Membrane Sialic Acid and the Mechanism of Insulin Action in Adipose Tissue Cells

Differential glycosylation of tissue non-specific alkaline phosphatase in mesenchymal stromal cells differentiated into either an osteoblastic or adipocytic phenotype

It has long been known that tissue non-specific alkaline phosphatase (TNAP) is essential for the correct formation of bone, as altered expression or function of this enzyme results in hypophosphatasia, a disease characterised by compromised bone structure, density and strength. However, recent evidence strongly suggests that the enzyme also has a role in lipid accrual and adipogenesis, a function that seems far removed from bone formation. Given that mesenchymal stromal cells (MSCs) are progenitors of both osteoblasts and adipocytes, the question arises of how TNAP is regulated to potentially have a different function when MSCs undergo either osteogenesis or adipogenesis. As the primary protein sequence is unchanged for the enzyme during both types of differentiation, any differences in function must be attributed to post-translational modification and/or localisation. We therefore examined the location of TNAP in bone- or adipose-derived MSCs differentiated into an adipocytic phenotype and compared the glycosylation state of the enzyme in MSCs differentiated into either osteoblasts or adipocytes. TNAP was found to co-locate with perilipin around lipid droplets in MSCs from bone, subcutaneous- and visceral adipose tissue during adipocytic differentiation. Treatment of TNAP with wheat germ lectin followed by electrophoresis showed minor differences in glycosylation between the phosphatase isolated from cells from these tissues, whereas electrophoresis after neuraminidase digestion highlighted differential glycosylation between cell types and during adipogenesis and osteoblastogenesis. This infers that post-translational modification of TNAP is altered during differentiation and is dependent on the eventual phenotype of the cells.

Sialic Acid: its importance to platelet function in health and disease

Platelets are known to contain sialic acid, most of which resides on glycoprotein carbohydrate side-chains. A large proportion of the platelet sialic acid is neuraminidase-labile. The sialic acid of human platelets is of the N-acetyl neuramininic acid type and is implicated in platelet functional events, such as aggregation and adhesion as well as contributing to the platelet's electrophoretic mobility. In this respect the platelet has an overall electronegative charge which is predominately the result of the carboxyl group of the sialic acid moiety. Platelet subpopulations have been isolated with differing amounts of sialic acid and cleavage of platelet sialic acid by neuraminidase is associated with loss of platelets from the circulation. This has relevance to platelet aging in vivo where removal of platelet sialic acid may reveal 'senescence antigens' which could facilitate platelet 'uptake' by the reticuloendothelial system. The importance of platelet sialic acid status has been described in certain platelet disorders such as some thrombocytopenias as well as particular disease states including malignancies and atherosclerosis.

Historical perspectives in fat cell biology: the fat cell as a model for the investigation of hormonal and metabolic pathways

For many years, there was little interest in the biochemistry or physiology of adipose tissue. It is now well recognized that adipocytes play an important dynamic role in metabolic regulation. They are able to sense metabolic states via their ability to perceive a large number of nervous and hormonal signals. They are also able to produce hormones, called adipokines, that affect nutrient intake, metabolism and energy expenditure. The report by Rodbell in 1964 that intact fat cells can be obtained by collagenase digestion of adipose tissue revolutionized studies on the hormonal regulation and metabolism of the fat cell. In the context of the advent of systems biology in the field of cell biology, the present seems an appropriate time to look back at the global contribution of the fat cell to cell biology knowledge. This review focuses on the very early approaches that used the fat cell as a tool to discover and understand various cellular mechanisms. Attention essentially focuses on the early investigations revealing the major contribution of mature fat cells and also fat cells originating from adipose cell lines to the discovery of major events related to hormone action (hormone receptors and transduction pathways involved in hormonal signaling) and mechanisms involved in metabolite processing (hexose uptake and uptake, storage, and efflux of fatty acids). Dormant preadipocytes exist in the stroma-vascular fraction of the adipose tissue of rodents and humans; cell culture systems have proven to be valuable models for the study of the processes involved in the formation of new fat cells. Finally, more recent insights into adipocyte secretion, a completely new role with major metabolic impact, are also briefly summarized.

Insulin receptors and insulin receptor antibodies: structure-function relationships

The insulin receptor has been purified by affinity chromatography and studied by affinity-labelling techniques. It appears to be a disulphide-linked heterotetramer, (alpha beta)2, composed of two copies of a 135,000 Mr subunit (alpha), and two copies of a 90,000 Mr subunit (beta). Beta is readily proteolysed to generate a 45,000 Mr fragment (beta 1). Alpha, beta and beta 1 all contain sialic acid and are, therefore, probably all exposed on the external surface of the membrane. Although alpha is predominantly labelled in affinity-labelling studies, beta and beta 1 can also be labelled. Therefore, alpha, beta and beta 1 are all in proximity to the insulin-binding site and may contain part of the binding site. Antibodies have been prepared against the intact, purified receptor and against the isolated alpha subunit. Both antibodies directly interact with the insulin receptor as indicated by their ability to immunoprecipitate the receptor. Neither antibody, however, directly competes with insulin binding. Therefore, they are probably directed against regions of the receptor distinct from the insulin-binding site. In spite of this, these antibodies have a wide range of insulin-like activities.

Glycosylation and the activation of proteinase-activated receptor 2 (PAR(2)) by human mast cell tryptase

1. Human mast cell tryptase appears to display considerable variation in activating proteinase-activated receptor 2 (PAR(2)). We found tryptase to be an inefficient activator of wild-type rat-PAR(2) (wt-rPAR(2)) and therefore decided to explore the factors that may influence tryptase activation of PAR(2). 2. Using a 20 mer peptide (P20) corresponding to the cleavage/activation sequence of wt-rPAR(2), tryptase was as efficient as trypsin in releasing the receptor-activating sequence (SLIGRL.). However, in the presence of either human-PAR(2) or wt-r PAR(2) expressing cells, tryptase could only activate PAR(2) by releasing SLIGRL from the P20 peptide, suggesting that PAR(2) expressed on the cells was protected from tryptase activation. 3. Three approaches were employed to test the hypothesis that PAR(2) receptor glycosylation restricts tryptase activation. (a) pretreatment of wt-rPAR(2) expressing cells or human embryonic kidney cells (HEK293) with vibrio cholerae neuraminidase to remove oligosaccharide sialic acid, unmasked tryptase-mediated PAR(2) activation. (b) Inhibiting receptor glycosylation in HEK293 cells with tunicamycin enabled tryptase-mediated PAR(2) activation. (c) Wt-rPAR(2) devoid of the N-terminal glycosylation sequon (PAR(2)T25(-)), but not rPAR(2) devoid of the glycosylation sequon located on extracellular loop-2 (PAR(2)T224A), was selectively and substantially (>30 fold) more sensitive to tryptase compared with the wt-rPAR(2). 4. Immunocytochemistry using antisera that specifically recognized the N-terminal precleavage sequence of PAR(2) demonstrated that tryptase released the precleavage domain from PAR(2)T25(-) but not from wt-rPAR(2). 5. Heparin : tryptase molar ratios of greater than 2 : 1 abrogated tryptase activation of PAR(2)T25(-). 6. Our results indicate that glycosylation of PAR(2) and heparin-inhibition of PAR(2) activation by tryptase could provide novel mechanisms for regulating receptor activation by tryptase and possibly other proteases.

Solubilization and characterization of motilin-receptor complexes from rabbit antral smooth muscle tissue

In the present study, the motilin receptor was characterized by enzymatic digestion studies and by solubilization of the motilin-receptor complex from prelabeled membranes using the anionic detergent cholic acid. Motilin binding was significantly decreased by preincubation of membranes of rabbit antral tissue with trypsin, phospholipase A2, C, D, dithiothreitol and 2-mercaptoethanol but not by neuraminidase and beta-galactosidase. Treatment of prelabeled membranes with 1% cholic acid resulted in solubilization of 24 +/- 5% of the proteins and 65 +/- 3% of the radioactivity. The latter was for 77 +/- 4% due to the presence of the motilin-receptor complex as estimated with PEG-precipitation. Upon gel-filtration on Superose 6 the complex partially dissociated but 43 +/- 3% eluted with macromolecular components in the void volume. This peak was not detected when membranes were first incubated with unlabeled motilin. Further disaggregation was accomplished by the addition of 0.5 M NaCl to the elution buffer. The chromatographic profile then showed a peak of about 370 kDa and a second one of 100 kDa. The latter value probably reflects the molecular mass of a single 125I-motilin-receptor-complex.

Kidney sialidase and sialyltransferase activities in spontaneously and experimentally diabetic rats. Influence of insulin and sorbinil treatments

Kidney cortex sialic acid level, sialidase and sialyltransferase activities have been measured in spontaneously diabetic BB rats and in streptozotocin-diabetic rats (STZ). In untreated diabetic BB rats, at the onset of the disease, sialidase specific activity was found to be increased by 21% when compared with diabetes-resistant BB controls (P less than 0.05) whereas sialyltransferase activity was not significantly modified and bound sialic acid concentration was diminished (P less than 0.05). In diabetic BB rats submitted to a minimal insulin therapy, during 3 months of disease, sialidase activity and sialic acid concentration were similar to those of Wistar age-matched controls. In STZ-diabetic Wistar rats, sialidase specific activity was increased by 76% after 5 months of disease when compared to age-matched Wistar controls (P less than 0.01); in contrast, specific sialyltransferase activity was decreased by 21% (P less than 0.05); these enzymatic alterations were associated with a decrease in bound sialic acid concentration (P less than 0.01); 1 month's insulin therapy, started 4 months after onset of the disease, normalized sialidase activity but had no effect on sialyltransferase activity and sialic acid concentration; treatment with sorbinil prevented cataract development but had no effect on sialidase activity whereas it emphasized the decrease in sialyltransferase activity and sialic acid concentration. The disturbances in the enzyme activities concerned with sialoglycoconjugate metabolism observed in experimental and spontaneous diabetes may be responsible for the decreased bound sialic acid content observed in the rat kidney cortex.

Asymmetric distribution of the phosphatidylinositol-linked phospho-oligosaccharide that mimics insulin action in the plasma membrane

We have investigated the topography of a glycosyl-phosphatidylinositol implicated in insulin action by a combination of two complementary methods: (a) chemical labelling with a non-permeable (isethionyl acetimidate) and a permeable (ethyl acetimidate) probe; and (b) enzymatic modifications with beta-galactosidase (EC 3.2.1.23) or phosphatidylinositol-specific phospholipase C (EC 3.1.4.3). Using the first approach the majority of the glycosyl-phosphatidylinositol is found in the outer surface of intact hepatocytes, adipocytes, fibroblasts and lymphocytes, but not in erythrocytes which presented only a 20% of the total labelled glycosyl-phosphatidylinositol to the exterior. Upon insulin addition (10 nM), about 60% of the total glycosyl-phosphatidylinositol was hydrolysed in both hepatocytes and adipocytes but not in erythrocytes. In agreement with the extracellular localization in hepatocytes and with the proposed role of this glycolipid in insulin action, treatment of rat hepatocytes with beta-galactosidase from Escherichia coli, an enzyme that hydrolyses the oligosaccharide moiety of the glycosyl-phosphatidylinositol, cleaved 65% of the total glycophospholipid and blocked the effect of insulin (but not of glucagon) on pyruvate kinase (EC 2.7.1.40). Similar treatment with phosphatidylinositol-specific phospholipase C from Bacillus cereus hydrolysed 62% of the total glycosyl-phosphatidylinositol. From the various approaches used it is concluded that the majority of this glycophospholipid is at the outer surface in a variety of insulin-sensitive cells.

Activation of glucose transport by activatory receptor agonists of adenylate cyclase in rat adipocytes

1. Catecholamine, glucagon, and adrenocorticotropic hormone stimulated 2-deoxyglucose (2-DG) uptake via an increase in glucose transporters in plasma membranes, similarly to insulin. 2. In contrast to the action of insulin, the stimulating effects of these agonists on 2-DG uptake were abolished when Gi was not activated. 3. The mode of the 2-DG uptake stimulation was partially different among these agonists.

Role of sialic acid in insulin action and the insulin resistance of diabetes mellitus

Adipocytes treated with neuraminidase show markedly reduced responsiveness to insulin without any alteration in insulin binding. In addition, several studies have separately demonstrated both insulin resistance and decreases in membrane sialic acid content and associated biosynthetic enzymes in diabetes mellitus. In the present study, we investigated the role that sialic acid residues may play in insulin action and in the hepatic insulin resistance associated with nonketotic diabetes. Primary cultures of hepatocytes from normal rats treated with neuraminidase demonstrated a dose-dependent decrease in insulin-stimulated lipogenesis. At a concentration of neuraminidase that decreases insulin action by 50%, 23% of total cellular sialic acid content was released. Neuraminidase-releasable sialic acid was significantly decreased in hepatocytes from diabetic rats and this was associated with significant insulin resistance. Treatment of hepatocytes from diabetic rats with cytidine 5'-monophospho-N-acetylneuraminic acid (CMP-NANA) enhanced insulin responsiveness 39%. The enhanced insulin responsiveness induced by CMP-NANA was blocked by cytidine 5'-monophosphate (CMP) suggesting that the CMP-NANA effect was catalyzed by a cell surface sialyltransferase. CMP reduced neuraminidase-releasable [14C]sialic acid incorporation into hepatocytes by 43%. The data demonstrate a role for cell surface sialic acid residues in hepatic insulin action and support a role for decreased cell surface sialic acid residues in the insulin resistance of diabetes mellitus.

Phorbol esters stimulate 2-deoxyglucose uptake in glia, but not neurons

Activation of protein kinase C by phorbol esters caused a time- and dose-dependent stimulation (270% of control) of glucose uptake in cultured glia, but not in neurons from rat brain. The phorbol ester stimulation of Vmax of glial glucose uptake occurred only in glia despite nearly 2.5-fold greater phorbol ester binding in neurons. These differences in cellular responses to protein kinase C activation may be the key to understanding brain glucose regulation.

Insulin stimulation of adipocyte membrane glucose transport. A graded biologic response insensitive to bilayer lipid disordering

Aspects of the mechanism by which insulin stimulates the membrane glucose transport system were examined by assessing the influence of the bilayer lipid structure on transport stimulation characteristics, and considering the form of the insulin dose-response curve. We tested the effects of membrane lipid perturbation on the insulin stimulation process. Benzyl alcohol, at concentrations (25 mM) that grossly fluidize lipids forming the adipocyte membrane bilayer matrix, caused 50% inhibition of intrinsic transporter activity. However, this membrane perturbation had no significant effect on either the insulin dose-response curve (conducted at 37 degrees) or the time-course of the insulin stimulation of hexose transport (conducted at 32 degrees). These data are difficult to rationalize in terms of a model in which transport stimulation involves interaction of transporters and hormone-bound receptors that is limited by lateral diffusion of these proteins in the fluid lipid bilayer. Curve-fitting experimental insulin dose-response data for stimulation of 2-deoxy-D-glucose and D-glucose uptake provided an estimate of an insulin "association constant" for transport regulation that may be compared with recent insulin receptor binding data. Similar magnitude constants were obtained whether estimated directly from plots of transport velocity versus arithmetic hormone dose, or by extrapolation from linear segments of sigmoidal velocity versus log dose plots, or from inverse (Lineweaver-Burk-type) plots of the insulin dose-response data. Insulin apparently regulates transport by associating with a binding site, having an apparent dissociation constant which is determinable through kinetic measurements of hexose uptake (KDapp approx. 17-40 pM). This is in good agreement with the dissociation constant, KD, determined from Scatchard plots of recent binding data to adipocytes, for a class of receptors representing the "high affinity" binding sites for insulin. Insulin dose-response curve simulations also indicated that the stimulation process may be classified in pharmacologic terms as a typical graded biologic response and may involve insulin association with a site that regulates transport rates in a manner kinetically analogous to allosteric modulation of a V-series enzyme by a noncompetitive ligand. From the results we suggest that a relatively close association occurs between transport and receptor proteins in the membrane, where the relative activation of transport depends on the fractional occupancy of functional high affinity receptors by insulin, and the insulin stimulation of transport involves regions of the membrane that are not influenced significantly by

Stimulation of 2-deoxyglucose uptake in rat adipocytes by a human growth hormone fragment (hGH 4-15)

The in vivo and in vitro effects of a hypoglycaemic fragment of human growth hormone containing the sequence H2N-Ile-Pro-Leu-Ser-Arg-Leu-Phe-Asp-Asn-Ala-Met-Leu-COOH (hGH 4-15) on 2-deoxy-[1-14C]-D-glucose uptake in adipocytes were studied. The isolated cells from rats after a single intravenous injection of hGH 4-15 (1 mg/kg) significantly increased uptake of 2-deoxyglucose (p less than 0.005). Adipocytes from untreated rats pre-incubated with the hGH fragment (10 micrograms/ml) at 37 degrees C for 30 min also clearly showed an elevated uptake of 2-deoxyglucose in the absence and the presence of exogenous insulin. The effect of hGH 4-15 was concentration-dependent, steadily increasing with the maximum effect at 10 micrograms/ml. The present findings suggest that the enhancement of glucose transport in target tissues may be a major contributing factor to the hypoglycaemic action of the amino-terminal fragments of human growth hormone.

The glycoprotein nature of A1 adenosine receptors

A1 adenosine receptors from different tissues and species were photoaffinity labelled and then the carbohydrate content was examined by both enzymatic and chemical treatment. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of the labelled membrane receptors shows that neuraminidase treatment alters the electrophoretic mobility of the receptor band indicating the presence of terminal neuraminic acids. Neuraminidase digestion does not influence the binding characteristics of the receptor. The totally deglycosylated receptor protein obtained by chemical treatment has an apparent molecular weight of 32,000.

Structural and functional characteristics of insulin receptors in rat neuroblastoma cells

Insulin receptors were detected in a variety of rat neuroblastoma and glioma cell lines. The binding of 125I-insulin to B103 neuroblastoma cells had characteristics typical of insulin receptors in other tissues, including high affinity for insulin, low affinity for insulin-like growth factor I (IGF-I), and curvilinear Scatchard plots. Using photoaffinity labeling procedures and sodium dodecyl sulfate (SDS) gel electrophoresis to analyze the subunit structure of insulin receptors in B103 cells, the predominantly labeled protein had an apparent molecular weight of 125K and the mobility of this protein was shifted after removal of sialic acid residues. On the basis of size and susceptibility to neuraminidase, the insulin binding subunit in neuroblastoma cells was identical to the alpha-subunit of insulin receptors in adipocytes and different from the 115K subunit found in brain. The presence of an "adipocyte" form of the insulin receptor in clonal cells derived from brain is probably a consequence of transformation and results from more extensive oligosaccharide processing of the 115K receptor expressed in normal brain cells. The fully glycosylated receptors in neuroblastoma cells were capable of exerting functions typical of insulin receptors in adipocytes such as internalization of insulin and stimulation of glucose transport.

Removal of sialic acids from the purified insulin receptor results in enhanced insulin-binding and kinase activities

Neuraminidase treatment of the purified insulin receptor resulted in an increase in both insulin-binding and kinase activities. Neuraminidase-treated alpha and beta subunits moved further than native subunits on sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) under reducing conditions. The enhancement of insulin-binding and kinase activities and increased mobility of the subunits on SDS-PAGE were not observed when the receptor was treated with neuraminidase in the presence of neuraminidase inhibitor. These results suggest that terminal sialic acid residues have a significant role in insulin-binding and kinase activities. The involvement of sialic acid residues in the activities of the receptor has not been detected by previous studies.

Insulin receptors on Chinese hamster ovary (CHO) cells: altered insulin binding to glycosylation mutants

We have identified and characterized insulin receptors on Chinese hamster ovary (CHO) cells. Insulin binds in a time, temperature and pH dependent fashion and insulin analogues compete for 125I-insulin binding in order of their biological potencies. Furthermore, two CHO cell glycosylation mutants, B4-2-1, lacking high mannose containing glycoproteins, and Lec 1.3c, lacking complex carbohydrate containing glycoproteins, bind insulin with a much higher and lower affinity respectively than wild type cells. This is the first report of insulin receptors on CHO cells and the first to use glycosylation mutants to study the effects of abnormal carbohydrates on insulin binding.

Unique features of the insulin receptor in rat brain

We examined the structure of the affinity-labeled insulin receptors in rat brain, rat liver, and human IM-9 lymphocytes using sodium dodecyl sulfate-polyacrylamide gel electrophoresis. In gels run under reducing conditions, the alpha-subunit of the insulin receptor in brain had an apparent Mr of 127,000 distinctly lower than that seen in both rat liver and human lymphocytes (apparent Mr = 136,000). Exposure to neuraminidase increased the electrophoretic mobility of the liver receptor, but had no effect on the insulin receptor in brain. The carbohydrate moieties of the insulin receptors in rat brain and liver were further examined by chromatography on wheat-germ agglutinin agarose. The receptors in both tissues adsorbed to the wheat-germ agglutinin; elution with 0.3 M N-acetyl glucosamine resulted in slightly better recovery of the brain than of the liver receptor. Exposure to neuraminidase virtually abolished the interaction of the liver receptor with the lectin, whereas adsorption of the brain receptor was unaffected by neuraminidase. These results indicate that the insulin receptor in brain is distinguished from those in peripheral tissues by structural alterations, including changes in the carbohydrate moiety of the receptor. Such alterations contrast sharply with the previously observed similarities in insulin binding properties between insulin receptors in brain and other tissues. The implications of such structural alterations for the program of insulin action expressed by the receptors in brain remain to be explored.

Agglutination kinetics of enzymatically treated normal and diabetic rat hepatocytes

The effect of trypsin and neuraminidase treatments on concanavalin A-induced agglutination of viable hepatocytes from normal and diabetic rats are reported. Trypsin (1.0 microgram/mL) treatment resulted in a increased rate of hepatocyte agglutination in both normal and diabetic cells in the presence of 100 micrograms/mL of concanavalin A. However, neuraminidase treatment resulted in a decrease in the rate of cytoagglutination in the normal cells and an increase in the rate in the diabetic counterpart. The results suggest that trypsin may have caused the removal of a surface protein and/or split a peptide bond on the agglutinin receptors resulting in identical receptor exposure and clustering in normal and diabetic cells. The neuraminidase data suggest that the arrangements of the neuraminic acid moieties on the receptors in normal cell membranes were different from those in the diabetic cells, eliminating the possible effect of changes in the surface charge density. In conclusion, normal cells carry numerous clustered (possibly some in the "cryptic" state) agglutinin receptors in the cell membrane as compared with cells from diabetic rats.

Effect of purified phospholipases on glucose transport, insulin binding, and insulin action in isolated rat adipocytes

The influence of alterations in phospholipid structure by phospholipase treatment on insulin action and glucose transport in rat adipocytes was studied. It appeared that phospholipase A2 from bee venom caused a breakdown of approximately 50% of phosphotidylcholine without lysis of the cells. Because of this treatment, insulin binding was increased, resulting in an increased sensitivity of glucose transport towards lower insulin concentrations. Moreover, an increased affinity of the transport system for 2-deoxyglucose was observed. Phospholipase C from Clostridium welchii caused complete lysis of adipocytes. Phospholipase A2 from Crotalus adamenteus was without effect.

Carbohydrate determinants involved in both the binding and action of insulin in rat adipocytes

The insulin receptor apparent affinity was markedly decreased in fat cells treated with lectins specific either for D-galactose (Ricinus communis agglutinin I, RCAI), D-mannose (concanavalin A, Con A, Lens culinaris agglutinin, LCA) or N-acetyl-D-glucosamine (wheat germ agglutinin, WGA), as indicated by a rightward shift of the binding competition curves and almost lineared Scatchard plots. Limulus polyphemus agglutinin (LPA), specific for sialic acid, was ineffective. All lectins enhanced 2-deoxy-D-glucose uptake with relative bioactivities (maximal lectin effect/maximal insulin effect) of 68-86%. Insulin and lectin stimulatory effects were antagonized by specific carbohydrates used as competitors and inhibited by cytochalasin B (70 microM). Maximal effects of insulin and lectins were not additive and were completely abolished in neuraminidase-treated fat cells. Lectins did not affect insulin degradation. These data show that sialylated glycosidic moieties containing D-galactose, D-mannose and N-acetyl-D-glucosamine units are involved in both processes of insulin 'high affinity' binding and activation of glucose transport but are not implicated in hormone degradation. They suggest that N-linked carbohydrate chains of the complex type may be essential for functional insulin receptor and post-receptor systems.

The lack of insulin mimetic or antagonistic effects of methyl-alpha-D-Mannoside in iso-osmolar solutions

Transduction of insulin binding into metabolic control in isolated rat adipocytes apparently requires intact cell surface carbohydrate. The ability of certain lectins and some glycosides to mimic and/or inhibit the actions of insulin had been cited as evidence supporting the hypothesis that a concanavalin A-like binding site on fat cells is crucial to this function. Such a binding site could explain the stimulation by methyl-alpha-D-mannoside of glucose oxidation or its ability to antagonize the effect of insulin on lipolysis. The present study corroborated these effects of methyl-alpha-D-mannoside in hyperosmolar medium, but shows that the effects vanish when osmolarity is maintained within physiological limits. Osmolarity alone could not explain all of the complex effect observed, but it can be concluded that earlier data suggesting methyl-alpha-D-mannoside mimics or antagonizes the actions of insulin cannot be used to support the above hypothesis.

Properties of the insulin receptor of rat pancreatic islet

Rat pancreatic islets have been shown to possess specific binding sites for 125I-labeled insulin. Enzymatic and chemical modification of islets are used to reveal important structures and chemical groups for insulin binding. pretreatment with trypsin, neuraminidase, 1-ethyl-3-(3-dimethylamino)carbodiimide (a carboxyl reagent), tetranitromethane (a tyrosyl and thiol reagent), and 1,3-difluoro-4,6-dinitrobenzene (modification of protein functional groups) decreased binding of insulin. This was due to the diminuation of the receptor number; in the case of trypsin-pretreatment also the receptor affinity was decreased. Inhibition of insulin binding was in each case associated with a decrease of the inhibitory effect of exogenous insulin on glucose-induced insulin secretion (not measured in the case of difluorodinitrobenzene and tetranitromethane). Phospholipase A2, (cleavage of phospholipids) did not affect these parameters. 5,5'-dithiobis(2-nitrobenzoic acid) (Ellman's reagent) and possibly p-chloromercuribenzoate (both thiol reagents) increased the number of receptors and decreased receptor affinity, but did not influence the inhibitory effect of insulin on insulin release. It is concluded that protein functional groups, sialic acid, carboxyl and tyrosyl groups, but no phospholipids and probably not sulfhydryl groups are important for the interaction of insulin with insulin receptors of rat pancreatic islets.

Insulin receptors on isolated heart cells: effect of temperature and hydrolytic enzymes

Isolated muscle cells from adult rat heart have been used to study the effect of temperature and enzymic digestion on the binding of 125I-labelled insulin. Equilibrium binding studies were performed at both 4 and 37 degrees C, using insulin concentrations ranging from 2.5 X 10(-11) mol/l to 10(-6) mol/l. The empty site affinity constant decreased by 51% from 1.0 X 10(8) l/mol at 4 degrees C to 4.9 X 10(7) l/mol at 37 degrees C, whereas the total receptor concentration remained unaltered at both temperatures. The rate of dilution induced dissociation was enhanced by the presence of native insulin at 37 degrees C, confirming the presence of negative cooperativity among the receptor sites at physiological temperatures. Treatment of isolated heart cells with trypsin and beta-galactosidase led to a decrease in specific binding of 125I-labelled insulin. Myocytes treated with neuraminidase exhibited a significant increase in insulin binding, which was shown to be due to an increase in insulin-receptor affinity. These studies provide new information on the molecular characteristics of insulin receptors in the heart muscle.

GnRH-receptors of rat pituitary is a glycoprotein: differential effect of neuraminidase and lectins on agonists and antagonists binding

The interaction of 125I-labeled gonadotropin releasing hormone (GnRH) agonist, [D-Ser-(t-Bu)6,des-Gly10-ethylamide]-GnRH, and antagonist [D-pGlu1,D-Phe2,D-Trp3,6]-GnRH with rat pituitary membranes was studied. Their binding was affected differently by pretreatment of membranes with neuraminidase and by wheat-germ agglutinin. Pretreatment of the membranes with neuraminidase abolished the specific binding of both antagonist and agonist, in a dose-response manner, with the former being less affected. Wheat-germ agglutinin affected antagonist binding slightly (22% inhibition at 40 microgram/ml) but inhibited agonist binding more markedly (50% at 40 microgram/ml). This inhibitory effect was specific since it was readily reversed by N-acetylglucosamine. Other lectins, such as concanavalin A and soybean agglutinin affected the binding of both agonist and antagonist to only a small degree. These results suggest that the GnRH-receptor of rat pituitary is a glycoprotein which contains sialic acid residue and that GnRH agonists and antagonists bind differently to the same receptor.

Lipid composition in liver and brain of genetically obese (ob/ob), heterozygote (ob/+)and normal (+/+) mice

Lipid composition was studied in liver and brain of normal (+/+), heterozygote (ob/+) and obese (ob/ob) mice. It was found that this genetic defect is expressed differently in the lipid composition of these organs. Cholesterol is increased in liver but strongly decreased in brain of obese animals. Phosphatide fatty acid composition is modified in liver and not in brain. In contrast, phospholipids and total ganglioside sialic acid are affected similarly in both organs. Although clinically normal, heterozygote (ob/+) mice already show an abnormal lipid composition in liver and brain. The potential importance of these results is presented.

Identification of liver cell membrane galactoglycoproteins involved in the process of insulin binding

The glycoproteinic nature of the insulin receptor was indicated using two different approaches: 1. [125I] insulin binding to soluble receptors from mouse liver was inhibited by digestion with beta-galactosidase or pretreatment with Ricinus communis I or concanavalin A. An other enzyme (neuraminidase) and lectins (wheat germ agglutinin, Dolichos biflorus) did not affect the binding reaction. These data confirmed that insulin directly interacts with the galactoglycoproteins of liver membranes. 2. The galactose oxidase-sodium boro[3H] hydride technique, previously used for labeling accessible membrane galactoglycoproteins, was again utilized to discern the components that interact with insulin. When liver membranes were equilibrated with 10-7 M insulin prior to labeling, the SDS gel radioactive profiles were specifically modified with two galactoglycoprotein of apparent molecular sizes 195 000 and 145 000, compatible with their participation in the insulin binding interaction. Membrane pretreatment with beta-galactosidase or Sophora japonica lectin reduced the labeling in most peaks, thus supporting the argument for labeling sensitivity. Preincubation of membranes with 10-7 M proinsulin slightly hindered labeling, while pretreatment with 10-7 M glucagon was ineffective, suggesting a specificity of the insulin effect. These data indicate that glycoprotein nature of the insulin receptor for two reasons: alteration of insulin binding after modification of the galactoglycoproteins, and alteration of galactoglycoprotein labeling after insulin binding. Two galactoglycoproteins, with apparent molecular weights 145 000 and 195 000, respectively, were identified and they are suggested to have insulin binding properties.

Endotoxin-induced alterations in glucose transport in isolated adipocytes

2-Deoxyglucose and 3-O-methylglucose were used to assess endotoxin-induced changes in glucose transport in rat adipocytes. 6 h after Escherichia coli endotoxin injection insulin-stimulated 2-deoxyglucose uptake was significantly depressed (V decreased, Km unaltered), phosphorylation of 2-deoxyglucose was seemingly unimpaired; basal 3-methylglucose entry was significantly increased, insulin-stimulated uptake was unaltered. Insulin significantly reduced Km in control and endotoxin-treated cells. Cytochalasin B-insensitive uptake of both 2-deoxy-glucose and 3-methylglucose, a small fraction of total transport, increased significantly in endotoxic cells. Endotoxin reduced spermine- and insulin-stimulated 2-deoxyglucose uptake to a similar extent. Results are consistent with the hypotheses that (1) a site of endotoxin-induced insulin resistance is at the cell membrane level and may reflect a decrease in number of activity of effective carrier units, rather than alterations in affinity, (2) endotoxin does not compromise the hexokinase system, (3) the cell membrane-localized effect of endotoxin on hexose transport is not necessarily mediated by the insulin receptor and (4) the entry of 2-deoxyglucose and 3-methyl-glucose may involve two separate transport systems.

Further characterization of the insulin receptor glycosidic moiety in rat adipocytes

Simultaneous or sequential treatment of rat adipocytes with neuraminidase plus beta-galactosidase decreased insulin binding by 43%. No modification was observed with either enzyme individually. alpha-Mannosidase enhanced insulin binding (38%), whereas beta-N-acetylglucosaminidase and alpha-L-fucosidase were ineffective. Lectins that interact with galactose (Ricinus communis I, RCAI), mannose, Lens culinaris agglutinin (LCA), Concanavalin A (Con A) or N-acetylglucosamine (wheat-germ agglutinin, WGA) decreased insulin binding by 43, 57, 59 and 85% respectively. Lectin inhibition was dose-dependent, saturable and prevented by specific monosaccharides. RCAI, LCA, Con A and WGA decreased the insulin dissociation process by 45, 90, 78 and 84% respectively. Lectins specific for sialic acid, terminal galactose, N-acetylgalactosamine or fucose (Limulus polyphemus, peanut, soybean and Ulex I agglutinins) did not modify either insulin binding or dissociation. These results indicate involvement of penultimate D-galactose, internal N-acetyl-D-glucosamine and D-mannose residues in both processes. They suggest that, in rat adipocytes, a glycosidic moiety participates in the insulin-receptor interaction through N-linked oligosaccharides of the 'complex type'.

Selective inhibition by neuraminidase of insulin action on hexose metabolism of mouse adipocytes

Mouse adipocytes treated with neuraminidase showed a decreased response of glucose oxidation to insulin although insulin binding to the cells was normal. The decreased response was associated with the release of sialic acids from the cells by enzyme digestion. The hormone action on 2-deoxyglucose uptake was also decreased. However, the hormone action on glyceride-glycerol synthesis or lipogenesis from glucose was unaltered when enzyme-treated cells were incubated with higher glucose concentration (greater than or equal to 5 mM). However, at lower glucose concentrations (< 5 mM), in which glucose transport was a rate-limiting step, the hormone action was markedly decreased. When fructose was used as a substrate, the enzyme-treated cells showed an impaired response to insulin in fructose oxidation but not in glyceride-glycerol synthesis and lipogenesis from fructose. These results suggest that the postreceptor systems of insulin action on glyceride-glycerol synthesis and lipogenesis from hexose are different from those of the hormone action on hexose transport and oxidation. Furthermore, alteration in insulin-sensitive metabolic profiles may be caused, in part, by changes in glycoproteins and/or glycoplipids on the cell surface.

Production of insulinomimetic antibodies against rat adipocyte membranes by hybridoma cells

SJL mice were injected intraperitoneally with adipocyte plasma membranes or with intrinsic membrane proteins obtained by extraction of plasma membranes with dimethylmaleic anhydride. Three days after the boost injection, the spleens were removed and fused with NS-a, a thioguanine-resistant myeloma cell line derived from P3X63 Ag8 (Balb/c). Following selection for hybrids with hypoxanthine, aminopterin, and thymidine, medium of the hybrid cells was tested for its ability to bind to the plasma membrane of the adipocyte and to stimulate the oxidation of D-(1-14C) glucose to 14CO2. Approximately 40% of the wells containing hybridomas derived from splenocytes of SJL mice immunized with plasma membranes produced immunoglobulin that bound to adipocyte plasma membranes. About 30% of these mimicked the ability of insulin to stimulate the oxidation of D-(1-14C) glucose to 14CO2 in adipocytes. Media from 51% of the wells containing hybridomas derived from splenocytes of SJL mice immunized with intrinsic membrane proteins produced immunoglobulin that bound to the plasma membrane and 48% of those stimulated glucose oxidation. The bioactivity of the hybrid cell media could be blocked by adsorption with intrinsic membrane proteins or by the removal of immunoglobulins using formalin-fixed Staphylococcus aureus. The hybrids generated in this study can be divided into three categories: 1) hybrids that secrete antibodies that can bind to plasma membranes and mimic insulin action of glucose transport; 2) hybrids that secrete antibodies that bind to plasma membranes but do not stimulate the oxidation of D-(1-14C) glucose to 14CO2; and 3) hybrids that produce no antimembrane antibodies. The data suggest that interaction of immunoglobulins with specific membrane proteins is essential in mimicking the action of insulin on glucose transport and oxidation in the rat adipocyte.