methylbutyryl-CoA was dehydrogenated on the longer chain producing tiglyl-CoA (Fig. 7). The data presented in this report represent the first scientific evidence that the two enantiomers of 2-methylbutyryl-CoA are, in fact, stereospecifically dehydrogenated.
The properties of 2-methyl-branched chain acyl-CoA dehydrogenase are summarized in Table VI, along with those of isovaleryl-CoA-and short chain acyl-CoA dehydrogenases. These three enzymes are similar in molecular size, prosthetic group, and basic mode of enzyme reaction, but they differ significantly from each other in catalytic and immunological properties. The native molecular weight of 2-methyl-branched chain acyl-CoA dehydrogenase is 170,000 as determined by gel filtration (Fig. 3). Its molecular weight is slightly larger than that of short chain acyl-CoA dehydrogenase. The subunit molecular weight of 2-methyl-branched chain acyl-CoA dehydrogenase was 41,500 on SDS-PAGE in the presence and absence of 2-mercaptoethanol (Fig. 4). These data indicate that the enzyme consists of four equal size subunits as in the case of isovaleryl-CoA-and short chain acyl-CoA dehydrogenases and that the binding between subunits is not through a disulfide linkage. However, unlike these other two enzymes which readily dissociate into four subunits in SDS-PAGE, 2methyl-branched chain acyl-CoA dehydrogenase gave a single protein band with M , = 85,000 wben analyzed without boiling the enzyme preparation (Fig. 4). This finding may suggest that the binding forces for four subunits are not equal and that the force between two subunits in a dimer is stronger than that which binds two dimers.
The absorption spectrum and fluorescence emission and excitation spectra of 2-methyl-branched chain acyl-CoA dehydrogenase are typical for FAD, indicating that this enzyme contains FAD as a prosthetic group. The FAD content was calculated to be 0.5 mol/subunit from the absorption spectrum. This FAD content is not a whole number, probably due to a partial loss of FAD in the purification process judging from the observation that the activity of the purified enzyme is enhanced 2.3-fold by the addition of 100 FM FAD. These results suggest that 2-methyl-branched chain acyl-CoA dehydrogenase originally contained 1 mol of FAD/mol of subunit in native form, as in the case of isovaleryl-CoA dehydrogenase (Table VI). In contrast, the activity of the purified short chain acyl-CoA dehydrogenase was not enhanced at all by the addition of FAD: its A275/A450 ratio in the absorption spectrum was 6.3, a typical value for an acyl-CoA dehydrogenase fully saturated with FAD (Table VI), indicating that the final short chain acyl-CoA dehydrogenase preparation contains 1 mol of FAD/mol of subunit. In catalytic properties, these three enzymes distinctly differ from each other. 2-Methyl-branched chain acyl-CoA dehydrogenase is specific for isobutyryl-CoA and S-2-methylbutyryl-CoA, isovaleryl-CoA dehydrogenase is for isovaleryl-CoA (9), and short chain acyl-CoA dehydrogenase is for n-butyryl-CoA and n-valeryl-CoA (8). There is essentially no cross-reactivity in these enzyme-substrate combinations except for n-valeryl-CoA, which is dehydrogenated by both short chain acyl-CoA and isovaleryl-CoA dehydrogenases. The high degree of substrate specificities of these three acyl-CoA dehydrogenases for short chain acyl coenzyme A esters is indicative of finely defined conformation surrounding the active and substrate-binding sites of these enzyme. These three enzymes are also immunologically distinct from each other (Table VI) (9).
The activity of 2-methyl-branched chain acyl-CoA dehydrogenase for either substrate is inhibited by low concentrations of organic sulfhydryl reagents such as N-ethylmaleimide, p-hydroxymercuribenzoate, and methyl mercury iodide ( Table V). The degrees of inhibition were essentially equal for the two substrates. The enzyme activity was severely inhibited by heavy metal ions such as Hg", Cu'+, and Ag' which are known to interact with sulfhydryl groups in proteins. These results suggest the existence of an essential cysteine residue at the active site. Similar inhibitory effects by organic sulfhydryl reagents have been observed on isovaleryl-CoA dehydrogenase (9) and apo-medium chain acyl-CoA dehydrogenase (22).