Studies on the Regulation of the Branched Chain a-Keto Acid Dehydrogenase in the Perfused Rat Liver*

The regulation of the branched chain a-keto acid dehydrogenase multienzyme complex was investigated in the isolated, perfused rat liver. The metabolic flux through the branched chain a-keto acid dehydrogenase was monitored by measuring the production of “COZ from infused 1-’”C-labeled branched chain a-keto acid substrates. The rate of decarboxylation of a-ketorl- ‘“C]isocaproate exceeded that of a-ket~[l-’~C]isovaler-ate at all concentrations of the substrates infused. Co- infusion of either a-ketoisovalerate or a-keto-fl-meth-ylvalerate inhibited the rate of a-ket~[l-’~C]isocaproate decarboxylation. The rate of a-ket~[l-’~C]isovalerate decarboxylation was enhanced during coinfusion of L(-)carnitine, while a-ket~[l-~~C]isocaproate decar- boxylation was unaffected. The presence of pyruvate in the perfusion medium resulted in an inhibition of the flux through the branched chain complex with either a-ketoisocaproate or a-ketoisovalerate as the substrate. DL-B-hydroxybutyrate infusion inhibited a- ket~[l-’~C]isocaproate decarboxylation by 18% but resulted in nearly a 100% stimulation of a-ket~[l-’~C]iso- (New England Nuclear Corp.) and counted. Knowing the correction for quenching and the specific radioactivity of the infused 1-I4C-labeled substrates, a continuous measurement of the metabolic flux through the branched chain a-keto acid dehydrogenase complex was obtained. Ketone bodies, P-hydroxybutyrate and acetoacetate, were measured in samples of the effluent perfusate using procedures described by Williamson and Corkey (19) and Mellanby and Wllliamson (20), respectively. [1-“C]Leucine, [l-‘4C]valine, aquasol, and phenethyla- mine were purchased from New England Nuclear (Boston, MA). 1-j4C-labeled a-ketoisocaproate and a-ketoisovalerate were prepared from the corresponding l-’4C-labeled amino acids (leucine and valine) using a procedure described by Rudiger et al. (21). Unlabeled a- ketoisocaproate, ct-ketoisovalerate, a-keto-P-methylvalerate, and L-leucine were purchased from Sigma (St. Louis, MO). Pyruvate and P-hydroxybutyrate dehydrogenase were obtained from Boehringer Mannheim (Indianapolis, IN). DL-P-Hydroxybutyrate was purchased from Calbiochem (La Jolla, CA). L(-)Carnitine chloride was the generous gift of the Otsuka Pharmaceutical Factory (Osaka, Japan). All other reagents and materials were of the highest purity available from commercial sources.

It has been established that during starvation the circulating blood levels of the branched chain amino acids, leucine, isoleucine, and valine increase transiently both in humans and in experimental animals (1,2). These amino acids represent a carbon source for the generation of energy, carbohydrate (e.g. valine and isoleucine), and ketone bodies (leucine) in various tissues (3). The branched chain amino acids are transaminated with a-ketoglutarate to form the corresponding a-keto acids which subsequently undergo oxidative decarboxylation in the mitochondrial compartment by the branched chain a-keto acid dehydrogenase mukienzyme complex which is an analogous enzyme complex to the pyruvate and a-ketoglutarate dehydrogenase complexes (4). The three branched chain cy-keto acids, a-ketoisocaproate, a-ketoisovalerate, and a-keto-P-methylvalerate, are oxidized by a single enzyme complex which requires CoASH, NAD', and thiamin pyrophosphate (1, 4-6). The various enzyme activities which are involved in branched chain amino acid metabolism are found in varying amounts in different tissues.
The demonstration that the transaminase activity is high in skeletal muscle and low in liver (7-9) while the branchec! chain a-keto acid dehydrogenase activity is high in liver and low in muscle (4) has been used as evidence that the primar) role of muscle in the metabolism of branched chain aminc acids is in transamination, while the liver is the primary sitc for oxidation of the branched chain a-keto acids (4, 8). Thir suggestion accounts for the observed elevation of the circulat. ing a-keto acid concentration during starvation (12). Hence an investigation of the regulation of the branched chain a. keto acid dehydrogenase in the liver is of special interest. Studies on the regulation of the branched chain a-keto acic dehydrogenase in rat liver have been performed using homog enates, isolated mitochondria, and hepatocytes. Few studie! of this enzyme complex have been performed using the is0 lated perfused rat liver. Walser et al. (13) studied the disap pearance of the three branched chain a-keto acids in thl perfused rat liver. However, the measurements in this stud! were conducted in the presence of 2 mM oleate which greatl! suppressed the uptake of the a-keto acids. The present stud! was performed in order to determine whether the metaball4 flux through the branched chain a-keto acid dehydrogenasc reaction could be determined continuously by measuring thl COZ production from infused 1-'%-labeled a-keto acids in i manner analogous to studies of the pyruvate dehydrogenasl complex in the perfused liver and heart (14,15). Previously we demonstrated that the flux through the branched chai~ dehydrogenase reaction could be monitored using this exper imental approach in the perfused rat heart (16).

MATERIALS AND METHODS
Male Sprague-Dawley rats weighing 180-200 g were used in thesl studies. Animals were given a standard laboratory chow and wate a d libitum. A nonrecirculating liver perfusion technique using i hemoglobin-free perfusion medium was employed and has been de scribed elsewhere (17). The perfusion medium was Krebs-Henseleit bicarbonate buffer (181, pH 7.4, saturated with an oxygen/carbo~ dioxide mixture (95%/5%) and maintained at 37 "C. The various 1 C-labeled as well as the unlabeled substrates were infused into thl perfusion circuit immediately prior to the liver. Oxygen consumptio~ by the liver was monitored using a Clark type oxygen electrode placet in the perfusion circuit immediately following the liver. Effluen perfusate samples were collected at 30-s intervals. Aliquots (2.5 ml of the perfusate were placed in 25-ml Erlenmeyer flasks fitted wit1 rubber serum stoppers equipped with plastic center wells containin, 0.3 ml of phenethylamine. Labeled carbon dioxide in the perfusat, which was produced from the infused l-"T-labeled substrates wa released by injecting 0.5 ml of 1 N HCl through the serum stoppe into the flasks. The flasks were agitated for 1 h and the center well were transferred to scintillation vials containing 10 rnl of aquasc 14

9009
This is an Open Access article under the CC BY license.

Branched Chain a-Keto Acid Metabolism
(New England Nuclear Corp.) and counted. Knowing the correction for quenching and the specific radioactivity of the infused 1-I4Clabeled substrates, a continuous measurement of the metabolic flux through the branched chain a-keto acid dehydrogenase complex was obtained.
Ketone bodies, P-hydroxybutyrate and acetoacetate, were measured in samples of the effluent perfusate using procedures described by Williamson and Corkey (19) and Mellanby and Wllliamson (20), respectively. [1-"C]Leucine, [l-'4C]valine, aquasol, and phenethylamine were purchased from New England Nuclear (Boston, MA). 1-j4C-labeled a-ketoisocaproate and a-ketoisovalerate were prepared from the corresponding l-'4C-labeled amino acids (leucine and valine) using a procedure described by Rudiger et al. (21). Unlabeled aketoisocaproate, ct-ketoisovalerate, a-keto-P-methylvalerate, and Lleucine were purchased from Sigma (St. Louis, MO). Pyruvate and P-hydroxybutyrate dehydrogenase were obtained from Boehringer Mannheim (Indianapolis, IN). DL-P-Hydroxybutyrate was purchased from Calbiochem (La Jolla, CA). L(-)Carnitine chloride was the generous gift of the Otsuka Pharmaceutical Factory (Osaka, Japan). All other reagents and materials were of the highest purity available from commercial sources.

RESULTS
In order to assess the validity of measuring 14C02 production from l-14C-labeled a-ketoisocaproate and a-ketoisovalerate as a monitor of the flux through the branched chain a-keto acid dehydrogenase reaction in the perfused rat liver, the experiment illustrated in Fig. lA was performed. The liver from a fed rat was perfused in the presence of 1 mM unlabeled aketoisocaproate for a period of 15 min. At 15 min, a tracer amount of a-ket~[l-'~C]isocaproate was infused for 10 min and samples were collected at 30-s intervals to monitor 14C02 production. Following the termination of the infusion of a-ket~[l-'~C]isocaproate, samples were collected every 12 s for 3 min and then every 30 s. The "off" or "washout" kinetics of 14C02 production shown in Fig. L4 indicates two distinct components with approximate half-times of 27 s and 4.8 min, respectively. The rapid kinetic component with a t l p of 27 s accounted for 94% of the I4CO2 washed out of the liver within the initial 2% min following the termination of the infusion of the labeled branched chain a-keto acid. This rapid kinetic component is very similar to the rapid component observed in the perfused rat heart with a-ketoisocaproate (16). In view of the fact that there are few metabolic fates for a-ketoisocaproate that would result in the rapid decarboxylation of aketoisocaproate, it was inferred that the rapid kinetic component observed in Fig be released at a slower rate; or (6) if the infused a-keto-[ l-'4C]isocaproate was transaminated to [ l-14C]leucine using glutamate as the amino group donor, the I4C-labeled leucine presumably present in an intracellular pool might be expected to be metabolized at a slower rate than the infused branched chain a-keto acid. In order to examine the second of these suggestions the experiment shown in Fig. 1B was performed.
A liver was perfused for 15 min with 1 m~ L-leucine following which a tracer amount of [l-'4C]leucine was infused during the 10-min interval indicated. Under these conditions, the metabolic flux through the branched chain a-keto acid dehydrogenase reaction was calculated as 1. (1 mM), was 59.5 pmo1.g". h" which is approximately 75% of the rate observed in the perfused liver from a fed rat with octanoate (0.5 m~) as the ketogenic substrate.
Infusing varying concentrations of a-keto[ 1-'4C]isovalerate into a perfused liver from a fed rat resulted in sequential increases in the flux through the branched chain a-keto acid dehydrogenase reaction (Fig. 3) and the rate of oxygen consumption by the liver. Again the flux through the enzyme complex saturated above 1 mM a-ketoisovalerate. It is evident that the flux through the branched chain complex was greater with a-ket~[l-'~C]isocaproate as the infused substrate than with a-ket~[l-'~C]isovalerate by nearly a factor of 2 (e.g. 30 compared to 54 pmo1.g". h").
This observation stands in contrast to the situation in the perfused rat heart (16) and in rat liver mitochondria (22) where a-ketoisovalerate was a better substrate than a-ketoisocaproate at the same concentration.
In a liver perfused with a-ket~[l-'~C]isocaproate (1 mM), co-infusion of either unlabeled u-ketoisovalerate (1 mM) or DL-a-keto-P-methylvalerate ( 2 mM) resulted in a 54% and 67% eto Acid Metabolism 901 1 decrease in the rate of I4CO2 production from the infused a-ket~[l-'~C]isocaproate, respectively (see Fig. 4). The rate of oxygen consumption by the liver was elevated by 10% upon infusion of either of the alternative branched chain substrates. Conflicting evidence has been published concerning the effects of L(-)-carnitine on the decarboxylation of branched chain a-keto acids. Although carnitine has been reported not to affect the rate of a-ketoisocaproate oxidation in liver (23) and muscle mitochondria (24), information from other laboratories indicates that L-(-)-carnitine significantly stimulates oxidation of leucine, valine, and a-ketoisocaproate in rat skeletal muscle mitochondria (25) and the rate of a-ketoisocaproate decarboxylation in rat liver homogenates and mitochondria (26,27). The effect of L-(-)-carnitine on the decarboxylation of I-'?-labeled a-ketoisocaproate and a-ketoisovalerate in the perfused liver is illustrated in Fig. 5, A and B, respec-tively. With cy-keto[l-'?]isocaproate as the substrate, L(-)carnitine infusion resulted in a 13% increase in oxygen consumption by the liver with no change in the rate of a-ketoisocaproate decarboxylation.
On the other hand, oxygen consumption by the liver and the decarboxylation of a-keto[l-"C]isovalerate were stimulated by approximately 11% and 50%, respectively, during the 8 min of L-(-)-carnitine infusion. We have demonstrated that pyruvate infusion into the perfused rat heart resulted in up to a 90% inhibition of the metabolic flux through the branched chain dehydrogenase reaction (16). While the exact mechanism of this pyruvatemediated inhibition of the branched chain complex is not known, it was shown that pyruvate infusion resulted in a decrease in the extractable branched chain cu-keto acid dehydrogenase activity in the perfused heart (28). The experiments depicted in Fig. 6, A and B, demonstrate the effect of pyruvate infusion (10 mM) on the flux through the branched chain dehydrogenase reaction in the perfused liver. Oxygen consumption increased by approximately 15% upon co-infusion of pyruvate and either branched chain substrate. However, pyruvate infusion inhibited the decarboxylation of a-keto[l-'?]isocaproate (1 IIIM) by 18% (Fig. 6A), while the decarbox-mM ylation of a-keto[l-'4C]isovalerate was inhibited by 12% (Fig.  6B). In the rat heart, pyruvate infusion inhibited ly-ketoisovalerate decarboxylation to a considerably greater extent than a-ketoisocaproate decarboxylation (16). The effects of fatty acids and ketone bodies on the decarboxylation of the branched chain amino acids and Lu-keto acids have been investigated in various tissues with somewhat conflicting results. While Odessey and Goldberg (10)  that octanoate caused a stimulation in the rate of decarboxylation of branched chain amino acids in the diaphragm and a recirculating heart perfusion system. Additionally, stimulation of valine oxidation by octanoate and palmitate in the perfused rat hindquarter has been reported (29). Our data in the nonrecirculating heart perfusion indicate that octanoate, P-hydroxybutyrate, and acetoacetate all inhibit strongly branched chain cY-keto acid decarboxylation (16). Crabb   when ,&hydroxybutyrate was withdrawn from the perfusion system. Increasing the perfusion concentration of a-keto[ l-"C]isovalerate to 5 lll~ did not change the stimulatory effect of /3-hydroxybutyrate on a-ketoisovalerate decarboxylation (data not shown). DISCUSSION The primary purpose of the present study was to investigate the regulation of the branched chain a-keto acid dehydrogenase complex in the isolated perfused rat liver. Nearly all of the information available in the literature concerning the regulation of this multienzyme complex in liver has been obtained using homogenates (26,27), isolated mitochondria (27), or isolated hepatocytes (31)(32)(33). The obvious advantage of using the perfused liver is that the regulation of the enzyme complex may be studied in a metabolic system under nearly in uivo conditions. The experiment illustrated in Fig. lA indicated that more than 90% of the 14C02 derived from the decarboxylation of infused a-keto[ l-'4C]isocaproate was "washed out" of the liver in the initial 2% min with a t112 of 27 s following cessation of the tracer infusion. The t1p of the kinetic component due to the pyruvate dehydrogenase reaction in the perfused liver is very similar but not identical with that calculated for the branched chain a-keto acids in the present study (e.g. compare tli2 = 18 s for pyruvate with tllz = 27s for a-ketoisocaproate) Studies of the concentration dependence of the rate of decarboxylation of infused branched chain a-keto acids in the perfused liver indicated that the flux through the enzyme complex is saturated at substrate concentrations in excess of 1 mM. A noticeable difference emerges between the decarboxylation rates of l-14C-labeled a-ketoisocaproate and a-ketoisovalerate in the perfused liver. At each concentration examined, the rate of a-ketoisocaproate decarboxylation exceeded the rate of a-ketoisovalerate decarboxylation. Evidence obtained in isolated liver mitochondria (22), in the perfused rat heart (16), and in a variety of other systems (5,34,35) has indicated that a-ketoisovalerate is decarboxylated at a more rapid rate than a-ketoisocaproate by the branched chain complex. It is possible that the various short and branched chain acyl-CoA intermediates in the pathway involved in the breakdown of a-ketoisocaproate and a-ketoisovalerate, respectively, may affect differentially the regulation of the branched chain complex in the intact liver. In the present experiments, it is suggested that acyl-CoA intermediates produced during aketoisovalerate oxidation may be more inhibitory to the enzyme complex than those produced during the oxidation of a-ketoisocaproate.
Furthermore, because the liver is capable of rapid rates of ketogenesis using a-ketoisocaproate as the substrate, the potentially inhibitory intermediates in the metabolism of a-ketoisocaproate may not accumulate to any appreciable extent. The rate of ketogenesis from a-ketoisocaproate (Fig. 2) exhibited saturation kinetics very similar to the kinetics of the increase in the rate of decarboxylation of a-keto[ l-14C]isocaproate as the concentration of the branched chain a-keto acid was increased in the perfusion medium. Such a correlation suggests that the measurement of 14C02 production from aketo[ 1-'*C]isocaproate as a monitor of metabolic flux through the branched chain complex is valid. Studies using either partially or highly purified preparations of the branched chain a-keto acid dehydrogenase complex have demonstrated that all three branched chain a-keto acids, a-ketoisocaproate, a-ketoisovalerate, and a-keto-P-methylvalerate, are decarboxylated oxidatively by a single multienzyme complex. The experiment illustrated in Fig. 4 indicated that co-infusion of either a-ketoisovalerate or a-keto-/3-methylvalerate with a-keto[l-'4C]isocaproate resulted in an increased hepatic oxygen consumption and an appreciable and reversible inhibition of the rate of 14C02 from the labeled a-ketoisocaproate. This experimental finding may be interpreted as a competition between the unlabeled and 14C-labeled a-keto acid at the level of the branched chain enzyme complex or at the level of the mitochondrial transport of the two branched chain a-keto acids.
We observed a differential effect of L-(--)-carnitine infusion on the rate of decarboxylation of a-ketoisocaproate and aketoisovalerate.
This effect (e.g. a stimulation of the rate of a-ketoisovalerate decarboxylation and no appreciable change in the rate of a-ketoisocaproate oxidation) likely is a result of the differential accumulation of acyl-CoA intermediates involved in the respective pathways for the metabolism of these two compounds in the absence and presence of L-(-)-carnitine. Indeed, Solberg and Bremer (36) demonstrated in rat liver mitochondria incubated in the presence of L-[Me-3H]carnitine and a-ketoisovalerate or a-ketoisocaproate that levels of propionylcarnitine and medium chain acylcarnitine were much greater with a-ketoisovalerate as the substrate than with a-ketoisocaproate.
Hence, L-(-)-carnitine may remove propionyl-CoA and other medium (branched) chain acyl-CoAs formed from a-ketoisovalerate oxidation which either may be inhibitory to the branched chain complex or may release free CoA which could accelerate the oxidation of the branched chain a-keto acid. If acyl-CoA intermediates in the pathway of a-ketoisocaproate oxidation do not accumulate or if the acyl-CoA species which accumulate are not as inhibitory to the branched chain complex, L-( -)-carnitine infusion would not cause an observable effect. It is unlikely that L-(-)-camitine causes an appreciable diversion of intermediates toward acylcarnitine formation during a-ketoisocaproate oxidation as the ketogenic rate in the presence and absence of L-(--Jcarnitine was unchanged (data not shown).
Infusion of relatively low levels of pyruvate into the perfused rat heart resulted in a 90% inhibition of the metabolic flux through the branched chain complex using a-ketoisovalerate as the substrate. A less extensive flux inhibition was noted using a-ketoisocaproate as the substrate (16). Additionally, the extractable activity of the branched chain complex was diminished by 72% upon infusion of pyruvate into a heart which had been perfused under conditions which maximally activated the branched chain complex (28). Infusion of pyruvate into the perfused liver resulted in an inhibition of aketo[ l-14C]isocaproate decarboxylation to a greater degree than a-keto[ l-'4C]isovalerate decarboxylation, just the reverse situation as was observed in the heart. Whether the flux inhibition observed in the liver was the result of an inactivation of the enzyme complex or was the result of a simple competitive inhibition of the branched chain complex by pyruvate or NADH is currently under investigation.
The decarboxylation of a-keto[ l-'4C]isocaproate was inhibited by the co-infusion of P-hydroxybutyrate into the liver.
On the surface, this effect would seem to be a simple NADHmediated inhibition of the enzyme complex since few other fates for P-hydroxybutyrate than oxidation to acetoacetate can occur in the liver. However, despite the increase in the mitochondrial NADH/ NAD' ratio upon infusion of P-hydroxybutyrate, the rate of decarboxylation of a-keto[ l-'4C]isovalerate was stimulated by 100%. Stimulation of the decarboxylation of [l-'4C]pyruvate in the perfused liver upon infusion of octanoate, oleate, or phydroxybutyrate has been reported (15,37).' It was proposed that the stimulation of pyruvate decarboxylation and the activation of the pyruvate dehydrogenase complex at low (ie. physiological) pyruvate levels were results of an accelerated acetoacetate/pyruvate exchange across the mitochondrial membrane on the monocarboxylate transporter. Under conditions of rapid ketogenesis ( i e . rapid acetoacetate efflux), the mitochondrial pyruvate level would be elevated and the pyruvate dehydrogenase kinase would be inhibited by pyruvate leading to an interconversion of the complex by the phosphoprotein phosphatase to its active form (15, 37). The mechanism by which a-ket~[l-'~C]isovalerate decarboxylation was stimulated upon infusion of P-hydroxybutyrate may occur in a similar fashion. First, both a-ketoisocaproate and a-ketoisovalerate have been shown to be transported across the liver mitochondrial membrane on the monocarboxylate transporter (22). While a-ketoisocaproate oxidation in the liver results in the formation of acetoacetate and acetyl-coA, a-ketoisovalerate produces only propionyl-CoA. Papa and Paradies (38) demonstrated that acetoacetate could replace OH-ions as an effective antiport for pyruvate on the monocarboxylate transporter. Hence, it is likely that as a-ketoisocaproate is oxidized, acetoacetate is generated and an exchange of the branched chain a-keto acid substrate for the product, acetoacetate, is established. Also, this exchange phenomena could explain why a-ketoisocaproate decarboxylation exceeds a-ketoisovalerate decarboxylation in the liver. In the case of the stimulation of the decarboxylation of a-ketoisovalerate by P-hydroxybutyrate, there may occur a situation analogous to the case for pyruvate. As acetoacetate is produced intramitochondrially, an exchange is established between acetoacetate and a-ketoisovalerate. Thus, an acceleration of the decarboxylation of this branched chain a-keto acid occurs either by mass action (e.g. an increase in the intramitochondrial concentration of the substrate) or by virtue of the fact that the branched chain a-keto acids tend to activate or to stabilize the enzyme complex (6, 28). This scenario requires that the transport of a-ketoisovalerate is limiting the oxidative system when infused alone into the perfusion system.