Purification and Characterization of an a-Ketoisocaproate Oxygenase of Rat Liver*

a-KIC oxygenase activity supernatant stable

Rat liver contains a cytosolic a-ketoisocaproate oxygenase which oxidatively decarboxylates and hydroxylates a-ketoisocaproate to form fi-hydroxyisovalerate. This oxygenase was purified to near homogeneity. The oxygenase is unstable during purification, unless 5% monothioglycerol is added. The purified enzyme is stable in the presence of 5% monothioglycerol for 3 weeks at 4 "C and at least 10 weeks at -80 "C. The molecular weight of the a-ketoisocaproate oxygenase was determined to be 46,000 and 51,000 using denaturing and nondenaturing conditions, respectively, indicating a monomer.
The a-ketoisocaproate oxygenase requires Fe*+; other metal ions did not replace Fe*+. Ascorbate activates the enzyme at subsaturating levels of Fe*+, by regenerating Fe*+. The activity is markedly affected by the type of buffer used. For example, the oxygenase activity increased 2-to 3-fold when 0.1 M maleate was used. Iron chelators, such as ADP and EDTA, are inhibitory.
The ratio of decarboxylation of 1 mu a-[l-14C] ketoisocaproate (as measured by '*CO2 release) to decarboxylation of 1 mu a-[1-14C]keto-y-methiolbutyrate is 1.0 for all purification fractions, indicating that a single enzyme catalyzes the decarboxylation of both substrates.
The apparent K,,, and V,, values of the aketoisocaproate oxygenase using optimized assay conditions are 0.32 mu and 130 nmol/min/mg of protein for a-ketoisocaproate and 1.9 mu and 247 nmol/min/ mg of protein for a-keto-y-methiolbutyrate. The principal product of the purified a-ketoisocaproate oxygenase, using a-ketoisocaproate as a substrate, is p-hydroxyisovalerate, although small amounts of a compound, which has the chromatographic properties of isovalerate, are also produced.
Wohlhueter and Harper (1) first reported the decarboxylation of a-ketoisocaproate, the cr-keto analogue of leucine, by a soluble fraction from rat liver. They recog&zed that this activity was not "leaked" mitochondrial branched-chain aketo acid dehydrogenase.
Subsequently, Grant and Connelly (2) reported that a-ketoisocaproate, but not a-ketoisovalerate or a-keto-P-methylvalerate, is decarboxylated by cytosolic preparations from liver and kidney of mouse, rat, rabbit, guinea pig, and cow and also from chicken liver. This decarboxylase did not use CoASH or NAD+, cofactors required for * This research was partially supported by Grant  the mitochondrial branched-chain cu-keto acid dehydrogenase. Cytosolic preparations from rat liver decarboxylate both a-KIC' and cu-keto-y-methiolbutyrate, the cY-keto acid of methionine, in the presence of 0~ (3), indicating that the enzyme is an oxygenase. Partially purified preparations from rat liver convert a-KIC to /?-hydroxyisovalerate.
Isovalerate is not an intermediate of this reaction (3). In order to further characterize this enzyme and to determine whether a single enzyme is responsible for the abovementioned observations, the a-KIC oxygenase from rat liver was purified to near homogeneity.
Some of the physical and kinetic properties and the purification of the or-KIC oxygenase are described herein.

Rat Liver a-Ketoisocaproate Oxygenase 7461
genase was purified from 10, OOO X g supernatant fractions of 10% w/v rat liver homogenates in 0.25 M sucrose, 1% isopropanol. The a-KIC oxygenase activity in the 10,000 x g supernatant fractions is stable at -80 "C for up to 2 years. Protein was determined using the method of Bradford (8). All purification steps were done at 4 "C unless noted otherwise. Powdered (NH4)2S04 was slowly added to 7.2 liters of the 10, OOO X g supernatant fraction with stirring until 45% of saturation was achieved. After stirring for 30 min, the preparation was centrifuged 30 min at 10, OOO X g. The 45% (NH4)2S04 supernatant was made to 75% saturation by a slow addition of powdered (NH&S04. After stirring for 30 min, the preparation was centrifuged 30 min at 10, OOO X g. The pellet (45-75% (NH&S04 fraction) was resuspended in 20 m~ Tris-HC1, pH 7.8, 1% isopropanol (Buffer A) and dialyzed 60 h against 10 liters of Buffer A, with 5 changes of buffer. Precipitated material was removed by centrifugation for 15 min at 8000 X g.
The dialysate was applied to a DEAE-cellulose (Whatman DE52) column (4.8 X 83 cm), equilibrated with Buffer A (Fig. lA) and was eluted with 2.5 liters of Buffer A followed by a 9-liter linear gradient of 0-0.2 M NaCl in Buffer A. The a-KIC oxygenase eluted at approximately 0.06 M NaC1, and fractions containing activity were pooled ("DEAE-cellulose pool"), concentrated to 250 ml using an Amicon PM 10 ultrafiltration membrane ("concentrated DEAE-pool"), and stored at -80 "C. Fractions from the initial DEAE-column, which eluted before and after the peak of a-KIC oxygenase activity, were pooled ("DEAE-side fractions"). During initial purification attempts, large losses of activity occurred when DEAE, phenyl-Sepharose and Sephacryl columns were used. For example, passage over Sephacryl S-200 yielded 10% of the applied activity. However, when columns were pretreated with the DEAE-side fractions, 90% of the applied a-KIC oxygenase activity was recovered. Therefore, all columns (except the first DEAE-column) were pretreated with the DEAE-side fractions and then washed extensively with the elution buffer until protein was not detected in the eluent.
A portion of the concentrated DEAE-pool (116 ml) was adjusted to 5% monothioglycerol and 2.5 M NaCl by slow addition while stirring and the solution was applied to a phenyl-Sepharose CL-4B column (4.0 X 40 cm) (Pharmacia). This column had been pretreated with 1.5 liters of Buffer A, 1.0 liter of Buffer A containing 5% monothioglycerol, 2.5 M NaCl, 1.8 liters of DEAE-side fractions containing 2.5 M NaC1, and 2 liters of Buffer A, and then it was equilibrated with 1.5 liters of Buffer A containing 5% monothioglycerol, 2.5 M NaC1. The column was eluted with 1.5 liters of Buffer A containing 5% monothioglycerol, 2.5 M NaC1, followed by a 2-liter linear gradient of 2.5-0 M NaCl in Buffer A containing 5% monothioglycerol (gradient started at fraction 70). When applied to phenyl-Sepharose in buffer containing 2.5 M NaC1, a-KIC oxygenase activity was retarded and eluted just after the void volume (Fig. 1B). Much of the protein was retained by the column. Fractions containing a-KIC oxygenase were pooled and concentrated to 20 ml ("phenyl pool concentrate").
The phenyl pool concentrate was then applied to a Sephacryl S-200 column (4.8 X 82 cm) which had been pretreated with DEAEside fractions and equilibrated with Buffer A containing 5% monothioglycerol, 0.1 M NaCl (Fig. 1 0 . The column was eluted with the equilibration buffer at a flow rate of 13.6 ml/h. Fractions containing a-KIC oxygenase were pooled (Sephacryl S-200 pool) and stored at Removal of Monothioglycerol-Since monothioglycerol at concentrations greater than 0.6% inhibits the a-KIC oxygenase (see "Results"), only small aliquots (10-20 p1) of the Sephacryl S-200 pool could be assayed accurately. For experiments requiring larger amounts of enzyme, the monothioglycerol was removed by passing 0.5 ml of the Sephacryl S-200 pool over a Bio-Gel P-6 (50-100 mesh) column (0.75 X 14 cm) which had been pretreated with DEAE-side fractions and equilibrated with Buffer A containing 0.1 M NaC1. Protein was monitored by adding 50 pl of Coomassie blue reagent (8) to 10 ,ul of each fraction and visually checking for blue color. Monothioglycerol was detected by adding 1 ml of 0.3 m~ 5,5'-dithiobis-2nitrobenzoic acid) in 0.25 M glycylglycine, pH 8.2, to aliquots of each fraction and the yellow color checked visually. Fractions containing protein, but not monothioglycerol, were pooled and stored at -80 "C (P-6 pool). The purified a-KIC oxygenase is stable for at least 1 week in the absence of monothioglycerol if kept at -80 "C.
Polyacrylamide Gel Electrophoresis Using Denaturing or Nondenaturing Conditions-SDS-polyacrylamide gel electrophoresis was done using slab gels according to the procedure of Laemmli (9). The 10% acrylamide gels were stained with Coomassie blue dye by the procedure of Bonner and Laskey (10).
Tube gels containing 7.5% acrylamide were prepared for native gel electrophoresis as described (9) except that SDS was omitted. Gels were prerun overnight at a constant current of 2 mA/gel. During the prerun the cathode chamber contained 0.375 M Tris, 0.08% L-cysteine, pH 9.0, and the anode chamber 2.5 m~ Tris-glycine, pH 8.3, 0.008% L-cysteine. Then the buffers were replaced with 25 m~ Tris-glycine, pH 8.3,0.08% L-cysteine (cathode chamber) and 2.5 m~ Tris-glycine, pH 8.3, 0.008% L-cysteine (anode chamber). The sample was applied in a solution containing 10% glycerol, 0.003% bromphenol blue. Electrophoresis was carried out at a constant current of 1 mA/gel for the 1st h, then 2 mA/gel for the 2nd h, and 3 mA/gel for the final 3 h.
The gels were stained with Coomassie blue dye as described (IO). a-KIC oxygenase activity was monitored in gels which were not stained, but were sliced into 2-mm cross-sections, each slice put into a separate test tube, and 50 ,ul of 20 m~ Tris-HC1, pH 7.8, 1% isopropanol, 0.1 M NaC1,5% monothioglycerol added. These were shaken for 2 days at 4 "C to extract the enzyme from the gel slices and then assayed for a-KIC oxygenase activity.
Determination of Molecular Weight of the a-Ketoisocaproate Oxygenase by Sephacryl S-200 Chromatography-A Sephacryl S-200 column (1.6 X 63 cm) was equilibrated with 20 mM Tris-HC1, pH 7.8, 1% isopropanol, 0.1 M NaCl and then treated with 40 ml of the DEAEside fractions (see "Purification of a-Ketoisocaproate Oxygenase") and washed extensively with equilibration buffer. To calibrate the column, bovine serum albumin, ovalbumin, and cytochrome c were applied separately to the column and eluted with the equilibration buffer, and their elution volumes were determined by measuring absorbance at 280 or 410 nm (cytochrome c). The flow rate of the column was 1.7 ml/h. In a separate experiment, 0.5 ml of the purified a-KIC oxygenase (Sephacryl S-200 pool) was applied to the column and eluted with the equilibration buffer containing 5% monothio-  glycerol, and the a-KIC oxygenase activity was determined in each fraction. The data were plotted and the molecular weight was determined by the method of Andrews (11).

RESULTS
Stabilization of the a-Ketoisocaproate Oxygenase-Initial attempts to purify the a-KIC oxygenase yielded very low recoveries. During these studies it was shown that inclusion of FeSO,, ascorbate, and dithiothreitol in assays produced a 3-to 4-fold stimulation of activity (3). With the partially purified a-KIC oxygenase, optimal concentrations of FeSO,, ascorbate, and dithiothreitol were 1.0,0.5, and 1.0 mM, respectively. Despite the inclusion of these compounds in all assays, large losses of activity occurred during purification when DEAE-cellulose, phenyl-Sepharose, or Sephacryl columns were used. For example, when 0.4 ml of the concentrated DEAE-pool (Table I) was applied to a Sephacryl S-200 column (1.3 X 56 cm), the recovery of a-KIC oxygenase was less than 10%. Elution with high concentrations of NaCl gave additional protein but did not increase the yield of a-KIC oxygenase. Pretreatment of the Sephacryl S-200 column with protein-containing fractions recovered from the DEAE-cellulose column, which did not contain a-KIC oxygenase (DEAEside fractions), increased the recovery to 82%. Kaufman and Fisher (12) used a similar technique for purification of phenylalanine hydroxylase from rat liver.
Despite pretreatment of phenyl-Sepharose and Sephacryl columns with DEAE-side fractions, large losses of a-KIC oxygenase activity occurred. Although the a-KIC oxygenase was quite stable at 4 "C in less pure preparations such as the concentrated DEAE-pool (9% loss of activity in 4 days), the purified preparations of a-KIC oxygenase rapidly lost activity (50% loss of activity in 4 days, data not shown). This loss was apparently not dependent on the protein concentration, since dilution of the concentrated DEAE-pool from 16.8 to 1.68 mg of protein/ml did not alter stability. The a-KIC oxygenase was stable at -80 "C for up to 20 days in all purification fractions; however, at -20 "C or room temperature activity was rapidly lost (5). The a-KIC oxygenase is more stable at 4 "C than at -20 "C or room temperature.
Use of 5% monothioglycerol (0.6 M) stabilized the a-KIC oxygenase at 4 "C for at least 6 days (Table 11), but 1.0% monothioglycerol (0.12 M) was not effective. Dithiothreitol at 1 or 5 m and in the presence or absence of 5% glycerol did not stabilize the activity. Although monothioglycerol stabilizes a-KIC oxygenase, it decreases the initial activity, because a-KIC oxygenase is inhibited by assay concentrations of monothioglycerol greater than 0.6%, possibly due to a depletion of O2 in the assay mixture. This can be prevented by restricting   the amount of monothioglycerol introduced into the assay to below 0.2%. Purification of a-Ketoisocaproate Oxygenase Activity- Table I summarizes a purification of a-KIC oxygenase using conditions which stabilize this enzyme. The final steps (Steps 6-9), but not the initial ones, contained 5% monothioglycerol. Although the recovery for the concentrated DEAE-pool was only 50% for this preparation, recoveries of 70-80% are usudy obtained at this stage. As indicated under "Methods," the phenyl-Sepharose and Sephacryl S-200 columns were pretreated with DEAE-side fractions and 5% monothioglycerol which increased the yields greatly. The overall yield of a-KIC oxygenase in the Sephacryl S-200 pool was 27% with a fmal specific activity of 104 nmol/min/mg of protein. This fraction is referred to as the purified a-KIC oxygenase herein.
Molecular Weight of Rat Liver a-Ketoisocaproate Oxygenase-SDS-gel electrophoresis of the purified a-KIC oxygenase showed one major protein band with several minor protein bands (Fig. 2 A ) . When electrophoresis was carried out under nondenaturing conditions, a-KIC oxygenase activity migrated with the major protein band (Fig. 2 0 . The subunit molecular weight of a-KIC oxygenase determined by SDS-gel electrophoresis was 46,000 (Fig. 2B). Molecular weight of the a-KIC oxygenase was also determined under nondenaturing conditions using Sephacryl 5-200 chro- matography, according to the method of Andrews (1 1). M, = 51,000 was obtained by this method (Fig. 3).
Storage and Stability of the Purified a-Ketoisocaproate Oxygenase-Purified a-KIC oxygenase was quite stable at 4 "C in the presence of 5% monothioglycerol. Only 15% of the activity was lost over 3 weeks, but by 70 days all of the activity was lost. This may be due to the gradual autooxidation of the monothioglycerol which stabilizes the enzyme.
At -80 "C, the oxidase is stable for at least 70 days. Routinely, small aliquots of the purified a-KIC oxygenase were stored separately at -80 "C and thawed prior to use.
Metal Requirement of the a-Ketoisocaproate Oxygenase-Since Fe2' is required by the oxygenase, the effects of several metals were tested (Table 111). In the presence of ascorbate and dithiothreitol, ferrous iron (FeSO., and FeC12) and ferric iron (FeC4) enhanced the a-KIC oxygenase activity. In the absence of added iron, but with added ascorbate and dithiothreitol the activity was 25% of that in the presence of iron. When ascorbate, dithiothreitol, and iron were omitted, the activity was negligible. o-Phenanthroline, an iron chelator, abolishes most of the a-KIC oxygenase activity even in the presence of ascorbate and dithiothreitol. A large nonenzymatic decarboxylation of a-KIC was obtained in the presence of 1 m~ o-phenanthroline. This high rate of nonenzymatic decarboxylation occurred only when both o-phenanthroline and ascorbate were included in the assay mixture.
The ability of ferric iron ( Fe3+) to replace ferrous iron (Fez+) in the activation of the a-KIC oxygenase could be due to reduction of Fe3' to Fe2' by ascorbate. To test this possibility, the effect of ascorbate and dithiothreitol on a-KIC oxygenase activity in the presence of FeC12, FeCb, or no added iron was determined (Fig. 4). In the presence of ascorbate and dithiothreitol, FeC12 and FeCh gave similar results, but in the absence of added iron the activity was reduced 72%. Omission of either ascorbate or dithiothreitol had little effect on the activity in the presence of FeC12 or FeCb. However, when both ascorbate and dithiothreitol were omitted, FeC12, but not

I11
Effect of metals on a-ketoisocaproate oxygenase a-Ketoisocaproate oxygenase was assayed as described under "Methods" except that FeS04 was replaced by the addition shown. All assays contained 1 mM dithiothreitol (DTT) and 0.5 mM ascorbate except where indicated. Each assay contained 10 p1 (12 pg of protein) of the Sephacryl S-200 pool. Values are the means of two replicate assays. Blank activity was measured in duplicate assays in which the Sephacryl S-200 pool was replaced by 10 pl of 20 mM Tris-HC1, pH 7.8, 1% isopropanol, 0.1 M NaCl, 5% monothioglycerol.  FeC13, stimulated the oxygenase. These results suggest that the reduced form of iron, Fe", is utilized by the enzyme. In the presence of ascorbate or dithiothreitol, Fe3+ may be reduced to Fez+.
Previous studies (3) indicated t,hat the a-KIC oxygenase required a sulfhydryl reducing agent for maximum activity, yet only a slight effect of dithiothreitol is apparent in Fig. 4. The samples used for these experiments were isolated in 5% monothioglycerol (Sephacryl S-200 pool, Table I). The monothioglycerol was rapidly removed by passage over a Bio-Gel P-6 column and the enzyme was stored at -80 "C (P-6 pool, see "Methods"). Sulfhydryl groups of the enzyme may remain reduced during this treatment, thus eliminating the requirement for a sulfhydryl reducing agent during the assay.
FeS04 gave optimal activation of the a-KIC oxygenase between 1.0 and 5.0 mM. Concentrations above 5 mM became progressively inhibitory. The effect of ascorbate on a-KIC oxygenase activity was tested at suboptimal (0.05 mM) and optimal (2 mM) concentrations of FeS04, data not shown. In the presence of 0.05 m~ FeSO4, 1.0 mM ascorbate increased a-KIC oxygenase activity about l.6-fold, but in the presence of 2 mM FeS04, ascorbate had very little effect on the activity. The stimulatory effect of ascorbate, therefore, is probably due to its capacity to keep iron in the reduced, ferrous state.
Optimal Assay Conditions for the a-Ketoisocaproate Oxygenase-With purified a-KIC oxygenase, the activity was linear with time for 60 min and with protein concentration (up to 6 pg) (data not shown). The pH optimum of the a-KIC oxygenase was determined using several different buffers (Fig.  5 ) at constant ionic strength. The pH optimum in a Trismaleate buffer was 6.0. The activity was much lower when 4morpholineethanesulfonic acid, 4-morpholinepropanesulfonic acid, or Tris buffers were used. Above pH 7.0 the assay mixtures turned a reddish-brown color, presumably due to oxidation of FeS04.
The variability of a-KIC oxygenase activity in different buffers was not an ionic strength effect since increasing the concentration of NaCl from 0.05 to 0.4 M had little effect. In contrast, changing the concentration of maleate from 50 mM to 0.1 M increased the activity 1.5-fold (Fig, 6). Optimal activity was obtained at 0.1-0.2 M maleate. Higher concentrations were inhibitory.
Maleate may activate the a-KIC oxygenase by forming a chelate of Fez+ which is favorable for the catalytic reaction. EDTA (1.0-5.0 m~) and ADP (1.0-10.0 mM), other iron chelators, were tested in the presence of 50 mM 4-morpholineethanesulfonic acid, a buffer which has very little tendency to bind metal ions (13). These compounds caused considerable inactivation of a-KIC oxygenase activity (data not shown). Therefore, the a-KIC oxygenase either prefers the Fe-maleate complex or maleate activates this enzyme by some other mechanism.

Rut Liver a-Ketoisocaproate
Oxygenase 7465 Substrate Specificity of the a-Ketoisocuproate Oxygenase-Crude preparations of rat liver cytosol (3) oxidatively decarboxylate both a-KIC and a-keto-y-methiolbutyrate. To determine whether both a-keto acids are decarboxylated by the same enzyme, decarboxylation of a- [

1-'4C]KIC and a-[l-'*C]
keto-y-methiolbutyrate was monitored in the various purification fractions (Table IV). The ratio of decarboxylation of a-KIC to that of a-keto-y-methiolbutyrate was approximately keto-y-methiolbutyrate (a-KyMB) was measured using the fractions from Table I. Values are the means of two determinations. 1.0 and did not vary significantly throughout the purification, indicating that one enzyme catalyzes the decarboxylation of both substrates.
The apparent K, values of the a-KIC oxygenase €or a-KIC and a-keto-y-methiolbutyrate were determined using optimized assay conditions (Fig. 7). The apparent K, for a-KIC was 0.32 0.02 m~. The apparent K,,, for a-keto-y-methiolbutyrate was 1.90 -C 0.12 m. In contrast, the V,,, for a-ketoy-methiolbutyrate is higher than the V,,, with a-KIC as the substrate (247 6 versus 130 k 3 nmol/min/mg of protein).
Product Identification-Previous studies (3) indicated that 6-hydroxyisovaleric acid is the major product of the reaction. However, when CI-[~,~-~H]KIC was incubated with a partially purified preparation of oxygenase several radioactive compounds were detected. Thus, the purified a-KIC oxygenase was incubated with a-[U-14C)KIC and the products were separated as previously described (3). Two peaks containing radioactivity were obtained by Dowex-1-chloride chromatography (solid line, Pig. 8). The major peak (Peak I) was identified as P-hydroxyisovalerate by gas chromatography as previously described The activation of the a-KIC oxygenase by ferrous iron (Fe"), ascorbate (NADH or NADPH), and a sulfhydryl cornpound is typical of many non-heme iron-requiring oxygenases (14). Although other metals did not substitute for Fez+, ferric iron (Fe3+) was as effective as Fez+ in the presence of either ascorbate or dithiothreitol. This activation was due to the reduction of Fe3+ to Fez+ by the reducing agents. The stimulatory affect of ascorbate on the a-KIC oxygenase appears to be due to reduction of iron to the ferrous state (Fez+) since ascorbate had no effect in the presence of high concentrations of Fe2+.
Previously we demonstrated that the a-KIC oxygenase requires a sulfhydryl compound, such as dithiothreitol or CoASH (3). The data in Table I1 show that monothioglycerol, at a concentration of 0.6 M (5%), stabilizes the purified enzyme when stored at 4 "C. The stabilization of the a-KIC oxygenase by monothioglycerol may not be entirely due to its ability to reduce enzyme sulfhydryl groups since 0.12 M monothioglycerol should be an adequate sulfhydryl reducing agent, yet it does not prevent a-KIC oxygenase inactivation. The high concentration of monothioglycerol (0.6 M) may protect the enzyme from inactivation by oxygen. Many dioxygenases are inactivated by oxygen (14,15).
Optimal activity of the purified a-KIC oxygenase occurs at pH 6.0. The low pH optimum for our assay conditions may be artifactual since Fez+ is rapidly oxidized at higher pH to ferric hydroxide. With reactions carried out at pH 7.0 or greater, the assay mixtures developed a reddish-brown color which was not detected at pH 6.0 or 6.5. Increasing the maleate concentration from 50 VM to 0.2 M gave almost a 2-fold increase in the a-KIC oxygenase activity which apparently was not an ionic strength effect since varying the concentrations of NaCl from 0 to 0.4 M had no effect on the a-KIC oxygenase activity (5). It is possible that maleate forms a complex with Fez+ favorable for catalysis. Other metal chelators such as EDTA and ADP did not replace maleate and actually inhibited the a-KIC oxygenase.
The cytosolic a-KIC oxygenase also uses a-keto-y-methiolbutyrate (the keto analog of methionine) as a substrate. The affinity of the enzyme for a-keto-y-methiolbutyrate (apparent K , = 1.9 mM) is less than that for a-KIC (apparent K , = 0.3 mM), but the V,,, is greater. The product formed from a-ketoy-methiolbutyrate has not yet been identified, but it migrates identically to /?-hydroxyisovalerate using Dowex-1-chloride chromatography? If the reaction is similar to the a-KIC oxygenase reaction, the expected product would be 3-hydroxy-3-methylthiopropionic acid. The product of the a-KIC oxygenase reaction using a-keto-y-methiolbutyrate did not partition into diethyl ether. Thus, the product does not appear to be 3-methylthiopropionic acid, since the latter is soluble in ether (16).
The branched-chain amino acids, leucine, isoleucine, and valine, are catabolized by very similar pathways. These amino acids are fmst transaminated to form a-keto acids which can be oxidatively decarboxylated by a mitochondrial branchedchain a-keto acid dehydrogenase (EC 1.2.4.3 and 1.2.4.4) and converted to acyl-CoAs. The branched-chain a-keto acid dehydrogenase is a highly regulated enzyme and has been extensively investigated (17)(18)(19)(20). The cytosolic a-KIC oxygenase of rat liver provides an alternate pathway for metabolism of a-KIC. The existence of a separate enzyme for the catabolism of a-KIC may be related to the unique effects of leucine and/ or its metabolites on important metabolic processes which P. J. Sabourin and L. L. Bieber, unpublished data. apparently are not affected by isoleucine or valine. These include protein synthesis and degradation (21-23), insulin secretion (24)(25)(26), and glucose and pyruvate oxidation (19,(27)(28)(29)(30). Recent investigations suggest that the a-KIC oxygenase may be important at least in physiological situations which cause elevated levels of the branched-chain amino acids. Dixon and Harper (31) found that when rats are fed a 50% casein diet for 6 days, cytosolic a-KIC-decarboxylating activity (presumably due to the a-KIC oxygenase) increases approximately 3.5-fold. Diabetes also elevates rat liver cytosolic a-KIC-decarboxylating activity (32). The product of the a-KIC oxygenase, P-hydroxyisovaleric acid, occurs in the urine of patients with a variety of clinical disorders such as /?-methylcrotonylglycinuria (33), biotin deficiency (34), isovaleric acidemia (35), and ketoacidosis (36, 37). This compound could also be formed by hydration of /3-methylcrotonyl-CoA, an intermediate in leucine metabolism (38), and subsequent hydrolysis of the /?-hydroxyisovaleryl-CoA.
The cytosolic a-KIC oxygenase occurs in human liver' and produces /?-hydroxyisovalerate directly from a-KIC. In isovaleric acidemia the block in leucine metabolism is at the isovaleryl-CoA dehydrogenase (39). Therefore, no /?-methylcrotonyl-CoA should be formed. Yet high levels of /?-hydroxyisovaleric acid occur in urine of isovaleric acidemia patients (35). The a-KIC oxygenase may form/?-hydroxyisovaleric acid in such situations.
The a-KIC oxygenase occurs in rat liver and kidney (40), but has not been detected in brain, heart, skeletal muscle, or pancreas. It could function as a "safety valve" to prevent excessive accumulation of a-KIC, which is quite toxic (29,30,(41)(42)(43), by converting this compound to P-hydroxyisovaleric acid, which can be excreted via the urine.
The significance of the u-KIC oxygenase in the metabolism of a-keto-y-methiolbutyrate is unknown. Rat liver homogenates transaminate methionine to form a-keto-y-methiolbutyrate and decarboxylate a-keto-y-methiolbutyrate to form 3methylthiopropionate (16). Dixon and Benevenga demonstrated that 75% of the a-keto-y-methiolbutyrate-decarboxylating activity is associated with the mitochondrial fraction and 15% is cytosolic (44). They found a high K , (>1 m) for the cytosolic a-keto-y-methiolbutyrate decarboxylase. This is consistent with the results presented here and indicates the activity could be of importance when methionine or a-ketoy-methiolbut-yrate levels are elevated, such as in hypermethioninemia (45). The mitochondrial a-keto-y-methiolbutyrate decarboxylase has a K,,, of 0.1-0.6 rrm for this substrate (M), but the activity apparently is due to the branched-chain a-keto acid dehydrogenase (46). The utilization of a-keto-ymethiolbutyrate by the same cytosolic and mitochondrial enzymes that metabolize a-KIC is interesting, especially since rat liver also contains an enzyme that transaminates both leucine and methionine to form these a-keto acids (47).