Search

Five mitochondrial protein kinases, all members of a new family of protein kinases, have now been identified, cloned, expressed as recombinant proteins, and partially characterized with respect to catalytic and regulatory properties. Four members of this unique family of eukaryotic protein kinases correspond to pyruvate dehydrogenase kinase isozymes which regulate the activity of the pyruvate dehydrogenase complex, an important regulatory enzyme at the interface between glycolysis and the citric acid cycle. The fifth member of this family corresponds to the branched-chain alpha-ketoacid dehydrogenase kinase, an enzyme responsible for phosphorylation and inactivation of the branched-chain alpha-ketoacid dehydrogenase complex, the most important regulatory enzyme in the pathway for the disposal of branched-chain amino acids. At least three long-term control mechanisms have evolved to conserve branched chain amino acids for protein synthesis during periods of dietary protein insufficiency. Increased expression of the branched-chain alpha-ketoacid dehydrogenase kinase is perhaps the most important because this leads to phosphorylation and nearly complete inactivation of the liver branched-chain alpha-ketoacid dehydrogenase complex. Decreased amounts of the liver branched-chain alpha-ketoacid dehydrogenase complex secondary to a decrease in liver mitochondria also decrease the liver's capacity for branched-chain keto acid oxidation. Finally, the number of E1 subunits of the branched-chain alpha-ketoacid dehydrogenase complex is reduced to less than a full complement of 12 heterotetramers per complex in the liver of protein-starved rats. Since the E1 component is rate-limiting for activity and also the component of the complex inhibited by phosphorylation, this decrease in number further limits overall enzyme activity and makes the complex more sensitive to regulation by phosphorylation in this nutritional state. The branched-chain alpha-ketoacid dehydrogenase kinase phosphorylates serine 293 of the E1 alpha subunit of the branched-chain alpha-ketoacid dehydrogenase complex. Site-directed mutagenesis of amino acid residues surrounding serine 293 reveals that arginine 288, histidine 292 and aspartate 296 are critical to dehydrogenase activity, that histidine 292 is critical to binding the coenzyme thiamine pyrophosphate, and that serine 293 exists at or in close proximity to the active site of the dehydrogenase. Alanine scanning mutagenesis of residues in the immediate vicinity of the phosphorylation site (serine 293) indicates that only arginine 288 is required for recognition of serine 293 as a phosphorylation site by the branched-chain alpha-ketoacid dehydrogenase kinase. Phosphorylation appears to inhibit dehydrogenase activity by introducing a negative charge directly into the active site pocket of the E1 dehydrogenase component of the branched-chain alpha-ketoacid dehydrogenase complex. A model based on the X-ray crystal structure of transketolase is being used to predict residues involved in thiamine pyrophosphate binding and to help visualize how phosphorylation within the channel leading to the reactive carbon of thiamine pyrophosphate inhibits catalytic activity. The isoenzymes of pyruvate dehydrogenase kinase differ greatly in terms of their specific activities, kinetic parameters and regulatory properties. Chemically-induced diabetes in the rat induces significant changes in the pyruvate dehydrogenase kinase isoenzyme 2 in liver. Preliminary findings suggest hormonal control of the activity state of the pyruvate dehydrogenase complex may involves tissue specific induced changes in expression of the pyruvate dehydrogenase kinase isoenzymes.

The branched-chain alpha-ketoacid dehydrogenase complex, catalyst for the rate-limiting step of branched-chain amino acid catabolism, is controlled by a highly specific protein kinase (branched-chain alpha-ketoacid dehydrogenase kinase) that associates tightly with the complex. The activity state (proportion of the enzyme in its active, dephosphorylated state) of the complex varies dramatically in different rat tissues. The activity state of the complex in the liver is greater than that in any other tissue, and liver contains the lowest amount of kinase protein and kinase mRNA. However, protein malnutrition, a condition under which the complex is largely phosphorylated and inactive, resulted in a three- to fourfold increase in hepatic kinase activity with an accompanying increase in amounts of kinase protein and mRNA. Refeeding a 50% protein diet restored the normal activity state and the original levels of kinase protein and mRNA. The amount of kinase protein associated with the complex rather than changes in specific activity of the kinase appears responsible for observed differences in activity states of the complex in several rat tissues tested. Accordingly, the levels of kinase protein and mRNA measured are highest in tissues with greatest kinase activity (heart > kidney > liver), correlating reasonably well inversely with activity state of the branched-chain alpha-ketoacid dehydrogenase complex in the respective tissues. These observations suggest that the amount of kinase protein expressed in various tissues and in response to dietary protein deficiency is an important factor determining the activity state of the complex.

3T3-L1 fibroblasts have limited enzymatic capacity to oxidize valine. Enzymes expressed in these cells allow efficient oxidation of only the first carbon of this branched chain amino acid. The pathway is effectively truncated at the level of 3-hydroxyisobutyrate because of very low expression of two enzymes required for the complete pathway, 3-hydroxyisobutyrate dehydrogenase and methylmalonate semialdehyde dehydrogenase. These two enzymes, as well as the branched chain alpha-ketoacid dehydrogenase, are markedly induced upon differentiation of 3T3-L1 fibroblasts into adipocytes. Flux through the first two decarboxylation steps of valine catabolism is increased dramatically after differentiation, particularly through the step catalyzed by methylmalonate semialdehyde dehydrogenase. Activation of the distal portion of the valine catabolic pathway correlates with significant increases in enzyme protein and mRNA levels for 3-hydroxyisobutyrate dehydrogenase and methylmalonate semialdehyde dehydrogenase, and this establishes the pathway in 3T3-L1 adipocytes for utilization of valine carbon for lipogenesis. The induction profiles of 3-hydroxyisobutyrate dehydrogenase and methylmalonate semialdehyde dehydrogenase are very similar, suggesting coordinate regulation of the expression of these two valine pathway-specific enzymes. Induction of valine catabolism in 3T3-L1 cells is solely differentiation dependent, suggesting regulation by the same factors that govern differentiation of 3T3-L1 fibroblasts into adipocytes.

Regulation of the branched chain alpha-ketoacid dehydrogenase complex, the rate-limiting enzyme of branched chain amino acid catabolism, involves phosphorylation of 2 amino acid residues (site 1, serine 293; site 2, serine 303). To directly assess the roles played by these sites, site-directed mutagenesis was used to convert these serines to glutamates and/or alanines. Functional E1 heterotetramers were expressed in Escherichia coli carrying genes for E1 alpha and E1 beta under control of separate T7 promoters in a dicistronic vector. Mutation of phosphorylation site 1 serine to glutamate inactivated E1 activity, i.e. mimicked the effect of phosphorylation of site 1. Replacement of the site 1 serine with alanine greatly increased Km for the alpha-ketoacid substrate but had no effect on maximum velocity. The site 1 serine to alanine mutant was phosphorylated at site 2, but phosphorylation had no effect upon enzyme activity. Mutation of site 2 serine to either glutamate or alanine also had no effect upon enzyme activity, but phosphorylation of these proteins at site 1 inhibited enzyme activity. E1 mutated to change both phosphorylation site serines to glutamates was without enzyme activity. The binding affinity of E1 to the E2 core was not affected by mutation of the phosphorylation sites to glutamates, suggesting no gross perturbation of the association of E1 with the E2 core. The results provide direct evidence that a negative charge at phosphorylation site 1 is responsible for kinase-mediated inactivation of E1. Site 2 is silent with respect to regulation of activity by phosphorylation.

Levels of expression of two subunits of the liver branched-chain alpha-ketoacid dehydrogenase complex in response to extremes of dietary protein intake (50% versus 0% protein diet) were determined by quantitative immunoblotting. Dietary protein deficiency decreased the amount of E1 alpha protein to a greater extent than E2 protein. The ratio of E1 alpha to E2 was below 1 in the liver of animals starved for protein and above 1 in the liver of animals fed the high-protein diet. Supplementation of the 0% protein diet with 5% leucine (but not 5% valine) had the same effect as the 50% protein diet. The extremes of dietary protein also resulted in a divergent pattern of expression of the mRNAs for the subunits of the complex. The E1 beta message showed the expected corollary of being greater in the liver of the high-protein-fed rats than the no-protein-fed rats. In contrast, the E2 message was not affected by the two extremes of dietary protein and the E1 alpha message was greater in the liver of the no-protein-fed rats than the high-protein-fed rats. Thus, coordinate regulation of gene expression of the subunits of the complex does not occur in response to dietary protein. Post-transcriptional regulatory mechanisms most likely determine the amount of the complex and the ratio of its subunits. The decrease in E1 alpha/E2 protein ratio that occurs in dietary protein deficiency may increase sensitivity of the complex to phosphorylation-mediated inhibition by branched-chain alpha-ketoacid dehydrogenase kinase.

Severe ketotic diabetes induced in rats by streptozotocin resulted in a reduction in activity of the hepatic branched-chain alpha-ketoacid dehydrogenase complex, regardless of whether activity was expressed on the basis of liver wet weight, total liver, liver protein, or liver DNA. A decrease in enzyme specific activity (units of enzyme activity per mg of enzyme protein) was found responsible for the reduction in measurable enzyme activity of the complex. Insulin treatment reversed the decrease in enzyme specific activity. Treatment of tissue extracts with phosphoprotein phosphatase had no effect, indicating that activity of the complex was decreased by some mechanism other than reversible phosphorylation. Specific protein components of the complex were also not found reduced by the diabetic state. Induction of severe ketotic diabetes in rats previously fed a low-protein diet resulted in activation of the enzyme as a consequence of dephosphorylation. Nevertheless, the specific activity of the dephosphorylated enzyme of diabetic, low-protein-fed rats was decreased relative to that of control, low-protein-fed animals. Reconstitution studies with tissue extracts fortified with the purified E1 component indicate that severe diabetes induces a defect in this component of the hepatic branched-chain alpha-ketoacid dehydrogenase complex.

Feeding clofibric acid to rats caused little or no change in total activity of the liver branched-chain 2-oxo acid dehydrogenase complex (BCODC). No change in mass of liver BCODC was detected by immunoblot analysis in response to dietary clofibric acid. No changes in abundance of mRNAs for the BCODC E1 alpha, E1 beta and E2 subunits were detected by Northern-blot analysis. Likewise, dietary clofibric acid had no effect on the activity state of liver BCODC (percentage of enzyme in the dephosphorylated, active, form) of rats fed on a chow diet. However, dietary clofibric acid greatly increased the activity state of liver BCODC of rats fed on a diet deficient in protein. No stable change in liver BCODC kinase activity was found in response to clofibric acid in either chow-fed or low-protein-fed rats. Clofibric acid had a biphasic effect on flux through BCODC in hepatocytes prepared from low-protein-fed rats. Stimulation of BCODC flux at low concentrations was due to clofibric acid inhibition of BCODC kinase, which in turn allowed activation of BCODC by BCODC phosphatase. Inhibition of BCODC flux at high concentrations was due to direct inhibition of BCODC by clofibric acid. The results suggest that the effects of clofibric acid in vivo on branched-chain amino acid metabolism can be explained by the inhibitory effects of this drug on BCODC kinase.

The mitochondrial kinases responsible for the phosphorylation and inactivation of rat heart pyruvate dehydrogenase complex and the rat liver and heart branched-chain alpha-ketoacid dehydrogenase complexes have been purified to homogeneity. The branched-chain alpha-ketoacid dehydrogenase kinase is composed of one subunit with a molecular weight of 44 kDa; pyruvate dehydrogenase kinase has two subunits with molecular weights of 48 (alpha) and 45 kDa (beta). Proteolysis maps of branched-chain alpha-ketoacid dehydrogenase kinase and the two subunits of pyruvate dehydrogenase kinase are different, suggesting that all subunits are different entities. The alpha subunit of the rat heart pyruvate dehydrogenase kinase was selectively cleaved by chymotrypsin with concomitant loss of kinase activity, as previously shown for the bovine kidney enzyme, suggesting that the catalytic activity of pyruvate dehydrogenase kinase resides in this subunit. Polyclonal antibodies against branched-chain alpha-ketoacid dehydrogenase kinase, purified by an epitope selection method, bound only to the 44 kDa polypeptide of the branched-chain alpha-ketoacid dehydrogenase complex, substantiating that the 44 kDa protein corresponds to the kinase for this complex. Both kinases exhibited strong substrate specificity toward their respective complexes and would not inactivate heterologous complexes. The kinases possessed slightly different substrate specificities toward histones. Phosphorylation and inactivation of the branched-chain alpha-ketoacid dehydrogenase complex by its purified kinase was inhibited by alpha-chloroisocaproate and dichloroacetate, established inhibitors of the phosphorylation of the complex. cDNAs encoding the branched-chain alpha-ketoacid dehydrogenase kinase have been isolated from rat heart and rat liver lambda gt11 libraries. This represents the first successful cloning of a mitochondrial protein kinase. Preliminary data suggest that two different isoforms of the kinase may exist in different ratios in various tissues. No evidence was found for induction of the branched-chain alpha-ketoacid dehydrogenase complex nor its kinase by clofibric acid. Rather, clofibric acid is a potent inhibitor of the activity of the branched-chain alpha-ketoacid dehydrogenase kinase and this may be the molecular mechanism responsible for the myotonic effects of clofibric acid in man.