Rapid Turnover of the mTOR Complex 1 (mTORC1) Repressor REDD1 and Activation of mTORC1 Signaling following Inhibition of Protein Synthesis*

mTORC1 is a complex of proteins that includes the mammalian target of rapamycin (mTOR) and several regulatory proteins. It is activated by a variety of hormones (e.g. insulin) and nutrients (e.g. amino acids) that act to stimulate cell growth and proliferation and repressed by hormones (e.g. glucocorticoids) that act to reduce cell growth. Curiously, mTORC1 signaling is reported to be rapidly (e.g. within 1-2 h) activated by inhibitors of protein synthesis that act on either mRNA translation elongation or gene transcription. However, the basis for the mTORC1 activation has not been satisfactorily delineated. In the present study, mTORC1 signaling was found to be activated in response to inhibition of either the initiation or elongation phases of mRNA translation. Changes in mTORC1 signaling were inversely proportional to alterations in the expression of the mTORC1 repressor, REDD1, but not the expression of TRB3 or TSC2. Moreover the cycloheximide-induced increase in mTORC1 signaling was significantly attenuated in cells lacking REDD1, showing that REDD1 plays an integral role in the response. Finally, the half-life of REDD1 was estimated to be 5 min or less. Overall, the results are consistent with a model in which inhibition of protein synthesis leads to a loss of REDD1 protein because of its rapid degradation, and in part reduced REDD1 expression subsequently leads to de-repression of mTORC1 activity.

The mammalian target of rapamycin (mTOR), 3 a Ser/Thr protein kinase, is an important regulator of cell growth (1). mTOR exists in two distinct signaling complexes referred to as mTOR complex (mTORC)1 and mTORC2 (2). mTORC1 contains G-protein ␤-subunit-like protein (G␤L), the regulatory associated protein of mTOR (raptor), the Ras homolog enriched in brain (Rheb), and proline-rich Akt substrate (PRAS) 40. In contrast, mTORC2 contains G␤L, rapamycin-insensitive companion of mTOR (rictor), and hSIN1. Repression of mTORC1 signaling using the selective inhibitor rapamycin not only leads to a reduction in the size of cells in culture (3) but also prevents cardiac hypertrophy associated with pressure overload (4 -6), resistance exercise-induced skeletal muscle hypertrophy (7)(8)(9), and regrowth of the liver after partial hepatectomy (10). In contrast, constitutive activation of mTORC1 can lead to uncontrolled cell growth and cancer (11). mTORC1 signaling is activated in response to growth-promoting hormones such as insulin (12), IGF-1 (13), or EGF (14). The signaling pathways through which these hormones act to increase mTOR signaling (e.g. the phosphatidylinositide (PI) 3-kinase and extracellular-regulated protein kinase (ERK) pathways) converge on a GTPase activator protein referred to as tuberous sclerosis complex (TSC)2 (Tuberin) (15). TSC2, in a complex with TSC1 (Hamartin), promotes the GTPase activity of the ras homolog enriched in brain (Rheb). Rheb binds directly to mTOR, and when present as a Rheb⅐GTP complex, activates mTOR. Conversely, the binding of Rheb⅐GDP to mTOR is inhibitory. By activating the GTPase activity of Rheb, TSC2 causes a redistribution of Rheb from the stimulatory Rheb⅐GTP complex into the inhibitory GDP-bound form. mTORC1 signaling is also activated by nutrients, particularly amino acids (16). Amino acids may also act through Rheb to activate mTORC1 (17)(18)(19); however, the mechanism through which they do so appears to be unrelated to TSC2 (19,20). In contrast to the activating effect of insulin/IGF-1 and amino acids, catabolic hormones such as glucocorticoids (21)(22)(23) and pro-inflammatory cytokines (24) and conditions that reduce the ATP: AMP ratio (25) repress mTORC1 activity. For example, glucocorticoids act rapidly (i.e. within 4 h) to upregulate the expression of the mTORC1 repressor, regulated in development and DNA damage responses (REDD1) (23). Increased REDD1 expression promotes the assembly of the active TSC1⅐TSC2 complex, leading to decreased mTORC1 signaling.
Reports in the literature suggest that mTORC1 signaling is up-regulated following the inhibition of protein synthesis (26 -31); however, a satisfactory explanation for this observation has not been forthcoming in regard to the regulators described in the preceding paragraph. In most cases (26 -30), inhibitors of the elongation phase of mRNA translation have been used to repress protein synthesis, and one report (31) suggests that accumulation of intracellular amino acids under these conditions might be responsible for the observed activation of mTORC1 signaling. Another possibility is that the activation of mTORC1 is mediated specifically through the inhibition of elongation, perhaps in a manner analogous to the generation of the signaling molecule ppGpp in bacteria (32). To date, there have been no reports to indicate whether or not inhibitors of the initiation phase of mRNA translation might produce a similar activation of mTORC1 signaling. Another condition under which activation of mTORC1 signaling occurs is following inhibition of gene transcription with actinomycin D treatment (30,33). In this case, the activation of mTORC1 signaling occurs prior to detectable inhibition of global rates of protein synthesis, so it is unlikely that the effect is due to an accumulation of intracellular amino acids. A potential explanation for the observed activation of mTORC1 signaling under all of these conditions is that inhibitors of protein synthesis acting either on the elongation or initiation phases of mRNA translation, or gene transcription, cause the rapid loss of a specific protein that acts to repress mTORC1. Such a protein would have to turnover rapidly because activation of mTORC1 signaling is observed within 1-2 h of inhibiting the elongation phase of mRNA translation (30).
In the present study, mTORC1 signaling was shown to be activated when either mRNA translation initiation or elongation was inhibited. Moreover, activation of mTORC1 signaling was proportional to the extent of inhibition of protein synthesis. A search for a potential upstream repressor of mTORC1 whose turnover was rapidly increased upon inhibition of mRNA translation revealed that REDD1, but not TSC2 or TRB3, expression was directly proportional to changes in protein synthesis and inversely proportional to alterations in mTORC1 signaling. In addition, the half-life of REDD1 was estimated to be Ͻ5 min. Together with the observation that induction of mTORC1 signaling by cycloheximide is dramatically attenuated in REDD1 Ϫ/Ϫ mouse embryo fibroblasts (MEFs), the results presented herein are consistent with a model in which inhibitors of protein synthesis acting on either the initiation or elongation phase of mRNA translation, or gene transcription, cause a rapid reduction in REDD1 protein expression, resulting in activation of mTORC1 activity.

Cell
Culture-Wild type (GCN2 ϩ/ϩ ) and GCN2 Ϫ/Ϫ MEFs (kindly provided by Drs. David Ron and Heather Harding, NY University School of Medicine) were maintained at 37°C in high glucose Dulbecco's modified Eagle's medium (Invitrogen) containing 10% fetal bovine serum (Atlas Biologicals) and 1% penicillin-streptomycin (Invitrogen). Wild type (REDD1 ϩ/ϩ ) and REDD1 Ϫ/Ϫ MEFs (kindly provided by Dr. Leif Ellisen, Harvard Medical School) were similarly maintained. In all experiments, cells were ϳ50 -70% confluent. On the day of the study, cells were randomly divided into five groups. One group of cells (control) was placed in serum-free medium. Two other groups of cells were placed in serum-free medium containing either 1 M cycloheximide (A. G. Scientific, Inc.) or 5 mM histidinol (Sigma-Aldrich). The final two groups of cells were placed in serum-free medium lacking His, and in one of those groups, the medium also contained 5 mM histidinol. Cells were returned to the incubator for 2 h prior to harvest unless otherwise indicated.
Measurement of Global Rates of Protein Synthesis-Global rates of protein synthesis were estimated by the incorporation of [ 35 S]methionine and [ 35 S]cysteine into protein as described previously (34). Western Blot Analysis-Cells were harvested by scraping in SDS sample buffer (0.0625 M Tris-HCl, pH 6.8; 12.5% v/v glycerol; 1.25% SDS; 1.25% v/v ␤-mercaptoethanol; 0.1% bromphe-nol blue), boiled for 5 min, and equal volumes of each sample were subjected to Western blot analysis as described previously (35). Briefly, samples were resolved by SDS-polyacrylamide gel electrophoresis (SDS-PAGE), transferred onto a 0.45-m polyvinylidene fluoride membrane (Pall Life Sciences), and the membrane was blocked with 5% nonfat dry milk. Membranes were then incubated overnight at 4°C with one of the following antibodies: anti-phospho-Ser 51 (BIOSOURCE International, Invitrogen) or total eIF2␣, anti-4E-BP1 (Bethyl Laboratory), anti-S6K1 (Bethyl), anti-TSC2 (Cell Signaling) anti-TRB3 (EMD Chemicals, Inc.), or anti-REDD1 (ProteinTech Group Inc.). Membranes were then incubated in secondary antibody (Bethyl) at room temperature for one hour and blots were developed with enhanced chemiluminescence (ECL) or ECL Plus reagents (Amersham Biosciences). Images were captured using a GeneGnome HR Bioimager (SynGene) and quantitated using GeneTools software (SynGene).
Sucrose Density Gradients-For polysome analysis, cells were scraped in buffer (50 mM HEPES, pH 7.4; 75 mM KCl; 5 mM MgCl 2 ; 250 mM sucrose; 1:10 volume 10% Triton X-100, and 13% sodium deoxycholate; 0.1 mg/ml cycloheximide), and the suspension was rocked for 10 min at 4°C to lyse the cells. The homogenate was centrifuged at 3,000 ϫ g for 15 min at 4°C centrifugation, and the supernatant was subjected to sucrose density gradient centrifugation as described previously (36). Briefly, discontinuous 9-step 20% (10 mM HEPES, pH 7.4; 250 mM KCl; 5 mM MgCl 2 ; 0.5 mM EDTA; 20% w/w sucrose) to 47% (10 mM HEPES, pH 7.4; 250 mM KCl; 5 mM MgCl 2 ; 0.5 mM EDTA; 47% w/w sucrose) sucrose density gradients were formed as previously described (36), except that the KCl concentration was changed to 75 mM. An equal amount of protein was loaded onto each gradient and centrifuged at 288,200 ϫ g at 4°C for 1 h 50 min. Gradients were fractionated using an ISCO gradient pump (Teledyne Isco, Inc.) while the absorbance at 254 nm was continuously recorded. The ratio of ribosomes present in the nonpolysomal fraction to those present in polysomes was calculated from the area under the curve for each fraction.
Measurement of Intracellular Amino Acid Content-For measurement of intracellular amino acid content, cells were washed twice in ice-cold phosphate-buffered saline, harvested in 3% perchloric acid, and centrifuged at 1,000 ϫ g for 3 min. The supernatant was neutralized with 1 M K 2 HPO 4 , mixed, and centrifuged at 1,000 ϫ g for 3 min. The supernatant was subjected to amino acid analysis by HPLC (37) in the General Clinical Research Center at The Pennsylvania State University College of Medicine.
Statistical Analysis-Statistical analyses were performed by Student's t test or one-way analysis of variance with the corresponding post-test using the InStat software program (GraphPad) as noted in the figure legends.

RESULTS
In the present study, wild type MEFs and MEFs lacking GCN2 (GCN2 Ϫ/Ϫ ) were employed as experimental model systems to distinguish mRNA translation initiationversus elongation-induced changes in mTORC1 signaling following selective inhibition of either process. Changes in mTORC1 signaling that occurred in response to inhibition of translation elongation by cycloheximide in MEFs were compared with changes associated with inhibition of initiation in cells deprived of His or in cells treated with the histidinyl-tRNA synthetase inhibitor, histidinol, to induce His-tRNA deacylation (38). Deprivation of essential amino acids such as His leads to activation of the protein kinase GCN2 by deacylated tRNA (39). GCN2 subsequently phosphorylates the ␣-subunit of eIF2 (eIF2␣) on Ser 51 , resulting in inhibition of initiation relative to elongation. Initially, the effectiveness of cycloheximide and His deprivation in inhibiting protein synthesis was assessed by measuring changes in [ 35 S]methionine and [ 35 S]cysteine incorporation into protein. As shown in Fig. 1A, global rates of protein synthesis were reduced to ϳ15% of the control value in wild type MEFs treated with cycloheximide and to about 30% of the control value in wild type MEFs deprived of His. Addition of histidinol to His-containing medium was significantly less effective in repressing protein synthesis compared with His deprivation. However, the combination of His deprivation and histidinol treatment reduced pro- tein synthesis to ϳ10% of the control value. In GCN2 Ϫ/Ϫ MEFs, the inhibition of protein synthesis caused by cycloheximide was similar in magnitude to the decrease observed in wild type MEFs (Fig. 1B). As reported previously using perfused livers from GCN2 Ϫ/Ϫ mice (40), the inhibition of protein synthesis in GCN2 Ϫ/Ϫ MEFs caused by His deprivation and/or histidinol treatment was also similar in magnitude to wildtype MEFs. To confirm that the GCN2 Ϫ/Ϫ MEFs lacked functional GCN2, the phosphorylation state of eIF2␣ was measured by Western blot analysis using an antibody that specifically recognizes the protein when Ser 51 is phosphorylated. As shown in Fig. 1C, His deprivation, treatment with histidinol, and His deprivation combined with histidinol treatment all led to a significant increase in eIF2␣ phosphorylation on Ser 51 compared with wild type control MEFs or wild type MEFs treated with cycloheximide. In contrast, eIF2␣ phosphorylation on Ser 51 was not increased in GCN2 Ϫ/Ϫ MEFs in response to any of the treatments.
A possible explanation for the decrease in protein synthesis that occurred in GCN2 Ϫ/Ϫ MEFs during His deprivation and/or histidinol treatment is that in the absence of GCN2, accumulation of deacylated tRNA resulting from deprivation of an essential amino acid leads to an inhibition of translation elongation rather than initiation. To assess this possibility, polysome aggregation was measured in wildtype and GCN2 Ϫ/Ϫ MEFs using sucrose density gradient centrifugation. In this type of analysis, inhibition of initiation relative to elongation leads to disaggregation of polysomes with an increase in the proportion of ribosomes present as 80 S monomers (41). In contrast, inhibition of elongation relative to initiation has the opposite effect, i.e. ribosomes accumulate in polysomes. As expected, cycloheximide treatment resulted in a decrease in the number of ribosomes present as 80 S monomers concomitant with accumulation of ribosomes in polysomes in both wild type and GCN2 Ϫ/Ϫ MEFs (Fig.  2), indicative of an inhibition of elongation. In wildtype MEFs, His deprivation, histidinol treatment, or a combination of His deprivation and histidinol treatment all caused a decrease in the number of ribosomes present in the polysomal fraction of the gradient with a corresponding increase in ribosomes present as 80 S monomers, a result consistent with inhibition of initiation. In contrast, in GCN2 Ϫ/Ϫ MEFs, His deprivation and/or histidinol treatment resulted in an inhibition of elongation as indicated by a shift of ribosomes from 80 S monomers into polysomes.
As an index of mTORC1 signaling, changes in phosphorylation of two proteins that are direct substrates of mTORC1, 4E-BP1, and S6K1 (42), were measured as altered migration during SDS-polyacrylamide gel electrophoresis. The electrophoretic migration of both proteins is inversely proportional to  FEBRUARY 8, 2008 • VOLUME 283 • NUMBER 6 their phosphorylation state, with the most highly phosphorylated forms of the proteins exhibiting the slowest migration during electrophoresis. As shown in Fig. 3, cycloheximide-and His deprivation-induced changes in phosphorylation of both proteins were inversely proportional to alterations in protein synthesis. Thus, all four treatments led to increased phosphorylation of both 4E-BP1 and S6K1. However, the magnitude of the increase was greater in cells treated with cycloheximide or treated with histidinol in medium lacking His compared with cells either deprived of His or treated with histidinol alone.

Protein Synthesis Inhibition Leads to Rapid REDD1 Turnover
Moreover, the pattern of response in GCN2 Ϫ/Ϫ MEFs (Fig. 3, B and D) was identical to that in wild type cells (Fig. 3, A and C). Interestingly, phosphorylation of both 4E-BP1 and S6K1 in untreated wild type MEFs tended to be higher than in untreated GCN2 Ϫ/Ϫ MEFs, a point discussed further below.
A previous report (26) suggested that activation of mTORC1 signaling that occurred in response to inhibition of translation elongation by agents such as cycloheximide was due to accumulation of intracellular concentrations of amino acids. To determine whether or not such a change would explain the activation of mTORC1 signaling observed in the present study, the intracellular content of His and the branchedchain amino acids was measured in control cells and cells deprived of histidine. The content of the branched-chain amino acids was measured because in many cells they are as potent as a complete mixture of amino acids in activating mTORC1 (43). It was found that the intracellular His content was reduced to 43% of the control value in His-deprived wild type MEFs and the intracellular contents of Leu, Ile, and Val in His-deprived wild type MEFs was 105, 93, and 91% of the control values, respectively (n ϭ 3 experiments; within each experiment three dishes of cells were independently analyzed). Similarly, in His-deprived GCN2 Ϫ/Ϫ MEFs, the intracellular contents of Leu, Ile, and Val were 116, 114, and 111% of the control value, respectively. Thus, an increase in the intracellular branched-chain amino acids is unlikely to explain activation of mTORC1 signaling associated with His deprivation in MEFs. An alternative explanation for the activation of mTORC1 signaling observed in the present study is that inhibition of protein synthesis resulted in increased turnover of a mTORC1 repressor. TSC2 is arguably the best known mTORC1 repressor. However, as shown in Fig. 4, A and B, TSC2 expression was unchanged in either wild type or GCN2 Ϫ/Ϫ MEFs, respectively, in response to cycloheximide treatment or His deprivation with or without histidinol treatment. TRB3 binds to and inhibits Akt, an upstream activator of mTORC1 signaling (44,45), and thus acts as a mTORC1 repressor. However, similar to TSC2, no consistent change in TRB3 expression was observed in either wild type (Fig. 4C) or GCN2 Ϫ/Ϫ MEFs (Fig. 4D) under any of the conditions used in the present study. REDD1 is induced in response to a variety of cell stresses (46) and acts upstream of TSC2 to repress mTORC1 (47,48). In the present study, the pattern of REDD1 expression in both wild type (Fig.  4E) and GCN2 Ϫ/Ϫ (Fig. 4F) MEFs mirrored exactly the changes in protein synthesis in response to cycloheximide treatment and His deprivation with or without histidinol treatment. Thus, cycloheximide treatment and a combination of His deprivation and histidinol treatment caused a larger reduction in REDD1 expression compared with either histidinol treatment or His deprivation alone.
To determine whether or not decreased REDD1 expression is critical for the activation of mTORC1 signaling under the conditions used in the present study, we considered exogenous expression of the protein prior to cycloheximide treatment or His deprivation. However, the success of such an approach would require that expression of REDD1 be maintained during the treatment period. Therefore, prior to initiating expression studies, an estimate of REDD1 protein half-life was obtained. In these studies, cycloheximide was added to wild type or GCN2 Ϫ/Ϫ MEFs, and the cells were harvested at various times and analyzed for REDD1 expression and mTORC1 signaling. As shown in Fig. 5, A and B, REDD1 expression was dramatically reduced within 5 min of cycloheximide addition to either wild type or GCN2 Ϫ/Ϫ MEFs, respectively, suggesting that the half-life of the protein was less than 2 min. The activation of mTORC1 signaling, as assessed by phosphorylation of 4E-BP1 (Fig. 5, C and D) and S6K1 (Fig. 5, E and F), was slightly delayed compared with the decrease in REDD1 expression, but was nonetheless rapidly induced after cycloheximide addition. Interestingly, the activation of mTORC1 signaling was delayed to a greater extent in GCN2 Ϫ/Ϫ MEFs compared with wild type cells, a finding in agreement with the greater REDD1 expression observed in GCN2 Ϫ/Ϫ MEFs compared with wild type cells.
To assess whether the observed changes in REDD1 content were associated with alterations in REDD1 mRNA, RNA was isolated from wild type and GCN2 Ϫ/Ϫ MEFs at various times after addition of cycloheximide to the cell culture medium. As shown in Fig. 6A, REDD1 mRNA content did not change during a one hour incubation with cycloheximide, a time at which mTORC1 signaling was maximally increased. Moreover, REDD1 mRNA content did not decrease in GCN2 Ϫ/Ϫ MEFs incubated for up to 1 h with cycloheximide, but instead was significantly increased (Fig. 6B). Interestingly, REDD1 mRNA expression was greater in GCN2 Ϫ/Ϫ MEFs compared with wild type cells, a finding that may explain the higher REDD1 protein expression in those cells (Fig. 5, A and  B). Overall, the rapid fall in REDD1 protein content is unlikely to be due to changes in REDD1 mRNA content.
Based on its apparently short half-life, we anticipated that REDD1 would need to be expressed at very high levels relative to the native protein to effectively repress mTORC1 signaling during inhibition of protein synthesis. Indeed, exogenously expressed REDD1 was rapidly degraded in MEFs (data not shown). Therefore, an alternative approach to assessing the role of changes in REDD1 expression in increasing TORC1 signaling under conditions of repressed mRNA translation was utilized. In this approach, wild type or REDD1 Ϫ/Ϫ MEFs were treated with cycloheximide, and the effect on mTORC1 signaling was assessed as changes in phosphorylation of 4E-BP1 and S6K1. The time point chosen for these studies was based on the finding that cycloheximide-induced changes in REDD1 expression are maximal by 15 min (Fig. 5, A and B). As shown in Fig. 7, A and C, signaling through mTORC1 was significantly increased within 15 min of exposure to cycloheximide in wild type MEFs, concomitant with decreased REDD1 expression (Fig. 7E). However, the increase in mTORC1 signaling induced by cycloheximide was dramatically reduced in MEFs lacking REDD1 compared with wild type MEFs. For example, the relative increase in 4E-BP1 phosphorylation in REDD1 Ϫ/Ϫ MEFs 15 min after addition of cycloheximide was only 26% of the increase in wild type MEFs. As one would predict, in REDD1 Ϫ/Ϫ MEFs, basal mTORC1 signaling was elevated compared with wild type MEFs, as assessed by a small, but significant, increase in phosphorylation of both 4E-BP1 and S6K1.

DISCUSSION
The results of the present study confirm earlier reports showing that inhibition of the elongation phase of mRNA translation results in activation of mTORC1 signaling (26 -31). The results extend the earlier reports to show that inhibition of the initiation phase of mRNA translation similarly results in activation of mTORC1. Thus, inhibition of mRNA translation per se, not decreased elongation, leads to activation of mTORC1 signaling. How does inhibition of translation activate mTORC1? Previously, it was suggested that the activation of mTORC1 signaling was due to accumulation of intracellular amino acids that could occur as a result of their reduced utilization for protein synthesis in conjunction with unchanged cellular uptake and appearance from protein degradation (26). The results of the present study showing that mTORC1 signaling was activated in MEFs deprived of His without a significant change in the intracellular content of any of the branched-chain amino acids suggests that activation of mTORC1 can occur without a detectable increase in intracellular amino acids.
An alternative explanation for the activation of mTORC1 signaling that occurs upon inhibition of mRNA translation is that synthesis of a short-lived protein that acts as a repressor of mTORC1 is reduced. The most proximal repressor of mTORC1 that has been identified thus far is TSC2. In cells in which TSC2 expression is repressed, the proportion of Rheb present in the stimulatory GTP complex is constitutively elevated, and mTORC1 signaling is activated (49,50). Moreover, in such cells, mTORC1 signaling is resistant to amino acid deprivation (18,51). However, in the present study, TSC2 expression in MEFs was unaffected by either cycloheximide treatment or His deprivation with or without treatment with histidinol. Another protein that acts to repress mTORC1 signaling is TRB3. TRB3 binds to and prevents activation of an upstream modulator of mTORC1 function, Akt (44). By inhibiting Akt, TRB3 represses insulin-induced activation of mTORC1 (45). Interestingly, the half-life of TRB3 is estimated to be less than 2 h (52), making it a potential candidate for mTORC1 regulation under conditions of reduced protein synthesis. However, similar to TSC2, there was no detectable change in TRB3 expression in the present study.
REDD1 was originally identified as a gene that is transcriptionally up-regulated in response to hypoxia (53)(54)(55)(56), and subsequently shown to be regulated by other cellular stresses such as exposure to arsenite (57), dexa- methasone (23,58), or agents that cause DNA damage (59). More recent studies have shown that REDD1 acts downstream of Akt (47,54) and upstream of TSC2 (48,53,60) to repress mTORC1. In the present study, REDD1 expression was dramatically decreased in both wild type and GCN2 Ϫ/Ϫ MEFs treated with cycloheximide as well as in response to His deprivation. Thus, mTORC1 signaling was indirectly proportional to REDD1 content (Fig. 8), strongly suggesting that decreased REDD1 expression was causative in the observed activation of mTORC1.
It is noteworthy that both REDD1 protein (Fig. 4, E and F) and mRNA (Fig. 6, A and B) were more highly expressed in GCN2 Ϫ/Ϫ compared with wild type MEFs. The basis for the increased REDD1 expression is unknown, but the finding that basal phosphorylation of both 4E-BP1 and S6K1 was lower in GCN2 Ϫ/Ϫ compared with wild type MEFs provides further support for the idea that up-regulated REDD1 expression leads to further repression of mTORC1 signaling.
Although not examined herein, it is likely that reduced REDD1 expression also explains the observation made in previous studies that inhibition of gene transcription results in mTORC1 activation. For example, mTORC1 signaling is significantly activated within 1-2 h of exposure to actinomycin D (30,33). The half-life of the REDD1 mRNA is relatively short, i.e. a few hours (58). Thus, inhibition of gene transcription would lead to a rapid reduction in REDD1 mRNA expression and subsequently to decreased REDD1 protein, thereby activating mTORC1 signaling.
An obvious question is why has not the link between inhibition of protein synthesis and activation of mTORC1 signaling been more widely observed? For example, in L6 myoblasts, His deprivation had no obvious effect on mTORC1 signaling (61). The answer to this question is unknown, but may in part involve the magnitude of the inhibition of protein synthesis engendered by activation of GCN2 in different cell types. For example, in many mammalian cells and tissues, activation of GCN2 in response to amino acid deprivation reduces protein synthesis to only about 70% of the control value (e.g. 62). In contrast, at the concentration used in the present and past studies, cycloheximide leads to a reduction in global rates of protein synthesis to ϳ15% of the control value. Thus, the relatively small changes in protein synthesis that occur in response to activation of GCN2 in some cells may allow REDD1 mRNA translation to proceed at a rate sufficient to sustain expression of the protein at levels necessary to maintain mTORC1 regulation. An alternative explanation for the dearth of studies reporting activation of mTORC1 signaling in response to amino acid deprivation is that in many cases cells were deprived of Leu, either alone or in combination with other amino acids. Leucine is thought to activate mTORC1 through a signaling pathway independent of TSC2 (19,20). In the present study, we found that Leu deprivation resulted in reduced mTORC1 signaling in MEFs, even though REDD1 expression was lower in Leu-deprived compared with control cells (data not shown). Thus, the negative influence of Leu deprivation on mTORC1 signaling appears to be dominant to the positive input provided by reduced REDD1. Therefore, it is likely that in many studies, the negative regulation of mTORC1 caused by Leu deprivation masks the up-regulation of mTORC1 signaling that would otherwise be induced by reduced REDD1 expression.
Acute inhibition of mRNA translation likely leads to decreased expression of a number of proteins that have short half-lives (e.g. a few minutes). Thus, although the changes in mTORC1 signaling that manifest in cells in which mRNA translation is inhibited are inversely proportional to changes in REDD1 expression (Fig. 8), such a correlation does not prove that REDD1 is directly involved in the effect. Therefore, in the present study, cycloheximide was used to inhibit mRNA translation in MEFs lacking REDD1. In those cells, the increase in mTORC1 signaling was significantly reduced compared with control MEFs, suggesting that, in large part, the increase in mTORC1 signaling was due to changes in REDD1 expression. However, the effect was not completely absent in REDD1 Ϫ/Ϫ MEFs, and therefore another protein(s) is likely to be involved. A likely candidate for the unknown regulatory protein is REDD2 (RTP801l). Like REDD1, REDD2 represses signaling through mTORC1 (47). Moreover, similar to REDD1, REDD2 mRNA expression is induced by hypoxia (54,63). Unfortunately, we are not aware of a source of anti-REDD2 antibodies, and examination of REDD2 mRNA content is unlikely to be informative, based on the assumption that, similar to REDD1 expression, rapid changes in REDD2 protein expression would probably be a result of repressed REDD2 protein expression rather than to changes in REDD2 mRNA content. In summary, the results of the present study reveal the novel finding that inhibitors of protein synthesis acting either on translation initiation or elongation cause a rapid increase in REDD1 turnover, leading to activated signaling through mTORC1. Moreover, the finding that mTORC1 signaling is inversely proportional to changes in REDD1 expression suggest that even small changes in REDD1 expression will have a significant effect on mTORC1 signaling. Finally, the results reveal that REDD1 has an exceptionally short half-life, making it a possible target for regulation under a variety of conditions that alter mTORC1 signaling. Future studies will be required to identify the other protein(s) that are involved in the increase in mTORC1 signaling that occur in response to inhibition of mRNA translation.