Activation of mRNA translation in rat cardiac myocytes by insulin involves multiple rapamycin-sensitive steps

Insulin acutely activates protein synthesis in ventricular cardiomyocytes from adult rats. In this study, we have established the methodology for studying the regulation of the signaling pathways and translation factors that may be involved in this response and have examined the effects of acute insulin treatment on them. Insulin rapidly activated the 70-kDa ribosomal S6 kinase (p70 S6k), and this effect was inhibited both by rapamycin and by inhibitors of phosphatidylinositol 3-kinase. The activation of p70 S6k is mediated by a signaling pathway involving the mammalian target of rapamycin (mTOR), which also modulates other translation factors. These include the eukaryotic initiation factor (eIF) 4E binding proteins (4E-BPs) and eukaryotic elongation factor 2 (eEF2). Insulin caused phosphorylation of 4E-BP1 and induced its dissociation from eIF4E, and these effects were also blocked by rapamycin. Concomitant with this, insulin increased the binding of eIF4E to eIF4G. Insulin also activated protein kinase B (PKB), which may lie upstream of p70 S6k and 4E-BP1, with the activation of the different isoforms being in the order alpha>beta>gamma. Insulin also caused inhibition of glycogen synthase kinase 3, which lies downstream of PKB, and of eEF2 kinase. The phosphorylation of eEF2 itself was also decreased by insulin, and this effect and the inactivation of eEF2 kinase were attenuated by rapamycin. The activation of overall protein synthesis by insulin in cardiomyocytes was substantially inhibited by rapamycin (but not by inhibitors of other specific signaling pathways, e.g., mitogen-activated protein kinase), showing that signaling events linked to mTOR play a major role in the control of translation by insulin in this cell type.     

Modification of the accessibility of ribosomal proteins after elongation factor 2 binding to rat liver ribosomes and during translocation

Free- and EF-2-bound 80 S ribosomes, within the high-affinity complex with the non-hydrolysable GTP analog: guanylylmethylenediphosphonate (GuoPP(CH2)P), and the low-affinity complex with GDP, were treated with trypsin under conditions that modified neither their protein synthesis ability nor their sedimentation constant nor the bound EF-2 itself. Proteins extracted from trypsin-digested ribosomes were unambiguously identified using three different two-dimensional gel electrophoresis systems and 5 S RNA release was checked by submitting directly free- and EF-2-bound 80 S ribosomes, incubated with trypsin, to two-dimensional gel electrophoresis. Our results indicate that the binding of (EF-2)-GuoPP[CH2]P to 80 S ribosomes modified the behavior of a cluster of five proteins which were trypsin-resistant within free 80 S ribosomes and trypsin-sensitive within the high-affinity complex (proteins: L3, L10, L13a, L26, L27a). As for the binding of (EF-2)-GDP to 80 S ribosomes, it induced an intermediate conformational change of ribosomes, unshielding only protein L13a and L27a. Quantitative release of free intact 5 S RNA which occurred in the first case but not in the second one, should be related to the trypsinolysis of protein(s) L3 and/or L10 and/or L26. Results were discussed in relation to structural and functional data available on the ribosomal proteins we found to be modified by EF-2 binding.     

Cellular mono(ADP-ribosyl) transferase inhibits protein synthesis

A reticulocyte translation system was depleted of functional EF-2 by treatment with diphtheria toxin (DT) fragment A and NAD. After dialysis to remove NAD, the system was reconstituted using preparations of EF-2 derived from pyBHK cells. Untreated and reconstituted lysates permitted similar rates of translation. As expected, when DT-treated EF-2 was used to reconstitute the system, no translation occurred. Furthermore EF-2, reacting with the endogenous ADP-ribosyl transferase from pyBHK cells, was also unable to restore protein synthesis in the reconstituted system. These studies suggest that eukaryotic cellular ADP-ribosyl transferases may play a role in regulating protein synthesis.     

Insulin and prolactin synergistically stimulate beta-casein messenger ribonucleic acid translation by cytoplasmic polyadenylation

Previous studies have shown that the synthesis and stability of milk protein mRNAs are regulated by lactogenic hormones. We demonstrate here in cultured mouse mammary epithelial cells (CID 9) that insulin plus prolactin also synergistically increases the rate of milk protein mRNA translation. Insulin alone stimulates synthesis of both milk and nonmilk proteins, whereas prolactin alone has no effect, but insulin plus prolactin selectively stimulate synthesis of milk proteins more than insulin alone. The increase in beta-casein mRNA translation is also reflected in a shift to larger polysomes, indicating an effect on translational initiation. Inhibitors of the phosphatidylinositol 3-kinase, mammalian target of rapamycin, and MAPK pathways block insulin-stimulated total protein and beta-casein synthesis but not the synergistic stimulation. Conversely, cordycepin abolishes synergistic stimulation of protein synthesis without affecting insulin-stimulated translation. The poly(A) tract of beta-casein mRNA progressively increases from approximately 20 to about 200 A residues over 30 min of treatment with insulin plus prolactin. The 3'-untranslated region of beta-casein mRNA containing an unaltered cytoplasmic polyadenylation element is sufficient for the translational enhancement and mRNA-specific polyadenylation, based on transient transfection of cells with a reporter construct. Insulin and prolactin stimulate cytoplasmic polyadenylation element binding protein phosphorylation with no increase of cytoplasmic poly(A) polymerase activity.     

Differential effects of insulin on peripheral and visceral tissue protein synthesis in neonatal pigs

We recently demonstrated in neonatal pigs that, with amino acids and glucose maintained at fasting levels, the stimulation of protein synthesis in longissimus dorsi muscle with feeding can be reproduced by a physiological rise in insulin alone. In the current report, we determine whether the response of protein synthesis to insulin in the neonatal pig is 1) present in muscles of different fiber types, 2) proportional in myofibrillar and sarcoplasmic proteins, 3) associated with increased translational efficiency and ribosome number, and 4) present in other peripheral tissues and in viscera. Hyperinsulinemic-euglycemic-amino acid clamps were performed in 7- and 26-day-old pigs infused with 0, 30, 100, or 1,000 ng. kg(-0.66). min(-1) of insulin to reproduce insulin levels present in fasted, fed, refed, and supraphysiological conditions, respectively. Tissue protein synthesis was measured using a flooding dose of L-[4-(3)H]phenylalanine. Insulin increased protein synthesis in gastrocnemius muscle and, to a lesser degree, masseter muscle. The degree of stimulation of protein synthesis by insulin was similar in myofibrillar and sarcoplasmic proteins. Insulin increased translational efficiency but had no effect on ribosome number in muscle. All of these insulin-induced changes in muscle protein synthesis decreased with age. Insulin also stimulated protein synthesis in cardiac muscle and skin but not in liver, intestine, spleen, pancreas, or kidney. The results support the hypothesis that insulin mediates the feeding-induced stimulation of myofibrillar and sarcoplasmic protein synthesis in muscles of different fiber types in the neonate by increasing the efficiency of translation. However, insulin does not appear to be involved in the feeding-induced stimulation of protein synthesis in visceral tissues. Thus different mechanisms regulate the growth of peripheral and visceral tissues in the neonate.     

Insulin-dependent covalent reassociation of isolated alpha beta heterodimeric insulin receptors into an alpha 2 beta 2 heterotetrameric disulfide-linked complex

The purified human placental alpha 2 beta 2 heterotetrameric insulin receptor complex was reduced and dissociated into functional alpha beta heterodimers by a combination of alkaline pH and dithiothreitol treatment. Insulin treatment of the isolated alpha beta heterodimeric complex was observed to induce the complete reassociation to an alpha 2 beta 2 heterotetrameric state when analyzed by nondenaturing Bio-Gel A-1.5m gel filtration chromatography. Nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis of 125I-insulin affinity cross-linked and 32P-autophosphorylated alpha beta heterodimers demonstrated that the insulin-dependent reassociation to the alpha 2 beta 2 heterotetrameric state occurred both covalently and noncovalently under these conditions. Comparison by reducing and nonreducing sodium dodecyl sulfate-polyacrylamide gel electrophoresis revealed that the insulin-dependent covalent reassociation to an alpha 2 beta 2 heterotetrameric complex was due to the formation of a disulfide linkage(s) between the alpha beta heterodimers. beta subunit autophosphorylation of the control alpha 2 beta 2 heterotetrameric insulin receptor preparation was maximally stimulated within 5 min of insulin preincubation and occurred exclusively in the Mr = 400,000 alpha 2 beta 2 complex. Similarly, maximal insulin-stimulated beta subunit autophosphorylation of the alpha beta heterodimeric preparation occurred within 5 min of insulin pretreatment in the Mr = 210,000 alpha beta complex. However, 4 h of insulin pretreatment of the alpha beta heterodimer preparation induced the formation (6-fold) of a covalent 32P-labeled alpha 2 beta 2 heterotetrameric complex. Maximal stimulation of substrate phosphorylation for the alpha 2 beta 2 heterotetrameric complex was also observed to occur within 5 min of insulin treatment, whereas maximal insulin-stimulated substrate phosphorylation of the alpha beta heterodimeric complex required greater than 4 h. These data demonstrate that (i) insulin treatment can induce the reassociation of the alpha beta heterodimeric complex into a covalent alpha 2 beta 2 heterotetrameric state, and (ii) insulin-dependent protein kinase activation of the alpha beta heterodimeric insulin receptor correlates with the covalent reassociation into a disulfide-linked alpha 2 beta 2 heterotetrameric complex.     

Insulin stimulates incorporation of 32Pi into nuclear lamins A and C in quiescent BHK-21 cells

The in vivo effect of insulin upon the incorporation of 32Pi into nuclear proteins was investigated in quiescent baby hamster kidney fibroblasts (BHK-21). Insulin has previously been shown to be mitogenic in these cells (Richman, R. A., Johnson, R. A., and Friedman, D. L. (1981) Proc. Soc. Exp. Biol. Med. 168, 196-203). Incorporation into two proteins (Mr = 62,000 and 72,000) in the 0.4 M salt-resistant nuclear fraction was enhanced 2-6-fold by insulin. The effect of insulin (20 nM) was observable within 5 min of treatment, reached a maximum at 15 min, and continued for at least 90 min. The half-maximal effect of insulin was obtained at a concentration of approximately 1 nM. Analysis of nuclear matrix preparations indicated that the two insulin-sensitive proteins were prominent nuclear matrix proteins and suggested that they were lamins A and C. This was confirmed by immunostaining with lamin antibodies and by two-dimensional electrophoresis. These studies indicate that insulin rapidly stimulates the incorporation of phosphate into nuclear lamins A and C in quiescent BHK-21 fibroblasts.     

Insulin modulation of hepatic synthesis and secretion of apolipoprotein B by rat hepatocytes

Insulin inhibition of apolipoprotein B (apoB) secretion by primary cultures of rat hepatocytes was investigated in pulse-chase experiments using [35S]methionine as label. Radioactivity incorporation into apoBH and apoBL, the higher and lower molecular weight forms, was assessed after immunoprecipitation of detergent-solubilized cells and media and separation of the apoB forms using sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Hepatocyte monolayers were incubated for 12-14 h in medium with and without an inhibitory concentration of insulin. Cells were then incubated for 10 min with label, and, after differing periods of chase with unlabeled methionine, cellular medium and media labeled apoB were analyzed; greater than 90% of labeled apoB was present in cells at 10 and 20 min after pulse, and labeled apoB did not appear in the medium until 40 min of chase. Insulin treatment inhibited the incorporation of label into total apoB by 48%, into apoBH by 62%, and into apoBL by 40% relative to other cellular proteins. Insulin treatment favored the more rapid disappearance of labeled cellular apoBH with an intra-cellular retention half-time of 50 min (initial half-life of decay, t1/2 = 25 min) compared with 85 min in control (t1/2 = 60 min). Intracellular retention half-times of labeled apoBL were similar in control and insulin-treated hepatocytes and ranged from 80 to 100 min. After 180 min of chase, 44% of labeled apoBL in control and 32% in insulin-treated hepatocytes remained cell associated. Recovery studies indicated that insulin stimulated the degradation of 45 and 27% of newly synthesized apoBH and apoBL, respectively. When hepatocyte monolayers were continuously labeled with [35S]methionine and then incubated in chase medium with and without insulin, labeled apoBH was secreted rapidly, reaching a plateau by 1 h of chase, whereas labeled apoBL was secreted linearly over 3-5 h of chase. Insulin inhibited the secretion of immunoassayable apoB but not labeled apoB. Results demonstrate that 1) insulin inhibits synthesis of apoB from [35S]methionine, 2) insulin stimulates degradation of freshly translated apoB favoring apoBH over apoBL, and 3) an intracellular pool of apoB, primarily apoBL, exists that is largely unaffected by insulin. Overall, insulin action in primary hepatocyte cultures reduces the secretion of freshly synthesized apoB and favors secretion of preformed apoB enriched in apoBL.     

Insulin and 12-O-tetradecanoylphorbol-13-acetate activation of two immunologically distinct myelin basic protein/microtubule-associated protein 2 (MBP/MAP2) kinases via de novo phosphorylation of threonine and tyrosine residues.

Two site-specific antibodies have been prepared by immunizing rabbits with chemically synthesized peptides derived from the partial cDNA-predicted amino acid sequence of extracellular signal-regulated kinase 1 (ERK1), which has been proposed to encode the microtubule-associated protein 2 (MAP2) kinase (Boulton, T. G., Yancopoulos, G. D., Gregory, J. S., Slauer, C., Moomaw, C., Hsu, J., and Cobb, M. H. (1990) Science 249, 64-67). With immunoprecipitation in the presence of sodium dodecyl sulfate (SDS) and Western blotting, an antibody to the peptide containing triple tyrosine residues (alpha Y91) resembling one of the insulin receptor autophosphorylation sites specifically recognized 42- and 44-kDa proteins. On the other hand, an antibody to the peptide corresponding to the COOH terminus portions (alpha C92) of the ERK1 cDNA gene product recognized the 44-kDa protein much more efficiently than the 42-kDa protein. With immunoprecipitation in the absence of SDS, alpha Y91 could barely recognize these two proteins and alpha C92 recognized the 44-kDa protein but failed to recognize the 42-kDa protein. Kinase assays in myelin basic protein (MBP)-containing gel, after SDS-polyacrylamide gel electrophoresis, revealed that insulin or 12-O-tetradecanoylphorbol-13-acetate (TPA)-stimulated MBP kinase activity in alpha Y91 immunoprecipitates comigrated at molecular mass 42 and 44 kDa. On the other hand, the stimulated MBP kinase activity in alpha C92 immunoprecipitates comigrated only at molecular mass 44 kDa. Insulin stimulated the MBP kinase activity in gels and phosphorylation of these two proteins by greater than 10-fold with a maximal level at 5 min. Insulin and TPA rapidly stimulate the phosphorylation of the 42- and 44-kDa proteins via de novo threonine and tyrosine phosphorylation. Tryptic phosphopeptide mapping analysis of the 42- and 44-kDa proteins, respectively, revealed a single major phosphopeptide containing phosphothreonine and phosphotyrosine, which was common to both insulin- and TPA-stimulated phosphoproteins. Protein phosphatase 2A treatment of these two phosphoproteins caused a complete loss of kinase activity with selective dephosphorylation of phosphothreonine. These data strongly suggest that these two proteins are highly related to the mitogen-activated protein (MAP) kinase with an apparent molecular mass of 42 kDa (Ray, L. B., and Sturgill, T. W. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 3753-3757) and that these two immunologically similar but distinct MBP/MAP2 kinases may represent isozymic forms of MBP/MAP2 kinases. These data also demonstrate that insulin and TPA activate MBP/MAP2 kinase activity by de novo phosphorylation of threonine and tyrosine residues via a very similar pathway.     

Effects of insulin and protein synthesis inhibitors on phospholipid metabolism, diacylglycerol levels, and pyruvate dehydrogenase activity in BC3H-1 cultured myocytes

BC3H-1 myocytes were cultured with 32PO4 for 3 days to label phospholipids to constant specific activity. Subsequent treatment with physiological concentrations of insulin provoked 40-70% increases in 32PO4 levels (reflecting increases in mass) in phosphatidic acid, phosphatidylinositol, and polyphosphoinositides, and, lesser, 20-25% increases in phosphatidylserine and the combined chromatographic area containing phosphatidylethanolamine plus phosphatidylcholine plus phosphatidylcholine. Insulin-induced increases in phospholipids were significant within 5 min and near-maximal at 15-30 min. Comparable rapid insulin-induced increases in [3H]phosphatidylinositol were observed in myocytes prelabeled with [3H]inositol. These insulin effects (as per prolonged pulse-chase experiments) were due to increase phospholipid synthesis rather than decreased phospholipid degradation. Cycloheximide (and puromycin) pretreatment prevented insulin-induced increases in phospholipids and rapidly reversed ongoing insulin effects on phospholipids and pyruvate dehydrogenase activity. Insulin also rapidly increased diacylglycerol levels. These findings suggest that: (a) insulin provokes rapid increases in de novo synthesis of phosphatidic acid and its derivatives, e.g. phosphoinositides and diacylglycerol; (b) protein synthesis inhibitors diminish phospholipid levels in insulin-treated (but not control) tissues by increasing phospholipid degradation (?phospholipase(s) activation); and (c) changes in phospholipids and diacylglycerol may be important for changes in pyruvate dehydrogenase and other enzymatic activities during treatment with insulin and/or protein synthesis inhibitors.     

Post-transcriptional regulation by insulin of Xenopus ribosomal protein S6 kinase

The activation of Xenopus oocyte ribosomal protein S6 kinase during oocyte maturation was investigated. Insulin treatment caused a rapid three-fold activation of S6 kinase that returned to near basal levels by 2 h postinsulin. This was followed by a later fivefold increase from 2 to 5 h with insulin, culminating with germinal vesicle breakdown. Pretreatment of oocytes with multiple protein synthesis inhibitors increased the level of basal activity, but did not greatly alter the time course of early activation of S6 kinase by insulin. In contrast, the later increase in S6 kinase activity was completely inhibited by pretreatment with cycloheximide. However, near maximal increases in S6 kinase activity occurred following injection of maturation-promoting factor, even in the presence of multiple protein synthesis inhibitors. Brief exposure to cycloheximide after 30 min or more of insulin stimulation increased the magnitude of insulin-stimulated activity without changing the overall pattern of activity increase. These results suggest that a rapidly turning-over inhibitor of S6 kinase exists, and the activation of S6 kinase by insulin occurs by protein synthesis-dependent and -independent mechanisms.     

Identification of the major Mr 100,000 substrate for calmodulin-dependent protein kinase III in mammalian cells as elongation factor-2

The major substrate for Ca2+/calmodulin-dependent protein kinase III in mammalian cells is a species of Mr 100,000 that has a primarily cytoplasmic localization. This substrate has now been identified as elongation factor-2 (EF-2), a protein that catalyzes the translocation of peptidyl-tRNA on the ribosome. The amino acid sequence of 18 residues from the N-terminal of the Mr 100,000 CaM-dependent protein kinase III substrate purified from rat pancreas was found to be identical to the N-terminal sequence of authentic rat EF-2 as previously deduced from nucleic acid sequencing of a cDNA (Kohno, K., Uchida, T., Ohkubo, H., Nakanishi, S., Nakanishi, T., Fukui, T., Ohtsuka, E., Ikehara, M., and Okada, Y. (1986) Proc. Natl. Acad. Sci. U.S.A. 83, 4978-4982). CaM-dependent protein kinase III phosphorylated EF-2 in vitro with a stoichiometry of approximately 1 mol/mol on a threonine residue. Amino acid sequencing of the purified tryptic phosphopeptide revealed that this threonine residue lies within the sequence: Ala-Gly-Glu-Thr-Arg-Phe-Thr-Asp-Thr-Arg (residues 51-60 of EF-2). The Mr 100,000 protein was stoichiometrically ADP-ribosylated in vitro by the addition of diphtheria toxin and NAD. The Mr 100,000 protein was photoaffinity labeled with a GTP analog and the protein had an endogenous GTPase activity that could be stimulated by the addition of salt-washed ribosomes. These properties are all characteristic of EF-2. Dephospho-EF-2 could support poly(U)-directed polyphenylalanine synthesis in a reconstituted elongation system when combined with EF-1. In the same system, phospho-EF-2 was virtually inactive in supporting polypeptide synthesis; this effect could be reversed by dephosphorylation of phospho-EF-2. These results suggest that intracellular Ca2+ inhibits protein synthesis in mammalian cells via CaM-dependent protein kinase III-catalyzed phosphorylation of EF-2.     

Glucose regulates EF-2 phosphorylation and protein translation by a protein phosphatase-2A-dependent mechanism in INS-1-derived 832/13 cells

The role of elongation factor (EF)-2 phosphorylation in the regulation of pancreatic beta-cell protein synthesis by glucose was investigated in the INS-1-derived cell line 832/13. Incubation of cells in media containing 1 mm glucose resulted in a progressive increase in EF-2 phosphorylation that was maximal by 1-2 h. Readdition of 10 mm glucose promoted a rapid dephosphorylation of EF-2 that was complete in 10 min and maintained over the ensuing 2 h. Similar results were obtained using primary rat islets or Min-6 insulinoma cells. The glucose effect in 832/13 cells was replicated by addition of pyruvate or alpha-ketocaproate, but not 2-deoxyglucose, suggesting that mitochondrial metabolism was required. Accordingly, glucose-mediated dephosphorylation of EF-2 was completely blocked by the mitochondrial respiratory antagonists antimycin A and oligomycin. The hyperglycemic effect was not mimicked by incubation of cells in 100 nm insulin, 30 mm potassium chloride, or 0.25 mm diazoxide, indicating that insulin secretion and/or depolarization of beta cells was not required. The locus of the high glucose effect appeared to be protein phosphatase-2A, the principal phosphatase acting on EF-2. Protein phosphatase-2A activity was stimulated by glucose addition to 832/13 cells, but neither protein phosphatase-1 nor calmodulin kinase III (EF-2 kinase) activity was affected under these conditions. The slower rephosphorylation of EF-2 during the transition from high to low glucose may involve effects on EF-2 kinase activity. Addition of 5-aminoimidazole-4-carboxamide 1-beta-d-ribofuranoside in high glucose led to a marked stimulation of EF-2 phosphorylation, consistent with the possibility that increased AMP kinase activity in low glucose stimulates EF-2 kinase. In parallel with the effects on EF-2 dephosphorylation, addition of high glucose to 832/13 cells markedly increased the incorporation of [(35)S]methionine into total protein. Taken together, these results suggest that modulation of extracellular glucose impacts protein translation rate in beta cells at least in part through regulation of the elongation step, via phosphorylation/dephosphorylation of EF-2.     

Mutations in Elongation Factor 1β, a Guanine Nucleotide Exchange Factor, Enhance Translational Fidelity

Translation elongation factor 1beta (EF-1beta) is a member of the family of guanine nucleotide exchange factors, proteins whose activities are important for the regulation of G proteins critical to many cellular processes. EF-1beta is a highly conserved protein that catalyzes the exchange of bound GDP for GTP on EF-1alpha, a required step to ensure continued protein synthesis. In this work, we demonstrate that the highly conserved C-terminal region of Saccharomyces cerevisiae EF-1beta is sufficient for normal cell growth. This region of yeast and metazoan EF-1beta and the metazoan EF-1beta-like protein EF-1delta is highly conserved. Human EF-1beta, but not human EF-1delta, is functional in place of yeast EF-1beta, even though both EF-1beta and EF-1delta have previously been shown to have guanine nucleotide exchange activity in vitro. Based on the sequence and functional homology, mutagenesis of two C-terminal residues identical in all EF-1beta protein sequences was performed, resulting in mutants with growth defects and sensitivity to translation inhibitors. These mutants also enhance translational fidelity at nonsense codons, which correlates with a reduction in total protein synthesis. These results indicate the critical function of EF-1beta in regulating EF-1alpha activity, cell growth, translation rates, and translational fidelity.     

cAMP-dependent activation of protein synthesis correlates with dephosphorylation of elongation factor 2

The addition of 5 mM cAMP to a cell-free translation system from rabbit reticulocytes increases the rate of protein synthesis 3 5-fold. Lower concentrations of cAMP (0.005, 0.05 and 0.5 mM) have no effect on translation in this system. cAMP at all the concentrations tested stimulates the phosphorylation of the same pattern of polypeptides, while 5 mM cAMP additionally stimulates dephosphorylation of the 95 kDa polypeptide identified as elongation factor 2 (EF-2). Testing of the preparations of EF-2 with a different content of the phosphorylated form in poly(U)-directed poly(Phe) synthesis reveals that the EF-2 activity correlates with the fraction of non-phosphorylated EF-2. Thus cAMP-dependent activation of protein synthesis seems to be due to dephosphorylation of EF-2.