Malonyl-CoA inhibits proteolysis of carnitine palmitoyltransferase

Topology of hepatic mitochondrial carnitine palmitoyltransferase I

Our earlier work using intact mitochondria and isolated mitochondrial outer membranes confirms the observations of Murthy and Pande that CPT-I is located on the mitochondrial outer membranes and supports the notion that this enzyme has a malonyl-CoA binding domain facing the cytosol and an acyl-CoA binding domain facing the inter membrane space. Our data also suggests that coenzyme A binds at the active site of CPT-I, as does acyl-CoA, 2-bromopalmitoyl-CoA, and (+)-hemipalmitoylcarnitinium, but malonyl-CoA does not bind at that site. Inhibition of CPT-I at the malonyl-CoA binding site by HPG and Ro 25-0187, which have no CoA moiety, contributes to a resolution of this question in that the CoA itself is not essential for the binding of malonyl-CoA to its regulatory site, but the dicarbonyl function which is present in malonyl-CoA, HPG, and Ro 25-0187 is absolutely essential. Our re-evaluation of the topology of hepatic mitochondrial CPT-I confirms the original observations that this enzyme has at least two different binding domains, one domain binding malonyl-CoA, HPG, and Ro-25-187 and the other domain binding acyl-CoA and other inhibitors of CPT-I. Furthermore, the malonyl-CoA binding domain is exposed to the cytosolic face of the membrane. Our data showing that treatment of the intact mitochondria with trypsin causes release of adenylate kinase which indicates that trypsin has damaged the mitochondrial outer membrane, possibly allowing trypsin to enter the intermembrane space and act on CPT from within the outer membrane. Since trypsin's action is limited to arginine and lysine residues, an alternative explanation could be that the portion of the protein domain responsible for malonyl-CoA inhibition may not contain these residues. The latter explanation is plausible, since malonyl-CoA was able to protect against loss of activity and sensitivity to inhibition, but did not protect against loss of adenylate kinase, suggesting that rupture of the outer membrane is not necessarily related to loss of CPT activity. These results suggest that some protein domain that is necessary for CPT activity is exposed on the outer surface of the outer membranes. Therefore, it seems likely that trypsin would have to be able to hydrolyse protein domains of CPT that are inaccessible to Nagarse and papain.

Expression and regulation of carnitine palmitoyltransferase-Ialpha and -Ibeta genes

Two genes control expression of mitochondrial carnitine palmitoyltransferase-I (CPT-I), the enzyme that catalyzes the primary rate-controlling step in fatty acid oxidation. Two CPT-I isoforms have been found--a "liver" isoform (CPT-Ialpha) expressed in most tissues, but not in skeletal muscles, and a "muscle" isoform (CPT-Ibeta) expressed in muscles and adipocytes. Liver CPT-Ialpha increases dramatically at birth, but heart CPT-Ialpha is abundant in the fetus and diminishes at birth. Insulin, thyroid hormone, and fatty acids regulate expression of CPT-Ialpha in liver, whereas electrical stimulation increases CPT-Ibeta and decreases CPT-Ialpha in cardiac myocytes. Both genes are TATA-less and contain Sp1 transcription factor binding sites upstream of the start site of transcription. Multiple transcripts of both CPT-Ialpha and CPT-Ibeta exist, some of which may have roles in regulating fatty acid oxidation.

Fatal carnitine palmitoyltransferase II deficiency in a newborn: new phenotypic features

We describe the term male infant of asymptomatic, healthy nonconsanguineous parents presenting on the first day of life with nonketotic hypoglycemia, seizures, hepatomegaly, cardiomegaly with biventricular hypertrophy, and ventricular arrhythmias. Cranial ultrasound revealed cystic dysplasia with several foci of hyperechogenicity within the right basal ganglia. Free carnitine was markedly decreased in the urine and plasma with a pronounced elevation of plasma long-chain acylcarnitines. Fibroblast carnitine palmitoyltransferase II activity was reduced to 26% and 38% in the father and mother, respectively. The infant expired on day 5 of life from malignant ventricular tachy-arrhythmias. Diffuse lipid accumulation was evident at autopsy, including in the liver, heart, kidney, adrenal cortex, skeletal muscle, and lungs. This new case of infantile CPT-II deficiency illustrates the severity of the early onset form of CPT-II deficiency.

Lipid molecular order in liver mitochondrial outer membranes, and sensitivity of carnitine palmitoyltransferase I to malonyl-CoA

Mitochondrial outer membranes were prepared from livers of rats that were in the normal fed state, starved for 48 h, or made diabetic by injection of streptozotocin. Membranes were also prepared from starved late-pregnant rats. The latter three conditions have previously been shown to induce varying degrees of desensitization of mitochondrial overt carnitine palmitoyltransferase (CPT I) to malonyl-CoA inhibition. We measured the fluorescence polarization anisotropy of two probes, 1,6-diphenyl-1,3,5-hexatriene (DPH) and 1-(4-trimethylammoniumphenyl)-6-phenyl-1,3,5-hexatriene-p-toluenes ulfonate (TMA-DPH) which, when incorporated into membranes, report on the hydrophobic core and on the peripheral regions of the bilayer, respectively. The corresponding polarization indices (rDPH and rTMA-DPH) were calculated. In membranes of all three conditions characterized by CPT I desensitization to malonyl-CoA, rDPH was decreased, whereas there was no change in rTMA-DPH, indicating that CPT I is sensitive to changes in membrane core, rather than peripheral, lipid order. The major lipid components of the membranes were analyzed. Although significant changes with physiological state were observed, there was no consistent pattern of changes in gross lipid composition accompanying the changes to membrane fluidity and CPT I sensitivity to malonyl-CoA. We conclude that CPT I kinetic characteristics are sensitive to changes in lipid composition that are localized to specific membrane microdomains.

Regulation of mitochondrial outer-membrane carnitine palmitoyltransferase (CPT I): role of membrane-topology

The topology of the outer membrane carnitine palmitoyltransferase (CPT I) of rat liver mitochondria was studied systematically using several experimental approaches. Studies with immobilized malonyl-CoA and octanoyl-CoA showed that functionally the active and regulatory sites of CPT I are exposed on the outer (cytosolic) surface of the mitochondrial outer membrane. Anti-peptide antibodies generated against three linear peptide sequences that occur in between and on either side of two hydrophobic, putative transmembrane domains were used to (a) ascertain which were bound by intact mitochondria and mitochondria in which the outer membrane was permeabilized to proteins; and (b) to determine the size of fragments generated by limited proteolysis (by trypsin or proteinase K) of CPT I in intact or outer membrane-ruptured mitochondria. The sizes and immunoreactivity of the proteolytic fragments generated were correlated with the effects of the proteases on CPT I activity and malonyl-CoA sensitivity. The results of all the different approaches suggested the following: (i) CPT I has two transmembrane domains; (ii) both the N- and C-termini are exposed on the cytosolic side of the membrane; (iii) the linker region between the two transmembrane domains protrudes into the intermembrane space; (iv) both the active site and the malonyl-CoA-binding site are exposed on the cytosolic side of the membrane; (v) the amino-terminus of the protein interacts with the C-terminal domain of the protein to maintain the optimal conformation required for activity of the enzyme and its sensitivity to malonyl-CoA.

The mitochondrial carnitine palmitoyltransferase system. From concept to molecular analysis

First conceptualized as a mechanism for the mitochondrial transport of long-chain fatty acids in the early 1960s, the carnitine palmitoyltransferase (CPT) system has since come to be recognized as a pivotal component of fuel homeostasis. This is by virtue of the unique sensitivity of the outer membrane CPT I to the simple molecule, malonyl-CoA. In addition, both CPT I and the inner membrane enzyme, CPT II, have proved to be loci of inherited defects, some with disastrous consequences. Early efforts using classical approaches to characterize the CPT proteins in terms of structure/function/regulatory relationships gave rise to confusion and protracted debate. By contrast, recent application of molecular biological tools has brought major enlightenment at an exponential pace. Here we review some key developments of the last 20 years that have led to our current understanding of the physiology of the CPT system, the structure of the CPT isoforms, the chromosomal localization of their respective genes, and the identification of mutations in the human population.

Thia fatty acids, metabolism and metabolic effects

(1) The chemical properties of thia fatty acids are similar to normal fatty acids, but their metabolism (see below: points 2-6) and metabolic effects (see below: points 7-15) differ greatly from these and are dependent upon the position of the sulfur atom. (2) Long-chain thia fatty acids and alkylthioacrylic acids are activated to their CoA esters in endoplasmatic reticulum. (3) 3-Thia fatty acids cannot be beta-oxidized. They are metabolized by extramitochondrial omega-oxidation and sulfur oxidation in the endoplasmatic reticulum followed by peroxisomal beta-oxidation to short sulfoxy dicarboxylic acids. (4) 4-Thia fatty acids are beta-oxidized mainly in mitochondria to alkylthioacryloyl-CoA esters which accumulate and are slowly converted to 2-hydroxy-4-thia acyl-CoA which splits spontaneously to an alkylthiol and malonic acid semialdehyde-CoA ester. The latter presumably is hydrolyzed and metabolized to acetyl-CoA and CO2. (5) Both 3- and 4-thiastearic acid are desaturated to the corresponding thia oleic acids. (6) Long-chain 3- and 4-thia fatty acids are incorporated into phospholipids in vivo, particularly in heart, and in hepatocytes and other cells in culture. (7) Long-chain 3-thia fatty acids change the fatty acid composition of the phospholipids: in heart, the content of n-3 fatty acids increases and n-6 fatty acids decreases. (8) 3-Thia fatty acids increase fatty acid oxidation in liver through inhibition of malonyl-CoA synthesis, activation of CPT I, and induction of CPT-II and enzymes of peroxisomal beta-oxidation. Activation of fatty acid oxidation is the key to the hypolipidemic effect of 3-thia fatty acids. Also other lipid metabolizing enzymes are induced. (9) Fatty acid- and cholesterol synthesis is inhibited in hepatocytes. (10) The nuclear receptors PPAR alpha and RXR alpha are induced by 3-thia fatty acids. (11) The induction of enzymes and of PPAR alpha and RXR alpha are increased by dexamethasone and counteracted by insulin. (12) 4-Thia fatty acids inhibit fatty acid oxidation and induce fatty liver in vivo. The inhibition presumably is explained by accumulation of alkylthioacryloyl-CoA in the mitochondria. This metabolite is a strong inhibitor of CPT-II. (13) Alkylthioacrylic acids inhibits both fatty acid oxidation and esterification. Inhibition of esterification presumably follows accumulation of extramitochondrial alkylthioacryloyl-CoA, an inhibitor of microsomal glycerophosphate acyltransferase. (14) 9-Thia stearate is a strong inhibitor of the delta 9-desaturase in liver and 10-thia stearate of dihydrosterculic acid synthesis in trypanosomes. (15) Some attempts to develop thia fatty acids as drugs are also reviewed.

Diabetes and proteolysis: effects on carnitine palmitoyltransferase-I and malonyl-CoA binding

Malonyl-CoA binding and malonyl-CoA inhibition of carnitine palmitoyltransferase-I (CPT-I) were measured in hepatic mitochondria from normal and diabetic rats and in protease-treated mitochondria from fed rats to test the hypothesis that proteolysis represents a mechanism by which diabetes produces changes in the sensitivity of CPT-I to inhibition by malonyl-CoA. As in diabetes, protease treatment increased the apparent Ki values for malonyl-CoA. Palmitoyl-CoA greatly diminished malonyl-CoA specific binding in the mitochondrial system being studied, suggesting strong competition at the malonyl-CoA binding site. Proteolysis decreased capacity for specific binding of malonyl-CoA by 60-80%, but it had no effect on binding affinity. In contrast, the decreased specific binding of malonyl-CoA seen in the diabetic state is accompanied by increased binding affinity. Furthermore, observed Kd values differed from Ki values by a factor of 10 or more, suggesting that measured Kd and Ki may represent different ligand-protein complexes. These data suggest that alterations in inhibition of CPT-I by malonyl-CoA occurring in the diabetic state may involve mechanisms other than simple proteolytic removal of malonyl-CoA binding sites.

Limited trypsin proteolysis renders carnitine palmitoyltransferase insensitive to inhibition by malonyl-CoA in patients with muscle carnitine palmitoyltransferase deficiency

Carnitine palmitoyltransferase (CPT) was studied in muscle homogenates of two patients with muscle CPT deficiency heterozygous for the Ser-113 Leu mutation in the CPT II gene. Total CPT activity was normal in both patients but was almost completely inhibited by malonyl-CoA and Triton X-100 whereas in controls 38% and 58% of total activity remained in the presence of malonyl-CoA and Triton X-100, respectively. The addition of 1% Tween 20 abolished about half of the activity in patients but not in controls. Preincubation of muscle homogenate with trypsin slightly increased the total activity and rendered the activity greatly insensitive to inhibition by malonyl-CoA in both patients and controls. The data support the view that in patients with muscle CPT deficiency both CPT I and II are active, but that CPT II is abnormally accessible to inhibition by malonyl-CoA.

Hepatic carnitine palmitoyltransferase-I has two independent inhibitory binding sites for regulation of fatty acid oxidation

Partial proteolysis of carnitine palmitoyltransferase (CPT-I) in intact mitochondria results in greatly diminished sensitivity to inhibition by its physiological inhibitor, malonyl-CoA, but inhibition by succinyl-CoA and methylmalonyl-CoA was affected to a lesser extent, whereas inhibition by coenzyme A, acetyl-CoA, and propionyl-CoA was not affected at all by proteinase treatment. These data suggested that inhibitors that are coenzyme A esters of short-chain dicarboxylic acids bind to a regulatory malonyl-CoA binding site located on the cytoplasmic face of the mitochondrial outer membrane while coenzyme A esters of monocarboxylic acids and free coenzyme A act at the active site in the mitochondrial intermembrane space. All inhibitors whose potency was altered by proteinase action provided protection against proteinases, whereas other inhibitors did not. Preincubation with the substrates carnitine, palmitoyl-CoA, or coenzyme A prior to proteolysis showed no protective effects against the loss of inhibition or loss of activity; however, preincubation with these substrates enhanced proteinase effects to more seriously diminish activity and inhibition by malonyl-CoA. Proteinases were also found to act on purified mitochondrial outer membranes to reduce inhibition by malonyl-CoA with little effect on activity. Using these outer membrane preparations it was found that the very potent inhibition of CPT-I by the active-site-directed substrate analog (+)-hemipalmitoylcarnitinium was not altered by proteinase treatment; however, inhibition by the malonyl-CoA analog Ro 25-0187, which is a more potent inhibitor than malonyl-CoA, was drastically reduced by proteinase treatment of mitochondrial outer membranes, confirming the different locations for the malonyl-CoA site and the active site. Coenzyme A and malonyl-CoA both act as competitive inhibitors with respect to the acyl-CoA substrate, but coenzyme A lacks cooperative effects seen with malonyl-CoA. For ligand binding to the malonyl-CoA regulatory site, there appears to be a requirement for two carbonyl groups in close juxtaposition, but there is apparently no requirement for the coenzyme A moiety per se. Current evidence, including the recently deduced primary structure for CPT-I, favors the hypothesis that (a) inhibitors of CPT-I may act at two distinct sites, (b) malonyl-CoA binds primarily to a regulatory site that is distinct from the active site of carnitine palmitoyltransferase-I, and (c) the two inhibitory sites are located on opposite sides of the mitochondrial outer membrane.

Myocardial carnitine palmitoyltransferase of the mitochondrial outer membrane is not altered by fasting

The regulation of heart carnitine palmitoyltransferase was studied during the transition to the fasting state. Using decanoyl-CoA or palmitoyl-CoA as substrates, we found no differences in carnitine palmitoyltransferase activity or in its sensitivity to inhibition by malonyl-CoA between fed and fasted states. No cooperativity was seen with either substrate, and the malonyl-CoA-induced shift to sigmoid kinetics normally observed with liver mitochondria was not obvious with heart mitochondria. Analysis of malonyl-CoA inhibition data revealed that mitochondria from rat heart exhibited incomplete maximum inhibition of carnitine palmitoyltransferase (partial inhibition). Homogenization of intact liver mitochondria resulted in a similar pattern of incomplete inhibition and suggested that the malonyl-CoA-insensitive carnitine palmitoyltransferase of the inner membrane was also being assayed. Carnitine palmitoyltransferase in mitochondrial outer membranes, isolated from the heart, proved to be extremely sensitive to malonyl-CoA inhibition and had maximum inhibition values of 90-100% with either decanoyl-CoA or palmitoyl-CoA as substrates, but fasting had no effect. Fasting produced no change in the Ki for malonyl-CoA (0.10 +/- 0.04 and 0.14 +/- 0.02 microM for the fed and fasted groups, respectively). Acyl-CoA chain length specificity was C10 greater than C16 greater than C14 greater than C12 greater than C18 = C8 for carnitine palmitoyltransferase in heart mitochondrial outer membranes. It is concluded that the regulation of carnitine palmitoyltransferase of heart mitochondrial outer membranes differs from regulation of the liver enzyme in three characteristics--the heart enzyme (a) has greater sensitivity to malonyl-CoA inhibition, (b) is resistant to the effects of fasting and (c) has somewhat different acyl-CoA substrate specificity.

Proteinase treatment of intact hepatic mitochondria has differential effects on inhibition of carnitine palmitoyltransferase by different inhibitors

Proteolysis of intact mitochondria by Nagarse (subtilisin BPN') and papain resulted in limited loss of activity of the outer-membrane carnitine palmitoyltransferase, but much greater loss of sensitivity to inhibition by malonyl-CoA. In contrast with a previous report [Murthy & Pande (1987) Proc. Natl. Acad. Sci. U.S.A. 84, 378-382], we found that trypsin had no effect on malonyl-CoA sensitivity. Even when 80% of activity was destroyed by trypsin, there was no difference in the malonyl-CoA sensitivity of the enzyme remaining. Trypsin caused release of the intermembrane-space enzyme adenylate kinase, indicating loss of integrity of the mitochondrial outer membrane, whereas Nagarse and papain caused no release of that enzyme. Citrate synthase was not released by any of the three proteinases, indicating no damage to the mitochondrial inner membrane. When we examined the effects of proteolysis on the inhibition of carnitine palmitoyltransferase by a wide variety of inhibitors having different mechanisms of inhibition, we found differential proteolytic effects that were specific for those inhibitors (malonyl-CoA and hydroxyphenylglyoxylate) that have their inhibitory potencies diminished by changes in physiological state. Both of those inhibitors protected carnitine palmitoyltransferase from the effects of proteolysis, but did not inhibit the proteinases directly. Inhibition by two other inhibitors (DL-2-bromopalmitoyl-CoA and N-benzyladriamycin 14-valerate) was not altered by proteinase treatment, even when most of the enzyme activity had been destroyed. Inhibition by glyburide, which is minimally affected by physiological state, was affected only to a slight extent at the highest concentration of trypsin tested. Proteolysis by Nagarse appeared to produce loss of co-operativity in malonyl-CoA inhibition. The effects of proteolysis are discussed and compared with changes in Ki occurring with changing physiological states.