Regulation of acetyl-CoA carboxylase: identity fo sites phosphorylated in intact cells treated with adrenaline and in vitro by cyclic AMP-dependent protein kinase

AMP-activated protein kinase - a journey from 1 to 100 downstream targets

A casual decision made one evening in 1976, in a bar near the Biochemistry Department at the University of Dundee, led me to start my personal research journey by following up a paper that suggested that acetyl-CoA carboxylase (ACC) (believed to be a key regulatory enzyme of fatty acid synthesis) was inactivated by phosphorylation by what appeared to be a novel, cyclic AMP-independent protein kinase. This led me to define and name the AMP-activated protein kinase (AMPK) signalling pathway, on which I am still working 46 years later. ACC was the first known downstream target for AMPK, but at least 100 others have now been identified. This article contains some personal reminiscences of that research journey, focussing on: (i) the early days when we were defining the kinase and developing the key tools required to study it; (ii) the late 1990s and early 2000s, an exciting time when we and others were identifying the upstream kinases; (iii) recent times when we have been studying the complex role of AMPK in cancer. The article is published in conjunction with the Sir Philip Randle Lecture of the Biochemical Society, which I gave in September 2022 at the European Workshop on AMPK and AMPK-related kinases in Clydebank, Scotland. During the early years of my research career, Sir Philip acted as a role model, due to his pioneering work on insulin signalling and the regulation of pyruvate dehydrogenase.

Novel mechanisms of the regulation of protein kinase B in adipocytes; implications for protein kinase A, Epac, phosphodiesterases 3 and 4

Crosstalk between insulin and cAMP signalling pathways has a great impact on adipocyte metabolism. Whilst Protein kinase B (PKB) is a pivotal mediator of insulin action, in some cells regulation of PKB by cAMP has also been demonstrated. Here we provide evidence that, in a phosphatidyl inositol 3-kinase dependent manner, beta3-adrenergic stimulation (using CL316243) in adipocytes induces PKB phosphorylation in the absence of insulin and also potentiates insulin-induced phosphorylation of PKB. Interestingly, insulin- and CL316243-induced PKB phosphorylation was found to be inhibited by pools of cAMP controlled by PDE3B and PDE4 (mainly in the context of insulin), whereas a cAMP pool controlling protein kinase A appeared to mediate stimulation of PKB phosphorylation (mainly in the context of CL316243). Furthermore, an Epac (exchange protein directly activated by cAMP) agonist (8-pCPT-2'-O-Me-cAMP) mimicked the effect of the PDE inhibitors, giving evidence that Epac has an inhibitory effect on PKB phosphorylation in adipocytes. Further, we put the results obtained at the level of PKB in the context of possible downstream signalling components in the regulation of adipocyte metabolism. Thus, we found that overexpression of PKB induced lipogenesis in a PDE3B-dependent manner. Furthermore, overexpression or inhibition of PDE3B was associated with reduced or increased phosphorylation of the key lipogenic enzyme acetyl-CoA carboxylase (ACC), respectively. These PDE3B-dependent effects on ACC correlated with changes in lipogenesis. The Epac agonist, 8-pCPT-2'-O-Me-cAMP, mimicked the effect of PDE3B inhibition on ACC phosphorylation and lipogenesis.

Roles of the AMP-activated and cyclic-AMP-dependent protein kinases in the adrenaline-induced inactivation of acetyl-CoA carboxylase in rat adipocytes

1. In isolated rat adipocytes, acetyl-CoA carboxylase is inactivated by treatment of the cells with adrenaline or the beta-agonist isoproterenol, but not by the alpha-agonist phenylephrine. The inactivation is stable during purification in the presence of protein phosphatase inhibitors, and is associated with a 30-40% increase in the labelling of enzyme isolated from 32P-labelled cells. 2. Increased phosphorylation occurs within peptide T1, which was identified by sequencing to be the peptide Ser-Ser77-Met-Ser79-Gly-Leu-His-Leu-Val-Lys, containing Ser-77 (phosphorylated by cyclic-AMP-dependent protein kinase) and Ser-79 (phosphorylated by the AMP-activated protein kinase). Analysis of the release of radioactivity as free phosphate during Edman degradation of peptide T1 revealed that all of the phosphate was in Ser-79 in both basal and hormone- or agonist-stimulated cells. Treatment of adipocytes with various agents which activate cyclic-AMP-dependent protein kinase by receptor-independent mechanisms (forskolin, cyclic AMP analogues, isobutylmethylxanthine) also produced inactivation of acetyl-CoA carboxylase and increased phosphorylation at Ser-79. 3. The (Rp)-[thio]phosphate analogue of cyclic AMP, which is an antagonist of binding of cyclic AMP to the regulatory subunit of cyclic-AMP-dependent protein kinase, opposes the effect of adrenaline on phosphorylation and inactivation of acetyl-CoA carboxylase. Together with the effects of isobutylmethylxanthine and the stimulatory cyclic AMP analogues, this strongly indicates that cyclic-AMP-dependent protein kinase is an essential component of the signal transduction pathway, although clearly it does not directly phosphorylate acetyl-CoA carboxylase. 4. As shown by okadaic acid inhibition, greater than 95% of the acetyl-CoA carboxylase phosphatase activity in extracts of rat adipocytes or liver is accounted for by protein phosphatase-2A, with less than 5% attributable to protein phosphatase-1. Inhibition of protein phosphatase-1 via phosphorylation of inhibitor-1 is therefore unlikely to be the mechanism by which cyclic-AMP-dependent protein kinase indirectly increases phosphorylation of acetyl-CoA carboxylase. Various other potential mechanisms are discussed.

Synergistic inhibition of hepatic ketogenesis in the presence of insulin and a cAMP antagonist

The separate or combined effects of insulin and the cAMP antagonist, the Rp-diastereomer of adenosine cyclic 3',5'-phosphorothioate (Rp-cAMPS), were examined on fatty acid-stimulated ketogenesis in hepatocytes from normal fasted rats. Addition of 0.4 mM oleic acid or 0.4 mM octanoic acid resulted in a linear increase in ketone production measured over 60 min. When oleic acid was the substrate, incubation with 1 to 30 microns Rp-cAMPS alone or 0.1 to 10 nM insulin alone caused a variable decrease in the production of ketones which did not exceed an average value of 30% in any one experiment. The simultaneous addition of Rp-cAMPS and insulin resulted in a greater than additive inhibition which reached average values between 47-60% when the theoretical combined inhibitory effect of the insulin alone plus the Rp-cAMPS alone was less than 18%. No significant effects of either insulin or Rp-cAMPS, alone or in combination, were seen when octanoic acid was the substrate. These data imply that Rp-cAMPS can potentiate insulin inhibition of hepatic ketogenesis through inhibition of a cAMP-mediated process.

Identification by amino acid sequencing of three major regulatory phosphorylation sites on rat acetyl-CoA carboxylase

We have examined the sites phosphorylated on acetyl-CoA carboxylase by three protein kinases which have been shown to inactivate the enzyme, i.e. cyclic-AMP-dependent protein kinase, acetyl-CoA carboxylase kinase-2 (ACK2, purified from rat mammary gland) and the AMP-activated protein kinase (formerly called acetyl-CoA carboxylase kinase-3, purified from rat liver). Each protein kinase phosphorylates two out of three sites (termed 1-3) which have been established by amino acid sequencing. The two sites phosphorylated by each kinase can be recovered on separate peptides, TC1 and TC2, derived by combined digestion of the native enzyme by trypsin and chymotrypsin: TC1 = Ser-2Ser(P)-Met-3Ser(P)-Gly-Leu; TC2 = Arg-Met-1Ser(P)-Phe- Cyclic-AMP-dependent protein kinase phosphorylates sites 1 and 2 exclusively, whereas the AMP-activated protein kinase phosphorylates sites 1 and 3, plus at least one other minor site. ACK2 phosphorylates site 1 and, more slowly, an unidentified site(s) within TC1. We have also established the structures of the single major phosphopeptides (T1 and C1 respectively) which are recovered by HPLC after acetyl-CoA carboxylase phosphorylated by cyclic-AMP-dependent protein kinase is digested with trypsin or chymotrypsin alone. T1 is related to TC1, and has the structure: Ser-Ser(P)-Met-Ser-Gly-Leu-His-Leu-Val-Lys. C1 is identical with TC2. We have carried out studies on the correlation of the activity of acetyl-CoA carboxylase with the occupancy of sites 1, 2 and 3 during phosphorylation by each of the three protein kinases. The results suggest that phosphorylation of site 3 is primarily responsible for the large decrease in Vmax produced by the AMP-activated protein kinase, while phosphorylation of site 1 may be primarily responsible for the increase in A0.5 for citrate and more modest depression of Vmax produced by cyclic-AMP-dependent protein kinase and ACK2. Our results emphasize that amino acid sequence information is essential in the unequivocal interpretation of data from phosphopeptide mapping experiments and allow a more complete interpretation of previous data on phosphorylation of acetyl-CoA carboxylase in intact cells. They also open the way to experiments which could establish the physiological roles of these protein kinases in the control of fatty acid synthesis.

Insulin and phorbol ester stimulate phosphorylation of acetyl-CoA carboxylase at similar sites in isolated adipocytes. Lack of correspondence with sites phosphorylated on the purified enzyme by protein kinase C

1. The phorbol ester 12-O-tetradecanoyl phorbol 13-acetate (TPA) stimulates fatty acid synthesis from glucose in isolated adipocytes with a half-maximal effect at 0.72 microM. In seven batches of cells, the maximal effects of TPA and insulin were 8.5 +/- 1.1-fold and 27.1 +/- 2.1-fold respectively. Insulin also stimulated fatty acid synthesis from acetate 8.9 +/- 0.5-fold (three experiments), but TPA did not significantly increase fatty acid synthesis from this precursor. 2. In contrast to insulin, TPA treatment of isolated adipocytes did not produce an activation of acetyl-CoA carboxylase which was detectable in crude cell extracts. 3. The total phosphate content of acetyl-CoA carboxylase, isolated from adipocytes in the presence of protein phosphatase inhibitors, was estimated by 32P-labelling experiments to be 2.6 +/- 0.1 (5), 3.4 +/- 0.2 (5), and 3.8 +/- 0.2 (3) mol/mol subunit for enzyme from control, insulin- and TPA-treated cells respectively. Insulin and TPA stimulated phosphorylation within the same two tryptic peptides. 4. Purified acetyl-CoA carboxylase is phosphorylated in vitro by protein kinase C at serine residues which are recovered in three tryptic peptides, i.e. peptide T1, which appears to be identical with the peptide Ser-Ser(P)-Met-Ser-Gly-Leu-His-Leu-Val-Lys phosphorylated by cyclic-AMP-dependent protein kinase, and peptides Ta and Tb, which have the sequences Ile-Asp-Ser(P)-Gln-Arg and Lys-Ile-Asp-Ser(P)-Gln-Arg respectively, and which appear to be derived from a single site by alternative cleavages. None of these correspond to the peptides whose 32P-labelling increase in response to insulin or TPA. Peptides Ta/Tb are not significantly phosphorylated in isolated adipocytes, even after insulin or TPA treatment. Peptide T1 is phosphorylated in isolated adipocytes, but this phosphorylation is not altered by insulin or TPA. 5. These results show that TPA mimics the effect of insulin on phosphorylation, but not activation, of acetyl-CoA carboxylase, i.e. that these two events can be dissociated. In addition, phorbol ester stimulates phosphorylation of acetyl-CoA carboxylase in isolated adipocytes, but this is not catalyzed directly by protein kinase C, and acetyl-CoA carboxylase does not appear to be a physiological substrate for this kinase.

The low activity of acetyl-CoA carboxylase in basal and glucagon-stimulated hepatocytes is due to phosphorylation by the AMP-activated protein kinase and not cyclic AMP-dependent protein kinase

Acetyl-CoA carboxylase purified from isolated hepatocytes is activated dramatically by protein phosphatase treatment, concomitant with a reduction of the phosphate content from 3.7 to 1.1 mol/subunit. Glucagon treatment of the cells produces a further inactivation of the enzyme that is totally reversed by phosphatase treatment, and is associated with an increase in phosphate content of 0.8 mol/subunit, distributed in two peptides which contain the sites phosphorylated in vitro by the cyclic AMP-dependent and AMP-activated protein kinases. Sequencing of these peptides shows that the low activity of acetyl-CoA carboxylase is due to phosphorylation by the AMP-activated protein kinase, and not cyclic AMP-dependent protein kinase, even after glucagon treatment.

Use of rapid gel-permeation chromatography to explore the inter-relationships between polymerization, phosphorylation and activity of acetyl-CoA carboxylase. Effects of insulin and phosphorylation by cyclic AMP-dependent protein kinase

Superose 6 chromatography was used to separate rapidly the polymeric and dimeric forms of acetyl-CoA carboxylase. With preparations of acetyl-CoA carboxylase purified by Sepharose-avidin chromatography, it is shown that citrate promotes polymerization and that the extent of polymerization is diminished, but not eliminated, after phosphorylation by cyclic-AMP-dependent protein kinase. After exposure of rat epididymal adipose tissue to insulin, evidence was obtained for a marked increase in polymerization. The polymeric form, which was active in the absence of citrate, exhibited increased phosphorylation, particularly on a tryptic peptide designated the I-peptide in an earlier study [Brownsey & Denton (1982) Biochem. J. 202, 77-86]. In contrast, in tissue exposed to the beta-agonist isoprenaline, most of the phosphorylated acetyl-CoA carboxylase appeared to be in the dimeric form if chromatography was carried out in the absence of citrate, whereas in the presence of citrate the degree of polymerization was diminished.

Modulation of intracellular cyclic AMP content and rate of lipogenesis in mammary acini in vitro

Relationships between the cyclic AMP content, the rate of lipogenesis and the activity of acetyl-CoA carboxylase in acini prepared from lactating rat mammary tissue were investigated by exposing them to agents that increase their cyclic AMP content in the presence or absence of insulin. The dose-dependent inhibition of lipogenesis by theophylline in acini isolated from fed rats was highly correlated with the induced increases in acinar cyclic AMP content. Cyclic AMP of acini from 24 h-starved lactating rats was more sensitive in its response to theophylline than that in acini from fed animals. Neither forskolin nor a mixture of isoprenaline and Ro 7-2956 were able significantly to change either the rate of lipogenesis or the activity of acetyl-CoA carboxylase in acini from fed rats when added to incubations in vitro, in spite of the large increases in cyclic AMP concentration produced by these agents. Insulin was without effect on the activity of acetyl-CoA carboxylase and on either the basal or isoprenaline-stimulated cyclic AMP content of acini. These results are discussed in terms of the possibility that the rate of lipogenesis and the cyclic AMP content in mammary acini can vary independently of one another and of the activity of acetyl-CoA carboxylase.

Evidence that activation of acetyl-CoA carboxylase by insulin in adipocytes is mediated by a low-Mr effector and not by increased phosphorylation

The activation of acetyl-CoA carboxylase (measured in a crude supernatant fraction) caused by insulin treatment of adipocytes was completely unaffected by the addition of a large amount of highly purified protein phosphatase to the supernatant fraction. Under the same conditions the inhibition of acetyl-CoA carboxylase by adrenaline was totally reversed. Experiments with 32P-labelled adipocytes showed that insulin increased the total phosphorylation of acetyl-CoA carboxylase from 2.7 to 3.5 molecules of phosphate/240 kDa subunit, and confirmed that this increase was partially accounted for by phosphorylation within a specific peptide (the 'I-site' peptide). Protein phosphatase treatment of the crude supernatant fractions removed over 80% of the 32P radioactivity from the enzyme and removed all detectable radioactivity from the I-site peptide. The effect of insulin on acetyl-CoA carboxylase activity, but not the effect on phosphorylation, was lost on purification of the enzyme on avidin-Sepharose. The effect on enzyme activity was also lost if crude supernatant fractions were subjected to rapid gel filtration after treatment under conditions of high ionic strength, similar to those used in the avidin-Sepharose procedure. These results show that, although insulin does increase the phosphorylation of acetyl-CoA carboxylase at a specific site, this does not cause enzyme activation. They suggest instead that activation of the enzyme by insulin is mediated by a tightly bound low-Mr effector which dissociates from the enzyme at high ionic strength.

Mapping of radiolabeled peptides derived from proteolysis of polypeptides bound to nitrocellulose after "Western" blotting

Sections of nitrocellulose containing bound 32P-labeled polypeptides were excised from "Western" blots and exhaustively digested by trypsin in order to analyze the distribution of phosphorylation sites between the products of limited proteolysis of the multifunctional protein CAD. Using the criterion of analytical isoelectric focusing, the 32P-peptides obtained by this method were found to be similar, although not identical, to peptides obtained by a more conventional digestion of trichloroacetic acid precipitates. Digestion on Western blots is more straightforward than electrophoretic elution of individual gel slices, gives better recoveries than direct digestion of gel slices, and is particularly suitable for peptide mapping of small peptides which bind to nitrocellulose but would diffuse out of polyacrylamide gels during the commonly used fixing and staining procedures.

The role of acetyl-CoA carboxylase phosphorylation in the control of mammary gland fatty acid synthesis during the starvation and re-feeding of lactating rats

Activation of acetyl-CoA carboxylase during incubation of crude extracts of lactating rat mammary gland with Mg2+ and citrate can be blocked by NaF, suggesting that it represents a dephosphorylation of the enzyme. The greater extent of activation in extracts from 24 h-starved rats (200%) compared with fed controls (70%) implies that the decrease in acetyl-CoA carboxylase activity in response to 24 h starvation may involve increased phosphorylation of the enzyme. Acetyl-CoA carboxylase was purified from the mammary glands of lactating rats in the presence of protein phosphatase inhibitors by avidin-Sepharose chromatography. Starvation of the rats for 24 h increased the concentration of citrate giving half-maximal activation by 75%, and decreased the Vmax. of the purified enzyme by 73%. This was associated with an increase in the alkali-labile phosphate content from 3.3 +/- 0.2 to 4.5 +/- 0.4 mol/mol of enzyme subunit. Starvation of lactating rats for 6 h, or short-term insulin deficiency induced by streptozotocin injection, did not effect the kinetic parameters or the phosphate content of acetyl-CoA carboxylase purified from mammary glands. The effects of 24 h starvation on the kinetic parameters and phosphate content of the purified enzyme were completely reversed by re-feeding for only 2.5 h. This effect was blocked if the animals were injected with streptozotocin before re-feeding, suggesting that the increase in plasma insulin that occurs on re-feeding was responsible for the activation of the enzyme. The effects of re-feeding 24 h-starved rats on the kinetic parameters and phosphate content of acetyl-CoA carboxylase could be mimicked by treating enzyme purified from 24 h-starved rats with protein phosphatase-2A in vitro. Our results suggest that, in mammary glands of 24 h-starved lactating rats, insulin brings about a dephosphorylation of acetyl-CoA carboxylase in vivo, which may be at least partly responsible for the reactivation of mammary lipogenesis in response to re-feeding.

Characterization of the phosphorylation of rat mammary ATP-citrate lyase and acetyl-CoA carboxylase by Ca2+ and calmodulin-dependent multiprotein kinase and Ca2+ and phospholipid-dependent protein kinase

ATP-citrate lyase and acetyl-CoA carboxylase purified from lactating rat mammary gland are phosphorylated stoichiometrically by the calmodulin-dependent multiprotein kinase from rabbit skeletal muscle. The reactions are completely dependent on the presence of both Ca2+ and calmodulin. ATP-citrate lyase and acetyl-CoA carboxylase are also phosphorylated stoichiometrically by the Ca2+- and phospholipid-dependent protein kinase (protein kinase C) purified from bovine brain. Phosphorylation of these substrates is stimulated 6-fold and 40-fold respectively by Ca2+ and phosphatidylserine. The calmodulin-dependent and phospholipid-dependent protein kinases phosphorylate the same serine residue on ATP-citrate lyase that is phosphorylated by cyclic-AMP-dependent protein kinase. The sequence of the tryptic peptide containing this site on the mammary enzyme is identical with the sequence of the peptide containing the site on ATP-citrate lyase that is phosphorylated in isolated hepatocytes in response to insulin and/or glucagon. The calmodulin-dependent, phospholipid-dependent and cyclic-AMP-dependent protein kinases phosphorylate distinct sites on acetyl-CoA carboxylase. However, one of the three phosphorylated tryptic peptides derived from enzyme treated with the phospholipid-dependent kinase is identical with the major phosphopeptide (T1) derived from enzyme treated with cyclic-AMP-dependent protein kinase. Phosphorylation of acetyl-CoA carboxylase by the phospholipid-dependent protein kinase inactivates acetyl-CoA carboxylase in a similar manner to cyclic-AMP-dependent protein kinase. With either protein kinase slightly greater phosphorylation and inactivation is seen after pretreatment of acetyl-CoA carboxylase with protein phosphatase-2A, but the effects of the protein phosphatase treatment are not completely reversed. Inactivation by the phospholipid-dependent protein kinase is Ca2+- and phospholipid-dependent, is reversed by protein phosphatase-2A, and correlates with the degree of phosphorylation. The relevance of these findings to insulin- and growth-factor-promoted phosphorylation of ATP-citrate lyase and acetyl-CoA carboxylase in intact cells is discussed.

Both insulin and epidermal growth factor stimulate lipogenesis and acetyl-CoA carboxylase activity in isolated adipocytes. Importance of homogenization procedure in avoiding artefacts in acetyl-CoA carboxylase assay

Epidermal growth factor (EGF) stimulates lipogenesis by 3-4-fold in isolated adipocytes, with a half-maximal effect at 10 nM-EGF. In the same batches of cells insulin stimulated lipogenesis by 15-fold. Freezing and prolonged homogenization of adipocytes results in release of large quantities of pyruvate carboxylase from broken mitochondria, and sufficient pyruvate can be carried through into assays for this enzyme to cause significant interference with assays of acetyl-CoA carboxylase in crude adipocyte extracts. This may account for the high amount of citrate-independent acetyl-CoA carboxylase activity reported to be present in adipocyte extracts in some previous publications. This problem may be eliminated by homogenizing very briefly without freezing. By using the modified homogenization procedure, EGF treatment of adipocytes was shown to produce an effect on acetyl-CoA carboxylase activity almost identical with that of insulin. Both messengers increase Vmax. without significant effect on the Ka for the allosteric activator, citrate.

Both insulin and epidermal growth factor stimulate fatty acid synthesis and increase phosphorylation of acetyl-CoA carboxylase and ATP-citrate lyase in isolated hepatocytes

Insulin and EGF cause identical stimulation (congruent to 40%) of fatty acid synthesis in hepatocytes isolated from rats which have been starved and then refed a low-fat diet. In both cases this stimulation is associated with increased phosphorylation of ATP-citrate lyase and of a specific site on acetyl-CoA carboxylase. However, the altered phosphorylation of acetyl-CoA carboxylase is not associated with a change in kinetic parameters which is detectable in the purified enzyme. Whatever the mechanism involved, stimulation of fatty acid synthesis by growth factors may have a role in providing new phospholipid for growth of membranes.

Evidence that glucagon-mediated inhibition of acetyl-CoA carboxylase in isolated adipocytes involves increased phosphorylation of the enzyme by cyclic AMP-dependent protein kinase

The kinetic parameters and phosphorylation state of acetyl-CoA carboxylase were analysed after purification of the enzyme by avidin--Sepharose chromatography from extracts of isolated adipocytes treated with glucagon or adrenaline. The results provide evidence that the mechanism of inhibition of acetyl-CoA carboxylase in adipocytes treated with glucagon [Zammit & Corstorphine (1982) Biochem. J. 208, 783-788] involves increased phosphorylation of the enzyme. Hormone treatment had effects on the kinetic parameters of the enzyme similar to those of phosphorylation of the enzyme in vitro by cyclic AMP-dependent protein kinase. Glucagon treatment of adipocytes led to increased phosphorylation of acetyl-CoA carboxylase in the same chymotryptic peptide as that containing the major site phosphorylated on the enzyme by purified cyclic AMP-dependent protein kinase in vitro [Munday & Hardie (1984) Eur. J. Biochem. 141, 617-627]. The dose--response curves for inhibition of enzyme activity and increased phosphorylation of the enzyme were very similar, with half-maximal effects occurring at concentrations of glucagon (0.5-1 nM) which are close to the physiological range. In general, the patterns of increased 32P-labelling of chymotryptic peptides induced by glucagon or adrenaline were similar, although there were quantitative differences between the effects of the two hormones on individual peptides. The results are discussed in terms of the possible roles of cyclic AMP-dependent and -independent protein kinases in the regulation of acetyl-CoA carboxylase activity and of lipogenesis in white adipose tissue.

Short-term regulation of acetyl CoA carboxylase: is the key enzyme in long-chain fatty acid synthesis regulated by an existing physiological mechanism?

Acetyl CoA carboxylase, the rate-limiting enzyme in regulating fatty acid synthesis, is thought to be controlled by allosteric effectors, its state of aggregation, covalent modulation and protein inhibitors. It is still obscure whether citrate, a positive allosteric effector, and long-chain fatty acyl CoA esters, negative allosteric effectors, function physiologically to regulate acetyl CoA carboxylase activity. New evidence from several laboratories reveals that the covalent phosphorylation may not involve regulation of acetyl CoA carboxylase activity. Protein inhibitors from liver cytosol and a peptide from fat cells were found to regulate acetyl CoA carboxylase both in vivo and in vitro. Coenzyme A, guanosine 5-monophosphate and phosphatidylinositol 4,5-bisphosphate may have an indirect effect, but certainly no direct involvement, on carboxylase activity.

Isolation of three cyclic-AMP-independent acetyl-CoA carboxylase kinases from lactating rat mammary gland and characterization of their effects on enzyme activity

Three cyclic AMP-independent acetyl-CoA carboxylase kinases (A, B1 and B2) have been isolated from lactating rat mammary gland, using phosphocellulose chromatography, high performance gel filtration, and affinity chromatography on casein-Sepharose and phosvitin-Sepharose. These protein kinases have been identified with previously described kinases by the following criteria. Kinase A phosphorylates the same sites on rabbit mammary acetyl-CoA carboxylase as acetyl-CoA carboxylase kinase 2, which was originally described as a contaminant of rabbit mammary acetyl-CoA carboxylase purified by the poly(ethylene glycol)procedure. Kinase A will henceforth be referred to as acetyl-CoA carboxylase kinase-2. Kinase B1 has been identified with casein kinase II by its heparin sensitivity, elution behaviour on phosphocellulose, molecular mass, substrate specificity and subunit composition. Kinase B2 has been identified with casein kinase I by its elution behaviour on phosphocellulose, molecular mass, substrate specificity and subunit composition. The three kinases phosphorylate distinct sites on acetyl-CoA carboxylase. Phosphorylation by either casein kinase I or II does not affect enzyme activity. However, acetyl-CoA carboxylase kinase 2 inactivates acetyl-CoA carboxylase reversibly, in an identical manner to cyclic-AMP-dependent protein kinase, and phosphorylates sites located on identical peptides. Acetyl-CoA carboxylase kinase-2 can, however, be distinguished from the free catalytic subunit of cyclic-AMP-dependent protein kinase by its molecular mass, its substrate specificity, its elution behaviour on phosphocellulose, and its complete lack of sensitivity to the protein inhibitor of cyclic-AMP-dependent protein kinase. We also present evidence that phosphorylation of acetyl-CoA carboxylase by cyclic-AMP-dependent protein kinase occurs directly and not via a bicyclic cascade system as proposed by other laboratories.

Cyclic AMP phosphodiesterase and cyclic GMP phosphodiesterase activities of rat mammary tissue

Cyclic nucleotide phosphodiesterase activity in mammary tissue from rats in midlactation was resolved by DEAE-cellulose chromatography into three functionally distinct fractions: a Ca2+/calmodulin-stimulated cyclic GMP phosphodiesterase, a cyclic GMP-stimulated low-affinity cyclic nucleotide phosphodiesterase, and a high-affinity cyclic AMP-specific phosphodiesterase. The absolute activities and relative proportions of high- and low-affinity enzymes resemble those found, for example, in liver, as distinct from those in excitable tissues. Three functional characteristics are described which are peculiar to mammary-tissue phosphodiesterases. Firstly, the concentration of free Ca2+ required to achieve half-maximal activation of the Ca2+/calmodulin-stimulated phosphodiesterase is somewhat higher than for the analogous enzyme in other tissues; secondly, the activity of this enzyme towards cyclic AMP relative to that towards cyclic GMP is unusually low, and thirdly, the low-affinity cyclic nucleotide phosphodiesterase is inhibited by low concentrations of free Ca2+.

Glucagon inhibits fatty acid synthesis in isolated hepatocytes via phosphorylation of acetyl-CoA carboxylase by cyclic-AMP-dependent protein kinase

If isolated rat hepatocytes are preincubated for 90 min before addition of hormone, glucagon causes a significant (50%) decrease in fatty acid synthesis without concomitant large decreases in the cellular content of the allosteric activator, citrate. We present evidence that this inhibition can be entirely accounted for by the phosphorylation of the rate-limiting enzyme, acetyl-CoA carboxylase, by cyclic-AMP-dependent protein kinase. In particular: (1) the effect is associated with a 50% decrease in acetyl-CoA carboxylase activity (measured at physiological citrate concentration) which survives purification of the enzyme; (2) the effect is associated with a selective increase in the phosphorylation of a chymotryptic peptide (peptide 1) which is identical to the peptide containing the major site phosphorylated on purified acetyl-CoA carboxylase by cyclic-AMP-dependent protein kinase; (3) the effects of glucagon on the kinetic parameters of the enzyme are very similar to the effect of phosphorylation of the purified enzyme, i.e. a decrease in V and an increase in Ka for citrate; and (4) all of these effects occur at physiological concentrations of glucagon identical to those producing inhibition of fatty acid synthesis.

Studies on insulin-stimulated phosphorylation of acetyl-CoA carboxylase, ATP citrate lyase and other proteins in rat epididymal adipose tissue. Evidence for activation of a cyclic AMP-independent protein kinase

Protein kinase activity in high-speed supernatant fractions prepared from rat epididymal adipose tissue previously incubated in the absence or presence of insulin was investigated by following the incorporation of 32P from [gamma-32P]ATP into phosphoproteins separated by sodium dodecyl sulphate/polyacrylamide-gel electro-phoresis. Incorporation of 32P into several endogenous proteins in the supernatant fractions from insulin-treated tissue was significantly increased. These included acetyl-CoA carboxylase and ATP citrate lyase (which exhibit increased phosphorylation within fat-cells exposed to insulin), together with two unknown proteins of subunit Mr 78000 and 43000. The protein kinase activity increased by insulin was distinct from cyclic AMP-dependent protein kinase, was not dependent on Ca2+ and was not appreciably affected by dialysis or gel filtration. The rate of phosphorylation of added purified fat-cell acetyl-CoA carboxylase and ATP citrate lyase was also increased by 60-90% in high-speed-supernatant fractions prepared from insulin-treated tissue. No evidence for any persistent changes in phosphoprotein phosphatase activity was found. It is concluded that insulin action on acetyl-CoA carboxylase, ATP citrate lyase and other intracellular proteins exhibiting increased phosphorylation involves an increase in cyclic AMP-independent protein kinase activity in the cytoplasm. The possibility that the increase reflects translocation from the plasma membrane, perhaps after phosphorylation by the protein tyrosine kinase associated with insulin receptors, is discussed.

Phosphorylation of different sites of acetyl CoA carboxylase by ATP-citrate lyase kinase and cyclic AMP-dependent protein kinase

Native acetyl CoA carboxylase was phosphorylated by catalytic subunit of cyclic AMP-dependent protein kinase and ATP-citrate lyase kinase to 1 and 0.5 mol/subunit respectively. Both protein kinases added together increased acetyl CoA carboxylase phosphorylation additively. Partial proteolysis of 32P-acetyl CoA carboxylase followed by electrophoretic analysis showed that the 32P-phosphopeptides generated from acetyl CoA carboxylase phosphorylated with lyase kinase were different from the peptides obtained from the enzyme phosphorylated by cyclic AMP-dependent protein kinase. Mapping of tryptic 32P-phosphopeptides by high performance liquid chromatography showed that the major phosphopeptides phosphorylated by ATP-citrate lyase kinase were different from the major phosphopeptides phosphorylated by cyclic AMP-dependent protein kinase. The results suggest that at least one different site on acetyl CoA carboxylase is preferentially phosphorylated by each protein kinase.

The role of phosphorylation in the regulation of fatty acid synthesis by insulin and other hormones

Insulin stimulates fatty acid synthesis in white and brown fat cells as well as in liver and mammary tissue. Hormones that increase cellular cyclic AMP concentrations inhibit fatty acid synthesis, at least in white adipose tissue and liver. These changes in fatty acid synthesis occur within minutes. In white fat cells, they are brought about not only by changes in glucose transport but also changes in the activities of pyruvate kinase, pyruvate dehydrogenase and acetyl-CoA carboxylase. The basis of the alterations in pyruvate kinase activity in fat cells is not understood. Unlike the liver isoenzyme, the isoenzyme present in fat cells does not appear to be phosphorylated either in the absence or presence of hormones. The changes in pyruvate dehydrogenase activity in fat cells are undoubtedly due to changes in phosphorylation of the alpha subunits. Insulin appears to act by causing the parallel dephosphorylation of all three sites. The persistence of the effect of insulin during the preparation and subsequent incubation of mitochondria has allowed the demonstration that insulin acts mainly by stimulating pyruvate dehydrogenase phosphatase rather than inhibiting the kinase. Acetyl-CoA carboxylase within fat cells is phosphorylated on a number of different sites. The exposure of cells to insulin leads to activation of the enzyme and this is associated with increased phosphorylation of a specific site on the enzyme. Exposure to adrenalin, which results in a marked diminution in activity, also causes a small increase in the overall level of phosphorylation, but this increase is due to an enhanced phosphorylation of different sites; probably those phosphorylated by cyclic-AMP-dependent protein kinase. Acetyl-CoA carboxylase is one of a number of proteins in fat cells that exhibit increased phosphorylation with insulin. Others include ATP-citrate lyase, the ribosomal protein S6, the beta subunit of the insulin receptor and a heat and acid stable protein of Mr 22000. Changes in phosphorylation of ATP-citrate lyase do not appear to result in any appreciable changes in catalytic activity. A central aspect of insulin action may be the activation and perhaps release of a membrane-associated protein kinase. Plasma membranes from fat cells have been shown to contain a cyclic-nucleotide-independent kinase able to phosphorylate and activate acetyl-CoA carboxylase. Furthermore, high-speed supernatant fractions from cells previously exposed to insulin contain elevated levels of the same or similar kinase activity capable of phosphorylating both ATP-citrate lyase and acetyl-CoA carboxylase.

The protein phosphatases involved in cellular regulation. 6. Measurement of type-1 and type-2 protein phosphatases in extracts of mammalian tissues; an assessment of their physiological roles

Methods were developed for quantifying protein phosphatases-1, 2A, 2B and 2C in cell extracts, and these procedures were exploited to determine their tissue and subcellular distributions. In addition, the contribution of each enzyme to the total protein phosphatase activity in skeletal muscle and liver extracts towards nine proteins involved in the control of glycogen metabolism, glycolysis/gluconeogenesis, fatty acid synthesis and cholesterol synthesis was assessed. Each protein phosphatase was present at significant concentrations in skeletal muscle, heart muscle, liver, brain and adipose tissue, although the relative amounts differed considerably. In skeletal muscle, protein phosphatase-1 was the major enzyme acting on phosphorylase, glycogen synthase and phosphorylase kinase (beta-subunit), and thus was the major protein phosphatase responsible for the inactivation of glycogenolysis and stimulation of glycogen synthesis. This idea was reinforced by the observation that 50% of the protein phosphatase-1 activity was associated with the protein-glycogen complex. In the liver, protein phosphatases-1, 2A and 2C each appear to play a role in the regulation of glycogen metabolism. Protein phosphatase-1 accounted for a significant fraction of the total potential activity towards phosphorylase and glycogen synthase, and was the major phosphorylase kinase (beta-subunit) phosphatase of this tissue. In addition, it was the only protein phosphatase present in the protein-glycogen complex. Protein phosphatase 2A was also a major phosphorylase phosphatase and glycogen synthase phosphatase in this tissue. Protein phosphatase 2C was a significant glycogen synthase phosphatase in the liver, but had negligible activity toward phosphorylase or phosphorylase kinase (beta-subunit). In the absence of Ca2+, protein phosphatase 2A was the major phosphorylase kinase (alpha-subunit) phosphatase and the only inhibitor-1 phosphatase, in skeletal muscle or liver. In the presence of Ca2+, protein phosphatase 2B accounted for most of the activity towards these substrates. Protein phosphatase 2A was the major enzyme acting on L-pyruvate kinase, ATP-citrate lyase and acetyl-CoA carboxylase in rat liver, suggesting an important role in the regulation of glycolysis/gluconeogenesis and fatty acid synthesis. Protein phosphatase 2C was the major enzyme acting on hydroxymethylglutaryl-CoA (HMG-CoA) reductase and HMG-CoA reductase kinase, suggesting an important role in the regulation of cholesterol synthesis. However, the observation that 20% of the protein phosphatase-1 in liver was associated with the microsomal fraction suggests that this enzyme may also be involved in regulating HMG-CoA reductase, which is tightly associated with microsomes. The activity of protein phosphatase-1 in dilute skeletal muscle and liver extracts was just as sensitive to inhibitor-1 and inhibitor-2 as the purified enzyme. In concentrated extracts, higher concentrations of the inhibitor proteins were required and the inhibition was time-dependent...

The protein phosphatases involved in cellular regulation. 3. Fatty acid synthesis, cholesterol synthesis and glycolysis/gluconeogenesis

The nature of the protein phosphatases involved in the regulation of glycolysis/gluconeogenesis, fatty acid synthesis and cholesterol synthesis in rat liver has been investigated using L-pyruvate kinase, ATP-citrate lyase, acetyl-CoA carboxylase and hydroxymethylglutaryl-CoA reductase as substrates. The results show that protein phosphatases-1, 2A and 2C are the only significant protein phosphatases in rat liver acting on these four substrates. The relationship of these three enzymes to other protein phosphatases described in the literature is discussed.

Inhibition of acetyl-CoA carboxylase activity in isolated rat adipocytes incubated with glucagon. Interactions with the effects of insulin, adrenaline and adenosine deaminase

1. Adipocytes isolated from epididymal fat-pads of fed rats were incubated with different concentrations of glucagon, insulin, adrenaline and adenosine deaminase, and the effects of these agents on the ;initial' activity of acetyl-CoA carboxylase in the cells were studied. 2. Glucagon (at concentrations between 0.1 and 10nm) inhibited acetyl-CoA carboxylase activity. Maximal inhibition was approx. 70% of the ;control' activity in the absence of added hormone, and the concentration of hormone required for half-maximal inhibition was 0.3-0.5nm-glucagon. 3. Incubation of cells with adenosine deaminase resulted in a similar inhibition of acetyl-CoA carboxylase activity. Preincubation of adipocytes with adenosine deaminase did not alter either the sensitivity of carboxylase activity to increasing concentrations of glucagon or the maximal extent of inhibition. 4. Adrenaline inhibited acetyl-CoA carboxylase to the same extent as glucagon. Preincubation of the cells with glucagon did not alter the sensitivity of enzyme activity to adrenaline or the degree of maximal inhibition. 5. Insulin activated the enzyme by 70-80% of ;control' activity. Preincubation of the cells with glucagon did not alter the concentration of insulin required to produce half the maximal stimulatory effect (about 12muunits of insulin/ml). The effects of insulin and glucagon appeared to be mediated completely independently, and were approximately quantitatively similar but opposite. These characteristics resulted in the mutual cancellation of the effects of the two hormones when they were both present at equally effective concentrations. 6. The implications of these findings with regard to current concepts about the mechanism of regulation of acetyl-CoA carboxylase and to the regulation of the enzyme in vivo are discussed.

The role of protein phosphorylation in neural and hormonal control of cellular activity

Protein phosphorylation is now recognized to be the major general mechanism by which intracellular events in mammalian tissues are controlled by external physiological stimuli. However, only recently has the idea that different cellular functions are controlled by common protein kinases and protein phosphatases started to gain widespread acceptance. Thus there is an integrated network of regulatory pathways, mediated by phosphorylation-dephosphorylation, that allows diverse cellular events to be coordinated by neural and hormonal stimuli. The evidence that supports this concept is reviewed, with emphasis on the role of protein phosphorylation in enzyme regulation.

Evidence that insulin activates fat-cell acetyl-CoA carboxylase by increased phosphorylation at a specific site

1. A new rapid method for the purification of fat-cell acetyl-CoA carboxylase is described; the key step is sedimentation after specific polymerization by citrate. 2. Incubation of epididymal fat-pads or isolated fat-cells with insulin or adrenaline leads to a rapid increase or decrease respectively in the activity of acetyl-CoA carboxylase measured in fresh tissue extracts. The persistence of the effect of insulin through high dilution of tissue extracts and through purification involving precipitation with (NH4)2SO4 suggests that the enzyme undergoes a covalent modification after exposure of intact tissue to the hormone. The opposed effects of insulin and adrenaline are not adequately explained through modification of a common site on acetyl-CoA carboxylase, since these hormones bring about qualitatively different alterations in the kinetic properties of the enzyme measured in tissue extracts. 3. The state of phosphorylation of acetyl-CoA carboxylase within intact fat-cells exposed to insulin was determined, and results indicate a small but consistent rise in overall phosphorylation of the Mr-230000 subunit after insulin treatment. 4. Acetyl-CoA carboxylase from fat-cells previously incubated in medium containing [32P]phosphate was purified by immunoprecipitation and then digested with performic acid and trypsin before separation of the released phosphopeptides by two-dimensional analysis. Results obtained show that the exposure of fat-cells to insulin leads to a 5-fold increase in incorporation of 32P into a peptide which is different from those most markedly affected after exposure of fat-cells to adrenaline. 5. These studies indicate that the activation of acetyl-CoA carboxylase in cells incubated with insulin is brought about by the increased phosphorylation of a specific site on the enzyme, possibly catalysed by the membrane-associated cyclic AMP-independent protein kinase described by Brownsey, Belsham & Denton [(1981) FEBS Lett. 124, 145-150].