Purification and partial characterization of 3-hydroxyisobutyryl-coenzyme A hydrolase of rat liver.

An unusual feature of valine catabolism is a reaction in which an intermediate of its catabolic pathway, (S)-3-hydroxyisobutyryl-CoA, is hydrolyzed to give the free acid and CoA-SH. The enzyme responsible for this reaction, 3-hydroxyisobutyryl-CoA hydrolase (EC 3.1.2.4), was purified 7200-fold from rat liver in this study. The purified enzyme consists of a single polypeptide with an M(r) of 36,000 in the native and denatured forms. The hydrolase is highly specific for (S)-3-hydroxyisobutyryl-CoA and 3-hydroxypropionyl-CoA (Km, 6 and 25 microM, respectively) with optimal activity around pH 8. The turnover rate of the enzyme for (S)-3-hydroxyisobutyryl-CoA is 270 s-1, which is high relative to other enzymes of the valine pathway. Likewise, activity of the enzyme expressed on a wet weight basis is also very high in the major tissues of the rat. These findings suggest that rapid destruction of (S)-3-hydroxyisobutyryl-CoA produced during valine catabolism is physiologically important. We propose that the need for a mechanism to protect cells against the toxic effects of methacrylyl-CoA, which is maintained in equilibrium with (S)-3-hydroxyisobutyryl-CoA by crotonase, explains why valine catabolism involves this enzyme and why its tissue activity is so high.

hydroxyisobutyric acid) that readily diffuses out of cells in which i t is formed. Indeed, it seems paradoxical that an acyl-CoA hydrolase should be required for this pathway when both proximal and distal parts of the pathway involve CoA ester intermediates. Furthermore, HIB-CoA hydrolase must be very specific for its substrate to avoid interference with catabolism of fatty acids, leucine, and isoleucine. Although there has been much interest in the interorgan traffic of (S)-3-hydroxyisobutyrate and its possible role as a substrate for various processes (2-61, we propose here that the reason for hydrolysis of (SI-HIB-CoA is to protect cells against toxic effects of methacrylyl-CoA, an intermediate in the valine pathway occurring upstream of (S)-3-HIB-CoA.
Composition of Bufers-Buffer A consisted of 50 m M potassium phosphate, pH 7.5, containing 0.1 m M EDTA, 0.1 p~ leupeptin, and 10 pg/ml trypsin inhibitor; Buffer B was 25 m M potassium phosphate, pH 7.5, containing 0.1 m M EDTA.
Purification Procedures-All procedures were performed at 0-4 "C. Frozen livers in 180-g portions were homogenized with an Oster blender (household type) at full speed for 4 min in 720 ml of Buffer A containing 1 nm EDTA, 1% bovine serum, 10 N-tosyl-L-phenylalanine chloromethyl ketone, 0.1 m M phenylmethanesulfonyl fluoride, and 0.5% Triton X-100. The homogenate was centrifuged at 9500 x g for 15 min. The supernatant was passed through four layers of cheesecloth, and the pH was adjusted to 7.5 with 2 M Tris.
The supernatant was made to 30% saturation in ammonium sulfate (176 g/liter) by slow addition of a solid salt with constant stirring and was allowed to stir for 20 min before centrifugation at 9500 x g for 20 min. Fat was effectively removed by this step. The supernatant was passed through four layers of cheesecloth, made to 45% saturation in ammonium sulfate (addition of 94 g/liter) as above, and allowed to stir for 20 min before centrifugation at 9500 x g for 30 min. The supernatant obtained was made to 75% saturation in ammonium sulfate (addition of 210 gfliter) as above and allowed to stir for 20 min before centrifugation at 9500 x g for 20 min. The pellet was dissolved in -100 ml of Buffer A and stored at -80 "C. The procedure described above was repeated once more.
The preparations were thawed, combined, and applied to a phenyl-Sepharose gel (350 ml) equilibrated with Buffer A containing 1 M ammonium sulfate on a sintered glass filter funnel (diameter, 9.5 cm). The gel was washed with the same buffer, and a fraction containing the HIB-CoA hydrolase activity not bound to the gel was collected. 332 g of solid ammonium sulfate (-80% final saturation) was slowly added to 870 ml of the fraction with constant stirring and followed by stirring for another 20 min before centrifugation at 9500 x g for 20 min. The pellet was dissolved in -60 ml of Buffer A and applied to an octyl-Sepharose gel (-100 ml) equilibrated with Buffer A containing 1 M ammonium sulfate on a sintered glass filter funnel (diameter, 6.5 cm). The gel was washed with the same buffer, and a fraction containing the hydrolase activity not bound to the gel was collected and concentrated by ammonium sulfate precipitation as above.
The pellet obtained was dissolved in -35 ml of 10 m M Tris-C1, pH 8.5, containing 0.1 m M EDTA and dialyzed against 1 liter of the same buffer for 18 h with two buffer changes. After removal of aggregated proteins formed during dialysis by centrifugation (9500 x g for 10 min), the dialysate was applied to a DEAE-Sephacel column (2.5 x 15 cm) equilibrated with the same buffer. The column was washed with the buffer at a flow rate of -65 ml/h until the absorbance of the eluate at 280 nm decreased almost to zero (-300 ml) and then eluted with 10 m M Tris-C1, pH 8.5, containing 0.1 m M EDTA and 0.1 M NaCl.
The fractions with the hydrolase activity were combined and dialyzed against 2 liters of 10 nm K+/citrate buffer, pH 5.6, containing 0.1 m M EDTA overnight. After removal of aggregated proteins as above, the dialysate was applied to a CM-Sepharose column (2.5 x 8 cm) equilibrated with the same buffer. The column was washed with the buffer at a flow rate of -75 mlh until the absorbance at 280 nm of the eluate was close to zero (-220 ml) and then eluted with 10 m M K+/citrate, pH 5.6, containing 0.1 m M EDTA and 0.1 M NaCl.
The fractions with the hydrolase activity were combined and concentrated to -5 ml by ultrafiltration (Y"10 membrane). The concentrate was diluted to -25 ml with Buffer B and again concentrated. This dilution and concentration cycle was repeated again to change buffer.
The concentrate obtained was applied to a hydroxylapatite column (1.5 x 4 cm) equilibrated with Buffer B. The column was washed with the buffer at a flow rate of -20 ml/h until the absorbance of the eluate at 280 nm decreased almost to zero (-100 ml) and eluted with Buffer B containing 50 m M potassium phosphate (total 75 m M potassium phosphate). The main hydrolase activity fractions ( Fig. 1) were combined, and Tween 20 was added to 0.01%.
After concentration to -1.5 ml by ultrafiltration, the preparation was made to 10% in glycerol and applied to a Sephacryl S-200 column (2.5 x 46.5 cm) equilibrated with 50 m M potassium phosphate, pH 7.5, con-' J. W. Hawes and R. A. Harris, unpublished method. taining 0.1 m M EDTA, 0.1 M KCl, and 0.05% Tween 20. The column was eluted with the buffer at a flow rate of 17.5 mlh. The fractions with the high ratio of the hydrolase activity to absorbance at 280 nm (more than 200) (Fig. 2) were combined and concentrated to -2 ml by ultrafiltration.
The buffer of the preparation was changed to 10 m M K'lcitrate buffer, pH 5.6, containing 0.1 m M EDTA and 0.05% Tween 20 by the ultrafiltration method as above. The preparation was applied to a CoA-Sepharose column (1 ml) equilibrated with the same buffer. The column was washed with 14 ml of the buffer at a flow rate of -30 ml/h and eluted with 10 nm K+/citrate buffer, pH 5.6, containing 0.1 m M EDTA, 0.05% Tween 20, and 0.1 M NaCl. The column was further eluted with the same buffer containing 0.5 M NaCl to remove proteins bound to the column and equilibrated again with 10 m M K+/citrate buffer, pH 5.6, containing 0.1 m M EDTA and 0.05% Tween 20. Fractions with the hydrolase activity eluted with the buffer containing 0.1 M NaCl were combined. The NaCl concentration was decreased to less than 10 m M by the ultrafiltration method, and the sample was applied again to the CoA-Sepharose column. The column was washed with 10 ml of the equilibrating buffer and eluted with Buffer B (pH 7.5) containing 0.05% Tween 20.
The eluate was concentrated to -1.5 ml by ultrafiltration, made 10% in glycerol, and applied again to the Sephacryl S-200 column as above.
Protein peak fractions consisting of a single polypeptide (molecular weight of 36,000) examined by SDS-PAGE were combined and stored at Assay of Hydrolase Activity-The hydrolase activity was assayed spectrophotometrically at 30 "C in a total volume of 1 ml with 0. Then, 0.1 ml of 1 m M DTNB was added and absorbance at 412 nm was measured. A control mixture without the hydrolase was prepared and treated in the same manner. The absorbance obtained after enzyme reaction was corrected by the absorbance of control mixture.
In the case of methylmalonyl-CoA (0.3 mM) used as substrate, crotonase was omitted from the reaction mixture, 0.34 pg of the hydrolase was used, and the incubation for the reaction was performed at 30 "C for 40 min.
Analytical Methods-3-Hydroxyisobutyrate was quantitated by the method of Rougraff et al. (12) with recombinant 3-hydroxyisobutyrate dehydrogenase as the enzyme source. SDS-PAGE was performed as described by Laemmli (13) except the acrylamide concentration was 12%. Samples for electrophoresis were treated as described previously (14). Protein determination was by the BCA method with bovine serum albumin as standard (15).

Purification of HIB-CoA Hydrolase from Rat Liuer-HIB-
CoA hydrolase was purified 7200-fold with an overall yield of 3.6% (Table I). Hydroxylapatite column chromatography proved to be one of the most effective steps of the purification with hydrolase activity being eluted with 75 mM phosphate buffer immediately after a major protein peak (Fig. 1). In the first Sephacryl S-200 column chromatography step (Fig. 2), it was found important to combine and recover only fractions with very high ratios of hydrolase activity to absorbance a t 280 nm. Tween 20 had to be added t o all buffers after hydroxylapatite column chromatography to minimize loss of the enzyme by adsorption to plastic. The final preparation consisted of a single Purification of HIB-CoA hydrolase from rat liver Refers to protein and activity of supernatant after centrifugation. polypeptide with a molecular weight of 36,000 as determined by SDS-PAGE (Fig. 3). The molecular weight of the hydrolase was also 36,000 by gel filtration, indicating that the native enzyme is a monomer.
pH Optimum and Lack of Cation and Nucleotide Effects-Partially purified HIB-CoA hydrolase was reported to have a pH optimum of about 5.6 (11, and that was the pH of the assay mixture used for the determination of activity in initial attempts to purify the enzyme. However, we found that the enzyme is active over a very wide range with the optimum being about pH 8 (Fig. 4). This was not an artifact of the coupled assay since crotonase is effective over a broad pH range and its activity at the concentrations used in the assay far exceeded that of HIB-CoA hydrolase. Thus, all subsequent work to purify and characterize the enzyme was carried out at pH 8.0.
The following salts and nucleotides were without effect upon HIB-CoA hydrolase activity: 60 mM KC1 and NaCl; 5 mM CaCI, and MgCI,; and 3 mM ATP, ADP, NAD+, and NADH.
For kinetic analysis, initial concentrations of (S)-HIB-CoA and 3-hydroxypropionyl-CoA were calculated from the extent of their formation from methacrylyl-CoA and acrylyl-CoA at equilibrium by the hydration reactions catalyzed by crotonase. According to the optical assay described previously (17), 37% of methacrylyl-CoA and nearly 100% of acrylyl-CoA were converted to their respective hydroxy-CoA esters within seconds under the standard conditions of the HIB-CoA hydrolase assay. Lineweaver-Burk plots were linear with both substrates with the lowest K,,, and turnover number being obtained with (SI-HIB-CoA (Table 111).
CoA esters that were hydrolyzed at slow rates inhibited (SI-HIB-CoA hydrolysis competitively (Table 11). K, values were relatively high compared with the K,,, value for (SI-HIB-CoA. pH Optimum for Methylmalonyl-CoA Hydrolysis by HIB-COA Hydrolase-As noted above, DL-methylmalonyl-CoA was hydrolyzed by HIB-CoA hydrolase, albeit slowly relative to (SI-HIB-CoA (Table 11). It is interesting to note that ( S )-methylmalonyl-CoA hydrolase (EC 3.1.2.17) was reported previously by Kovachy et al. (18,19) to have almost the same molecular weight (35,000) as found in the present study for HIB-CoA hydrolase. Optimal activity of (S)-methylmalonyl-CoA hydrolase was found at pH 6 by Kovachy et al. (19), and this was also found to be the optimum pH for methylmalonyl-CoA hydrolysis by HIB-CoA hydrolase (Fig. 4).
Tissue Distribution of HZB-CoA Hydrolase-Distribution of HIB-CoA hydrolase in major rat tissues is given in Table IV. Liver has the highest activity of the enzyme, followed closely by heart and then kidney. Muscle and brain have the lowest activities among the tissues examined.
The HIB-CoA hydrolase activity of mitochondria prepared from rat liver was found to be 152 milliunitdmg of protein.
Since liver contains -60 mg of mitochondrial proteidg, wet weight (20), this value fits well with the finding of an enzyme activity of -10 unitslg, wet weight, thereby confirming the view that HIB-CoA hydrolase is a mitochondrial enzyme. DISCUSSION HIB-CoA hydrolase has been purified to homogeneity from rat liver. The purified enzyme exhibits very high activity with ( S )-HIB-CoA and 3-hydroxypropionyl-CoA as substrates. Activity was detectable with several short-chain CoA esters, but ( R 1-HIB-CoA is not a substrate. No cofactors are necessary for enzyme activity. The enzyme exists as a monomer with a molecular weight of 36,000. HIB-CoA hydrolase had been previously purified %fold from an alcohol-KC1 extract of pig heart (1). A much lower pH optimum for enzyme activity was reported for the partially purified enzyme (1) than found in the present study with the purified enzyme. The reason for the apparent discrepancy is not known, but a pH optimum slightly on the alkaline side makes good physiological sense considering the intracellular location of HIB-CoA hydrolase (mitochondrial matrix space).
HIB-CoA hydrolase shows great substrate specificity for ( S ) -HIB-CoA and 3-hydroxypropionyl-CoA, thereby restricting its action to the valine catabolic pathway and a minor pathway for propionate catabolism involving the latter CoA ester. The degree of specificity exhibited by the enzyme undoubtedly is important in preventing interference with numerous metabolic   Values were calculated from approximately 37 and 100% conversion of methacrylyl-CoA to (SI-HIB-CoA and acrylyl-CoA to 3-hydroxypropionyl-CoA, respectively, a t equilibrium in the presence of crotonase. processes involving CoA ester intermediates. Although the enzyme was found to hydrolyze nine additional CoA esters, all were poor substrates relative to (S)-HIB-CoA. It is not likely that HIB-CoA hydrolase can hydrolyze any of these CoA esters effectively under physiological conditions. However, their hydrolysis may be catabolized in the event that their concentration became markedly elevated due to a defect in a downstream enzyme of their catabolic pathway. The low but significant capacity of HIB-CoA hydrolase to hydrolyze methylmalonyl-CoA is a case in point. Kovachy et al. (18,19) have purified and partially characterized rat liver D(S)-methylmalonyl-CoA hydrolase. This enzyme is not known to have a function in normal metabolism, but rather it seems to be present in cells as a safeguard to prevent CoA sequestration in the event of blockage of the propionyl-CoA catabolic pathway by an enzyme defect, e.g. propionyl-CoA carboxylase deficiency or vitamin B,, deficiency. The characteristics of the enzyme described by Kovachy et al. (18,19) are quite similar to those of HIB-CoA hydrolase with respect to the following:  (18,19). Although more definitive evidence is needed, the present study suggests that the (SI-methylmalonyl-CoA hydrolase activity of the enzyme previously purified by Kovachy et al. may correspond to a minor activity of HIB-CoA hydrolase. Hydrolysis of 2-methyl-3-hydroxybutyryl-CoA by HIB-CoA hydrolase may occur under some conditions. This CoA ester is formed by crotonase from tiglyl-CoA, a n intermediate in the isoleucine catabolic pathway. The normal pathway calls for conversion of 2-methyl-3-hydroxybutyryl-CoA to 2-methylacetoacetyl-CoA. However, 2-methyl-3-hydroxybutyrate is found in the urine of ketotic rats and humans (2), suggesting the latter conversion may be impeded by the redox state in ketosis. Although HIB-CoA hydrolase hydrolyzes 2-methyl-3-hydroxybutyryl-CoA at a relatively slow rate (3 units/mg protein), it most likely is the enzyme responsible for the production of the It is interesting to consider why a step is included in the valine catabolic pathway (Fig. 5) in which a CoA ester is hydrolyzed to a free carboxylic acid. Destruction of a n activated acyl group is rare in metabolic pathways, particularly when subsequent steps of the pathway also involve activated intermediates. Since it seems that (S)-HIB-CoA could be converted to (SI-methylmalonyl-CoA in just two steps (by a n alcohol dehydrogenase followed by a n aldehyde dehydrogenase as depicted by dashed lines in Fig. 5) rather than the four steps actually used by the pathway, including one that requires ATP, it seems odd that nature chose to sacrifice a CoA ester in the middle of what could have been a much simpler pathway. However, hydrolysis of (S)-HIB-CoA may be an important strategy for disposal of methacrylyl-CoA by cells. The latter compound is a thiol-reactive molecule that undoubtedly would inactivate numerous enzymes in the absence of a mechanism designed to minimize its intramitochondrial concentration (22,23). In a simpler pathway involving two dehydrogenases for direct conversion of (S)-HIB-CoA to (S)-methylmalonyl-CoA, the methacrylyl-CoA concentration would likely vary with the mitochondrial redox state, perhaps allowing toxic concentrations of this thiol-reactive CoA ester to accumulate under conditions of reducing equivalent overload.
The clinical experience with an infant born with a n almost complete deficiency of HIB-CoA hydrolase lends credence to the above interpretation (22). The child exhibited multiple congenital physical malformations, suggesting that a defect in this enzyme may be teratogenic. Death from a cardiac lesion occurred at 3 months. During life, the patient excreted large amounts of cysteinekysteamine conjugates of methacrylic acid, indicating that conjugation between methacrylyl-CoA and glutathione occurred. Methacrylyl-CoA (but not the free acid) reacts readily with free thiol groups of proteins (22,231, suggesting that high concentrations could cause inhibition of enzymes with sensitive sulfhydryl groups. Methacrylate oxygen esters have been reported to be teratogenic (24) and are recognized as genotoxidclastogenic agents from studies with F344/N rats, B6C3F1 mice, and mouse lymphoma cells (25,26). Because of the electron-withdrawing carboxylic acid oxygen ester group, compounds such as ethyl acrylate and methyl methacrylate (and the well known carcinogen acrylamide CH,=CHCONH,) readily react with the nucleophiles of proteins, DNA, and glutathione (Michael addition). Since thioesters are more electron-withdrawing than oxygen esters (271, we propose that methacrylyl-CoA and acrylyl-CoA are particularly reactive compounds with considerable potential for cytogenic, mutagenic, and clastogenic actions, making it important to maintain their intracellular concentrations extremely low. Acrylyl-CoA is a naturally occurring compound produced in small amounts from propionyl-CoA(1). Thus, it is interesting to note that cells use the same strategy to protect against acrylyl-CoA toxicity as methacrylyl-CoA toxicity, i.e. conversion of acrylyl-CoA to 3-hydroxypropionyl-CoA by crotonase followed by thioester cleavage of the latter compound by HIB-CoA hydrolase to give the less reactive compound 3-hydroxypropionate.
In contrast to the relatively low activities and turnover rates for the enzymes in distal and proximal parts of the valine pathway (activity and turnover numbers of 1.2 pmol/midg, wet weight, of liver and 18 s-l for liver branched chain a-ketoacid dehydrogenase complex (14); 1.0 ymollmidg, wet weight, and 7 s-' for 3-hydroxyisobutyrate dehydrogenase (28); 0.7 pmol/ midg, wet weight, and 2 s-l for methylmalonate semialdehyde dehydrogenase (29)), HIB-CoA hydrolase has markedly higher tissue activity and turnover number (9.9 pmol/min/g, wet weight, and 270 d ) , thereby accomplishing the rapid destruction of (5')-HIB-CoA as well as methacrylyl-CoA, the latter being due to the reaction catalyzed by crotonase, another enzyme with high tissue activity and turnover number. As a consequence, (S)-HIB-CoA and methacrylyl-CoA are not detectable in liver cells even when incubated under conditions that should maximize the concentrations of intermediates of the valine pathway (30).
Formation of (S)-3-hydroxyisobutyric acid by the action of HIB-CoA hydrolase produces a carboxylic acid that readily dif-fuses from its intracellular site of formation. In some tissues, the enzymes that catalyze steps of the valine pathway distal to the reaction catalyzed by HIB-CoA hydrolase are quite poorly expressed, thereby establishing interorgan trafficking of (S)-3hydroxyisobutyrate (2)(3)(4)(5)(6)21). Although there has been much interest in establishing a physiological role for circulating (SI-3-hydroxyisobutyrate, analogous to the important roles of circulating lactate and ketone bodies, its presence in blood may simply reflect the mechanism that has evolved to minimize methacrylyl-CoA toxicity.