Kinetic and regulatory properties of pyruvate dehydrogenase from Ehrlich ascites tumor cells.

Pyruvate dehydrogenase was partially purified from Ehrlich ascites tumor cell mitochondria and its kinetic properties were determined. The apparent KM values for pyruvate, nicotinamide adenine dinucleotide, and coenzyme A (CoA) were 46 muM, 110 muM, and 36 muM, respectively. Reduced nicotinamide adenine dinucleotide and acetyl-CoA inhibited enzyme activity competitively to nicotinamide adenine dinucleotide (Ki = 22 muM) and CoA (Ki = 58 muM), respectively. Copurified alpha-ketoglutarate dehydrogenase displayed apparent KM values for alpha-ketoglutarate, nicotinamide adenine dinucleotide, and CoA of 1.25 mM, 67 muM, and 50 muM, respectively. Pyruvate dehydrogenase, but not alpha-ketoglutarate dehydrogenase, was inactivated specifically by adenosine triphosphate with concomitant phosphorylation, and it was reactivated at 10 mM Mg2+ by a protein fraction separated from the complex during purification. The rate of inactivation was decreased by pyruvate or pyrophosphate. The existence of active and inactive forms of pyruvate dehydrogenase in Ehrlich ascites tumor cells was demonstrated. Active form and total activity were determined to be 74.0 +/- 1.5 and 93.6 +/- 4.9 munits/g packed cells (mean +/- S.E., n = 25), respectively.

Kinetic analysis of glycation as a tool for assessing the half-life of proteins.

Glycation of proteins, a common postribosomal modification, proceeds via Amadori rearrangement to yield a stable ketoamine linkage of glucose with the protein. Kinetic analysis of the reaction shows that the amount of glycation at steady state is proportional to the glucose concentration, to protein half-life and to the rate of glycation. Thus, when the rate of glycation is determined in vitro and the extent of glycation of a given protein isolated from euglycemic subjects is measured, the half-life may be calculated. As the in vivo situation may not be simulated accurately in vitro, the calculated values may be considered as approximation. When the calculated values were compared with values reported in the literature fairly good agreement was found except for hemoglobin. Studies on stability of glycated albumin show that ketoamine decreases by about 20% when incubated under physiological conditions for 20 days. The method described by us is especially valuable when turnover of proteins in normal and pathophysiological states are compared. The half-life of plasma low-density lipoprotein is longer in patients with hypothyroidism or a high plasma low-density lipoprotein level than in normal subjects. Extending our studies to tissue proteins we did not find a significant increase in half-life of tendon collagen with age. Basement membrane collagen turnover is faster in diabetic patients in bad metabolic control. Thus, the procedure using fructosylamine as endogenous label of protein offers a method of great potential to study the turnover of human body proteins.

Evidence for a cholera-toxin-sensitive G-protein involved in the regulation of phosphatidylinositol 4-phosphate kinase of rat liver membranes.

Studies on the phosphorylation of inositol phospholipids of rat liver membranes have shown that [gamma S]pppG stimulates 32P incorporation from [gamma-32P]ATP into PI and PIP. This effect appeared specific for stable GTP analogues and could not be reproduced by other compounds. ADP-ribosylation of the membranes with cholera toxin resulted in a large decrease of PIP2 without changes in the level of PIP. Since an activation of phospholipase C can be ruled out, the lowering of PIP2 is explained on the basis of an inhibition of PIP kinase (EC 2.7.1.68). From these results it appears that a novel cholera-toxin-sensitive G-protein is involved in the regulation of PIP kinase.

Purification and partial characterization of phosphatidylinositol-4-phosphate kinase from rat liver plasma membranes. Further evidence for a stimulatory G-protein.

Phosphatidylinositol 4-phosphate (PIP) kinase (E.C. 2.7.1.68) has been purified about 1200-fold from rat liver plasma membranes, taking advantage of affinity chromatography on quercetin-Sepharose as a novel step. The purified PIP kinase showed no contamination by the following enzyme activities: phosphatidylinositol (PI) kinase (EC 2.7.1.67), protein kinase C (EC 2.7.1.-), diacylglycerol kinase (EC 2.7.1.-), phospholipase C (EC 3.1.4.11), protein-tyrosine kinase (EC 2.7.1.112), alkaline phosphatase (EC 3.1.3.1), triphosphoinositide phosphomonoesterase (EC 3.1.3.36), adenylate kinase (EC 2.7.4.3) and cAMP-dependent protein kinase (EC 2.7.1.37). The liver membrane enzyme requires high Mg2+ concentrations with a KM value of 10 mM. Ca2+ or Mn2+ could replace Mg2+ to a certain, though small, extent. Apparent KM values with respect to PIP and ATP were 10 and 65 microM, respectively. GTP was slightly utilized by the kinase as phosphate donor while CTP was not. Quercetin inhibited the enzyme with Ki = 34 microM. Extending our previous observations (Urumow, T. and Wieland, O.H. (1986) FEBS Lett. 207, 253-257 and Urumow, T. and Wieland, O.H. (1988) Biochim. Biophys. Acta 972, 232-238) [gamma S]pppG still stimulated the PIP kinase in extracts of solubilized liver membranes. 20-40% (NH4)2SO4 precipitation of the membrane extracts yielded a fraction that contained the bulk of enzyme activity but did not respond to stimulation by [gamma S]pppG any longer. This was restored by recombination with a protein fraction collected at 40-70% (NH4)2SO4 saturation, presumably containing a GTP binding protein and/or some other factor separated from the PIP kinase. In the reconstituted system [gamma S]pppG stimulated PIP kinase in a concentration dependent manner with maximal activation at 5 microM. This effect was not mimicked by [gamma S]pppA and was blocked by [beta S]ppG. These results strongly support our view that in liver membranes PIP kinase is regulated by a G-protein.

Limited nonenzymatic glucosylation of low-density lipoprotein does not alter its catabolism in tissue culture.

This study examines the effects of various degrees of chemical modification of low-density lipoprotein (LDL) on its catabolism by various cell types. Moderate glucosylation of LDL does not alter its interaction with the high-affinity receptor present on human fibroblasts at concentration of 5-2000 micrograms LDL-cholesterol/ml. Only heavily glucosylated LDL (more than 12 lysine residues glucosylated per apolipoprotein B) or LDL glucosylated in the presence of Na(CN)BH3, i.e., conditions not expected to occur in diabetes, inhibit receptor-mediated internalisation and degradation. Moderately glucosylated LDL is also readily recognized by cultured rat hepatocytes and porcine endothelial cells. Human monocyte-derived macrophages accumulate cholesteryl ester when incubated with acetylated LDL for 12 days but no enhanced cholesteryl ester formation was found when native or glucosylated LDL (3.3 lysines glucosylated per apolipoprotein B) were used.

Separation of the protein-tyrosine kinase and phosphatidylinositol kinase activities of the human placental insulin receptor.

On immunoprecipitation using a specific antiphosphotyrosine antibody, phosphatidylinositol kinase (EC 2.7.1.67) activity was separated from the protein-tyrosine kinase (EC 2.7.1.112) activity of the wheat germ agglutinin (WGA) -purified insulin receptor from human placenta. This clearly indicates that protein-tyrosine kinase and phosphatidylinositol kinase activity do not reside on the same polypeptide chain as previously has been suggested. Quantitatively, the fraction of phosphatidylinositol kinase that was bound to WGA sepharose and eluted together with the insulin receptor amounted to 2% of the Triton X-100 soluble phosphatidylinositol kinase. The apparent Km values of the bound and unbound phosphatidylinositol kinase with respect to PI and ATP were very similar (0.4 and 0.3 mmol/l and 8 and 7 mumol/l, respectively) suggesting that the WGA-bound phosphatidylinositol kinase is not a different enzyme, but rather represents a small portion of the bulk Triton X-100-soluble phosphatidylinositol kinase that is bound to the lectin tightly associated with the insulin receptor. The synthetic polymer (Glu80Tyr20)n, a model substrate of the insulin receptor tyrosine kinase, at 0.5 mmol/l, inhibited phosphatidylinositol kinase of WGA-purified insulin receptor by 70-90%. This inhibition was not overcome by increasing the concentrations of ATP or PI as one would expect if a functional interrelationship of the protein-tyrosine kinase and the phosphatidylinositol kinase would exist.

Increased glycosylation of serum albumin in diabetes mellitus.

The level of glycosylated albumin has been determined in the serum of normal and diabetic subjects after purification of the albumin to apparent homogeneity. The sugar was released from the albumin preparations as 5-hydroxymethylfurfural (HMF) after 24 h of hydrolysis in 2 N acetic acid at 92 degrees C, and it was assayed by the thiobarbituric acid reaction. The mean value for glycosylated albumin, expressed as picomoles of HMF per nanomole of albumin, obtained from 10 normal control and 65 diabetic subjects, was 64 and 124, respectively. The level of glycosylated albumin correlates with the mean blood glucose concentration (n = 55, r = 0.715), but not with the fasting blood sugar concentrations. Moreover, a linear relationship was observed between the amounts of glycosylated hemoglobin (HbAIa-c) and glycosyl-albumin (n = 74, r = 0.88). In an insulin-treated diabetic patient, there was a different temporal relationship between blood glucose concentrations and glycosylated hemoglobin and albumin levels. While HbAIa-c was lowered by only 15% after 20 days, glycosylated albumin had dropped by more than 50% during the same time. Our results indicate that glycosylated albumin might provide a valuable tool to assess the average blood sugar levels between shorter intervals, since the turnover of serum albumin is considerably faster than that of HbAIa-c.

epsilon-Amino-lysine-bound glucose in human tissues obtained at autopsy. Increase in diabetes mellitus.

The level of nonenzymatically epsilon-lysine-bound glucose (NEBG) of tissue proteins obtained at autopsy has been determined with a specific method recently described. The mean values expressed as nmol lysine-bound glucose per mumol phenylalanine were 34 for tendon, 13 for aorta, 16 for coronary artery, 23.5 for femoral nerve, 10 for glomerular basement membrane, and 30 for lung parenchyma for tissues from nondiabetics and 86, 39.5, 34, 60, 29, and 57 for tissues form diabetics, respectively. NEBG was below detection limit in skeletal muscle from either nondiabetic or diabetic subjects. Tendon and aorta NEBG levels correlated well with those from other tissues (r = 0.8-0.94, except r = 0.68 for glomerular basement membrane). A fairly good correlation was also found when NEBG of aorta was compared with the mean blood sugar level determined during the last weeks of hospitalization. Finally, an arbitrary index of diabetic late complication was compared with NEBG of aorta. There appears to be a tendency of an increase of the index of complications along with an increase in tissue glucosylation. The possible role of nonenzymatic glucosylation in the long-term complication of diabetes is discussed.

Clinical utility of nonenzymatically glycosylated blood proteins as an index of glucose control.

This study compares the utility of nonenzymatically glycosylated serum proteins (lys-GSP) to glycosylated hemoglobins (HbA1a-c) as control indices of glucose homeostasis in patients with IDDM. The diagnostic value of lys-GSP was also examined in patients with non-insulin-dependent diabetes mellitus, in subjects with impaired glucose tolerance, and in two patients with insulinoma. The intraindividual fluctuation of lys-GSP in normoglycemic subjects is very small, resulting in an interindividual range of 3.0 +/- 0.3 lysine-bound glucose/mg protein (means +/- SD, N = 52). HbA1a-c with a normal range of 6.4 +/- 0.9% (N = 52) shows greater variability. In IDDM there is no overlap of lys-GSP levels between the normal and the diabetic range at the 95% confidence level. In patients treated with an open-loop insulin delivery system failure of normalization of the glucose balance was clearly discernible by an elevation of GSP. In contrast, in about 40% of the patients with incomplete glycemic control the HbA1a-c levels fell within the normal range. The utility of lys-GSP for diagnosis of diabetes is compared with the results of 60 oral glucose tolerance tests. Two patients suffering from insulinoma displayed decreased lys-GSP values. From these results it appears that determination of lys-GSP represents a more sensitive parameter for long-term control than HbA1a-c and is suitable for monitoring even small fluctuations of blood glucose.

Essential role of coenzyme A in pyruvate dehydrogenase kinase activity.

The rate of phosphorylation and concomitant inactivation of purified pig heart muscle pyruvate dehydrogenase complex by intrinsic kinase (EC 2.7.1.99) is markedly accelerated by the addition of coenzyme A to the incubation medium, showing a half-maximum effect at 1.8 microM. The pantetheine moiety is the effective part of the coenzyme A molecule. The free thiol group is prerequisite for the stimulatory action, acetyl-CoA, benzoyl-CoA or CoAS-SCoA being ineffectual. The thiol's specificity is evidenced by showing that dithiothreitol, 2-mercaptoethanol or glutathione up to 5 mM failed to replace coenzyme A. The possibility is considered that coenzyme A might act as a physiological modifier of pyruvate dehydrogenase kinase activity.

Evidence that (a) serine specific protein kinase(s) different from protein kinase C is responsible for the insulin-stimulated actin phosphorylation by placental membrane.

Partially purified phospholipid- and Ca2+-dependent protein kinase C from human placenta catalyzes the Mg-ATP-dependent phosphorylation of serine residues of purified rabbit muscle actin. Two tryptic [32P]-phosphopeptides were found on HPLC separation. Confirming the previous report by Machicao and Wieland [(1985) Curr. Top. Cell. Regul. 27, 95-105], actin is phosphorylated at serine residues by human placental membranes, and this is stimulated by insulin. In the absence of insulin trypsin treatment yielded eight [32P]phosphopeptides, two of which coincided with the ones due to protein kinase C. Insulin led to the appearance of three new [32P]phosphopeptides. These results suggest that insulin stimulates (a) serine protein kinase(s) which, like protein kinase C, is present in placental membranes.

Stimulation of phosphatidylinositol 4-phosphate phosphorylation in human placenta membranes by GTP gamma S.

In human placenta membranes the rate limiting enzyme for PIP2 formation from PI is PIP kinase. GTP gamma S is shown to activate PIP kinase by increasing Vmax of the enzyme. It is suggested that a guanine nucleotide regulatory protein is involved in the activation of PIP kinase although coupling with a specific receptor is not yet known. Since PIP2 is the preferred substrate of phospholipase C, the possibility exists that an increase of PIP2 due to activation of PIP kinase leads to an enhancement of phospholipase C activity and hence to an increased production of IP3 and DAG.

A small G-protein involved in phosphatidylinositol-4-phosphate kinase activation.

A guanine nucleotide binding protein tentatively designated Gs(PIPK) which activates purified phosphatidylinositol-4-phosphate kinase in vitro, has been partially purified from rat liver membranes and identified as a small G-protein with a molecular mass between 20 and 25 kDa.

Inositol trisphosphate activates pyruvate dehydrogenase in isolated fat cells.

Inositol trisphosphate, when added to permeabilized rat fat cells, led to a several-fold increase of pyruvate dehydrogenase activity due to conversion of the inactive phospho form (PDHb) to the active, dephospho form (PDHa). It is suggested that inositol trisphosphate, probably through intracellular Ca2+-mobilisation, acts as a physiological mediator of insulin for activation of the mitochondrial PDH-complex.

Decrease by glucagon in peroxide generation by isolated hepatocytes.

Exposure of isolated rat liver cells to glucagon or dibutyryl cyclic AMP leads to a prompt decrease in the rate of cellular peroxide generation as evidenced by (i) a reduced rate of [14C]formate oxidation and (ii) a lowered steady-state concentration of catalase Compound I.

Evidence that the insulin receptor-associated protein kinase acts as a phosphatidylinositol kinase.

Insulin receptor preparations from human placenta at various states of purity were shown to catalyze insulin-stimulated phosphate incorporation from [gamma-32P]ATP into endogenous (membrane) and exogenous phosphatidylinositol. Our data suggest that the insulin receptor associated protein (tyrosine) kinase can act as a phosphatidylinositol kinase, and that this mechanism may be of physiological importance in signal transduction of insulin.

Phosphorylation--dephosphorylation of purified insulin receptor from human placenta. Effect of insulin.

The insulin receptor of human placenta even after extensive purification is phosphorylated in the presence of [gamma-32P]ATP and NaF, and is dephosphorylated again on incubation in NaF-free medium. Insulin stimulates phosphate incorporation into the Mr 95 000 subunit probably by activation of the phosphorylation step. Our data suggest that the insulin receptor contains both kinase and phosphatase activities that may control the phosphorylation state of the receptor.

Evidence for phosphorylation of actin by the insulin receptor-associated protein kinase from human placenta.

A purified preparation of rabbit muscle actin (43-kDa protein) is phosphorylated at tyrosine residues when incubated with solubilized insulin receptor from human placenta. Phosphorylation of the 95-kDa receptor subunit and of 43-kDa protein is stimulated by insulin and vanadate, respectively; however, the mode of action of the two agents is distinguishable.

Insulin activates phospholipase C in fat cells: similarity with the activation of pyruvate dehydrogenase.

Phospholipase C (PHL-C) activity determined in homogenates of fat cells treated with physiological concentrations of insulin showed a 2-3-fold increase as compared to controls in the absence of insulin. The changes of PHL-C and pyruvate dehydrogenase (PDH) activity which was measured concomitantly exhibited very similar characteristics as to insulin sensitivity, saturability, time dependence and glucose requirement. Exogenous PHL-C as an activator of PDH in fat cells (Honeyman et al., 1983) also showed a striking similarity to insulin. Our findings strongly suggest that, in fat cells, PHL-C is susceptible to short-term activation by insulin. This effect may be relevant to the mechanism of PDH activation and perhaps to other metabolic actions of insulin.

Endothelial plasma membrane is a glucocorticoid-regulated barrier for the uptake of glucose into the cell.

The effect of glucose concentrations and hormones on glucose consumption, lactate, pyruvate, sorbitol and fructose formation of porcine aortic endothelial cells and human umbilical vein endothelial cells has been investigated. Endothelial cells have a high glycolytic activity which is saturated far below physiologic blood glucose levels (KM apparent less than 1 mmol/l). Glucocorticoids reduce glucose catabolism as a function of their concentration. Insulin, adrenaline, triiodothyronine and glucagon do not influence glucose consumption. Studies with the non-metabolizable analogue 3-O-methyl-D-glucose revealed that glucocorticoids slow down glucose transport into the endothelial cell. The passage of glucose through the cell membrane is the rate-limiting step of glucose utilization. Consequently, the intracellular glucose level is independent of the ambient glucose concentration and endothelial cells do not accumulate sorbitol under hyperglycaemic conditions since the affinity of aldose reductase for glucose is low.