Mechanistic Target of Rapamycin Complex 1 (mTORC1)-mediated Phosphorylation Is Governed by Competition between Substrates for Interaction with Raptor*

Background: mTORC1 targets p70S6K1 and 4E-BP1 are phosphorylated in response to nutrient and hormonal signals. Results: Loss of 4E-BP1/2 enhances the rate of p70S6K1 phosphorylation. Conclusion: By competitively interacting with raptor, 4E-BP1 impairs p70S6K1 phosphorylation. Significance: Altered 4E-BP1 expression potentially governs phosphorylation of p70S6K1 under pathophysiological states. In this study, the interaction of mTORC1 with its downstream targets p70S6K1 and 4E-BP1 was evaluated in both mouse liver and mouse embryonic fibroblasts following combined disruption of the genes encoding 4E-BP1 and 4E-BP2. Phosphorylation of p70S6K1 was dramatically elevated in the livers of mice lacking 4E-BP1 and 4E-BP2 following feeding-induced activation of mTORC1. Immunoprecipitation of mTORC1 suggested that elevated phosphorylation was the result of enhanced interaction of p70S6K1 with raptor. These findings were extended to a cell culture system wherein loss of 4E-BP1 and 4E-BP2 resulted in elevated interaction of p70S6K1 with IGF1-induced activation of mTORC1 in conjunction with an enhanced rate of p70S6K1 phosphorylation at Thr-389. Furthermore, cotransfecting HA-p70S6K1 with 4E-BP1, but not 4E-BP1(F114A), reduced recovery of mTORC1 in HA-p70S6K1 immunoprecipitates. Together, these findings support the conclusion that, in the absence of 4E-BP proteins, mTORC1-mediated phosphorylation of p70S6K1 is elevated by a reduction in competition between the two substrates for interaction with raptor.

In this study, the interaction of mTORC1 with its downstream targets p70S6K1 and 4E-BP1 was evaluated in both mouse liver and mouse embryonic fibroblasts following combined disruption of the genes encoding 4E-BP1 and 4E-BP2. Phosphorylation of p70S6K1 was dramatically elevated in the livers of mice lacking 4E-BP1 and 4E-BP2 following feeding-induced activation of mTORC1. Immunoprecipitation of mTORC1 suggested that elevated phosphorylation was the result of enhanced interaction of p70S6K1 with raptor. These findings were extended to a cell culture system wherein loss of 4E-BP1 and 4E-BP2 resulted in elevated interaction of p70S6K1 with IGF1-induced activation of mTORC1 in conjunction with an enhanced rate of p70S6K1 phosphorylation at Thr-389. Furthermore, cotransfecting HA-p70S6K1 with 4E-BP1, but not 4E-BP1(F114A), reduced recovery of mTORC1 in HA-p70S6K1 immunoprecipitates. Together, these findings support the conclusion that, in the absence of 4E-BP proteins, mTORC1-mediated phosphorylation of p70S6K1 is elevated by a reduction in competition between the two substrates for interaction with raptor.
The mechanistic target of rapamycin (mTOR 2 ; previously known as the mammalian target of rapamycin) is an evolutionarily conserved serine/threonine protein kinase that regulates the cell signaling response to nutrient sufficiency and growth factors (1). The polypeptide mTOR functions in two distinct multiprotein complexes: mTORC1 (mTOR complex 1) and mTORC2 (2). Both complexes contain mTOR and mLST8; however, only mTORC1 contains raptor (regulatory-associated protein of mTOR), which serves as a scaffolding protein for recruitment of substrates for phosphorylation by the kinase domain of mTOR. Overexpression of raptor stimulates mTORC1-mediated phosphorylation of its substrates, whereas the inhibitor rapamycin is believed to act by disruption of the mTOR-raptor interaction (3,4). Activation of the mTORC1 signaling pathway in response to insulin or growth factors occurs primarily through the PI3K/Akt signaling pathway (5)(6)(7). The activation of Akt is responsible for phosphorylation of at least two proteins involved in the regulation of mTORC1, TSC2 (tuberous sclerosis complex 2) and PRAS40 (proline-rich Akt substrate of 40 kDa). TSC2 functions in a heterodimeric complex with TSC1 as a GTPase-activating protein toward the small GTPase Rheb (Ras homolog enriched in brain). Association of Rheb in its GTP-bound, but not GDP-bound, state with mTORC1 leads to its activation through an incompletely defined mechanism. Phosphorylation of PRAS40 by Akt results in its dissociation from mTORC1, allowing raptor to recruit the two best characterized downstream targets of mTORC1, p70S6K1 (p70 ribosomal protein S6 kinase 1) and 4E-BP1 (eukaryotic initiation factor 4E-binding protein 1) (8 -10).
Both p70S6K1 and 4E-BP1 contain a common mTORC1 signaling (TOS) motif that plays a critical role in substrate recognition by mediating direct interaction with raptor and consequently phosphorylation by mTORC1 (11,12). The TOS motif contains an essential Phe residue, followed by four alternating acidic and small hydrophobic residues, and is found in the N terminus of p70S6K1 (FDIDL) and C terminus of 4E-BP1 (FEMDI). Mutation of the essential Phe residue impairs the interaction of both 4E-BP1 and p70S6K1 with raptor and thus eliminates the rapamycin-sensitive phosphorylation of the two substrates (11). Intriguingly, the interaction of PRAS40 with raptor (and thus, its inhibitory effect on 4E-BP1 and p70S6K1 phosphorylation) is also dependent on a TOS motif variant (10,13). Overexpression of PRAS40 suppresses phosphorylation of 4E-BP1 and p70S6K1, whereas mutant PRAS40 that is deficient in binding raptor does not affect their phosphorylation (13). Furthermore, overexpression of p70S6K1 impairs mTORC1catalyzed phosphorylation of 4E-BP1 (14); however, this effect is absent when the TOS motif is ablated (11). Thus, direct inter-action of 4E-BP1 and p70S6K1 with raptor functions as a critical step in mediating substrate phosphorylation by mTORC1.
In this study, we demonstrate accelerated phosphorylation of p70S6K1 occurring concomitant with increased interaction of p70S6K1 with mTORC1 in the livers of mice lacking 4E-BP1 and 4E-BP2 compared with control mice. Moreover, in mouse embryonic fibroblasts (MEFs) lacking 4E-BP1/2, the interaction of exogenously expressed HA-p70S6K1 with mTORC1 was elevated, and this effect was repressed by cotransfection with exogenous 4E-BP1, but not with 4E-BP1 in which the essential Phe residue in the TOS motif was altered. Together, these findings demonstrate that competitive substrate binding to raptor governs mTORC1-mediated phosphorylation of p70S6K1 and 4E-BP1.

EXPERIMENTAL PROCEDURES
Materials-Protease inhibitor mixture was purchased from Sigma, and ECL Western blotting detection reagent was from Pierce. Preparation of the anti-4E-BP1 and anti-eIF4E antibodies has been described previously (15,16). Anti-mTOR, anti-p70S6K1, anti-eIF3c, and horseradish peroxidase-conjugated goat anti-rabbit IgG antibodies were purchased from Bethyl Laboratories. Anti-GAPDH antibody was purchased from Santa Cruz Biotechnology, and all other antibodies were purchased from Cell Signaling Technology. Protein content was measured using the DC protein assay (Bio-Rad).
Animals-Eif4ebp1; Eif4ebp2 mutant mice were a kind gift from Dr. Nahum Sonenberg (McGill University). Male mice weighing ϳ30 g were maintained on a 12/12-h light/dark cycle with food (Harlan Teklad) and water provided ad libitum. All procedures involving these mice were approved by the Pennsylvania State University College of Medicine Institutional Animal Care and Use Committee. Mice were fasted for 10 h, followed by a refeeding period of 45 min where indicated.
Processing of Liver Samples-For analysis of the protein phosphorylation state, a portion (ϳ0.3 g) of liver was homogenized in 7 volumes of CHAPS lysis buffer (40 mM HEPES, 0.3% CHAPS, 1 mM EDTA, 50 mM NaF, 40 mM NaCl, 50 mM ␤-glycerophosphate, 10 mM sodium pyrophosphate, 1 mM benzamidine, 200 mM sodium vanadate, and 10 l/ml protease inhibitor mixture) using a Polytron homogenizer. The homogenate was centrifuged at 1000 ϫ g for 3 min at 4°C, and the resulting supernatant fraction was subjected to SDS-PAGE and Western blot analysis as described previously (17). Phosphorylation of p70S6K1, Akt, 4E-BP1, and IRS1 was measured in the supernatant fraction using phosphospecific antibodies as described previously (18).
Cell Culture and Transfections-Cultures of wild-type and Eif4ebp1;Eif4ebp2 double knock-out (DKO) MEFs, a kind gift from Dr. Nahum Sonenberg, and Rps6K1;Rps6k2 DKO MEFs, a kind gift from Dr. Sara Kozma (University of Cincinnati), were maintained in Dulbecco's modified Eagle's medium lacking sodium pyruvate and containing high glucose (Invitrogen) supplemented with 10% fetal bovine serum (Atlas Biologicals) and 1% penicillin/streptomycin (Invitrogen). Transfections were performed using X-tremeGENE HP (Roche Applied Science) with a 3:1 ratio of reagent to DNA (l/g) according to the manufacturer's instructions. Cells were deprived of serum for 3 h and treated with IGF1 (10 ng/ml) as indicated. Cells were harvested in 1ϫ SDS sample buffer for analysis of cell lysate or in CHAPS lysis buffer for immunoprecipitation. Overexpression of HA-S6K1 was achieved using pRK7-HA-S6K1, which was generously provided by Dr. John Blenis (Harvard University). 4E-BP1 was exogenously expressed in cells using the pCMV6 -4E-BP1 plasmid (purchased from OriGene Technologies). Generation of pCMV6 -4E-BP1(F114A) was performed using the QuikChange Lightning site-directed mutagenesis kit (Stratagene) with the following primers: 5Ј-CAGGCGGTGA-AGAGTCACAAGCTGAGATGGACATTTAAGG-3Ј and 5Ј-CCTTAAATGTCCATCTCAGCTTGTGACTCTTCACCG-CCTG-3Ј. Confirmation of mutagenesis was obtained by sequencing.
Immunoprecipitations-Immunoprecipitations were performed by incubating the 1000 ϫ g supernatant fractions of liver homogenates or cell lysates with polyclonal anti-mTOR or monoclonal anti-raptor antibody (Cell Signaling Technology). Supernatant fractions or cell lysates containing ϳ1 mg of protein were incubated with 2 g of antibody for 1 h at 4°C. A 200-g portion of BioMag goat anti-rabbit or anti-mouse IgG beads (Qiagen), previously blocked with CHAPS lysis buffer containing 1% BSA, was added to each sample, and the suspension was rocked at 4°C for 2 h. For anti-HA immunoprecipitations, 10 l of anti-HA-agarose affinity resin (Sigma) was washed twice with CHAPS lysis buffer and blocked for 1 h in CHAPS lysis buffer containing 1% BSA. Cell lysate was then added and incubated for 2 h. For all immunoprecipitations, beads were washed twice with 1 ml of ice-cold CHAPS lysis buffer, suspended in 1ϫ SDS sample buffer, and boiled for 5 min. Supernatants were subjected to Western blot analysis using antibody to mTOR, raptor, p70S6K1, eIF3c, or 4E-BP1, and the results were normalized for the amount of the target protein in the immunoprecipitate.
Statistical Analysis-Data are expressed as means Ϯ S.E. One-way analysis of variance and Student's t test were used to compare differences among groups. p Ͻ 0.05 was considered statistically significant.

Ablation of 4E-BP1 and 4E-BP2 Enhances Phosphorylation of p70S6K1 in Response to Activation of mTORC1-Feeding-in-
duced activation of mTORC1 led to enhanced phosphorylation of p70S6K1 (Fig. 1A) and 4E-BP1 (Fig. 1B) in the livers of wildtype mice. However, in the livers of mice lacking 4E-BP1/2, phosphorylation of p70S6K1 at Thr-389 was elevated by 22-fold compared with a 7-fold increase in wild-type mice when evaluated 45 min after feeding-induced activation of mTORC1 (Fig. 1A). One possible explanation for the enhanced phosphorylation of p70S6K1 at Thr-389 would be up-regulated stimulatory signals to mTORC1 in mice lacking 4E-BP1/2; however, it has been previously reported that signaling to mTORC1 is actually impaired in the muscle, liver, and adipose tissue of 4E-BP1/2 DKO mice compared with wild-type mice due to a p70S6K1-mediated negative feedback loop through the phosphorylation of IRS1 at Ser-636/639 and impaired phosphorylation of Akt at Ser-473 (19). In this study, there was no significant difference in either the basal or feeding-induced phosphoryla-tion of Akt at Ser-473 or of PRAS40 at Ser-246, an Akt target site, in the livers of DKO mice compared with wild-type mice ( Fig. 1, C and D). Taken together, these findings suggest that greater activation of mTORC1 by signaling inputs upstream of Akt is not likely responsible for the feeding-induced elevation in the phosphorylation of p70S6K1 in the livers of DKO mice compared with wild-type mice.
To further evaluate mTORC1 signaling to p70S6K1 in the absence of 4E-BP1 and 4E-BP2, we employed wild-type and 4E-BP1/2 DKO mouse MEFs. IGF1-induced activation of mTORC1 enhanced p70S6K1 phosphorylation at Thr-389 in 4E-BP1/2 DKO MEFs to an extent that was nearly 2-fold greater than the values observed in wild-type MEFs ( Fig. 2A). Similar to the findings in DKO mouse liver, phosphorylation of p70S6K1 did not appear to be the result of a greater signaling input from Akt ( Fig. 2A, compare the third and fourth lanes with the seventh and eighth lanes). Notably, unlike in the liver, where fasting levels of p70S6K1 phosphorylation at Thr-389 were almost undetectable in wild-type and DKO mice (Fig. 1A), p70S6K1 phosphorylation at Thr-389 was elevated in serum-starved DKO MEFs compared with wild-type MEFs ( Fig. 2A). Whereas fasting eliminates signaling inputs to mTORC1 from both growth factors and amino acids, serum deprivation of cell culture medium eliminates only those from growth factors. Thus, the input from amino acids maintained mTORC1 signaling in serum-deprived MEFs at levels sufficient to detect differential signaling to p70S6K1 in wild-type MEFs compared with DKO MEFs. To further evaluate phosphorylation of p70S6K1 in response to IGF1-induced activation of mTORC1, we performed a time course experiment in wild-type and DKO MEFs. At early time points (0 -20 min), phosphorylation of p70S6K1 was elevated in DKO MEFs compared with wild-type MEFs. However, after 30 min, phosphorylation of p70S6K1 did not appear to differ in wild-type and DKO MEFs (Fig. 2, B and C, compare the seventh and eighth lanes; and Fig. 2D). Thus, IGF1induced p70S6K1 phosphorylation at Thr-389 occurred more rapidly in DKO MEFs compared with wild-type MEFs. In both cell types, a similar maximal level of p70S6K1 phosphorylation was achieved, but this occurred more rapidly in the absence of 4E-BP1/2 (i.e. ϳ40 min for wild-type MEFs and ϳ20 min for DKO MEFs) (Fig. 2D).
Because the enhanced feeding-induced phosphorylation of p70S6K1 in the absence of 4E-BP1 did not appear to be due to upstream signal inputs from Akt, we sought to evaluate activation of mTORC1. The interaction of raptor with mTOR is required for the recruitment of substrates under nutrient-rich conditions; however, in the absence of nutrient-induced signaling to mTORC1, a second high-affinity interaction with raptor represses mTORC1 activity (3). In this study, the interaction of raptor with mTOR was evaluated by immunoprecipitating either mTOR or raptor from the supernatant fraction of liver homogenates and assessing the amount of each protein in the immunoprecipitate. As shown in Fig. 3A and supplemental Fig.  1, in the fasted state, the mTOR-raptor interaction was similar in the livers of both wild-type and 4E-BP1/2 DKO mice. Moreover, feeding reduced the mTOR-raptor interaction by ϳ25% in the livers of both wild-type and DKO mice. As a second measure of mTORC1 activation, we also assessed autophosphorylation of mTOR at Ser-2481 in the raptor immunoprecipitate. Autophosphorylation of mTOR at Ser-2481 has been previously shown to function as a biomarker for monitoring mTORC-specific catalytic activity (20). In the livers of both wild-type and DKO mice, feeding increased the autophosphorylation of the mTOR present in the mTORC1 complex by ϳ80% (Fig. 3B). Taken together, these findings imply that the elevation in the feeding-induced phosphorylation of p70S6K1 in the livers of 4E-BP1/2 DKO mice was not the result of enhanced mTORC1 activation but was likely due to the enhanced ability of active mTORC1 to phosphorylate p70S6K1 in the absence of 4E-BP1/2.
p70S6K1 Exhibits Increased Interaction with mTORC1 in the Absence of 4E-BP1/2-Phosphorylation of p70S6K1 and 4E-BP1 by mTORC1 involves recruitment of the substrate protein, mTOR-catalyzed phosphorylation, and release of the phosphorylated substrate. An attractive molecular mechanism for explaining enhanced phosphorylation of p70S6K1 in the

4E-BP1 and p70S6K1 Compete for Interaction with mTORC1
JANUARY 4, 2013 • VOLUME 288 • NUMBER 1 absence of 4E-BP1/2 is a reduction in competition between the two substrates for binding to raptor, as phosphorylation of 4E-BP1 and p70S6K1 by mTORC1 depends on the interaction of raptor with their respective TOS motifs (12). To evaluate the interaction of p70S6K1 with mTORC1, we immunoprecipitated raptor from the supernatant fraction of liver homogenates and accessed the presence of p70S6K1 in the immunoprecipitate. The interaction of p70S6K1 with raptor was enhanced in the livers of both wild-type and DKO mice upon feeding-induced activation of mTORC1 relative to fasted mice (Fig. 3C). However, in the livers of 4E-BP1/2 DKO mice, the magnitude of the increase was 3-fold greater than that observed in wild-type mice (Fig. 3C, compare the first and second lanes with the third   and fourth lanes). We also observed phosphorylation of p70S6K1 at Thr-389 when bound to raptor (Fig. 3D).
To further evaluate the enhanced interaction of p70S6K1 with mTORC1, we performed immunoprecipitations on the supernatant fractions of cell lysates from wild-type and 4E-BP1/2 DKO MEFs. Initially, we sought to evaluate the interaction of endogenous p70S6K1 and raptor. Unfortunately, unlike with the liver samples, the recovery of endogenous p70S6K1 from the raptor immunoprecipitate of cell lysates was poor, and the presence of background bands made analysis difficult. This result is consistent with previous reports of a weak interaction between endogenous p70S6K1 and mTORC1 in cell lysates (21,22). To perform a more thorough analysis, we made use of exogenously overexpressed HA-tagged p70S6K1 and evaluated its interaction with mTORC1 upon serum deprivation and following treatment with IGF1. The immunoprecipitate of HA-p70S6K1 from serum-starved MEFs contained both mTORC1 and eIF3 (Fig. 4A). When inactive, p70S6K1 interacts with eIF3 (23). In a previously reported model, activation of mTORC1 promoted its recruitment to eIF3, where mTORC1 interacted with p70S6K1 to mediate phosphorylation of multiple residues (23). Thus, the presence of mTOR and raptor in the immunoprecipitate of exogenously expressed HA-p70S6K1 potentially reflects both mTORC1 bound directly to HA-p70S6K1 and mTORC1 bound indirectly through eIF3. It seems likely that exposure to serum-free medium for 3 h was not sufficient to produce full dissociation of mTORC1 from eIF3, as there was no change in the interaction of mTORC1 with either HA-eIF3f (supplemental Fig. 2A) or HA-eIF3h (supplemental Fig. 2B) when serum-deprived wild-type or DKO MEFs were treated with IGF1. This interpretation is supported by the finding that hypophosphorylated 4E-BP1 was also present in the HA-p70S6K1 immunoprecipitate from serum-starved MEFs (Fig. 4A), as in the hypophosphorylated state, 4E-BP1 is presumably bound to eIF4E that is associated with eIF3. Alternatively, the presence of 4E-BP1 in the HA-p70S6K1 immunoprecipitate may reflect mTORC1 dimers (24) in which one raptor subunit was bound to HA-p70S6K1 and the other interacted with 4E-BP1. This interpretation is less appealing for two reasons. 1) Co-immunoprecipitation of 4E-BP1 with p70S6K1 was principally observed in serum-deprived cells. 2) Co-immunoprecipita- Wild-type and 4E-BP1/2 DKO MEFs were exposed to serum-free medium (SFM) for 3 h to repress phosphorylation of p70S6K1 at Thr-389. A and B, activation of mTORC1 signaling was achieved by stimulating cells with IGF1 as indicated. The interaction of endogenous p70S6K1 with mTORC1 was evaluated by immunoprecipitating (IP) raptor and measuring the amount of p70S6K1 by Western blot analysis. Autophosphorylation of mTOR at Ser-2481 was also measured in the HA-p70S6K1 immunoprecipitate to assess mTORC1 activation. Expression of mTOR, raptor, eIF3c, HA-p70S6K1, and 4E-BP1 was evaluated in whole cell lysate (WCL) by Western blot analysis. Phosphorylation of p70S6K1 at Thr-389 was assessed by Western blot analysis with a phosphospecific antibody. Quantitation of mTOR (C) and raptor (D) interaction with HA-p70S6K1 was evaluated 5-25 min after IGF1 administration. Values are means Ϯ S.E. (n ϭ 2-4). *, p Ͻ 0.05. Representative blots are shown.
When serum-deprived wild-type or DKO MEFs were treated with IGF1, the amount of eIF3c, mTOR, and raptor recovered in the HA-p70S6K1 immunoprecipitate was reduced (Fig. 4A). As a measure of mTORC1 activation status, we evaluated autophosphorylation of mTOR at Ser-2481 in the HA-p70S6K1 immunoprecipitate (Fig. 4A). Whereas mTOR phosphorylation at Ser-2481 was low in the HA immunoprecipitate relative to the amount of total mTOR under the serum-deprived condition, IGF1 treatment produced an increase in the ratio of phosphorylated to total mTOR in the HA-p70S6K1 immunoprecipitate (Fig. 4A, compare mTOR Ser-2481-P with mTOR (total)). On the basis of these findings and the previously described model (23), we propose that, in the serum-deprived condition, mTORC1 and HA-S6K1 bind independently to eIF3, whereas upon IGF1 treatment, HA-S6K1 interacts directly with raptor to allow for phosphorylation by mTOR (supplemental Fig. 2E).
To evaluate the influence of 4E-BP1/2 on the interaction of HA-p70S6K1 with mTORC1, we treated serum-deprived wildtype and DKO MEFs with IGF1 for 10, 15, or 20 min and evaluated the presence of mTOR and raptor in the HA-p70S6K1 immunoprecipitate (Fig. 4, A and B). Under the serum-deprived condition, the absence of 4E-BP1/2 did not significantly alter the interaction of either protein with HA-p70S6K1 (Fig. 4,  C and D, respectively). This finding suggests that p70S6K1 and 4E-BP1 possess unique binding sites under this condition; HA-p70S6K1 was likely associated directly with eIF3, with 4E-BP1 being bound indirectly to eIF3 via eIF4E. When wildtype cells were treated with IGF1, the amount of mTORC1 in the HA-p70S6K1 immunoprecipitate was reduced; however, in cells lacking 4E-BP1/2, the presence of both mTOR and raptor in the HA-p70S6K1 immunoprecipitate following 15 or 20 min of IGF1 treatment was significantly elevated compared with that observed in wild-type cells (Fig. 4, C and D). This result demonstrates enhanced interaction of HA-p70S6K1 with active mTORC1 in cells lacking 4E-BP1/2 compared with the wild-type condition. We also evaluated the interaction of mTORC1 with exogenously overexpressed HA-tagged 4E-BP1 in S6K1 Ϫ/Ϫ /S6K2 Ϫ/Ϫ MEFs (supplemental Fig. 2C). Whereas the amount of mTORC1 detected in the HA-4E-BP1 immunoprecipitate was not significantly different in the absence of S6K1/2 expression under serum-deprived conditions, the interaction of both mTOR and raptor with HA-4E-BP1 was elevated following IGF1 treatment (supplemental Fig. 2D). Together, these findings support a model wherein the interaction of p70S6K1 and 4E-BP1 with mTORC1 and eIF3 occurs through different mechanisms that can be distinguished by signaling inputs that activate mTORC1 (supplemental Fig. 2E).
Overexpression of 4E-BP1 with a Functional TOS Motif Impairs the Interaction of p70S6K1 with mTORC1-To validate the inhibitory role of 4E-BP1 in the interaction of p70S6K1 with mTORC1, we cotransfected wild-type and DKO MEFs with HA-p70S6K1 and 4E-BP1. Overexpression of 4E-BP1 reduced the phosphorylation of p70S6K1 at Thr-389 (Fig. 5A, WCL panels) in both wild-type and DKO MEFs. Furthermore, cotrans-fection of 4E-BP1 with HA-p70S6K1 repressed recovery of raptor and mTOR in the HA-p70S6K1 immunoprecipitates from both wild-type and DKO MEFs following IGF1 treatment (Fig.  5A, compare the first lane with the second lane and the third lane with the fourth lane). A functional TOS motif is required for 4E-BP1 to interact with raptor, as mutation of a single Phe residue (F114A) has been previously shown to disrupt binding to mTORC1 and impair phosphorylation (22). When 4E-BP1(F114A) was overexpressed in DKO MEFs, IGF1 failed to induce phosphorylation of this mutant to the same extent as wild-type 4E-BP1, as some 4E-BP1(F114A) could still be observed in the hypophosphorylated ␣-isoform (Fig. 5B, WCL  panel). Moreover, unlike wild-type 4E-BP1, overexpression of 4E-BP1(F114A) had no effect on the phosphorylation of p70S6K1 at Thr-389 following stimulation with IGF1, and it did not repress the interaction of HA-p70S6K1 with mTORC1 (Fig.  5B). This result provides direct evidence that 4E-BP1 impairs the interaction of active mTORC1 with p70S6K1 in a manner that is mediated by the TOS motif of 4E-BP1.

DISCUSSION
In this study, we used 4E-BP1/2 DKO mice and MEFs generated from these animals to evaluate the interaction of mTORC1 with its downstream targets p70S6K1 and 4E-BP1. Phosphorylation of p70S6K1 and 4E-BP1 by mTORC1 is facilitated by a common TOS motif (11). Mutation of the TOS domain impairs the ability of both 4E-BP1 and p70S6K1 to interact with raptor (12) and thus eliminates the rapamycin-sensitive phosphorylation of mutant isoforms (11). Furthermore, the integrity of the TOS domain appears to be necessary for p70S6K1 and 4E-BP1 to compete as substrates for mTORC1-mediated phosphorylation, as overexpression of p70S6K1 that contains an altered TOS motif fails to inhibit 4E-BP1 phosphorylation (11). Although 4E-BP1 and p70S6K1 have been shown to compete as substrates for mTORC1-mediated phosphorylation (11,14), it has not been previously demonstrated that this phenomenon is the result of mutually exclusive interaction with the substratebinding domain of raptor upon activation of mTORC1. In the livers of mice lacking 4E-BP1 and 4E-BP2, enhanced phosphorylation of p70S6K1 was observed following feeding in conjunction with increased interaction of p70S6K1 with mTORC1. Moreover, in cells lacking 4E-BP1/2, the interaction of HA-p70S6K1 with mTORC1 was elevated in a manner that could be repressed by cotransfection with exogenous 4E-BP1, but not with 4E-BP1 containing a disruption of the TOS domain. Together, these findings suggest that mTORC1-mediated phosphorylation of p70S6K1 and 4E-BP1 is co-regulated by competition between the two substrates for interaction with mTORC1.
In response to nutrients and environmental signals, mTORC1 and p70S6K1 associate with or dissociate from eIF3 in a coordinated manner (23). When inactive, p70S6K1 is bound to eIF3, whereas mTORC1 only weakly interacts with eIF3 in HEK293E cells after 20 h of serum starvation or following treatment with rapamycin (23). Upon stimulation with insulin, but not with insulin plus rapamycin, mTORC1 is recruited to eIF3, where mTORC1 interacts with p70S6K1 to mediate phosphorylation of multiple residues (23). In support of this model, we observed a feeding-induced enhancement in the interaction of p70S6K1 with mTORC1 in the livers of both wild-type and DKO mice. However, unlike in this previously reported model (23), we were initially surprised to observe a reduction in the interaction of exogenously expressed HA-S6K1 with both mTORC1 and eIF3 in serum-deprived wild-type and DKO MEFs upon stimulation with IGF1. We observed that mTORC1 remained associated with subunits of eIF3 following 3 h of serum deprivation despite a marked reduction in phosphorylation of p70S6K1. Furthermore, following serum deprivation, IGF1 treatment enhanced autophosphorylation of mTOR at Ser-2481 but was unable to increase the interaction of mTORC1 with eIF3 (supplemental Fig. 2). Thus, differences in cell types and length of serum deprivation between this study and the previous report (23) are likely responsible for the observed discrepancies in the observed interaction of mTORC1 with eIF3.
In this study, we observed co-immunoprecipitation of p70S6K1 and mTORC1 in subpopulations containing both phosphorylated and unphosphorylated p70S6K1 and mTORC1. In the livers of fasted mice, p70S6K1 interacted with mTORC1; however, upon feeding, autophosphorylation of mTOR at Ser-2481 was enhanced, and the interaction of p70S6K1 with mTORC1 was elevated. More interestingly, feeding induced phosphorylation of p70S6K1 in the raptor immunoprecipitate (Fig. 3D), an observation that suggests that phosphorylation of p70S6K1 proceeds more rapidly than release of p70S6K1 from mTORC1. Upon mTORC1-mediated phosphorylation, p70S6K1 is released from the eIF3-mTORC1 complex such that p70S6K1 can be phosphorylated at Thr-229 in its activation loop by PDK1 and become fully activated (23). Following dissociation of phosphorylated p70S6K1 from mTORC1, the TOS motif of raptor is free to interact with either a new unphosphorylated p70S6K1 molecule or hypophosphorylated 4E-BP. Wild-type and 4E-BP1/2 DKO MEFs were exposed to serum-free medium for 3 h. Activation of mTORC1 signaling was achieved by treating the cells with IGF1 for 15 min. A, the inhibitory effect of 4E-BP1 on the interaction of HA-p70S6K1 with mTORC1 was evaluated by immunoprecipitating (IP) HA-p70S6K1 from cells cotransfected with wild-type 4E-BP1 and measuring the amount of the corresponding proteins by Western blot analysis. B, to evaluate the role of the TOS motif, HA-S6K1 and 4E-BP1(F114A) were cotransfected into wild-type and 4E-BP1/2 DKO MEFs. Expression of 4E-BP1, mTOR, raptor, eIF3c, and GAPDH was evaluated in whole cell lysate (WCL). Total 4E-BP1 expression was measured by treating cell supernatants with -phosphatase (PPase), followed by Western blot analysis. Representative blots are shown.
Thus, in the absence of 4E-BP1/2, there is a lack of competition for binding to mTORC1 in its activated state, as evidenced by increased recovery of mTOR phosphorylated at Ser-2481 in the HA-p70S6K1 immunoprecipitate from DKO MEFs compared with wild-type MEFs following stimulation with IGF1 (Fig. 4B). A similar result was obtained in vivo, where we observed enhanced interaction of p70S6K1 with mTORC1 in the livers of 4E-BP1/2 DKO mice compared with wild-type mice upon feeding (Fig. 3C). However, unlike the activated state, we observed no difference in the association of p70S6K1 with mTORC1 in the absence of 4E-BP1/2 in the livers of fasted mice or in serumdeprived MEFs. One explanation for these results is that mTORC1, p70S6K1, and the 4E-BP proteins bind independently to eIF3 when mTORC1 is inactive, but upon activation of mTORC1, the conformation of the complex is altered such that mTORC1 can interact directly with either p70S6K1 or the 4E-BP proteins, but not both at the same time (supplemental Fig. 2). Moreover, in such a model, the interaction between mTORC1 and its substrates is mediated primarily through the binding of the substrates to raptor. On the basis of this model, we conclude that enhanced phosphorylation of p70S6K1 in the absence of 4E-BP1 occurs due to a reduction in competition for binding to activated mTORC1.
The findings of this study demonstrate that competition between substrates for binding to raptor plays an important role in mTORC1-mediated phosphorylation of p70S6K1 and 4E-BP1. It is important to point out that increased substrate binding alone has been recently shown to produce limited increases in mTORC1-mediated phosphorylation (25). However, when mTORC1 is activated by upstream signaling inputs, enhanced binding of p70S6K1 is achieved by the elimination of 4E-BP1/2 and substantially increases p70S6K1 phosphorylation. Thus, conditions that alter expression of p70S6K1 or 4E-BP1 potentially modulate signaling through both pathways. We have recently shown that the expression of 4E-BP1 is elevated in the livers of mice with streptozotocin-induced type 1 diabetes (26). Furthermore, a majority of large advanced breast cancers overexpress 4E-BP1 in a manner that facilitates tumor angiogenesis and hypoxia responses (27). In these pathophysiological states, elevated 4E-BP1 expression potentially impairs phosphorylation of p70S6K1 by acting as a competitive substrate for interaction with raptor and thus impairs the insulin signaling cascade downstream of mTORC1.