Induction of sepsis in rats causes an inhibition of protein synthesis in skeletal muscle that is resistant to the stimulatory actions of insulin. To gain a better understanding of the underlying reason for this lack of response, the present study was undertaken to investigate sepsis-induced alterations in insulin signaling to regulatory components of mRNA translation. Experiments were performed in perfused hindlimb preparations from rats 5 days after induction of a septic abscess. Sepsis resulted in a 50% reduction in protein synthesis in the gastrocnemius. Protein synthesis in muscles from septic rats, but not controls, was unresponsive to stimulation by insulin. The insulin-induced hyperphosphorylation response of the translation repressor protein 4E-binding protein 1 (4E-BP1) and of the 70-kDa S6 kinase (S6K1) (1), two targets of insulin action on mRNA translation, was unimpaired in gastrocnemius of septic rats. Hyperphosphorylation of 4E-BP1 in response to insulin resulted in its dissociation from the inactive eukaryotic initiation factor (eIF)4E · 4E-BP1 complex in both control and septic rats. However, assembly of the active eIF4F complex as assessed by the association of eIF4E with eIF4G did not follow the pattern predicted by the increased availability of eIF4E resulting from changes in the phosphorylation of 4E-BP1. Indeed, sepsis caused a dramatic reduction in the amount of eIF4G associated with eIF4E in the presence or absence of insulin. Thus the inability of insulin to stimulate protein synthesis during sepsis may be related to a defect in signaling to a step in translation initiation involved in assembly of an active eIF4F complex.
- translation initiation
- eukaryotic initiation factors 4G and 4E
- 4E-binding protein 1
- S6 kinase
insulin resistance in skeletal muscle is one characteristic of the host's response to severe infection. Both glucose and protein metabolism in skeletal muscle show a diminished responsiveness to the anabolic actions of insulin during sepsis. With respect to protein metabolism, the ability of insulin to limit the release of amino acids from extrasplanchnic organs of septic rats is less than that observed in nonseptic rats (31). The release of amino acids from skeletal muscle during sepsis is dependent on a balance between rates of protein synthesis and protein degradation. Insulin is known to modulate both of these metabolic processes in various physiological and pathological conditions (for review see Refs. 4 and 21). Proteolysis in skeletal muscle consistently shows a relative resistance to inhibition by insulin during the immediate posttrauma period (30), acute (16-h) peritonitis (10, 11), or chronic bacteremia (33). Likewise, insulin fails to augment skeletal muscle protein synthesis after chronic (5-day) intra-abdominal sepsis (15).
The mechanisms through which insulin stimulates protein synthesis in skeletal muscle are beginning to be elucidated. Insulin stimulates protein synthesis, in part, by accelerating the rate of mRNA translation. Two steps involved in translation initiation appear important as major regulatory points in the overall control of protein synthesis in vivo. The first is the binding of met-tRNA to the 40S ribosomal subunit to form the 43S preinitiation complex. This step is mediated by eukaryotic initiation factor 2 (eIF2) and is regulated by the activity of another initiation factor, eIF2B. The second regulatory step involves the binding of mRNA to the 43S preinitiation complex, which is mediated by the cap-binding protein eIF4E in association with a second factor, eIF4G. Insulin has been shown to modulate both of these steps in the skeletal muscle of nonseptic rats (18-21).
Insulin stimulates the mRNA binding step in translation initiation, in part, by enhancing the availability of eIF4E, which is modulated by its association with a small acid- and heat-labile protein, termed 4E-BP1 (PHAS-I) (19, 20). The interaction between 4E-BP1 and eIF4E is regulated by the extent of phosphorylation of 4E-BP1, which in turn is under control by insulin (19, 20, 22). Hypophosphorylated 4E-BP1 binds to eIF4E, sequestering the protein in an inactive 4E-BP1 · eIF4E complex. When eIF4E is bound to 4E-BP1, eIF4E binds to mRNA but cannot form an active eIF4E · eIF4G complex (23, 26), thereby preventing binding of mRNA to the ribosome. Insulin enhances the phosphorylation of 4E-BP1, thereby releasing eIF4E from the 4E-BP1 · eIF4E complex and allowing the eIF4E · mRNA complex to bind to eIF4G and then to the 43S preinitiation complex (23).
Insulin also acts on translation initiation, in part, by enhancing the activity of S6K1, a serine/threonine kinase that phosphorylates the ribosomal protein S6 (8). The ribosomal protein S6 is uniquely positioned to regulate translation of mRNA by its location at the tRNA binding site on the 40S ribosome. Increased phosphorylation of ribosomal S6 through activation of S6K1 correlates with accelerated rates of translation of mRNAs containing a 5′-terminal oligopyrimidine (TOP) sequence (12, 13). Activation of S6K1 by insulin correlates with hyperphosphorylation of the protein (9).
The present investigations were designed to examine the ability of insulin to modulate the phosphorylation status of both 4E-BP1 and S6K1 in the gastrocnemius of septic rats and to compare the response with that of controls by use of a perfused hindlimb preparation. The results indicate that insulin signaling to 4E-BP1 and S6K1 is relatively unimpaired in skeletal muscle of septic rats. In contrast, assembly of the active eIF4F complex, as measured by the interaction of eIF4E with eIF4G, was diminished in both the basal and insulin-stimulated conditions during sepsis. Thus a sepsis-induced defect in insulin signaling to mediate the assembly of the active eIF4F complex may account for the lack of action of the hormone on protein synthesis under these circumstances.
Adult male Sprague-Dawley rats were maintained on a 12:12-h light-dark cycle. Rats weighing 200–300 g were anesthetized with a combination of ketamine (110 mg/kg body wt) and acepromazine (1 mg/kg body wt). Sepsis was then induced by implanting a sterilized fecal-agar pellet that was inoculated with 103 colony-forming units (CFU) Escerichia coli and 109 CFUBacteroides fragilis into the abdominal cavity (5, 15, 16, 20, 31, 32, 38). The intra-abdominal septic abscess was allowed to develop for 5 days. Laparotomies were not performed in control rats. Previous studies provided evidence that rates of protein synthesis in skeletal muscle are not depressed 5 days after laparotomy or creation of a sterile abscess compared with muscle in nonoperated animals (15, 32, 38).
Hindlimb perfusion technique.
Hindlimb perfusions were carried out according to the method described by Bylund-Fellenius et al. (3) as modified in our laboratory (15-17, 20). Five days after implantation of the fecal-agar pellet, rats were anesthetized with pentobarbital sodium (50 mg/kg body wt), and the skin covering the right and left hindlimbs was removed. A midline incision was made, and both the inferior vena cava and the abdominal aorta were exposed. The abdominal aorta was cannulated, and perfusate was delivered immediately via the abdominal aorta at a rate of 0.32 ml · min−1 · g−1(3) to the hindlimb musculature. The inferior vena cava was then cannulated. The first 50 ml of perfusate passing through the hindlimb were discarded. The cannula in the inferior vena cava was then connected to the perfusion system, and recirculation of the perfusate was begun. After perfusion for an additional 35 min, [3H]phenylalanine was introduced into the perfusate to give a final concentration of 2 μCi/ml, and perfusion continued for 60 min. After perfusion, gastrocnemius muscles were frozen between aluminum blocks precooled to the temperature of liquid nitrogen or used directly for analysis of eukaryotic initiation factors. A perfusate sample was withdrawn and centrifuged to remove red blood cells. The perfusate plasma samples were stored at −20°C until analyzed for the determination of phenylalanine specific radioactivity, as described previously (15-17, 20).
The perfusate (250 ml/hindlimb) consisted of a modified Krebs-Henseleit bicarbonate buffer containing 30% (vol/vol) washed bovine erythrocytes, 4.5% (wt/vol) bovine serum albumin (BSA fraction V), 11 mM glucose, 1 mM phenylalanine, and all other amino acids at normal rat plasma concentration, as previously described (14, 26). When present, insulin was added to give a final concentration of 100 μU/ml. The medium was maintained at 37°C and gassed with humidified 95% O2-5% CO2.
Measurement of protein synthesis.
Rates of protein synthesis were estimated by the incorporation of [3H]phenylalanine into muscle proteins (5,15-17, 32, 33). Frozen tissue was powdered under liquid nitrogen. A portion (0.5 g) of powdered tissue was homogenized in 5 volumes of ice-cold 10% (wt/vol) trichloroacetic acid (TCA), and the homogenate was centrifuged at 10,000 g for 11 min at 4°C. The resulting supernatant was decanted, and the pellet was washed five times with 10% TCA to remove any acid-soluble radioactivity. After a wash with acetone and then with water, the pellet was dissolved in 0.1 N NaOH. Aliquots were assayed for protein using the Biuret method, with crystalline BSA serving as the protein standard. Another aliquot was assayed for radioactivity by liquid scintillation spectrometry with corrections for quenching (in dpm).
The rate of protein synthesis, expressed as nanomoles of phenylalanine incorporated into protein per hour per gram of muscle (nmol Phe · g−1 · h−1), was determined by dividing the disintegrations per hour incorporated in muscle by the perfusate phenylalanine specific radioactivity. At perfusate concentrations >0.8 mM, the specific radioactivity of tRNA-bound phenylalanine is the same as that of the extracellular and intracellular pools of free phenylalanine (3). Therefore, the specific radioactivity of perfusate phenylalanine provides an accurate estimate of the specific radioactivity of phenylalanine t-RNA.
Determination of phosphorylation state of 4E-BP1.
The various phosphorylated forms of 4E-BP1 were measured after immunoprecipitation of 4E-BP1 from muscle homogenates (19,20). 4E-BP1 was immunoprecipitated from the supernatants as described previously (20, 35). The immunoprecipitates were solubilized with SDS sample buffer. The various phosphorylated forms of 4E-BP1 were separated by electrophoresis. After electrophoresis, the proteins were transferred electrophoretically onto polyvinylidene difluoride (PVDF) membranes. The membranes were incubated with polyclonal antibodies specific for 4E-BP1 for 1 h at room temperature. The blots were then developed using an enhanced chemiluminescence (ECL) Western blotting kit by following the manufacturer's instructions. Films were scanned using a Microtek ScanMaker III scanner equipped with a transparent media adapter connected to a Macintosh computer. Images were obtained using the ScanWizard Plugin (Microtek) for Adobe Photoshop, quantitated using National Institutes of Health Image 1.60 software and by protein immunoblot analysis as described previously (20, 22, 35).
Determination of phosphorylation state of S6K1.
For determination of the phosphorylation state of S6K1, supernatants from a 10,000-g centrifugation were subjected to SDS-PAGE, transferred to PVDF membranes, and immunoblotted with a polyclonal antibody that recognizes S6K1 (Santa Cruz Biotechnology, Santa Cruz, CA) as described previously (1, 8). The blots were then developed using ECL. Films were scanned and quantitated as described previously (6, 34). There were no significant differences in the amount of SK61 in muscles from septic rats [146 ± 18 arbitrary units (AU)/mg protein; n = 9] compared with control animals (127 ± 18 AU/mg protein; n = 10).
Quantification of 4E-BP1 · eIF4E and eIF4G · eIF4E complexes.
The association of eIF4E with 4E-BP1 and eIF4G was determined as described previously (29). Briefly, hindlimb muscles were rapidly removed, immediately weighed, and homogenized in 7 volumes ofbuffer A [20 mM HEPES, pH 7.4, 100 mM KCl, 0.2 mM EDTA, 2 mM EGTA, 1 mM dithiothreitol, 50 mM NaF, 50 mM β-glycerolphosphate, 0.1 mM phenylmethylsulfonyl fluoride (PMSF), 1 mM benzamidine, 0.5 mM sodium vanadate, and 1 μM microcystin LR] with a Polytron homogenizer. The homogenate was centrifuged at 10,000 g for 10 min at 4°C. eIF4E and 4E-BP1 · eIF4E and eIF4G · eIF4E complexes were immunoprecipitated from aliquots of 10,000-g supernatants with an anti-eIF4E monoclonal antibody. The antibody/antigen complex was collected by incubation for 1 h with goat anti-mouse Biomag IgG beads (PerSeptive Diagnostics). Before use, the beads were washed in 1% nonfat dry milk in buffer B (50 mM Tris · HCl, pH 7.4, 150 mM NaCl, 5 mM EDTA, 0.1% β-mercaptoethanol, 0.5% Triton X-100, 50 mM NaF, 50 mM β-glycerolphosphate, 0.1 mM PMSF, 1 mM benzamidine, and 0.5 mM sodium vanadate). The beads were captured using a magnetic stand and were washed twice with buffer B and once with buffer B containing 500 mM NaCl rather than 150 mM NaCl. Resuspending the beads in SDS-sample buffer and boiling the sample for 5 min eluted protein bound to the beads. The beads were collected by centrifugation, and the supernatants were subjected to electrophoresis, either on a 10% polyacrylamide gel for quantitation of eIF4G or on a 15% polyacrylamide gel for quantitation of 4E-BP1 and eIF4E. Proteins were then electrophoretically transferred to a PVDF membrane, as described previously (19, 20). The membranes were incubated with a mouse anti-human eIF4E antibody, a rabbit anti-rat 4E-BP1 antibody, or a rabbit anti-eIF4G antibody for 1 h at room temperature. The blots were then developed using an ECL Western blotting kit by following the manufacturer's instructions. Films were scanned and quantitated as described in Determination of phosphorylation state of 4E-BP1.
Determination of eIF4E phosphorylation state.
The phosphorylated and unphosphorylated forms of eIF4E in extracts of hindlimb muscles were separated by isoelectric focusing on a slab gel and quantitated by protein immunoblot analysis, as described previously (19, 20, 29).
Values shown are means ± SE. Statistical evaluation of the data was performed using analysis of variance (ANOVA) to test for overall differences among groups, followed by the Sidak test for multiple comparisons to determine significance between means only when ANOVA indicated a significant difference among the group means. Differences among the means were considered significant when P was <0.05.
The effect of insulin on protein synthesis in gastrocnemius muscle during sepsis was studied using the isolated perfused hindlimb. The use of the perfused hindlimb preparation allows for the direct examination of the effects of insulin on the regulation of translation initiation in muscle independent of alterations in other hormones and/or metabolites that occur in vivo. The gastrocnemius was chosen because previous studies have established that protein synthesis is regulated differently in various muscle types during sepsis (38). Sepsis causes an inhibition of protein synthesis in muscle composed primarily of fast-twitch fibers (i.e., gastrocnemius, psoas), whereas in muscles composed of slow-twitch fibers (i.e., soleus), protein synthesis is unaffected (4, 5, 38). More importantly, insulin fails to stimulate protein synthesis in perfused gastrocnemius of septic rats (15).
The stimulatory effect of insulin on protein synthesis in gastrocnemius from control, fed rats was confirmed (Fig.1). Insulin at a concentration of 100 μU/ml stimulated protein synthesis ∼50% compared with muscle perfused in the absence of the hormone, when physiological concentrations of amino acids were provided (P < 0.05). Protein synthesis was reduced in gastrocnemius from septic rats in both the basal and insulin-stimulated conditions compared with controls. In the basal state, protein synthesis was reduced to ∼50% of control values in gastrocnemius from septic rats (P< 0.05). In contrast to control animals, insulin failed to significantly stimulate protein synthesis in gastrocnemius of septic rats. Moreover, the rate of protein synthesis in gastrocnemius of septic rats treated with insulin remained significantly (P < 0.05) less than that observed in control muscle.
One explanation for the failure of insulin to stimulate protein synthesis in gastrocnemius of septic rats would be that there is a general defect in insulin signaling. 4E-BP1 undergoes multiple phosphorylations, which are characterized by reduced mobility during electrophoresis. 4E-BP1 is resolved into multiple electrophoretic forms termed α, β, and γ, representing differently phosphorylated forms of the protein (23, 26). In control rats, insulin augments the abundance of the most highly phosphorylated form of the protein, 4E-BP1γ (23, 26). Therefore, we investigated the effects of sepsis and insulin on the abundance of 4E-BP1γ (Fig.2). Under basal conditions, sepsis induced a 50% decrease in the proportion of 4E-BP1 in the γ-form. Insulin caused a significant increase in the abundance of 4E-BP1γ in muscles from both control and septic rats. There were no significant differences in the proportion of 4E-BP1 in the γ-form in insulin-treated muscles between control (55 ± 7%) and septic rats (63 ± 3%). Thus insulin signaling to 4E-BP1 appears unimpaired in sepsis.
Phosphorylation of S6K1.
Insulin also activates S6K1. Like 4E-BP1, S6K1 is activated by multisite phosphorylation that results in hyperphosphorylated forms of the protein. When subjected to SDS-PAGE, S6K1 resolves into multiple electrophoretic forms (for review see Ref. 9), with increased phosphorylation corresponding to a reduced electrophoretic mobility. The slowest migrating form represents S6K1 phosphorylated on multiple residues including Thr389, a residue whose phosphorylation is associated with activation of the protein (41). We used this property (assessed by immunoblotting techniques) as an indicator of the effect of insulin on the activation of S6K1 in perfused skeletal muscle from control and septic rats. Perfusion with buffer containing insulin enhanced the prominence of the more electrophoretically retarded bands in skeletal muscles from both control and septic rats treated with insulin (Fig.3). The percentage of S6K1 in the most phosphorylated form was significantly elevated in gastrocnemius perfused in the presence of insulin from either control [+Insulin: 11 ± 3% (n = 5) vs. −Insulin: 0.19 ± 0.19% (n = 5); P < 0.05] or septic [+Insulin: 24 ± 6% (n = 4) vs. −Insulin: 0 ± 0% (n = 4); P < 0.01] compared with values obtained in muscles perfused in the absence of the hormone (Fig. 3). Likewise, the percentage of S6K1 in the least phosphorylated form was significantly reduced in gastrocnemius perfused in the presence of insulin from either control [+Insulin: 15 ± 3% (n = 5) vs. −Insulin: 61 ± 9% (n = 5); P < 0.05] or septic [+Insulin: 7 ± 2% (n = 4) vs. −Insulin: 53 ± 13% (n = 4); P < 0.01] compared with values obtained in muscles perfused in the absence of the hormone. These observations indicate a net increase in the phosphorylation state of the protein from either control or septic rats. Thus insulin signaling to S6K1 also appears unimpaired in sepsis.
Regulation of eIF4F complex assembly.
Insulin administration has been proposed to stimulate global rates of protein synthesis by enhancing the assembly of the active eIF4F complex through increased availability of eIF4E (19, 20, 22). The changes in the extent of phosphorylation of 4E-BP1γ would be expected to modulate the association of 4E-BP1 with eIF4E in sepsis and after insulin treatment. Therefore, we examined whether eIF4E availability for eIF4F complex assembly was increased to a similar extent after stimulation by insulin in control or septic rats. To investigate this possibility, eIF4E immunoprecipitates were analyzed for 4E-BP1 content (Fig. 4). The relative amount of 4E-BP1 was normalized by dividing the values for 4E-BP1 by the amount of eIF4E present in the immunoprecipitate. As shown in Fig. 4, an increased association of 4E-BP1 with eIF4E was observed in muscles from septic rats compared with controls. Perfusion with insulin resulted in a reduction in the amount of 4E-BP1 bound to eIF4E in both control and septic rats. There were no significant differences in the amount of 4E-BP1 bound to eIF4E in septic rats perfused with insulin compared with controls perfused with insulin. In addition, the differences in association of 4E-BP1 with eIF4E were not the result of significant differences in the amounts of 4E-BP1 [Control = 3,127 ± 258 arbitrary units (n = 10); Septic = 3,794 ± 242 arbitrary units (n = 9)] between the two groups.
The reduced amounts of 4E-BP1 bound to eIF4E in both control and septic rats after perfusion with insulin would be expected to increase the availability of eIF4E and thereby enhance the formation of the eIF4G · eIF4E complex. To investigate the effects of insulin on the association of eIF4G with eIF4E during sepsis, eIF4E immunoprecipitates were used to measure the amount of eIF4G that coprecipitated with eIF4E (Fig. 5). The relative amount of eIF4G was normalized by dividing the values for eIF4G by the amount of eIF4E present in the immunoprecipitate. Perfusion of muscles from control animals with insulin resulted in a twofold elevation in the amount of eIF4G bound to eIF4E. Sepsis significantly decreased the amount of eIF4G associated with eIF4E in the basal state (P < 0.05) compared with control rats. Insulin caused a 5.5-fold increase (P < 0.005) in the amount of eIF4G associated with eIF4E in muscles from septic rats compared with the basal state. Even with this large increase in the binding of eIF4G with eIF4E in muscles from septic rats, the values were only 50% of those obtained in control animals perfused with insulin.
The differences in the amount of eIF4G associated with eIF4E could have been due to a reduced abundance of eIF4G in muscles from septic rats treated with insulin compared with controls. Therefore, we measured the amount of eIF4G in gastrocnemius from control and septic rats treated with insulin. There were no significant differences in the eIF4G content in muscles from these two groups [Control+Insulin 165 ± 22 AU/mg protein (n = 6) vs. Septic+Insulin 158 ± 22 AU/mg protein (n = 5)]. Thus differences in the relative abundance of eIF4G associated with eIF4E were not the result of a reduced amount of eIF4G in the muscles from septic rats treated with insulin. In addition, the diminished amount of eIF4E · eIF4G complex did not result from a decreased amount of eIF4E, because total gastrocnemius eIF4E content was not significantly altered (data not shown).
Assembly of the eIF4E · eIF4G complex was significantly diminished in skeletal muscle from septic rats perfused in the absence of insulin. Reduced amounts of eIF4E associated with eIF4G after chronic sepsis would be expected to result in a diminished association of mRNA with the ribosome and, hence, to limit protein synthesis. Therefore, we examined the relationship between rates of protein synthesis and amount of eIF4G associated with eIF4E in muscles from control animals treated with and without insulin and from septic rats perfused in the absence of insulin (Fig.6). There was a positive linear relationship (r 2 = 0.9997) between rates of protein synthesis and amount of eIF4G associated with eIF4E, with a slope significantly different from zero (F = 3,531;P < 0.01). Although this correlation does not prove cause and effect, the relationship between protein synthesis and amount of eIF4G associated with eIF4E is consistent with the proposed role of the eIF4E · eIF4G complex in the overall regulation of protein synthesis.
To further define potential mechanisms responsible for the sepsis-induced inhibition of protein synthesis, the effect of sepsis on phosphorylation of eIF4E was examined (Fig.7). In these studies, slab gel isoelectric focusing was used to separate phosphorylated and nonphosphorylated forms of eIF4E, followed by immunoblot analysis to quantitate the amount of eIF4E in the two forms. In muscles from control rats, ∼85% of eIF4E was in the phosphorylated form. Neither insulin nor sepsis significantly affected the proportion of eIF4E in the phosphorylated form. Therefore, it is unlikely that the failure of insulin to augment rates of protein synthesis during sepsis occurred as a result of alterations in the phosphorylation state of eIF4E.
We have previously established that chronic (5-day) sepsis causes a reduction in global rates of protein synthesis in skeletal muscle (15-17, 32, 38) that results from decreased translation initiation (15, 36, 38-40). It has been suggested that one way to augment protein synthesis during sepsis would be through administering anabolic hormones. Insulin enhances protein synthesis in skeletal muscle of nonseptic rats through an acceleration of translation initiation (for review see Refs. 4 and 21). For example, decreased rates of protein synthesis secondary to fasting or diabetes can be stimulated by provision of insulin in both rat heart (14) and skeletal muscle (18, 19). Despite the ability of insulin to increase protein synthesis in a variety of conditions, skeletal muscle from septic animals shows a relative resistance to the anabolic effects of the hormone on protein synthesis (15). The response of protein synthesis to increasing insulin concentrations in the gastrocnemius of septic rats is characterized by both a decreased sensitivity and a decreased maximal responsiveness compared with control rats (15). The potential mechanism for the failure of insulin to enhance translation initiation in muscles from septic rats has not been elucidated.
A general defect in insulin signaling would be one explanation for the failure of insulin to stimulate protein synthesis in skeletal muscle of septic animals. In nonseptic rats, insulin induces the phosphorylation of 4E-BP1 (19, 20, 29). Hence, functional insulin signaling in skeletal muscle from septic rats perfused with the hormone should be associated with an increase in the extent of phosphorylation of 4E-BP1. In both control and septic animals, insulin caused a marked increase in the proportion of 4E-BP1 in the γ-phosphorylated form to a similar extent. The insulin-induced stimulation of 4E-BP1 phosphorylation is thought to occur after activation of the phosphatidylinositol 3-kinase pathway through increased activity of a protein kinase termed the mammalian target of rapamycin (mTOR). The results of the present set of experiments suggest that insulin signaling through mTOR activity is maintained in septic rats.
S6K1 is a serine/threonine kinase that phosphorylates the ribosomal protein S6, and it has been implicated in stimulating skeletal muscle protein synthesis after insulin treatment (8). Insulin stimulates the activity of S6K1 and results in multisite phosphorylation of the protein (9). It is currently thought that phosphorylation of S6K1, like that of 4E-BP1, lies distal to mTOR in the signaling pathway for insulin. Hence, stimuli that affect the activity of mTOR should influence the extent of phosphorylation of both S6K1 and 4E-BP1. Because the phosphorylation of 4E-BP1 was enhanced to a similar extent after stimulation with insulin in control and septic rats, we also examined the extent of phosphorylation of S6K1. In muscles from both control and septic rats, insulin caused an increased phosphorylation of S6K1. Thus the insulin-signaling pathway leading to enhanced phosphorylation of both S6K1 and 4E-BP1, presumably through mTOR, is activated by insulin in muscles from septic rats despite a failure of the hormone to augment protein synthesis.
Anabolic hormones, such as insulin or insulin-like growth factor I, stimulate global rates of protein synthesis in skeletal muscle, in part, through changes in the regulation of eIF4F (20, 35). eIF4F is a protein complex composed of eIF4E, the 7-methyl-GTP cap-binding protein, and eIF4G, a scaffold protein with binding sites for other proteins, including eIF4E and the third component of eIF4F, the ATP-dependent helicase eIF4A. The translation of mRNA involves recruitment of the mRNA to the 43S preinitiation ribosomal subunit. The efficacy of the mRNA binding to the ribosome may be controlled through the distribution of eIF4E between active eIF4E · eIF4G and inactive eIF4E · 4E-BP1 complexes (19, 20, 22,36). When eIF4E is bound to 4E-BP1, eIF4E binds to mRNA but cannot form an active eIF4E · eIF4G complex (for review see Refs. 4 and 27). In this scenario, formation of the 4E-BP1 · eIF4E complex prevents binding of mRNA to the ribosome. Insulin causes release of eIF4E from 4E-BP1 and promotes the assembly of the active eIF4E · eIF4G complex (19,20). Therefore, it might be expected that modulation of eIF4E distribution between the active eIF4E · eIF4G complex and inactive eIF4E · 4E-BP1 complex would contribute to the failure of insulin to stimulate protein synthesis in skeletal muscle during sepsis.
In the present set of experiments, insulin caused an increase in the abundance of eIF4G bound to eIF4E compared with the basal state in muscles from both control and septic rats. The amount of eIF4G bound to eIF4E in muscles from septic rats perfused with insulin was approximately the same as was observed in controls in the basal condition and 50% of that observed in controls treated with insulin. Thus insulin failed to raise the amount of eIF4G bound to eIF4E to values observed in skeletal muscle from control animals treated with insulin. These findings indicate that sepsis prevents insulin from enhancing the association of eIF4E to eIF4G to the same extent as controls.
One potential mechanism to account for the inability of insulin to increase the amount of eIF4E bound to eIF4G would be through a failure of eIF4E to be released from the inactive 4E-BP1 · eIF4E complex. Addition of insulin to the perfusion buffer diminished the abundance of 4E-BP1 associated with eIF4E to a similar extent in muscles from both control and septic rats. The decrease in the amount of 4E-BP1 associated with eIF4E after insulin was inversely correlated with increases in the extent of 4E-BP1 phosphorylation. Thus insulin induced a marked reduction in the binding of eIF4E to 4E-BP1 to the same extent in control and septic rats, presumably by freeing eIF4E for binding to eIF4G. However, although eIF4E binding to 4E-BP1 was reduced in muscles from septic rats treated with insulin, formation of the eIF4E · eIF4G complex was not restored to values observed in control rats perfused with insulin. Thus, in muscles from septic rats, association of eIF4E with 4E-BP1 accounts for only a portion of the change in eIF4F assembly. Moreover, the results suggest that additional mechanisms must be operative to repress eIF4E binding to eIF4G during sepsis.
For a given amount of eIF4G bound to eIF4E (Fig. 6) or enhanced phosphorylation of S6K1 in muscles from septic animals perfused with insulin, the rate of protein synthesis was reduced compared with that of controls. Several possible scenarios can be hypothesized to account for this finding. First, the changes in eIF4E binding to eIF4G may be associated with altered translation of specific mRNAs during sepsis. Enhanced formation of the eIF4E · eIF4G complex in cells in culture results in the preferential stimulation in the synthesis of specific proteins (e.g., ornithine decarboxylase or myc) encoded by mRNAs that typically contain a high degree of secondary structure in the 5′-untranslated region. These proteins may reflect only a small percentage of the proteins synthesized, and, hence, an accelerated rate of translation of mRNAs encoding for these proteins would not be of significant magnitude to augment global rates of protein synthesis. Likewise, phosphorylation of S6K1 enhances synthesis of proteins whose mRNAs contain an oligopyrimidine tract in the 5′ terminus (TOP mRNAs). These proteins include elongation factors and ribosomal proteins. In support of this scenario, inactivation of S6K1 caused by inhibition of mTOR with rapamycin results in only a modest decrease in the global rate of protein synthesis in skeletal muscles after treatment with insulin (7, 8). These findings indicate that much of the stimulatory effect of insulin on global rates of protein synthesis in skeletal muscle may be independent of rapamycin-sensitive pathways (7, 8). More importantly, the data suggest that sepsis modulates an mTOR-independent pathway, limiting not only protein synthesis in the basal condition but also the ability of insulin to stimulate the process.
Second, sepsis may induce modifications in eIF4G or other proteins that render the protein less likely to bind to eIF4E. For example, incubation of 293T cells with serum induces phosphorylation of eIF4G (28). In addition to phosphorylation, eIF4G can be cleaved by several proteases, including caspase 3 (2, 24) and calpains (25). There may also be unknown proteins induced or activated during sepsis that interact with eIF4G to prevent the ability of eIF4G to bind with eIF4E. However, whether any of these potential changes modulates the ability of eIF4G to initiate translation remains to be determined.
Third, we have previously reported that formation of the 43S preinitiation complex is markedly reduced in skeletal muscle during sepsis (38-40). Furthermore, the decreased formation of the 43S preinitiation complex is associated with a diminished activity and expression of eIF2B during chronic sepsis (37, 39,40). Thus a reduced amount of eIF2B protein may limit translation initiation and, hence, protein synthesis in the gastrocnemius during sepsis (6, 39, 40). In the present set of experiments, the abundance of eIF2B was reduced by ∼40% in muscles from septic rats perfused under basal conditions without insulin present, as assessed by Western blot techniques. Perfusion of muscles from either control or septic rats with insulin present in the buffer did not cause the amount of eIF2B to be increased, as expected from such a short time of exposure to the hormone (data not shown). Therefore, the abundance of eIF2B remained significantly lower in gastrocnemius from septic rats compared with controls after perfusion with buffer containing insulin (P < 0.01; ANOVAF = 14.49, P < 0.0001). The changes in the expression of eIF2Bε were consistent with the alterations in protein synthesis under the different experimental conditions. Thus the failure of insulin to stimulate protein synthesis even though the amount of eIF4E bound to eIF4G is increased may be related to a block at the level of eIF2B.
This work was supported by National Institute of General Medical Sciences Grant GM-39277 (T. C. Vary) and National Institute of Diabetes and Digestive and Kidney Diseases Grant DK-15658 (L. S. Jefferson).
Address for reprint requests and other correspondence: T. C. Vary, Dept. of Cellular and Molecular Physiology, The Pennsylvania State Univ. College of Medicine, Hershey, PA 10733 (E-mail:).
The costs of publication of this article were defrayed in part by the payment of page charges. The article must therefore be hereby marked “advertisement” in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
- Copyright © 2001 the American Physiological Society