Recent progress in our understanding of the mechanism of action of insulin

Use of an antisense strategy to dissect the signaling role of protein-tyrosine phosphatase alpha

The protein-tyrosine phosphatase PTPalpha has been proposed to play an important role in controlling the dephosphorylation of a number of key signaling proteins and in regulating insulin signaling. To examine the potential cellular functions and physiological substrates of PTPalpha, a potent phosphorothioate oligonucleotide-based antisense strategy was developed that specifically depleted endogenous PTPalpha from 3T3-L1 adipocytes. The antisense probe, alphaAS1, achieved PTPalpha depletion levels normally of >/=85% and which varied up to levels where PTPalpha was not detected at all. Elimination of PTPalpha by 85% inhibited c-Src activity by 80%. Abolishing PTPalpha to levels undetected did not alter the tyrosine dephosphorylation of the insulin receptor or insulin receptor substrate proteins. Moreover, the ability of insulin to activate ERK2 or to stimulate DNA synthesis was not altered by alphaAS1. It is concluded that endogenous PTPalpha is a key regulator of c-Src activity in 3T3-L1 adipocytes and that PTPalpha is not required for the dephosphorylation of the insulin receptor or the insulin receptor substrate proteins or for the regulation of several downstream insulin signaling events in 3T3-L1 adipocytes. Finally, the development of the antisense probe, alphaAS1, provides an important molecular tool of general applicability for further dissecting the roles and precise targets of endogenous PTPalpha.

Fatty acid modification of membrane fluidity in Chinese hamster ovary (TR715-19) cells

Dietary saturated fatty acids, especially lauric (12:0), myristic (14:0) and palmitic (16:0) acids, which are hypercholesterolemic, influence cell membrane fatty acid composition and affect LDL receptor function. When membrane phospholipid fatty acids in Chinese hamster ovary cells, containing the human LDL receptor, were modified (Hannah J. S. et al., 1995 Metabolism 44, 1428-1434), LDL receptor function was affected, but correlations with DPH-determined membrane fluidity were weak. The role of fluidity in various membrane domains with respect to the LDL receptor is examined here. Membrane fluidity was assessed by measuring steady-state fluorescence polarization of diphenylhexatriene (DPH) and its polar propionic acid (DPH-PA) and trimethylammonium (TMA-DPH) derivatives from 38 to 4 degrees C in fatty acid modified Chinese hamster ovary cells. Fatty acid changes modulated mid-bilayer fluidity as determined with DPH, but fluidity in phospholipid headgroup domains, assessed with DPH-PA and TMA-DPH, was independent of fatty acyl composition. The DPH fluidity was related to membrane unsaturation (P < 0.02), oleate contents (P < 0.009) in particular, but inversely related (P < 0.0002) to the longer chain (> or = 20 C atoms) unsaturated fatty acids with from four to six double bonds. The LDL binding was independent of fluidity, but there were weak relations between LDL internalization and DPH-PA anisotropy and between LDL degradation and TMA-DPH anisotropy. It was concluded that LDL binding was not related to mid-bilayer fluidity, but the results with the polar probes suggest a role of fluidity in modulating vertical displacement of the LDL/LDL receptor complex across the plasma membrane.

In vivo modulation of N-myristoyltransferase activity by orthovanadate

N-Myristoyltransferase (NMT) catalyses the transfer of myristate from myristoyl-CoA to the NH2-terminal glycine residue of several proteins and are important in signal transduction. STZ-induced diabetes (an animal model for insulin-dependent diabetes mellitus, IDDM) resulted in a 2-fold increase in rat liver NMT activity as compared with control animals. In obese Zucker (fa/fa) rats (an animal model for non-insulin dependent diabetes mellitus, NIDDM) there was a approximately 4.7-fold lower liver particulate NMT activity as compared with the control lean rat livers. Administration of sodium orthovanadate to the diabetic rats normalised liver NMT activity. These results would indicate that the rat liver particulate N-myristoyltransferase activity appears to be inversely proportional to the level of plasma insulin, implicating insulin in the control of N-myristoylation.

Studies on an insulin-stimulated insulin receptor serine kinase activity: separation of the kinase activity from the insulin receptor and its reconstitution back to the insulin receptor

In cells insulin stimulates autophosphorylation of the insulin receptor on tyrosine and its phosphorylation on serine and threonine by poorly characterized kinases. Recently we have achieved co-purification of the insulin receptor with insulin-stimulated insulin receptor serine kinase activity. We now show that the co-purified serine kinase activity can be removed by NaCl washing and reconstituted by adding back the NaCl eluate. Reconstitution enabled higher serine phosphorylation than achieved with the co-purified preparation. Myelin basic protein was discovered to be a potent substrate for insulin-stimulated serine phosphorylation by the co-purified preparation, with the activity responsible having similar properties to the serine kinase activity towards the receptor. Myelin basic protein was also phosphorylated on serine by the NaCl eluate. Myelin basic protein phosphorylated by the co-purified preparation or the NaCl eluate gave the same set of phosphoserine peptides. The major myelin basic protein serine kinase activity in the NaCl eluate co-purified exactly on Mono Q with the activity that restored insulin-stimulated insulin receptor serine phosphorylation. These results provide strong evidence for the true separation of the serine kinase from the insulin receptor and the distinctiveness of the serine kinase activity from the insulin receptor tyrosine kinase and mitogen-activated protein kinases. The procedures developed for the isolation of the serine kinase and the establishment of an effective in vitro substrate should allow purification of the kinase. The protocols also provide flexible systems for identifying the functions of the insulin-stimulated serine phosphorylations and the respective kinase(s).

Regulation of an hepatic low-M(r) membrane-associated protein-tyrosine phosphatase

Protein-tyrosine phosphatases (PTPases), active against autophosphorylated insulin and epidermal growth factor (EGF) receptors in rat liver, are predominantly membrane associated. Fasting of rats for 48 h decreased hepatic particulate PTPase activity by 15.0-26.9%. This reduction in particulate PTPase activity was due to a rather specific decrease in activity of > 85% of a single species of PTPase, termed PTPase I. Disappearance of PTPase I activity from the particulate fraction was not accounted for by its translocation to the cytosol. PTPase I displayed the highest activity against autophosphorylated insulin and EGF receptors, relative to activity against a 32P-labelled peptide substrate, of three PTPases resolved from the liver particulate fraction. The M(r) value of PTPase I, as determined by gel filtration on a Superose 12 column was approx. 42,000, indicating that PTPase I belongs to the low-M(r) class of PTPases. An antibody raised against PTPase 1B, the prototype of this class of PTPases, did not react with PTPase I in Western blots. The potential importance of the novel change in activity of PTPase I in the regulation of insulin-receptor signal transduction is discussed.

Insulin-induced enhancement of cell morphological dynamics: non-specific biophysical mechanisms for the generalized stimulation of metabolism?

Time series analysis of fluctuations in the intensity of light scattered by cells indicates the existence of a range of oscillations in cell morphology that usually occur in (pseudo) periodic bursts. Insulin has a frequency-dependent effect on the rhythms, predominantly that of stimulating the short period components believed to reflect surface movements. It is suggested that the non-specific action of the hormone on transport is due to an effective decrease in the net thickness of the surrounding diffusion layer (resulting from increased stirring of the micro-environment) and to accompanying membrane and cytoplasmic vibrations. From published data it seems likely that other mitogens and chemotactic agents act in a similar manner.

Insulin induces changes in the subcellular distribution of actin and 5'-nucleotidase

An increase in the amount of actin associated with the plasma membrane was visualized by immunocytochemistry 5 min after the addition of insulin to Krebs II ascites tumour cells maintained in serum-free medium. At 1 h of incubation the rim of fluorescence at the plasma membrane as measured by image analysis, was about 30% more intense than in control cells indicating that the initial accumulation of actin at the plasma membrane was not of a transient nature. Since an increase in the total cellular actin content in ascites cells did not occur until after a lag period of about 15 min then the increased amount of actin at the plasma membrane seen at 5 min was attributed to a stimulation of the polymerization of actin. An increase in the association of actin at the plasma membrane was also observed in 3T3 fibroblasts in areas of membrane ruffling, while in some cells there was also increased actin accumulation in the perinuclear area. The putative plasma membrane-microfilament linking protein 5'-nucleotidase was shown to be present in association with actin in the cytoskeletal fraction. Incubation of cells with insulin resulted in a shift of the enzyme toward the bottom of gradients indicating association with actin filaments of a greater length. The results demonstrate that insulin causes a stimulation of actin polymerization and that the hormone can be therefore assigned a role in the regulation of the cytoskeleton.

Site-specific dephosphorylation and deactivation of the human insulin receptor tyrosine kinase by particulate and soluble phosphotyrosyl protein phosphatases

Insulin receptor tyrosine kinase activation, induced by insulin-stimulated autophosphorylation, was measured using a synthetic peptide containing residues 1142-1153 of the insulin receptor and shown to be reversed by both particulate and soluble phosphotyrosyl protein phosphatases from rat liver. Deactivation of the tyrosine kinase was highly sensitive to phosphatase action and was correlated best with disappearance of insulin receptors triphosphorylated in the tyrosine-1150 domain. Dephosphorylation of the di- and mono-phosphorylated forms of the tyrosine-1150 domain generated during dephosphorylation or of phosphorylation sites in the C-terminal or putative juxta-membrane domains occurred 3- greater than 10-fold more slowly than deactivation of the tyrosine kinase, and these phosphorylated species did not appear to appreciably (less than 20%) contribute to tyrosine kinase activation. These results indicate that the transition from the triply to the doubly phosphorylated form of the tyrosine-1150 domain acts as an important switch for deactivation of the insulin receptor tyrosine kinase during dephosphorylation. The exquisite sensitivity of this dephosphorylation/deactivation event to phosphotyrosyl protein phosphatase action, combined with the high affinities of this phosphatases for substrates and the high activities of the phosphatases in cells, suggests that the tyrosine kinase activity expressed by insulin-stimulated insulin receptors is likely to be stringently regulated.

Insulin activates guanosine 5'-[gamma-thio] triphosphate (GTP gamma S) binding to a novel GTP-binding protein, GIR, from human placenta

A novel GTP-binding protein, GIR, along with insulin receptor (IR), has been partially purified from human placenta. A non-hydrolyzable substrate, GTP gamma S which is known to bind to GTP-binding proteins with high affinity, reduces insulin binding to IR-GIR fraction by approximately 29% and 100 nM insulin stimulates GTP gamma S binding to IR-GIR fraction by approximately five-fold. The molecular weight of the protein (may be subunit) that binds to 8-azido-GTP, a photoaffinity label for G-proteins, is approximately 66,000.

The effect of putative protein kinase C inhibitors, K252a and staurosporine, on the human neutrophil respiratory burst activated by both receptor stimulation and post-receptor mechanisms

1. Two compounds, reported to be potent inhibitors of protein kinase C (PKC), K252a and staurosporine, have been examined in order to gain further information as to the possible role played by PKC in the signal transduction sequence of the neutrophil respiratory burst as determined by superoxide (O2-) production. 2. A number of stimuli were used in the study, some acting at receptors i.e. fMet-Leu-Phe (fMLP), opsonized zymosan and heat-aggregated IgG (HAGG), one acting on a G-protein, fluoride, and two direct PKC activators, dioctanoylglycerol (diC8) and phorbol myristate acetate (PMA). 3. K252a and staurosporine inhibited the respiratory burst with all the stimuli but the order of agonist sensitivity was very different with the two inhibitors. 4. For K252a-induced inhibition of O2- release, the order of potency was fluoride greater than fMLP, HAGG greater than opsonized zymosan greater than PMA, DiC8. For staurosporine-induced inhibition of O2- release, the order of potency changed to fluoride greater than DiC8, PMA greater than HAGG, fMLP greater than opsonized zymosan. The significance of this unexpected difference in relative rank order of potency is discussed with reference to the reported mechanism of action of the two inhibitors and the events involved in the oxidative burst. 5. Staurosporine at low concentrations increased the fMLP-stimulated O2- response by 100%, the maximum effect occurring at 35 nM. 6. To the extent that the compounds used are specific inhibitors of PKC, these findings support a role for the enzyme PKC in stimulus-activation coupling in O2- generation with all the stimuli used in this study.

Dephosphorylation of insulin-receptor autophosphorylation sites by particulate and soluble phosphotyrosyl-protein phosphatases

Insulin stimulates autophosphorylation of the insulin receptor on multiple tyrosines in three domains: tyrosines 1316 and 1322 in the C-terminal tail, 1146, 1150 and 1151 in the tyrosine-1150 domain, and possibly 953, 960 or 972 in the juxtamembrane domain. In the present work the sequence of dephosphorylation of the various autophosphorylation sites by particulate and cytosolic preparations of phosphotyrosyl-protein phosphatase from rat liver was studied with autophosphorylated human placental insulin receptor as substrate. Both phosphatase preparations elicited a broadly similar pattern of dephosphorylation. The tyrosine-1150 domain in triphosphorylated form was found to be exquisitely sensitive to dephosphorylation, and was dephosphorylated 3-10-fold faster than the di- and monophosphorylated forms of the tyrosine-1150 domain or phosphorylation sites in other domains. The major route for dephosphorylation of the triphosphorylated tyrosine-1150 domain involved dephosphorylation of one of the phosphotyrosyl pair, 1150/1151, followed by phosphotyrosyl 1146 to generate a species monophosphorylated mainly (greater than 80%) at tyrosine 1150 or 1151. Insulin receptors monophosphorylated in the tyrosine-1150 domain disappeared slowly, and overall the other domains were completely dephosphorylated faster than the tyrosine-1150 domain. Dephosphorylation of the diphosphorylated C-terminal domain yielded insulin receptor in which the domain was singly phosphorylated at tyrosine 1322. Triphosphorylation of the insulin receptor in the tyrosine-1150 domain appears important in activating the receptor tyrosine kinase to phosphorylate other proteins. The extreme sensitivity of the triphosphorylated form of the tyrosine-1150 domain to dephosphorylation may thus be important in terminating or regulating insulin-receptor tyrosine kinase action and insulin signalling.

Characterization of sites of serine phosphorylation in human placental insulin receptor copurified with insulin-stimulated serine kinase activity by two-dimensional thin-layer peptide mapping

Insulin receptor was copurified from human placenta together with insulin-stimulated kinase activity that phosphorylates the insulin receptor on serine residues. Analysis of phosphorylated insulin receptor by two-dimensional tryptic peptide mapping showed that sites of insulin stimulated serine phosphorylation in the insulin receptor were recovered in the same peptides as those known to be phosphorylated on serine in vivo in response to insulin. This indicates that the serine kinase copurified with the insulin receptor represents a physiologically important enzyme involved in the insulin triggered serine phosphorylation of the insulin receptor in vivo.

Insulin-receptor phosphotyrosyl-protein phosphatases

Calmodulin-dependent protein phosphatase has been proposed to be an important phosphotyrosyl-protein phosphatase. The ability of the enzyme to attack autophosphorylated insulin receptor was examined and compared with the known ability of the enzyme to act on autophosphorylated epidermal-growth-factor (EGF) receptor. Purified calmodulin-dependent protein phosphatase was shown to catalyse the complete dephosphorylation of phosphotyrosyl-(insulin receptor). When compared at similar concentrations, 32P-labelled EGF receptor was dephosphorylated at greater than 3 times the rate of 32P-labelled insulin receptor; both dephosphorylations exhibited similar dependence on metal ions and calmodulin. Native phosphotyrosyl-protein phosphatases in cell extracts were also characterized. With rat liver, heart or brain, most (75%) of the native phosphatase activity against both 32P-labelled insulin and EGF receptors was recovered in the particulate fraction of the cell, with only 25% in the soluble fraction. This subcellular distribution contrasts with results of previous studies using artificial substrates, which found most of the phosphotyrosyl-protein phosphatase activity in the soluble fraction of the cell. Properties of particulate and soluble phosphatase activity against 32P-labelled insulin and EGF receptors are reported. The contribution of calmodulin-dependent protein phosphatase activity to phosphotyrosyl-protein phosphatase activity in cell fractions was determined by utilizing the unique metal-ion dependence of calmodulin-dependent protein phosphatase. Whereas Ni2+ (1 mM) markedly activated the calmodulin-dependent protein phosphatase, it was found to inhibit potently both particulate and soluble phosphotyrosyl-protein phosphatase activity. In fractions from rat liver, brain and heart, total phosphotyrosyl-protein phosphatase activity against both 32P-labelled receptors was inhibited by 99.5 +/- 6% (mean +/- S.E.M., 30 observations) by Ni2+. Results of Ni2+ inhibition studies were confirmed by other methods. It is concluded that in cell extracts phosphotyrosyl-protein phosphatases other than calmodulin-dependent protein phosphatase are the major phosphotyrosyl-(insulin receptor) and -(EGF receptor) phosphatases.

Evidence that a novel serine kinase catalyses phosphorylation of the insulin receptor in an insulin-dependent and tyrosine kinase-dependent manner

Insulin receptor was co-purified from human placenta together with insulin-stimulated kinase activity that phosphorylates the insulin receptor on serine residues. By using this 'in vitro' system, the mechanism of activation of the serine kinase by insulin was explored. Peptide 1150, histone, poly(Glu-Tyr), eliminating Mn2+ (Mg2+ only), treatment at 37 degrees C (1 h), N-ethylmaleimide, phosphate, beta-glycerol phosphate and anti-phosphotyrosine antibody all inhibited insulin-receptor tyrosine kinase activity and the ability of insulin to stimulate phosphorylation of the insulin receptor on serine. Additionally, direct stimulation of the receptor tyrosine kinase by vanadate increased serine phosphorylation of the insulin receptor. Insulin-stimulated tyrosine phosphorylation preceded insulin-stimulated serine phosphorylation of the insulin receptor. The activity of the insulin-sensitive receptor serine kinase was not augmented by cyclic AMP, cyclic GMP, Ca2+, Ca2+ + calmodulin, Ca2+ + phosphatidylserine + diolein or spermine, or inhibited appreciably by heparin. Additionally, the serine kinase phosphorylated casein or phosvitin poorly and was active with Mn2+. This indicates that it is distinct from Ca2+, Ca2+/phospholipid, Ca2+/calmodulin, cyclic AMP- and cyclic GMP-dependent protein kinases, casein kinases I and II and insulin-activated ribosomal S6 kinase. Taken together, these data indicate that a novel species of serine kinase catalyses the insulin-dependent phosphorylation of the insulin receptor and that activation of this receptor serine kinase by insulin requires an active insulin-receptor tyrosine kinase.