Site-specific mTOR phosphorylation promotes mTORC1-mediated signaling and cell growth

Abstract

The mammalian target of rapamycin (mTOR) complex 1 (mTORC1) functions as a rapamycin-sensitive environmental sensor that promotes cellular biosynthetic processes in response to growth factors and nutrients. While diverse physiological stimuli modulate mTORC1 signaling, the direct biochemical mechanisms underlying mTORC1 regulation remain poorly defined. Indeed, while three mTOR phosphorylation sites have been reported, a functional role for site-specific mTOR phosphorylation has not been demonstrated. Here we identify a new site of mTOR phosphorylation (S1261) by tandem mass spectrometry and demonstrate that insulin-phosphatidylinositol 3-kinase signaling promotes mTOR S1261 phosphorylation in both mTORC1 and mTORC2. Here we focus on mTORC1 and show that TSC/Rheb signaling promotes mTOR S1261 phosphorylation in an amino acid-dependent, rapamycin-insensitive, and autophosphorylation-independent manner. Our data reveal a functional role for mTOR S1261 phosphorylation in mTORC1 action, as S1261 phosphorylation promotes mTORC1-mediated substrate phosphorylation (e.g., p70 ribosomal protein S6 kinase 1 [S6K1] and eukaryotic initiation factor 4E binding protein 1) and cell growth to increased cell size. Moreover, Rheb-driven mTOR S2481 autophosphorylation and S6K1 phosphorylation require S1261 phosphorylation. These data provide the first evidence that site-specific mTOR phosphorylation regulates mTORC1 function and suggest a model whereby insulin-stimulated mTOR S1261 phosphorylation promotes mTORC1 autokinase activity, substrate phosphorylation, and cell growth.

FIG. 1.

Identification of S1261 as a novel in vivo mTOR phosphorylation site in intact cells. (A) mTOR undergoes extensive phosphorylation in vivo. HEK293 cells cultured in DMEM-FBS (steady state) were lysed in CHAPS buffer, and mTOR was immunoprecipitated. Immunoprecipitates were either resuspended immediately in sample buffer (IP) or washed in phosphatase buffer and incubated in vitro with various units of λ-phosphatase, resolved on 6% SDS-PAGE, and immunoblotted as indicated (IB). (B) Low-energy collision-induced dissociation spectrum of the doubly charged mTOR pS1261 phosphopeptide. HEK293 cells were transfected with AU1-mTOR, cultured in DMEM-FBS, lysed in CHAPS buffer, and immunoprecipitated with AU1 antibodies. A Coomassie-stained band of AU1-mTOR was digested with trypsin after SDS-PAGE. Liquid chromatography-MS/MS and data analysis were conducted as described in Materials and Methods. Note that the y₆ singly charged ion distinguishes between phosphorylation at S1261 and T1262. (C) Localization of P-S1261 and previously identified P-sites within mTOR's domain structure (4). S1261 maps to a central region of mTOR within HEAT repeat unit 25T, which lies N-terminal to the FAT domain (48), while T2446, S2448, and S2481 lie in the extreme C terminus after the FRB and kinase domains but prior to the FATC domain (7, 43, 49). The tryptic peptide in which P-S1261 was identified is shown. (D) Alignment of mTOR S1261 from various organisms using the algorithm Clustal W. The Caenorhabditis elegans sequence was omitted due to poor alignment resulting from large regions of nonhomology. (E) P-S1261 antibodies are site specific. HEK293 cells were transfected with vector control (V) or with WT or S1261A AU1-mTOR alleles, as indicated, and lysed in CHAPS buffer. WCL were immunoprecipitated with AU1 antibodies and immunoblotted with the indicated antibodies. (F) P-S1261 antibodies are phospho-specific. mTOR immunoprecipitates from HEK293 cells (lysed in CHAPS buffer) were either resuspended immediately in sample buffer (IP) or washed in phosphatase buffer and incubated in the absence or presence of λ-phosphatase (250 units). The immunoprecipitates were immunoblotted as indicated. (G) S1261 is not an autophosphorylation site. HEK293 cells were transfected with vector control, WT (1.0 μg), or KD (1.0 μg) Myc-mTOR plasmids and incubated in the absence or presence of rapamycin (20 ng/ml) for 2 h. Cells were lysed in CHAPS buffer, and WCLs were immunoprecipitated with Myc antibodies and immunoblotted as indicated.

FIG. 2.

Insulin stimulates mTOR S1261 phosphorylation in mTORC1 and mTORC2. (A) Regulation of mTORC1-associated mTOR P-S1261 in 3T3-L1 adipocytes: Adipocytes were serum deprived (24 h), pretreated with rapamycin (R; 20 ng/ml), or wortmannin (W; 100 nM) for 30 min, incubated in the absence or presence of insulin (100 nM) for 30 min, and lysed in CHAPS buffer. WCL were immunoprecipitated with preimmune sera (PI), raptor antibodies (left panel), and mTOR antibodies (right panel) (IP). Immunoprecipitates were immunoblotted as indicated (upper panels), and WCL was immunoblotted to confirm the expected activation and/or inhibition of mTORC1 signaling by the various treatments (lower panels) (IB). Note: the left and right panels represent independent experiments. (B) Regulation of mTORC2-associated mTOR P-S1261 in 3T3-L1 adipocytes. Adipocytes were treated, lysed, and analyzed as described above, except that rictor antibody was used for immunoprecipitation to isolate mTORC2. (C) Regulation of mTOR P-S1261 in HEK293 cells. Cells were transfected with vector control or with Myc-raptor (0.5 μg) and serum deprived (24 h). Cells were pretreated with rapamycin (R; 20 ng/ml), and wortmannin (W; 100 nM), incubated in the absence or presence of insulin (100 nM) for 30 min, and lysed in CHAPS buffer. WCL were immunoprecipitated with Myc antibodies (lanes 1 to 10). Immunoprecipitates (upper panels) and WCL (lower panels) were immunoblotted (IB) as indicated. Note: the left and right panels represent two distinct experiments that were performed identically. On the left, rapamycin partially dissociated mTOR from raptor, while in the experiment shown on the right, rapamycin failed to dissociate mTOR from raptor. While rapamycin dissociated mTOR from raptor in a majority of experiments, a minority failed to show such dissociation, for reasons unclear to us. (D) Amino acids are required to maintain mTOR P-S1261 in mTORC1 in HEK293 cells. HEK293 cells were transfected with Myc-raptor (0.5 μg) and cultured under steady-state conditions. Amino acids were withdrawn via incubation in D-PBS/Glc for 60 min. Myc-raptor was immunoprecipitated, and the level of P-S1261 on Myc-raptor-associated mTOR was assessed by immunoblotting. WCL was immunoblotted as indicated.

FIG. 3.

TSC inhibits and Rheb promotes mTOR P-S1261. (A) TSC1^−/− MEFs exhibit high levels of mTORC1-associated mTOR P-S1261 compared to WT MEFs. Littermate-matched, 3T3 immortalized mouse embryonic fibroblasts derived from TSC1^+/+ or TSC^−/− animals were serum deprived (24 h) and lysed in CHAPS buffer. Triplicate lysates were immunoprecipitated with raptor antibodies and immunoblotted as indicated. Note: as we found that TSC1^−/− fibroblasts express higher levels of total raptor and mTOR protein, two-thirds of the immunoprecipitate from TSC1^−/− cells was loaded relative to that for TSC1^+/+ cells in order to normalize the amounts of raptor and mTOR between the two cell lines. WCL were loaded similarly. (B) Amino acids are required to maintain mTORC1-associated mTOR P-S1261 in TSC1^−/− MEFs. Littermate-matched, TSC1^+/+, and TSC1^−/− MEFs were cultured under steady-state conditions. TSC1^−/− MEFs were also amino acid deprived by incubation in D-PBS/Glc for 60 min. Raptor was immunoprecipitated, and the level of P-S1261 on raptor-associated mTOR was assessed by immunoblotting (IB). WCL were immunoblotted as indicated. (C) Knockdown of Rheb in TSC1^−/− MEFs reduces mTORC1-associated mTOR P-S1261. Parental TSC1^−/− MEFs (Par) were infected with a control scrambled shRNA lentivirus (Scr) or with shRNA lentiviruses targeting two distinct regions of Rheb (Rheb-1 and Rheb-3). Infected cells were selected in puromycin and serum deprived. Raptor was immunoprecipitated, and the level of P-S1261 on raptor-associated mTOR was assessed by immunoblotting. WCL were immunoblotted to confirm Rheb knockdown and the expected decrease in mTORC1 signaling as well as to observe the level of P-S1261 on total mTOR.

FIG. 4.

Phosphorylation of mTOR S1261 promotes mTORC1-mediated phosphorylation and activation of S6K1. (A) P-S1261 promotes S6K1 phosphorylation under steady-state conditions. HEK293 cells were transiently transfected with vector control or cotransfected with HA-S6K1 (0.5 μg) without or with various AU1-mTOR alleles (5 μg) as indicated and cultured in DMEM-FBS. Transfected cells were pretreated without or with rapamycin (20 ng/ml) for 2 h prior to lysis in NP-40-Brij buffer, where indicated. WCL were immunoprecipitated with HA antibodies and immunoblotted as indicated. WCL were also immunoblotted directly. To quantitate the level of S6K1 phosphorylation, the P-S6K1 signal was analyzed by densitometry. The numbers on the uppermost image indicate the densities (percentages) of each P-S6K1 band relative to the mean (set to 100%) of the four RR bands (lanes 6 to 8 and 14). The relative P-S6K1 signals in the graph were normalized to the total amount of HA-S6K1 in each immunoprecipitate (mean ± standard deviation). *, P < 0.005. (B) P-S1261 promotes S6K1 phosphorylation in response to insulin. Results are shown for an experiment similar to that shown in Fig. 3A except that transfected cells were serum deprived (24 h), pretreated without or with rapamycin (20 ng/ml) for 30 min, and then incubated in the absence or presence of insulin (100 nM) for 30 min, as indicated. (C) P-S1261 promotes S6K1 activation in response to insulin. Results are shown for an experiment similar to that shown in Fig. 3B above except that HEK293 cells were cotransfected with 0.2 μg HA-S6K1 together with 5 μg AU1-mTOR alleles. After HA-S6K1 immunoprecipitation, in vitro immune complex kinase (IVK) assays were performed using recombinant GST-S6 as substrate and [³²P]ATP. In vitro kinase reactions were resolved on SDS-PAGE. The portion of the gel containing HA-S6K1 was transferred to a PVDF membrane and immunoblotted with HA antibodies, while the portion of the gel containing GST-S6 was dried and exposed to a phosphorimager screen. The relative amount of radioactive ³²P added to GST-S6 by HA-S6K1 is shown by the numbers on the upper image. The mean activity of HA-S6K1 in response to insulin-stimulated RR-mTOR signaling was set to 100%, and all other activities are relative to this value. The graph in the lower panel shows the combined results of two independent experiments (mean ± standard deviation). The mean activity of HA-S6K1 in response to insulin-stimulated RR-mTOR signaling was set to 100%, and all other activities are relative to this value. *, P < 0.005 by unpaired t test. (D) P-S1261 is required for Rheb-driven S6K1 phosphorylation in the absence of serum growth factors. HEK293 cells were cotransfected with HA-S6K1 (0.5 μg) together with various AU1-mTOR alleles (2.5 μg) and Flag-Rheb (2.5 μg). Cells were serum deprived (24 h), pretreated without or with rapamycin (20 ng/ml) for 1 h, stimulated in the absence or presence of insulin (100 nM) for 30 min, and lysed in NP-40-Brij buffer, as indicated. WCL were immunoprecipitated with HA antibodies and immunoblotted, as indicated. WCL were also immunoblotted directly. (E) In the absence and presence of insulin, the S1261D allele does not behave in a phospho-mimetic manner. The experiment was performed similarly to that in panel B.

FIG. 5.

Phosphorylation of mTOR S1261 promotes mTORC1-mediated 4EBP1 phosphorylation. HEK293 cells were transfected with various AU1-mTOR alleles (4 μg) together with 3HA-4EBP1 (1 μg), deprived of serum (24 h), pretreated without or with rapamycin (20 ng/ml) for 30 min, incubated in the absence or presence of insulin (100 nM) for 30 min, and lysed in NP-40-Brij buffer. WCL were resolved on SDS-PAGE and immunoblotted with the indicated antibodies. SE, short exposure; LE, long exposure. The positions of 3HA-4EBP1 (*) versus endogenous 4EBP1 (**) are indicated.

FIG. 6.

mTOR S1261 phosphorylation is required for mTORC1's in vivo catalytic activity. (A) In mTORC1, S2481 autophosphorylation increases in response to insulin in a PI3K-dependent manner and upon Rheb overexpression. HEK293 cells were cotransfected with Myc-raptor (0.5 μg) together with Flag-Rheb (2.5 μg) and AU1-mTOR (2.5 μg), as indicated. Cells were serum deprived (24 h), pretreated with wortmannin (W; 100 nM), incubated in the absence or presence of insulin (100 nM), and lysed in NP-40-Brij buffer. WCL were immunoprecipitated with Myc antibodies and immunoblotted (IB) as indicated (upper panels). WCL were also immunoblotted (lower panels). (B) In mTORC1, P-S1261 is required for S2481 autophosphorylation. Results are for an experiment similar to that shown in panel A except that additional AU1-mTOR mutant alleles were cotransfected. (C) Phosphorylation of mTOR S1261 does not alter the interaction of mTOR with its TORC1 partners raptor and mLST8/GβL. HEK293 cells were cotransfected with HA-mLST8/GβL (0.5 μg) and HA-raptor (0.5 μg) together with various Myc-mTOR alleles (4 μg), cultured in DMEM-FBS, serum deprived (24 h), incubated in the absence or presence of insulin (100 nM), and lysed in CHAPS buffer. WCL were immunoprecipitated with anti-Myc antibodies and immunoblotted with the indicated antibodies.

FIG. 7.

Phosphorylation of mTOR S1261 promotes mTORC1-mediated cell growth with increased cell size. (A) Cells expressing RR/S1261A-mTOR exhibit impaired rescue of rapamycin-inhibited cell growth. HEK293 cells were cotransfected with GFP-spectrin (1 μg) together with various Myc-mTOR alleles (10 μg). One day posttransfection, the 60-mm plates were split to 10-cm plates and allowed to proliferate for 72 h in the absence or presence of rapamycin. Relative size of subconfluent cells was determined using a flow cytometer and the parameter mean FSC-H. The graph in panel A shows the combined, mean FSC-H values (± standard deviations) of transfected (GFP-positive), G₁-phase cells from three experiments, each performed in quadruplicate (n = 12). The expression levels of the various Myc-mTOR alleles from one of the three experiments are shown (three of the four lysates are shown). The size of cells transfected with WT-mTOR and cultured in the absence of rapamycin was set at 100%. All other samples are shown relative to this value. Statistical significance was determined using a one-way ANOVA followed by Tukey's post hoc tests. The letters (a to d) indicate that the FSC-H means are significantly different at a P level of <0.01. (B) In cycling cells, the phospho-mimetic S1261D allele weakly mimics phosphorylation at S1261. The experiment was performed under steady-state conditions similar to the experiment shown in Fig. 4A except for the use of Myc-tagged mTORs.

FIG. 8.

Model for regulation of mTORC1 signaling by mTOR S1261 phosphorylation. We propose that insulin activates mTORC1 via a series of tightly regulated biochemical steps, allowing mTORC1 to respond appropriately to divergent and dynamic environmental cues. Insulin signals via PI3K and Akt to suppress TSC and thus activate Rheb. Akt directly phosphorylates TSC2 (27) and PRAS40 (54, 67) to cooperatively activate mTORC1. Insulin signaling via PI3K/TSC/Rheb promotes mTOR S1261 phosphorylation (step 1) (see Discussion for more details), which likely cooperates with other inputs to activate the mTORC1 kinase (step 2). Active mTORC1 phosphorylates itself (on S2481) as well as its partners (e.g., raptor [our data and reference 70] and PRAS40 [17, 45, 68]) (step 3), culminating in the phosphorylation of substrates (e.g., S6K1, 4EBP1) (step 4) and cell growth (step 5). Not shown is the recently described AMPK-mediated phosphorylation of raptor upon induction of energy stress (23) and the RSK-mediated phosphorylation of raptor upon mitogenic stimulation of the MAPK pathway (5), events that inhibit and activate mTORC1 signaling, respectively. Taken together, the emerging data suggest that multiple phosphorylation events on mTOR and its partners cooperate to regulate the activation state of mTORC1, both positively and negatively.